Impact of Age, Injury Severity Score, and Medical Comorbidities on Early Complications After Fusion and Halo-Vest Immobilization for C2 Fractures in Older Adults A Propensity Score Matched Retrospective Cohort Study

Article (PDF Available)inSpine 37(10):854-9 · September 2011with12 Reads
DOI: 10.1097/BRS.0b013e3182377486 · Source: PubMed
Abstract
Propensity score matched retrospective cohort study. To report early complication rates and associated risk factors in patients with C2 fractures who underwent fusion or halo immobilization. There is limited data on the impact of age, injury severity score, and medical comorbidities on overall complication rates from surgical fixation versus halo-vest immobilization of C2 fractures. The Nationwide Inpatient Sample database from 2002 to 2008 was queried to identify cohorts of adult patients (age ≥ 18 years) with C2 fractures without spinal cord injury who were treated with either fusion or halo-vest immobilization. Complication rates, hospital length of stay, and costs were compared in a propensity score matched sample. Multivariate analysis was used to identify predictors of in-hospital complications. A total of 3758 patients (1627 fusion and 2131 halo) were identified. Fusion was associated with greater overall complication rates (20.2% vs. 10.1%, P < 0.0001), increased length of stay (8.9 d vs. 6.4 d, P < 0.0001), higher charges ($80,000 vs. $41,000, P < 0.0001), but a lower rate of nonroutine discharge (52.6% vs. 62.6%, P < 0.0001). There was no difference in mortality between the fusion group (2.75%) and the halo group (3.33%). Age, injury score, and comorbidity increased complication rates by a similar degree (odds ratio) in both cohorts. Patients aged 80 years and older were 3.5 times more likely to have a complication than those younger than 60 years. Fusion patients had greater overall complication rates, increased length of stay, and greater resource utilization but were discharged home in a greater proportion. Both fusion and halo were associated with significant (more than 3-fold) increase in complication rates in elderly patients aged 80 years or older. Given the similar mortality rate between the fusion group and the halo group and the higher cost and complication rate in the fusion group, our study supports the use of halo-vest immobilization in patients where operative therapy is contraindicated.

Full-text (PDF)

Available from: Yi-Ren Chen, Aug 12, 2016
EPIDEMIOLOGY
SPINE Volume 37, Number 10, pp 854–859
©2012, Lippincott Williams & Wilkins
854 www.spinejournal.com May 2012
Impact of Age, Injury Severity Score, and Medical
Comorbidities on Early Complications After Fusion
and Halo-Vest Immobilization for C2 Fractures in
Older Adults
A Propensity Score Matched Retrospective Cohort Study
Maxwell Boakye , MD,* Robert T. Arrigo , BS , Paul S. A. Kalanithi , MD , and Yi-Ren Chen , BA, BS
Study Design. Propensity score matched retrospective cohort
study.
Objective. To report early complication rates and associated risk
factors in patients with C2 fractures who underwent fusion or halo
immobilization.
Summary of Background Data. There is limited data on the
impact of age, injury severity score, and medical comorbidities on
overall complication rates from surgical xation versus halo-vest
immobilization of C2 fractures.
Methods. The Nationwide Inpatient Sample database from 2002
to 2008 was queried to identify cohorts of adult patients (age 18
years) with C2 fractures without spinal cord injury who were treated
with either fusion or halo-vest immobilization. Complication rates,
hospital length of stay, and costs were compared in a propensity
score matched sample. Multivariate analysis was used to identify
predictors of in-hospital complications.
Results. A total of 3758 patients (1627 fusion and 2131 halo) were
identi ed. Fusion was associated with greater overall complication
rates (20.2% vs . 10.1%, P < 0.0001), increased length of stay (8.9 d
vs . 6.4 d, P < 0.0001), higher charges ($80,000 vs . $41,000, P <
0.0001), but a lower rate of nonroutine discharge (52.6% vs . 62.6%,
P < 0.0001). There was no difference in mortality between the
fusion group (2.75%) and the halo group (3.33%). Age, injury score,
Cervical spine fractures are a signifi cant cause of disabil-
ity and pain, accounting for 60% of all spinal injuries
in the older adults.
1
5
The incidence of axis (C2) frac-
tures, which account for the majority of cervical spine frac-
tures in the older adults, is increasing because of the aging
US population.
1
,
6
The management of C2 fractures is fraught
with diffi culty because of lack of adequate information in
the peer-reviewed literature regarding the morbidities associ-
ated with treatment. Management options for C2 fractures
include anterior odontoid fi xation, posterior cervical fusion,
and external immobilization with halo or hard collar, all of
which are associated with signifi cant rates of morbidity and
mortality.
3
,
7
,
8
However, there is a lack of data on compara-
tive risks of surgery versus halo in the older adults. In par-
ticular, the incremental risk of age in octogenarians treated
with surgery or halo compared with younger cohorts has not
been quantifi ed. Although it is generally assumed that halo
complications are signifi cantly increased in octogenarians,
9
11
few studies have actually compared complication rates in
octogenarians with those of patients in their sixties. Similarly,
the extent to which risks are mitigated in elderly patients with
From the * Center for Advanced Neurosurgery, University of Louisville,
Louisville, KY; School of Medicine, Stanford University Stanford, CA; and
Department of Neurosurgery, Stanford University Medical Center, Palo Alto,
CA.
Acknowledgment date: January 14, 2011. First revision date: August 8, 2011.
Second revision date: September 4, 2011. Acceptance date: September 11,
2011.
The manuscript submitted does not contain information about medical
device(s)/drug(s).
No funds were received in support of this work.
No bene ts in any form have been or will be received from a commercial
party related directly or indirectly to the subject of this manuscript.
Address correspondence and reprint requests to Maxwell Boakye, MD, FACS,
Center for Advanced Neurosurgery, University of Louisville, 220 Abraham
Flexner Way, 1105, Louisville, KY 40202; E-mail: max.boakye@louisville.edu
and comorbidity increased complication rates by a similar degree
(odds ratio) in both cohorts. Patients aged 80 years and older were
3.5 times more likely to have a complication than those younger
than 60 years.
Conclusion. Fusion patients had greater overall complication
rates, increased length of stay, and greater resource utilization but
were discharged home in a greater proportion. Both fusion and
halo were associated with signi cant (more than 3-fold) increase in
complication rates in elderly patients aged 80 years or older. Given
the similar mortality rate between the fusion group and the halo
group and the higher cost and complication rate in the fusion group,
our study supports the use of halo-vest immobilization in patients
where operative therapy is contraindicated.
Key words: C2 fracture , complications , comorbidities , injury
severity score , Nationwide Inpatient Sample , propensity score
matching. Spine 2012 ; 37 : 854 859
DOI: 10.1097/BRS.0b013e3182377486
Copyright © 2012 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
BRS204730.indd 854BRS204730.indd 854 07/04/12 1:07 PM07/04/12 1:07 PM
EPIDEMIOLOGY Outcomes for C2 Fractures • Boakye et al
Spine www.spinejournal.com 855
no comorbidities is also poorly understood and the studies
that do exist in the literature are limited by relatively small
sample sizes in single-institution studies ( < 40 patients).
9
,
12
14
In this study, we used the Healthcare Cost and Utiliza-
tion Project’s Nationwide Inpatient Sample (NIS) databases
to identify cohorts of patients with acute C2 fractures who
underwent halo fi xation or spinal fusion. Our intent was to
perform a retrospective matched cohort comparison of in-
hospital major complication rates and to assess the impact of
age, comorbidities, and injury score within each cohort. It is
hoped that the results will fi ll some current gaps in knowledge
in the peer-reviewed literature regarding management of C2
fractures.
MATERIALS AND METHODS
Data Source and Inclusion/Exclusion Criteria
The NISs, produced by the Healthcare Cost and Utilization
Project, contains inpatient discharge data on approximately
20% of hospitalizations in the United States and employs a
single-stage, stratifi ed, and clustered (around hospitals) design
allowing accurate derivation of national estimates. The 2002
through 2008 NIS databases were assayed for adult patients
(age 18 years) with a primary International Classifi cation of
Disease, Ninth Revision, Clinical Modifi cation ( ICD-9-CM )
diagnosis code of closed fracture of the second cervical ver-
tebra (805.02). The primary procedure code was used to
identify patients receiving fusion ( International Classifi cation
of Disease, Ninth Revision [ ICD-9 ] procedure code 81.01,
81.02, or 81.03), other fracture repair (03.53), or halo-vest
placement (02.94 or 93.41). Patients were included in the
study if their primary procedure code was either surgical
(fusion or other fracture repair) or halo placement. Patients
receiving both surgery and halo-vest placement were excluded
from the analysis. Patients were also excluded if they had any
diagnosis code indicating spinal cord injury (806) or patho-
logical fracture (733.1).
Recorded and Derived Data
Age, sex, discharge status (routine [ i.e ., to home] versus
nonroutine [ i.e ., to anywhere other than home]), length of
inpatient stay, total charges, and inpatient mortality were
noted for each hospitalization. The Elixhauser method,
15
a
well-established technique for identifying comorbidities from
ICD-9 diagnosis codes found in administrative databases,
such as NIS, was used to compute an overall comorbidity
score for each patient by adding 1 point per comorbidity
(maximum possible, 24); we excluded paralysis as a possible
comorbidity. Age was treated as a categorical variable with
groups: younger than 60 years, 60 to 69 years, 70 to 79 years,
and 80 years or older.
We used the International Classifi cation of Disease –9
based Injury Severity Score (ICISS)
16
combined with the most
recent survival risk ratios provided by the American College
of Surgeons
17
to derive an estimated probability of death for
each patient. ICISS defi nes a survival risk ratio for each ICD-
9 diagnosis code in a set of diagnosis codes associated with
traumatic injury. The product of the individual survival risk
ratios provides an overall estimate of the probability of sur-
vival ranging from 0 (certain death) to 1 (certain survival).
To provide a more intuitive measure of trauma severity ( i.e .,
where increasing trauma score is associated with increasing
expected probability of death), we subtracted the calculated
overall survival probability from 1 and then multiplied by
100 to yield an ICISS-based traumatic injury severity score
ranging from 0 (certain survival) to 100 (certain death) for
each patient.
In-hospital postoperative complications were identifi ed by
ICD-9-CM diagnosis codes as follows: renal (584, 997.5), car-
diac (997.1, 410.0–410.91), neurological (997.00–997.09),
deep-vein thrombosis or pulmonary embolism (415.1, 415.11,
415.19, 451.1, 451.11, 451.19, 451.2, 451.81, 451.9, 453.4,
453.40, 453.41, 453.42, 453.8, 453.9), pulmonary (507.0,
518.4, 518.5, 518.81, 518.82, 997.3, 997.31, 997.39), infec-
tion (038, 320, 510, 513.1, 519.2, 590.1, 590.80, 683), and
wound complication (998.1–998.9).
Statistical Analysis
All statistical analyses were performed using complex survey
design analysis to derive variance estimates properly using
discharge weights, clustering around hospitals and stratifi ca-
tion over hospital, and regional factors as specifi ed by NIS.
The raw number of cases ( i.e ., unweighted) are presented
in Tables 1 and 2 and are labeled as such; throughout the
rest of the study and in all statistical comparative analyses,
complex survey analysis methods (SAS Proc Surveyfreq, Proc
Surveymeans, Proc Surveylogistic) were employed to estimate
standard error, variance, P values, etc .
The application of propensity score modeling (PSM) to
retrospective data can mitigate imbalances between treat-
ment groups and produce matched cohorts for comparative
analysis of complications. PSM helps minimize confound-
ing caused by the uneven distribution of patient character-
istics between treatment groups. We adapted an optimal,
1:1 PSM matching algorithm from Kleinman and Horton
18
and Kosanke and Bergstralg
19
to balance age, comorbid-
ity burden, primary diagnosis, and traumatic injury sever-
ity score, using 0.01 as the maximum allowable propensity
score distance between each case-control pair. In this way,
a new cohort was created of 1404 fusion and 1404 halo
PSM-matched patients. To assess balance, histograms of the
propensity scores by treatment group were visually assessed
and standardized differences (SD) were calculated for each
balancing variable; a value less than 0.10 was considered
well-balanced.
20
,
21
We present the patient characteristics for
the entire cohort (including unmatched cases) in Table 1 ;
throughout the rest of the study, all analyses were performed
on the matched cases with the exception of 2 separate mul-
tivariate models that were also fi tted to either all surgery
patients or all halo-vest patients.
Patient characteristics and outcomes were compared by
treatment group using the Rao-Scott χ
2 test, Fischer exact
Copyright © 2012 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
BRS204730.indd 855BRS204730.indd 855 07/04/12 1:07 PM07/04/12 1:07 PM
EPIDEMIOLOGY Outcomes for C2 Fractures • Boakye et al
856 www.spinejournal.com May 2012
test, or linear regression as appropriate. Multivariate logistic
regression modeling was performed to identify predictors of
in-hospital complications. In all analyses, a P value less than
0.01 was considered statistically signifi cant. All calculations
were performed using SAS software (version 9.2; SAS Insti-
tute, Inc., Cary, NC) running on Windows XP Pro.
RESULTS
Using the NIS, a total of 3758 patients with C2 fractures
from 2002 to 2008 were identifi ed. A total of 1627 (43.3%)
patients received fusion and 2131 (56.7%) patients received
halo-vest immobilization. For patients who underwent sur-
gery, 18.1% received posterior fusion, 10.9% received ante-
rior fusion, and 71% had “repair of vertebral fracture” or
“atlax-axis fusion” where the approach was unspecifi ed. The
mean age was higher for the fusion group at 66.9 (confi dence
interval [CI]: 65.7–68.0) than for the halo group at 56.7 (CI:
55.2–58.2; P < 0.0001; Table 1 ). In the fusion group, 50.4%
of patients were male compared with 54.7% in the halo
group ( P = 0.0069). The mean Elixhauser comorbidity score
was also higher for the fusion group at 1.84 (CI: 1.75–1.93)
than for the halo group at 1.29 (CI: 1.23–1.36; P < 0.0001).
In contrast, the injury severity score was lower for the fusion
group at 4.53 (CI: 3.28–49.1) than for the halo group at
5.54 (CI: 3.28–78.2; P < 0.0001). In addition, more patients
receiving halo treatment had private insurance (40.7%) than
patients who received fusion (29.5%), whereas more patients
in the fusion group had Medicare (57.2%) than patients in
the halo group (35.3%).
To better compare the risk factors and outcomes associated
with the fusion group versus the halo group, a propensity-
matched analysis was performed to adjust for the uneven dis-
tribution of variables. A total of 1404 patients in the fusion
group were matched with 1404 patients in the halo group.
After propensity score matching, SD were computed for
each balancing variable. All SD were less than 0.10 and are
therefore considered well balanced between the fusion group
and the halo group for the PSM-matched cohort. There were
TABLE 1. Patient Characteristics of Closed C2 Fractures Treated With Halo or Fusion (Unweighted)
Fusion Halo
PCases % Cases %
Overall 1627 100 2131 100
Age group (yr) <0.0001*
<60 486 29.9 1,090 51.1
60–69 189 11.6 254 11.9
70–79 405 24.9 378 17.7
80 547 33.6 409 19.2
Elixhauser comorbidity score <0.0001*
0 345 21.2 722 33.9
1 401 24.6 629 29.5
2 378 23.2 448 21.0
3 291 17.9 179 8.40
4 or more 212 13.0 153 7.18
Expected payer <0.0001*
Private insurance 480 29.5 868 40.7
Medicare 931 57.2 752 35.3
Other including Medicaid 216 13.3 511 24.0
Sex 0.0069*
Female 807 49.6 963 45.3
Male 819 50.4 1,161 54.7
Injury severity score 4.53 [3.28, 49.1] 5.54 [3.28, 78.2] <0.0001*
*Statistically signi cant.
An estimate of injury severity derived from the ICD - 9 –Based Injury Severity Score; ranges from 0 (no chance of death) to 100 (certain death).
Copyright © 2012 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
BRS204730.indd 856BRS204730.indd 856 07/04/12 1:07 PM07/04/12 1:07 PM
EPIDEMIOLOGY Outcomes for C2 Fractures • Boakye et al
Spine www.spinejournal.com 857
and complications in both the fusion and halo groups
( Figure 1 ). Complication rates for fusion versus halo were
11.6% and 5.78% for the age group younger than 60 years
( P = 0.002), 14.8% and 7.71% for the 60- to 69-year-old
no statistically signifi cant differences in age (SD = 0.009),
comorbidity (SD = 0.010), expected payer (SD = 0.035), sex
(SD = 0.003), and injury severity (SD = 0.032) between the
treatment groups of the matched cohort.
PSM-matched analysis showed no difference in mortal-
ity between patients who received fusion (2.75%) and those
who received halo (3.33%) ( P = 0.40; Table 2 ). However,
the rate of in-hospital complications was signifi cantly higher
for the fusion group at 20.2% compared with 10.7% for the
halo group ( P < 0.0001). In particular, complication rates for
cardiac (2.59%, 0.79%, P = 0.0005), pulmonary (14.9%,
7.78%, P < 0.0001), and wound complications (2.26%,
0.41%, P < 0.0001) were signifi cantly higher in the fusion
group than in the halo group. However, the rate of nonrou-
tine discharge was higher for the halo group (62.6%) than
for the fusion group (52.6%; P < 0.0001). In addition, the
length of stay was longer in the fusion group at 8.9 days (CI:
8.4–9.5) compared with 6.4 days (CI: 6.0–6.7) for the halo
group ( P < 0.0001), with the total charges also higher for the
fusion group at $80,000 (CI: 74–86) compared with $41,000
(CI: 38–45) for the halo group ( P < 0.0001).
Multivariate analysis on all PSM-matched patients with
C2 fractures showed that the risk factors for in-hospital com-
plications included age, pre-existing comorbidities, injury
severity, and male sex ( Table 3 ). Separate multivariate analy-
ses on patients who received fusion ( Table 4 ) or halo ( Table 5 )
also showed the same variables associated with in-hospital
complications. In addition, stratifi cation of complication rates
by the age group showed a direct correlation between age
TABLE 2. Outcomes of PSM-Matched Patients
by Treatment Group
Fusion (%) Halo (%) P
Died 2.75 3.33 0.3975
Any complications 20.2 10.7 < 0.0001*
Cardiac 2.59 0.79 0.0005*
Infection 1.50 1.19 0.4803
Neurological 0.93 0 0.0009
Pulmonary 14.9 7.78 < 0.0001*
Renal 2.51 2.08 0.4341
Venous
thromboembolism
2.10 1.14 0.0420
Wound complication 2.26 0.41 < 0.0001*
Nonroutine discharge 52.6 62.6 < 0.0001*
Length of stay (d),
mean (95% CI)
8.94
(8.4–9.5)
6.4
(6.0–6.7)
< 0.0001*
Total charges
(thousands of dollars),
mean (95% CI)
80 (74–86) 41 (38–45) < 0.0001*
*Statistically signi cant.
CI indicates con dence interval; PSM, propensity score modeling.
TABLE 3. Multivariate Logistic Regression for
Predictors of In-Hospital Complication
in a Propensity Score Matched Cohort
Age group (yr) Odds Ratio 95% CI
60–69 vs . under 60 1.220 0.814–1.828
70–79 vs . under 60 2.005 1.398–2.875*
80 or older vs . under
60
3.458 2.465–4.849*
Elixhauser Comorbidity Score
Single point increase 1.319 1.220–1.427*
Injury Severity Score†
Single point increase 1.073 1.048–1.098*
Sex
Male vs . female 1.650 1.310–2.079*
Treatment group
Fusion vs . halo 2.238 1.752–2.858*
*Statistically signi cant.
†An estimate of injury severity derived from the ICD - 9 –Based Injury Severity
Score; ranges from 0 (no chance of death) to 100 (certain death).
CI indicates con dence interval.
TABLE 4. Multivariate Logistic Regression for
Predictors of In-Hospital Complication
for All Fusion Patients
Odds Ratio 95% CI
Age group (yr)
60–69 vs . under 60 1.236 0.752–2.032
70–79 vs . under 60 2.412 1.585–3.671*
80 or older vs . under 60 3.272 2.251–4.756*
Elixhauser comorbidity score
Single point increase 1.312 1.199–1.436*
Injury severity score†
Single point increase 1.083 1.051–1.116*
Sex
Male vs . female 1.847 1.412–2.417*
*Statistically signi cant.
†An estimate of injury severity derived from the ICD-9 –Based Injury Severity
Score; ranges from 0 (no chance of death) to 100 (certain death).
CI indicates con dence interval.
Copyright © 2012 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
BRS204730.indd 857BRS204730.indd 857 07/04/12 1:07 PM07/04/12 1:07 PM
EPIDEMIOLOGY Outcomes for C2 Fractures • Boakye et al
858 www.spinejournal.com May 2012
was used primarily in patients aged 60 years or younger.
Higher comorbidities were present to a greater proportion in
the surgery cohort, whereas halo patients were younger and
healthier patients but had more severe overall injuries. This is
not surprising given the fact that effi cacy of halo placement
and halo fusion rates are much greater in younger cohorts.
22
,
23
Halo immobilization is therefore more likely used in younger
patients with severe systemic injuries who have the greatest
likelihood of benefi t from halo placement. Differences in fre-
quency of use of surgery and halo in the older adults are likely
due to feared complications of halo in this age group. In fact,
the use of halo in patients aged 80 years or older has been
described as a “death sentence.”
22
Third, the increasing risk of complication with age seems
to have an age “threshold,” which differs between halo and
operative fi xation. The risks of in-hospital complications
from halo occur primarily in octogenarians, whereas surgi-
cal risks are signifi cantly elevated in patients aged 70 years
or older with a much smaller incremental increase in compli-
cation rates going from ages 70 to 79 and to older than 80
years. Although this may be an artifact of the way in which
age groups were created, it bears further investigation.
The management of elderly fractures is fraught with dif-
culty with an array of confusing and contradictory recom-
mendations in the literature.
4
,
12
14
,
24
28
Overall, the results sup-
port recent recommendations proposed by the spine trauma
study group.
28
For stable fractures, a hard cervical collar pro-
vides the most favorable risk to benefi t ratio. For fractures
that need either surgery or halo, the group made a weak rec-
ommendation in favor of surgery.
Limitations of this study include the well-known limita-
tions of the NIS database.
29
The sensitivity and specifi city of
our identifi cation of patients with C2 fractures depend on
the accuracy of input in the NIS database. Coding practices
may vary by hospital, and factors such as reimbursement may
infl uence the coding.
30
Furthermore, because the NIS data-
base is limited to the patient’s in-hospital stay, no information
regarding long-term morbidity and mortality can be obtained.
This may particularly misrepresent the risk of halo fi xation,
because length of stay is shorter, and its complications may be
more likely to present in delayed fashion after discharge. In
addition, the complication rates reported in halo patients may
group ( P = 0.027), 24.0% and 8.93% for the 70- to 79-year-
old group ( P < 0.0001), and 29.7% and 19.6% for the 80
years or older group ( P = 0.0006).
DISCUSSION
The main fi ndings are that both halo and surgical complica-
tions were associated with signifi cantly increased in-hospital
complications in the older adults, with a much higher over-
all complication rate in elderly surgical patients. Age was the
strongest predictor of complications in both the halo and sur-
gery cohorts, independent of comorbidity and injury severity
score. Patients aged 80 years or older had a 3.5-fold increased
risk of in-hospital complications compared with patients aged
60 years or younger. Age, comorbidity, and injury severity
score had similar adverse effect (odds ratio) on in-hospital
complications in both cohorts. In addition, surgery was asso-
ciated with greater hospital length of stay and hospitalization
costs, but a higher rate of discharge to home.
The results highlight several important fi ndings. Given
that surgery was associated with greater overall complication
rates, increased length of stay, and higher charges, our data
support the usage of halo-vest immobilization. However, it
is important to note that halo treatment is associated with
early complications in patients aged 80 years or older
10
and
the risk of nonunion may be increased in the older adults.
9
,
11
Physicians should exercise best clinical judgment to determine
whether patients are appropriate operative candidates.
Second, the data show that signifi cant differences exist
in the pattern of halo or surgery use in the treatment of C2
fractures in the older adults. Surgery was used with almost
equal frequency in young patients (younger than 60 yr) com-
pared with old patients (aged 80 yr or older), whereas halo
Figure 1. Effect of age on the complication rate by treatment group.
CI indicates con dence interval; LCL, lower con dence interval; OR,
odds ratio. UCL, upper con dence interval.
TABLE 5. Multivariate Logistic Regression
for Predictors of In-Hospital
Complication for All Halo Patients
Odds Ratio 95% CI
Age group (yr)
60–69 vs . under 60 1.145 0.673–1.949
70–79 vs . under 60 1.390 0.889–2.173
80 or older vs . under 60 3.198 2.103–4.864*
Elixhauser comorbidity score
Single point increase 1.268 1.134–1.419*
Injury severity score†
Single point increase 1.061 1.035–1.088*
Sex
Male vs . female 1.554 1.147–2.105*
*Statistically signi cant.
†An estimate of injury severity derived from the ICD-9 –Based Injury Severity
Score; ranges from 0 (no chance of death) to 100 (certain death).
CI indicates con dence interval.
Copyright © 2012 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
BRS204730.indd 858BRS204730.indd 858 07/04/12 1:07 PM07/04/12 1:07 PM
EPIDEMIOLOGY Outcomes for C2 Fractures • Boakye et al
Spine www.spinejournal.com 859
underestimate the true rate, given that halo dislodgement and
pin site complications, dysphagia, or delayed complications
after discharge, because the information was not available in
the NIS database.
31
There is also a lack of detail about frac-
ture characteristics in the NIS database, and thus we were
not able to control for differences in fracture displacement
between the fusion group and the halo group. Similarly, we
were also not able to break down the subtype of C2 fractures
in our study. Although injury severity score was controlled for
between groups, it may not be an adequate proxy for fracture
displacement or type of C2 fracture.
Despite these limitations, our study leveraged the large
number of records in the NIS to provide much-needed popu-
lation-based estimates of in-hospital complications after sur-
gery or halo immobilization in elderly patients with C2 frac-
tures. We show that in-hospital complications are increased
to a similar extent by age, comorbidities, and injury sever-
ity, regardless of treatment modality. We also show that
fusion patients had a higher in-hospital complication rate,
increased length of stay, and higher charges, but similar mor-
tality rate compared with that of halo patients. We identifi ed
age ranges associated with signifi cant increases in morbidity
from both treatment groups and quantifi ed the impact of age
on in-hospital complications. Further prospective, random-
ized studies with follow-up information are needed to assess
the long-term impact of fusion and halo-vest placement in
patients with C2 fractures.
Key Points
Patients with fusion had greater complication rates
and resource utilization than halo patients but were
discharged home in a greater proportion.
There was no di erence in mortality rates between
the fusion group and the halo group.
Complication rates increased 3-fold in elderly pa-
tients aged 80 years or older, regardless of whether
they received fusion or halo.
Acknowledgment
The opinions expressed are those of the authors and are not
necessarily those of the Department of Veterans Affairs or the
U.S. Government.
References
1. Ryan MD , Henderson JJ . The epidemiology of fractures and
fracture-dislocations of the cervical spine . Injury 1992 ; 23 : 38 40 .
2. Ryan MD , Taylor TK . Odontoid fractures in the elderly . J Spinal
Disord 1993 ; 6 : 397 401 .
3. Pepin JW , Bourne RB , Hawkins RJ . Odontoid fractures, with
special reference to the elderly patient . Clin Orthop Relat Res
1985 ; 193 : 178 83 .
4. Weller SJ , Malek AM ,
Rossitch E . Cervical spine fractures in the
elderly . Surg Neurol 1997 ; 47 : 274 80 .
5. Olerud C , Andersson S , Svensson B , et al. Cervical spine fractures
in the elderly: factors infl uencing survival in 65 cases . Acta Orthop
Scand 1999 ; 70 : 509 13 .
6. Day JC. Population Projections of the United States by Age, Sex,
Race, and Hispanic Origin: 1993: 2050 . Washington, DC : Govern-
ment Printing Offi ce, US Bureau of the Census, Current Population
Reports, Series P25-1104 ; 1993 .
7. Lennarson PJ , Mostafavi H , Traynelis VS , et al. Management of
type II dens fractures: a case-control study . Spine 2000 ; 25
: 1234 7 .
8. Muller EJ , Wick M , Russe OJ , et al. Management of odontoid frac-
tures in the elderly . Eur Spine J 1999 ; 8 : 360 5
9. Frangen TM , Zilkens C , Muhr G , et al. Odontoid fractures in the
elderly: dorsal C1/C2 fusion is superior to halo-vest immoboliza-
tion . J Trauma 2007 ; 63 : 83 9 .
10. Smith HE , Kerr SM , Maltenfort M , et al. Early complications of
surgical versus conservative treatment of isolated type II odontoid
fractures in octogenarians . J Spinal Disord 2008 ; 21
: 535 9 .
11. Lind B , Nordwall A , Sihlbom H . Odontoid fractures treated with
halo-vest . Spine 1987 ; 12 : 173 7 .
12. Omeis I , Duggal N , Rubano J , et al. Surgical treatment of C2
fractures in the elderly: a multicenter retrospective analysis . J Spinal
Disord Tech 2009 ; 22 : 91 5 .
13. Fagin AM , Cipolle MD , Barraco RD , et al. Odontoid fractures in
the elderly: should we operate ? J Trauma 2010 ; 68 : 583
6 .
14. Tashjian RZ , Majercik S , Biffl WL , et al. Halo-vest immobilization
increases early morbidity and mortality in elderly odontoid frac-
tures . J Trauma 2006 ; 60 : 199 203 .
15. Elixhauser A , Steiner C , Harris DR , et al. Comorbidity measures
for use with administrative data . Med Care 1998 ; 36 : 8 27 .
16. Osler T , Rutledge R , Deis J , et al. ICISS: an international classifi ca-
tion of disease-9 based injury severity score . J Trauma 1996 ; 41 :
380 6 .
17. American College of Surgeons, Committee on Trauma . Survival
risk ratios. Available at: http://www.facs.org/trauma/ntdb/srr.zip .
Accessed September 15, 2010.
18. Kleinman K , Horton N . Example 7.35: propensity score matching .
Available at: http://sas-and-r.blogspot.com/2010/05/example-735-
propensity-score-matchingn.html . Accessed August 31, 2010.
19. Kosanke J , Bergstralg E . vmatch . Available at: http://mayoresearch.
mayo.edu/biostat/sasmacros.cfm . Accessed August 31, 2010.
20. Rosenbaum PR , Rubin DB . Constructing a control group using
multivariate matched sampling methods that incorporate the pro-
pensity score . Am Statist 1985 ; 39 : 33 9 .
21. Austin PC , Mamdani MM . A comparison of propensity-score
methods: a case study estimating the effectiveness of post-AMI
statin use . Stat Med 2006
; 25 : 2084 106 .
22. Majercik S , Tashjian RZ , Biffl WL , et al. Halo vest immobilization
in the elderly: a death sentence? J Trauma 2005 ; 59 : 350 7 .
23. Platzer P , Thalhammer G , Sarahrudi K , et al. Nonoperative man-
agement of odontoid fractures using a halothoracic vest . Neurosur-
gery 2007 ; 61 : 522 9 .
24. Bednar DA , Parikh J , Hummel J . Management of type II odon-
toid process fractures in geriatric patients: a prospective study of
sequential cohorts with attention to survivorship . J Spinal Disord
1995
; 8 : 166 9 .
25. Hanigan WC , Powell FC , Elwood PW , et al. Odontoid fractures in
elderly patients . J Neurosurg 1993 ; 78 : 32 5 .
26. Kuntz C , Mirza SK , Jarell AD , et al. Type II odontoid fractures in
the elderly: early failure of nonsurgical treatment . Neurosurg Focus
2000 ; 8 : e7 .
27. Taitsman LA , Altman DT , Hecht AC , et al. Complications of cervi-
cal halo-vest orthosis in elderly patients . Orthopedics 2008 ; 31 :
446 .
28. Harrop JS , Hart RA , Anderson PA . Optimal treatment for odon-
toid fractures in the elderly . Spine 2010 ; 35 : S219 27 .
29. Lawthers AG , McCarthy EP , Davis RB , et al. Identifi cation of in-
hospital complications from claims data. Is it valid? Med Care
2000 ; 38 : 785 95 .
30. Hsia D , Krushat WM , Fagan AB , et al. Accuracy of diagnosis cod-
ing for Medicare patients under the Prospective Payment System . N
Engl J Med 1988 ; 318 : 352
5 .
31. Hanigan WC , Powell FC , Elwood PW , et al. Odontoid fractures in
cervical orthosis: a study comparing their effectiveness in restrict-
ing cervical motion in normal subjects . J Bone Joint Surg Am
1977 ; 59 : 332 9 .
Copyright © 2012 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
BRS204730.indd 859BRS204730.indd 859 07/04/12 1:07 PM07/04/12 1:07 PM
    • "This review also could not identify the risk factors of mortality. Boakye claims that mortality is independent of the type of treatment and that age is the main risk factor [52]. For others age, a neurological deficit and concomitant injuries may be risk factors of mortality and the type of treatment alone was not the only cause of death. "
    [Show abstract] [Hide abstract] ABSTRACT: The frequency of cervical spine trauma in elderly patients is increasing with most injuries occurring in the upper cervical spine. These fractures are associated with a risk of sometimes life-threatening complications, although very few studies have specifically analyzed this. The goal of this study was to identify the incidence of complications in the literature (mortality and morbidity) following upper cervical spine trauma in elderly patients. A systematic search was performed on the MEDLINE database without limiting the search by language or date to identify all studies reporting the rate of complications after upper cervical spine trauma in patients over the age of 60. Twenty-four observational studies were included, four were comparative. These studies included a total of 857 patients, mean age 76. Nearly all traumas were odontoid process fractures, and most were treated surgically (57%). The median mortality rate was 9.2% (Q1-Q3: 2.5-19.6) and the median rate of short-term complications was 15.4% (Q1-Q3: 5.8-26.9). The main late stage complication was nonunion, which developed in a mean 10 to 12% depending on the type of treatment. Complications following cervical spine trauma are frequent in elderly patients whatever the type of treatment. Knowledge of the rate of complications in the literature and the potential risk factors is essential for the clinician to improve the information provided to patients and to prevent complications. Systematic review of the literature. Level of evidence IV.
    Full-text · Article · Aug 2013
  • [Show abstract] [Hide abstract] ABSTRACT: In small series, endoscopic third ventriculostomy (ETV) has been shown to potentially have similar efficacy as ventriculoperitoneal shunting (VPS) for idiopathic normal pressure hydrocephalus (iNPH). Therefore, some clinicians have advocated for ETV to avoid the potential long-term complications associated with VPS. Complication rates for these procedures vary widely based on limited, small series data. We utilized a nationwide database that provides a comprehensive investigation of peri-operative safety of ETV for iNPH as compared to VPS. We identified discharges with the primary diagnosis of idiopathic Normal Pressure Hydrocephalus (iNPH) (International Classification of Diseases (ICD-9) code 331.5) with ICD-9 primary procedure codes for ventriculoperitoneal shunt (02.34) and endoscopic third ventriculostomy (02.2) from 2007-2010. We analyzed short-term safety outcomes using univariate and hierarchical logistic regression analyses. There were a total of 652 discharges for ETV for iNPH and 12,845 discharges for VPS for iNPH over the study period. ETV was associated with a significantly higher mortality (3.2% vs 0.5%) and short-term complication rate (17.9% vs. 11.8%) than VPS despite similar mean modified comorbidity scores. In multivariate analysis, ETV alone predicted increased mortality and increased length of stay when adjusted for other patient and hospital factors. This is the first study that robustly assesses the peri-operative complications and safety outcomes of ETV for iNPH. Compared to VPS, ETV is associated with greater peri-operative mortality and complication rates. This consideration is important to weigh against the potential benefit of ETV: avoiding long-term shunt dependence. Prospective, randomized studies are needed.
    Article · Aug 2013
  • [Show abstract] [Hide abstract] ABSTRACT: La fréquence des traumatismes du rachis cervical chez le sujet âgé est en constante augmentation avec une prédominance des lésions du rachis cervical supérieur. Ces fractures exposent à des complications parfois mortelles mais peu d’études les analysent précisément. L’objectif de ce travail était l’identification dans la littérature de l’incidence des complications (mortalité et morbidité) après traumatisme du rachis cervical supérieur du sujet âgé.
    Article · Oct 2013
Show more