Table 4.
Parameter | Univariate Odds Ratio (95% CI) |
P-value | Multivariate Odds Ratio (95% CI) |
P-value |
---|---|---|---|---|
Operational theater | ||||
Afghanistan | Reference | Reference | ||
Iraq | 0.56 (0.42–0.75) | <0.0001 | 1.32 (0.87–1.99) | 0.193 |
Combat-related injury | 7.13 (5.10–9.99) | <0.0001 | 0.60 (0.32–1.15) | 0.124 |
Branch of Service1 | ||||
Army | Reference | – | ||
Marine | 1.66 (1.41–1.95) | <0.0001 | – | |
Other | 0.83 (0.62–1.12) | 0.220 | – | |
Mechanism of Injury | ||||
Gunshot wound | Reference | Reference | ||
IED blast | 3.26 (2.60–4.08) | <0.0001 | 1.84 (1.35–2.49) | <0.0001 |
Non-IED blast | 0.99 (0.72–1.37) | 0.937 | 0.91 (0.60–1.37) | 0.639 |
Other | 0.45 (0.33–0.62) | <0.0001 | 1.49 (0.84–2.66) | 0.174 |
Injury Severity Score | ||||
≤15 | Reference | Reference | ||
16–25 | 7.7 (5.9–10.0) | <0.0001 | 1.72 (1.23–2.42) | 0.002 |
≥ 26 | 28.5 (23.0–35.4) | <0.0001 | 2.48 (1.81–3.41) | <0.0001 |
RBC transfusion requirements | ||||
0/missing units2 | Reference | Reference | ||
1–9 units | 9.19 (7.45–11.33) | <0.0001 | 2.50 (1.89–3.30) | <0.0001 |
10 – 20 units | 38.69 (29.57–50.64) | <0.0001 | 5.66 (3.97–8.08) | <0.0001 |
> 20 units | 83.58 (59.50–117.41) | <0.0001 | 10.68 (6.73–16.94) | <0.0001 |
Injury Type | ||||
Closed | Reference | Reference | ||
Open | 6.84 (5.21–9.00) | <0.0001 | 1.37 (0.93–2.03) | 0.114 |
MDR Gram-negative Colonization at LRMC admission | 2.72 (2.15–3.44) | <0.0001 | 1.39 (0.97–1.99) | 0.075 |
Use of prophylactic antibiotics within 1st 48 hours | 6.43 (5.16–8.01) | <0.0001 | 1.42 (1.03–1.97) | 0.033 |
ICU Admission | ||||
Non-ICU | Reference | Reference | ||
LRMC only | 4.65 (3.49–6.19) | <0.0001 | 1.98 (1.41–2.76) | <0.0001 |
U.S. MTFs ± LRMC | 13.59 (10.82–17.06) | <0.0001 | 3.80 (2.85–5.05) | <0.0001 |
CI – Confidence Interval; ICU – intensive care unit; IED – improvised explosive device; LRMC – Landstuhl Regional Medical Center; MDR – multidrug-resistant; MTFs – military treatment facilities; RBC – packed red blood cells plus whole blood
Due to stepwise selection, the branch of service parameters was not included in the multivariate model.
Missing RBC transfusion data are not randomly distributed. Patients with missing RBC data are characterized by lower injury severity scores and shock indices. In addition, the majority of patients with missing RBC data did not sustain a traumatic amputation and were not admitted to the LRMC ICU.