Perspectives on Modern Orthopaedics

Thoracoscopic Anterior Instrumentation

and Fusion for Idiopathic Scoliosis
Daniel J. Sucato, MD, MS

Abstract
Thoracoscopically assisted surgery is a new approach to access the anterior spine to ning so that the appropriate thoracic
perform biopsies, anterior releases, diskectomies, and anterior instrumentation and curve correction is achieved, especial-
fusion for idiopathic thoracic scoliosis. This approach compromises the chest wall ly in the setting of a so-called selec-
less than an open thoracotomy does because it uses several small portal incisions. tive thoracic fusion in the IC curve
It has been suggested that this approach allows fusion of fewer motion segments type. The ideal patient for thoraco-
and better correction of curvature than does posterior spinal fusion and instrumen- scopic anterior instrumentation and
tation. The technique, which is still evolving, is technically demanding, requiring fusion is one who has a relatively small
advanced training and special instrumentation and anesthesia techniques. curve size (50° to 65°) of relative flex-
J Am Acad Orthop Surg 2003;11:221-227 ibility (>50% flexibility index); is thin
(40 to 60 kg), which makes placement
and utilization of the portals easier;
and is tall, because the sizable chest
Surgical treatment for idiopathic scolio- In the last decade, indications have provides a greater working space and
sis has changed rapidly in the last 20 increased for endoscopic approach- larger vertebral bodies for easier in-
years. Posterior segmental spinal in- es to thoracic spine surgery. Endos- sertion of screws. For surgeons with
strumentation was an advance over copy was first used for biopsy and disk- experience in the technique, the in-
Harrington instrumentation because ectomy as well as for anterior release dications can include stiffer curves of
it improved correction in the sagittal and fusion, in combination with pos- up to 75°. The primary contraindica-
and coronal planes.1,2 The single solid terior spinal fusion and instrumenta- tion for the procedure is poor pulmo-
rod used with anterior surgery was tion, to treat severe curves or when nary function, which limits the pa-
an improvement over the Dwyer ca- there was risk for the development of tient’s ability to tolerate single-lung
ble, especially for thoracolumbar and the crankshaft phenomenon.5-7 The en- ventilation. All patients should have
lumbar curves, because it allowed sur- doscopic approach also has been used preoperative pulmonary function tests
geons to use a rotational maneuver to perform an anterior instrumen- to assess their ability to tolerate the
to correct both the sagittal and coro- tation, correction, and fusion. Early
nal deformities.1,2 Recently, the single results are encouraging, but the tech-
solid rod placed through an open tho- nique requires further study and im-
racotomy has been used to correct tho- provement. Dr. Sucato is Assistant Professor, Department of
racic curves.3,4 Anterior correction of Orthopaedic Surgery, University of Texas at
thoracic scoliosis offers the theoretic Southwestern, and Staff Orthopaedist, Texas Scot-
tish Rite Hospital, Dallas, TX.
advantage of better coronal correction Patient Selection and
because it permits the surgeon to per- Preoperative Planning The author or the department or departments with
form diskectomies, provides improve- which he is affiliated has received something of val-
ment in the thoracic hypokyphosis seen The indications for anterior instrumen- ue from a commercial or other party related di-
in idiopathic scoliosis, and saves mo- tation and fusion include single tho- rectly or indirectly to the subject of this article.
tion segments. In a prospective study, racic curves or thoracic curves with
Reprint requests: Dr. Sucato, 2222 Welborn Street,
Betz et al4 demonstrated that anterior a compensatory lumbar and/or up- Dallas, TX 75219.
surgery improved sagittal plane align- per thoracic curve, that is, type IA, IB,
ment while saving an average of 2.5 or IC curves using the Lenke classi- Copyright 2003 by the American Academy of
Orthopaedic Surgeons.
distal motion segments compared with fication.8 It is important to determine
posterior surgery. the curve type for preoperative plan-

Vol 11, No 4, July/August 2003 221

Thoracoscopic Anterior Instrumentation and Fusion for Idiopathic Scoliosis
procedure and to help predict the post-
operative course. Pulmonary function
test findings below 60% of predicted
results are a relative contraindication
to anterior thoracic surgery.
Preoperative assessment of the pa-
tient should include a physical exam-
ination to confirm radiographic find-
ings that the upper thoracic and lumbar
curves are compensatory without any
structural characteristics. Imaging
should include standing lateral and
posteroanterior and supine bending
right and left radiographs. The lateral
radiograph should be used to ensure
that excessive kyphosis (>40°) is not
present. This is a contraindication for
anterior correction because, when com-
pression is used, anterior correction
can increase kyphosis.9 Fusion levels
for the thoracic curve are determined
on the posteroanterior radiograph, us-
ing the superior and inferior end ver-
tebrae of the Cobb measurement as
the upper and lower end instrumented
vertebrae. In a smaller patient or one
who has marked tilt of the upper end
instrumented vertebrae, a level supe-
rior to the end vertebra may be cho-
sen to provide greater fixation because
of the risk of cutout of the superior
screws. Analysis of the lower end ver-
tebra may reveal that the disk prox-
imal to it is in fact neutral. If so, the
more proximal level may be chosen
as the lower end instrumented ver-
tebra. Supine bending radiographs are
important to confirm that the lumbar
and upper thoracic curves are truly
compensatory (bend to <25°). Bend-
ing radiographs are used to determine
the flexibility of the thoracic curve so
that a coronal bend may be placed in
the rod if the curve is stiff.
Anesthesia Considerations
Maintaining a proper airway during
anesthesia is critical to the success of
thoracoscopic surgery. To perform an-
terior instrumentation and fusion, the
lung on the convexity of the curve
must be deflated, and single-lung ven-
222
tilation techniques are used. This is
typically accomplished with a double-
lumen endotracheal tube, which has
a bronchial lumen that sits in the de-
pendent mainstem bronchus and a tra-
cheal lumen that lies just proximal to
the carina (Fig. 1). The dependent lung
is ventilated through the bronchial lu-
men, while the lung on the convex-
ity of the curve becomes deflated when
the tracheal lumen is occluded. It is
important to recheck tube placement
after the patient is in the lateral de-
cubitus position because, in up to 80%
of cases, the tube tends to move dis-
tally.10
Patients undergoing single-lung
ventilation are subjected to significant
stresses from the right-to-left shunt
through the dependent lung and from
that lung’s decreased functional ca-
pacity, the result of increased intra-
abdominal pressure and compression
from the weight of the mediastinal
structures. The high pressures that re-
sult can lead to airway leaks or trau-
ma, which can cause pneumothorax.
The so-called down lung syndrome,
seen most frequently with lengthy
surgeries, is characterized by absorp-
tion atelectasis, accumulation of se-
cretions, and formation of transudate
in the dependent lung. The anesthe-
siologist needs to be skilled in the
technique to minimize the chance of
anesthetic complications.11,12
Patient Positioning and
Operating Room Setup
The patient is positioned in the lat-
eral decubitus position on a radio-
lucent operating table with the con-
vexity of the curve up (Fig. 2). An
absolutely lateral position is critical,
especially during screw placement,
and should be checked periodically
to ensure that it is maintained through-
out the procedure. The patient may
be secured using an inflatable radio-
lucent beanbag or other positioning
system. Whatever method is used, the
patient’s spine must be palpable pos-
Journal of the American Academy of Orthopaedic Surgeons
Figure 1 Correct positioning of the double-
lumen endotracheal tube to ventilate the left
lung. The bronchial lumen should be just dis-
tal to the carina and the tracheal lumen just
proximal to the carina. The tracheal lumen is
occluded to allow for selective ventilation of
the left lung.
teriorly, and the umbilicus visible an-
teriorly, to allow orientation and ex-
posure in case conversion to an open
thoracotomy is necessary. The thora-
cotomy tray should be available in the
operating suite. The arm on the con-
vexity of the curve can usually be po-
sitioned out of the sterile field, espe-
cially when the upper instrumented
level is at T5 or below. However, when
the upper instrumented level is above
T5, the arm may be incorporated into
the sterile field to provide better con-
trol of the patient’s arm and scapula,
making proximal portal placement
easier.
One or two surgeons are positioned
on the posterior aspect and one on the
anterior aspect of the patient. The scrub
assistant is usually anterior. The video
monitors should be at the head of the
table on both sides of the patient to
give the surgeons on each side a di-
rect view. The fluoroscopy C-arm unit
is brought in anteriorly when screws
are placed, with the monitor at the foot
of the table.
Although some surgeons perform
the diskectomy on the posterior as-
pect, the anterior position allows bet-
ter control of posterior penetration be-
yond the posterior anulus fibrosus and

Figure 2 Operating room setup.
posterior longitudinal ligament. Dur-
ing screw placement and instrumen-
tation, the surgeon may be more com-
fortable at the posterior aspect of the
patient because leaning over the op-
erating room table is then unneces-
sary, and it is safer to direct the screws
slightly anteriorly.
Surgical Procedure
Portal Placement
Accurate placement of the portals
is critical because they determine the
approach for the diskectomies and,
more important, the screw starting
points and directions. Before the pa-
tient is prepared and draped, the spi-
nal levels to be instrumented are lo-
cated fluoroscopically in the coronal
and sagittal planes, and the skin is
marked. In general, the incision for
Vol 11, No 4, July/August 2003
Daniel J. Sucato, MD, MS
A single anterolateral portal is
placed at the apex of the curve in the
anterior-to-midaxillary line, and the
thoracoscope is placed through this
portal. The thoracoscope consists of
a camera and a scope that is angled
at 30° or 45°. Seen from the antero-
lateral portal, the spine is horizontal
on the monitor; seen from the postero-
lateral portal, the spine is vertical, giv-
ing a good “pipeline” view (Fig. 3).
The scope should be oriented to see
the disks straight on when the tho-
racoscope is in the anterolateral por-
tal. This is best achieved by keeping
the orientation light from the lens per-
pendicular to the spine, with the scope
handle at the 3-o’clock position when
looking at the most cephalad disk (Fig.
4, A) and at the 9-o’clock position
when looking at the most caudad disk
(Fig. 4, B). This position allows visu-
alization down the axis of the disk
space and provides a true anteropos-
terior view of the vertebral bodies.
The posterolateral portals are made
under direct visualization. The place-
ment of the most cephalad portal is
very important for proper instrumen-
tation. The skin mark initially made
under fluoroscopic visualization is
the portals should be directly over the used to place a guide pin, which is
rib so that two portals (above and be- then assessed using the camera in the
low the rib) can be used for each in- anterolateral portal. The ribs should
cision. be counted to check the level of the
Figure 3 Thoracoscopic video images. A, With the thoracoscope in the anterolateral portal,
the spine is horizontal on the monitor and provides a good assessment of the superior and
inferior extent of each vertebral body. B, With the thoracoscope in one of the posterolateral
portals, the spine is more vertical on the monitor “pipeline” view, providing a good assess-
ment of the anterior and posterior aspects of the spine. The diaphragm is visible at the top
of the image.
223
Thoracoscopic Anterior Instrumentation and Fusion for Idiopathic Scoliosis

Figure 4 A, Use of the thoracoscope in the anterolateral portal. To view the proximal (cephalad) portion of the spine, the camera is po-
sitioned parallel to the floor and the light source handle is at the 3-o’clock position. The spine appears horizontal on the monitor (inset).
B, To view the distal (caudad) aspect of the spine, the camera is positioned parallel to the floor and the light source handle is at the 9-o’clock
position. This keeps the spine horizontal on the monitor (inset). C, Portal placement for a typical thoracic idiopathic curve. The anterolateral
portal is made in the anterior axillary line at the apex of the curve.

guide pin. If the pin is not sufficient-
ly superior or posterior to allow the
surgeon to place the proximal screw,
the pin is moved and the portal in-
serted. The camera may then be placed
through that portal to check the po-
sition further.
The remaining posterolateral por-
tals are then placed, with close atten-
tion paid to the distances between
portals and their positions in the an-
teroposterior and superoinferior di-
rections. Positioning is assessed with
the thoracoscope in the anterior por-
tal to ensure that the portals are made
directly over the vertebral bodies. A
typical portal configuration for a
seven- or eight-level instrumentation
is a single anterolateral portal and
four posterolateral portals (Fig. 4, C).
Various portal configurations have
been described, including posterolat-
eral portals only or a combination of
three anterolateral with three pos-
terolateral portals.
Disk Excision and Bone Grafting Autologous rib or iliac crest bone
Disk excision is the most important grafts can be used and probably are
aspect of the procedure. The surgeon best placed immediately upon com-
incises the pleura in the midvertebral pletion of the diskectomy at each lev-
body, then coagulates the segmental el. Bone funnels are used to place the
vessels. The pleura should be bluntly grafts and should start in the depths
teased posteriorly past the rib heads
and anteriorly around the front of the
spine to allow access to the anterior
longitudinal ligament and contralat-
eral anulus. Sharp incision of the disk
can be made with a scalpel blade or
harmonicscalpel.Diskshavers,rongeurs,
and curettes are used to excise the disk
as completely as possible (Fig. 5). An-
imal studies comparing open thora-
cotomy with thoracoscopic techniques
havedemonstratedcomparableamounts
of diskectomy.13,14 A quantitative anal- Figure 5 Axial CT scan of a thoracic disk
ysis of computed tomography (CT) space after diskectomy and bone grafting
in 12 adolescent patients (mean age, done as part of an anterior thoracoscopic in-
strumentation and fusion. Note the bone graft
13.3 years) demonstrated that a mean material (arrows) packed all the way to the
of 73% of the disk and end plate was opposite side of the disk space and posteri-
removed, allowing correction from a orly. The rod is seen on the right side of the
vertebral body.
mean of 55° to a mean of 9°.15

224 Journal of the American Academy of Orthopaedic Surgeons


of the disk space to ensure that the
grafts are packed completely.
Screw Placement
Before screws are placed, the pa-
tient’s position should be rechecked
toensureitisdirectlylateral.Thefluo-
roscopic image should be at right an-
gles to the vertebral bodies in the an-
teroposterior projection and is used
to confirm that the screw is oriented
parallel to the end plate. The thora-
coscope is placed in the anterior por-
tal initially to direct the guidewire
with respect to the superoinferior
starting point and orientation. The
thoracoscope is then moved to a pos-
terolateral portal to check the antero-
posterior starting point and its direc-
tion. The anteroposterior fluoroscopic
images are then used to fine-tune the
starting point in the superoinferior di-
rection.
Screws are placed beginning at the
apex of the curve, with the starting
point of the screw just anterior to the
rib head. The screws are directed
slightly anteriorly to avoid the spinal
canal and to be in the midaxial plane
of the rotated apical vertebral bodies.
This screw orientation allows for ro-
tational correction during rod inser-
tion and compression. As screws are
placed proximal and distal to the apex,
the starting holes move slightly more
anteriorly. The cephalad screws are the
most difficult to place accurately with
good purchase because the vertebral
bodies are smaller, the rib heads ob-
scure more of the vertebral bodies, and
the proximal portals are often not ide-
ally placed. The proximal screws must
be placed with great care and atten-
tion to anatomic landmarks to ensure
that these screws are not too poste-
rior, which could lead to spinal canal
penetration, but are posterior enough
to allow secure purchase in good bone
stock (Fig. 6). It is often necessary to
remove the rib heads at T5 and T6 to
gain good access to the vertebral bod-
ies at these levels.
Present instrumentation systems
are modifications of open anterior in-
Vol 11, No 4, July/August 2003
strumentation systems, with all in-
struments made to fit through a 10.5-
mm–diameter portal. Screws in sizes
from 5.5 to 7.5 mm and rods in 4.0-,
4.5-, and 4.75-mm diameters are avail-
able. The proximity of the aorta to the
vertebral bodies in the upper and
midthoracic spine limits the amount
of bicortical screw purchase that can
be achieved16 (Fig. 6). In the lower
thoracic spine in a patient with idio-
pathic scoliosis, the aorta is posi-
tioned more anterior to the vertebral
body. Newer instruments allow the
surgeon to place screws without the
use of the guide wire, which can lead
to complications with inadvertent ad-
vance across the vertebral body.
Rod Insertion and Correction
Maneuvers
The stiffness of the curve, the pur-
chase of the most proximal screws,
and whether maximum correction is
desired (Lenke IA curve) will deter-
mine whether a small coronal bend
should be placed in the rod before in-
serting it into the chest. In taller pa-
tients with smaller, more flexible
curves and larger vertebral bodies, no
coronal bend in the rod is necessary.
In patients with a very lordotic tho-
racic segment, a kyphotic bend can
be placed in the rod.
The rod is inserted through the dis-
tal or proximal posterolateral portal
and grasped within the chest with a
rod grabber so that it can be seated
into the screws in one step. The rod
is initially seated distally to help con-
trol the length of rod that protrudes
distal to the screw and prevent it from
pushing against the diaphragm.
Two correction maneuvers are per-
formed: compression and cantilever.
Because the rod is essentially straight
in the coronal plane, in contrast with
the deformity, the rod can be seated
only in the distal three or four screws.
Initially, compression is performed
across these screws, followed by can-
tilevering the rod down into the re-
maining proximal screws. After the
rod is captured in the proximal screw
Daniel J. Sucato, MD, MS
Figure 6 Axial CT scan of a thoracic ver-
tebral body after anterior thoracoscopic in-
strumentation and fusion. The starting posi-
tion of the screw is just anterior to the rib head.
The outline of the aorta is seen at approxi-
mately 1 o’clock, just posterior to the left
mainstem bronchus. The screw has one to two
threads engaging the opposite cortex; how-
ever, the screw tip is close to the aorta.
heads, compression is then complet-
ed at these levels with care taken to
avoid excessive force on the top
screws. The securing plugs are then
tightened fully. The surgeon must be
sure to place the guide sleeve over the
screw or grasp the rod to produce a
countertorque to prevent screw mi-
gration or “plowing.” Anteroposterior
and lateral radiographs or fluoroscop-
ic images should be checked to en-
sure that all screws are safely posi-
tioned and that correction is adequate
in the coronal and sagittal planes.
Pleural Closure and Chest Tube
Insertion
The pleura can be closed to help
decrease chest tube output, limit de-
velopment of lung adhesions, and
contain the bone graft in the disk
space. Diaphragmatic repair is incor-
porated into the pleural closure when
the instrumentation extends to T12 or
L1. The pleura is closed with an En-
dostitch device (US Surgical, Nor-
walk, CT), running a suture begin-
ning distally and another beginning
proximally, which then meet in the
center so that they can then be tied
easily. A chest tube is placed through
the incision of the most distal poste-
rior portal skin incision. Because of
the single, small-diameter rod, all pa-
225
Thoracoscopic Anterior Instrumentation and Fusion for Idiopathic Scoliosis

tients should wear a brace during the
day (when not sleeping) for the first
3 months.
Early Results
In one series of 28 girls (average age,
12.1 years) with a mean preoperative
curve of 55° (range, 46° to 78°), the
mean postoperative curve at 1 year
was 14° (74.5% correction)15 (Fig. 7).
Complications included six proximal
screws that partially pulled from the
vertebral body at the time of compres-
sion in four patients; two screws that
cut out at the time of insertion because
of small vertebral bodies in two pa-
tients; guidewire migration into the
spinal canal in one patient, with re-
sultant dural leak without neurologic
sequelae; and asymptomatic pseudar-
throsis in one patient who underwent
a posterior spinal fusion.15
Picetti and Bueff17 reported follow-
ups over 2 years on 50 patients (mean
age, 12.7 years) with a mean preop-
erative curve of 58°. Improvements in
techniques resulted in enhanced cor-
rection and fewer complications over
the course of this series. Mean curve
correction was 50.1% in the first 10 pa-
tients and 68.6% in the last 10. Sur-
gical time improved from a mean of
6 hours 6 minutes in the initial 30 cas-
es to 3 hours 58 minutes in the last
10 cases. Mean blood loss was 266
mL. The chest tube was in place for
a mean of 2.25 days (range, 1 day to
5 days), and hospital stay averaged
2.9 days (range, 2 to 7 days). Report-
ed complications included one screw
pullout, three patients with chest wall
numbness, five mucous plugs, one
wound revision, and two rod frac-
tures. A demineralized bone matrix
product was used in the initial pa-
tients, resulting in a high incidence
of pseudarthrosis; however, only 1
patient of the remaining 35 had a
pseudarthrosis when autologous rib
graft was used.17
Complications
There are no published series of pa-
tients who have had thoracoscopic in-
strumentation and fusion for idio-
pathic scoliosis, so the prevalence of
complications is not known. How-
ever, complications that have been
presented and discussed at scientific
meetings can be categorized as
anesthesia-related and surgical. The
anesthesia-related complications in-
clude the down lung syndrome, with
significant atelectasis present on the
initial chest radiograph; inability to
tolerate single-lung ventilation and
conversion to an open technique or
posterior spinal fusion; inability to ob-
tain single-lung ventilation because
of difficulty in tube placement; and
pneumothorax secondary to high air-
way pressures.12 Because this proce-
dure is new and technically demand-
ing, the incidence of complications
can be high, especially early in the
surgeon’s experience. Complications
that can occur during surgery include
blood vessel injury, lymphatic injury
with resultant chylothorax, guide-pin
migration into the opposite side of the
chest with resultant pneumothorax,18
distal migration or plowing of the
screw when the rod is seated prox-

Figure 7 Preoperative anteroposterior (A) and lateral (B) radiographs of a 13-year-old girl with a 56° right thoracic idiopathic curve with
a notable trunk shift to the right and hypokyphosis (panel B). Anteroposterior (C) and lateral (D) radiographs 1 year after anterior tho-
racoscopic instrumentation from T5 to T12, with near-complete correction of the coronal plane deformity and restoration of the normal sag-
ittal profile.

226 Journal of the American Academy of Orthopaedic Surgeons


imally or is compressed, and screw
cutout at the time of screw insertion.
Summary
The endoscopic approach to curve cor-
rection, instrumentation, and fusion
for spinal deformity is a new technique
that promises improved patient care
because it limits the surgical incision
and chest wall compromise, improves
postoperative pain and pulmonary
function, and enhances cosmesis.
Compared with posterior instrumen-
References
1. Lenke LG, Bridwell KH, Blanke K, Bal-
dus C, Weston J: Radiographic results
of arthrodesis with Cotrel-Dubousset
instrumentation for the treatment of
adolescent idiopathic scoliosis: A five to
ten-year follow-up study. J Bone Joint
Surg Am 1998;80:807-814.
2. Richards BS, Herring JA, Johnston CE,
Birch JG, Roach JW: Treatment of ado-
lescent idiopathic scoliosis using Texas
Scottish Rite Hospital instrumentation.
Spine 1994;19:1598-1605.
3. Lenke LG, Betz RR, Bridwell KH, Harms
J, Clements DH, Lowe TG: Spontaneous
lumbar curve coronal correction after se-
lective anterior or posterior thoracic fu-
sion in adolescent idiopathic scoliosis.
Spine 1999;24:1663-1672.
4. Betz RR, Harms J, Clements DH III, et
al: Comparison of anterior and posteri-
or instrumentation for correction of ad-
olescent thoracic idiopathic scoliosis.
Spine 1999;24:225-239.
5. Wall EJ, Bylski-Austrow DI, Shelton FS,
Crawford AH, Kolata RJ, Baum DS: En-
doscopic discectomy increases thoracic
spine flexibility as effectively as open
diskectomy: A mechanical study in a
porcine model. Spine 1998;23:9-16.
6. Newton PO, Wenger DR, Mubarak SJ,
Meyer RS: Anterior release and fusion
in pediatric spinal deformity: A com-
parison of early outcome and cost of
thoracoscopic and open thoracotomy
approaches. Spine 1997;22:1398-1406.
7. Regan JJ, Guyer RD: Endoscopic tech-
Vol 11, No 4, July/August 2003
tation, anterior instrumentation by ei-
ther open or thoracoscopic approach
can save fusion levels while improv-
ing three-dimensional correction.
However, no studies have directly
compared thoracoscopic instrumen-
tation and fusion with open anterior
and/or posterior procedures, making
any conclusive statements impossible.
A multicenter prospective study may
be needed to fully elucidate the ad-
vantages this technique may have and
to help define the exact indications for
a thoracoscopic approach to treat
scoliosis.
niques in spinal surgery. Clin Orthop
1997;335:122-139.
8. Lenke LG, Betz RR, Harms J, et al: Ad-
olescent idiopathic scoliosis: A new
classification to determine extent of spi-
nal arthrodesis. J Bone Joint Surg Am
2001;83:1169-1181.
9. Clements DH, Betz RR, Lowe TG,
Lenke LG, Newton PO: Abstract: Ado-
lescent idiopathic scoliosis with exces-
sive thoracic kyphosis: Comparison of
anterior versus posterior instrumenta-
tion for maintaining correction. Scolio-
sis Research Society 35th Annual Meeting
Book. Rosemont, IL: Scoliosis Research
Society, 2000, p 97.
10. Desiderio DP, Burt M, Kolker AC, Fis-
cher ME, Reinsel R, Wilson RS: The ef-
fects of endobronchial cuff inflation
on double-lumen endobronchial tube
movement after lateral decubitus posi-
tioning. J Cardiothorac Vasc Anesth 1997;
11:595-598.
11. Dieter RA Jr, Kuzycz GB: Complica-
tions and contraindications of thoracos-
copy. Int Surg 1997;82:232-239.
12. Sucato DJ, Girgis M: Bilateral pneumotho-
races, pneumomediastium, pneumoperi-
toneum, pneumoretroperitoneum, and
subcutaneous emphysema following in-
tubation with a double-lumen endotra-
cheal tube for thoracoscopic anterior spi-
nal release and fusion in a patient with
idiopathic scoliosis. J Spinal Disord Tech
2002;15:133-138.
13. Huntington CF, Murrell WD, Betz RR,
Daniel J. Sucato, MD, MS
Several important issues must be
kept in mind. First, the proposed ad-
vantages have not been confirmed
through scientific study. Second, the
technique continues to evolve to de-
crease the duration of surgery while
maintaining the safety of the proce-
dure. Third, screw migration and
proximity of screws to important soft-
tissue structures need further study.
Finally, this is a technically demand-
ing procedure with a steep learning
curve and may not be appropriate for
all surgeons who treat spinal defor-
mity.
Cole BA, Clements DH III, Balsara RK:
Comparison of thoracoscopic and open
thoracic discectomy in a live ovine
model for anterior spinal fusion. Spine
1998;23:1699-1702.
14. Newton PO, Cardelia JM, Farnsworth
CL, Baker KJ, Bronson DG: A biome-
chanical comparison of open and thora-
coscopic anterior spinal release in a
goat model. Spine 1998;23:530-536.
15. Sucato D, Kassab F, Dempsey M: Ab-
stract: Thoracoscopic anterior spinal in-
strumentation and fusion for idiopathic
scoliosis: A CT analysis of screw place-
ment and completeness of discectomy.
Scoliosis Research Society 36th Annual
Meeting Book. Rosemont, IL: Scoliosis
Research Society, 2001, p 90.
16. Sucato DJ, Duchene C: MRI analysis of
the position of the aorta relative to the
spine: A comparison between normal pa-
tients and those with idiopathic right tho-
racic curves. J Bone Joint Surg Am, in press.
17. Picetti GD III, Bueff HU: Abstract: En-
doscopic instrumentation, correction
and fusion of thoracic curves in idio-
pathic adolescent scoliosis. Scoliosis Re-
search Society 35th Annual Meeting Book.
Rosemont, IL: Scoliosis Research Soci-
ety, 2000, p 110.
18. Roush TF, Crawford AH, Berlin RE,
Wolf RK: Tension pneumothorax as a
complication of video-assisted thora-
scopic surgery for anterior correction of
idiopathic scoliosis in an adolescent fe-
male. Spine 2001;26:448-450.
227
 Spinal Manipulative Therapy for Low Back Pain
Rand Swenson, DC, MD, PhD, and Scott Haldeman, DC, MD, PhD, FRCPC

Abstract
Growing interest in complementary and alternative medicine in the United States was an average of 41 visits per 100
has been paralleled by increased use of spinal manipulative therapy in an attempt person-years.
to manage symptoms of low back pain, spinal stenosis, and spondylolisthesis. Chi- The increased volume of research
ropractors have been the main practitioners of spinal manipulative therapy, with about the effectiveness and mecha-
osteopaths and physical therapists providing a smaller fraction of these services. The- nisms of action of SMT have led to
ories explaining the mode of action of spinal manipulative therapy are largely pre- the inclusion of SMT on a very short
liminary and have focused on the mechanical effects of manipulative forces on the list of national guidelines for manag-
spine and neurologic responses to manipulation. The effects of spinal manipulation ing acute low back pain (eg, those de-
on patients with both acute and chronic low back pain have been investigated in veloped by the Agency for Healthcare
randomized clinical trials. Most reviews of these trials indicate that spinal manip- Research and Quality5 in the United
ulative therapy provides some short-term benefit to patients, especially with acute States and similar guidelines devel-
low back pain. oped in Great Britain and Denmark).
J Am Acad Orthop Surg 2003;11:228-237 Many patients under the care of prac-
titioners of conventional medical tech-
niques (eg, orthopaedic surgeons, neu-
rosurgeons, physiatrists) have already
Spinal manipulative therapy (SMT) is tact health care clinicians; in addition, undergone manipulation or will con-
one of the oldest therapeutic proce- their services are covered by most health sider spinal manipulation, especial-
dures offered to patients with lumbar insurance systems, including Work- ly if more commonly practiced med-
spine pain. Although it has been used ers’ Compensation, Medicare, and the ical procedures prove to be ineffective.
for more than 2,000 years, until recent- Military Health System. However, spi- Therefore, it is important for any phy-
ly there has been little scientific ev- nal manipulation is not practiced ex- sician who treats patients with back
idence comparing its mode of action clusively by chiropractors. Some os- pain to be knowledgeable enough to
and effectiveness with those of other teopathic physicians, physical therapists, advise patients about SMT and to con-
techniques. Interest in and use of com- and medical physicians are provid- sider whether there is a role for such
plementary and alternative therapies ing and specializing in SMT. The ex- treatment in selected patients.
increased in the United States during act utilization rate of spinal manip-
the last decade of the 20th century.1 ulation in the United States is not
Patients with low back pain often seek known, but it is estimated that >90%
Dr. Swenson is Associate Professor, Section of
care from clinicians who offer relax- of spinal manipulations in the United Neurology, Dartmouth Medical School, Lebanon,
ation therapy, yoga, nutritional sup- States are done by chiropractors.3 Al- NH. Dr. Haldeman is Clinical Professor, Depart-
plements, herbal therapy, massage, though many chiropractors also in- ment of Neurology, University of California–
acupuncture, and spinal manipulation. clude other physical modalities, nu- Irvine, Santa Ana, CA.
Spinal manipulation is the most com- tritional advice, and rehabilitation in
One or more of the authors or the departments
mon and widely recognized of the com- their practices, most chiropractic vis- with which they are affiliated has received some-
plementary and alternative therapies its include SMT. Approximately 30% thing of value from a commercial or other party
used to manage low back pain. In fact, of patients with low back pain use the related directly or indirectly to the subject of this
it is so commonplace that there is now services of chiropractors, with addi- article.
debate about whether it should be con- tional patients receiving SMT from other
Reprint requests: Dr. Haldeman, 1125 East 17th
sidered part of mainstream health care. health care practitioners.2 Shekelle and Street, Santa Ana, CA 92701.
Chiropractors, who provide the great- Brook4 studied insurance industry data
est percentage of manipulative treat- and reported that 7.5% of patients in Copyright 2003 by the American Academy of
ments,2,3 are now licensed in all states their population group used chiroprac- Orthopaedic Surgeons.
and are recognized as primary con- tic services each year and that there

228 Journal of the American Academy of Orthopaedic Surgeons


Spinal Manipulation
Many treatment methods are encom-
passed within SMT. These techniques
can be divided into three broad cat-
egories: therapeutic massage, mobi-
lization, and manipulative procedures
(or adjustments). Massage includes
manual procedures applied to the spi-
nal soft tissues without causing joint
movement, such as deep tissue mas-
sage, acupressure, muscle stimulation,
and relaxation methods. Mobilization
generally refers to procedures that ex-
ert stretching, traction, or pressure on
the spine within or at the limit of ac-
tive range of motion.6 Mobilization
procedures typically involve slow, oc-
casionally rhythmic stressing of the
joint, ligaments, and muscles. They
usually consist of passive force exert-
ed by the clinician on the joint but may
require active contraction of specific
muscles by the patient to assist in the
motion. Massage and mobilization
often are used in preparation for spi-
nal manipulation as a way of re-
ducing the force needed to do the ma-
nipulation and to relax the patient.
Spinal manipulation usually re-
quires application of a quick, high-
velocity, short-amplitude force, either
directly or indirectly, to the spine.6
Manipulative forces can be applied
with the patient lying either on his or
her side or prone. Most of these tech-
niques involve so-called short levers
directed at the vertebral processes (eg,
spinous, transverse, mammillary, ar-
ticular) to apply force to a specific spi-
nal segment. So-called long lever
techniques consist of force applied to
an extremity (eg, arm, shoulder, hip,
leg) to move the spine indirectly. Spi-
nal manipulative procedures also
may involve traction or a mobilizing
force at the end of the physiologic
joint range of motion, followed by a
short-amplitude impulse to move the
joint into the paraphysiologic range,
beyond the passive range of motion
but short of the point of anatomic dis-
ruption of the joint (Fig. 1). The typ-
ical forces and movements produced
Vol 11, No 4, July/August 2003
Rand Swenson, DC, MD, PhD, and Scott Haldeman, DC, MD, PhD, FRCPC
Zone C
Paraphysiologic
range of Zone A
motion Active
Zone B
Passive
range of
motion
range of
motion
Joint disruption
Increasing joint range of motion
Figure 1 Mobilization and manipulation
take the spine beyond the active range of mo-
tion (Zone A). Ligaments and muscles resist
motion beyond that point; articular mobili-
zation passively stretches them into Zone B,
although it may be actively assisted by the
patient. Manipulation involves bringing the
joint beyond this passive range of motion into
a paraphysiologic range of motion (Zone C).
Often there is cavitation of the joint being ma-
nipulated. Motion beyond that point may
cause injury to the structures supporting the
joint.
by such types of manipulation have
been characterized in studies of vol-
unteers.7 The movement is often ac-
companied by an audible sound that
may be attributable to joint cavita-
tion,8 which is the release of tissue gas
(probably nitrogen) into the joint
space as the result of a vacuum cre-
ated by the manipulative thrust. This
audible sign indicates that there has
been motion in the joint. However,
many manipulative practitioners do
not consider an audible release essen-
tial to obtain a good manipulative re-
sponse.
A common spinal manipulative
procedure is the so-called hold-thrust
adjustment, in which the force is
maintained briefly before and after
the application of the impulse. The re-
coil technique involves an impulse
from a neutral position followed by
an immediate termination of contact.
In several techniques, very low force
is applied in precise directions to ver-
tebral structures, under the theory
that the more specific the thrust, the
less force necessary. Acupressure and
soft-tissue massage may be done in
preparation for manipulation or as
the sole method of treatment. In ad-
dition, several mechanical devices,
such as the handheld, spring-loaded
Activator (Fig. 2), have been devel-
oped to deliver force directly to the
spine with or without manual guid-
ance. It remains to be determined
whether the mechanical and clinical
effects produced by these instruments
are similar to those of the manual
techniques.9
Theoretic Basis for
Symptom Relief
The theoretic basis for SMT has
evolved with increased understand-
ing of spinal pathology, spinal biome-
chanics, and pain physiology. Initial-
ly, the manipulable lesion (osteopathic
lesion or subluxation in chiropractic
terminology) was considered a bone
out of place impacting either the vas-
cular structures or spinal nerve roots.10
Most manipulative practitioners and
all chiropractic schools have broad-
ened this concept to encompass cur-
rent theories of spinal pathology that
incorporate concepts of abnormal spi-
nal biomechanics and include neuro-
physiologic theories about reflex func-
tion and pain physiology.11
The most common theory of the
manipulable lesion is that a vertebral
Figure 2 Manipulation with a handheld de-
vice (Activator; Activator Methods Interna-
tional, Phoenix, AZ) designed to deliver force
to the tissue of the back.
229
Spinal Manipulative Therapy for Low Back Pain
unit can display abnormal motion or
become fixated (ie, mobility can be re-
stricted within the normal physiolog-
ic range of joint motion). Addition-
ally, it is thought that biomechanical
changes are often associated with pain
and abnormal spinal reflex function,
including muscle spasms and auto-
nomic nervous system responses. The
manipulation is thought to have a di-
rect effect on muscles and joints as well
as an indirect effect on the nervous
system. Limited experimental and clin-
ical evidence supports this theory.
Magnetic resonance imaging studies
have indicated a direct effect on spi-
nal joints consistent with reports de-
scribing increased spinal range of mo-
tion after spinal manipulation.12 Spinal
manipulation also is accompanied by
a reflex contraction of paraspinal mus-
culature.13 Whether this is of thera-
peutic benefit and whether these re-
flex responses produce long-term
changes in muscle tone or muscle
spasm are unknown.
Recent neurophysiologic research
has focused on possible effects of SMT
on the central nervous system.11 Al-
tered pain thresholds have been re-
ported after SMT, possibly related to
activation of endogenous pain-
suppression mechanisms. In addition,
decreased pain response after lumbar
manipulation has been associated
with abnormal somatosensory-
evoked potentials from paraspinal
musculature of patients with low back
pain, suggesting a central effect on
sensory processing.14 Activation of
zygapophyseal joint receptors in rats
is capable of markedly attenuating the
reflex response in paraspinal muscles
to noxious stimulation of nerves in the
intervertebral disk,15 which suggests
interaction between spinal joint recep-
tors and the processing mechanisms
for spinal pain reflexes, at least in an-
imals. Suter et al16 investigated the ef-
fect of manipulation of the sacroiliac
joint on the degree of quadriceps mus-
cle inhibition produced by knee joint
pathology. They showed that manip-
ulation of the sacroiliac joint de-
230
creased the inhibitory effect, suggest-
ing interaction between manipulation
and the inhibition of voluntary activ-
ity produced by pain. Despite exper-
imental observations, however, the
underlying mechanisms proposed to
explain the therapeutic effects of SMT
remain poorly understood and re-
quire further investigation.
Clinical Trials
Many patients with uncomplicated
low back pain do not have clear and
quantifiable underlying pathophysi-
ology, which has largely prevented
the use of physiologic outcome mea-
sures in studies of back pain treat-
ment. For this reason, studies of the
therapeutic efficacy of most treatment
approaches to low back pain have
tended to be empiric, relying on clin-
ical outcome measures such as pain
scores, functional capacity, patient sat-
isfaction, time lost from work, and
cost (mostly from insurance and
Workers’ Compensation data). Many
of these researchers have attempted
to compare chiropractic with conven-
tional medical treatment, while others
have compared spinal manipulation
with other nonsurgical interventions
or placebo treatments. Although most
patients treated by chiropractors re-
ceive spinal manipulation, chiroprac-
tic treatment is rarely limited to this
modality, and caution must be exer-
cised when extrapolating from trials
of chiropractic care. For example,
studies of patient satisfaction have
been very favorable to chiropractic
treatment, but that may have more to
do with doctor-patient interaction
than with SMT.17 Although compar-
ison studies of Workers’ Compensa-
tion18 and private health insurance19
data have mostly shown similar or
lower costs for patients treated by chi-
ropractors, this may reflect a differ-
ence in the population that seeks al-
ternative providers.
The most widely accepted model
for overcoming the problem of differ-
Journal of the American Academy of Orthopaedic Surgeons
ent patient populations is the ran-
domized clinical trial. Forty-four ran-
domized clinical trials evaluating the
efficacy of spinal manipulation in pa-
tients with low back pain have been
published, and there are more than
50 reviews of these trials, each using
different criteria to determine their
value. There are even systematic stud-
ies that rate the quality of these clin-
ical trial reviews.3,20-28
Many of the published clinical tri-
als are of relatively low quality, often
involving small numbers of patients,
differing outcome measures, short
follow-up periods, heterogeneous
study populations, a range of meth-
ods of manipulation or manual ther-
apy, and varying degrees of blinding
of patients and assessors. It is ex-
tremely difficult to blind patients in
randomized controlled trials of phys-
ical interventions such as spinal ma-
nipulation and, therefore, to develop
appropriate placebo controls. Many
researchers have avoided placebo and
instead have tried to use more prag-
matic approaches, such as examining
and comparing the outcomes of two
or more common clinical procedures,
one of which is manipulation or man-
ual therapy. However, the clinical val-
ue of many of the treatments against
which manipulation can be compared
is mostly unknown or has minimal
research support. In addition, many
of the manipulative treatment proto-
cols in these studies do not model typ-
ical clinical practice, making it diffi-
cult to extrapolate the results to the
clinical setting. This is particularly
problematic in studies that attempt to
establish rigorous controls or use pla-
cebo treatments such as detuned dia-
thermy, sham laser treatments, or
sham manipulations.
Acute Low Back Pain
The studies by Glover et al29 and
Hadler et al30 of the effect of manip-
ulation in patients with acute low
back pain have received high qual-
ity scores in most reviews and are
widely quoted to have established

some legitimacy for manipulation.
Glover et al29 compared two groups
of patients, one that received a sin-
gle manipulative treatment followed
by four sessions of detuned diather-
my (placebo) and the other five treat-
ments of the placebo only. Significant
(P < 0.05) improvement was report-
ed in patients immediately after the
initial manipulative treatment. How-
ever, there was no difference between
groups after 3 or 7 days. Hadler et
al30 randomized patients to receive ei-
ther a single manipulative treatment
or mobilization procedure. In patients
with pain of 2 to 4 weeks’ duration,
there was significant (P = 0.009) dif-
ference in favor of SMT at 3 days, but
not at 6 days, after treatment. Despite
the positive short-term beneficial out-
come of manipulation in these two
studies, the use of a single manipu-
lative treatment in both has been crit-
icized as not modeling usual clinical
practice. Mathews et al31 attempted
to remedy this shortcoming by com-
paring a group of patients with acute
lower back pain treated with up to 10
manipulative treatments with a
group receiving 6 treatments of infra-
red heat applied to the lumbar spine.
A significant (P < 0.05) difference was
reported in recovery at 2 weeks for
patients with lower back pain who
also had leg pain, but not for patients
with lower back pain only. There was
no difference in relapse rate after 1
year.
MacDonald and Bell32 compared
the effect of five spinal manipulative
treatments done in combination with
a low back education program with
the effects of five low back educa-
tion sessions alone. The results were
not statistically significant in favor of
manipulation but favored manipula-
tion. Shekelle et al3 pooled the results
of seven trials, using pain and func-
tional outcome as clinical measures,
and concluded that spinal manipula-
tion had significant (P < 0.05) over-
all benefit, averaging approximately
a 34% improvement in recovery com-
pared with several alternative treat-
Vol 11, No 4, July/August 2003
Rand Swenson, DC, MD, PhD, and Scott Haldeman, DC, MD, PhD, FRCPC
ment methods. They calculated a
pooled odds ratio of 0.17 probability
of faster recovery at 3 weeks com-
pared with other therapies. In a sim-
ilar meta-analysis, Brønfort23 con-
cluded that there was moderate
evidence of short-term efficacy when
SMT was used to manage acute low
back pain. van Tulder and Waddell27
concluded that there is moderate ev-
idence that manipulation is more ef-
fective than placebo for short-term re-
lief of acute low back pain; however,
they also stated that it was not yet
possible to judge whether manipula-
tion is more effective than nonsteroi-
dal anti-inflammatory drugs or phys-
ical therapy. This type of review led
the Agency for Health Care Policy
and Research in 1994 to add manip-
ulation to its short list of recommend-
ed treatments for acute low back
pain.5
Chronic Low Back Pain
The study by Koes et al33 of SMT
in patients with chronic low back pain
has received high quality scores. The
authors compared a course of 14 SMT
treatments with three other treatment
approaches. One included massage,
heat, and modalities such as electro-
therapy, ultrasound, and diathermy;
another consisted of medical manage-
ment with anti-inflammatory medi-
cations and advice; the third invloved
detuned modalities (placebo). At
1-year follow-up, they found signif-
icant (P = 0.05) benefit in terms of
pain, but not physical functioning, for
patients treated with spinal manipu-
lation compared with the group that
received massage, heat, and modal-
ities. There was significant benefit for
spinal manipulation over medical
management (P = 0.05) and over pla-
cebo modalities (P = 0.02) at 6 weeks
but not at 12 weeks.
Pope et al34 conducted a four-arm
clinical trial in patients with chronic
low back pain. They compared a
group of patients treated with nine
spinal manipulative treatments with
three other treatment protocols (mas-
sage, transcutaneous electrical mus-
cle stimulation, or corsets). At 3-week
follow-up, SMT showed a significant
(P < 0.05) benefit in terms of disabil-
ity score. However, this improvement
was not found to extend to pain
scores, where the only significant (P
= 0.05) benefit was found in the com-
parison with transcutaneous muscle
stimulation.
Triano et al35 compared conven-
tional spinal manipulation provided
daily for 2 weeks with a similar num-
ber of placebo manipulations using
measured forces below a previously
identified manipulative threshold. A
third group of patients received the
same number of structured education
sessions. The differences in Oswestry
disability scores between those treat-
ed with manipulation and placebo
reached statistical significance (P =
0.004) at 2 weeks but not at 4 weeks.
No comparison was made with the
back education program. At 4 weeks,
the difference in pain scores between
manipulation and placebo manipula-
tion were not statistically significant.
However, the authors placed con-
straints on the manipulative treat-
ments to adequately mimic the pla-
cebo treatments.
The most comprehensive study of
the long-term effects of manipulation
was done by Meade et al.36 They com-
pared nine spinal manipulative treat-
ments done in private chiropractic of-
fices with six hospital-based physical
therapy clinic sessions that included
conventional physical therapy as well
as spinal manipulation in a cohort of
741 patients. They reported signifi-
cant (P < 0.05) benefits for the group
treated by the chiropractors in terms
of pain at 6-week, 6-month, and 1-,
2-, and 3-year follow-ups and in terms
of disability at 6-month and 2- and
3-year follow-ups. However, this
study has been criticized for having
excessive numbers of patients lost to
follow-up as well as for differences in
treatment settings.
Cherkin et al37 conducted a ran-
domized clinical trial in a group of 321
231
Spinal Manipulative Therapy for Low Back Pain
patients with acute or chronic low
back pain of at least 7 days’ duration.
Patients were assigned to SMT or
physical therapy according to the
McKenzie method or were provided
an educational booklet about back
pain. The group treated by chiroprac-
tors did significantly better than the
group receiving the booklet at 4
weeks (P = 0.02) but not at 12 weeks,
1 year, or 2 years. However, there was
no difference between patients who
received SMT and those treated with
physical therapy. Although disabili-
ty scores appeared to be better for the
groups treated with manipulation
and physical therapy than for the
group given the booklet, this was not
statistically significant. In the groups
that received active treatment, ap-
proximately 75% of patients report-
ed that their care was good or excel-
lent, while only 30% of those in the
group given the booklet reported re-
ceiving good or excellent care (P <
0.001). One limitation of this study
was the rather low level of initial
symptoms that, combined with the
usual trend to spontaneous improve-
ment of low back pain, resulted in a
statistical floor effect restricting the
ability to detect differences between
treatment groups.
Until recently, systematic reviews
and meta-analyses of trials in patients
with chronic low back pain have not
been as positive as those of patients
with acute pain. The issue has been
complicated by inconsistency between
studies, with some showing a bene-
ficial effect on pain but not disability,
and others reaching the opposite con-
clusion. However, in none of the stud-
ies has it been shown that manipu-
lation is less effective than any
treatment approach with which it has
been compared. In their systematic re-
view, van Tulder et al24 concluded that
there was strong evidence that ma-
nipulation was more effective than pla-
cebo and moderate evidence that ma-
nipulation was more effective than
several other treatments with which
it had been compared. Brønfort23 con-
232
cluded that there was moderate ev-
idence of a short-term effect of ma-
nipulation in chronic low back pain
but inconclusive evidence of a long-
term effect. In the review by the Swed-
ish Council on Technology Assessment
in Health Care in 2000, using the
Cochrane Collaboration methodology,
van Tulder and colleagues concluded
that “there is strong evidence that man-
ual therapy is more effective than a
placebo treatment for short-term re-
lief”27 of acute low back pain, although
they found insufficient evidence to de-
termine whether it is better than other
physical therapeutic interventions or
drug therapy. In an examination of the
literature on chronic low back pain,
the authors found that “there is strong
evidence that manual therapy provides
more effective short-term pain relief
than a placebo treatment” and mod-
erate evidence that “manual therapy
is more effective than usual care by
the general practitioner, bed rest, an-
algesics and massage for short-term
pain relief.”28 However, they also con-
cluded that “there is limited and con-
flicting evidence of any long-term ef-
fects.”28
Two recent studies have compared
the treatment of back pain patients by
chiropractors with that by various
other providers. Skargren et al38 ran-
domized 323 patients with back and
neck pain to chiropractic treatment or
treatment by physical therapists. The
authors found no significant overall
difference in the cost or outcome of
treatment. However, chiropractic
treatment was favored in the group
with a short duration and high level
of symptoms, whereas physical ther-
apy was favored in those with more
chronic symptoms. Some questions
have been raised about their conclu-
sions because of restrictions that the
experimental protocol placed on the
chiropractic treatment. Nonetheless,
the findings of Skargren et al38 sug-
gest overall similarity in benefit be-
tween treatments incorporating SMT
and those using intensive physical
therapy.
Journal of the American Academy of Orthopaedic Surgeons
Hurwitz et al39 randomized 681 pa-
tients with low back pain in a man-
aged care setting to conventional med-
ical treatment with or without physical
therapy or to chiropractic care with
or without physical modalities. They
found that all groups improved over
6 months of follow-up without any
significant differences in disability or
pain between the patients treated by
chiropractors and those treated by
medical providers with or without
physical therapy. Although these stud-
ies38,39 do not specifically address the
question of benefit of SMT as an iso-
lated intervention, their results argue
that care incorporating spinal manip-
ulation does at least as well as the best
medical management.
Radiculopathy and Disk
Herniation
While most studies on the effects
of spinal manipulation on patients
with radiculopathy and/or disk her-
niation have been individual case
reports or uncontrolled small case se-
ries, there have been three random-
ized clinical trials40-42 and one nonran-
domized clinical trial.43 Most of the
small studies had no control or com-
parison group. The authors of the ran-
domized clinical trials compared ma-
nipulation with other nonsurgical
measures (eg, traction, exercise, heat
treatments), and all reported an ad-
vantage for spinal manipulation at 2
to 6 weeks.40-42 In one of the studies,
patients were followed for 1 year, and
no difference in relapse rates was re-
ported.42 The authors of the nonran-
domized trial compared manipula-
tion with surgical intervention and
reported greater benefit with sur-
gery.43 A trial of single-session lum-
bar rotary manipulation was report-
ed to significantly (P = 0.0045)
improve the abnormal H-reflex am-
plitude in patients with unilateral
lumbar disk herniation and radicu-
lopathy; the abnormal H-reflex laten-
cy showed insignificant improvement
(P = 0.3877).44 No complications or se-
rious side effects that would preclude

the use of manipulation in patients
with radiculopathy were reported in
any of the case series or trials. Manip-
ulation is of uncertain benefit in pa-
tients with radiculopathy, and well-
designed studies in this area are
needed.45
Spinal Stenosis and
Spondylolisthesis
No randomized clinical trials of
patients with spinal stenosis, spondy-
lolisthesis, or spondylolysis have
been done. In their case series of pa-
tients with spinal stenosis, Kirkaldy-
Willis and Cassidy46 reported favor-
able results in the group treated with
SMT. They also reported the respons-
es of 283 consecutive patients with
low back pain to spinal manipulation,
25 of whom had spondylolisthesis.46
The response was similar to that not-
ed in the other patients, and no ad-
verse outcomes were reported.
Management of Common
Disorders
Premanipulation Examination
The premanipulation evaluation of
patients is similar to the routine or-
thopaedic assessment of patients with
low back pain, but with certain char-
acteristics unique to practitioners of
manipulation. The three-step process
of evaluation consists of establishing
a diagnosis, determining whether the
patient is a candidate for spinal ma-
nipulation, and deciding on the type
of manipulation that should be used.
Elements of the clinical history, in-
cluding the mechanism and timing of
onset of the symptoms, as well as ag-
gravating and relieving activities and
the effects of posture on symptoms,
are useful for establishing the treat-
ment plan and predicting the likeli-
hood of effectiveness of a course of
treatment.
Many practitioners of spinal ma-
nipulation, especially chiropractors,
use radiographs to evaluate patients.
Radiographs may help the practi-
Vol 11, No 4, July/August 2003
Rand Swenson, DC, MD, PhD, and Scott Haldeman, DC, MD, PhD, FRCPC
tioner determine the integrity of the
underlying osseous structure to with-
stand the forces used in the treatment.
They also may help in the analysis of
spinal distortions and postural rela-
tionships that have been perceived as
important in planning the type of
manipulation. Although the former
use of radiographs is clearly impor-
tant, the latter is controversial. Radio-
graphs, with certain exceptions, have
not been shown to be predictive of
spinal symptoms. Although there
appears to be reasonable reliability in
chiropractic interpretation of radio-
graphic findings, it has yet to be
established that the postural mark-
ings of relative vertebral position con-
tribute to treatment effect and out-
comes.
Radiographs often are necessary to
aid in the appropriate premanipula-
tion evaluation of patients with back
pain. Before January 2000, Medicare
required radiographic demonstration
of a subluxation for all patients treat-
ed by chiropractors, forcing most of-
fices to maintain their own facilities.
With the change in Medicare regula-
tions and the revision of medical eth-
ics allowing referral to chiropractors,
medical radiographic facilities have
become more accessible to chiroprac-
tors. Also, radiologists have become
more aware of the studies required
by chiropractors, and there is a trend
in the chiropractic community to re-
fer patients to medical facilities for the
necessary studies.
The primary diagnostic proce-
dures used to determine the type of
manipulation that may be most ben-
eficial are postural analysis and man-
ual palpation of the spine and
paraspinal tissues. Of these two pro-
cedures, palpation is the most fre-
quently applied for the identification
of the manipulable lesion (eg, sublux-
ation, fixation, osteopathic lesion).
Static palpation, done with the patient
in a relaxed (usually prone) position,
is used to detect areas of muscle ten-
sion, tenderness, spasm, and segmen-
tal differences in tissue consistency.
This is often followed by motion pal-
pation, in which the clinician palpates
individual spinal motion segments
through various ranges of motion,
evaluating each direction of motion
for restriction and asymmetry of
movement. Interexaminer reliability
and reproducibility of these proce-
dures is not high.47 Intraexaminer
reliability is higher for some pro-
cedures, but this is of uncertain
value.47
Spinal Manipulation Techniques
SMT for low back pain can be di-
vided into three procedures: those
preparatory to manipulation, manip-
ulation itself, and those subsequent
to manipulation. Preparatory treat-
ments, such as cryotherapy, various
types of stretching, and soft-tissue
massage or reflex techniques, are usu-
ally done to relax the patient, reduce
tension, and ease acute pain. When
the pain is acute and severe or when
significant muscle spasm is evident,
such treatments often are followed by
mobilization procedures rather than
a specific manipulation.
If the preparatory therapies relax
the patient sufficiently, spinal manip-
ulation with impulse is applied to one
or more of the abnormal spinal seg-
ments identified during examination.
In the lumbar spine, the manipulative
force may be directed in a posteroan-
terior direction on one of several spi-
nal contacts, with the patient prone
on a segmented table or plinth (Fig.
3). However, the most common lum-
bar manipulation is done with the pa-
tient in a side-lying position (Fig. 4),
with the lower leg straight and the
upper leg bent at the hip and knee.
The upper body is braced by putting
one hand on the patient’s shoulder,
traction is placed on the uppermost
leg to produce rotational locking of
the facet joints, and a thrusting im-
pulse is administered at one of sev-
eral spinal contact points. Depending
on the findings in the postural and
palpatory examination, the manipu-
lative thrust can be applied to the
233
Spinal Manipulative Therapy for Low Back Pain
Figure 3 Prone, short-lever technique with
a primary vector of thrust from posterior to
anterior. Contact is directly on one of several
spinal contact points, and the thrust can be
delivered with torque, depending on the type
of movement desired.
spinous or transverse process, the
sacrum, the ilium, or, in the case of
long-lever techniques, the leg (Fig. 5).
It is not uncommon for spinal manip-
ulation to be applied to the thoracic
and cervical spine even when the pri-
mary symptoms are in the lumbar re-
gion. The rationale for this approach
is that there is interdependence be-
tween regions of the spine to allow
proper movement and that each area
contributes to correct posture and
motion. Although attractive, this rea-
soning is largely hypothetical.
Flexion-distraction treatment has
been recommended for patients with
disk herniation and has found favor
with many manipulative practi-
tioners. The patient is placed on a spe-
Figure 4 Side-lying, short-lever spinal ma-
nipulative technique.
234
cial table that introduces continuous
mobilization of the lumbar spine
while manual contact and force are
applied to introduce an element of di-
rected intersegmental lumbar trac-
tion. Although there is some evidence
that this approach can increase the
mobility of the lumbar spine and also
cause some expansion of the interver-
tebral disk in the same manner as oth-
er traction procedures, its efficacy has
yet to be demonstrated in random-
ized clinical trials.
Spinal manipulative treatment is
usually followed by postural and er-
gonomic advice, recommendations
for home exercise, and nutritional and
lifestyle recommendations. Many
Figure 5 Side-lying, long-lever spinal ma-
practitioners also offer supports or
nipulative technique.
orthotics in the hope of reducing pos-
tural factors that may cause recur-
rence of pain. Increasing numbers of nonsurgical treatments for low back
chiropractors and practitioners in in- pain.
terdisciplinary treatment centers are The primary difference between
incorporating spinal manipulation as the treatment protocols for chronic
part of a comprehensive rehabilita- low back pain and those for acute
tion process. pain is that, with chronic back pain,
A major topic of contention and a lesser intensity of initial treatment
confusion is the number of manipu- and a longer duration of overall ther-
lative treatments that are reasonable apy is common. Ergonomics, lifestyle,
and necessary to achieve optimal re- and exercise recommendations also
sults. The consensus is that some de- play a greater role in the treatment of
gree of improvement may be expect- patients with chronic back pain.
ed within 2 weeks of treatment at a The treatment of asymptomatic or
frequency of three to five treatments minimally symptomatic individuals
per week.48 Lack of improvement af- on a prophylactic or maintenance ba-
ter 4 weeks, or 12 treatments, is rea- sis has been controversial among
son to discontinue treatment until fur- practitioners of SMT in general and
ther examination of the patient has chiropractors in particular. Current-
been done.48 If the patient shows pro- ly, there is no evidence that prophy-
gressive improvement of symptoms lactic treatment either prevents or
at 4 weeks, treatments may be tapered modifies the occurrence of subse-
to two per week, then to one. How- quent back problems or the appear-
ever, most patients with acute or ep- ance or course of other disorders.
isodic uncomplicated low back pain
respond to just a few treatments and
tend to return only when there is re- Complications
currence of pain. Standard treatment
protocols suggest that patients should Serious complications caused by spi-
be reevaluated every 2 weeks to de- nal manipulation are rare and very
termine whether the treatment is suc- poorly documented.2 There have
cessful.48 Such protocols are not very been a few case reports of herniated
different from those for physical ther- intervertebral disks and even of cau-
apy, exercise, medication, and other da equina syndrome3 after manipu-
Journal of the American Academy of Orthopaedic Surgeons

lation. The precise incidence of these
complications is unknown and is
probably underreported, although
even the degree of underreporting is
not known. Shekelle et al3 estimated
the serious complication rate for lum-
bar manipulation at 1 in 100 million
manipulations, but that conclusion
was based solely on reported cases.
Senstad et al49 reported that minor,
self-limited side effects of manipula-
tion are quite common. The most typ-
ical reactions, usually lasting less than
24 hours but occasionally as long as
48, were local discomfort in the area
of treatment, pain in areas other than
that of treatment, fatigue, and head-
ache. None of the clinical trials of
lumbar manipulation reported any
complications, which indicates that
spinal manipulation is a relatively
safe form of lumbar spine treatment
and is considerably safer than many
of the medications that can be used
either alone or in combination with
SMT to manage back pain.
Contraindications
Physical force should not be placed
on a spine that cannot structurally
withstand it. Therefore, destructive
lesions of the spine, acute fracture, or
osteomyelitis are absolute contrain-
dications to lumbar manipulation.
Osteoporosis and bleeding disorders
are conditions that mandate caution,
especially when considering impulse
manipulation. Rheumatoid disease
with acute inflammation or ligamen-
References
1. Eisenberg DM, Kessler RC, Foster C,
Norlock FE, Calkins DR, Delbanco TL:
Unconventional medicine in the United
States: Prevalence, costs, and patterns
of use. N Engl J Med 1993;328:246-252.
2. Cherkin DC, Mootz RD (eds): Chiroprac-
tic in the United States: Training, Practice,
and Research. AHCPR Publication 98-
N002. Rockville, MD: US Department of
Health and Human Services, 1998.
Vol 11, No 4, July/August 2003
Rand Swenson, DC, MD, PhD, and Scott Haldeman, DC, MD, PhD, FRCPC
tous instability is also a contraindi-
cation to forceful manipulation, but
mobilization and massage may be
considered in specific situations. Pa-
tients who take medications that
could affect the appropriateness of
SMT, such as long-term glucocorti-
coids or anticoagulants, also may be
unsuitable.
The presence of disk herniation or
severe spondylosis often requires
modification of the treatment ap-
proach, but they are not absolute con-
traindications to spinal manipulation
in the absence of neurologic deficit.
Progressive neurologic deficits or
cauda equina syndrome are absolute
contraindications to spinal manipu-
lation. However, patients with stable,
long-standing, thoroughly evaluated
neural defects may be considered as
candidates.
Summary
Spinal manipulation is a common and
generally safe method of lower back
pain therapy. Research indicates that
SMT for acute and chronic back pain
provides at least short-term benefits.
A growing number of patients with
back pain are treated with manipu-
lation, and most express a high de-
gree of acceptance of and satisfaction
with such procedures. Because of this
patient satisfaction, as well as a grow-
ing awareness of the clinical trials sup-
porting the effectiveness of manipu-
lative procedures to treat specific
categories of patients with back pain,
3. Shekelle PG, Adams AH, Chassin MR,
Hurwitz EL, Brook RH: Spinal manip-
ulation for low-back pain. Ann Intern
Med 1992;117:590-598.
4. Shekelle PG, Brook RH: A community-
based study of the use of chiropractic ser-
vices. Am J Public Health 1991;81:439-442.
5. Bigos SJ (ed): Clinical Practice Guideline
Number 14: Acute Low Back Problems in
Adults. AHCPR Publication 94-0040.
SMT and its practitioners have been
incorporated into conventional med-
ical settings. This assimilation is ex-
pected to continue, and models of treat-
ment participation and cooperation
are being developed and refined. Ma-
nipulation therapies in general and chi-
ropractic treatments in particular are
now covered by Medicare and most
health insurance and HMO plans. The
primary motivation for inclusion has
been demand by patients and insur-
ance plan subscribers for such services.
The intense scrutiny of spinal ma-
nipulative procedures that has been
the hallmark of the past decade can
be expected to continue. With a more
established research infrastructure, it
should be easier to conduct studies
in multidisciplinary settings. Many
questions remain unanswered, such
as whether particular subgroups of pa-
tients with back pain are more likely
to benefit from manipulative treatments
and whether certain manipulative tech-
niques are markedly more efficacious
or hazardous than others. Identifica-
tion of the mechanism through which
manipulation affects symptoms is un-
der increasing investigation. Compar-
ative studies with other forms of treat-
ment, as well as studies to determine
whether there are added benefits of
combining treatment modalities, are
ongoing. Further research should de-
fine the exact role of spinal manipu-
lation in the treatment of patients with
low back pain, the optimal duration
and intensity of treatment, and the cost-
effectiveness of SMT relative to other
treatment options.
Rockville, MD: Agency for Health Care
Policy and Research, US Department of
Health and Human Services, 1994.
6. Peterson DH, Bergmann TF: Principles
of adjustive technique, in Peterson DH,
Bergmann TF: Chiropractic Technique:
Principles and Procedures, ed 2. St. Louis,
MO: Mosby, 2002, pp 97-174.
7. Harms MC, Innes SM, Bader DL: Forces
measured during spinal manipulative
235
Spinal Manipulative Therapy for Low Back Pain
procedures in two age groups. Rheuma-
tology (Oxford) 1999;38:267-274.
8. Herzog W, Zhang YT, Conway PJ, Kaw-
chuk GN: Cavitation sounds during spi-
nal manipulative treatments. J Manipu-
lative Physiol Ther 1993;16:523-526.
9. Gemmell HA, Jacobson BH: The imme-
diate effect of activator vs. meric adjust-
ment on acute low back pain: A ran-
domized controlled trial. J Manipulative
Physiol Ther 1995;18:453-456.
10. Gatterman MI (ed): What’s in a word?
in Gatterman MI (ed): Foundations of
Chiropractic: Subluxation. St. Louis, MO:
Mosby, 1995, pp 5-17.
11. Haldeman S: Neurologic effects of the
adjustment. J Manipulative Physiol Ther
2000;23:112-114.
12. Cramer GD, Tuck NR Jr, Knudsen JT, et
al: Effects of side-posture positioning
and side-posture adjusting on the lum-
bar zygapophysial joints as evaluated
by magnetic resonance imaging: A be-
fore and after study with randomiza-
tion. J Manipulative Physiol Ther 2000;23:
380-394.
13. Herzog W, Scheele D, Conway PJ: Elec-
tromyographic responses of back and
limb muscles associated with spinal
manipulative therapy. Spine 1999;24:
146-153.
14. Zhu Y, Haldeman S, Starr A, Seffinger
MA, Su SH: Paraspinal muscle evoked
cerebral potentials in patients with uni-
lateral low back pain. Spine 1993;18:
1096-1102.
15. Indahl A, Kaigle AM, Reikerås O, Holm
SH: Interaction between the porcine
lumbar intervertebral disc, zygapophy-
sial joints, and paraspinal muscles.
Spine 1997;22:2834-2840.
16. Suter E, McMorland G, Herzog W, Bray
R: Conservative lower back treatment re-
duces inhibition in knee-extensor mus-
cles: A randomized controlled trial. J Ma-
nipulative Physiol Ther 2000;23:76-80.
17. Cherkin DC, MacCornack FA: Patient
evaluations of low back pain care from
family physicians and chiropractors.
West J Med 1989;150:351-355.
18. Assendelft WJ, Bouter LM: Does the
goose really lay golden eggs? A meth-
odological review of Workmen’s Com-
pensation studies. J Manipulative Phys-
iol Ther 1993;16:161-168.
19. Stano M, Smith M: Chiropractic and
medical costs of low back care. Med
Care 1996;34:191-204.
20. Ottenbacher K, DiFabio RP: Efficacy of
spinal manipulation/mobilization ther-
apy: A meta-analysis. Spine 1985;10:
833-837.
21. Di Fabio RP: Efficacy of manual thera-
py. Phys Ther 1992;72:853-864.
236
22. Koes BW, Assendelft WJ, van der
Heijden GJ, Bouter LM: Spinal manip-
ulation for low back pain: An updated
systematic review of randomized clin-
ical trials. Spine 1996;21:2860-2873.
23. Brønfort G: Efficacy of spinal manipu-
lation and mobilisation for low back
and neck pain: A systematic review and
best evidence synthesis, in Efficacy of
Manual Therapies of the Spine. Amster-
dam, The Netherlands: Thesis Publish-
ers Amsterdam, 1997, pp 117-146.
24. van Tulder MW, Koes BW, Bouter LM:
Conservative treatment of acute and
chronic nonspecific low back pain: A
systematic review of randomized con-
trolled trials of the most common inter-
ventions. Spine 1997;22:2128-2156.
25. Anderson R, Meeker WC, Wirick BE,
Mootz RD, Kirk DH, Adams A: A meta-
analysis of clinical trials of spinal ma-
nipulation. J Manipulative Physiol Ther
1992;15:181-194.
26. Assendelft WJ, Koes BW, Knipschild
PG, Bouter LM: The relationship be-
tween methodological quality and con-
clusions in reviews of spinal manipula-
tion. JAMA 1995;274:1942-1948.
27. van Tulder MW, Waddell G: Conserva-
tive treatment of acute and subacute
low back pain, in Nachemson A, Jons-
son E (eds): Neck and Back Pain: The Sci-
entific Evidence of Causes, Diagnosis and
Treatment. Philadelphia, PA: Lippincott
Williams & Wilkins, 2000, pp 241-269.
28. van Tulder MW, Goosen M, Waddell G,
Nachemson A: Conservative treatment
of chronic low back pain, in Nachem-
son A, Jonsson E (eds): Neck and Back
Pain: The Scientific Evidence of Causes,
Diagnosis and Treatment. Philadelphia,
PA: Lippincott Williams & Wilkins,
2000, pp 271-304.
29. Glover JR, Morris JG, Khosla T: Back
pain: A randomized clinical trial of ro-
tational manipulation of the trunk. Br
J Ind Med 1974;31:59-64.
30. Hadler NM, Curtis P, Gillings DB, Stin-
nett S: A benefit of spinal manipulation
as adjunctive therapy for acute low-
back pain: A stratified controlled trial.
Spine 1987;12:702-706.
31. Mathews JA, Mills SB, Jenkins VM, et
al: Back pain and sciatica: Controlled
trials of manipulation, traction, sclero-
sant and epidural injections. Br J Rheu-
matol 1987;26:416-423.
32. MacDonald RS, Bell CM: An open con-
trolled assessment of osteopathic manip-
ulation in nonspecific low-back pain.
Spine 1990;15:364-370.
33. Koes BW, Bouter LM, van Mameren H,
et al: Randomised clinical trial of ma-
nipulative therapy and physiotherapy
Journal of the American Academy of Orthopaedic Surgeons
for persistent back and neck com-
plaints: Results of one year follow up.
BMJ 1992;304:601-605.
34. Pope MH, Phillips RB, Haugh LD,
Hsieh CY, MacDonald L, Haldeman S:
A prospective randomized three-week
trial of spinal manipulation, transcuta-
neous muscle stimulation, massage and
corset in the treatment of subacute low
back pain. Spine 1994;19:2571-2577.
35. Triano JJ, McGregor M, Hondras MA,
Brennan PC: Manipulative therapy ver-
sus education programs in chronic low
back pain. Spine 1995;20:948-955.
36. Meade TW, Dyer S, Browne W, Frank
AO: Randomised comparison of chiro-
practic and hospital outpatient man-
agement for low back pain: Results
from extended follow up. BMJ 1995;311:
349-351.
37. Cherkin DC, Deyo RA, Battié M, Street
J, Barlow W: A comparison of physical
therapy, chiropractic manipulation,
and provision of an educational book-
let for the treatment of patients with
low back pain. N Engl J Med 1998;339:
1021-1029.
38. Skargren EI, Carlsson PG, Oberg BE:
One-year follow-up comparison of the
cost and effectiveness of chiropractic
and physiotherapy as primary manage-
ment for back pain: Subgroup analysis,
recurrence, and additional health care
utilization. Spine 1998;23:1875-1883.
39. Hurwitz EL, Morgenstern H, Harber P,
et al: A randomized trial of medical care
with and without physical therapy and
chiropractic care with and without
physical modalities for patients with
low back pain: 6-month follow-up out-
comes from the UCLA low back pain
study. Spine 2002;27:2193-2204.
40. Coxhead CE, Inskip H, Meade TW,
North WR, Troup JD: Multicentre trial
of physiotherapy in the management of
sciatic symptoms. Lancet 1981;1:1065-
1068.
41. Nwuga VC: Relative therapeutic effica-
cy of vertebral manipulation and con-
ventional treatment in back pain man-
agement. Am J Phys Med 1982;61:273-278.
42. Mathews W, Morkel M, Mathews J:
Manipulation and traction for lumbago
and sciatica: Physiotherapeutic tech-
niques used in two controlled trials.
Physiother Pract 1988;4:201-206.
43. Siehl D, Olson DR, Ross HE, Rockwood
EE: Manipulation of the lumbar spine
with the patient under general anesthe-
sia: Evaluation by electromyography
and clinical-neurologic examination of
its use for lumbar nerve root compres-
sion syndrome. J Am Osteopath Assoc
1971;70:433-440.

44. Floman Y, Liram N, Gilai AN: Spinal
manipulation results in immediate
H-reflex changes in patients with uni-
lateral disc herniation. Eur Spine J 1997;
6:398-401.
45. Bronfort G, Haldeman S: Spinal manip-
ulation in patients with lumbar disc dis-
ease. Semin Spine Surg 1999;11:97-103.
46. Kirkaldy-Willis WH, Cassidy JD: Spinal
Vol 11, No 4, July/August 2003
Rand Swenson, DC, MD, PhD, and Scott Haldeman, DC, MD, PhD, FRCPC
manipulation in the treatment of low-
back pain. Can Fam Physician 1985;31:
535-540.
47. Hestbœk L, Leboeuf-Yde C: Are chiro-
practic tests for the lumbo-pelvic spine
reliable and valid? A systematic critical
literature review. J Manipulative Physiol
Ther 2000;23:258-275.
48. Haldeman S, Chapman-Smith D, Pe-
tersen DM Jr (eds): Guidelines for Chiro-
practic Quality Assurance and Practice Pa-
rameters: Proceedings of the Mercy Center
Consensus Conference. Gaithersburg, MD:
Aspen Publishers, 1993.
49. Senstad O, Leboeuf-Yde C, Borchgre-
vink C: Frequency and characteristics
of side effects of spinal manipulative
therapy. Spine 1997;22:435-441.
237
 Extensor Mechanism Failure Associated With Total Knee
Arthroplasty: Prevention and Management
David A. Parker, MBBS, Michael J. Dunbar, MD, and Cecil H. Rorabeck, MD

Abstract
Extensor mechanism complications are the most commonly reported reasons for re- onstrated contact stresses higher than
vision surgery after total knee arthroplasty and are a frequent source of postoper- the yield strength of polyethylene.
ative morbidity. Patellofemoral instability is the most commonly reported extensor Nevertheless, satisfactory long-term
mechanism complication and has multiple etiologies, including prosthetic malalign- results have been reported with both
ment and soft-tissue imbabalce. Patellar fracture or rupture of either the quadriceps conforming and dome-shaped com-
or patellar tendon can cause catastrophic disruption of the extensor mechanism. Al- ponents. Whichever style is used, it
though some stable fractures can be successfully managed nonsurgically, displaced is important that the patellar compo-
fractures or tendon rupture often lead to poor results. Other complications include nent be an appropriate match to the
patellar clunk and soft-tissue adhesions, prosthetic wear or loosening, and osteone- femoral component throughout the
crosis. Increased understanding of implant alignment, rotation, and soft-tissue bal- full range of motion.5,6
ance, as well as improved design of the trochlear groove of femoral implants and The high failure rate of metal-
patellar components, has resulted in a decline in extensor mechanism complications. backed prostheses during the 1980s
Appropriate prosthetic selection and meticulous surgical technique remain the keys led to the development of all-
to avoiding unsatisfactory results and revision surgery. polyethylene patellar components
J Am Acad Orthop Surg 2003;11:238-247 (Fig. 1). Failure of the metal-backed
implants was predominantly caused
by wear of the thin polyethylene with
resultant metallosis and, less com-
As many as 50% of revision total knee Initial experiences with unresurfaced monly, by loosening of the uncement-
arthroplasty (TKA) procedures have patellae revealed a high rate of per- ed implants. Also, patellar implants
been attributed to patellofemoral com- sistent patellofemoral discomfort should have multiple small pegs be-
plications.1 The incidence of these com- among patients with rheumatoid ar- cause a single large peg produces a
plications has decreased markedly as thritis or osteoarthritis, which stim-
surgical techniques and component ulated the development of a patellar
design have been refined. Earlier implant. Based on an anatomic study Dr. Parker is Orthopaedic Fellow, University Hos-
studies2-4 reported incidences of 10% of 80 arthritic patellas, Aglietti et al7 pital, University of Western Ontario, London,
ON, Canada. Dr. Dunbar is Assistant Professor,
to 35%; more recent studies5,6 cite rates initially described a dome-shaped pa-
Department of Orthopaedics, QE II Health Sci-
of 1% to 12%. However, patellofemo- tellar prosthesis for use with TKA. ences Centre, Dalhousie University, Halifax, NS,
ral complications continue to be a sig- Subsequently, both symmetric dome- Canada. Dr. Rorabeck is Professor, Department
nificant source of postoperative mor- shaped and asymmetric, conforming, of Orthopaedic Surgery, University Hospital, Uni-
bidity and revision surgery. Possible anatomic-shaped designs have be- versity of Western Ontario.
complications include patellofemoral come widely used. Hsu and Walker8
One or more of the authors or the departments
instability, extensor mechanism dis- reported that increased conformity of with which they are affiliated has received some-
ruption, soft-tissue impingement, the patellar component decreased the thing of value from a commercial or other party
prosthetic wear or loosening, and os- predicted amount of deformation and related directly or indirectly to the subject of this
teonecrosis. wear, whereas Matsuda et al9 found article.
that conforming patellas had higher
Reprint requests: Dr. Rorabeck, 339 Windermere
contact stresses than did dome- Road, London, ON N6A 5A5, Canada.
Prosthetic Design shaped implants, which reduced the
load sharing between the patellar Copyright 2003 by the American Academy of
Early prosthetic designs did not al- component and the quadriceps ten- Orthopaedic Surgeons.
low the option of patellar resurfacing. don. Both types of component dem-

238 Journal of the American Academy of Orthopaedic Surgeons


Figure 1 Three types of patellar implant components. Left, All-polyethylene component
with single central peg. Center, All-polyethylene component with three pegs. Right, Metal-
backed component.
greater increase in patellar strain and
may cause fracture.2
Femoral component design is also
an important factor in patellofemo-
ral function. Kirk et al10 compared two
different prosthetic designs and con-
cluded that the higher incidence of pa-
tellofemoral complications with a spe-
cific implant was caused by the less
anatomic design of the femoral com-
ponent. Mont et al6 attributed the elim-
ination of patellofemoral complications
primarily to improvements in femo-
ral component design, especially in
relation to patellofemoral contact. Early
improvements in the femoral compo-
nent included the addition of a troch-
lear flange and a more anatomic troch-
lear groove that improves patellar
tracking. A broad, deep trochlear
groove is important to accommodate
the patellar component congruently
throughout the full range of motion.
Eckhoff et al11 showed that the trochlear
groove lies lateral to the midplane be-
tween the femoral condyles along a
line between the anatomic and me-
chanical axes of the femur. Therefore,
the component groove should extend
far enough distally to contain the pa-
tella in deep flexion. Such prosthetic
design improvements have positively
affected patellofemoral function after
TKA. Improved surgical technique also
has helped decrease the incidence of
complications, as has greater appreci-
ation of the importance of femoral and
tibial component rotation, joint line
alteration, and soft-tissue balancing.
Although both anatomic and clin-
ical studies support the use of an
Vol 11, No 4, July/August 2003
asymmetric femoral component, Har-
win5 reported only a 0.6% rate of sub-
luxation using a symmetric femoral
component design. No advantage has
been shown to using an asymmetric
trochlear groove and/or asymmetric
component; in fact, surgical technique
may be the most critical factor.12 In a
patient whose patella is not resur-
faced, the femoral component should
be compatible with the native patel-
la, including a long, deep trochlear
groove. However, no designs have
been able to exactly reproduce nor-
mal kinematics, and no femoral com-
ponent has been designed specifical-
ly for the native patella. Thus, the
longevity of the articulation between
the native patella and the prosthetic
trochlea requires further study.
Patellofemoral Instability
Patellofemoral instability has been one
of the most common reasons for re-
vision surgery after TKA (Fig. 2), with
Figure 2 Patellar instability 3 years after initial TKA. A, Sunrise radiograph of a lateral sub-
luxation of an unresurfaced patella. B, Sunrise view of a lateral dislocation of a resurfaced
patella.
David A. Parker, MBBS, et al
a reported incidence of subluxation
and/or dislocation as high as 27% in
earlier series.3,4 Recent studies5,6 have
shown the benefit of improved sur-
gical technique and prosthetic design,
with reported incidences of sublux-
ation and/or dislocation ≤1%. Insta-
bility may be related to several fac-
tors, including femoral and patellar
component design, as well as mal-
alignment of any of the three compo-
nents, malrotation of the femoral or
tibial component, or soft-tissue imbal-
ance. Instability also may be related
to overstuffing of the patellofemoral
joint or to asymmetric resection of the
patella. Preoperative patellar sublux-
ation or dislocation should be assessed
because it may indicate an anatomic
imbalance that would predispose to
postoperative instability, which should
be addressed intraoperatively.
Patellar component positioning
has an important effect on patellar
tracking. In a cadaveric model, cen-
tral placement achieved optimal pa-
tellofemoral mechanics, and medial
placement produced increased patel-
lar tilt.13 However, most authors5,6 still
advocate some medialization to avoid
a laterally placed component, which
would increase the risk of sublux-
ation. Reproducing the normal patel-
lar thickness is important because ex-
cessive resection can predispose to
fracture, and inadequate resection
will result in limited flexion and pa-
tellar maltracking. An oversized or
anteriorly positioned femoral compo-
nent also can lead to increased patel-
239
Extensor Mechanism Failure Associated With TKA
lofemoral pressure. Asymmetric pa-
tellar resection frequently leads to
patellar tilt and instability.
Femoral and tibial component
alignment also is critical. Correct ax-
ial alignment is important for all as-
pects of TKA function; in an ideal
alignment, the weight-bearing line is
placed at the center of the knee joint.
Knees with fixed preoperative valgus
or varus alignment can develop post-
operative patellofemoral instability if
the deformity remains uncorrected.
Lynch et al2 found an increased inci-
dence of subluxation in patients with
preoperative valgus alignment that
was not corrected intraoperatively.
Kirk et al14 reported on a series of 15
knees with patellar dislocation after
TKA, of which 13 had preoperative
varus alignment. Preservation of the
level of the joint line also has been
shown to be important in decreasing
patellofemoral complications by
avoiding elevation of the joint line
and patella baja.15
Femoral and tibial component ro-
tation is critical to patellar stability.
Internal rotation of the femoral com-
ponent will medialize the trochlear
groove relative to the patella, result-
ing in lateral subluxation. Similarly,
excessive internal rotation of the tib-
ial component will cause lateraliza-
tion of the tibial tubercle, also result-
ing in lateral patellar subluxation.
Berger et al16 found a direct correla-
tion between combined internal ro-
tation of the two components and the
severity of patellofemoral complica-
tions. Akagi et al17 found improved
patellar tracking and less frequent lat-
eral retinacular releases done in knees
with femoral components externally
rotated to the posterior condylar axis
compared with those parallel to the
axis.
Femoral component rotation can be
assessed in several ways. Techniques
for cutting the femur parallel to the
transepicondylar axis, perpendicular
to the anteroposterior axis, or exter-
nally rotated 3° to the posterior condy-
lar axis have been described,18 and
240
each of these axes should be assessed
(Fig. 3). The posterior condyles also
can be cut parallel to the tibial cut with
the flexion gap tensioned and the knee
at 90° of flexion. When using the pos-
terior condylar axis, it is important to
assess for lateral condylar wear and
hypoplasia, particularly in the valgus
knee, because these can lead to inter-
nal rotation of the femoral component.
Poilvache et al18 found the transepi-
condylar axis to be the most reliable
guide; this axis also has been shown
to most closely approximate the flex-
ion axis of the knee. In practice, it is
prudent to observe all of the available
anatomic landmarks; the transepi-
condylar axis is probably the best
guide if significant disparity exists.
Tibial component rotation also is
important, particularly in more con-
forming designs. The center of the
component should be in line with the
medial third of the tibial tubercle or
internally rotated 18° compared with
the plane of the tubercle.16 Sufficient
surgical exposure of the entire tibial
plateau is necessary to visualize land-
marks for accurate assessment of com-
ponent alignment. Excessive internal
rotation must be avoided because it
will lateralize the tubercle. Placement
of femoral and tibial components in
the coronal plane also is important.
Medial placement of the femoral com-
ponent medializes the trochlear
groove, and medial placement of the
tibial component lateralizes the tuber-
cle; both can cause lateral subluxation
of the patella. Although these com-
ponents generally should be centered
over the condyles, lateralization of
both components is indicated if the
patient’s anatomy cannot be perfect-
ly matched.
Assessment of patellar tracking
should be done intraoperatively with
trial implants and after implantation
of definitive implants. Use of a thigh
tourniquet can alter patellofemoral
tracking; if maltracking exists, the
tourniquet should be deflated and
tracking reassessed before lateral ret-
inacular release. The patella should
Journal of the American Academy of Orthopaedic Surgeons
Figure 3 Axes of femoral component rota-
tion. APA= anteroposterior axis, TEA= trans-
epicondylar axis, TFG = tensioned flexion gap,
PCA = posterior condylar axis. (Reprinted
with permission from Lonner JH, Lotke PA:
Aseptic complications after total knee arthro-
plasty. J Am Acad Orthop Surg 1999;7:311-
324.)
independently track centrally within
the trochlear groove without any tilt-
ing or subluxation. This overesti-
mates the need for lateral release be-
fore closure of the knee joint because
of the detachment of the vastus me-
dialis muscle and medial retinaculum
during exposure. Therefore, it is ap-
propriate to place a single suture or
towel clip in the extensor mechanism
when testing tracking intraoperative-
ly to better mimic the postoperative
setting. If tilting or subluxation re-
mains, all facets of component align-
ment should be checked and, if nec-
essary, corrected (Table 1).
If there is no identifiable compo-
nent malalignment or malrotation,
then the etiology most likely relates
to soft-tissue imbalance. A tight lat-
eral retinaculum can lead to sublux-
ation and should be addressed with
a lateral release. Lynch et al2 advised
doing retinacular release well lateral
to the patella to avoid the circum-
patellar anastomosis. To avoid in-
creasing the risk of quadriceps ten-
don rupture, the release should not
extend medially at its proximal ex-

Table 1
Potential Problems in the Presence of Patellar Subluxation Before
Considering Lateral Retinacular Release
Area Potential Problems
Femoral component Internally rotated
Medially translated
Oversized
Tibial component Internally rotated
Medially translated
Patellar component Original thickness not
reproduced (over-
stuffed)
Tourniquet Quadriceps restricted
tent, and division of the vastus lat-
eralis insertion to the superolateral
patella should be avoided. If this does
not correct the problem, the proximal
realignment can be completed by ad-
vancing the vastus medialis muscle
as part of the closure. Lateral release
and proximal realignment usually
correct most degrees of patellar sub-
luxation. If frank dislocation or per-
sistent subluxation is not corrected by
proximal realignment, tibial tubercle
malposition may be the cause, and a
tubercle transfer may be required.
Earlier studies indicated a high com-
plication rate with distal realignment
and recommended proximal realign-
ment only, cautioning against the use
of tubercle transfer.19 However, in lat-
er studies, satisfactory correction of
patellar maltracking was achieved
with this technique with minimal
complications.14,20 Kirk et al14 report-
ed 15 cases of patellar dislocation af-
ter TKA that were all successfully
treated using a modification of the
Trillat procedure. The osteotomy
should consist of a long fragment and
should be tapered distally to avoid a
stress riser. Fixation can be done with
either screws or wires (Fig. 4). Trans-
fer of the tibial tubercle is rarely nec-
essary, and correct component align-
ment and soft-tissue balance should
ensure good patellar tracking in most
cases.
Vol 11, No 4, July/August 2003
David A. Parker, MBBS, et al
Goldberg et al23 described a classifi-
cation for patellar fractures compli-
cating TKA based on the integrity of
the extensor mechanism and fixation
Solutions of the patellar implant. They found
that fractures not associated with
Revise position componentloosening,extensormech-
Revise position anism disruption, or major malalign-
Downsize component
ment generally had good results with
Revise position
nonsurgical management. Other frac-
Revise position
tures required surgery, with a high
Measure thickness
before resection and proportion of unsatisfactory results.
reproduce with implant Most fractures usually are vertical
Release and recheck (similar to stress fractures), without
tracking disruption of the extensor mecha-
nism, and often are incidental find-
ings that require no specific manage-
ment. In most series3,5,6,24 of
Extensor Mechanism postoperative patellar fractures, more
Disruption than half were managed nonsurgical-
ly with good or excellent results.
Patellar Fracture Intraoperative fractures are more
Patellar fracture is an uncommon common in revision surgery but can
complication of TKA (Fig. 5). Al- occur in primary cases, especially if
though the reported incidence rang- the patella is particularly thin. It is im-
es from 0%4 to 6%,21 most series2,5,6,22 portant to avoid overreaming or ec-
report an incidence of 1% to 2%, with centric reaming, overcompression of
occurrence more common in men the patellar clamps during reaming,
than in women. Such fractures usu- and slippage of the reamer. Resurfac-
ally are described on the basis of lo- ing patellas that are <10 mm thick re-
cation, integrity of the extensor mech- quires extra care to avoid fracture; im-
anism, and stability of the implant. plants with increased thickness can
Figure 4 Tibial tubercle osteotomy used for exposure during revision of failed TKA caused
by metal-backed patellar implant wear with extensive metallosis. A, Long tapered fragment
with intact lateral soft tissues. B, Wires through the medial tibial cortex, with proximal wire
through the tubercle and the distal two wires around the tubercle. C, Wires tightened and
tubercle secured.
241
Extensor Mechanism Failure Associated With TKA

implants23). Technical factors include

Figure 5 Patellar fractures. A, Sunrise radiograph of a vertical fracture with stable implant.
B, Sunrise view of a vertical fracture with unstable implant. C, Lateral view of a transverse
fracture with unstable implant.
help minimize resection and restore
patellar height in such cases.25
Postoperative fractures may be
nontraumatic or traumatic. Nontrau-
matic fractures may be associated
with any of several possible risk fac-
tors. Resurfacing of the patella has
been shown to be associated with a
higher rate of fracture.3,26 Other risk
factors include those related to the pa-
tient (eg, osteoporosis1) and implant
(eg, central peg design,25 cementless
excessive or inadequate resection,
devascularization of the patella, and
patellar subluxation.26 Malalignment
of the limb and implant malrotation
also have been associated with in-
creased risk of patellar fracture.16,27
Nontraumatic fractures are usually
associated with one or several of these
risk factors, especially osteonecrosis.
The role of lateral retinacular release
in nontraumatic fractures is unclear.
Healy et al22 found increased inci-
dence of fracture after lateral release,
whereas Ritter and Campbell,24 in a
large series, did not. The technical
procedures that add to the risk of os-
teonecrosis also increase the risk of
fracture (eg, quadriceps turndown).
Traumatic postoperative fractures
may result from a direct trauma or an
indirect cause, such as eccentric quad-
riceps muscle contraction resulting in
a proximal avulsion.
Management depends on the frac-
ture pattern, stability of the implant,
and integrity of the extensor mech-
anism (Fig. 6). Many fractures are
asymptomatic and occur with stable
implants and an intact extensor
mechanism. Such fractures can be

Patellar fracture

Transverse Vertical

Stable implant Unstable implant Stable implant Unstable implant

Extensor Extensor Extensor Extensor Extensor Extensor Extensor Extensor

mechanism mechanism mechanism mechanism mechanism mechanism mechanism mechanism
intact disrupted intact disrupted intact disrupted intact disrupted

Closed Open Remove Remove Closed Open Remove Remove

treatment fixation prosthesis prosthesis treatment repair prosthesis prosthesis
• Cylinder • Wires or and internal • Early motion • Early motion and
• Splint • Screws Revise fixation or internal fixation
Patellectomy or Patellectomy

Figure 6 London Health Sciences Centre algorithm for the management of patellar fractures.

242 Journal of the American Academy of Orthopaedic Surgeons


managed nonsurgically, usually with
good or excellent results. Fractures
with extensor mechanism disruption
or dislocation and/or unstable im-
plants require surgical treatment and
have poor results in more than half
of cases.23,25 Alignment of the implant
also is important because knees with
major malalignment have more se-
vere fractures and the poorest out-
comes.23,27
Quadriceps Tendon Rupture
Quadriceps tendon rupture is an
extremely rare complication.5,6,22
Lynch et al2 reported an incidence of
1.1% in a series of 281 TKAs; other
authors26 have presented single case
reports. With such small numbers, the
etiology can only be speculated on
and could include factors such as
overresection of the patella with dam-
age to the quadriceps tendon as well
as vascular injury and incomplete
healing after extended approaches
such as V-Y turndown, manipulation,
or trauma, particularly if there is pre-
existing tendon degeneration. In the
study by Lynch et al,2 all three pa-
tients with quadriceps tendon rup-
ture had a lateral release, possibly in-
dicating reduced vascularity as an
etiologic factor. Anterior extension of
lateral release also may contribute to
rupture and should be avoided.
Treatment of the rupture requires
direct repair of the tendon. Rupture
usually occurs near the distal inser-
tion; thus, the tendon can be repaired
directly to bone via drill holes or su-
ture anchors using nonabsorbable su-
tures. Suture anchors are preferable
if patellar bone stock is limited be-
cause drill holes may compromise the
implant. Supplementation with an al-
lograft may be required if there is any
deficiency in the extensor mechanism
that would prevent satisfactory re-
pair. The knee is held in full exten-
sion for 6 weeks postoperatively be-
fore beginning gradual restoration of
motion, with the priority being du-
rable healing of the repair rather than
rapid restoration of full preoperative
Vol 11, No 4, July/August 2003
movement. Unfortunately, surgical
repair has had largely unsatisfactory
results in the small series reported,
with persistent extensor lag and lim-
ited range of motion.2
Patellar Tendon Rupture
Patellar tendon rupture is also un-
common, but it is more frequently re-
ported than quadriceps tendon rup-
ture. Although earlier published
incidences ranged from 0.2% to 5%,2,28,29
reported incidences in later series were
<1%,5,6 suggesting a decreasing fre-
quency of rupture as surgical technique
improves. As with quadriceps tendon
rupture, the low frequency makes
study of etiology difficult. However,
rupture should be preventable. In gen-
eral, the patients most at risk are those
with multiply operated knees. Possi-
ble specific etiologies include a stiff
knee that causes difficulty everting the
patella during exposure; trauma with
hyperflexion, including postoperative
manipulation; multiple procedures
with subsequent devascularized tis-
sue; and patient factors (eg, chronic
steroid use, systemic disease).2,30 Dis-
tal realignment procedures have been
implicated,29 as has excessive patel-
lar resection with damage to the ex-
tensor mechanism.31 Component mal-
alignment27 and hinged implants4 also
are thought to place increased stress
on the extensor mechanism.
Rupture can occur intraoperative-
ly, in the immediate postoperative
period, or as a delayed complication.
Intraoperative avulsion can occur
during exposure of a stiff knee if ex-
cessive force is applied to the tendon
attachment while the patella is evert-
ed and the knee flexed. Rupture can
occur in the early postoperative pe-
riod, such as during manipulation, at
any stage as a result of trauma, or as
a delayed complication because of
chronic attrition, such as may occur
with impingement against the tibial
insert.31
Preventing this complication re-
quires vigilance during exposure
when everting the patella and flexing
David A. Parker, MBBS, et al
the knee. If this is difficult, ancillary
methods should be used. Careful pos-
teromedial dissection from the tibia
is important to allow external rotation
of the tibia, which markedly decreas-
es tension at the tendon insertion. If
excessive tension remains, a quadri-
ceps snip can be done; in revision sur-
gery, a lateral release in association
with division of the lateral gutter and
infrapatellar scar tissue also is useful.
Occasionally, a tubercle osteotomy or
a quadriceps turndown may be nec-
essary.
Treatment of tendon rupture can
be difficult and has generally had un-
satisfactory results, with few patients
regaining full active extension or a
satisfactory degree of flexion.29,30,32
Repair can be primary, with or with-
out autograft augmentation, or with
allograft reconstruction. If there is
partial avulsion or avulsion with an
intact periosteal sleeve, the tendon
can be reattached primarily to bone
either through drill holes or with su-
ture anchors33 or staples.29 This repair
can be augmented using a semiten-
dinosus or gracilis tendon autograft
left attached distally, particularly if
the quality of the primary repair is
poor or if there is soft-tissue defect.30
Primary repair of late ruptures gen-
erally has had poor results.29 Abril et
al34 reported two cases of tendon rup-
ture 1 month after TKA, with success-
ful primary repair through drill holes
and support by a figure-of-8 wire for
3 months. These results are difficult
to reproduce with a direct repair, and
a more extensive reconstruction is
usually indicated.
Cadambi and Engh30 described a
technique in which a semitendinosus
tendon autograft is left attached dis-
tally and used to augment the patel-
lar tendon by passing it along the me-
dial border of the tendon through a
drill hole in the patella and suturing
it to itself distally. Although the mean
extensor lag was 10° and flexion only
79°, the authors concluded that this
technique was superior to primary re-
pair or allograft reconstruction.
243
Extensor Mechanism Failure Associated With TKA
Allograft reconstruction was first
described by Emerson et al,31 who
used an allograft of quadriceps ten-
don, a patella with a cemented pros-
thesis, the patellar tendon, and the
tibial tubercle. The tibial tubercle was
attached to the tibia with two screws
and a tension band wire. The quad-
riceps tendon allograft was then
placed on slight tension and attached
to host tendon, with the patella main-
tained in the appropriate position on
the femoral component. One third of
patients had extensor lag ranging
from 20° to 40°, and there was a high
complication rate. However, the au-
thors still considered this technique
a satisfactory option for extensor
mechanism deficiency. They suggest-
ed that resurfacing of the allograft pa-
tella was unnecessary.
In a larger series, Nazarian and
Booth35 described a modification of
this technique using fresh-frozen al-
lograft with unresurfaced patellas
and the allograft tensioned with the
knee in full extension. The mean ex-
tensor lag was 13° in 15 of 36 patients
and the mean flexion, 98° in all pa-
tients. Although the authors quoted
a success rate of 34 of 36 patients, 8
required repeat allograft and 12, a
walking aid. Leopold et al32 also re-
ported a high rate of failure using the
technique described by Emerson et
al,31 with progressive extensor lag and
dependence on walking aids. They
suggested that improvements were
required to tension the graft intraop-
eratively and that alternative tech-
niques should be considered. Other
techniques have been described,
including the use of a medial gastroc-
nemius flap,36 use of synthetic liga-
ment augmentation,37 and patello-
tibial fusion,38 all with small numbers,
short follow-up, notable complica-
tions, and persistent extensor lag.
Delayed rupture of the patellar
tendon also can be repaired using an
Achilles tendon allograft. The al-
lograft, attached to a fragment of cal-
caneus, is inset into the tibia and fixed
in that position. The Achilles tendon
244
is split and wound in a figure-of-8
fashion, through either the extensor
mechanism or the patella. The former
approach minimizes the risk of fur-
ther devitalizing the blood supply to
the patella. The allograft is then su-
tured back onto itself with the knee
in full extension. Our experience with
this technique in seven patients has
been positive in terms of restoring ex-
tensor function to the point that a
brace is not required. Most patients
are left with an extensor lag, but this
has not proved to be problematic.
Patellar Clunk and
Soft-Tissue Impingement
Patellar clunk is a well-recognized
complication of posterior stabilized
TKA, with a reported incidence of up
to 3.5%.39 It is caused by a prolifer-
ation of synovial and fibrous tissue
at the superior pole of the patella at
the quadriceps tendon insertion. This
proliferation of tissue is itself caused
by articulation of the region with the
sharp anterior flange of the inter-
condylar notch in flexion. Presum-
ably that, in turn, causes an inflam-
matory reaction that subsequently
leads to development of a fibrous
nodule in the notch when the knee
is flexed. When the nodule is of suf-
ficient size, painful dislodgement oc-
curs as the knee is actively extended
from a flexed position. The dislodge-
ment usually occurs at approximate-
ly 30° of flexion and causes the pain-
ful clunk for which the syndrome is
named. Symptoms usually present a
mean of 1 year after the procedure.39
Prosthetic design seems to be the
main risk factor for patellar clunk.
Most newer prostheses have a small-
er box with a deeper patellar groove
and a more posterior position of the
femoral cam, thereby decreasing the
chance of soft tissue articulating with
this region. The incidence of patellar
clunk with these newer prostheses
seems to be greatly reduced, although
longer follow-up is necessary. Preven-
Journal of the American Academy of Orthopaedic Surgeons
tion may be helped by excision of the
synovium on the posterior aspect of
the quadriceps tendon in this region.
Although a trial of nonsurgical
management may be undertaken,
most patients with an established
clunk require surgery to resolve the
symptoms. Excellent results with ar-
throscopic resection through a supero-
lateral portal have been described,39
although care must be taken to avoid
scratching the femoral component or
damaging the patellar polyethylene.
An arthrotomy to remove this tissue
is a simple procedure with a relative-
ly rapid recovery time, especially if
significant adhesions are expected to
make arthroscopic visualization dif-
ficult or if other problems with the
prosthesis must be addressed. Recur-
rence after successful removal is rare.
Other complications related to ab-
normal soft-tissue formation have
been published. Thorpe et al40 report-
ed 11 patients (in a series of 635 ar-
throplasties) who had painful patel-
lofemoral dysfunction caused by
intra-articular peripatellar fibrous
bands. Nine of 11 implants were sta-
bilized posteriorly and, although the
etiology of the pathology is unknown,
all patients had resolution of symp-
toms after arthroscopic removal of the
bands.
Patellar Component Wear
and Loosening
Patellar component wear is usually
secondary to either maltracking or im-
plant design.4 Markedly higher fail-
ure rates exist with polyethylene im-
plants with metal-backed components
than with cemented all-polyethylene
components.22 However, it is unusu-
al for wear of an all-polyethylene com-
ponent to be sufficient to require
revision surgery. Although metal
backing of the patella improves load
distribution, such implants have had
a high rate of failure,41 primarily be-
cause the thin polyethylene rapidly
wears and delaminates (Fig. 7), leav-

Figure 7 Failed metal-backed patellar implant. Left, Burnished metal backing of patellar
component from articulation with femoral component. Center, Worn delaminated polyeth-
ylene. Right, Marked burnishing (arrow) of femoral component.
ing a metal-on-metal articulation. Of-
ten, revision of the entire TKA is re-
quired. Although some surgeons have
had good results with metal-backed
implants, most now advocate use of
a cemented all-polyethylene compo-
nent.
Patellar component loosening is rare.
Incidence of up to 2% has been not-
ed,42 but no cases of loosening have
been reported in most recent large se-
ries using cemented all-polyethylene
implants.5,6 However, loosening has
been a reported problem with cement-
less, metal-backed implants. Healy et
al22 found an increased rate of loos-
ening in cementless implants and also
reported loosening secondary to os-
teonecrosis. Patients with high activ-
ity levels and good range of motion
are thought to be at increased risk, as
are patients with malpositioned com-
ponents or those with small central
fixation lugs. Loosening also can oc-
cur secondary to fracture, maltrack-
ing, and osteolysis; in such cases, the
underlying problem requires manage-
ment. Improvements in femoral com-
ponent rotation and femoral trochlear
design should help decrease the prob-
lem of patellar loosening.
Osteonecrosis
The vascular supply to the patella has
been well described in anatomic stud-
Vol 11, No 4, July/August 2003
ies that have demonstrated extensive
extraosseous and intraosseous sys-
tems with contributions from all ge-
nicular vessels. The extraosseous ves-
sels form an anastomotic ring, which
is damaged to some extent during ar-
throplasty. The standard medial para-
patellar approach divides the three
medial contributors; lateral meniscec-
tomy and lateral release can divide
the two lateral contributors; and ex-
cision of the infrapatellar fat pad can
damage the inferior part of the ring.
Patellar resurfacing can cause dam-
David A. Parker, MBBS, et al
age to the intraosseous supply, put-
ting the patella at risk for osteonecro-
sis. Scuderi et al43 reported decreased
patellar vascularity on bone scan af-
ter lateral release, but follow-up stud-
ies showed possible revascularization
within 60 days. In a large series of
TKAs, Ritter and Campbell24 report-
ed no increase in osteonecrosis in pa-
tients who had a lateral release. Healy
et al,22 in one of the few series about
osteonecrosis in TKA, reported an in-
cidence of 1.4% in 211 TKAs. Failure
to recognize this complication is prob-
ably the cause of the low reported in-
cidence. The highest incidence is in
patients who required a quadriceps
turndown procedure for exposure in
revision surgery. In our experience,
radiographic evidence of osteonecro-
sis with sclerosis and flattening or
fragmentation developed in 8 of 29
TKAs in which quadriceps turndown
procedures were done as part of the
revision procedure (Fig. 8).
The natural history of osteonecro-
sis is poorly defined, except in symp-
tomatic cases in which the sclerotic
appearance and secondary fracture or
fragmentation can be associated with
prosthetic loosening. Prevention
Figure 8 Patellar osteonecrosis after
quadriceps turndown. A, Lateral radio-
graph demonstrating sclerosis and flatten-
ing of the patella. B, Sunrise radiograph
showing patellar flattening, fragmenta-
tion, and lateral subluxation.
245
Extensor Mechanism Failure Associated With TKA
should include avoiding the turn-
down approach and, theoretically,
minimizing lateral releases and fat
pad resection. Use of the subvastus
approach also preserves most of the
medial supply, although these mea-
sures remain to be proved as effec-
tive osteonecrosis prevention tech-
niques. Management of established
osteonecrosis involves treatment of
secondary complications. If asymp-
tomatic, nonsurgical management is
indicated, whereas patellar fragmen-
tation and prosthetic loosening neces-
sitate removal of the implant and
loose bony fragments. The remaining
patella should be preserved as much
as possible, although the prognosis is
guarded.
Management of the Failed
Patellar Component
Appropriate management of a failed
patellar component depends primar-
ily on the cause of failure. If a com-
ponent requires revision because of
wear or loosening, considerations
should include techniques for pros-
References
1. Brick GW, Scott RD: The patellofemo-
ral component of total knee arthroplas-
ty. Clin Orthop 1988;231:163-178.
2. Lynch AF, Rorabeck CH, Bourne RB:
Extensor mechanism complications fol-
lowing total knee arthroplasty. J Arthro-
plasty 1987;2:135-140.
3. Cameron HU, Fedorkow DM: The pa-
tella in total knee arthroplasty. Clin
Orthop 1982;165:197-199.
4. Mochizuki RM, Schurman DJ: Patellar
complications following total knee ar-
throplasty. J Bone Joint Surg Am 1979;61:
879-883.
5. Harwin SF: Patellofemoral complica-
tions in symmetrical total knee arthro-
plasty. J Arthroplasty 1998;13:753-762.
6. Mont MA, Yoon TR, Krackow KA, et al:
Eliminating patellofemoral complica-
tions in total knee arthroplasty: Clinical
and radiographic results of 121 consec-
utive cases using the Duracon system.
J Arthroplasty 1999;14:446-455.
246
thesis removal, evaluation of the pre-
dicted remaining bone stock, the con-
dition of the remainder of the
extensor mechanism, and the state of
the femoral component.
If technically possible, revision of
the patellar component is preferred
because it restores the extensor mech-
anism and provides better pain relief
than does patellectomy or patello-
plasty. However, when there is insuf-
ficient bone stock to seat a compo-
nent, the chance of failure is high and
such revision should not be undertak-
en. As mentioned, at least 10 mm of
residual bone is required for resurfac-
ing; the patella should be left unre-
surfaced when <10 mm of bone re-
mains. In such a situation, to avoid
further loss of tension of the exten-
sor mechanism, resection arthroplas-
ty and patelloplasty with reshaping
of the residual patella to match the
femoral component is preferable to a
patellectomy. Complete revision and
synovectomy is necessary if a failed
metal-backed component has caused
damage to the femoral component
with subsequent metallosis.41 How-
ever, Barrack et al44 found that retain-
7. Aglietti P, Insall JN, Walker PS, et al: A
new patella prosthesis: Design and ap-
plication. Clin Orthop 1975;107:175-187.
8. Hsu HP, Walker PS: Wear and deforma-
tion of patellar components in total
knee arthroplasty. Clin Orthop 1989;246:
260-265.
9. Matsuda S, Ishinishi T, White SE, et al:
Patellofemoral joint after total knee ar-
throplasty: Effect on contact area and con-
tact stress. J Arthroplasty 1997;12:790-797.
10. Kirk PG, Rorabeck CH, Bourne RB:
Clinical comparison of the Miller
Galante I and AMK total knee systems.
J Arthroplasty 1994;9:131-136.
11. Eckhoff DG, Burke BJ, Dwyer TF, et al:
Sulcus morphology of the distal femur.
Clin Orthop 1996;331:23-28.
12. Bindelglass DF, Dorr LD: Symmetry
versus asymmetry in the design of to-
tal knee femoral components: An unre-
solved controversy. J Arthroplasty 1998;
13:939-944.
Journal of the American Academy of Orthopaedic Surgeons
ing well-fixed, undamaged, well-
aligned patellar components at the
time of revision surgery for other
components resulted in equivalent
outcome to that achieved with suc-
cessful reimplantation.
Summary
Extensor mechanism failure is the pri-
mary reason for revision TKA. Ad-
vances in prosthetic design and sur-
gical technique have led to a marked
decrease in the incidence of such com-
plications, but they continue to be a
notable source of morbidity and un-
satisfactory results. Because many
complications are difficult to manage
and often have relatively poor results,
prevention is the cornerstone of man-
agement. Most of these complications
can be avoided with appropriate
prosthetic selection and attention to
detail in surgical technique. When
management is contemplated, a sys-
tematic assessment of the specific
causative factors of the complication
should be done to determine and ap-
ply the appropriate treatment.
13. Lee TQ, Budoff JE, Glaser FE: Patellar
component positioning in total knee ar-
throplasty. Clin Orthop 1999;366:274-281.
14. Kirk P, Rorabeck CH, Bourne RB, et al:
Management of recurrent dislocation of
the patella following total knee arthro-
plasty. J Arthroplasty 1992;7:229-233.
15. Figgie HE III, Goldberg VM, Heiple
KG, et al: The influence of tibial-
patellofemoral location on function of
the knee in patients with the posterior
stabilized condylar knee prosthesis.
J Bone Joint Surg Am 1986;68:1035-1040.
16. Berger RA, Crossett LS, Jacobs JJ, et al:
Malrotation causing patellofemoral
complications after total knee arthro-
plasty. Clin Orthop 1998;356:144-153.
17. Akagi M, Matsusue Y, Mata T, et al: Ef-
fect of rotational alignment on patellar
tracking in total knee arthroplasty. Clin
Orthop 1999;366:155-163.
18. Poilvache PL, Insall JN, Scuderi GR, et
al: Rotational landmarks and sizing of

the distal femur in total knee arthro-
plasty. Clin Orthop 1996;331:35-46.
19. Grace JN, Rand JA: Patellar instability
after total knee arthroplasty. Clin
Orthop 1988;237:184-189.
20. Whiteside LA: Distal realignment of the
patellar tendon to correct abnormal pa-
tellar tracking. Clin Orthop 1997;344:
284-289.
21. Clayton ML, Thirupathi R: Patellar com-
plications after total condylar arthroplasty.
Clin Orthop 1982;170:152-155.
22. Healy WL, Wasilewski SA, Takei R, et
al: Patellofemoral complications follow-
ing total knee arthroplasty: Correlation
with implant design and patient risk fac-
tors. J Arthroplasty 1995;10:197-201.
23. Goldberg VM, Figgie HE III, Inglis AE,
et al: Patellar fracture type and progno-
sis in condylar total knee arthroplasty.
Clin Orthop 1988;236:115-122.
24. Ritter MA, Campbell ED: Postoperative
patellar complications with or without
lateral release during total knee arthro-
plasty. Clin Orthop 1987;219:163-168.
25. Bourne RB: Fractures of the patella af-
ter total knee replacement. Orthop Clin
North Am 1999;30:287-291.
26. Grace JN, Sim FH: Fracture of the pa-
tella after total knee arthroplasty. Clin
Orthop 1988;230:168-175.
27. Figgie HE III, Goldberg VM, Figgie MP,
et al: The effect of alignment of the im-
plant on fractures of the patella after
condylar total knee arthroplasty. J Bone
Joint Surg Am 1989;71:1031-1039.
28. Lettin AW, Kavanagh TG, Scales JT:
Vol 11, No 4, July/August 2003
The long-term results of Stanmore total
knee replacements. J Bone Joint Surg Br
1984;66:349-354.
29. Rand JA, Morrey BF, Bryan RS: Patellar
tendon rupture after total knee arthro-
plasty. Clin Orthop 1989;244:233-238.
30. Cadambi A, Engh GA: Use of a semi-
tendinosus tendon autogenous graft
for rupture of the patellar ligament af-
ter total knee arthroplasty: A report of
seven cases. J Bone Joint Surg Am 1992;
74:974-979.
31. Emerson RH Jr, Head WC, Malinin TI:
Extensor mechanism reconstruction
with an allograft after total knee arthro-
plasty. Clin Orthop 1994;303:79-85.
32. Leopold SS, Greidanus N, Paprosky WG,
et al: High rate of failure of allograft re-
construction of the extensor mechanism
after total knee arthroplasty. J Bone Joint
Surg Am 1999;81:1574-1579.
33. Sinha RK, Crossett LS, Rubash HE: Ex-
tensor mechanism disruption after total
knee arthroplasty, in Insall JN, Scott
WN (eds): Surgery of the Knee, ed 3. New
York, NY: Churchill Livingstone, 2001,
vol 2, pp 1863-1873.
34. Abril JC, Alvarez L, Vallejo JC: Patellar
tendon avulsion after total knee arthro-
plasty: A new technique. J Arthroplasty
1995;10:275-279.
35. Nazarian DG, Booth RE Jr: Extensor
mechanism allografts in total knee arthro-
plasty. Clin Orthop 1999;367:123-129.
36. Jaureguito JW, Dubois CM, Smith SR, et
al: Medial gastrocnemius transposition
flap for the treatment of disruption of
David A. Parker, MBBS, et al
the extensor mechanism after total knee
arthroplasty. J Bone Joint Surg Am 1997;
79:866-873.
37. Aracil J, Salom M, Aroca JE, et al: Ex-
tensor apparatus reconstruction with
Leeds-Keio ligament in total knee arthro-
plasty. J Arthroplasty 1999;14:204-208.
38. Kempenaar JW, Cameron JC: Patel-
lotibial fusion for patellar tendon rup-
ture after total knee arthroplasty. J Ar-
throplasty 1999;14:115-117.
39. Lucas TS, DeLuca PF, Nazarian DG, et
al: Arthroscopic treatment of patellar
clunk. Clin Orthop 1999;367:226-229.
40. Thorpe CD, Bocell JR, Tullos HS: Intra-
articular fibrous bands: Patellar com-
plications after total knee replacement.
J Bone Joint Surg Am 1990;72:811-814.
41. Bayley JC, Scott RD, Ewald FC, et al: Fail-
ure of the metal-backed patellar compo-
nent after total knee replacement. J Bone
Joint Surg Am 1988;70:668-674.
42. Colizza WA, Insall JN, Scuderi GR: The
posterior stabilized total knee prosthe-
sis: Assessment of polyethylene dam-
age and osteolysis after a ten-year min-
imum follow-up. J Bone Joint Surg Am
1995;77:1713-1720.
43. Scuderi G, Scharf SC, Meltzer LP, et al:
The relationship of lateral releases to
patella viability in total knee arthro-
plasty. J Arthroplasty 1987;2:209-214.
44. Barrack RL, Rorabeck C, Partington P,
et al: The results of retaining a well-
fixed patellar component in revision to-
tal knee arthroplasty. J Arthroplasty
2000;15:413-417.
247
 Magnetic Resonance Imaging
of the Pediatric Spine
A. Jay Khanna, MD, Bruce A. Wasserman, MD,
and Paul D. Sponseller, MD

Abstract
Magnetic resonance is an excellent modality for imaging the pediatric spine. Its suc- anatomic detail, including that of os-
cessful use requires understanding both the basic physics and the sedation protocols seous structures, disk, and soft tissues.
necessary for acquiring high-resolution images. Interpreting the images accurately T2-weighted images are used primar-
depends on appreciating the differences between the normal anatomy of the pedi- ily to evaluate the spinal cord and to
atric and the adult spine. Evaluating the images requires familiarity with the dif- enhance lesion conspicuity. Agradient-
ferential diagnosis of pediatric spine disease, including the most common processes recalled echo sequence typically is used
(infections, neoplasms, and trauma) as well as spinal dysraphism. Despite the ac- when thin axial images are needed,
knowledged usefulness of magnetic resonance imaging of the pediatric spine, con- such as for evaluating foraminal nar-
troversies remain related to its safety in this age group and its limitations in di- rowing in the cervical spine, because
agnosing and evaluating scoliosis and tethered cord syndrome. its three-dimensional acquisition al-
J Am Acad Orthop Surg 2003;11:248-259 lows for very thin sections.
Standard pulse sequences for spi-
nal imaging include spin echo T1-
weighted images and fast spin echo
Magnetic resonance is an excellent MRI Techniques (FSE) T2-weighted images. The FSE
modality for imaging pathologic technique allows acquisition of scans
processes in the pediatric spine. It The major factors that influence the without prolonged imaging times. Be-
allows high-resolution views of not MRI appearance of various tissues are cause cerebrospinal fluid (CSF) is bright
only osseous structures (including the density of protons in the tissue, on T2-weighted images and the spi-
the vertebral body, spinal canal, and the chemical environment of the pro- nal cord retains its intermediate sig-
posterior elements) but also soft-tis- tons, and the magnetic field strength nal, the images maximize the contrast
sue structures (including the spinal of the scanner. Unlike computed to- between CSF and neural tissue, allow-
cord, intervertebral disk, and nerve mography (CT), which produces im- ing optimal delineation of the spinal
roots). Magnetic resonance imaging ages based on the density of various cord and nerve roots. T2-weighted im-
(MRI) can show these structures in tissues, MRI produces images based ages are very sensitive to pathologic
various planes using different pulse on free water content and on other changes in tissue, including any pro-
sequences that allow optimal char- magnetic properties of water, yield-
acterization of the tissues in and ing superior soft-tissue contrast.
around the pediatric spine. Indica- Various sequences are produced by Dr. Khanna is Chief Resident, Department of Or-
tions for MRI in children (<18 years) manipulating the strength of the ra- thopaedic Surgery, The Johns Hopkins Hospital,
Baltimore, MD. Dr. Wasserman is Assistant Pro-
are gradually expanding as technol- diofrequency (RF) pulses, the inter-
fessor, Department of Radiology, The Johns Hop-
ogy improves. Properly interpreting val between the pulses, the repetition kins Hospital. Dr. Sponseller is Professor and Vice
MRI scans in these age groups de- time (TR), and the echo time (TE), that Chairman, Department of Orthopaedic Surgery,
pends on understanding the MRI is, the time between applying the RF The Johns Hopkins Hospital.
appearance of the normal pediatric pulse and measuring the signal emit-
Reprint requests: Dr. Sponseller, c/o Elaine P.
spine anatomy at various stages of ted by the patient. By manipulating
Henze, Room A672, 4940 Eastern Avenue, Bal-
development. For entities such as these variables, the images can be timore, MD 21224-2780.
spinal dysraphism, left thoracic weighted to emphasize the T1, T2,
curves, and juvenile scoliosis, spe- gradient-recalled echo, or proton den- Copyright 2003 by the American Academy of
cific recommendations can help cli- sity characteristics of a tissue. T1- Orthopaedic Surgeons.
nicians use MRI effectively. weighted images allow evaluation of

248 Journal of the American Academy of Orthopaedic Surgeons


cesses in which cells and the extra-
cellular matrix have an increase in wa-
ter content. This pathologic change is
usually shown as an increase in sig-
nal intensity on T2-weighted images.
The signal from fat may be sup-
pressed by a variety of techniques, in-
cluding chemical saturation of its sig-
nal or application of an inversion
pulse, and imaging at a short time of
inversion (TI) when there is no fat sig-
nal present (short TI recovery [STIR]).
Chemical suppression typically is
used in sequences that result in high
fat signal, such as FSE T2-weighted
images or postcontrast T1-weighted
images. Fat suppression is of little val-
ue for noncontrast T1-weighted im-
ages because the signal from most
pathologic lesions, whether inflam-
matory, neoplastic, or infectious, is of-
ten low and better visualized because
of contrast against the adjacent bright
fat signal. Fat suppression on post-
contrast T1-weighted images of the
vertebral body is useful in adults who
have fatty transformation of marrow.
Fat-suppressed images may be par-
ticularly useful for evaluating liga-
mentous injuries or lesions involving
the paraspinal tissues. The usefulness
of STIR imaging is more limited be-
cause the imaging parameters are re-
stricted and cannot be optimized to
maximize contrast between adjacent
tissues of interest.
Gradient-recalled echo images ap-
pear to be T2-weighted because CSF
is relatively bright; however, paren-
chymal lesions typically are more con-
spicuous on FSE T2-weighted images.
The gradient-recalled echo sequence
is sensitive to local inhomogeneities
of the magnetic field, and signal loss
is exaggerated in the presence of such
inhomogeneities. Field inhomogene-
ities may be caused by metallic im-
plants (eg, pedicle screws or paraspi-
nal rods), differences in the magnetic
susceptibilities of adjacent tissues (eg,
air-tissue interfaces), and paramagnetic
substances (eg, gadolinium). Blood-
breakdown products cause local field
distortions resulting in signal loss, mak-
Vol 11, No 4, July/August 2003
ing this technique very sensitive for
the detection of blood.
Open MRI systems are being used
more frequently, especially for chil-
dren. These systems have notably
lower field strengths than do closed
systems and therefore usually pro-
duce studies of inferior overall qual-
ity, especially of the spine. However,
open MRI systems allow easier access
to the sedated or otherwise compro-
mised patient. Young patients and pa-
tients with claustrophobia have ac-
cess to parents and the environment,
making the procedure less intimidat-
ing. However, whenever possible,
spinal MRI should be done using
closed, 1.5-T systems.
Pediatric Sedation
Protocols
Sedation is often required for success-
ful MRI in young children. Many
studies have evaluated specific seda-
tion protocols.1,2 The American Acad-
emy of Pediatrics (AAP) has pub-
lished guidelines for the elective
sedation of pediatric patients,3,4 but
compliance with these guidelines is
not mandatory. The AAP has stated
that careful medical screening and pa-
tient selection by knowledgeable
medical personnel are needed to ex-
clude patients at high risk of life-
threatening hypoxia.4 Also, monitor-
ing using AAP guidelines is necessary
for the early detection and manage-
ment of life-threatening hypoxia.3 The
AAP recommends that before an ex-
amination in which sedation is to be
used, children from newborn to age
3 years take nothing by mouth for 4
hours and those aged 3 to 6 years take
nothing by mouth for 6 hours.4
Pediatric sedation practices vary,
but a few agents are common to most
protocols. Oral chloral hydrate is of-
ten recommended for children young-
er than 18 months. However, its use
is controversial because of its variable
absorption, paradoxical effects, and
nonstandardized dosing. Older chil-
A. Jay Khanna, MD, et al
dren usually receive intravenous pen-
tobarbital with or without fentanyl.
Although studies have reported
successful administration of sedatives
by trained nurses,1,2 an anesthesiol-
ogist’s expertise can be beneficial for
patients with substantial comorbidi-
ties, including cardiopulmonary dis-
ease, skeletal dysplasias, neuromus-
cular disease, and abnormal airway
anatomy. Because of the potential risks
of anesthesia and sedation in children,
there is a trend toward referring those
who require sedation to hospitals with
pediatric anesthesiologists.
An important consideration after
sedation for pediatric MRI is the need
for strict adherence to established dis-
charge criteria, including return to
baseline vital signs, level of con-
sciousness close to baseline, and abil-
ity to maintain a patent airway.5 Be-
cause of the inherent risks of sedation,
alternative techniques have been de-
vised, including sleep deprivation
and rapid, segmental scanning. The
latter permits acquisition of high-
quality images without the use of se-
dation.
Normal MRI Anatomy
Appreciating normal MRI anatomy
(Fig. 1) is essential for understanding
and predicting the MRI appearance
of pathologic processes.6
Adolescents and Adults
The lumbar spine is more fre-
quently imaged than the cervical and
thoracic area in both children and
adults. In adolescents and adults, the
lumbar spinal canal appears round
proximally and triangular distally.
The lumbar facet joints, best visual-
ized in the axial plane, are covered
with 2 to 4 mm of hyaline cartilage.
This cartilage can be well visualized
with FSE and gradient-recalled echo
pulse sequences. The epidural space
and ligaments also should be evalu-
ated carefully. Epidural fat is seen as
high signal intensity on T1-weighted
249
Magnetic Resonance Imaging of the Pediatric Spine

Figure 1 A, Sagittal T1-weighted MRI scan of a normal lum-

bar spine in a 2-year-old boy shows the rectangular shape of
the vertebral bodies. The conus medullaris is seen at the L1-L2
level (arrow). B, T2-weighted image shows the long, thin ap-
pearance of the intervertebral disk. C, Sagittal T1-weighted scan
of a normal lumbar spine in a 10-year-old girl. D, T2-weighted
scan. Lordosis is normal. The posterior elements are well
formed, with a resultant decrease in the canal diameter. E, Sag-
ittal T1-weighted scan of a normal lumbar spine in a 16-year-
old girl shows dark CSF (thin arrow), the conus medullaris at
the L1-L2 level (open arrow), and the basivertebral channel
(arrowhead). Note the normal rectangular appearance of the
vertebral bodies and the lumbar lordosis compared with the
10-year-old girl. F, Sagittal T2-weighted scan shows bright CSF
(thin arrow) and a bright nucleus pulposus (arrowhead).

images; the ligamentum flavum
shows minimally higher T1-weighted
signal compared with the other lig-
aments. The conus medullaris is usu-
ally located at the L1-L2 level. The tra-
versing nerve roots pass distally from
the conus medullaris and extend an-
teriorly and laterally, exiting lateral-
ly underneath the pedicle and extend-
ing into the neural foramen. The
intervertebral disk, consisting of the
cartilaginous end plates, anulus fibro-
sus, and nucleus pulposus, normally
shows increased T2-weighted signal
in its central portion. CSF, well im-
aged as low T1-weighted and high
T2-weighted signal, often can be used
to determine the type of pulse se-

250 Journal of the American Academy of Orthopaedic Surgeons


quence that is being used. CSF pul-
sations often create artifacts that de-
grade the image in the lumbar spine;
these artifacts must not be mistaken
for a pathologic process.
The cervical spine shows a mild
lordosis on sagittal images. On axial
images, the spinal canal is triangular,
with the base located anteriorly. A
dark band at the base of the dens is
a normal variant that is a remnant of
the subdental synchondrosis and
should not be mistaken for a fracture.
In adults, the facet joints are small and
triangular, whereas in children they
are large and flat. The spinal cord is
elliptical in cross section in the cer-
vical spine. There is a difference in sig-
nal between the normal gray and
white matter of the spinal cord. This
signal heterogeneity should not be
mistaken for intramedullary pathol-
ogy. The intervertebral disks are sim-
ilar in appearance to, but smaller
than, those seen at the thoracic and
lumbar levels. An important anatom-
ic feature of the cervical spine is the
prominent epidural venous plexus,
which is not present in the thoracic
or lumbar spine.
The thoracic vertebral bodies are
relatively constant in size, and the spi-
nal canal is almost round. Abundant
epidural fat is present posteriorly, but
there is less anteriorly than in the lum-
bosacral region. The cord is more
round than in the cervical or lumbar
regions, and the cord segment lies
two to three levels above the corre-
sponding vertebral body. The inter-
vertebral disks are thinner than the
disks in the lumbar spine. The ap-
pearance of the CSF is more variable
in the thoracic spine than in the lum-
bar region because of more prominent
CSF pulsations, but on T1-weighted
images, it is commonly seen as a re-
gion of low signal dorsal to the spi-
nal cord. This artifact is often most se-
vere at the apex of curves, including
the thoracic kyphosis. Certain tech-
niques can minimize this artifact, in-
cluding gating to the pulse or cardi-
ac cycle.
Vol 11, No 4, July/August 2003
Children
Differences Between the Pediatric and
the Adult Spine
The MRI appearance of the grow-
ing spine is complex. Substantial
changes occur in the vertebral ossi-
fication centers and the intervertebral
disks, changing the overall appear-
ance of the spine markedly, especial-
ly between infancy and age 2 years.7
In general, the vertebral ossification
centers are incompletely ossified ear-
ly in childhood, and the disks are
thicker and have a higher water con-
tent than those in adults. The spinal
canal and neural foramina are larger,
and there is less curvature. In addi-
tion, the overall signal intensity of the
vertebral bodies is lower than that of
the adult spine on T1-weighted im-
ages because of the abundance of red
(hematopoietic) marrow relative to
yellow (fat) marrow in the pediatric,
adolescent, and young adult spine.
Full-Term Infant
In the newborn, the overall size of
the vertebral body is small relative to
the spinal canal, and the spinal cord
ends at approximately the L2 level.
The lumbar spine does not exhibit the
usual lordosis and is straight. The ver-
tebral bodies show a markedly low
signal intensity on T1-weighted im-
ages, with a thin, central, hyperin-
tense band that likely represents the
basivertebral plexus. The spongy
bone of the ossification center is el-
lipsoid rather than rectangular and
often mistaken for disk. The interver-
tebral disk is relatively narrow and
often contains a thin, bright central
band on T2-weighted images that
represents the notochordal rem-
nants.6,7
Age 3 Months
At age 3 months, the osseous com-
ponent of the vertebral body has in-
creased and the amount of hyaline
cartilage has decreased, giving the
vertebral bodies a rectangular appear-
ance. The ossification centers begin to
A. Jay Khanna, MD, et al
gain in signal intensity, starting at the
end plates and progressing centrally.
The neural foramina have not sub-
stantially changed at this age, remain-
ing relatively large and ovoid.6,7
Age 2 Years
At age 2 years, the spine has be-
gun to show its normal sagittal align-
ment, most likely because of weight
bearing (Fig. 1, A and B). The ossified
portion of the vertebral body increas-
es substantially and begins to assume
its adult appearance, with near-
complete ossification of the pedicles
and the articular processes. The disk
space and nucleus pulposus become
longer and thinner. The cartilaginous
end plate has decreased in size and
is often difficult to identify. The neu-
ral foramen also begins to take its
adult appearance as its inferior por-
tion narrows.7
Age 10 Years
At age 10 years, sagittal alignment
resembles that of an adult (Fig. 1, C
and D). Ossification of the vertebral
bodies and posterior elements is near-
ly complete, with a resultant decrease
in the spinal canal diameter. The ver-
tebral bodies also develop concave
superior and inferior contours. The
nucleus pulposus becomes smaller at
this age and spans approximately half
the disk space in the sagittal plane.
The neural foramina continue to nar-
row inferiorly.6
The Conus Medullaris
In early fetal life, the spinal cord
extends to the inferior aspect of the
bony spinal column.6 Because the ver-
tebral bodies grow more rapidly lon-
gitudinally than the spinal cord does,
by birth the conus medullaris is re-
positioned in the upper lumbar spine.
It is important to note the location of
the conus medullaris on every pedi-
atric spine MRI study (Fig. 1, A and
E). Aconus medullaris level below the
L2-3 interspace in children older than
5 years is abnormal and indicates pos-
sible tethering.8,9 Saifuddin et al10 re-
251
Magnetic Resonance Imaging of the Pediatric Spine
viewed the MRI findings in 504 nor-
mal adult spines and found that the
average position of the conus med-
ullaris was the lower third of L1
(range, middle third of T12 to upper
third of L3).
Pathologic Processes in the
Pediatric Spine
Infection
Infectious processes involving the
pediatric spine include osteomyelitis,
diskitis, and epidural and paraspinal
abscess.11-13 In general, the MRI sig-
nal characteristics of infection include
a region of low T1 and high T2 sig-
nal intensity in bone and soft tissue.
In identifying vertebral osteomy-
elitis, MRI is more sensitive than con-
ventional radiographs or CT and
more specific than nuclear scintigra-
phy.14,15 Marrow edema can be detect-
ed on precontrast, fat-suppressed,
FSE T2-weighted images. Postgado-
linium enhancement of the disk and
adjacent vertebral bodies on postcon-
trast, fat-suppressed, T1-weighted
images helps confirm the diagnosis.
The specificity of MRI for infection is
higher in children than adults because
one of the primary confounders, de-
generative arthritis, is not part of the
differential diagnosis. Differentiating
osteomyelitis from neoplastic disease
is a common dilemma; generally, in-
fectious processes are more likely to
cross and destroy intervertebral disks
than are neoplastic conditions.
Diskitis is seen as a disruption of
the normally well-defined disk-
vertebral borders on T1-weighted im-
ages and as an increase in signal of
the disk on T2-weighted images.12 On
T2-weighted images, diskitis may
obliterate the normally seen horizon-
tal cleft within the intervertebral disk.
The abnormal signal seen in infec-
tious diskitis is associated classically
with surrounding soft-tissue inflam-
mation and reactive end-plate chang-
es. Primary diskitis is more likely to
develop in children than adults be-
252
cause of the greater blood supply to
the disk. Secondary diskitis after dis-
kography or surgery is more likely to
develop in adults.
Epidural abscesses are rare, but when
they do develop, it is usually after sur-
gery or vertebral osteomyelitis. Epi-
dural abscesses are diagnosed based
on the MRI findings of a collection in
the epidural space and the appropri-
ate clinical setting.11 Gadolinium-
enhanced T1-weighted images often
show a peripheral rim of enhancement
that represents the abscess wall.
Paraspinal abscesses occur adja-
cent to the spinal column, most com-
monly in the paraspinal musculature.
They may be secondary to a primary
infection in the spine or may arise
spontaneously in the paraspinal mus-
culature. These abscesses may be seen
as retropharyngeal abscesses in the
cervical spine, paraspinous or retro-
mediastinal abscesses in the thoracic
spine, or psoas abscesses in the lum-
bar spine. The MRI characteristics of
paraspinal abscesses include a well-
defined wall and peripheral enhance-
ment on postgadolinium, T1-
weighted images.
Trauma
MRI can be used to evaluate the
pediatric spinal trauma victim who
has an abnormal neurologic exami-
nation or is unresponsive. The patient
is first evaluated with conventional
radiographs, which may be normal,
even in a child with a neural deficit.
Although CT allows for better eval-
uation of osseous detail and displaced
fractures, MRI provides improved
evaluation of nondisplaced fractures
because of its ability to detect
marrow-signal abnormalities.
Spinal cord injury without radio-
graphic abnormality (SCIWORA) is
a well-defined entity seen in the pe-
diatric age group.16,17 The character-
istic hypermobility and ligamentous
laxity of the pediatric bony cervical
and thoracic spine predispose children
to this type of injury.16 The elasticity
of the bony pediatric spine and the
Journal of the American Academy of Orthopaedic Surgeons
relatively large size of the head allow
the musculoskeletal structures to de-
form beyond physiologic limits, which
results in cord trauma followed by
spontaneous reduction of the spine.16
As with other types of spinal cord
injuries, the most important predic-
tor of outcome is the severity of neu-
rologic injury. A patient with a com-
plete neurologic deficit after SCIWORA
has a poor prognosis for recovery of
neurologic function. The role of MRI
in SCIWORA syndrome is to define
the location and the degree of neural
injury, rule out occult fractures and
subluxation that may require surgi-
cal intervention, and evaluate for the
presence of ligamentous injury. T2-
weighted images typically show in-
creased signal in the cord, vertebral
body, or ligaments. The increased T2
signal in the cord is compatible with
edema and can range from a partial,
reversible contusion to complete
transection of the cord.
Two other traumatic entities can oc-
cur in children, usually as the result
of participation in sports. The first is
acute disk herniation. This is often a
fracture with a hingelike displacement
of fibrocartilage and slipping of the
entire disk with vertebral end-plate
fracture rather than extrusion of a disk
fragment from the nucleus, as is seen
in adults.18 Such avulsion fractures are
often occult on conventional radio-
graphs and are better detected with
CT and MRI.18 Axial MRI scans dem-
onstrate the fracture fragment as an
area of low signal intensity protrud-
ing into the spinal canal, and sagittal
images demonstrate a low signal in-
tensity region in the shape of a Y or
7 on all pulse sequences.18
The second entity is a spondylo-
lysis as a cause of back pain in young
athletes. MRI, however, is not the op-
timal method for evaluating spondy-
lolysis. CT offers increased spatial res-
olution and the ability to accurately
define the osseous defect, whereas ra-
dionuclide imaging can demonstrate
increased radiotracer activity in the
region of the defect.
 A. Jay Khanna, MD, et al

Neoplasms
MRI is the modality of choice for
evaluating neoplasms in and around
the pediatric spine.19 An effective and
commonly used approach is to clas-
sify the lesion as extradural, intra-
dural-extramedullary (Fig. 2), or in-
tradural-intramedullary (Fig. 3). With
this anatomic classification system,
the primary role of the MRI exami-
nation is to define the location of the
suspected neoplasm, which is best
achieved with axial and sagittal T1-
and T2-weighted images. Once the
lesion has been classified, the T2-
weighted images can be used to char-
acterize the lesion further. Specifical-
ly, the degree of surrounding edema
and tissue infiltration and the pres-
ence or absence of a cystic component Figure 2 A schwannoma in an 8-year-old boy. A, Sagittal T1-weighted MRI scan shows an
intradural-extramedullary mass impressing on the anterior cervical cord at the C5 level (ar-
can be determined. Next, postgado- row). B, Axial T2-weighted image shows the lesion herniating through the right C5-C6 neu-
linium enhancement images should ral foramen (arrows).
be compared with unenhanced T1-
weighted images. The final step in ob-
taining a diagnosis is to correlate the within the neural arch with partial or diagnosis then can be made based on
imaging findings with the patient’s complete absence of the spinous pro- the lesion’s MRI characteristics.
age and other criteria to narrow the cesses, laminae, or other components Myelomeningocele is the most
differential diagnosis. of the posterior elements. MRI has common form of spinal dysraphism
been shown to be the best modality (Fig. 4). It usually presents in the lum-
for evaluating spinal dysraphism.20,21 bosacral region (although it can be
Spinal Dysraphism A classification system has been seen at higher levels) as a back mass
proposed for evaluating a patient not covered with skin. The mass may
Spinal dysraphism is a general term with a suspected spinal dysraphism or may not be covered by lepto-
used to describe a wide range of (Table 1).21 The differential diagnosis meninges containing a variable
anomalies resulting from incomplete can be narrowed to one of three types: amount of neural tissue. The sac her-
fusion of the midline mesenchyma, spinal dysraphism with a back mass niates through a defect in the poste-
bone, and neural elements. The os- either covered or not covered with rior elements of the spine. The spinal
seous abnormalities consist of defects skin, or with no back mass. The final cord usually contains a dorsal cleft,

Figure 3 An astrocytoma in a 6-year-old boy. A, Sagittal T1-weighted MRI scan shows an intradural-intramedullary lesion within the spi-
nal cord at the T3-T5 levels (arrow). B, Sagittal T2-weighted image shows the partially cystic nature of the lesion. C, Axial T2-weighted
image confirms that the lesion (arrow) is within the center of the spinal cord.

Vol 11, No 4, July/August 2003 253

Magnetic Resonance Imaging of the Pediatric Spine
is splayed open, and is often tethered
within the sac.21 Progressive scolio-
sis is seen in 66% of patients with my-
elomeningocele, Arnold-Chiari type
II malformation in 90% to 99%, di-
astematomyelia in 30% to 40%, and
syringohydromyelia in 40% to
80%.22 Scarring can occur at the sur-
gical site after sac closure, so it is im-
portant to monitor these patients for
signs and symptoms of tethered cord
syndrome.
Of the entities presenting with a
skin-covered back mass in the pres-
ence of spinal dysraphism, lipomen-
ingocele is the most common.6,21 The
lipomeningocele consists of lipoma-
tous tissue that is continuous with the
subcutaneous tissue of the back and
also insinuates through the dysraph-
ic defect and dura and into the spi-
nal canal. The spinal cord often con-
tains a dorsal defect at the level of the
lipomatous tissue and may be teth-
ered at this level. The essential MRI
feature of this lesion is that the li-
pomatous tissue follows the signal
characteristics of subcutaneous fat on
all pulse sequences, including fat-
suppressed pulse sequences.
Occult spinal dysraphism pre-
sents without a back mass. Diastema-
tomyelia is characterized by a sagit-
tal splitting into two segments of the
spinal cord, conus medullaris, or
Figure 4 A myelomeningocele in a 6-year-old girl. A, Sagittal T1-weighted MRI scan shows a low-back mass contiguous with the contents
of the spinal canal (arrows). B, T2-weighted image shows that the mass is filled with high-signal-intensity fluid, compatible with CSF (ar-
rows). C, Axial T1-weighted image confirms that the mass communicates with the spinal canal through a defect in the posterior elements
(arrows).
254
Table 1
Classification of Spinal Dysraphism
Category
Back mass not covered with skin
Back mass covered with skin
No back mass (occult)
(Adapted with permission from Byrd SE, Darling CF, McLone DG, Tomita T: MR im-
aging of the pediatric spine. Magn Reson Imaging Clin North Am 1996;4:797-833.)
filum terminale, often in the thoracic
or lumbar spine. The dural tube and
arachnoid are undivided in approx-
imately half these patients; clinical
findings are rare, and surgery is not
indicated. In the remaining patients,
the dural tube and arachnoid are
completely or partially split at the
level of the spinal cord cleft, which
results in tethering of the cord and
subsequent clinical symptoms. Coro-
nal T1- and T2-weighted images
best define the sagittal split in the
Journal of the American Academy of Orthopaedic Surgeons
Types
Myelomeningocele
Myelocele
Lipomyelomeningocele
Myelocystocele
Simple posterior meningocele
Diastematomyelia
Dorsal dermal sinus
Intradural lipoma
Tight filum terminale
Anterior sacral meningocele
Lateral thoracic meningocele
Hydromyelia
Split notochord syndrome
Caudal regression syndrome
cord; the findings should be con-
firmed on axial images.
Another entity often seen in pa-
tients with spinal dysraphism is sy-
ringohydromyelia, or a syrinx (Fig. 5).
A syrinx is a longitudinal cavity with-
in the spinal cord that may or may
not communicate with the central ca-
nal. Attempts to explain the etiology
include developmental, traumatic, in-
flammatory, ischemic, and pressure-
related causes. Sagittal MRI scans
show a linear, low T1 and high T2 sig-

Figure 5 A large syrinx involving the entire spine in a 2-year-old boy. A, Sagittal T1-weighted MRI scan shows the syrinx to be largest
at the level of the lower thoracic spine (arrows). Axial T1-weighted (B) and T2-weighted (C) images confirm that the syrinx is located within
the center of the spinal cord.
nal intensity within the parenchyma
of the spinal cord.
Gibbs artifact, or truncation arti-
fact, can mimic a syrinx on sagittal
images (Fig. 6). Gibbs artifact is seen
on sagittal T1- and T2-weighted im-
ages as a linear region of altered sig-
nal intensity in the center of the spi-
nal cord. Thus, it is important to
Vol 11, No 4, July/August 2003
evaluate serial axial T1- and T2-
weighted images to confirm findings.
Gibbs artifact results from not using
a sufficiently high spatial frequency
for sampling data. It can be corrected
by using a higher-resolution matrix.
Chiari Malformations
Chiari malformations are seen fre-
A. Jay Khanna, MD, et al
quently in patients with spinal dys-
raphism. Chiari type I malformations
consist of cerebellar tonsillar ectopia,
in which the cerebellar tonsils extend
below the level of the foramen mag-
num. The common measurement for
the degree of herniation of the ton-
sils below the foramen magnum is 5
mm. Mikulis et al23 reported a vari-
Figure 6 A 5-year-old girl had a history of
neck and arm pain. A, Sagittal T2-weighted
MRI scan shows a long linear region of high
signal intensity within the center of the cer-
vical spinal cord (arrow). This finding can
easily be mistaken for a syrinx. B, Sagittal
T1-weighted image also suggests low signal
intensity in the same region but fails to
show a syrinx, demonstrating normal cord
anatomy. C, Axial T2-weighted image also
demonstrates normal anatomy. These find-
ings are compatible with a Gibbs artifact.
255
Magnetic Resonance Imaging of the Pediatric Spine

ation by age in the upper limit of nor-

mal: 6 mm in the first decade of life,
5 mm in the second and third de-
cades, and 3 mm by the ninth decade.
In Chiari I malformations, the brain-
stem is spared and the fourth ventri-
cle remains in its normal location.
Chiari I malformations are associat-
ed with syringohydromyelia, cranio-
vertebral junction anomalies, and
basilar invagination. Chiari II malfor-
mations are more advanced and con-
sist of downward displacement of the
brainstem and inferior cerebellum into
the cervical spinal canal, with a de-
crease in size of the posterior fossa.

Tethered Cord Syndrome

Tethered cord syndrome is seen in
a substantial number of patients with
spinal dysraphism, especially those
who have undergone surgical closure
of the defect.24,25 During fetal life, the
spinal cord extends to the sacrococ-
cygeal level. Because of the rapid growth
of the vertebral column after birth, the
cord ascends to the L1-L2 level in the
newborn. During the formation of a
spinal dysraphic defect such as my-
elomeningocele, the open neural el-
ements often attach to the peripheral
ectoderm, resulting in spinal cord teth-
ering. After surgical closure of the sac,
there is a tendency for the spinal cord
to become adherent at the repair site.
As the child grows, this adherence may
tether the cord and prevent cephalad
cord migration, with eventual symp-
toms. Thus, in patients with spinal dys-
raphic and related conditions, includ-
ing myelomeningoceles, myeloceles,
lipomeningoceles, and diastematomy-
elia, tethered cord should be ruled out
as the potential cause of any deteri-
oration in neurologic function.
MRI has been proposed as the ini-
tial, and possibly only, imaging study
for a patient with a suspected teth-
ered spinal cord.9 Sagittal images
should be evaluated to determine the
level of the conus medullaris (Fig. 7).
A conus level below the L2-L3 inter-
space in children older than 5 years
is abnormal and an indication of pos-
Figure 7 A 14-year-old boy had a history of lipomeningocele. After surgical resection, bowel
and bladder dysfunction and new lower-extremity paresthesias developed. A, Sagittal T2-
weighted image shows the conus medullaris extending to approximately the L4 level and
the filum terminale extending to the S1 level (arrow), compatible with tethered cord syn-
drome. B, Axial T2-weighted image at the L4 level shows the cord located posteriorly within
the thecal sac (arrow). C, Axial T2-weighted image at the L5 level shows the placode (thin
arrow) with a right-side nerve root (thick arrow) coursing anteriorly and laterally.
sible tethering.8,9 In addition, the teth- syndrome, as well as with spinal in-
ered cord is often displaced posteri- strumentation. Safety is also a concern.
orly in the spinal canal. Other findings
include lipoma or scar tissue within Scoliosis
the epidural space and increased thick- The use of MRI imaging in scoli-
ness of the filum terminale.9 Although osis is primarily to detect intraspinal
MRI can determine whether a spinal abnormalities, which are more fre-
cord is anatomically tethered, these quently associated with uncommon
findings should be correlated with the curve patterns such as left thoracic
patient’s symptoms and serial phys- curves, an abnormal neurologic ex-
ical examinations before surgical re- amination, or young age at pre-
lease is considered. sentation.26-30 Recently, Do et al26 con-
cluded that MRI is not indicated
before spine arthrodesis in a patient
Controversies in MRI of with an adolescent idiopathic scoli-
the Pediatric Spine osis curve pattern and a normal phys-
ical and neurologic examination.
MRI of the pediatric spine remains One area of particular controversy
controversial in several conditions, in- is back pain in the presence of scolio-
cluding scoliosis and tethered cord sis. In a retrospective study of 2,442

256 Journal of the American Academy of Orthopaedic Surgeons


patients, Ramirez et al31 found that a
left thoracic curve or abnormal result
on neurologic examination best pre-
dicted an underlying pathologic con-
dition. They found a significant asso-
ciation between back pain and age older
than 15 years (P < 0.001), skeletal ma-
turity (P < 0.001), postmenarcheal sta-
tus (P < 0.001), and history of injury
(P < 0.018). The authors concluded that
it is unnecessary to perform extensive
diagnostic studies on every patient with
scoliosis and back pain. MRI should
be reserved for patients with infan-
tile or juvenile scoliosis, left thoracic
curves, or abnormal neurologic find-
ings. Because coronal views are espe-
cially useful in evaluating patients with
scoliosis, they should be a part of the
routine imaging protocol.
Tethered Cord Syndrome
The rate of MRI in tethered cord
syndrome remains controversial.
When MRI demonstrates a tethered
Vol 11, No 4, July/August 2003
cord, a choice between surgical and
nonsurgical treatment must be made.
Although anatomic tethering of the
cord is detected easily on MRI, indi-
cations for surgery depend on the
clinical history and results of serial
physical examinations.
Imaging in the Presence of
Implants
MRI of the spine in the presence
of instrumentation is generally safe
but is limited by the image artifacts
the implants produce. The pulse se-
quence used for imaging titanium
produces less degradation from arti-
fact because it is less ferromagnetic
than stainless steel (Fig. 8).32,33 Thus,
titanium may be the better choice of
implant in a patient who may require
follow-up with MRI. However, with
appropriate imaging techniques, clin-
ically useful information can be ob-
tained safely in the presence of both
types of implants.34 Specialized pulse
A. Jay Khanna, MD, et al
sequences such as the metal artifact
reduction sequence (MARS) can help
reduce the degree of tissue-obscuring
artifact produced by spinal hardware
and improve image quality compared
with conventional T1-weighted spin-
echo pulse sequences.35
MRI Safety
MRI may be contraindicated in pa-
tients with ferromagnetic implants,
materials, or devices because of the
risk of implant dislodgement, heat-
ing, and induction of current.36 Shel-
lock et al36 reviewed and compiled the
results of more than 80 studies and
described the ferromagnetic qualities
of 338 objects, including 30 ortho-
paedic implants, materials, and devic-
es. They found that most orthopaedic
implants are made from nonferro-
magnetic materials and therefore are
safe for MRI procedures. Another
concern is that of safety within the
MRI suite. Areas surrounding and
Figure 8 A 6-year-old boy had a
history of high-grade astrocytoma.
A, Anteroposterior radiograph 6 weeks
after resection, multilevel laminectomy,
and posterior spinal arthrodesis from
T4 to L3 with titanium pedicle screws,
hooks, and rods. B, Midline sagittal
postgadolinium T1-weighted MRI scan
allows visualization of the canal con-
tents with minimal artifact from the
pedicle screws (arrows). C, Parasag-
ittal postgadolinium T1-weighted im-
age shows a rod (thick arrow) and
pedicle screw (thin arrow). Neither
obscures the MRI scan. D, Axial post-
gadolinium T1-weighted image also
shows the pedicle screws (arrows) and
a patent spinal canal.
257
Magnetic Resonance Imaging of the Pediatric Spine
within the suite should be carefully
monitored for the presence of ferro-
magnetic equipment that may act as
a projectile and injure the patient or
hospital personnel. A recent report
described a series of projectile cylin-
der accidents when ferromagnetic ni-
trous oxide or oxygen tanks were in
the MRI suite.37 Other equipment (eg,
References
1. Beebe DS, Tran P, Bragg M, Stillman A,
Truwitt C, Belani KG: Trained nurses
can provide safe and effective sedation
for MRI in pediatric patients. Can J
Anaesth 2000;47:205-210.
2. Sury MR, Hatch DJ, Deeley T, Dicks-
Mireaux C, Chong WK: Development
of a nurse-led sedation service for pae-
diatric magnetic resonance imaging.
Lancet 1999;353:1667-1671.
3. Vade A, Sukhani R, Dolenga M,
Habisohn-Schuck C: Chloral hydrate
sedation of children undergoing CT
and MR imaging: Safety as judged by
American Academy of Pediatrics
guidelines. AJR Am J Roentgenol 1995;
165:905-909.
4. American Academy of Pediatrics Com-
mittee on Drugs: Guidelines for moni-
toring and management of pediatric
patients during and after sedation for
diagnostic and therapeutic procedures.
Pediatrics 1992;89(6 pt 1):1110-1115.
5. Malviya S, Voepel-Lewis T, Prochaska
G, Tait AR: Prolonged recovery and de-
layed side effects of sedation for diag-
nostic imaging studies in children.
Pediatrics 2000;105:e42. Available at
http://www.pediatrics.org/cgi/reprint/
105/3/342pdf. Accessed May 2, 2003.
6. Goske MJ, Modic MT, Yu S: Pediatric
spine: Normal anatomy and spinal dys-
raphism, in Modic MT, Masaryk TJ,
Ross JS (eds): Magnetic Resonance Imag-
ing of the Spine, ed 2. St. Louis, MO:
Mosby-Year Book, 1994, pp 352-387.
7. Sze G, Baierl P, Bravo S: Evolution of
the infant spinal column: Evaluation
with MR imaging. Radiology 1991;181:
819-827.
8. Barson AJ: The vertebral level of termi-
nation of the spinal cord during normal
and abnormal development. J Anat
1970;106:489-497.
9. Moufarrij NA, Palmer JM, Hahn JF,
Weinstein MA: Correlation between
magnetic resonance imaging and surgi-
cal findings in the tethered spinal cord.
258
intravenous pumps, hospital beds,
handheld instruments) also should be
compatible with MRI.
Summary
MRI is an excellent modality for ad-
vanced imaging of the pediatric
Neurosurgery 1989;25:341-346.
10. Saifuddin A, Burnett SJ, White J: The
variation of position of the conus med-
ullaris in an adult population: A mag-
netic resonance imaging study. Spine
1998;23:1452-1456.
11. Auletta JJ, John CC: Spinal epidural ab-
scesses in children: A 15-year experi-
ence and review of the literature. Clin
Infect Dis 2001;32:9-16.
12. du Lac P, Panuel M, Devred P, Bollini
G, Padovani J: MRI of disc space infec-
tion in infants and children: Report of
12 cases. Pediatr Radiol 1990;20:175-178.
13. Modic MT, Feiglin DH, Piraino DW, et
al: Vertebral osteomyelitis: Assessment
using MR. Radiology 1985;157:157-166.
14. Fernandez M, Carrol CL, Baker CJ: Dis-
citis and vertebral osteomyelitis in chil-
dren: An 18-year review. Pediatrics 2000;
105:1299-1304.
15. Miller GM, Forbes GS, Onofrio BM:
Magnetic resonance imaging of the
spine. Mayo Clin Proc 1989;64:986-1004.
16. Kriss VM, Kriss TC: SCIWORA (spinal
cord injury without radiographic ab-
normality) in infants and children. Clin
Pediatr (Phila) 1996;35:119-124.
17. Pang D, Pollack IF: Spinal cord injury
without radiographic abnormality in
children: The SCIWORA syndrome.
J Trauma 1989;29:654-664.
18. Banerian KG, Wang AM, Samberg LC,
Kerr HH, Wesolowski DP: Association
of vertebral end plate fracture with pe-
diatric lumbar intervertebral disk her-
niation: Value of CT and MR imaging.
Radiology 1990;177:763-765.
19. Walker HS, Dietrich RB, Flannigan BD,
Lufkin RB, Peacock WJ, Kangarloo H:
Magnetic resonance imaging of the pe-
diatric spine. Radiographics 1987;7:1129-
1152.
20. Altman NR, Altman DH: MR imaging
of spinal dysraphism. AJNR Am J
Neuroradiol 1987;8:533-538.
21. Byrd SE, Darling CF, McLone DG, To-
mita T: MR imaging of the pediatric
Journal of the American Academy of Orthopaedic Surgeons
spine. A basic understanding of the
normal MRI appearance of the spine
at various ages, the signal character-
istics of various pathologic changes,
and the differential diagnosis of spi-
nal pathology can help the clinician
correlate the history and physical
examination with MRI findings to
establish the most likely diagnosis.
spine. Magn Reson Imaging Clin N Am
1996;4:797-833.
22. Modic MT, Yu S: Normal anatomy, in
Modic MT, Masaryk TJ, Ross JS (eds):
Magnetic Resonance Imaging of the Spine,
ed 2. St. Louis, MO: Mosby-Year Book,
1994, pp 37-79.
23. Mikulis DJ, Diaz O, Egglin TK, Sanchez
R: Variance of the position of the cere-
bellar tonsils with age: Preliminary re-
port. Radiology 1992;183:725-728.
24. Hall WA, Albright AL, Brunberg JA: Di-
agnosis of tethered cords by magnetic
resonance imaging. Surg Neurol 1988;
30:60-64.
25. Heinz ER, Rosenbaum AE, Scarff TB,
Reigel DH, Drayer BP: Tethered spinal
cord following meningomyelocele re-
pair. Radiology 1979;131:153-160.
26. Do T, Fras C, Burke S, Widmann RF,
Rawlins B, Boachie-Adjei O: Clinical
value of routine preoperative magnetic
resonance imaging in adolescent idio-
pathic scoliosis: A prospective study of
three hundred and twenty-seven pa-
tients. J Bone Joint Surg Am 2001;83:577-
579.
27. Evans SC, Edgar MA, Hall-Craggs MA,
Powell MP, Taylor BA, Noordeen HH:
MRI of ‘idiopathic’ juvenile scoliosis: A
prospective study. J Bone Joint Surg Br
1996;78:314-317.
28. Gupta P, Lenke LG, Bridwell KH: Inci-
dence of neural axis abnormalities in
infantile and juvenile patients with spi-
nal deformity: Is a magnetic resonance
image screening necessary? Spine 1998;
23:206-210.
29. Mejia EA, Hennrikus WL, Schwend
RM, Emans JB: A prospective evalua-
tion of idiopathic left thoracic scoliosis
with magnetic resonance imaging.
J Pediatr Orthop 1996;16:354-358.
30. Schwend RM, Hennrikus W, Hall JE,
Emans JB: Childhood scoliosis: Clinical
indications for magnetic resonance im-
aging. J Bone Joint Surg Am 1995;77:
46-53.

31. Ramirez N, Johnston CE, Browne RH:
The prevalence of back pain in children
who have idiopathic scoliosis. J Bone
Joint Surg Am 1997;79:364-368.
32. Rudisch A, Kremser C, Peer S, Kathrein
A, Judmaier W, Daniaux H: Metallic ar-
tifacts in magnetic resonance imaging
of patients with spinal fusion: A com-
parison of implant materials and imag-
ing sequences. Spine 1998;23:692-699.
33. Rupp R, Ebraheim NA, Savolaine ER, Jack-
son WT: Magnetic resonance imaging eval-
Vol 11, No 4, July/August 2003
uation of the spine with metal implants:
General safety and superior imaging with
titanium. Spine 1993;18:379-385.
34. Lyons CJ, Betz RR, Mesgarzadeh M,
Revesz G, Bonakdarpour A, Clancy M:
The effect of magnetic resonance imag-
ing on metal spine implants. Spine 1989;
14:670-672.
35. Chang SD, Lee MJ, Munk PL, Janzen
DL, MacKay A, Xiang QS: MRI of spi-
nal hardware: Comparison of conven-
tional T1-weighted sequence with a
A. Jay Khanna, MD, et al
new metal artifact reduction sequence.
Skeletal Radiol 2001;30:213-218.
36. Shellock FG, Morisoli S, Kanal E: MR
procedures and biomedical implants,
materials, and devices: 1993 update.
Radiology 1993;189:587-599.
37. Chaljub G, Kramer LA, Johnson RF III,
Johnson RF Jr, Singh H, Crow WN: Pro-
jectile cylinder accidents resulting from
the presence of ferromagnetic nitrous
oxide or oxygen tanks in the MR suite.
AJR Am J Roentgenol 2001;177:27-30.
259
 Injury to the Tarsometatarsal Joint Complex
Michael C. Thompson, MD, and Matthew A. Mormino, MD

Abstract
Tarsometatarsal joint complex fracture-dislocations may result from direct or in- ment and treatment of injuries. Sta-
direct trauma. Direct injuries are usually the result of a crush and may involve as- bility of the complex is achieved by
sociated compartment syndrome, significant soft-tissue injury, and open fracture- a combination of bony architecture and
dislocation. Indirect injuries are often the result of an axial load to the plantarflexed ligamentous support. The medial, mid-
foot. Midfoot pain after even a minor forefoot injury should raise suspicion; up to dle, and lateral cuneiforms articulate
20% of tarsometatarsal joint complex injuries are missed on initial examination. distally with the first, second, and third
An anteroposterior radiograph with abduction stress may reveal subtle injury, but metatarsals, respectively14 (Fig. 1, A).
computed tomography is the preferred imaging modality. The goal of treatment is The cuboid articulates distally with
the restoration of a pain-free, functional foot. The preferred treatment is open re- the fourth and fifth metatarsals. The
duction and internal fixation, using screw fixation for the medial three rays and middle cuneiform is recessed proxi-
Kirschner wires for the fourth and fifth tarsometatarsal joints. Satisfactory outcome mally relative to the medial and lat-
can be expected in approximately 90% of patients. eral cuneiforms. This mortise config-
J Am Acad Orthop Surg 2003;11:260-267 uration accommodates the base of the
second metatarsal and lends additional
osseous stability at this articulation.
In the coronal plane, stability is fur-
Lisfranc described amputations through painful posttraumatic arthritis and pla- ther enhanced by the so-called Roman
the tarsometatarsal (TMT) joint for the novalgus deformity.3,4 A high index arch configuration of the metatarsal
treatment of severe, gangrenous mid- of suspicion should be maintained when bases, with the second metatarsal base
foot injuries, and his name has been examining a patient with an injured acting as the keystone (Fig. 1, B).
associated with many different inju- foot because delayed or missed diag- Ligaments supporting the TMC
ries to this region.1 Myerson2 described nosis occurs in up to 20% of cases.5-7 are grouped according to anatomic lo-
such injuries as involving the tarsometa- The goal of treating TMC injury is cation (dorsal, plantar, and in-
tarsal complex (TMC), which includes to obtain a plantigrade, stable, pain- terosseous). The lesser metatarsals are
the metatarsals and TMT joints, the less foot. Successful outcome largely bound together by dorsal and plan-
cuneiforms, the cuboid, and the na- is related to obtaining and maintain- tar intermetatarsal ligaments (Fig. 1,
vicular.2 The spectrum of TMC injury ing an anatomic reduction.5,6,8,9 Ear- A). Similarly, dorsal and plantar in-
ranges from low-energy trauma, such ly studies documented the failure of tertarsal ligaments hold the cunei-
as a misstep, to high-energy crush in- closed reduction to maintain an an- forms and cuboid together. There are
juries characterized by extensive os- atomic reduction.10-12 In 1982, Hard-
seous comminution and soft-tissue com- castle et al13 reported that open tech-
promise. Accordingly, the pattern of niques with temporary, nonrigid Dr. Thompson is Chief Resident, Department of
TMC injury is highly variable and may fixation occasionally resulted in late Orthopaedic Surgery and Rehabilitation, Creighton-
Nebraska Health Foundation, University of Ne-
involve purely ligamentous disrup- displacement. Rigid screw fixation, the
braska Medical Center, Omaha, NE. Dr. Mormino
tions without fracture, associated meta- technique reported by Arntz et al6 in is Assistant Professor and Director, Orthopaedic
tarsal fractures, or fractures of the cu- 1988, has become the preferred meth- Trauma, Department of Orthopaedic Surgery and
neiforms, cuboid, or navicular. od for stabilization of these injuries.5 Rehabilitation, University of Nebraska Medical
Accurate diagnosis of these inju- Center.
ries is paramount. Although only min-
Reprint requests: Dr. Mormino, 981080 Nebraska
imal displacement may be present on Anatomy and Medical Center, Omaha, NE 68198-1080.
initial radiographs, severe ligamen- Biomechanics
tous disruption might still exist. Left Copyright 2003 by the American Academy of
untreated, such disruption may result Understanding the anatomy of the Orthopaedic Surgeons.
in marked disability characterized by TMC is imperative for accurate assess-

260 Journal of the American Academy of Orthopaedic Surgeons

 Michael C. Thompson, MD, and Matthew A. Mormino, MD

no ligamentous connections between at the third TMT joint is approximately

Figure 1 A, Anteroposterior view of the
bony and ligamentous anatomy of tarsometa-
tarsal joint complex. I through V = metatar-
sal bones. (Adapted with permission from
Myerson MS: Fractures of the midfoot and
forefoot, in Myerson MS: Foot and Ankle Dis-
orders. Philadelphia, PA: WB Saunders, 2000,
vol 2, pp 1265-1296.) B, Coronal section
through the metatarsal bases illustrating the
Roman arch configuration. (Adapted with
permission from Lenczner EM, Waddell JP,
Graham JD: Tarsal-metatarsal [Lisfranc] dis-
location. J Trauma 1974;14:1012-1020.)
the first and second metatarsal bases.
The largest and strongest interos-
seous ligament in the TMC is the so-
called Lisfranc ligament, which aris-
es from the lateral surface of the
medial cuneiform and inserts onto the
medial aspect of the second metatar-
sal base near the plantar surface.14 The
first metatarsal base is anchored to the
dorsal and plantar aspects of the me-
dial cuneiform by two longitudinal
ligaments. The peroneus longus and
tibialis anterior tendon insertions fur-
ther stabilize the first TMT joint. A
variable network of longitudinal and
oblique ligaments secures the remain-
der of the metatarsals to the cunei-
forms and cuboid on the dorsal and
plantar aspects of the complex. In
general, the dorsal ligaments are
weaker than their plantar counter-
parts. To a lesser extent, the plantar
fascia and intrinsic musculature of the
foot add stability to the TMC.
Because of the unique bony and lig-
amentous anatomy of the TMC, nor-
mal motion of the individual compo-
nents varies. Having articular contact
with all three cuneiforms, the base of
the second metatarsal demonstrates
very little motion under normal cir-
cumstances, with an average dorsi-
flexion-plantarflexion arc of 0.6°.15 In
comparison, dorsiflexion-plantarflexion
1.6°, and, at the first joint, 3.5°. The
fourth and fifth TMT joints are the most
mobile, demonstrating an average
of 9.6° and 10.2° of dorsiflexion-
plantarflexion, respectively.15
Injury to the
Tarsometatarsal Joint
Complex
The overall annual incidence of TMC
injuries is approximately 1 per 60,000
persons,13,16 and the injury is two to
three times more common in males
(Table 1). Motor vehicle accidents are
the most frequently cited mechanism,
accounting for about 40% to 45% of
injuries. Low-energy mechanisms ac-
count for approximately 30%. Falls
from a height and crush injuries are
also commonly reported causes.
The mechanism of TMC injury may
be either direct or, more commonly,
indirect trauma. The direct mechanism
involves high-energy blunt trauma,
usually applied to the dorsum of the
foot. Crush injuries constitute most of
these injuries, and many are associ-
ated with notable soft-tissue trauma.
Associated compartment syndromes
and open fracture-dislocations are
more often present with direct inju-
ry mechanisms. In part as a result of

Table 1
Tarsometatarsal Joint Complex: Mechanisms of Injury

No. of Injuries (%)

No. of Patients/ Motor Vehicle Fall From

Study Injuries (M/F) Accident Height Crush Other

Kuo et al5 48/48 (32/16) 20 (42) 7 (14.5) 6 (12.5) 15 (31)

Arntz et al6 40/41 (28/12) 21 (51) 7 (17) 0 (0) 13 (32)
Vuori et al16 66/66 (46/20) 22 (33) 9 (14) 14 (21) 21 (32)
Myerson et al9 52/55 (NA) 34 (62) 8 (14.5) 8 (14.5) 5 (9)
Hesp et al36 23/23 (16/7) 19 (83) 3 (13) 1 (4) 0 (0)
Hardcastle et al13 119/119 (86/33) 48 (40.3) 16 (13.5) 0 (0) 55 (46.2)
Wilppula et al12 26/26 (21/5) 7 (27) 0 (0) 8 (31) 11 (42)
NA = not available.

Vol 11, No 4, July/August 2003 261

Injury to the Tarsometatarsal Joint Complex

the associated soft-tissue trauma and
greater degree of articular injury, di-
rect injuries often result in a worse clin-
ical outcome compared with indirect
injuries.8,9
The indirect mechanism of injury
usually involves axial loading of the
plantarflexed foot. An example is a
football player falling onto the heel
of another player whose foot is planted
and plantarflexed. This type of injury
also can occur with soccer, basketball,
and gymnastics.17 Falls from a height
may result in forefoot plantarflexion
at the time of impact. In automobile
accidents, injury to the plantarflexed
foot occurs with a combination of de-
celeration and floorboard intrusion.
Less commonly, violent abduction or
twisting of the forefoot may result in
fracture-dislocation around the TMC.
The fracture pattern and direction
of dislocation in direct injuries are
highly variable and depend on the
force vector applied. In contrast, the
most frequent pattern seen in indirect
injuries involves failure of the weak-
er dorsal TMT ligaments in tension,
with subsequent dorsal or dorsolat-
eral dislocation of the metatarsals. Mi-
nor displacement at the TMT joint
level results in a marked reduction in
articular contact. Dorsolateral dis-
placement of the second metatarsal
base of 1 or 2 mm results in the re-
duction of the TMT articular contact
area by 13.1% and 25.3%, respective-
ly.18 Although fractures of the cune-
iforms are relatively common, the
most frequent fracture in TMC inju-
ries involves the second metatarsal
base.16 Less common are associated
fractures of the cuboid, navicular, or
other metatarsals.
subcutaneous tissue. Inspection of the
foot may reveal gross morphologic ab-
normalities such as widening or flat-
tening. A gap between the first and
second toes is suggestive of intercu-
neiform disruption as well as TMT
joint injury.19,20 Palpation of the dor-
salis pedis artery may not be pos-
sible, depending on the extent of
swelling and deformity. Although dis-
ruption of the dorsalis pedis artery has
been reported, the incidence of vas-
cular injury appears to be rare.7,21,22
Significant pain on passive dorsiflex-
ion of the toes in a tensely swollen foot
is suggestive of a compartment syn-
drome; however, evaluation may be
hampered by pain associated with the
osseous injury.23,24 When there is un-
certainty about the presence of a com-
partment syndrome, pressures should
be measured. An absolute pressure >40
mm Hg is diagnostic and an indica-
tion for emergent compartment re-
lease. Particularly in the hypotensive
patient, a compartment pressure with-
in 30 mm Hg of the diastolic pressure
also is an indication for release.
Findings after a low-energy TMC
injury may be relatively subtle. Ahigh
index of suspicion should be main-
tained in the patient with forefoot
pain after even a minor traumatic
event. Patients usually have notable
pain on weight bearing or are unable
to bear weight on the affected foot.
Swelling is present to a variable ex-
tent, and ecchymosis occasionally is
found along the plantar aspect of the
midfoot.25 Palpation of the affected
TMT joints usually reveals tender-
ness. Notable pain on passive abduc-
tion and pronation of the forefoot also
is suggestive of TMC injury.17
The initial radiographic examina-
tion should include anteroposterior,
lateral, and 30° oblique views of the
foot. To visualize the Lisfranc joint in
the tangential plane, the anteropos-
terior radiograph should be taken
with the beam approximately 15° off
vertical. Standing radiographs are
ideal but may be difficult to obtain
secondary to pain (Fig. 2, A and B).
If weight-bearing views are not pos-
sible, a stress view with the forefoot
in abduction often will reveal subtle
instability, especially at the first TMT
joint.17,26 All radiographs should be
evaluated for signs of instability. On
the anteroposterior view, the distance
between the first and second metatar-
sal bases varies among uninjured in-
dividuals, with up to 3 mm consid-

Diagnosis
The diagnosis of high-energy or crush
injuries to the TMC is relatively
straightforward. Examination typical-
ly reveals moderate to severe swell-
ing of the forefoot and, in open inju-
ries, disruption of the skin and
Figure 2 A, Anteroposterior non–weight-bearing radiograph of a patient with forefoot pain
after an axial load injury. Note the subtle widening (arrow) between the bases of the first and
second metatarsals. B, Anteroposterior standing view of the same patient as in Panel A dem-
onstrating subluxation (arrow) at the base of the second metatarsal. C, Anteroposterior view
of a patient with avulsion of the Lisfranc ligament, or fleck sign (arrow), at the base of the
second metatarsal.

262 Journal of the American Academy of Orthopaedic Surgeons

 Michael C. Thompson, MD, and Matthew A. Mormino, MD

ered normal.26,27 In subtle cases,

radiographs of the contralateral foot
should be obtained for comparison.
Stein28 reviewed 100 radiographs
of normal feet and noted several con-
stant anatomic relationships. On the
anteroposterior view, the medial bor-
der of the second metatarsal is in line
with the medial border of the mid-
dle cuneiform, the first metatarsal
aligns with the medial and lateral bor-
ders of the medial cuneiform, and the
first and second intermetatarsal space
is continuous with the intertarsal space
of the medial and middle cuneiforms
(Fig. 1, A). On the 30° oblique view,
the medial border of the fourth meta-
tarsal is in line with the medial bor-
der of the cuboid, the lateral border
of the third metatarsal is aligned with
the lateral border of the lateral cune-
iform, and the third and fourth inter-
metatarsal space is continuous with Figure 3 Medial column line. On an anteroposterior radiograph with the forefoot stressed
the intertarsal space of the lateral cu- in abduction (dashed outline of first metatarsal), a line is drawn tangential to the medial bor-
neiform and the cuboid.28 ders of the navicular and medial cuneiform (heavy dashed line). Failure of this line to in-
tersect the base of the first metatarsal is strongly suggestive of TMC injury. A, Normal foot.
Other radiographic findings may B, First, second, and third TMT joint disruption (heavy dark line). Arrows indicate direction
assist with diagnosis. The fleck sign, of forces. (Adapted with permission from Coss HS, Manos RE, Buoncristiani A, Mills WJ:
or avulsion of Lisfranc’s ligament at Abduction stress and AP weightbearing radiography of purely ligamentous injury in the tar-
sometatarsal joint. Foot Ankle Int 1998;19:537-541.)
the base of the second metatarsal, is
diagnostic of TMC injury9 (Fig. 2, C).
Analysis of the medial column line on dislocations, a preoperative CT may published classification system, pub-
an anteroposterior abduction stress facilitate surgical planning by delin- lished by the Orthopaedic Trauma
view may reveal subtle injury26 (Fig. eating the extent of osseous injury. Association,35 is similar to the orig-
3). Flattening of the longitudinal arch The role of magnetic resonance im- inal Quenu and Kuss classification.
may suggest injury to the TMC and aging (MRI) in evaluating TMC inju- These classification systems are all
can be evaluated by comparing the ries has yet to be defined. MRI is more based on the congruency of the TMT
weight-bearing lateral view to that of sensitive than plain radiographs in joints and the direction of displace-
the uninjured foot.29 detecting small fractures and joint ment of the metatarsal bases. Com-
Computed tomography (CT) has malalignment and in assessing liga- mon to all classification systems is
proved to be a valuable tool in the di- mentous structures around the that none appears to be helpful in
agnosis of injuries to the TMC. It is TMC.33,34 However, with regard to di- terms of management or prognosis.9
more sensitive than plain radiographs agnosis and decision-making, CT is
in detecting minor displacement and superior to MRI.30 Therefore, MRI is
small fractures.30-32 Displacement of not routinely recommended in the as- Management
up to 2 mm may not be detectable on sessment of these injuries.
plain radiographs but is visible on Nonsurgical management of TMC in-
CT.31 Axial and coronal views of both juries should be limited to those that
feet should be made for comparison. Classification are without fracture, nondisplaced,
Subtle widening or dorsal sublux- and stable under radiographic stress
ation of the metatarsals are CT find- The earliest classification system was examination. As little as 2 mm of dis-
ings suggestive of TMC disruptions, published in 1909 by Quenu and placement or the presence of a frac-
and avulsion fracture of the second Kuss12 and subsequently modified by ture within the TMC warrants fixa-
metatarsal base is diagnostic of in- Hardcastle et al13 in 1982 and Myer- tion. Nondisplaced, stable ligamentous
jury33 (Fig. 4). In high-energy fracture- son et al9 in 1986. The most recently injuries may be treated in a non–

Vol 11, No 4, July/August 2003 263

Injury to the Tarsometatarsal Joint Complex
Figure 4 A, Coronal CT scan demonstrating subtle widening (arrow) of the first and sec-
ond metatarsal bases. B, Coronal CT scan showing an avulsion fracture (arrow) of the sec-
ond metatarsal base.
weight-bearing short leg cast for a
minimum of 6 weeks. Radiographic
examination should be done 1 to 2
weeks after injury to ensure that align-
ment and stability are maintained.
Gradual weight bearing in a protec-
tive brace may begin at 6 weeks. Per-
mission for unrestricted activity, such
as running and jumping, should be
withheld for 3 to 4 months.
Although displaced or unstable
TMC injuries have been treated by
closed reduction and casting, loss of
reduction was common and outcomes
were variable, with a high incidence
of poor results. Currently accepted sur-
gical techniques involve either closed
reduction with percutaneous Kirsch-
ner wire (K-wire) or screw fixation2
or open reduction with screw and/
or K-wire fixation.4-6 For fixation of
the medial three TMT joints, screw fix-
ation may be preferable to K-wires be-
cause ligamentous healing may re-
quire as much as 12 to 16 weeks of
immobilization to occur, and K-wires
can become loose, necessitating re-
moval as early as 6 weeks. Regard-
less of the technique used, the goal
should be anatomic reduction of the
affected joints because numerous stud-
ies have documented that clinical out-
come correlates with accuracy of
reduction.1,5-9,12,21,36,37
264
Ideally, surgical management of
closed injuries is undertaken when
soft-tissue swelling is at a minimum,
either immediately or after swelling
has abated. This delay may take up
to 2 weeks and can be identified by
the return of wrinkles to the skin. The
initial incision is made dorsally be-
tween the first and second web space.
The extensor hallucis longus tendon,
deep peroneal nerve, and dorsalis pe-
dis artery are identified and retract-
ed as a unit, allowing deep, sharp dis-
section to expose the first and second
TMT joints. Small, irreducible bone
fragments are débrided from the
joints. The reduction should begin
medially and progress laterally.
Aligning the medial aspect of the first
metatarsal and the medial cuneiform
reduces the first TMT joint. The en-
tire medial aspect of this joint is ex-
posed to ensure that plantar gapping
is not present. The reduction is pro-
visionally held with a K-wire, and the
joint is stabilized with a countersunk
3.5- or 2.7-mm screw placed from the
base of the first metatarsal into the
medial cuneiform. Using fully thread-
ed cortical screws placed for position-
ing, rather than compression, is pref-
erable. Screws crossing otherwise
normal joints result in little, if any,
long-term morbidity. If rotational in-
Journal of the American Academy of Orthopaedic Surgeons
stability of the first TMT joint persists
after placement of the first screw, a
second screw or K-wire may be
placed from the medial cuneiform
into the base of the first metatarsal.
The second metatarsal is then re-
duced to the medial border of the
middle cuneiform and temporarily
held with a K-wire. Definitive fixation
follows with a 3.5- or 2.7-mm coun-
tersunk screw directed from the base
of the second metatarsal into the mid-
dle cuneiform. A 3.5-mm screw is
usually appropriate for most patients;
a 2.7-mm screw may be used for pa-
tients of small stature or when there
is concern about the size of the 3.5-
mm screw relative to the diameter of
the second metatarsal. Medial column
fixation is then completed by placing
a 3.5- or 2.7-mm screw from the me-
dial cuneiform into the base of the
second metatarsal.
If the third TMT joint is disrupted
and remains unstable after fixation of
the first and second TMT joints, a sec-
ond dorsal incision is made between
the third and fourth metatarsals to ex-
pose the third TMT joint. This joint
is similarly reduced and fixed with a
3.5- or 2.7-mm screw directed from
the base of the third metatarsal into
the lateral cuneiform. Reduction of
the fourth and fifth TMT joints usu-
ally occurs with reduction of the me-
dial three TMT joints and is secured
with percutaneous K-wire fixation
(Fig. 5). Alternative fixation, although
typically unnecessary, is done with
screw fixation.
Occasionally, an associated impact-
ed (nutcracker) fracture of the cuboid
may require treatment. The technique
described by Sangeorzan and Swiont-
kowski38 involves restoration of cuboid
length by distraction bone grafting and
plating. Failure to restore length re-
sults in lateral column shortening and
a persistently abducted and pronated
forefoot. A distractor or external fix-
ator may be used intraoperatively to
facilitate distraction before plating (Fig.
6). Associated fractures of the navicu-
lar may be exposed and stabilized by

Figure 5 Typical fixation scheme for a TMC
disruption.
extending the dorsal medial incision
proximally. In most cases, fragments
are large enough to accommodate 3.5-
or 2.7-mm screws placed using a lag
technique.
Rarely, severely comminuted or
contaminated injuries of the TMC may
not be amenable to internal fixation
using standard techniques. Temporary
or definitive spanning external fixa-
tion is an option for these difficult cas-
es. Limited percutaneous fixation with
K-wires or screws may augment sta-
bilization but should be used with cau-
tion in contaminated cases.
Wound closure should be accom-
plished with meticulous soft-tissue
handling and closure. Ashort leg, non–
weight-bearing cast is maintained for
6 weeks. Any percutaneous pins are
then removed, and the patient is ad-
vanced to full weight bearing in a
walking boot for an additional 4 to 6
weeks. The indication for screw re-
moval remains controversial.2,5 Most
authors recommend routine remov-
al of the screws either on weight bear-
ing or approximately 16 weeks after
fixation.2 We prefer to remove screws
only if patients are symptomatic but
no sooner than 16 weeks postopera-
Vol 11, No 4, July/August 2003
Michael C. Thompson, MD, and Matthew A. Mormino, MD
tively. Broken screws seem to occur
in only a minority of patients. Further-
more,affectedpatientsareoftenasymp-
tomatic, although broken screws may
be problematic if salvage by fusion is
necessary.
When a compartment syndrome is
diagnosed at the initial evaluation, emer-
gent fasciotomy should be done.23 Us-
ing the two dorsal incisions described,
the interosseous compartments are each
released. Dissection between the meta-
tarsals is done to achieve release of
the medial, central, and lateral com-
partments (Fig. 7). Rarely, associated
hindfoot injuries such as a calcaneus
fracture may be present and may re-
quire release of the calcaneal compart-
ment. This may be achieved through
a longitudinal medial incision over the
compartment. After fasciotomy, defin-
itive fixation should be done. Fascial
compartments and wounds should be
left open, and the patient may undergo
redébridement and attempted wound Figure 6 Restoration of cuboid length with
bone graft and a plate. An external fixator or
closure within 48 to 72 hours. Delayed distractor may be used intraoperatively to fa-
primary wound closure may not be cilitate distraction. (Adapted with permission
possible, and coverage with split- from Hansen ST Jr: Acute fractures in the foot,
in Hansen ST Jr: Functional Reconstruction of
thickness skin graft may be necessary.23,24 the Foot and Ankle. Philadelphia, PA: Lippin-
Open TMC fracture-dislocations cott Williams & Wilkins, 2000, pp 65-103.)
should be treated as surgical emergen-
cies. Débridement and irrigation should
be done within 6 hours of injury, if Results
possible. In addition to tetanus pro-
phylaxis, Gustilo and Anderson type In 1986, Myerson et al9 published a
I and II open injuries should receive retrospective study of 76 TMT joint
a first-generation cephalosporin, with injuries treated over a 10-year peri-
an aminoglycoside added for type III od. Six open injuries were included.
injuries. Severe contamination or vas- Treatment methods comprised immo-
cular compromise requires adding pen- bilization alone, closed reduction and
icillin G to the antibiotic regimen. casting, closed reduction and percu-
Wounds are left open and covered with taneous K-wire fixation, and open re-
saline gauze or an equivalent dress- duction followed by K-wire fixation.
ing. Repeat débridement and irriga- Fifty-five injuries were followed up
tion are done every 48 hours until a at a mean of 4.2 years (range, 1.6 to
clean, viable wound bed is achieved. 11 years). Immobilization alone or
Ideally, wound closure is achieved by closed reduction and casting result-
delayed primary closure. In the foot, ed in 0 of 5 and 3 of 15 (20%) good
however, this is often not possible. and excellent results, respectively. In
Coverage may be achieved by split- contrast, good to excellent clinical re-
thickness skin graft, free tissue trans- sults were documented in 9 of 17 pa-
fer, or local rotation flaps, according tients (53%) who underwent closed
to surgeon preference and institution reduction and percutaneous pinning
capabilities. as well as in 14 of 18 patients (78%)
265
Injury to the Tarsometatarsal Joint Complex
Figure 7 Release of compartment syndrome through dorsal incisions. (Adapted with per-
mission from Myerson MS: Experimental decompression of the fascial compartments of the
foot: The basis for fasciotomy in acute compartment syndromes. Foot Ankle 1988;8:308-314.)
treated with open reduction and
K-wire fixation. Seven of the eight di-
rect crush injuries had fair to poor
functional outcomes (88%). Overall,
the quality of reduction, which was
a subjective assessment of TMT joint
alignment, correlated with the clin-
ical result. Good to excellent results
were achieved in 22 of 26 patients
(85%) with an acceptable reduction
and in only 5 of 29 patients (17%) with
an unacceptable reduction. The au-
thors concluded that the major deter-
minants of unacceptable results are
the damage to the articular surface
at the time of injury and the quality
of the initial reduction.9
In 1988, Arntz et al6 published their
results of 41 TMC injuries in 40 pa-
tients treated with open reduction and
screw fixation. Seven of the injuries
were open fracture-dislocations.At sur-
gery, intra-articular fracture or peri-
articular comminution was noted in
54% of injuries (22/41). Anatomic re-
duction (within 2 mm) was achieved
in 97% of the closed injuries (33/34)
and in 88% overall (36/41). Hardware
was removed from all patients at a min-
imum of 12 weeks. Thirty-four patients
(35 injuries) were followed up at a
266
mean of 3.4 years after injury. Good
or excellent functional results were re-
ported for 93% of closed injuries (27/
29). In contrast, four of the six patients
with open fractures had a fair or poor
functional result. In all patients, the
presence of degenerative changes on
follow-up radiographs negatively cor-
related with functional outcome. Ra-
diographic evidence of posttraumatic
degenerative changes was absent or
minimal in 26 of the 30 injuries with
an anatomic reduction (87%). Con-
versely, all five injuries with nonana-
tomic reduction after surgery devel-
oped moderate or severe posttraumatic
arthritis. In general, patients who sus-
tained open injuries were more likely
to have periarticular comminution
noted intraoperatively, more advanced
posttraumatic degenerative changes
at follow-up, and a worse functional
outcome. The authors concluded that
injury to the articular cartilage and fail-
ure to achieve an anatomic reduction
were the most important determinants
in the development of posttraumatic
arthritis. Furthermore, they stressed
the importance of open anatomic re-
duction followed by rigid screw fix-
ation in optimizing outcome.6
Journal of the American Academy of Orthopaedic Surgeons
More recently, Kuo et al5 reported
on 92 TMC injuries treated over a
7-year period. Six open injuries were
included in the study. All patients
were treated surgically with the me-
dial three joints stabilized with screws
and the fourth and fifth joints, with
Kirschner wires. Postoperatively,
screws were removed only when
painful. Forty-eight patients were ex-
amined at a mean of 4.3 years after
injury (range, 1.1 to 9.5 years), for a
follow-up rate of 52%. The prevalence
of radiographic posttraumatic arthri-
tis was significantly (P = 0.004) lower
in patients with an anatomic reduc-
tion within 2 mm (6/38 [16%]) com-
pared with those with nonanatomic
reduction (6/10 [60%]). In addition,
patients with anatomic reduction had
a statistically significant (P = 0.05) bet-
ter average functional score, as mea-
sured by the American Orthopaedic
Foot and Ankle Society score for the
midfoot. Purely ligamentous injuries
tended to have a higher prevalence
of osteoarthritis, but without statis-
tical significance. The authors con-
cluded that the overall outcomes af-
ter surgical treatment of these injuries
are good and that anatomic reduction
is important for long-term outcome.5
Complications
Posttraumatic arthritis remains the
most common complication after TMC
injury. Not all patients who develop
degenerative radiographic changes are
symptomatic.9 In the series by Kuo et
al,5 12 of 48 patients (25%) had symp-
tomatic arthritis at final follow-up. Of
these, six underwent arthrodesis.Arntz
et al6 reported moderate to severe de-
generative changes on follow-up ra-
diographs in 9 of 35 patients (26%).
Cushioned inserts, shoe modifications,
and nonsteroidal anti-inflammatory
medications are the mainstay of non-
surgical treatment for posttraumatic
arthritis after TMC injury. If these mo-
dalities fail, arthrodesis of the affected
joints is the treatment of choice.

Other complications occur with
less frequency. Arntz et al6 and Kuo
et al5 reported an incidence of broken
screws of 2% and 25%, respectively.
Superficial infection, residual dyses-
thesias, late displacement, and deep
vein thrombosis have been reported
in <4% of cases.5,6,9
References
1. Cassebaum WH: Lisfranc fracture-
dislocations. Clin Orthop 1963;30:116-129.
2. Myerson MS: The diagnosis and treat-
ment of injury to the tarsometatarsal
joint complex. J Bone Joint Surg Br 1999;
81:756-763.
3. Brunet JA, Wiley JJ: The late results of
tarsometatarsal joint injuries. J Bone
Joint Surg Br 1987;69:437-440.
4. Sangeorzan BJ, Veith RG, Hansen ST Jr:
Salvage of Lisfranc’s tarsometatarsal
joint by arthrodesis. Foot Ankle 1990;10:
193-200.
5. Kuo RS, Tejwani NC, DiGiovanni CW,
et al: Outcome after open reduction and
internal fixation of Lisfranc joint injuries.
J Bone Joint Surg Am 2000;82:1609-1618.
6. Arntz CT, Veith RG, Hansen ST Jr: Frac-
tures and fracture-dislocations of the
tarsometatarsal joint. J Bone Joint Surg
Am 1988;70:173-181.
7. Goossens M, De Stoop N: Lisfranc’s
fracture-dislocations: Etiology, radiolo-
gy, and results of treatment. A review of
20 cases. Clin Orthop 1983;176:154-162.
8. Wiss DA, Kull DM, Perry J: Lisfranc
fracture-dislocations of the foot: A clin-
ical kinesiological study. J Orthop Trau-
ma 1987;1:267-274.
9. Myerson MS, Fisher RT, Burgess AR,
Kenzora JE: Fracture dislocations of the
tarsometatarsal joints: End results cor-
related with pathology and treatment.
Foot Ankle 1986;6:225-242.
10. Aitken AP, Poulson D: Dislocations of
the tarsometatarsal joint. J Bone Joint
Surg Am 1963;45:246-260.
11. Bassett FH III: Dislocations of the tar-
sometatarsal joints. South Med J 1964;57:
1294-1302.
12. Wilppula E: Tarsometatarsal fracture-
dislocation: Late results in 26 patients.
Acta Orthop Scand 1973;44:335-345.
13. Hardcastle PH, Reschauer R, Kutscha-
Lissberg E, Schoffman W: Injuries to the
tarsometatarsal joint: Incidence, classi-
fication and treatment. J Bone Joint Surg
Br 1982;64:349-356.
Vol 11, No 4, July/August 2003
Michael C. Thompson, MD, and Matthew A. Mormino, MD
Summary
Injuries to the tarsometatarsal joint
complex are often overlooked and can
be misunderstood. An appreciation of
the complex bony and ligamentous
anatomy is necessary to make an ac-
curate diagnosis from the appropri-
14. de Palma L, Santucci A, Sabetta SP, Ra-
pali S: Anatomy of the Lisfranc joint
complex. Foot Ankle Int 1997;18:356-364.
15. Ouzounian TJ, Shereff MJ: In vitro de-
termination of midfoot motion. Foot
Ankle 1989;10:140-146.
16. Vuori J-P, Aro HT: Lisfranc joint inju-
ries: Trauma mechanisms and associat-
ed injuries. J Trauma 1993;35:40-45.
17. Curtis MJ, Myerson M, Szura B: Tar-
sometatarsal joint injuries in the ath-
lete. Am J Sports Med 1993;21:497-502.
18. Ebraheim NA, Yang H, Lu J, Biyani A:
Computer evaluation of second tar-
sometatarsal joint dislocation. Foot An-
kle Int 1996;17:685-689.
19. Davies MS, Saxby TS: Intercuneiform
instability and the “gap” sign. Foot An-
kle Int 1999;20:606-609.
20. Leenen LP, van der Werken C: Fracture-
dislocations of the tarsometatarsal joint:
A combined anatomical and computed
tomographic study. Injury 1992;23:51-55.
21. Wilson DW: Injuries of the tarso-
metatarsal joints: Etiology, classifica-
tion and results of treatment. J Bone
Joint Surg Br 1972;54:677-686.
22. Del Sel JM: The surgical treatment of
tarso-metatarsal fracture-dislocations.
J Bone Joint Surg Br 1955;37:203-207.
23. Myerson MS: Management of compart-
ment syndromes of the foot. Clin Orthop
1991;271:239-248.
24. Myerson M: Split-thickness skin exci-
sion: Its use for immediate wound care
in crush injuries of the foot. Foot Ankle
1989;10:54-60.
25. Ross G, Cronin R, Hauzenblas J, Juliano
P: Plantar ecchymosis sign: A clinical
aid to diagnosis of occult Lisfranc tar-
sometatarsal injuries. J Orthop Trauma
1996;10:119-122.
26. Coss HS, Manos RE, Buoncristiani A,
Mills WJ: Abduction stress and AP
weightbearing radiography of purely
ligamentous injury in the tarsometatar-
sal joint. Foot Ankle Int 1998;19:537-541.
27. Potter HG, Deland JT, Gusmer PB, Car-
ate radiographic studies. Open ana-
tomic reduction and rigid internal
fixation is the preferred method of
management. The keys to maximiz-
ing outcome are maintaining anatom-
ic reduction (<2 mm) and avoiding
complications with safe soft-tissue
handling.
son E, Warren RF: Magnetic resonance
imaging of the Lisfranc ligament of the
foot. Foot Ankle Int 1998;19:438-446.
28. Stein RE: Radiological aspects of the
tarsometatarsal joints. Foot Ankle 1983;
3:286-289.
29. Faciszewski T, Burks RT, Manaster BJ:
Subtle injuries of the Lisfranc joint.
J Bone Joint Surg Am 1990;72:1519-1522.
30. Preidler KW, Peicha G, Lajtai G, et al:
Conventional radiography, CT, and MR
imaging in patients with hyperflexion
injuries of the foot: Diagnostic accuracy
in the detection of bony and ligamen-
tous changes. AJR Am J Roentgenol 1999;
173:1673-1677.
31. Lu J, Ebraheim NA, Skie M, Porshinsky
B, Yeasting RA: Radiographic and com-
puted tomographic evaluation of Lis-
franc dislocation: A cadaver study. Foot
Ankle Int 1997;18:351-355.
32. Goiney RC, Connell DG, Nichols DM:
CT evaluation of tarsometatarsal fracture-
dislocation injuries. AJR Am J Roentgenol
1985;144:985-990.
33. Preidler KW, Brossmann J, Daenen B,
Goodwin D, Schweitzer M, Resnick D:
MR imaging of the tarsometatarsal joint:
Analysis of injuries in 11 patients. AJR
Am J Roentgenol 1996;167:1217-1222.
34. Preidler KW, Wang Y-C, Brossmann J,
Trudell D, Daenen B, Resnick D: Tar-
sometatarsal joint: Anatomic details on
MR images. Radiology 1996;199:733-736.
35. Orthopaedic Trauma Association Com-
mittee for Coding and Classification:
Fracture and dislocation compendium.
J Orthop Trauma 1996;10(suppl 1):150.
36. Hesp WL, van der Werken C, Goris RJ:
Lisfranc dislocations: Fractures and/or
dislocations through the tarso-metatarsal
joints. Injury 1984;15:261-266.
37. van der Werf GJ, Tonino AJ: Tarsometa-
tarsal fracture-dislocation. Acta Orthop
Scand 1984;55:647-651.
38. Sangeorzan BJ, Swiontkowski MF: Dis-
placed fractures of the cuboid. J Bone
Joint Surg Br 1990;72:376-378.
267
 Chronic Exertional Compartment Syndrome
Michael J. Fraipont, MD, and Gregory J. Adamson, MD

Abstract
Chronic exertional compartment syndrome is an often overlooked and uncommon been reported in the hand, forearm,
cause of pain in the extremities of individuals who engage in repetitive physical ac- leg, thigh, gluteus, and foot. In the
tivity. A thorough history, a careful physical examination, and compartment pres- rare untreated case, an acute compart-
sure testing are essential to establish the diagnosis. Catheter measurements can pro- ment syndrome may result in myo-
vide useful information on baseline resting compartment pressures as well as necrosis, causing release of myoglo-
compartment pressures after exercise or trauma. Patients with chronic exertional bin into the vascular circulation,
compartment syndrome usually do not respond to nonsurgical therapy other than which can lead to renal failure. Treat-
completely ceasing the activities that cause the symptoms. Surgical intervention en- ment of myonecrosis consists of
tails fasciotomies of the involved compartments. Although obtaining accurate com- prompt hydration, restoration of flu-
partment pressure measurements can be difficult and fascial releases must be done id deficits, and concomitant diuresis
carefully, patients typically have satisfactory functional results and are able to re- (maintained at 100 to 200 mL/h).3
turn to their usual physical activities after fasciotomy.
J Am Acad Orthop Surg 2003;11:268-276
Anatomy
The three areas most commonly af-
Compartment syndrome is defined as leg are most commonly involved in fected by CECS are the lower leg,
increased pressure within a closed CECS, but it has been described in all thigh, and forearm. The lower leg
fibro-osseous space, causing reduced compartments of the leg, shoulder, consists of four compartments: ante-
blood flow and tissue perfusion in upper arm, forearm, hand, gluteus, rior, lateral, superficial posterior, and
that space, leading to ischemic pain thigh, and foot. deep posterior (Fig. 1). Each compart-
and possible damage to the tissues of In contradistinction, acute com- ment contains one major nerve, and
the compartment. Compartment syn- partment syndromes, whether in- two compartments (anterior and
drome may be either chronic or acute. duced by trauma or repeated exer- deep posterior) house major blood
Chronic compartment syndrome is tion, are commonly progressive and vessels, which may be affected by
often recurrent and is associated with require urgent attention to avoid ir- CECS. The anterior compartment
repetitive exertion. It is typically seen reversible damage to the tissues of the contains the anterior tibial artery and
in athletes whose exercise level ele- affected compartment. Patients with the deep peroneal nerve. The lateral
vates the intramuscular pressure to acute compartment syndrome pre- compartment contains the superficial
a point that the tissues within the af- sent with severe pain that is exacer- peroneal nerve. The superficial pos-
fected compartment become tight and bated by passive stretch of the terior compartment contains the sural
painful, thus preventing further ac- muscles and does not resolve spon-
tivity. The pain disappears quickly taneously with rest. Development of
Dr. Fraipont is Assistant Clinical Professor,
after rest, and there are usually no paresthesia and pallor can be fol-
Department of Orthopaedic Surgery, University
permanent sequelae in the affected lowed by the loss of pulse in the dis- of Southern California, Pasadena, CA. Dr.
tissue. In 1962, French and Price1 doc- tal extremity. The typical presentation Adamson is Associate Clinical Professor, Depart-
umented elevated compartment pres- of an acute compartment syndrome ment of Orthopaedic Surgery, University of
sures as the cause of chronic exertion- most often occurs after a high-energy Southern California.
al compartment syndrome (CECS) of trauma with or without fracture or
Reprint requests: Dr. Fraipont, Suite 201, 39
the tibia. Previously, in 1956, Mavor2 reperfusion of an ischemic limb. In an Congress Street, Pasadena, CA 91105.
had successfully treated a CECS by exercise-induced acute compartment
widening the fascia of the anterior syndrome, symptoms may not devel- Copyright 2003 by the American Academy of
compartment of the tibia. The ante- op until 24 to 48 hours after the pre- Orthopaedic Surgeons.
rior and lateral compartments of the cipitating event. The syndrome has

268 Journal of the American Academy of Orthopaedic Surgeons


Figure 1 Cross section of the lower left leg.
nerve. The deep posterior compart-
ment contains the posterior tibial
nerve and both the posterior tibial
and peroneal arteries and veins.
The thigh consists of three compart-
ments that can be affected by CECS:
anterior, medial, and posterior (Fig.
2). The anterior compartment contains
the femoral nerve. The medial com-
partment contains the obturator nerve
and both the femoral and femoral pro-
fundus arteries. The posterior com-
partment contains the sciatic nerve.
The forearm consists of three com-
partments: volar (superficial and deep)
and dorsal, and the mobile wad (Fig.
3). The volar compartment consists of
the six muscles responsible for flex-
ion, pronation, and supination: flex-
or carpi radialis, flexor pollicis lon-
gus, palmaris longus, flexor digitorum
superficialis, flexor carpi ulnaris, and
flexor digitorum profundus. This com-
partment also contains the median and
ulnar nerves along with the radial, ul-
nar, and anterior interosseous arter-
ies. The dorsal compartment contains
the extensor pollicis brevis, extensor
Vol 11, No 4, July/August 2003
Michael J. Fraipont, MD, and Gregory J. Adamson, MD
digitorum communis, and extensor
carpi ulnaris muscles, as well as the
posterior interosseous nerve and ar-
tery and perforators off the anterior
interosseous artery. The mobile wad
consists of three muscles: the brachio-
radialis, extensor carpi radialis lon-
gus, and extensor carpi radialis brevis.
Figure 2 Cross section of the left thigh, 10 to 15 cm inferior to the inguinal ligament. Note
the intermuscular fascial septa.
Pathophysiology
During strenuous exercise, muscle fi-
bers can swell to up to 20 times their
resting size, leading to a 20% increase
in the muscle volume and weight.4 In-
creased perfusing blood volume,
muscle hypertrophy, and interstitial
fluid volume within a nonexpanding
compartment increase pressure in ac-
cordance with Laplace’s law (a cap-
illary membrane subjected to internal
and external pressure reaches an
equilibrium based on those forces).
The blood flow through muscles is
chiefly regulated by the resistance of
the arteriole, which depends on the
tension in the vascular wall. The in-
crease in intramuscular pressure
causes a decrease in arteriolar blood
flow. Even though the circulation may
not be totally arrested, venous return
is markedly reduced and some cap-
illaries may become occluded.
When the blood flow is insufficient
to meet the requirements of the mus-
cle, the patient experiences pain with
continued activity. The symptoms of
CECS, which result from this is-
chemia, are caused by inadequate tis-
sue oxygenation from the decreased
venous return and insufficient perfu-
sion of muscle tissue. Because mus-
cles have blood flow only during the
269
Chronic Exertional Compartment Syndrome
Figure 3 Cross section of the middle of the left forearm distal to the level of the pronator
teres insertion.
relaxation phase of exercise, increased
intracompartmental pressures during
the relaxation phase are thought to
have the greatest effect on muscle is-
chemia. The most critical intracom-
partmental pressures are those
present when the muscle is not in a
contractile state. During this phase,
the balance between intramuscular
compartment pressure and the mi-
crovascular pressure determines the
adequacy of perfusion and, hence, the
oxygenation of the muscle. These val-
ues are best reflected by measuring
postexercise pressure. The patient
will continue to experience pain in the
affected extremity after exercise un-
til the total intramuscular pressure
decreases to a level at which the blood
flow can again meet the muscle’s re-
quirements.
Of patients with CECS involving
the legs, 39% to 46% have fascial de-
fects over the anterolateral lower leg
compared with asymptomatic indi-
viduals, who have <5% incidence.4,5
These fascial hernias or defects are
usually 1 to 2 cm2 in size and occur
near the intermuscular septum be-
tween the anterior and lateral com-
partments, often at the exit of the su-
270
perficial peroneal nerve. The fascial
hernia is approximately at the junc-
tion of the middle and distal thirds
of the leg. The superficial peroneal
nerve can be compressed by either the
edge of the fascial defect itself or the
muscle bulging through the defect. At
rest, no palpable abnormality may be
apparent, but with exercise, local ten-
derness and swelling may occur. Oc-
casionally Tinel’s sign may be found
at the site of the hernia.
It is not clear why patients with
CECS have increased total intramus-
cular pressure at rest and higher than
normal intramuscular pressure with
exercise compared with normal indi-
viduals. It is unlikely that a limited
osseofascial expansion can be the sole
explanation of this increase because,
after fasciotomy, the total intramus-
cular pressure at rest usually remains
higher than that in normal individ-
uals. In addition, while fascial herni-
as are a contributing anatomic find-
ing, fascial hernias are not present in
all patients with CECS. Arteriole reg-
ulation also may be a factor; howev-
er, it is likely that a combination of
anatomic limitations contributes to
the presence and severity of CECS.6
Journal of the American Academy of Orthopaedic Surgeons
Evaluation
History
During physical exertion, a patient
with CECS often notices pain that ini-
tially begins as a dull ache. If it is ig-
nored and the patient continues to
train, the pain increases to the point
that the activity must be stopped. The
onset and degree of the pain often be-
come both predictable and reproduc-
ible because the pain begins at about
the same time during the exercise ac-
tivity. The pain typically is well local-
ized to the entire affected compart-
ment.
Patients experience a feeling of
fullness or a cramplike sensation in
the affected compartment when they
attempt to exercise. They also may
complain of transient numbness, tin-
gling, or weakness in the motor and
sensory distributions of nerves with-
in the involved compartments. In
some cases, patients may have had a
recent increase in training time or in-
tensity that now takes them over their
threshold level for generating symp-
toms. Rest usually relieves the pain,
but it takes some time for complete
relief to occur, especially as the CECS
becomes more severe. Patients typi-
cally will not have persistent pain the
following day unless they exercise
again. Generally, they have no histo-
ry of trauma, and if they return to
their sport after discontinuing it for
some time, the symptoms typically
recur. Most patients present with bi-
lateral symptoms.7
Patients with CECS of the fore-
arm complain of a feeling of firm-
ness or cramping associated with
weakness in the hands and wrists
during vigorous athletic or repeti-
tive grasping activities. In addition,
they may experience numbness and
tingling. These symptoms can man-
ifest in the thenar, interosseous, or
hypothenar regions as well as in the
forearm. Symptoms resolve quickly
when the activity is discontinued
but recur with resumption of the ac-
tivity.

Physical Examination
Results of the physical examina-
tion of the lower extremity at rest are
usually normal. However, Rowdon et
al8 showed that athletes with CECS
demonstrated a contradictory elec-
tromyographic finding: decreased
postexercise potentiation of the pe-
roneal motor amplitude and mild im-
pairment in vibratory sensation. Di-
rect inspection and circumference
measurements are typically normal;
however, muscle atrophy may be
found if the condition is unilateral.
Results of physical examination of the
extremity after it has been provoked
by exercise may reveal tenderness
and increased tension in the involved
compartment. In addition, there may
be an associated decreased sensation
or tingling in the distal region.
In the upper extremity, results of
physical examination usually reveal
neither signs of nerve entrapment (eg,
a Tinel sign at the wrist or elbow) nor
abnormal two-point discrimination.
Results of the neurodiagnostic eval-
uation, including nerve-conduction
studies and electromyographs of the
ulnar and median nerves, also should
be normal, although Kutz et al9 re-
ported slowed median nerve conduc-
tion in one case. While muscle ten-
derness may be noted, symmetrically
functioning muscles in the hands and
forearms are usually found.
Differential Diagnosis
A number of different conditions
may overlap with the diagnosis of
CECS (Table 1). When the patient his-
tory, physical examination results,
and pressure measurements are not
diagnostic for CECS, consideration
should be given to further imaging,
neurophysiologic testing, and/or lab-
oratory studies.
Testing
Equipment and Criteria
Patients with CECS demonstrate
increased intracompartmental pres-
Vol 11, No 4, July/August 2003
Michael J. Fraipont, MD, and Gregory J. Adamson, MD
Table 1
Differential Diagnosis for Chronic Exertional Compartment Syndrome
Diagnosis Findings Confirmatory Studies
Stress fracture Localized tenderness Plain radiograph, bone
directly over the tibia; scan, MRI
pain with torsional or
bending stress
Medial tibial stress Manual resistance to Bone scan, MRI
syndrome (periostitis at active plantarflexion and
the muscular attachment inversion leading to pain
site along the along the distal
posteromedial tibia) posteromedial aspect of
the tibia; localized to
diffuse tibial tenderness
Chronic regional pain Allodynia and trophic Triple-phase bone
syndrome (reflex skin changes scan, thermography,
sympathetic dystrophy) sympathetic block
Tenosynovitis of the Tenderness along the MRI
ankle dorsiflexors or the extent of the tendon
posterior tibialis tendon aggravated by flexion
and extension maneuvers
Peripheral nerve Tingling or numbness EMG, nerve
entrapment syndromes associated with a specific conduction study
location (Tinel’s sign)
Venous stasis disease Trophic skin changes Duplex ultrasound
Deep vein thrombosis Palpable cords or pain Duplex ultrasound,
with plantarflexion; calf venogram
swelling
Radiculopathy Sensory losses, weakness EMG, central nervous
system evaluation
Arterial vascular disease Pain, paresthesias, and Ankle-brachial index
coolness with activities;
claudication
Popliteal artery Pain and coolness; Arteriogram
entrapment syndrome paradoxical claudication
EMG = electromyogram; MRI = magnetic resonance imaging
sures in the affected extremity at rest microcapillary infusion,5 and needle
and during and after exercise. Mea- manometer.16
suring intracompartmental pressures Many authors use the criteria of
during exercise is difficult and im- Pedowitz et al10 to evaluate patients.
practical; resting and postexercise These criteria are appropriate for eval-
measurements have been shown to be uation of both the upper and lower
the best method of confirming the di- extremities: a resting pressure mea-
agnosis of CECS.4,5,10-12 The type of surement ≥15 mm Hg, and/or a
exercise used during measurement measurement taken 1 minute after ex-
taking can vary, but it must be suf- ercise ≥30 mm Hg, and/or a measure-
ficiently provocative to induce symp- ment taken 5 minutes after exercise
toms. The following different com- ≥20 mm Hg. The criteria of Whitesides
partment measurement methods and Heckman17 for acute compartment
show equal effectiveness, assuming syndrome have been applied to CECS.
correct use: slit catheter,13 microtip Compartment ischemia is considered
pressure method,14 wick catheter,15 to occur when a compartment pres-
271
Chronic Exertional Compartment Syndrome

sure increases to 20 mm Hg below the

diastolic pressure.

Measurement Limitations
Factors that can affect the accura-
cy of pressure measurements include
proper use of the equipment, correct
anatomic placement of the catheter
tip, depth of needle insertion, posi-
tion of the extremity during pressure
measurement, and the contractile
force of the muscle. This process can
be especially difficult to control and
interpret in the clinical setting. There-
fore, care must be taken to place the
limb in a relaxed and consistent po-
sition for accurate, reproducible mea-
surements.
Although the measurement of in-
tracompartmental pressures of the
anterior compartment of the leg is rel-
atively simple, the same cannot be
said for the deep posterior compart-
ment or for the so-called fifth com-
partment, the tibialis posterior mus-
cle. When measuring deep posterior
compartment and tibialis posterior
muscle pressures, the exact location
of the tip of the catheter may vary.
Schepsis et al18 described a method
of placing the catheter medially, par-
allel to the posterior surface of the tib-
ia at the junction of the middle and
distal thirds of the leg, into the flexor
digitorum longus muscle. Wiley et
al19 proposed using ultrasound as a
guide for catheter placement into the
deep posterior compartment. Mollica
and Duyshart20 advocated placing in-
tracompartmental pressure measure-
ment apparatus in the medial foot
compartment. Upper extremity cath-
eter placement is determined by the
affected compartment.21,22
Figure 4 The single-incision or perifibular approach allows access to all four compartments
Other Testing Modalities through a lateral incision. A, The skin incision is made in line with and directly over the
fibula. B, Release of the lateral compartment (2) can be done directly after identification of
Alternative methods of testing for the intermuscular septum. Care must be taken to preserve the superficial peroneal nerve,
elevated compartment pressures are and the fasciotomy is done 1 cm posterior to the intermuscular septum. C, If the anterior
being considered, especially because compartment (1) needs to be released, the skin is retracted anteriorly and the fasciotomy is
done 1 cm anterior to the intermuscular septum. D, For the fasciotomy of the superficial pos-
of the difficulty in measuring the deep terior compartment (3), the skin is retracted posteriorly for exposure. E, For fasciotomy of
compartment pressures. Mohler et al23 the deep posterior compartment (4), the lateral and superficial posterior compartments are
found that patients with CECS of the retracted and the compartment is reached by following the interosseous membrane from the
posterior aspect of the fibula. (Adapted with permission from Rorabeck CH: A practical ap-
anterior compartment had greater proach to compartment syndromes: III. Management. Instr Course Lect 1983;32:102-113.)
deoxygenation of the muscle during

272 Journal of the American Academy of Orthopaedic Surgeons

 Michael J. Fraipont, MD, and Gregory J. Adamson, MD

exercise and delayed reoxygenation Surgical Techniques anterolateral aspect of the leg in its
of the muscle after exercise compared midportion between the tibial crest
with patients who did not have CECS, Anterior and Lateral Leg and the fibula (Fig. 4, A). After iden-
as measured by infrared spectrosco- Compartment Fasciotomy tification of the anterior intermuscu-
py. A more promising and practical Surgical release of the anterior and lar septum between the anterior and
measurement of elevated pressure is lateral compartments is done through lateral compartments (Fig. 4, B and
by magnetic resonance imaging, which a 10-cm longitudinal incision over the C), the fascia is divided proximally
can be used in diagnosing CECS.24 The
affected compartment shows an in-
crease in T2-weighted signal intensi-
ty during exercise. Although the as-
sistance of an experienced radiologist
in reviewing these subtle findings can
be helpful, the intracompartmental sig-
nal intensity can be normalized with
the signal intensity from surrounding
tissue not affected by CECS.24 Bone
scan technology using thallium Tl 201
single-photon emission computed to-
mography (SPECT) has been shown
to localize an ischemic compartment.25

Management
CECS occurs when athletes perform
an activity above their threshold lev-
el. Therefore, nonsurgical treatment
of CECS can be successful only when
the patient gives up the activity or the
activity level that causes the symp-
toms. However, it is not unreasonable
to offer a treatment plan that includes
stopping the activities that provoke
the symptoms while introducing a
different program of appropriate con-
ditioning. Nevertheless, because most
patients with CECS who seek med-
ical attention are unwilling to mod-
ify their exercise programs, subcuta-
neous fasciotomy of the involved
compartment should be considered.
It is the mainstay of treatment and is
successful in relieving pain and al-
lowing a return to full activities.12,26
Endoscopically assisted, two-
incision fasciotomy is an alternative
Figure 5 Either or both incisions from the two-incision fasciotomy technique can be
technique purported to be as safe and utilized depending on the number and location of affected compartments. A, Position of two
effective as single-incision fasciotomy.27 incisions (dotted line = posteromedial incision). B, Cross section of lower leg showing the
The advantages of endoscopic release relationship of the two incisions to the four compartments. 1 = anterior compartment,
2 = lateral compartment, 3 = superficial posterior compartment, 4 = deep posterior compart-
in the lower extremity are access to ment. C, The anterior intermuscular septum. D, The fascia is divided to separate the anterior
the entire length of the compartment and lateral compartments. The superficial peroneal nerve will be visualized. (Adapted with
and visualization of the superficial permission from Rorabeck CH: A practical approach to compartment syndromes: III. Man-
agement. Instr Course Lect 1983;32:102-113.)
peroneal nerve and its branches.27

Vol 11, No 4, July/August 2003 273

Chronic Exertional Compartment Syndrome

and distally in both compartments Thigh Compartment Fasciotomy between the lateral and posterior com-
under direct visualization (Fig. 4, D). Tarlow et al28 described a two- partments, the fascia is divided prox-
Care must be taken to identify the su- incision fasciotomy release in which imally and distally in both compart-
perficial peroneal nerve before re- the lateral incision is through the fas- ments under direct visualization. Care
lease. Fasciotomy should include in- cia lata and the iliotibial band (Fig. 6, must be taken to identify and palpate
spection for and release of any fascial A). Both the anterior and posterior the sciatic nerve. A separate medial
hernias. compartments can be addressed by incision is required to address the vas-
releasing the lateral intermuscular sep- tus medialis and the adductor mus-
Superficial and Deep Posterior Leg tum (Fig. 6, B). After identification of cles. After identification of the medi-
Compartment Fasciotomy the lateral intermuscular septum al intermuscular septum between the
The superficial posterior, deep
posterior, and tibialis posterior mus-
cle compartments can be released
through either an extended dissection
from the lateral approach (Fig. 4, D
and E) or more easily through a sep-
arate 10-cm medial incision (Fig. 5 ).
Once the muscular fascia is identified,
the superficial posterior compartment
can be released directly because it lies
more posterior to the other compart-
ments (Fig. 5, B). To reach the deep
posterior compartment, it is necessary
to undermine anteriorly to reach
the posterior tibial margin, thereby
avoiding the saphenous vein and
nerve as well as reaching the soleus
muscle. The soleus originates from
the entire proximal upper half of the
tibia and fibula, creating a soleus
bridge under which the deep poste-
rior compartment resides. The prox-
imal soleus attachment to the tibia
and fibula must be completely de-
tached to visualize the deep posteri-
or compartment. In addition to per-
forming a fasciotomy of the deep
posterior compartment, it is recom-
mended that a specific fasciotomy of
the tibialis posterior muscle compart-
ment be done, as well.11,12
Release of the deep posterior com-
partment of the leg has not been as
successful as that of the superficial
posterior compartment. The reasons
for this are not clear. Published expla-
nations11,12,18 for these reported fail-
ures include the fact that the patients Figure 6 Thigh compartment fasciotomy. A, Lateral incision site. B, Cross section of thigh
did not have CECS; the fasciotomy showing opening of the anterior compartment and release of the posterior compartment through
was incomplete, specifically not iden- the lateral intermuscular septum. C, Lateral view of the thigh showing the two-incision fas-
ciotomy release technique. (Panels A and B adapted with permission from Tarlow SD,
tifying and releasing the posterior tib- Achterman CA, Hayhurst J, Ovadia DN: Acute compartment syndrome in the thigh com-
ialis muscle within the deep compart- plicating fracture of the femur: A report of three cases. J Bone Joint Surg Am 1986;68:1439-1443.
ment; and dense scar tissue had Panel C adapted with permission from Azar FM, Pickering RM: Traumatic disorders, in Ca-
nale ST [ed]: Campbell’s Operative Orthopaedics, ed 9. St. Louis, MO: Mosby, 1998, vol 2, p 1408.)
formed after surgery.

274 Journal of the American Academy of Orthopaedic Surgeons


anterior and posterior compartments,
the fascia is divided proximally and
distally in both compartments under
direct visualization. Care must be tak-
en to identify and palpate the femo-
ral artery and nerve.
Forearm Compartment Fasciotomy
In a superficial volar forearm com-
partment fasciotomy, the incision be-
gins just above the elbow over the me-
dial antecubital fossa through the
entire length of the volar forearm in
a curvilinear fashion to the wrist. It
is important to release the lacertus fi-
brosus at the elbow and the carpal
tunnel at the wrist to decompress the
median nerve. The mobile wad com-
partment also may be released
through this incision and can be ad-
dressed as needed.29 In a dorsal com-
partment fasciotomy, a dorsal incision
is made in a line with the lateral as-
pect of the forearm connecting the lat-
eral epicondyle to the distal radioul-
nar joint.
Postoperative Care
Ice and elevation of the extremity
are used for 3 to 5 days after surgery
to help limit pain and excessive swell-
ing. Active range-of-motion exercises
should be instituted immediately af-
ter surgery. Crutches or an upper-
extremity sling may be used as nec-
essary for the first few postoperative
days, but patients should be encour-
aged to walk and perform light ac-
tivities without assistance. Weight bear-
ing as tolerated may be begun directly
after fasciotomies of the lower extrem-
References
1. French EB, Price WH: Anterior tibial
pain. BMJ 1962;2:1290-1296.
2. Mavor GE: The anterior tibial syndrome.
J Bone Joint Surg Br 1956;38:513-517.
3. Matava MJ, Whitesides TE Jr, Seiler JG
III, Hewan-Lowe K, Hutton WC: Deter-
mination of the compartment pressure
threshold of muscle ischemia in a ca-
nine model. J Trauma 1994;37:50-58.
4. Fronek J, Mubarak SJ, Hargens AR, et
Vol 11, No 4, July/August 2003
Michael J. Fraipont, MD, and Gregory J. Adamson, MD
ity. Full activities may begin as soon
as tolerated, usually 3 to 4 weeks af-
ter surgery.
Results
The results of compartment releases
indicate that most patients surgically
treated for CECS in the leg experience
a high level of pain relief and are sat-
isfied with the results of surgery. Re-
ports of improvement range from 81%
to 100%.4,11,12,18,26,30-32 However, authors
who differentiate the results of ante-
rior versus deep posterior compart-
ment releases report notably different
outcomes for the deep posterior com-
partment releases. Success of deep pos-
terior compartment release of the lower
extremity ranges from 50% to 65%.11,12,32
CECS in the deep posterior compart-
ment is multifactorial, and a fasciotomy
may not fully alleviate the cause of
the pain.11,12,18 Therefore, these out-
comes underscore the need to perform
compartment pressure measurements
before compartment releases are done
so that the correct compartment or
compartments can be identified and
adequate expectations can be relayed
to the patient.
Generally, patients have noted a
high level of pain relief and satisfac-
tion with the results of fasciotomy. In
their report on the subjective percent-
age of pain relief experienced by pa-
tients, Howard et al32 stated that re-
lief may come in increments of
improvement from the preoperative
level and is dependent on the indi-
al: Management of chronic exertional
anterior compartment syndrome of the
lower extremity. Clin Orthop 1987;220:
217-227.
5. Styf JR, Korner LM: Microcapillary in-
fusion technique for measurement of
intramuscular pressure during exer-
cise. Clin Orthop 1986;207:253-262.
6. Reneman RS: The anterior and the lat-
eral compartmental syndrome of the
vidual. Patients can expect to return
to light activity by 2 to 4 weeks and
to full activity by 4 to 6 weeks.
Complications
Complications of surgery for CECS
include hemorrhage, wound infec-
tion, nerve entrapment, swelling, ar-
tery injury, hematoma/seroma, lym-
phocele, peripheral cutaneous nerve
injury, and deep vein thrombosis.
Incidence ranges from 4.5% to
13%.4,11,12,19,31 In addition to postop-
erative complications, recurrence of
symptoms has been reported in 7%
to 17% of patients after surgical com-
partment release.11,12,18
Summary
Recurrent CECS is diagnosed with
accuracy when there is a history of
reproducible exertional pain associ-
ated with increased compartment
pressure measurements at rest and/
or after exercise. In patients with re-
current CECS, fasciotomy is advis-
able to allow a return to all activities.
At surgery, particular attention
should be paid to the careful release
of fascial defects anteriorly and lat-
erally in the leg and posteriorly in
the posterior tibialis muscle. As
well, when releasing compartments,
whether in the foot, leg, thigh, or
forearm, care must be taken to avoid
injuring the surrounding neurovas-
cular structures.
leg due to intensive use of muscles. Clin
Orthop 1975;113:69-80.
7. Jones DC, James SL: Overuse injuries of
the lower extremity: Shin splints, ilio-
tibial band friction syndrome, and ex-
ertional compartment syndromes. Clin
Sports Med 1987;6:273-290.
8. Rowdon GA, Richardson JK, Hoffmann
P, Zaffer M, Barill E: Chronic anterior
compartment syndrome and deep per-
275
Chronic Exertional Compartment Syndrome
oneal nerve function. Clin J Sports Med
2001;11:229-233.
9. Kutz JE, Singer R, Linday M: Chronic
exertional compartment syndrome of
the forearm: A case report. J Hand Surg
[Am] 1985;10:302-304.
10. Pedowitz RA, Hargens AR, Mubarak SJ,
Gershuni DH: Modified criteria for the
objective diagnosis of chronic compart-
ment syndrome of the leg. Am J Sports
Med 1990;18:35-40.
11. Rorabeck CH, Bourne RB, Fowler PJ: The
surgical treatment of exertional compart-
ment syndrome in athletes. J Bone Joint
Surg Am 1983;65:1245-1251.
12. Rorabeck CH, Fowler PJ, Nott L: The
results of fasciotomy in the manage-
ment of chronic exertional compart-
ment syndrome. Am J Sports Med 1988;
16:224-227.
13. Rorabeck CH, Castle GS, Hardie R,
Logan J: Compartmental pressure mea-
surements: An experimental investiga-
tion using the slit catheter. J Trauma
1981;21:446-449.
14. McDermott AG, Marble AE, Yabsley RH,
Phillips MB: Monitoring dynamic ante-
rior compartment pressures during ex-
ercise: A new technique using the STIC
catheter. Am J Sports Med 1982;10:83-89.
15. Mubarak SJ, Hargens AR, Owen CA,
Garetto LP, Akeson WH: The wick cath-
eter technique for measurement of in-
tramuscular pressure: A new research
and clinical tool. J Bone Joint Surg Am
1976;58:1016-1020.
16. Brace RA, Guyton AC, Taylor AE: Re-
evaluation of the needle method for
measuring interstitial fluid pressure.
Am J Physiol 1975;229:603-607.
276
17. Whitesides TE, Heckman MM: Acute
compartment syndrome: Update on di-
agnosis and treatment. J Am Acad Or-
thop Surg 1996;4:209-218.
18. Schepsis AA, Martini D, Corbett M:
Surgical management of exertional
compartment syndrome of the lower
leg: Long-term followup. Am J Sports
Med 1993;21:811-817.
19. Wiley JP, Short WB, Wiseman DA,
Miller SD: Ultrasound catheter place-
ment for deep posterior compartment
pressure measurements in chronic
compartment syndrome. Am J Sports
Med 1990;18:74-79.
20. Mollica MB, Duyshart SC: Analysis of
pre- and postexercise compartment pres-
sures in the medial compartment of the
foot. Am J Sports Med 2002;30:268-271.
21. Berlemann U, al-Momani Z, Hertel R:
Exercise-induced compartment syn-
drome in the flexor-pronator muscle
group: A case report and pressure mea-
surements in volunteers. Am J Sports
Med 1998;26:439-441.
22. Tompkins DG: Exercise myopathy of
the extensor carpi ulnaris muscle: Re-
port of a case. J Bone Joint Surg Am 1977;
59:407-408.
23. Mohler LR, Styf JR, Pedowitz RA, Har-
gens AR, Gershuni DH: Intramuscular
deoxygenation during exercise in pa-
tients who have chronic anterior com-
partment syndrome of the leg. J Bone
Joint Surg Am 1997;79:844-849.
24. Verleisdonk EJ, van Gils A, van der
Werken C: The diagnostic value of MRI
scans for the diagnosis of chronic exer-
tional compartment syndrome of the low-
er leg. Skeletal Radiol 2001;30:321-325.
Journal of the American Academy of Orthopaedic Surgeons
25. Takebayashi S, Takazawa H, Sasaki R,
Miki H, Soh R, Nishimura J: Chronic ex-
ertional compartment syndrome in
lower legs: Localization and follow-up
with thallium-201 SPECT imaging.
J Nucl Med 1997;38:972-976.
26. Styf JR, Korner LM: Chronic anterior-
compartment syndrome of the leg: Re-
sults of treatment by fasciotomy. J Bone
Joint Surg Am 1986;68:1338-1347.
27. Leversedge FJ, Casey PJ, Seiler JG III,
Xerogeanes JW: Endoscopically assist-
ed fasciotomy: Description of tech-
nique and in vitro assessment of lower-
leg compartment decompression. Am
J Sports Med 2002;30:272-278.
28. Tarlow SD, Achterman CA, Hayhurst J,
Ovadia DN: Acute compartment syn-
drome in the thigh complicating fracture
of the femur: Areport of three cases. J Bone
Joint Surg Am 1986;68:1439-1443.
29. Hoppenfeld S, de Boer P: The forearm,
in Surgical Exposures in Orthopaedics: The
Anatomic Approach, ed 2. Philadelphia,
PA, Lippincott-Raven, 1994, pp 125-127.
30. Qvarfordt P, Christenson JT, Eklof B,
Ohlin P, Saltin B: Intramuscular pres-
sure, muscle blood flow, and skeletal
muscle metabolism in chronic anterior
tibial compartment syndrome. Clin
Orthop 1983;179:284-290.
31. Detmer DE, Sharpe K, Sufit RL, Girdley
FM: Chronic compartment syndrome:
Diagnosis, management, and outcomes.
Am J Sports Med 1985;13:162-170.
32. Howard JL, Mohtadi NG, Wiley JP:
Evaluation of outcomes in patients fol-
lowing surgical treatment of chronic ex-
ertional compartment syndrome in the
leg. Clin J Sport Med 2000;10:176-184.
 Proximal Row Carpectomy and Intercarpal Arthrodesis
for the Management of Wrist Arthritis
John D. Wyrick, MD

Abstract
For advanced noninflammatory wrist arthritis, the most common surgical treatments tolunate articulation (Fig. 1, B).
to preserve motion are proximal row carpectomy and scaphoid excision with Chronic scaphoid nonunion fol-
capitohamate-lunotriquetral arthrodesis. Both procedures have documented success- lows a similar progression and has
ful outcomes. Proximal row carpectomy is simpler but typically is contraindicated been described as the scaphoid non-
when degeneration of the capitate head cartilage exists. Scaphoid excision with union advanced collapse (SNAC)
capitohamate-lunotriquetral arthrodesis is more complex but may provide greater wrist.6,7 Degeneration occurs between
grip strength and can be successful in the presence of capitate degeneration. Treat- the radius and the distal pole of the
ment selection should be based on surgeon preference and experience as well as on scaphoid and progresses to involve
the patient’s understanding of the possible complications and benefits of each procedure. the capitate–proximal pole articula-
J Am Acad Orthop Surg 2003;11:277-281 tion, while sparing the proximal pole–
radius articulation. The radiolunate
joint typically is spared, even with ad-
vanced degenerative changes. This
Advanced arthritis of the wrist can row acts as a mobile, intercalated seg- predictable distribution of arthritis is
cause notable disability. It typically af- ment between the rigid distal carpal the basis of Watson’s SLAC wrist pro-
fects middle-aged men during their row and the radius. When ligament cedure (capitohamate-lunotriquetral,
prime working years. Because joint re- disruption occurs between the bones or four-corner, arthrodesis combined
placement is associated with marked of the proximal row, wrist instability with scaphoid excision), which trans-
complications in active individuals, develops. Disruption of the scapholu- fers the loads across the wrist to the
and because total wrist arthrodesis nate interosseous ligament causes the preserved radiolunate joint.4
may result in unacceptable compro- lunate to dorsiflex or extend, creating
mise of function, other motion-pre- the so-called dorsiflexed intercalated
serving procedures, such as proximal segment instability.2 The scaphoid Management
row carpectomy (PRC) and scaphoid flexes palmarly, causing an abnormal
excision with intercarpal arthrodesis, distribution of forces across the ellip- A staged approach is commonly rec-
have become popular options.1 Acom- tical radioscaphoid joint and increased ommended for treatment of the SLAC
parison of the indications, techniques, contact pressures, resulting in progres- wrist.7-9 A stage I SLAC wrist often
and results of these two procedures sive degenerative changes.3 The se- can be managed with splints, nonste-
can help elucidate their relative dif- quential progression of arthritis that roidal anti-inflammatory medica-
ferences and aid in appropriate pa- follows is known as the scapholunate
tient selection. advanced collapse (SLAC) wrist.4,5
Radiographic progression of the
Dr. Wyrick is Assistant Professor, Department of
SLAC wrist is divided into three stages. Orthopaedic Surgery, Director, Division of Up-
Pathoanatomy Stage I consists of degeneration be- per Extremity Surgery, and Associate Director,
tween the radial styloid and the dis- Division of Orthopaedic Trauma, University of
Primary osteoarthritis is relatively rare. tal pole of the scaphoid. Stage II de- Cincinnati, Cincinnati, OH.
Most cases of radiocarpal arthritis are scribes the progression to involvement
Reprint requests: Dr. Wyrick, 231 Albert Sabin
secondary to structural changes that of the proximal pole of the scaphoid Way, Cincinnati, OH 45267-0212.
result in abnormal mechanics and are and scaphoid fossa (Fig. 1, A). In stage
usually caused by chronic scapholu- III, the capitate migrates proximally Copyright 2003 by the American Academy of
nate dissociation or rotary subluxation between the scaphoid and the lunate, Orthopaedic Surgeons.
of the scaphoid. The proximal carpal resulting in degeneration at the capi-

Vol 11, No 4, July/August 2003 277

Proximal Row Carpectomy and Intercarpal Arthrodesis

capitate articulation relies on trans-

lation as well as rotation for wrist mo-
tion; this is because the curvature radius
of the capitate is smaller than that of
the radiolunate fossa. The translation
may allow more dissipation of forces
and decreased wear at the radiocapi-
tate joint.11 Reports of wrist weakness
and instability were the main argu-
ment against PRC. Although this sub-
jective complaint was reported in older
studies, more recently good recovery
of grip strength and return to occu-
pations requiring manual labor have
been demonstrated.8,11-17 Degeneration
may occur over time at the new ar-
ticulation, but symptoms usually are
mild (Fig. 3).
The primary indication for PRC is
Figure 1 A, Anteroposterior radiograph of a stage II SLAC wrist with arthritis at the radial pain relief in patients with radiocar-
styloid and joint space narrowing between the proximal pole of the scaphoid and the radius
(arrow). (Reprinted with permission from Wyrick JD, Stern PJ, Kiefhaber TR: Motion-preserving pal arthritis, which usually results
procedures in the treatment of scapholunate advanced collapse wrist: Proximal row carpec- from a SLAC or SNAC wrist after fail-
tomy versus four-corner arthrodesis. J Hand Surg [Am] 1995;20:965-970.) B, Stage III SLAC ure of nonsurgical treatment. The
wrist with degenerative changes at the capitolunate joint (arrow) and the radioscaphoid joint
(asterisk). (Reprinted with permission from Wyrick JD: Wrist reconstruction: Intercarpal ar- SLAC or SNAC wrist can have mul-
throdesis and proximal row carpectomy. Seminars in Arthroplasty 1997;8:172-179.) tiple etiologies, including scapho-
lunate dissociation, scaphoid non-
union, and residual incongruencies
tions, or cortisone injections, although wrist by converting it from a complex from a perilunate dislocation. Patients
a radial styloidectomy can relieve link system to a simple ball-and-socket with calcium pyrophosphate deposi-
symptoms in older, less active pa- joint. Despite favorable reports, PRC tion disease also may develop a SLAC
tients. For both stages I and II, PRC remained controversial and poorly ac- pattern of arthritis that may be ame-
is preferred because it is simpler to cepted until the last decade. PRC works nable to PRC. Other inflammatory ar-
do than the more technically difficult well and endures because the radio- thritides, however, such as rheuma-
four-corner arthrodesis, has less po-
tential for complications, has better
motion recovery, and has proven du-
rable results. The four-corner arthro-
desis is preferred in the stage III SLAC
wrist with advanced capitolunate ar-
thritis. This obviates concerns regard-
ing the degenerative capitolunate
joint and uses the preserved articu-
lation between the lunate and radius
to transfer joint forces. Either proce-
dure can be relatively easily convert-
ed to a full wrist arthrodesis should
the primary procedure fail and symp-
toms persist (Fig. 2).

Figure 2 Failed four-corner arthrodesis. A, Preoperative anteroposterior radiograph of a

Proximal Row Carpectomy
Stamm10 in 1944 described PRC as a
method of restoring function to the
stage II SLAC wrist. B, Patient continued to have pain 1 year after successful arthrodesis.
C, Pain relief achieved after conversion to total wrist fusion. (Reprinted with permission from
Wyrick JD, Stern PJ, Kiefhaber TR: Motion-preserving procedures in the treatment of scapho-
lunate advanced collapse wrist: Proximal row carpectomy versus four-corner arthrodesis.
J Hand Surg [Am] 1995;20:965-970.)

278 Journal of the American Academy of Orthopaedic Surgeons

 John D. Wyrick, MD

to performing a four-corner fusion, as cision also has been described.20 The

Figure 3 Anteroposterior radiograph of a
stage II SLAC wrist 8 years after undergoing
PRC (same patient as in Figure 1, A). (Reprint-
ed with permission from Wyrick JD, Stern PJ,
Kiefhaber TR: Motion-preserving procedures
in the treatment of scapholunate advanced
collapse wrist: Proximal row carpectomy ver-
sus four-corner arthrodesis. J Hand Surg
[Am] 1995;20:965-970.)
toid arthritis, have not responded to
PRC, and the procedure therefore is
not indicated.17 Prerequisites for PRC
are preservation of the capitate head
articular cartilage and the lunate fac-
et. However, Eaton et al18 reported sat-
isfactory results with the use of fas-
cial interposition even in patients with
more advanced changes.
they are for performing a PRC. (3) The
muscle-tendon units that cross the
wrist maintain their relative lengths,
thus decreasing the risk of weakness.4,5
Calandruccio et al19 described a mod-
ification of the procedure in which
only the capitate and lunate are fused
and both the scaphoid and triquetrum
are excised.
Arthritis secondary to either SLAC
or SNAC is the primary indication for
the four-corner arthrodesis with ex-
cision of the scaphoid (Fig. 4). Sim-
ilar to PRC, four-corner arthrodesis
is not indicated in the patient with in-
flammatory wrist arthritis and is con-
traindicated if the radiolunate artic-
ulation is degenerated.
Surgical Techniques
Proximal Row Carpectomy
The carpus is typically exposed
through a dorsal longitudinal incision
through the third dorsal compart-
ment; however, a transverse dorsal in-
lunate is excised first because it is usu-
ally the easiest. Whenever possible,
the triquetrum and scaphoid are then
excised sharply; if that is not possi-
ble, they can be excised piecemeal. A
threaded Kirschner wire (K-wire) can
be used as a joystick to gain purchase
on these small bones. Care should be
taken to preserve the radiocarpal lig-
aments to prevent postoperative ul-
nar translocation of the carpus. Most
surgeons no longer routinely do a ra-
dial styloidectomy during this proce-
dure. Some authors have reported a
slight increase in radial deviation af-
ter styloidectomy, but overall out-
come is not notably improved.11,17 The
dorsal wrist capsule is then repaired
with sutures. Pin fixation of the car-
pus to the radius is not recommend-
ed because it does not improve out-
come and is a frequent source of
complications (eg, infection, migra-
tion).8 The patient is immobilized
postoperatively for a minimum of 2
weeks in a short arm cast, primarily
for comfort, although slightly longer

Scapholunate Advanced
Collapse Procedure
The management procedure proposed
by Watson and Ballet4 takes advan-
tage of the spared radiolunate joint
by redirecting the forces across the
wrist through the remaining healthy
articulation. The proposed advantag-
es of the SLAC procedure over PRC
include the following: (1) There is a
better match of the radius of curva-
ture of the lunate with the radius com- Figure 4 Four-corner arthrodesis. A, Preoperative anteroposterior radiograph in a patient
with stage III SLAC wrist. B, Three-year follow-up anteroposterior radiograph demonstrates
pared with the capitate (ie, the nat- successful arthrodesis with excellent results. (Reprinted with permission from Wyrick JD,
ural articulation is retained). (2) Stern PJ, Kiefhaber TR: Motion-preserving procedures in the treatment of scapholunate ad-
Degenerative changes at the capitolu- vanced collapse wrist: Proximal row carpectomy versus four-corner arthrodesis. J Hand Surg
[Am] 1995;20:965-970.)
nate joint are not a contraindication

Vol 11, No 4, July/August 2003 279

Proximal Row Carpectomy and Intercarpal Arthrodesis
periods of immobilization have been
recommended. Radiographs are ob-
tained after 2 weeks to make sure
the capitate is located in the lunate
facet of the radius. The patient is
placed in a removable splint and
started on range-of-motion and
strengthening exercises; activities
are progressed as tolerated. Maxi-
mum strength may take 1 year or
longer to achieve.
Four-Corner Arthrodesis With
Scaphoid Excision
The wrist is exposed in a fashion
similar to that described for a PRC.
The scaphoid is then excised either
sharply or, more often, piecemeal. Care
should be taken not to transect the ra-
dioscaphocapitate ligament. If it is cut,
there is a risk of destabilizing the re-
maining carpus and of possible ulnar
translocation. The opposing surfaces
of the capitate, hamate, lunate, and
triquetrum are decorticated and the
interstices filled with cancellous bone
graft. Iliac bone graft is preferred, but
the distal radius can be used as the
donor site. The anatomic position of
the four bones is maintained by leav-
ing the volar capsular attachments and
ligaments intact. The lunate is reduced
relative to the capitate head to correct
any extension malalignment of the lu-
nate or carpal collapse. The bones are
then stabilized with K-wires, staples,
or small bone screws. Some newer im-
plant devices have been developed spe-
cifically for this procedure, but there
are no published series evaluating their
efficacy.ASilastic scaphoid was inserted
in the original description of the SLAC
procedure.4 However, scaphoid re-
placement is unnecessary and to avoid
the potential for silicone synovitis is
not recommended.1,7-9 An alternative
technique is to excise the scaphoid and
triquetrum and fuse the capitate to the
lunate with small bone screws.19 Re-
sults with this technique are compa-
rable to those in other series of scaphoid
excision and capitohamate-lunotri-
quetral arthrodesis. The wrists are im-
mobilized postoperatively for 6 to 8
280
Table 1
Proximal Row Carpectomy Results
Study No. of Wrists
Krakauer et al7 12
Wyrick et al8 11
Imbriglia et al11 27
Neviaser12 31
Crabbe13 20
Jorgensen14 22
Inglis and Jones15 12
Tomaino et al16 23
Culp et al17 17
Totals 175
*Success generally equated with good or excellent study results. Failure generally equat-
ed with poor or fair study results.
(Reprinted with permission from Wyrick JD: Wrist reconstruction: Intercarpal arthro-
desis and proximal row carpectomy. Seminars in Arthroplasty 1997;8:172-179.)
weeks in a short arm cast, after which
the pins are removed. The duration
of immobilization may be modified
depending on radiographic evidence
of fusion. A removable splint is then
fashioned, and the patient is started
on range-of-motion and strengthen-
ing exercises.
Results
The PRC and four-corner arthrode-
sis with scaphoid excision are both
good procedures for preserving mo-
tion in the painful posttraumatic ar-
thritic wrist. However, there is con-
troversy as to which procedure is
more appropriate. The PRC has been
used longer, and clinical results
have been documented in multiple
series.7-17 The most compelling evi-
dence in support of the procedure is
the number of series that corrobo-
rate excellent results (Table 1). Be-
cause it was introduced later, the
SLAC procedure has shorter fol-
low-up periods in published re-
ports, although potential advan-
tages have been indicated.4
Three clinical series have directly
compared the two procedures and
Journal of the American Academy of Orthopaedic Surgeons
Success* Failure*
9 3
11 0
26 1
29 2
17 3
19 3
12 0
20 3
14 3
157 (90%) 18 (10%)
help to define the criteria for choos-
ing between the two.7-9 The data from
these series are similar. Based on the
mean of the outcomes from the three
studies, range of motion was consis-
tently better in the PRC group, aver-
aging 64% of the opposite side com-
pared with 45% in the four-corner
arthrodesis group.7-9 Grip strength
varied between the series, but the
mean of approximately 75% of the op-
posite side was not statistically dif-
ferent.7-9 Pain relief and patient sat-
isfaction were good for both groups;
however, the failure rate differed some-
what between the series. Tomaino et
al9 and Krakauer et al7 reported an ap-
proximate failure rate of 20% after
PRC, with failure defined as an un-
satisfied patient or conversion to
wrist fusion. The failure rate after
four-corner arthrodesis was much
lower, with reported rates of 0%9 and
7%.7 Wyrick et al8 had a 30% failure
rate after four-corner arthrodesis and
unanimous patient satisfaction after
PRC. The small patient populations
probably account for the sometimes
substantial percentage difference in
outcomes between the three series.
However, Cohen and Kozin21 report-
ed no clear advantage of one method

over the other in their comparative
study of the two procedures.
Risk of complications is the other
factor to be considered when choos-
ing a procedure. A markedly higher
number of patients have required fur-
ther surgery to treat complications af-
ter four-corner arthrodesis. Most of
the complications are implant-related,
many caused by the use of staples.7,8
Symptomatic nonunions also can oc-
cur, but they are uncommon (<4%) as
long as the arthrodesis includes the
hamate and triquetrum rather than
just fusion of the capitate to the lu-
nate.19,22 Calandruccio et al19 suggest
that capitolunate fusion can be reli-
ably achieved, but Kirschenbaum et
al22 found the nonunion rate to be un-
acceptable. The difference may be re-
References
1. Smith RJ, Atkinson RE, Jupiter JB: Sili-
cone synovitis of the wrist. J Hand Surg
[Am] 1985;10:47-60.
2. Linscheid RL, Dobyns JH, Beabout JW,
Bryan RS: Traumatic instability of the
wrist: Diagnosis, classification, and
pathomechanics. J Bone Joint Surg Am
1972;54:1612-1632.
3. Burgess RC: The effect of rotatory sub-
luxation of the scaphoid on radio-
scaphoid contact. J Hand Surg [Am]
1987;12(5 pt 1):771-774.
4. Watson HK, Ballet FL: The SLAC wrist:
Scapholunate advanced collapse pat-
tern of degenerative arthritis. J Hand
Surg [Am] 1984;9:358-365.
5. Watson HK, Ryu J: Evolution of arthri-
tis of the wrist. Clin Orthop 1986;202:
57-67.
6. Vender MI, Watson HK, Wiener BD, Black
DM: Degenerative change in symptom-
atic scaphoid nonunion. J Hand Surg [Am]
1987;12:514-519.
7. Krakauer JD, Bishop AT, Cooney WP:
Surgical treatment of scapholunate ad-
vanced collapse. J Hand Surg [Am] 1994;
19:751-759.
8. Wyrick JD, Stern PJ, Kiefhaber TR:
Motion-preserving procedures in the
treatment of scapholunate advanced
Vol 11, No 4, July/August 2003
lated to the fixation method, with
headless screws performing better
than staples or K-wires.
Summary
A staged approach can be used to
manage advanced wrist arthritis. A
stage I SLAC wrist often can be man-
aged nonsurgically with splints, non-
steroidal anti-inflammatory drugs, or
cortisone injections. Occasionally, a
radial styloidectomy can relieve the
symptoms in the older, less active pa-
tient. If this fails or, more commonly,
when the patient presents with a
stage II SLAC wrist with a preserved
capitolunate joint, PRC is preferred
because it is technically less demand-
collapse wrist: Proximal row carpecto-
my versus four-corner arthrodesis.
J Hand Surg [Am] 1995;20:965-970.
9. Tomaino MM, Miller RJ, Cole I, Burton
RI: Scapholunate advanced collapse
wrist: Proximal row carpectomy or
limited wrist arthrodesis with scaphoid
excision? J Hand Surg [Am] 1994;19:
134-142.
10. Stamm TT: Excision of the proximal
row of the carpus. Proc R Soc Med 1944;
38:74-75.
11. Imbriglia JE, Broudy AS, Hagberg WC,
McKernan D: Proximal row carpecto-
my: Clinical evaluation. J Hand Surg
[Am] 1990;15:426-430.
12. Neviaser RJ: On resection of the prox-
imal carpal row. Clin Orthop 1986;202:
12-15.
13. Crabbe WA: Excision of the proximal
row of the carpus. J Bone Joint Surg Br
1964;46:708-711.
14. Jorgensen EC: Proximal-row carpecto-
my: An end-result of twenty-two cases.
J Bone Joint Surg Am 1969;51:1104-1111.
15. Inglis AE, Jones EC: Proximal-row
carpectomy for diseases of the proximal
row. J Bone Joint Surg Am 1977;59:
460-463.
16. Tomaino MM, Delsignore J, Burton RI:
John D. Wyrick, MD
ing and yields durable results. The
patient should be made aware that as
much as 1 year may be needed to
achieve full rehabilitation of the hand
and wrist. When salvage is necessary,
a PRC can be relatively easily convert-
ed to a wrist fusion.7 For the patient
with a stage III SLAC wrist, the crit-
ical decision is whether a complete or
a partial wrist arthrodesis is more ap-
propriate. This decision is usually
predicated on patient factors (eg,
whether motion needs to be pre-
served). If motion is a priority,
scaphoid excision and four-corner ar-
throdesis can provide excellent re-
sults. Conversion to a total wrist ar-
throdesis can be relatively easily
accomplished as a salvage procedure
when pain relief is unsatisfactory.
Long-term results following proximal
row carpectomy. J Hand Surg [Am] 1994;
19:694-703.
17. Culp RW, McGuigan FX, Turner MA,
Lichtman DM, Osterman AL, McCar-
roll HR: Proximal row carpectomy: A
multicenter study. J Hand Surg [Am]
1993;18:19-25.
18. Eaton RG, Akelman E, Eaton BH: Fas-
cial implant arthroplasty for treatment
of radioscaphoid degenerative disease.
J Hand Surg [Am] 1989;14:766-774.
19. Calandruccio JH, Gelberman RH, Dun-
can SF, Goldfarb CA, Pae R, Gramig W:
Capitolunate arthrodesis with scaphoid
and triquetrum excision. J Hand Surg [Am]
2000;25:824-832.
20. Taleisnik J (ed): The Wrist. New York,
NY: Churchill Livingstone, 1985, pp
131-137.
21. Cohen MS, Kozin SH: Degenerative ar-
thritis of the wrist: Proximal row carpec-
tomy versus scaphoid excision and four-
corner arthrodesis. J Hand Surg [Am] 2001;
26:94-104.
22. Kirschenbaum D, Schneider LH, Kirk-
patrick WH, Adams DC, Cody RP:
Scaphoid excision and capitolunate ar-
throdesis for radioscaphoid arthritis.
J Hand Surg [Am] 1993;18:780-785.
281
 Metastatic Bone Disease of the Humerus
Frank J. Frassica, MD, and Deborah A. Frassica, MD

Abstract
Metastatic bone disease is the most common cause of destructive bone lesions in secondary to bone metastases or to
adults, and involvement of the humerus is common. Patients with destructive le- nononcologic sources, such as gleno-
sions involving <50% of the cortex are treated nonsurgically with external beam humeral arthritis or rotator cuff ten-
irradiation. Patients with diaphyseal lesions involving ≥50% of the cortex or those dinosis, which also manifest as dif-
with pain after irradiation can be treated with intramedullary nailing to achieve fuse discomfort, night pain, difficulty
rigid fixation. Although closed intramedullary nailing is used most often, open nail- with sleeping on the affected side, and
ing with methylmethacrylate is appropriate for destructive lesions in which rigid limited use of the upper extremity
fixation cannot be achieved with closed nailing. Plate fixation is acceptable when secondary to pain. Plain radiographs
adequate proximal and distal cortical bone is present for screw purchase, although are the first step in evaluation. An an-
proximal humeral lesions usually are treated with prosthetic arthroplasty. Postop- teroposterior view of the shoulder
erative external beam irradiation can help prevent disease progression and subse- and humerus is done to assess the
quent loss of fixation. However, when disease progression persists or rigid internal proximal half of the humerus and
fixation is not feasible because of extensive bone destruction, wide resection and re- scapula, and a scapular axillary view
construction with a custom prosthesis can be done. is made to evaluate the glenoid and
J Am Acad Orthop Surg 2003;11:282-288 coracoid process. If the discomfort ex-
hibits a radicular pattern or if the
shoulder pain extends proximally
into the neck, radiographs of the cer-
Metastases to bone are the most fre- Diagnosis vical spine should be obtained. If
quent cause of destructive lesions to plain radiographs do not show a de-
the skeleton in adults. The most com- Presentation and Evaluation structive lesion and bone metastases
mon primary malignancies that me- Several distinct presentations of up- are suspected, technetium Tc 99m
tastasize to bone are breast, lung, kid- per extremity metastatic disease can (99mTc) bone scanning or magnetic res-
ney, and prostate carcinoma. The typical occur. Patients with known metastatic onance imaging (MRI) may be done.
99mTc bone scanning is an excellent
distribution of metastatic lesions is to disease may have either an asymp-
the spine, ribs, pelvis, and proximal tomatic or a painful and disabling le- modality for screening the cervical
limb girdles.1 However, almost any sion. A small percentage without a spine and shoulder girdle, especially
primary malignancy may metastasize known history of cancer may present when previous scans are available for
to bone, and any bone in the body may with a destructive bone lesion second- comparison. MRI of the shoulder can
be involved. In the upper extremity, ary to an occult primary tumor. Com- differentiate pain secondary to rota-
the most common location is the hu- plete fractures secondary to metastatic
merus, usually the proximal third or lesions often occur after very minor
the diaphysis. Lesions in the distal third trauma, such as rolling over the arm Dr. F. Frassica is Chairman and Robert A. Rob-
inson Professor, Department of Orthopaedic Sur-
of the humerus are less common and in bed, a minor fall, opening a jar, or
gery, Johns Hopkins University, Baltimore, MD.
typically occur in patients with my- other simple activities. The presence Dr. D. Frassica is Assistant Professor of Oncol-
eloma or lung or renal carcinomas. Di- of substantial arm or shoulder pain ogy, Department of Radiation Oncology, Johns
agnosis of metastatic humeral bone in the patient with a history of can- Hopkins University.
lesions and alternatives for treatment cer indicates the possibility of bone
Reprint requests: Dr. Frank J. Frassica, c/o Elaine
vary depending on whether the pa- metastases. Common characteristics
P. Henze, Room A672, 4940 Eastern Avenue,
tient has a large impending lesion or include pain at rest, night pain, and Baltimore, MD 21224-2780.
complete fracture, which region of the pain unresponsive to anti-inflam-
bone is affected, the extent of overall matory medications and narcotics. Copyright 2003 by the American Academy of
disease, the histologic diagnosis, and During evaluation, the clinician Orthopaedic Surgeons.
the nature of prior treatment. must determine whether the pain is

282 Journal of the American Academy of Orthopaedic Surgeons


tor cuff tendinosis from that of met-
astatic disease and can identify bone
marrow infiltration by tumor cells
and rotator cuff inflammation. Short
tau inversion recovery (STIR) or a T2-
weighted fast spin-echo fat-saturated
sequence is the optimal method for
assessing tumor presence. Although
computed tomography (CT) is sensi-
tive in detecting cortical bone destruc-
tion in the scapula and humerus, its
use is limited because it cannot de-
tect marrow invasion in the absence
of bone destruction. However, CT is
more accurate than other diagnostic
tools in determining structural com-
promise.
Criteria for Impending Fracture
Determining the risk for patholog-
ic fracture is subjective and depends
on many factors, including the pat-
tern of bone destruction, location in
the bone, response of the host bone,
and anticipated loading conditions.
Most classification systems2-6 used to
predict areas at risk of fracture are
based on the amount of cortical bone
destruction measured on anteropos-
terior and lateral radiographs. Bone
destruction of 25% of the cortical di-
ameter has a low risk of fracture (Fig.
1, A); bone destruction of 75% of the
cortical diameter is associated with a
high risk of fracture.2-4 When bone de-
struction is between these extremes,
the risk of fracture is more difficult to
predict. Most surgeons consider that
≥50% bone destruction indicates im-
pending fracture (Fig. 1, B).
Mirels4 developed a 12-point scor-
ing system based on the location,
type, and amount of bone destruction
and the presence or absence of ac-
tivity-related pain. Combined scores
of 9, 8, and 7 respectively had a 33%,
15%, and 4% risk of fracture. There
was a low risk of fracture in patients
with <50% cortical bone destruction.
Although this is an objective scoring
method, many surgeons do not use
the Mirels system because of the sub-
jectivity of the variables and the sub-
stantial overlap between the fracture
Vol 11, No 4, July/August 2003
Frank J. Frassica, MD, and Deborah A. Frassica, MD
Figure 1 A, Anteroposterior radiograph of
the proximal humerus in a patient with breast
cancer showing a predominantly blastic le-
sion with a small amount of cortical bone de-
struction. This patient would be an excellent
candidate for external beam irradiation.
B, Anteroposterior radiograph of a humeral
lesion in a patient with metastatic prostate
cancer showing ≥50% cortical bone destruc-
tion. This lesion meets the criteria of impend-
ing fracture.
and nonfracture groups, which un-
derscores the difficulty in predicting
risk of fracture.
Mirels4 showed that purely lytic
bone metastases have a much higher
risk of fracture than do purely blas-
tic metastases (often seen in patients
with metastatic breast and prostate
carcinoma). However, many patients
have a combination of lytic and blas-
tic metastases. Lesions in the proxi-
mal humeral metaphysis are less
prone to fracture than are those in the
diaphysis or in the transition zone be-
tween the metaphysis and diaphysis
(in the region of the insertion of the
pectoralis major muscle).
The anticipated loading of the up-
per extremity is also an important
consideration. A patient with an iso-
lated metastasis in the humerus can
easily protect against overloading the
upper extremity during the course of
irradiation and/or chemotherapy. In
contrast, a patient with substantial
concomitant lower-extremity disease
requiring protected weight bearing
with crutches or a walker may sub-
ject the upper extremity to increased
loading. In such a situation, internal
fixation may be necessary to prevent
fracture.
Nonsurgical Management
Most humeral metastases that have
not fractured can be managed with
external beam irradiation without
surgery. The decision may be depen-
dent on the histology of the tumor.
Patients are counseled to avoid pro-
vocative activities that may lead to
fracture while the lesion heals. Pa-
tients generally undergo 1 week to 2
weeks of external beam irradiation
(usually 3,000 cGy in 10 fractions).7,8
If the isolated lesion occurs in a pa-
tient with an excellent prognosis, a
longer course of treatment may be
recommended in an effort to provide
more durable local control. Treatment
of humeral metastases with a single
dose of 800 cGy can be successful in
reducing pain in patients with termi-
nal disease and short life expec-
tancies;9-12 this is especially useful
when patient transport to the radia-
tion therapy facility is difficult. Al-
though tumor progression generally
halts after the completion of irradi-
ation, activity should be modified for
2 to 3 months. Patients can continue
activities of daily living such as eat-
ing, cleaning, bathing, and changing
clothes, but exertions such as tennis,
changing a tire, opening tight jars,
overhead throwing, and swinging an
ax are discouraged.
Patients with complete fractures
are poor candidates for nonsurgical
treatment. Fracture braces and casts
are not effective in controlling dis-
comfort, and patients avoid using
their extremities because of pain. In
contrast with nonpathologic humer-
al fractures, which heal quickly, com-
plete fractures secondary to metastat-
ic bone disease heal very slowly if it
all. Flemming and Beals13 and Doug-
lass et al14 reported poor results with
283
Metastatic Bone Disease of the Humerus
nonsurgical management of patho-
logic humeral fractures and later
characterized the results of closed
management as “unsatisfactory, pro-
ducing limited use, incomplete pain
relief, and unpredictable healing.”13
Surgical Management
Preoperative Planning
The general medical condition of
the patient must be assessed before
surgery. Terminally ill patients (antic-
ipated survival, <2 to 3 weeks) are
poor candidates, unlikely to benefit
from the surgery. Although there are
no absolute criteria to predict surviv-
al of the patient with metastatic bone
disease, poor prognostic factors in-
clude hypercalcemia, substantial cy-
topenia from bone marrow failure,
cachexia, and poor performance sta-
tus.
Careful preoperative planning is
essential. The cervical spine should
be carefully assessed for destructive
lesions so that injury can be avoided
while anesthetizing or positioning the
patient, and plain radiographs or a re-
cent 99mTc bone scan should be re-
viewed. The entire humerus should
be viewed with plain radiography in
two orthogonal planes to determine
if there are multiple lesions. Position-
ing the end of the fixation device at
a site of diseased bone must be avoid-
ed so that fracture does not occur in
this transition zone when the patient
begins using the extremity. CT and
MRI scans generally are not needed
for preoperative planning. However,
if plain radiographs are equivocal as
to the presence of cortical destruction
in regions designated to receive the
fixation device, MRI can be used to
confirm the presence of disease.
Because patients with metastatic
bone disease may survive only 3 to
12 months, the goal of surgery is to
attain rigid and durable internal fix-
ation and, accordingly, immediate
postoperative use of the upper ex-
tremity. Rigid fixation can be
284
achieved with a variety of internal fix-
ation or prosthetic devices. These de-
vices can be used with or without
methylmethacrylate. Fracture healing
should not be necessary to achieve
functional stability.
Device Selection
The selection of the reconstruction
device, such as an intramedullary
nail, plate, or prosthesis, depends on
the area of humeral involvement and
the degree of bone destruction. When
selecting a fixation method, it is con-
venient to divide the humerus into
three regions: (1) proximal metaphy-
seal, (2) metadiaphyseal and diaphy-
seal, and (3) distal metadiaphyseal
and metaphyseal (supracondylar)
(Fig. 2).
Proximal Metaphyseal Region
Complete or impending fractures
of the proximal humerus usually are
managed with a humeral endopros-
thesis. Intramedullary nails are diffi-
cult to use in the proximal metaphy-
sis because rigid proximal fixation
cannot be achieved. Plate fixation is
likewise ineffective for solid fixation
because of the thin and compromised
cortical bone.
The surgical procedure is similar
to that for a nonpathologic fracture,
in which a deltopectoral approach is
used to osteotomize the humeral
head. For a pathologic fracture, the
proximal fragments are excised. The
proximal humeral metaphysis is
carefully curetted to remove all of
the gross tumor but not the cortical
shell or periosteal tissues. The hu-
merus is prepared through the en-
tire diaphysis to receive a long-stem
prosthesis. The prosthesis is care-
fully cemented in place so that the
cement does not enter the soft tis-
sues; extravasation through hu-
meral defects might result in neuro-
logic or vascular injury. A long-stem
prosthesis spanning the diaphysis to
the supracondylar region is gener-
ally used to maximize protection of
the entire humeral shaft.
Journal of the American Academy of Orthopaedic Surgeons
Figure 2 Different regions of the humerus
are amenable to fixation with various devic-
es. (1) Proximal metaphyseal region—head
and anatomic neck region: prosthetic arthro-
plasty with a Neer-type endoprosthesis; sur-
gical neck: either prosthetic arthroplasty with
aNeer-typeprosthesisorRushrodwithmeth-
ylmethacrylate supplementation. (2) Metadi-
aphyseal and diaphyseal region: intramedul-
lary nailing or plate fixation. (3) Distal
metadiaphyseal and metaphyseal (supra-
condylar) region: either plate fixation or
crossed flexible nails.
Metadiaphyseal and Diaphyseal
Region
Metadiaphyseal and diaphyseal
lesions can be managed with either
intramedullary nail or plate fixation.
Both methods are effective, and the
choice of implant rests with the indi-
vidual surgeon. Each technique has
specific advantages and disadvan-
tages.
Intramedullary Nail Fixation In-
tramedullary nailing, the most pop-
ular method used for humeral shaft
lesions, can be done either closed or
open through an anterograde or ret-
rograde approach. The major advan-
tage of intramedullary nail fixation is
that it can protect a long segment of
the humerus. When augmented with
methylmethacrylate, it also can pro-
vide rigid fixation of a long segment
of diseased bone. Other advantages
include a low risk of implant failure
and the fact that the nail can be placed

Figure 3 A, Anteroposterior radiograph
showing a destructive humeral diaphyseal le-
sion with ≥50% cortical bone destruction.
B, The lesion was managed with closed in-
tramedullary nailing with proximal interlock-
ing. Distal interlocking was not used because
the medial cortex was intact after nail place-
ment.
in a closed manner. The major disad-
vantage of anterograde intramedul-
lary nailing is the mandatory incision
and repair of the rotator cuff. Many
patients experience residual rotator
cuff tendinitis and weakness. Prom-
inent hardware (proximal interlock-
ing screws or the tip of the nail) can
cause persistent symptoms.
Intramedullary nail fixation can be
used for destructive bony lesions
from 2 to 3 cm below the level of the
greater tuberosity (proximal one sixth
of the humerus) to approximately 5
cm above the olecranon fossa.15 An-
terograde or retrograde nailing may
be used; care must be taken to pro-
tect areas of bone destruction.16 To
achieve rigid fixation, there must be
at least 4 to 5 cm of intramedullary
nail on either side of the lesion with
intact cortices surrounding its prox-
Vol 11, No 4, July/August 2003
Frank J. Frassica, MD, and Deborah A. Frassica, MD
imal and distal ends. An intramedul-
lary nail can be used for more prox-
imal or distal lesions if the fixation can
be made rigid with an interlocking
screw or methylmethacrylate. If the
bone destruction occurs in the supra-
condylar region of the distal humer-
us, plate fixation or crossed flexible
nails can be used.
The amount and location of the
bone destruction must be carefully as-
sessed. Closed nailing is an excellent
technique for both impending and
complete fractures. Proximal and dis-
tal locking is recommended to ensure
rigid fixation for complete fractures.
Patients with intact cortices after nail-
ing may be treated with proximal in-
terlocking alone or with proximal and
distal interlocking (Fig. 3). With the
closed technique, augmentation with
methylmethacrylate generally is not
necessary. However, for severe bone
destruction with no remaining corti- Figure 4 A, Anteroposterior radiograph
showing a pathologic fracture with >90% cor-
ces over a length of 3 to 6 cm, open tical bone destruction. B, The fracture was man-
nailing with curettage of the tumor aged with open nailing, methylmethacrylate
and methylmethacrylate can be con- supplementation, and proximal interlocking.
sidered to supplement the fixation.
Open nailing is done with a tech- where the cement augmentation is to
nique similar to closed nailing. The end. The cement is mixed, placed in
fracture site can be approached the cement gun, and injected through
through an anterolateral or posterior a small insertion tube first into the
incision or, if the entire humerus distal fragment, then the proximal
needs to be exposed, through a del- fragment. The guide wire is general-
topectoral approach proximally and ly left in the proximal and distal frag-
the anterolateral approach distally. A ments. The fracture is reduced and
portion of the deltoid insertion is el- the nail advanced into the distal frag-
evated off the humerus, and the bra- ment. Interlocking screws then can be
chialis muscle is split to expose the placed if necessary. If sufficient ce-
humeral shaft. After either exposure, ment has been injected into the dis-
a cortical window is made through tal fragment to secure the nail, an
the area of bone destruction, and all interlocking screw can be placed
of the gross tumor is removed with proximally to augment stability in the
curettes. The humerus is prepared to proximal fragment.
receive the nail, as in the closed tech- Plate Fixation Plate fixation is also
nique. The open fixation can be sup- an acceptable technique for impend-
plemented with methylmethacrylate ing and complete fractures of the
(Fig. 4). proximal metadiaphyseal and di-
After the nail is inserted over a aphyseal region (Fig. 5). A major ad-
guide wire, the entire humerus is im- vantage of plate fixation is that the ro-
aged to verify satisfactory length and tator cuff is not disturbed as it is with
fracture reduction. The nail is then anterograde intramedullary nailing.
withdrawn into the proximal frag- Disadvantages of plate fixation in-
ment to a level just above the point clude more blood loss than with
285
Metastatic Bone Disease of the Humerus

can be applied before or after the in-
ternal fixation, but it is easier to re-
duce the fracture and place the inter-
nal fixation device first. To obtain
good apposition of bone ends, irreg-
ular fracture ends can be shortened
if necessary. Once the plate is applied,
the screws that span the defect are re-
moved, and cement is placed by hand
into the defect. The screws can be re-
placed while the cement is curing or
by drilling and tapping once the ce-
ment hardens.
weeks). Treatment is generally de-
layed to 10 days after surgery so that
the skin incision can heal.
Patients may begin range of mo-
tion movement of the elbow and
shoulder during the first postopera-
tive week. If the patient has not pre-
viously undergone radiation therapy,
the sutures or staples are removed 2
weeks after surgery. If the patient has
previously received radiation, the su-
tures are left in place for approximate-
ly 4 weeks.

Distal Metadiaphyseal and Alternative Surgical Techniques

Figure 5 A, Anteroposterior radiograph of
a patient with prostate cancer showing a met-
astatic diaphyseal lesion with >90% cortical
bone destruction. B, The lesion was managed
with plate fixation and methylmethacrylate
augmentation.
Metaphyseal Region Rush rods can be used for very
Lesions within 2 to 4 cm of the proximal (within 3 cm of the humer-
olecranon fossa are best managed with al head) fractures at the surgical
plate fixation. Neither anterograde nor neck.18 There is too little proximal
retrograde nails can provide rigid fix- bone in this region for intramedullary
ation in these distal lesions. Plating nail fixation. The proximal hook of the
through a posterior triceps muscle– Rush rod is anchored in the rotator
splitting approach is very effective cuff, and the fixation is supplement-
when the distal fragment is large
enough to receive three 4.5-mm
screws. Distal lesions that involve the
supracondylar area are difficult to
manage. Fixation can be achieved with
medial and lateral plates, flexible nails
inserted from the epicondyles, or pros-
thetic arthroplasty.

closed nailing, the potential for radi- Postoperative Management

al nerve injury, and inability to pro-
tect as much humeral length as with
intramedullary nailing.16 When con-
sidering plate fixation, the radio-
graphs must be studied carefully to
determine that at least three screws
can be placed in normal cortical bone
on either side of the fracture. When
there is diffuse involvement of the hu-
merus, intramedullary nailing is a
better choice because a plate may not
provide solid fixation.
The exposure for plate fixation can
be done through an anterolateral or
posterior approach. A cortical win-
dow is created large enough to curette
the gross tumor. Care must be taken
not to remove an excessive amount
of normal bone; otherwise, it will be
difficult to achieve rigid fixation in the
remaining cortical bone. The cement
After prophylactic fixation or sur-
gical treatment of a pathologic frac-
ture, radiation therapy to the site of
the lesion and the implanted device
is recommended to decrease the risk
of continued bone destruction (which
could lead to increased pain), loosen-
ing of the fixation, and the need for
additional surgery.17 In a study of 64
procedures in 60 patients, Townsend
et al17 found that the addition of ex-
ternal beam irradiation to surgery sig-
nificantly (P = 0.02) improved func-
tional outcome. There was also a
significantly (P = 0.035) higher risk of Figure 6 A humeral lesion in a patient with
the need for a second surgical proce- metastatic renal cell cancer was treated with
dure in patients who did not receive intramedullary nailing and external beam ir-
radiation. A, Anteroposterior radiograph show-
postoperative radiation. The dose of ing disease progression and fracture (arrow)
radiation is similar to that used when around the nail just below the locking screw.
treating patients nonsurgically (3,000 B, The patient was treated with resection and
custom proximal humeral arthroplasty.
cGy in 10 fractions over 1 week to 2

286 Journal of the American Academy of Orthopaedic Surgeons


ed with methylmethacrylate. If the
hook protrudes, it can cause symp-
toms with overhead activity; if it is
driven through the proximal cortical
bone, there will be less purchase in
the proximal fragment. Flexible nails
also have been used for diaphyseal
and proximal metadiaphyseal le-
sions.19
Custom proximal humeral prosthe-
ses can be used selectively when large
segments of bone need to be resected.20
Resection is usually reserved for pa-
tients with progressive disease after
external beam irradiation or in patients
with failed internal fixation and poor
bone stock. Some custom modular de-
vices allow the restoration of length
of the humerus with immediate rigid
fixation after resection of variable
lengths of the proximal humerus (Fig.
6). Other devices have been designed
to permit resection and reconstruction
of the diaphysis in the presence of an
intact proximal humeral segment.21
These designs, with medullary stems
cemented into the proximal and dis-
tal intramedullary canals, are still in
development and are associated with
complication rates as high as 25%.22
References
1. Frassica FJ, Gitelis S, Sim FH: Metastat-
ic bone disease: General principles,
pathophysiology, evaluation, and biop-
sy. Instr Course Lect 1992;41:293-300.
2. Fidler M: Prophylactic internal fixation
of secondary neoplastic deposits in
long bones. BMJ 1973;1:341-343.
3. Fidler M: Incidence of fracture through
metastases in long bones. Acta Orthop
Scand 1981;52:623-627.
4. Mirels H: Metastatic disease in long
bones: A proposed scoring system for
diagnosing impending pathologic frac-
tures. Clin Orthop 1989;249:256-264.
5. Beals RK, Lawton GD, Snell WE: Pro-
phylactic internal fixation of the femur
in metastatic breast cancer. Cancer 1971;
28:1350-1354.
6. Hipp JA, Springfield DS, Hayes WC:
Predicting pathologic fracture risk in
the management of metastatic bone de-
fects. Clin Orthop 1995;312:120-135.
7. Arcangeli G, Giovinazzo G, Saracino B,
Vol 11, No 4, July/August 2003
Frank J. Frassica, MD, and Deborah A. Frassica, MD
Results
Pain relief can be reliably obtained in
>90% of patients treated with rigid
fixation.21-23 Redmond et al15 de-
scribed good to excellent pain relief
in 12 of 13 patients treated with in-
tramedullary nail fixation for humer-
al pathologic fractures. In 10 patients
with documented postoperative
range of motion, there was a mean of
101° of abduction (range, 55° to 180°)
and a mean of 98° of forward flexion
(range, 45° to 170°). Of the 13 patients,
11 recovered use of the arm for activ-
ities of daily living.15 Dijkstra et al16
retrospectively compared nail and
plate fixation in 37 patients with 38
pathologic humeral fractures. There
was good to excellent subjective pain
relief in approximately 90% of the pa-
tients treated with either method. The
authors also reported early fixation
failure secondary to angular deformi-
ty and rotational instability in patients
treated with intramedullary nails
without proximal and distal lock-
ing.16 Most treatment failures are sec-
ondary to disease progression or ear-
ly loss of fixation.17 Patients with renal
et al: Radiation therapy in the manage-
ment of symptomatic bone metastases:
The effect of total dose and histology on
pain relief and response duration. Int J
Radiat Oncol Biol Phys 1998;42:1119-
1126.
8. Tong D, Gillick L, Hendrickson FR: The
palliation of symptomatic osseous me-
tastases: Final results of the Study by
the Radiation Therapy Oncology
Group. Cancer 1982;50:893-899.
9. Cole DJ: A randomized trial of a single
treatment versus conventional fraction-
ation in the palliative radiotherapy of
painful bone metastases. Clin Oncol
1989;2:59-62.
10. Hoskin PJ, Price P, Easton D, et al: A
prospective randomised trial of 4 Gy or
8 Gy single doses in the treatment of
metastatic bone pain. Radiother Oncol
1992;23:74-78.
11. Nielsen OS, Bentzen SM, Sandberg E,
Gadeberg CC, Timothy AR: Random-
cell carcinoma are especially prone to
disease progression and often are
treated with higher initial doses of ra-
diation (4,500 cGy) to reduce the risk
of early failure.
Summary
Metastatic disease of the humerus is
common, and effective management
can improve the quality of life for can-
cer patients. Nonsurgical treatment
with external beam irradiation is used
for symptomatic lesions with <50%
cortical bone destruction. With ≥50%
cortical bone destruction, intramed-
ullary nailing is the most common
method of both prophylactic and frac-
ture fixation. Locked intramedullary
nails can provide rigid fixation with
early pain relief. Methylmethacrylate
can be used to aid in the reconstruc-
tion of defects caused by the surgi-
cal treatment of large lesions and to
improve fixation. Plate fixation also
can be used for diaphyseal and dis-
tal lesions. Immediate rigid fixation
is necessary to achieve consistently
good pain relief.
ized trial of single dose versus fraction-
ated palliative radiotherapy of bone
metastases. Radiother Oncol 1998;47:233-
240.
12. Price P, Hoskin PJ, Easton D, Austin D,
Palmer SG, Yarnold JR: Prospective
randomised trial of single and multi-
fraction radiotherapy schedules in the
treatment of painful bony metastases.
Radiother Oncol 1986;6:247-255.
13. Flemming JE, Beals RK: Pathologic frac-
ture of the humerus. Clin Orthop 1986;
203:258-260.
14. Douglass HO Jr, Shukla SK, Mindell E:
Treatment of pathological fractures of
long bones excluding those due to
breast cancer. J Bone Joint Surg Am 1976;
58:1055-1061.
15. Redmond BJ, Biermann JS, Blasier RB:
Interlocking intramedullary nailing of
pathological fractures of the shaft of the
humerus. J Bone Joint Surg Am 1996;78:
891-896.
287
Metastatic Bone Disease of the Humerus
16. Dijkstra S, Stapert J, Boxma H, Wiggers
T: Treatment of pathological fractures
of the humeral shaft due to bone me-
tastases: A comparison of intramedul-
lary locking nail and plate osteosynthe-
sis with adjunctive bone cement. Eur J
Surg Oncol 1996;22:621-626.
17. Townsend PW, Rosenthal HG, Smalley
SR, Cozad SC, Hassanein RE: Impact of
postoperative radiation therapy and
other perioperative factors on outcome
after orthopedic stabilization of im-
pending or pathologic fractures due to
metastatic disease. J Clin Oncol 1994;12:
2345-2350.
288
18. Lewallen RP, Pritchard DJ, Sim FH:
Treatment of pathologic fractures or
impending fractures of the humerus
with Rush rods and methylmethacry-
late: Experience with 55 cases in 54 pa-
tients, 1968-1977. Clin Orthop 1982;166:
193-198.
19. Lin J, Hou SM, Hang YS, Chao EY:
Treatment of humeral shaft fractures by
retrograde locked nailing. Clin Orthop
1997;342:147-155.
20. Sim FH, Frassica FJ, Chao EY: Ortho-
paedic management using new devices
and prostheses. Clin Orthop 1995;312:
160-172.
Journal of the American Academy of Orthopaedic Surgeons
21. Chin HC, Frassica FJ, Hein TJ, et al:
Metastatic diaphyseal fractures of the
shaft of the humerus: The structural
strength evaluation of a new method of
treatment with a segmental defect pros-
thesis. Clin Orthop 1989;248:231-239.
22. Damron TA, Sim FH, Shives TC, An
KN, Rock MG, Pritchard DJ: Intercalary
spacers in the treatment of segmentally
destructive diaphyseal humeral lesions
in disseminated malignancies. Clin
Orthop 1996;324:233-243.
23. Sim FH, Pritchard DJ: Metastatic dis-
ease in the upper extremity. Clin Orthop
1982;169:83-94.
 Web and Wireless Review
MD Consult
The effective use of the Web by physicians should fulfill
a few criteria. First, the time spent should be proportion-
al to the value obtained from the information sought. The
material obtained should be able to add measurably to
the care of patients, office efficiency, fund of information,
or bottom line of the practice. Ideally, the time spent find-
ing information should be an investment if the use of a
site is ongoing.
One such type of reference Website is MD Consult (www.
mdconsult.com), which provides reference textbooks, a
MEDLINE journal search, drug information, a review of
medical news of the week, clinical topic tours, practice
guidelines, patient handouts, online CME programs,
downloads for PDAs, and a drug information module.
Some specialties, such as infectious disease, have a sep-
arate collection of materials, called editions, but current-
ly there is not one for orthopaedic surgery. The edition
relevant to orthopaedic surgeons is the Core Collection,
which can be accessed easily for a 10-day free trial by
submitting your name and e-mail address.
Although the offerings for primary care physicians and
internists are somewhat broader, only two of the forty
reference texts on this site are directly relevant to the or-
thopaedic surgeon: the tenth edition (2003) of Campbell’s
Operative Orthopaedics and the second edition (2003) of
Delee and Drez’s Orthopaedic Sports Medicine. It does give
access to all of the text and most of the figures for view-
ing and printing. The MEDLINE search is similar to that
found in many other sites. For fifty journals, full text is
available free through this site, but only one, the Amer-
ican Journal of Sports Medicine, is an orthopaedic title. MD
Consult has a catalogue of 3,500 rather straightforward
patient information brochures that can be downloaded
and customized. Twenty-five relate to musculoskeletal
conditions; a few have diagrammatic illustrations, and
several are also available in Spanish. There are two hun-
dred grand rounds topics available online for CME cred-
it in fifteen specialties; however, none is related to ortho-
paedic surgery. Similarly, there are approximately a
thousand clinical practice guidelines but, again, none re-
lating to orthopaedics.
What are the practical applications for the orthopaedic
surgeon? There is immediate access from any computer
with Internet access to MEDLINE for a quick topic search
or to the text of Campbell’s Operative Orthopaedics to rap-
idly review surgical techniques. With problems such as
antithrombolic prophylaxis or septic arthritis, immedi-
ate access to texts such as Conn’s Current Therapy or Man-
Vol 11, No 4, July/August 2003
dell, Douglas, and Bennett’s Principles and Practice of Infec-
tious Diseases allows a broader perspective than that in
texts more commonly available to us. Information about
various drugs, the spectrum of antimicrobial agents, and
recommendations for appropriate antibiotic coverage are
all rapidly accessible through this site.
Following are a few practical hints to maximize your
use of MD Consult (or a site such as this). While viewing
the reference text you have searched for, copy the ma-
terial on the screen and paste it into a document in a stan-
dard program such as Microsoft Word (MS Word). Call
the file Personal Medical Textbook (as an example). By
cutting and pasting selected sections of the material, you
can choose the groups of text you want so that the past-
ed matter represents your own “review” and personal
interests. After you have it in the file, type a heading at
the top of that section, for example, Biceps Ruptures. High-
light Biceps Ruptures. Then, in the Styles section of Word,
assign that phrase the style Heading 1. When you do this
with all of your downloads from the Web, you can, with
only a few keystrokes, have Word make a hyperlinked
personal Table of Contents for rapid access at a later date.
Over time, you will have your own ongoing review text
from multiple reference sources as well as the published
literature.
You can view your Personal Medical Textbook on your
PDA (Pocket PC or Palm). Any MS Word file can be con-
verted into a format called Microsoft Reader. Microsoft
Reader conversion programs, and the Microsoft Reader
program for your PDA, are distributed free by Microsoft.
When you load the conversion program onto your com-
puter, there will be a single Microsoft Reader icon on your
MS Word toolbar that will convert the file to PDA for-
mat so that you can transfer your Personal Medical Text-
book to your PDA.
Microsoft Reader can be found at www.microsoft.com/
reader/downloads/default.asp. The program to allow
you to convert any MS Word file to PDA Microsoft Read-
er format can be found at www.microsoft.com/reader/
downloads/rmr.asp. Detailed step-by-step instructions
and models on all steps reviewed can be found at
www.officemedical.com.
The cost of MD Consult for residents is $119.95 a year
or $14.95 a month. For attendings, it is $219.00 a year or
$24.95 a month. Considering the cost of just one of the
textbooks available, this is quite a bargain and a tremen-
dous educational and reference resource.
Ira H. Kirschenbaum, MD
289