Retrospective Evaluation of
Patients with Pelvic Fractures:
Analysis of Mortality, ISS, and
Hemodynamic Parameters
Lucas Anissian,MD,PhD,FACS
James T. Nichols,MD
First Coast Orthopaedics,
Orange Park, FL
Impetus For Investigation and
Objectives




LSUHSC-S population has 100% mortality
with ISS > 54 and 86% mortality with ISS
> 43
Statistical validation and investigation of
observed trend
Evaluate probability of death given ISS
Determine other independent predictors of
mortality in our population
Introduction



Approximately 50% of
patients who die following
pelvic fracture have
associated hemorrhage
Study goal: Examine the
LSUHSC-S experience with
the pelvic fracture cohort
Effect of ISS, blood
transfusion and
hemodynamic parameters
on mortality
Introduction: ISS







Anatomical scoring system
Values from 1-75
Based on the AIS
Sum of squares for 3 most significant injured
regions
Different injury patterns can result in same score
Scores not weighted based on region injured
Extent of injuries not known until full
investigation completed: limits triage capabilities
Introduction (Cont)



ISS is correlated with
mortality (1)
Death frequently
secondary to non
pelvic injury
Early significant
transfusion volumes
often required (3)
Introduction (Cont)



The majority of transfusion in the trauma
patient is given in the first 24 hours (6)
Efficacy of massive transfusions has been
questioned (4,5)
Continued transfusion supported in the
setting of massive requirement and
associated injuries (6)
Methods



IRB approval
obtained
LSUHSC-S trauma
registry 2005-2010
Polytrauma Patients
admitted with pelvic
fractures
Methods (cont)


1.
2.
3.
4.
5.
Patients divided into two
groups based on mortality
Each group analyzed with
respect to
Age and race
Hemodynamic parameters
Transfusion requirements in
first 24h
ISS
GCS
Results: Descriptive
Characteristic / Outcome
Number (%) or
Mean±SD, Median, Range
Male sex
Race: White
Black
Hispanic
Other (Asian, Arabic)
Died
Age at admission (years)
Length of hospital stay (days)
GCS
ISS
Total Units of blood
Respiratory rate
Pulse rate
SBP
DBP
MAP
624 (65.1)
635 (66.2)
300 (31.3)
18 (1.9)
6 (0.6)
55 (5.7)
37.1±18.4, 0 - 104
16.1±22.4, 0 – 366
12.8±4.4, 3 - 15
17.8±11.1, 4 - 75
0.23±1.10, 0 - 14.0
20.2±5.4, 3 - 48
99.4±22.7, 18 - 190
128.9±25.3, 47 - 217
75.8±17.7, 16 - 177
102.3±19.5, 32.5 - 175.5
Univariate Analyses
*
Variable
Died (N=55)
Did not die (N=904)
p-value
White race
55 (81.8%)
590 (65.3%)
0.01
Hospital stay (days)
5.4±12.9
16.8±22.7 (N=903)
<0.01
Age (years)
48.2±23.8
36.4±17.9
<0.01
GCS
6.8±22.7
13.1±4.0 (N=902)
<0.01
ISS
40.8 ±1
16.4±8.8
<0.01
SBP
112.9±39.6 (N=45)
129.8±24.1 (N=890)
<0.01
DBP
59.0±22.3 (N=44)
76.6±17.0 (N=889)
<0.01
MAP
85.2±29.2 (N=44)
103.1±18.5 (N=890)
<0.01
Total units of blood
1.53±2.84
*Factors Significantly Associated with Mortality
0.15±0.83
<0.01
Distribution of ISS
1.2
1.0
.8
.6
.4
.2
0.0
-.2
0
ISS
20
40
60
80
ISS
80
60
610
641
256
142
404
98
456
74
760
382
403
119
48
800
528
440
901
668
40
20
353
ISS
0
-20
N=
904
55
Lived
Died
mortality 0=Lived, 1=Died
Multiple Logistic Regression
Analysis



Used to determine independent and significant
risk factors for mortality
A patient would be predicted to die if probability
of death given values for the independent risk
factors was greater than observed mortality rate
ISS threshold for predicting death was
determined by equating observed mortality with
estimated probability of death given ISS
Multivariate Analysis
Factor
GCS
ISS
Age at admission
Odds Ratio (OR)
0.88
1.15
1.05
95% CI for OR
0.82 to 0.95
1.11 to 1.19
1.03 to 1.06
p-value
<0.01
<0.01
<0.01
Adjusting for the effects of other significant factors, odds for death:
1. Increases by 12% for each unit decrease in GCS
2. Increases by 15% for each unit increase in ISS
3. Increases by 5% for each year increase in age
Odds for death increases by 15% for each unit increase in ISS
Probability of Death


Patient will be predicted to die if
probability of death using multiple logistic
regression model is greater than 0.057
(observed mortality)
Estimated probability of death is 5.6% at
ISS of 26
Observed and Predicted Mortality
Using Independent Variables *
Predicted to
Die
Live
Total
Died
46
9 (16.4%)
55
Lived
Total
111 (12.3%)
157
791
800
902
957
-Among all patients, total error rate = (9+111)/957 = 12.5%
-Among the 55 patients who died, error rate = 9/55 = 16.4% (wrongly predicted to live)
-Among the 902 who lived, error rate = 111/902 = 12.3% (wrongly predicted to die)
*ISS, GCS, Age
Observed and Predicted Mortality
Using ISS And Total Units
Transfused
Predicted to
Die
Live
Total
Died
47
8 (14.6%)
55
Lived
Total
100 (11.1%)
147
804
812
904
959
-Among all 959 patients, total prediction error rate = 11.3%
-Among the 55 patients who died, 8 were predicterd to live; error rate 14.6%
-Among the 904 patients who lived, 100 were predicted to die; error rate 11.1%
Conclusion

1.
2.

Two sets of independent risk factors for
death identified:
GCS, ISS, age at admission
ISS and total units blood transfused in
first 24 hours
Using the second group (ISS and TBU)
to predict death results in lowest
prediction errors (11.1%)
Limitations




Population included acetabular fractures and
isolated sacral fractures
Significant potential for recording bias given
retrospective analysis
Any error in AIS scoring increases ISS error
significantly
Full description of patient injuries not known
until full investigation completed; limits triage
capabilities
References
1.
2.
3.
4.
5.
6.
Balogh J, Varga E, Tomka J, Suveges G, Toth L, Simonka J. The New Injury
Severity Score is a Better Predictor of Extended Hospitalization and Intensive Care
Unit Admission Than the Injury Severity Score in Patients With Multiple
Orthopaedic Injuries. Journal of Orthopaedic Trauma. 17: 508-512, August 2003.
Demetriades D, Karaiskakis M, Toutouzas K, Alo K, Velmahos G, Chan L. Pelvic
Fractures: Epidemiology and Predictors of Associated Abdominal Injuries and
Outcomes. Journal of the American College of Surgeons. 195: 1-10, July 2002.
Tachibana T, Yokoi H, Kirita M, Marukawa S, Yoshiya S. Instability of the Pelvic
Ring and Injury Severity can be Predictors of Death in Patients with Pelvic Ring
Fractures: A Retrospective Study. Journal of Orthopaedic Traumatology. 10(2):
79-82, April 2009.
Velmahos G, Chan L, Chan M, Tatevossian R, Cornwell E, Asensio J, Berne T,
Demetriades D. Is There a Limit to Massive Blood Transfusion After Severe
Trauma? Archives of Surgery. 133: 947-952, September 1998.
Vaslef S, Dnudsen N, Neligan P, Sebastian M. Massive Transfusion Exceeding 50
Units of Blood Products in Trauma Patients. Journal of Trauma-Injury Infection &
Critical Care. 53(2): 291-296, August 2002.
Como J, Dutton R, Scalea T, Edelman B, Hess J. Blood Transfusion Rates in the
Care of Acute Trauma. Transfusion. 44:809-813, June 2004.
Thank You