1748

C OPYRIGHT 2015 BY T HE J OURNAL OF B ONE AND J OINT S URGERY, I NCORPORATED

Morbidity and Mortality Associated with Geriatric

Ankle Fractures
A Medicare Part A Claims Database Analysis
Raymond Y. Hsu, MD, Yoojin Lee, MS, MPH, Roman Hayda, MD, Christopher W. DiGiovanni, MD,
Vincent Mor, PhD, and Jason T. Bariteau, MD

Investigation performed at the Department of Orthopaedic Surgery, Brown University, Providence, Rhode Island

Background: The purpose of this study was to examine the incidence of adverse events in elderly patients who required
inpatient admission after sustaining an ankle fracture and to consider these data in relation to geriatric hip fracture and
other geriatric patient admissions.
Methods: A retrospective cohort study of patients admitted with an ankle fracture, a hip fracture, or any other diagnosis
was performed with the Medicare Part A database for 2008. The primary outcome measure was the one-year mortality
rate, examined with multivariate analysis factoring for both patient age and preexisting comorbidity. Secondary outcome
measures analyzed additional morbidity as reected by length of stay, discharge disposition, readmissions, and medical
complications.
Results: There were 19,648 patients with ankle fractures, 193,980 patients with hip fractures, and 5,801,831 patients
with other admitting diagnoses. Signicant differences (p < 0.001) were noted in both age and comorbidity status between
the group with ankle fractures and the group with hip fractures. The one-year mortality after admission was 11.9% for
patients with ankle fracture, 28.2% for patients with hip fracture, and 21.5% for patients with any other admission. Upon
using multivariate analysis to account for both age and comorbidity, the hazard ratio for one-year mortality associated with
fracture was 1.088 for patients with hip fracture and 0.557 for patients with ankle fracture.
Conclusions: Even after selecting for admitted patients and accounting for both age and comorbidity, geriatric patients
with ankle fractures were found to have a lower one-year morbidity compared with geriatric patients who had sustained a
hip fracture or alternative admitting diagnoses. Geriatric patients with ankle fractures are likely healthier and more active
in ways that are not captured by simply accounting for age and comorbidity. These ndings may support more aggressive
denitive management of such injuries in this population.
Level of Evidence: Prognostic Level III. See Instructions for Authors for a complete description of levels of evidence.

Peer Review: This article was reviewed by the Editor-in-Chief and one Deputy Editor, and it underwent blinded review by two or more outside experts. It was also reviewed
by an expert in methodology and statistics. The Deputy Editor reviewed each revision of the article, and it underwent a nal review by the Editor-in-Chief prior to publication.
Final corrections and clarications occurred during one or more exchanges between the author(s) and copyeditors.

A
mong the U.S. geriatric population, ankle fractures are
the third most common extremity fracture1. Only hip
and distal radial fractures occur at a denitively higher
rate in this population1-3. Previous work using Medicare Part B
claims demonstrated an annual incidence of ankle fractures
among geriatric patients of 1.4 per 1000 beneciaries4. Re-
cently, several studies based on Scandinavian registries have
demonstrated that the burden of geriatric ankle fractures is
expected to continue to grow given an aging population5,6.
The general trend toward open reduction and internal
xation as standard treatment for displaced, unstable ankle
fractures includes the geriatric population7,8. In earlier studies,

Disclosure: None of the authors received payments or services, either directly or indirectly (i.e., via his or her institution), from a third party in support of any
aspect of this work. One or more of the authors, or his or her institution, has had a nancial relationship, in the thirty-six months prior to submission of this
work, with an entity in the biomedical arena that could be perceived to inuence or have the potential to inuence what is written in this work. No author has
had any other relationships, or has engaged in any other activities, that could be perceived to inuence or have the potential to inuence what is written in this
work. The complete Disclosures of Potential Conicts of Interest submitted by authors are always provided with the online version of the article.

J Bone Joint Surg Am. 2015;97:1748-55 d http://dx.doi.org/10.2106/JBJS.O.00095

 1749
TH E JO U R NA L O F B O N E & JO I N T SU RG E RY J B J S . O RG
d
M O R B I D I T Y A N D M O R TA L I T Y A S S O C I AT E D WITH G E R I AT R I C
V O LU M E 9 7-A N U M B E R 2 1 N O V E M B E R 4, 2 015
d d
AN K L E FR AC T U R E S

TABLE I ICD-9-CM Codes Used to Analyze Ninety-Day Complications

Ninety-Day Complication ICD-9-CM Codes

Surgical site infection 998.51, 998.59, 996.60, 996.67, 996.69

Deep venous thrombosis 453.40, 453.41, 453.42, 997.2
Pulmonary embolism 415.11, 997.3
Congestive heart failure 428.0, 428.1, 428.20, 428.21, 428.22, 428.23, 428.30
Pneumonia 480.0-480.9, 481, 482.0, 482.1, 482.2 482.30, 482.31, 482.32, 482.39
Urinary tract infection 599.0
Pressure ulcer 707.00-707.07
Myocardial infarction 410, 410.0-410.6
Clostridium difcile infection 8.45
Gastrointestinal bleeds 578, 578.0, 578.1, 578.9

both operative and nonoperative treatment of ankle fractures their counterparts with hip fracture8,21-23. However, ankle fractures
in the geriatric population reected unacceptably high com- in geriatric patients frequently pose technical challenges that may
plication rates9,10. More recently, multiple retrospective series of be more difcult to manage than in patients with hip fracture
geriatric ankle fractures treated by modern methods demon- secondary to tenuous soft-tissue envelopes and osteopenic bone.
strated improved short-term outcomes and complication rates This study is not designed to specically investigate the potential
similar to those of younger cohorts11-17. wound or xation complications. However, these fractures fre-
However, to our knowledge, very few data are available quently require prolonged distal extremity immobilization and
concerning the one-year morbidity and mortality of geriatric non-weight-bearing, placing additional stresses on these patients,
patients with ankle fractures. In contrast, the one-year mortality which we hypothesized might increase morbidity and mortality.
and morbidity associated with hip fractures is well documented.
Multiple hip fracture studies have demonstrated a one-year Materials and Methods
mortality of up to 30%18,19. In 2011, Shivaranthre et al. reported
on a retrospective series on ninety-two operative ankle fractures F ollowing institutional review board approval, access to complete (100%)
2008 data from the Centers for Medicare and Medicaid Services was ob-
tained through Data Use Agreement #25776. Medicare Part A claims were used
of patients who were older than eighty years of age12. They found
to select for inpatient admissions. Codes from the International Classication
82% of patients returning to pre-injury mobility by three to six of Diseases, Ninth Revision, Clinical Modication (ICD-9-CM) were used to
months post-injury but also a 12% one-year postoperative create two cohorts reecting patients with new diagnoses of ankle fractures
mortality rate12. Koval et al. used a 20% sample of Medicare Part (824.0-824.9) and hip fractures (820.0-820.9). Patients younger than sixty-ve
B claims data, representing patients who are sixty-ve years years of age at the time of their diagnosed injury were excluded to only select
of age and older, and similarly demonstrated an 8.4% one-year geriatric patients. Patients in each cohort who had sustained a previous fracture
mortality rate after ankle fracture20. of the same type in the preceding year were excluded to eliminate any con-
founding by the previous injury. All Medicare inpatient admissions in 2008 for
The purpose of this study was to better characterize the
any reason other than ankle or hip fractures served as a control cohort.
morbidity and mortality of ankle fractures in the elderly pop- The primary outcome measure selected was the one-year post-injury
ulation. Hip fractures were chosen as a comparator because of mortality. Each Medicare beneciary is represented by a unique Denominator
the substantial literature that has been dedicated to this injury
in the geriatric population. To more evenly compare with pa-
tients with hip fracture and to select geriatric patients who were
physiologically elderly, we selected patients who required a
hospitalization after their ankle injury. As the population older
than sixty-ve years of age increasingly includes physiologically
younger individuals who may not need inpatient hospitaliza-
tion after an ankle fracture, this study intentionally excluded
those patients.
We hypothesized that, despite modern treatment advances,
geriatric ankle fractures are associated with signicant one-year
morbidity and mortality. In spite of this, these fractures have not
garnered the same public health attention as geriatric hip frac-
tures. We proposed this hypothesis understanding that evidence
suggests that geriatric ankle fractures tend to occur in a cohort Fig. 1
of younger, heavier, and more active individuals compared with Flowchart of inclusion criteria for geriatric ankle and hip fracture cohorts.
 1750
TH E JO U R NA L O F B O N E & JO I N T SU RG E RY J B J S . O RG
d
M O R B I D I T Y A N D M O R TA L I T Y A S S O C I AT E D WITH G E R I AT R I C
V O LU M E 9 7-A N U M B E R 2 1 N O V E M B E R 4, 2 015
d d
AN K L E FR AC T U R E S

Results

A mong the more than 7 million patients identied with a

Fig. 2
One-year survivorship following inpatient admission in the geriatric Medi-
care population for ankle fracture, hip fracture, and all other admitting
diagnoses.
File, which contains the requisite demographic characteristics, including date of
birth and date of death. Mortality rates for patients with ankle fracture were
evaluated by decade of life in comparison with patients with hip fracture and all
other admission cohorts.
As ankle fractures and hip fractures are believed to occur in non-
equivalent cohorts of geriatric patients, age differences and preexisting conditions
could therefore confound the primary outcome of mortality. Age subgroup
analysis was therefore performed to account for the difference in age distribution.
To identify differences in preexisting health status, the mean Elixhauser and
Charlson-Deyo comorbidity indices were determined on the basis of ICD-9-CM
24,25
codes prior to injury . Finally, Cox regression multivariate analysis was per-
formed to isolate the relative contributions of the injury, ankle or hip fracture, to
one-year mortality independent of age and comorbidities.
Secondary outcome measures were focused on injury morbidity as
represented by length of inpatient hospitalization, discharge to nursing home,
readmissions, and medical complications. The primary claim le was analyzed
for inpatient length of stay. The Minimum Data Set, a required form upon
admission to a nursing home, was used to determine discharge to nursing
home. Readmissions within thirty days and complications within ninety days
following the denitive fracture claim were determined by linking the De-
nominator File to further claims. To exclude transfers between facilities from
being represented as readmissions, additional admission claims were counted
as readmissions only if there was a minimum one-day separation from a pre-
vious discharge. The following medical complications were identied by their
respective ICD-9-CM codes: surgical site infection, deep venous thrombosis,
pulmonary embolism, congestive heart failure, pneumonia, urinary tract in-
fection, pressure ulcer, myocardial infarction, Clostridium difcile infection, and
gastrointestinal bleeds (Table I).
Medicare fee-for-service inpatient Part A claim in 2008, we
identied 19,648 patients who were sixty-ve years of age or
older, had been diagnosed with an ankle fracture in 2008, and
had not had an ankle fracture inpatient admission in 2007. We
identied 193,980 patients with hip fracture meeting the same
criteria to serve as the comparison group (Fig. 1). There were
180 patients who sustained both an ankle fracture and a hip
fracture in 2008, and these patients were included in both
cohorts. The control cohort of Medicare patients older than
sixty-ve years of age admitted with any diagnosis other than a
hip or ankle fracture included 5,801,831 patients. The cohorts
differed signicantly (p < 0.001) with respect to mean age, sex,
and comorbidity burden (Table II). The ankle fracture cohort
was both younger and less aficted by comorbidity.
Among these cohorts of admitted patients, the one-year
mortality was 11.9% after an ankle fracture and 28.2% after a
hip fracture. In the control cohort, the one-year mortality after
any other admission diagnosis was 21.5% (Fig. 2). Parsing out
the age subgroups by decade of life reveals the large variation in
mortality across age groups (Fig. 3). This correlation suggests
that the difference in mortality rates in the fracture cohorts may
largely be due to the difference in age distribution.
Without factoring in patient age or any comorbidities,
the hazard ratio for one-year mortality after admission with
an ankle fracture was 0.520 compared with the control (all
other admissions) cohort. The same hazard ratio for patients
with hip fracture was found to be 1.361 (Table III). As sug-
gested by the results above, these hazard ratios do not ac-
count for the substantial contributions of age and preexisting
comorbidities.

Statistical Analysis
Cohort characteristics were summarized by calculating both means and stan-
dard deviations for the continuous variables as well as frequency distributions
for the categorical variables. Unadjusted Cox proportional hazard survival models
were used to examine the one-year mortality for each cohort. Age-adjusted,
Fig. 3
Elixhauser score-adjusted, and both age and Elixhauser score-adjusted models
One-year mortality after geriatric ankle fracture, hip fracture, and all other
were built. All statistical analysis was performed with use of SAS version 9.3 (SAS
Institute, Cary, North Carolina). admissions in the Medicare population, combined and separated by age
subgroups. The one-year mortalities for the ankle fracture cohort and the
Source of Funding hip fracture cohort are signicantly different overall and within each age
There were no external funding sources for this study. subgroup at p < 0.001.
 1751
TH E JO U R NA L O F B O N E & JO I N T SU RG E RY J B J S . O RG
d
M O R B I D I T Y A N D M O R TA L I T Y A S S O C I AT E D WITH G E R I AT R I C
V O LU M E 9 7-A N U M B E R 2 1 N O V E M B E R 4, 2 015
d d
AN K L E FR AC T U R E S

TABLE II Cohort Demographic Characteristics

Cohort
Ankle Fracture Hip Fracture All Other Admissions

Total no. of patients 19,648 193,980 5,801,831

Age* (yr) 77.5 8.0 (77.4 to 77.6) 83.6 7.6 (83.6 to 83.7) 78.9 8.1 (78.9 to 78.9)
Age distribution
Sixty-ve to less than seventy-ve years 42.7% 14.6% 35.9%
Seventy-ve to less than eight-ve years 37.4% 38.9% 39.3%
Eighty-ve to less than ninety-ve years 18.2% 41.2% 22.4%
Ninety-ve years or older 1.7% 5.3% 2.4%
Sex*
Male 21.6% 26.6% 42.4%
Female 78.4% 73.4% 57.6%
Nursing home stay within one year prior to fracture* 12.6% 22.0%
Comorbidities*
Elixhauser score (points) 2.3 1.3 (2.3 to 2.3) 2.5 1.3 (2.5 to 2.5) 2.4 1.3 (2.4 to 2.4)
Deyo-Charlson score (points) 1.1 1.3 (1.1 to 1.1) 1.2 1.4 (1.2 to 1.2) 1.6 1.8 (1.6 to 1.6)

*There was a signicant difference at p < 0.001 between the ankle fracture cohort and the hip fracture cohort. The values are given as the mean
and the standard deviation, with the 95% condence interval in parentheses. The condence intervals were very tight because the standard errors
of the mean all approached 0. The values are given as the percentage of patients.

Sequential multivariate analyses separate and reveal the The secondary outcomes of morbidity were reected
relative contributions of age, comorbidity, and injury to mor- by length of stay, disposition after discharge, readmissions, and
tality. Factoring both fracture and age quanties the substantial thirty-day and ninety-day complications (Table IV). The mean
contribution of patient age to one-year mortality. Dening length of stay was signicantly lower (p < 0.0001) for the ankle
patients in the rst decade of Medicare age eligibility of sixty- fracture cohort (4.6 days) compared with the hip fracture co-
ve to seventy-four years as the reference group, the hazard hort (6.0 days). The mean discharge rate to a nursing home
ratio for mortality increases with each additional decade. Once was also signicantly lower (p < 0.0001) for the ankle frac-
the contribution of age is factored separately, the hazard ratio ture cohort (59.2%) compared with the hip fracture cohort
for a hip fracture falls to 1.088 and that of an ankle fracture
changes only slightly to 0.551.
To avoid duplication, only one comorbidity index, the
Elixhauser score, was chosen because of some evidence that it
may be a better predictor of mortality than the Charlson-Deyo
score26-28. Factoring comorbidities as represented by Elixhauser
score demonstrates a substantial contribution but to a lesser
degree than patient age.
Standardizing for both age and comorbidity, the hazard
ratio for one-year mortality after admission was 1.081 for hip
fracture and 0.557 for ankle fracture (compared with 1 for all
other admissions). In other words, the hazard ratio for one-
year mortality after admission is higher for hip fracture but
substantially lower for ankle fracture when compared with the
average admission. This is best visually demonstrated by cal-
culating survivorship standardizing to median age and median
Elixhauser score to eliminate their contributions (Fig. 4). In
comparison with Figure 2, which did not factor for age and
comorbidities, the survivorship of patients with hip fracture Fig. 4
in this standardized model approaches that of all other admis- Calculated one-year survivorship standardizing for age and comorbidities,
sions, but that of patients with ankle fracture is relatively set to a median age of seventy-ve to eighty-four years and a median
unchanged. Elixhauser score of 2 points.
 1752
TH E JO U R NA L O F B O N E & JO I N T SU RG E RY J B J S . O RG
d
M O R B I D I T Y A N D M O R TA L I T Y A S S O C I AT E D WITH G E R I AT R I C
V O LU M E 9 7-A N U M B E R 2 1 N O V E M B E R 4, 2 015
d d
AN K L E FR AC T U R E S

TABLE III Sequential Multivariate Analysis to Account for Age and Comorbidities*

One-Year Mortality
Fracture and Fracture, Age,
Covariate Factors Fracture Only Fracture and Age Comorbidities and Comorbidities

Reference group
Age in years
75 to 84
85 to 94
95
Ankle fracture group
Hip fracture group
Comorbidity (for each
additional point in the
Elixhauser score)
All other admissions All other admissions
and age of 65 to 74 yr
NA 1.604 (1.597 to 1.612)
NA 2.744 (2.731 to 2.757)
NA 4.521 (4.484 to 4.558)
0.520 (0.499 to 0.542) 0.551 (0.529 to 0.574)
1.361 (1.350 to 1.373) 1.088 (1.079 to 1.097)
NA NA
All other admissions, All other admissions,
Elixhauser score of 0 age of 65 to 74 yr, and
Elixhauser score of 0
NA 1.567 (1.560 to 1.574)
NA 2.644 (2.632 to 2.656)
NA 4.367 (4.331 to 4.403)
0.528 (0.507 to 0.549) 0.557 (0.535 to 0.580)
1.333 (1.321 to 1.344) 1.081 (1.072 to 1.091)
1.176 (1.174 to 1.177) 1.150 (1.149 to 1.152)

*NA = not applicable. The variables are given as the hazard ratio, with the 95% condence interval in parentheses. The hazard ratio for this
group is 1.

(71.4%). There were clinically similar but signicantly different
thirty-day readmission rates (p < 0.0001) between the ankle
fracture cohort (13.3%) and the hip fracture cohort (15.0%).
However, there was a signicantly lower association (p < 0.0001)
with thirty-day mortality rates for ankle fractures (1.9%) com-
pared with hip fractures (6.8%). The two cohorts had essentially
equivalent deep venous thrombosis rates (p = 0.5053), at 2.6%
for the ankle fracture cohort compared with 2.7% for the hip
fracture cohort. However, the rate of pulmonary embolus was
signicantly higher (p < 0.0001) in the hip fracture cohort (1.9%)
compared with the ankle fracture cohort (0.9%). The rate of
surgical site infections was signicantly higher (p < 0.0001) in the

TABLE IV Morbidity Following Admission

Cohort
Ankle Fracture Hip Fracture All Other Admissions P Value*

Length of stay (d) 4.6 4.1 6.0 4.0 4.9 5.4 <0.0001
Discharge to nursing home 59.2% 71.4% NA <0.0001
Thirty-day readmission 13.3% 15.0% NA <0.0001
Thirty-day mortality 1.9% 6.8% NA <0.0001
Ninety-day complications
Deep venous thrombosis 2.6% 2.7% 2.9% 0.5053
Pulmonary embolus 0.9% 1.9% 0.7% <0.0001
Surgical site infection 3.6% 1.5% NA <0.0001
Pneumonia 1.2% 1.8% 2.6% <0.0001
Congestive heart failure 15.5% 20.6% 22.4% <0.0001
Urinary tract infection 16.5% 29.1% 16.7% <0.0001
Pressure ulcer 3.4% 7.5% 3.3% <0.0001
Myocardial infarction 3.0% 4.9% 6.8% <0.0001
C. difcile infection 1.6% 2.5% 1.7% <0.0001
Gastrointestinal bleed 2.3% 3.4% 4.6% <0.0001

*The p value was determined by a comparison between the ankle fracture cohort and the hip fracture cohort. The values are given as the mean
and the standard deviation. The values are given as the percentage of patients. NA = not available.
 1753
TH E JO U R NA L O F B O N E & JO I N T SU RG E RY J B J S . O RG
d
V O LU M E 9 7-A N U M B E R 2 1 N O V E M B E R 4, 2 015
d d
ankle fracture cohort (3.6%) compared with the hip fracture
cohort (1.5%). Otherwise, the ninety-day complication rate for
the hip fracture cohort was signicantly higher (p < 0.0001) for
all other outcomes investigated.
Discussion
T he purpose of this study was to characterize the one-year
mortality and morbidity rates associated with ankle frac-
ture occurrence in the geriatric population when these injuries
led to inpatient admission. Similarly available data on the elderly
population with hip fracture were used as a comparator. Using
Medicare Part A claims data helps to select geriatric patients with
ankle fracture who require at least temporary hospitalization
or other institutionalization secondary to a combination of their
injury, immobilization, and non-weight-bearing status. This
purposefully biases the data set and omits patients who are able
to return home without any type of admission and who we
believed to be physiologically younger than the patients whom
this work was intended to assess.
Even after selecting for patients with ankle fracture who
required inpatient admission, the demographic characteristics
of the cohorts in this study revealed that when patients with
ankle fracture were compared with patients with hip fracture,
they tended to be younger, healthier, and less likely to have
already had a nursing home admission. This difference in de-
mographic characteristics is consistent with previously pub-
lished work on geriatric ankle fractures that has identied
younger age and higher physical activity level to be risk fac-
tors8,21,22,29. Further support for differences between the two
groups, although not elucidated in this study, is that obese
geriatric women are at a threefold increased risk of sustaining
ankle fractures but they are at a decreased risk of sustaining
hip fractures29.
Given the large differences in age and comorbidities
among the cohorts, a direct comparison of one-year mortality
is not particularly revealing, but is a good starting point and
provides prognostic information. The primary outcome of
one-year mortality of the ankle fracture cohort in this study
(11.9%) was signicantly less than both that of the hip fracture
cohort (28.2%) and that of all other admissions (21.5%). The
hip fracture mortality rate is consistent with that of previously
published results that have hovered around 30%18,19. The ankle
fracture mortality rate is also similar to that of previously
published studies (8.4% to 12%)12,20. The mortality rate from
the Medicare Part A claims in this study is on the higher end
likely because of the intentional bias toward patients who re-
quired hospitalization.
To the best of our knowledge, this work represents the
rst effort to use a database for examining mortality after ankle
fracture while factoring for both preexisting comorbidity burden
and age cohort. After examination of the variation in one-year
mortality across age groups for ankle fractures, hip fractures, and
all other admissions, it becomes readily apparent that age is a
major factor.
Factoring out the contributions of age and comorbidity
by multivariate analysis brings the one-year mortality hazard
M O R B I D I T Y A N D M O R TA L I T Y A S S O C I AT E D WITH G E R I AT R I C
AN K L E FR AC T U R E S
ratio of hip fractures closer to that of any other admission. A
substantial portion of the mortality associated with a hip fracture
is therefore presumably due to a combination of the preexisting
age and relative comorbidities of each patient. These results re-
inforce the common notion of a hip fracture as a sentinel event
representing preexisting conditions. In contrast, the same mul-
tivariate analysis did not dramatically adjust the one-year mor-
tality hazard ratio of ankle fractures. This difference suggests that
selecting for patients with ankle fractures may also select for
traits protective of morbidity beyond younger age and fewer
comorbidities. The most likely candidate traits are general im-
proved tness and higher activity level, which are not reected by
simply adjusting for age and Elixhauser score. Previous studies
have suggested that higher activity level is a risk factor for ankle
fractures8. Any such difference in characterization between pa-
tients with ankle fracture and those with hip fracture becomes
even more impressive when considering that this study was
purposefully biased to exclude patients with ankle fracture
who never required admission.
The secondary outcomes of ninety-day complications
used as a proxy for morbidity need to be interpreted in the
context of the underlying differences between the two cohorts,
but can still be used to provide guidance to patients. The higher
rate of surgical site infections in patients with ankle fracture
compared with patients with hip fracture is consistent with
general expectations for distal lower-extremity surgery. The
equivalent rates of deep venous thrombosis between the hip
fracture cohort and the ankle fracture cohort but the higher
rate of pulmonary embolisms in the hip fracture cohort stand
in contrast to prevailing opinion and are difcult to interpret. A
difference in anatomic location of the deep venous thrombosis
may be contributory, as distal deep venous thrombosis is be-
lieved to be less likely to propagate30. The choice of prophylactic
regimens between the two cohorts remains unknown and likely
differs, which further complicates any interpretation of this
information.
The rate of pressure ulcers in the ankle fracture cohort
was less than half that of the hip fracture cohort. This is inter-
esting because the nature of immobility after these respective
injuries is often quite different: distal immobilization with non-
weight-bearing occurs most commonly for the ankle fracture
population but weight-bearing as tolerated, albeit with discom-
fort, occurs most commonly for hip fractures postoperatively.
The rate for all other ninety-day complications was lower
for ankle fractures compared with hip fractures. Similar to the
ndings for one-year mortality, these differences are likely
largely due to the preexisting differences in age and comor-
bidity between the two cohorts.
This investigation had several strengths. All reported data
were based on large population numbers and a well-captured
data set. Incorporation of multivariate analysis accounted for
the confounding variables of age and comorbidity, enabling
better comparison of the different cohorts.
There were several limitations to this study that should be
emphasized. The Medicare database presumably provides accu-
rate data concerning age, mortality, discharges, and readmission.
 1754
TH E JO U R NA L O F B O N E & JO I N T SU RG E RY J B J S . O RG
d
M O R B I D I T Y A N D M O R TA L I T Y A S S O C I AT E D WITH G E R I AT R I C
V O LU M E 9 7-A N U M B E R 2 1 N O V E M B E R 4, 2 015
d d
AN K L E FR AC T U R E S

However, the comorbidity indices rely on accurate ICD-9 presents with diabetes, neuropathy, or osteopenia continues to
coding. Furthermore, although the indices provide a conve- remain a treatment challenge. n
nient method of accounting for comorbidity, the individual
contributions of specic conditions of interest such as pe-
ripheral vascular disease, diabetes, and nicotine use were not
investigated.
Raymond Y. Hsu, MD
The ninety-day complications, which were predominantly Roman Hayda, MD
chosen as proxies for overall health, also depend on ICD-9 coding. Department of Orthopaedic Surgery,
Additionally, other than querying for surgical site infections, this Brown University,
study did not examine other local wound complications, failure of Suite 200, 2 Dudley Street,
xation, osteomyelitis, reoperation, or amputation. Providence, RI 02903.
This work conrms previously published one-year mor- E-mail address for R.Y. Hsu: Raymond_Hsu@brown.edu
tality rates for patients with ankle or hip fractures but creates an
Yoojin Lee, MS, MPH
important distinction between patients with hip fracture and Vincent Mor, PhD
those with ankle fracture in the elderly population. Although the Center for Gerontology and Health Care Research,
one-year morbidity of geriatric patients with hip fracture is School of Public Health,
largely a sign of preexisting age and health, ankle fractures in this Brown University,
population do not seem to represent this same type of sentinel Box G-S121-6,
event. In fact, it appears that ankle fractures are more likely to 121 South Main Street,
Providence, RI 02912
occur in geriatric patients with a lower risk of one-year mortality.
These ndings may support more aggressive and less expectant Christopher W. DiGiovanni, MD
management of geriatric patients with ankle fractures because Department of Orthopaedic Surgery,
their mortality is lower than that of the average admitted Harvard Medical School, Massachusetts General Hospital,
geriatric patient. Overall, the results of this study might have 55 Fruit Street,
important implications for providing additional guidance to Boston, MA 02114
patients and their providers during the care and decision-
Jason T. Bariteau, MD
making process for their injuries. However, it is still recom- Department of Orthopaedic Surgery,
mended that these data be interpreted with caution, because the Emory University School of Medicine,
database represents a very large and heterogeneous patient 59 Executive Park South,
population. The geriatric patient with ankle fracture who also Atlanta, GA 30329

References
1. Sporer SM, Weinstein JN, Koval KJ. The geographic incidence and treatment
variation of common fractures of elderly patients. J Am Acad Orthop Surg. 2006
Apr;14(4):246-55.
2. Karagas MR, Baron JA, Barrett JA, Jacobsen SJ. Patterns of fracture among the
United States elderly: geographic and uoride effects. Ann Epidemiol. 1996 May;
6(3):209-16.
3. Bell JE, Leung BC, Spratt KF, Koval KJ, Weinstein JD, Goodman DC, Tosteson
ANA. Trends and variation in incidence, surgical treatment, and repeat surgery of
proximal humeral fractures in the elderly. J Bone Joint Surg Am. 2011 Jan 19;
93(2):121-31.
4. Koval KJ, Lurie J, Zhou W, Sparks MB, Cantu RV, Sporer SM, Weinstein J. Ankle
fractures in the elderly: what you get depends on where you live and who you see.
J Orthop Trauma. 2005 Oct;19(9):635-9.
5. Kannus P, Palvanen M, Niemi S, Parkkari J, Jarvinen M. Increasing number and
incidence of low-trauma ankle fractures in elderly people: Finnish statistics during
1970-2000 and projections for the future. Bone. 2002 Sep;31(3):430-3.
6. Thur CK, Edgren G, Jansson KA, Wretenberg P. Epidemiology of adult ankle
fractures in Sweden between 1987 and 2004: a population-based study of 91,410
Swedish inpatients. Acta Orthop. 2012 Jun;83(3):276-81. Epub 2012 Mar 8.
7. Strauss EJ, Egol KA. The management of ankle fractures in the elderly. Injury.
2007 Sep;38(Suppl 3):S2-9.
8. Olsen JR, Hunter J, Baumhauer JF. Osteoporotic ankle fractures. Orthop Clin
North Am. 2013 Apr;44(2):225-41. Epub 2013 Feb 22.
9. Beauchamp CG, Clay NR, Thexton PW. Displaced ankle fractures in patients over
50 years of age. J Bone Joint Surg Br. 1983 May;65(3):329-32.
10. Litcheld JC. The treatment of unstable fractures of the ankle in the elderly.
Injury. 1987 Mar;18(2):128-32.
11. Srinivasan CM, Moran CG. Internal xation of ankle fractures in the very elderly.
Injury. 2001 Sep;32(7):559-63.
12. Shivarathre DG, Chandran P, Platt SR. Operative xation of unstable ankle
fractures in patients aged over 80 years. Foot Ankle Int. 2011 Jun;32(6):599-602.
13. Little MTM, Berkes MB, Lazaro LE, Sculco PK, Cymerman RM, Pardee N, Helfet
DL, Lorich DG. Comparison of supination external rotation type IV ankle fractures in
geriatric versus nongeriatric populations. Foot Ankle Int. 2013 Apr;34(4):512-7.
Epub 2013 Feb 27.
14. Fong W, Acevedo JI, Stone RG, Mizel MS. The treatment of unstable ankle frac-
tures in patients over eighty years of age. Foot Ankle Int. 2007 Dec;28(12):1256-9.
15. Pagliaro AJ, Michelson JD, Mizel MS. Results of operative xation of unstable
ankle fractures in geriatric patients. Foot Ankle Int. 2001 May;22(5):399-402.
16. Vioreanu M, Brophy S, Kearns S, Kelly E, Hurson B, ORourke SK, Quinlan W.
Ankle fractures in the geriatric population: Operative or non-operative treatment.
J Bone Joint Surg Br. 2006;88-B(SUPPL II):281.
17. Lynde MJ, Sautter T, Hamilton GA, Schuberth JM. Complications after open
reduction and internal xation of ankle fractures in the elderly. Foot Ankle Surg. 2012
Jun;18(2):103-7. Epub 2011 Apr 30.
18. Moran CG, Wenn RT, Sikand M, Taylor AM. Early mortality after hip fracture:
is delay before surgery important? J Bone Joint Surg Am. 2005 Mar;87(3):483-9.
19. Roche JJW, Wenn RT, Sahota O, Moran CG. Effect of comorbidities and post-
operative complications on mortality after hip fracture in elderly people: prospective
observational cohort study. BMJ. 2005 Dec 10;331(7529):1374. Epub 2005
Nov 18.
20. Koval KJ, Zhou W, Sparks MJ, Cantu RV, Hecht P, Lurie J. Complications after
ankle fracture in elderly patients. Foot Ankle Int. 2007 Dec;28(12):1249-55.
21. Seeley DG, Kelsey J, Jergas M, Nevitt MC; The Study of Osteoporotic Fractures
Research Group. Predictors of ankle and foot fractures in older women. J Bone Miner
Res. 1996 Sep;11(9):1347-55.
22. Hasselman CT, Vogt MT, Stone KL, Cauley JA, Conti SF. Foot and ankle fractures
in elderly white women. Incidence and risk factors. J Bone Joint Surg Am. 2003
May;85(5):820-4.
23. Pritchard JM, Giangregorio LM, Ioannidis G, Papaioannou A, Adachi JD, Leslie
WD. Ankle fractures do not predict osteoporotic fractures in women with or without
diabetes. Osteoporos Int. 2012 Mar;23(3):957-62. Epub 2011 May 12.
 1755
TH E JO U R NA L O F B O N E & JO I N T SU RG E RY J B J S . O RG
d
V O LU M E 9 7-A N U M B E R 2 1 N O V E M B E R 4, 2 015
d d
24. Quan H, Sundararajan V, Halfon P, Fong A, Burnand B, Luthi JC, Saunders LD,
Beck CA, Feasby TE, Ghali WA. Coding algorithms for dening comorbidities in ICD-9-
CM and ICD-10 administrative data. Med Care. 2005 Nov;43(11):1130-9.
25. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying
prognostic comorbidity in longitudinal studies: development and validation. J Chronic
Dis. 1987;40(5):373-83.
26. Menendez ME, Neuhaus V, van Dijk CN, Ring D. The Elixhauser comorbidity
method outperforms the Charlson index in predicting inpatient death after ortho-
paedic surgery. Clin Orthop Relat Res. 2014 Sep;472(9):2878-86. Epub 2014
May 28.
27. Southern DA, Quan H, Ghali WA. Comparison of the Elixhauser and Charlson/
Deyo methods of comorbidity measurement in administrative data. Med Care. 2004
Apr;42(4):355-60.
M O R B I D I T Y A N D M O R TA L I T Y A S S O C I AT E D WITH G E R I AT R I C
AN K L E FR AC T U R E S
28. Chu YT, Ng YY, Wu SC. Comparison of different comorbidity measures for use
with administrative data in predicting short- and long-term mortality. BMC Health
Serv Res. 2010 May 27;10:140.
29. Armstrong MEG, Cairns BJ, Banks E, Green J, Reeves GK, Beral V; Million
Women Study Collaborators. Different effects of age, adiposity and physical activity
on the risk of ankle, wrist and hip fractures in postmenopausal women. Bone. 2012
Jun;50(6):1394-400. Epub 2012 Mar 21.
30. Kearon C, Akl EA, Comerota AJ, Prandoni P, Bounameaux H, Goldhaber SZ,
Nelson ME, Wells PS, Gould MK, Dentali F, Crowther M, Kahn SR, American College
of Chest Physicians. Antithrombotic therapy for VTE disease: Antithrombotic Therapy
and Prevention of Thrombosis, 9th ed: American College of Chest Physicians
Evidence-Based Clinical Practice Guidelines. Chest. 2012 Feb;141(2 Suppl):
e419S-94S.