NORMAL

Ulnar>0.4
Median>0.31 ABSTRACT: Criteria for the diagnosis of chronic inflammatory demyelin-
ating polyneuropathy (CIDP) are met by the polyneuropathy associated with
immunoglobulin M (IgM) paraproteinemia and anti-myelinassociated gly-
coprotein (MAG) antibody (MAG-CIDP). However, MAG-CIDP differs from
other types of CIDP, mainly in its poorer response to treatment. The utility of
terminal latency index (TLI) as an electrophysiological marker for MAG-
TLI: distal distance/ CIDP has been debated. In this study we confirmed its diagnostic usefulness
and evaluated TLI threshold values for motor nerves investigated in routine

DML/Proximal CV nerve conduction studies. Median, ulnar, peroneal, and tibial TLIs of 11
subjects with MAG-CIDP, 18 with CIDP, and 76 healthy controls were com-
pared, and threshold values for MAG-CIDP evaluated as the lowest value
70mm/4.1ms/ with a likelihood ratio higher than 10. Mean TLI values and TLIs of all but the
peroneal nerve were significantly lower in MAG-CIDP. Median nerve TLI of
0.26 and ulnar nerve TLI of 0.33 were identified as the threshold TLI values
44ms=.388 for MAG-CIDP.
2001 John Wiley & Sons, Inc. Muscle Nerve 24: 12781282, 2001

TERMINAL LATENCY INDEX IN POLYNEUROPATHY

WITH IgM PARAPROTEINEMIA AND
ANTI-MAG ANTIBODY
DARIO COCITO, MD,1 GIANLUCA ISOARDO, MD,2
PALMA CIARAMITARO, MD,2 GIUSEPPE MIGLIARETTI, PhD,3
ANTONIO PIPIERI, MD,2 PIERANGELO BARBERO, MD,2
ANGELE CUCCI, PhD,2 and LUCA DURELLI, MD2

1
Unita` Operativa di Neurofisiologia Clinica, Dipartimento di Neuroscienze,
Universita` di Torino, Via Cherasco 15, 10126 Torino, Italy
2
Clinica Neurologica I, Dipartimento di Neuroscienze, Universita` di Torino, Italy
3
Unita` Operativa di Statistica, Dipartimento di Sanita` Pubblica e Microbiologia,
Universita` di Torino, Torino, Italy

Accepted 11 May 2001

Myelin-associated glycoprotein (MAG) antibody has
been found in patients with immunoglobulin M
(IgM) paraproteinemia and chronic polyneuropathy
fulfilling criteria for chronic inflammatory demyelin-
ating polyneuropathy (CIDP).1,9,14,16 However, the
CIDPs associated with IgM paraproteinemia and
anti-MAG antibody (MAG-CIDP) have a different
pathogenesis, clinical and electrophysiological fea-
tures, response to treatment, and prognosis than
Abbreviations: CIDP, chronic inflammatory demyelinating polyneuropa-
thy; CMAP, compound muscle action potential; DADS-I, distal acquired
demyelinating symmetric polyneuropathy without paraproteinemia;
DADS-M, distal acquired demyelinating symmetric polyneuropathy with
paraproteinemia; IgM, immunoglobulin M; LR, likelihood ratio; MAG, my-
elin-associated glycoprotein; MGUS, monoclonal gammopathy of unde-
termined significance; NCS, nerve conduction study; TLI, terminal latency
index
Key words: anti-MAG antibody; chronic inflammatory demyelinating
polyneuropathy; paraproteinemia; terminal latency index
Correspondence to: D. Cocito; e-mail: dariococito@yahoo.it
2001 John Wiley & Sons, Inc.
other forms of CIDP, regardless of whether there is
an association with paraproteinemia.9,12,13 Differen-
tial diagnosis between MAG-CIDP and other forms
of CIDP is sometimes difficult on clinical and elec-
trophysiological grounds, particularly when large-
fiber sensory loss is the prominent clinical fea-
ture,7,14,18 but it is very important because sensory
CIDP without anti-MAG antibody responds better to
immunomodulating treatment.18
A disproportionate increase in distal motor la-
tency was identified as a typical feature of MAG-CIDP
on nerve conduction studies (NCS), and Kaku et al.
identified a low value of terminal latency index
(TLI) as a useful electrophysiological marker for
MAG-CIDP.6 Some later studies confirmed these re-
sults,9,17 but others did not.4,7 The aims of our study
were to determine whether TLI is a suitable electro-
physiological marker to differentiate MAG-CIDP
from other forms of CIDP, and to find the TLI value

1278 TLI in MAG-CIDP MUSCLE & NERVE October 2001

for each motor nerve that best differentiates MAG-
CIDP from other forms of CIDP.
PATIENTS AND METHODS
Patient Selection. We included 29 consecutive pa-
tients with CIDP (mean age 60.8 15.2 years, 21 men
and 8 women) who were evaluated in our depart-
ment between December 1, 1998 and July 1, 2000.
All patients satisfied the electrophysiological criteria
described by Rotta et al.14 Exclusion criteria were
history or laboratory findings suggestive of heredi-
tary sensorimotor neuropathy types 1A or 1B, heredi-
tary neuropathy with liability to pressure palsy, hypo-
thyroidism, vasculitis, amyloidosis, leukodystrophy,
Refsums disease, retinitis pigmentosa, or optic atro-
phy. Serum protein electrophoresis or immunofixa-
tion was obtained from each patient at the time of
first evaluation. Patients with detectable paraprotein-
emia underwent urine electrophoresis to detect
Bence-Jones proteinuria as well as a skeletal bone
survey; they were then referred to a hematologist for
bone marrow examination. Differential diagnosis be-
tween IgM monoclonal gammopathy of undeter-
mined significance (MGUS) and Waldenstroms dis-
ease was made according to Dimopoulos et al.3
All patients were tested for anti-MAG antibodies
by an enzyme-linked immunosorbent assay (Buhl-
mann Laboratories AG, Allshwil, Switzerland).
Patients with both IgM paraproteinemia and anti-
MAG antibodies were diagnosed as having MAG-
CIDP. The remaining patients were diagnosed as
having another form of CIDP. Sensory and motor
disturbances were quantified according to Van Dijk
et al.18 Sensorimotor polyneuropathy was defined by
the presence of weakness in at least one tested
muscle in patients with a bilateral and symmetric
sensory deficit with a peripheral nerve distribution.
We used 76 healthy subjects (55.5 12.8 years, 46
men and 30 women) as controls. All patients and
volunteers gave their informed consent, and the
study was approved by the local ethics committee.
Electrophysiological Assessment. Bilateral motor
NCS of median, ulnar, peroneal, and tibial nerves
were performed using standard techniques.8 All
studies were performed in a warm room. Skin tem-
perature was maintained above 35C in both patients
and controls. Distal motor latency was measured at
the onset of compound muscle action potential
(CMAP). Area of CMAP was measured with the ap-
propriate software. TLI was calculated as distal con-
duction distance (mm)/(conduction velocity [m/s]
distal motor latency [ms]).2,4 Distal conduction
TLI in MAG-CIDP
distance (distance between recording and stimulat-
ing electrode at the most distal site) was 70 mm for
the median, ulnar, and peroneal nerves, and 100
mm for the tibial nerve. We also evaluated median,
ulnar, peroneal, and tibial TLIs in 76 healthy con-
trols, and the mean value of all these nerves com-
bined.
Statistical Analysis. Categorical data were com-
pared using Fishers exact test. Means were com-
pared using Students t-test. Likelihood ratio (LR)15
was calculated for each TLI value of all nerves under
study according to the formula: LR = sensitivity/(1
specificity).
Sensitivity identifies the proportion of patients
with disease in whom the test is positive, and speci-
ficity identifies the proportion of patients without
disease in whom the test is negative. The LR indi-
cates how much the result of a given diagnostic test
will increase or decrease the probability of a given
disorder. An LR of 1 means that the posttest prob-
ability is exactly the same as the pretest probability.
LRs greater than 1 increase the probability that the
target disorder is present; the higher the LR, the
greater the increase.5 The increase in pretest prob-
ability of having the disorder can be calculated by
converting pretest probability to odds, multiplying
the calculated odds by LR, and then converting the
obtained posttest odds into a probability again.5
Change of pretest probability can also be evaluated
in a simpler way with the normogram illustrated in
Figure 1.15 We defined the threshold TLI for MAG-
CIDP as the highest TLI values with an LR higher
than 10. Such a high LR yields a very large increase
in pretest probability.5 We then evaluated the in-
crease in probability of having MAG-CIDP for the
TLI threshold value of each nerve. Sensitivity and
specificity of TLI for each nerve were calculated for
MAG-CIDP.
RESULTS
Among our 39 patients, 13 (45%) had paraprotein-
emia: 11 had MGUS (10 IgM, 1 IgG); and 2 had
Waldenstroms disease. Eleven of 12 patients with
IgM paraproteinemia had IgM anti-MAG antibodies
and were diagnosed as MAG-CIDP cases. The re-
maining 18 patients were diagnosed with another
form of CIDP. No patient without IgM paraprotein-
emia was positive for anti-MAG antibody. Demo-
graphic and clinical features, presence of increased
cerebrospinal fluid protein level, and response to
immunomodulating treatment are summarized in
Table 1. The MAG-CIDP patients were significantly
MUSCLE & NERVE October 2001 1279

FIGURE 1. Normogram for evaluating increase in probability of a
target disease after the finding of a test results with known like-
lihood ratio (LR). From left to right, the first column represents the
pre-test probability, the second column represents the LR and the
third column represents the post-test probability. Post-test prob-
ability can be evaluated by anchoring a ruler at the pre-test prob-
ability and rotating it until lines up with the observed LR of test
results. The straight line refers to an example: the finding of an
ulnar TLI lower than 0.33 (LR 10) in patient with low probability of
having MAG-CIDP (30%) will increase to 80% the probability of
such diagnosis.
in 100 (91%) nerves of those with other forms of
CIDP. Peroneal (P = 0.02) and tibial (P = 0.04)
nerves were more frequently inexcitable in the MAG-
CIDP group than in other CIDP group.
Mean TLI values of each nerve under study and
of all nerves are summarized in Table 2. Mean TLI,
and median, ulnar, and tibial TLI of MAG-CIDP pa-
tients were significantly lower than those of healthy
controls or the other CIDP group (Table 2). We
evaluated LR only for TLIs of median and ulnar
nerves because of the high proportion of inexcitable
peroneal and tibial nerves in the MAG-CIDP group.
The threshold value of the median TLI was 0.26 (LR
= 16) and of the ulnar TLI was 0.33 (LR = 10). The
LR of combined median nerve TLI lower than 0.26
and ulnar nerve TLI lower than 0.33 was 10. Sensi-
tivity and specificity of TLIs at these threshold levels
are summarized in Table 3. Patients with MAG-CIDP
had a significantly greater number of median TLIs
lower than 0.26 (82%, P < 0.0001) and ulnar TLIs
lower than 0.33 (44%, P = 0.002) than did the group
with other forms of CIDP.
Table 4 summarizes the increase in probability of
having MAG-CIDP after finding at least one motor
nerve with a TLI value below the identified thresh-
old.
DISCUSSION

This study showed that TLI was significantly lower in

older (P = 0.003) than those with other forms of
CIDP.
We examined 480 nerves: 119 median; 119 ulnar;
121 peroneal; and 121 tibial. Of these, 66 motor
nerves were evaluated in patients with MAG-CIDP,
110 in patients with other forms of CIDP, and 304 in
healthy controls. CMAPs could be elicited in 48
(73%) of the nerves in patients with MAG-CIDP and
patients with MAG-CIDP than in patients with other
CIDPs, thus reflecting disproportionate distal con-
duction slowing. We identified a median nerve TLI <
0.26 or an ulnar nerve TLI < 0.33 as threshold values
for MAG-CIDP. Finding at least one ulnar nerve in
combination or not with one median nerve TLI be-
low the threshold in a patient with CIDP greatly in-
creases the probability of having MAG-CIDP.

Table 1. Demographic, clinical, and immunological features of MAG-CIDP and the different subgroups of CIDP.

Disease duration Frequency of increased

Number of Age (months) CSF protein level Response to
Type of CIDP patients (mean SD) (mean SD) (normal <0.50 mg/dl) treatment*

MAG 11 70 8.5 60.2 48.9 8/10

(2/11)
Sensory 4 51.2 146 18.7 14.3 1/4 ++
(3/4)
Distal motor 4 46.5 23.3 12.0 8.8 4/4 ++
(3/4)
Diffuse 10 60.4 12.4 55.7 61.0 9/9 +++
weakness (8/10)

CIDP, chronic inflammatory demyelinating polyneuropathy; CSF, cerebrospinal fluid; MAG, myelin-associated glycoprotein.
*Response to immunomodulating treatment: , poor; +, mild; ++, moderate; +++, high.

Sensory polyneuropathy in 10, sensorimotor polyneuropathy with diffuse weakness in 1.

Frequency of patients with at least a mild response to immunomodulating treatment.

1280 TLI in MAG-CIDP MUSCLE & NERVE October 2001

 Table 2. Comparison of mean terminal latency index ( SD).

MAG-CIDP vs.
MAG-CIDP Other CIDP MAG-CIDP vs. Healthy controls healthy controls
(n = 11) (n = 18) other CIDP (P-value*) (n = 76) (P-value*)

Median nerve 0.22 0.04 0.37 0.08 <0.0001 0.40 0.04 <0.0001
[0.200.24] [0.340.40] [0.390.41]
(17/17) (26/26) (76/76)
Ulnar nerve 0.36 0.06 0.44 0.09 0.0017 0.48 0.06 <0.0001
[0.330.39] [0.410.47] [0.470.49]
(16/16) (27/27) (76/76)
Peroneal nerve 0.27 0.12 0.36 0.06 0.09 0.35 0.05 0.0
[0.190.35] [0.340.38] [0.340.36]
(8/16) (24/29) (76/76)
Tibial nerve 0.29 0.09 0.44 0.10 0.0009 0.47 0.06 <0.0001
[0.230.35] [0.400.48] [0.460.48]
(8/17) (23/28) (76/76)
All nerves 0.29 0.09 0.41 0.09 <0.0001 0.42 0.06 <0.0001
combined [0.270.31] [0.390.43] [0.410.43]
(49/66) (100/110) (304/304)

CIDP, chronic inflammatory demyelinating polyneuropathy; MAG, myelin-associated glycoprotein; MAG-CIDP, chronic demyelinating polyneuropathy
associated with IgM paraproteinemia and anti-MAG antibody. Ranges in brackets are the 95% confidence intervals. The number of nerves with
excitable compound muscle action potential/number of evaluated nerves is shown in parentheses.
*Comparisons are by Students t-test.

Previous studies have described MAG-CIDP as a a higher proportion of patients with serum parapro-
distinct clinical entity,9,12,13 characterized by a slowly teinemia (45% versus 20%), which was probably due
progressive sensory deficit with mild distal weakness to the higher mean age and proportion of patients
at the later stage of disease and a poor response to submitted to serum protein electrophoresis or im-
immunomodulating treatment.9,11,13 Kaku et al.6 munofixation in our cohort. Our results were similar
identified a TLI < 0.25 as a highly specific electro- to those of previous studies that found a significantly
physiological feature of MAG-CIDP. More recently, lower TLI in MAG-CIDP than in other forms of
Trojaborg et al.17 and Maisonobe et al.9 found a CIDP.9,17
significantly lower TLI in MAG-CIDP than in other Discrepancies between our results and those of
CIDPs. However, others4,7 did not confirm these re- Katz et al.7 could be due to either differences in the
sults. Low values of TLI were also found in heredi- number of patients who underwent anti-MAG anti-
tary neuropathy with liability to pressure palsy.2 Dif-
ferences between studies could be due to differences Table 4. Probability of having MAG-CIDP related to terminal
in patient selection and classification. First, we in- latency index value.
cluded patients according to electrophysiological cri-
Pretest Posttest
teria less stringent than those of the Ad Hoc Sub- probability probability
committee of the American Academy of Neurology.1 Nerve TLI value (%) (%)
The AAN criteria were, in fact, mainly designed for
Median <0.26 20 80
the trials, whereas we aimed at defining an electro- 30 87.2
physiological parameter that could be employed in 40 91.4
the diagnosis of all patients with CIDP. Our patient 50 94.1
cohort was similar to that of Rotta et al.,14 except for 60 96
Ulnar <0.33 20 71.4
30 81.1
Table 3. Sensitivity and specificity of median and ulnar 40 86.9
terminal latency index (TLI) for the diagnosis of MAG-CIDP. 50 90.9
Sensitivity Specificity 60 93.7
Nerve TLI value (%) (%) Both nerves As above 20 71.4
30 81.1
Median <0.26 90.9 94.4 40 86.9
Ulnar <0.33 54.5 94.4 50 90.9
Both nerves As above 54.5 94.4 60 93.7

MAG-CIDP, chronic demyelinating polyneuropathy associated with IgM MAG-CIDP, chronic demyelinating polyneuropathy associated with IgM
paraproteinemia and anti-MAG antibody. paraproteinemia and anti-MAG antibody.

TLI in MAG-CIDP MUSCLE & NERVE October 2001 1281

body studies or in patient classification. With regard
to patient classification, Katz et al. did not find any
significant difference in TLI between distal acquired
demyelinating symmetric neuropathy and CIDP,
whether associated with paraproteinemia (DADS-M)
or not (DADS-I). However, these investigators, un-
like us, included patients with IgG and IgM parapro-
teinemia in the DADS-M group and did not include
patients with predominantly distal weakness in the
CIDP group.
To our knowledge, no previous study, with the
exception of Kaku et al.,6 has reported threshold
values of TLI for MAG-CIDP. Although median
nerve TLI < 0.26 had the highest sensitivity, specific-
ity, and LR for MAG-CIDP, low values of TLI were
also detected in patients with carpal tunnel syn-
drome. Because the frequency of carpal tunnel syn-
drome in patients with MAG-CIDP is unknown, we
suggest that only an ulnar nerve TLI < 0.33, whether
or not in combination with a median nerve TLI <
0.26, should be considered in clinical practice as
highly suggestive of MAG-CIDP. Our results demon-
strated a very large increase in the probability of
having MAG-CIDP after finding ulnar and median
TLIs below the identified thresholds. However, the
pretest probabilities shown in Table 4 are only hy-
pothetical, and do not refer to our series. They rep-
resent MAG-CIDP probabilities that a physician may
calculate from the combination of a patients history
and physical examination, and weighing, according
to personal clinical experience, all available informa-
tion. Multivariate analyses combining clinical/
immunological, and electromyographic parameters
to generate a very precise probability of MAG-CIDP
diagnosis are lacking, and pretest TLI probabilities
of MAG-CIDP can only be evaluated very approxi-
mately, relying on each neurologists clinical experi-
ence and intuition. Further studies will address this
issue. The differential diagnosis between MAG-CIDP
and sensory CIDP, or distal motor CIDP, is often
difficult on clinical and electrophysiological
grounds, but it is mandatory because treatment strat-
egies are different. MAG-CIDP is poorly responsive
to immunomodulating treatment,11 unlike sensory
and distal motor CIDP,14,18 which may suggest the
possibility of Waldenstroms disease.3,10
In our opinion, TLI is an easy, inexpensive,
highly informative test that is helpful in clarifying
this differential diagnosis and is therefore extremely
useful in the diagnostic work-up of patients with
CIDP.
Presented in part at the 10th European Congress of Clinical Neu-
rophysiology, Lyon, France, August 2000.
1282 TLI in MAG-CIDP
REFERENCES
1. Ad Hoc Subcommittee of the American Academy of Neurol-
ogy AIDS Task Force. Research criteria for diagnosis of
chronic inflammatory demyelinating polyneuropathy (CIDP).
Neurology 1991;41:617618.
2. Amato AA, Gronseth GS, Callerame KJ, Kagan-Hallet KS,
Bryan WW, Barohn RJ. Tomaculous neuropathy: a clinical
and electrophysiological study in patients with and without
1.5-Mb deletions in chromosome 17p11.2. Muscle Nerve
1996;19:1622.
3. Dimopulos MA, Panayiotidis P, Moulopulos M, Stikakis P, Dal-
akas M. Waldenstrom macroglobulinemia: clinical features,
complications and management. J Clin Oncol 2000;18:
214226.
4. Ellie E, Vital A, Steck A, Boiron JM, Vital C, Julien J. Neurop-
athy associated with benign anti-myelin-associated glycopro-
tein IgM gammopathy: clinical, immunological, neurophysi-
ological, pathological findings and response to treatment in
33 cases. J Neurol 1996;243:3443.
5. Jaeschke R, Guyatt GH, Sackett DL, for the Evidence-Based
Medicine Working Group. Users guide to the medical litera-
ture. III. How to use an article about a diagnostic test B. What
are the results and will they help me in caring for my patients?
The Evidence-Based Medicine Working Group. JAMA 1994;
271:703707.
6. Kaku DA, England JD, Summer AJ. Distal accentuation of
conduction slowing in polyneuropathy associated with anti-
bodies to myelin-associated glycoprotein and sulphated gluc-
uronyl paragloboside. Brain 1994;117:941947
7. Katz JS, Saperstein DS, Gronseth G, Amato AA, Barohn RJ.
Distal acquired demyelinating symmetric neuropathy. Neurol-
ogy 2000;54:615619.
8. Kimura J. Electrodiagnosis in disease of the nerve and muscle:
principles and practice. Philadelphia: FA Davis; 1989. p 103138.
9. Maisonobe T, Chassande B, Verin M, Jouni M, Leger JM,
Bouche P. Chronic dysimmune demyelinating polyneuropa-
thy: a clinical and electrophysiological study of 93 patients. J
Neurol Neurosurg Psychiatry 1996;61:3642.
10. Nobile-Orazio E, Francomano E, Daverio R, Barbieri S,
Marmiroli P, Manfredini E, Carpo M, Moggio M, Legname M,
Baldini L, Scarlato G.Anti-myelin associated glycoprotein IgM
antibody titers in neuropathy associated with macroglobulin-
emia. Ann Neurol 1989;26:543550.
11. Nobile-Orazio E, Meucci N, Baldini L, Di Troia A, Scarlato G.
Long term prognosis of neuropathy associated with anti-MAG
IgM M-proteins and its relationship to immune therapies.
Brain 2000;123:710717.
12. Notermans NC, Franssen H, Eurelings M, Van der Graaf Y,
Wokke JH. Diagnostic criteria for demyelinating polyneurop-
athy associated with monoclonal gammopathy. Muscle Nerve
2000;23:7379.
13. Ropper AH, Gorson KC. Neuropathies associated with para-
proteinemia. N Engl J Med 1998; 338:16011607.
14. Rotta FT, Sussman AT, Bradley WG, Ayyar DR, Sharma KR,
Shebert RT. The spectrum of chronic inflammatory demye-
linating polyneuropathy. J Neurol Sci 2000;173:129139.
15. Sackett DL. A prime on the precision and accuracy of the
clinical examination. JAMA 1992;267:26382644.
16. Tagawa Y, Yuki N, Hirata K. Anti-SGPG antibody in CIDP:
nosological position of IgM anti-MAG/SGPG antibody-
associated neuropathy. Muscle Nerve 2000;23:895899.
17. Trojaborg W, Hays AP, Van den Berg L, Younger DS, Latov N.
Motor conduction parameters in neuropathies associated
with anti-MAG antibodies and other types of demyelinating
and axonal neuropathies. Muscle Nerve 1995;18:730735.
18. Van Dijk GW, Notermans NC, Franssen H, Wokke JH. Devel-
opment of weakness in patients with chronic inflammatory
demyelinating polyneuropathy and only sensory symptoms at
presentation: a long term follow-up study. J Neurol 1999;246:
11341139.
MUSCLE & NERVE October 2001