Page 29 - JSOM Winter 2025

Methods TABLE 1 Study Variables Included in the Analysis with Description
and Abbreviations
A retrospective chart review was conducted with the ap-
proval of the University of Alberta Research Ethics Office Variable Abbreviation
(Pro00088974). STROBE reporting guidelines were followed. Primary outcome
7
The study used Shock Trauma Air Rescue Society (STARS) First measured fibrinogen level in the ED*
Electronic Patient Care Reports (ePCR), and records from the Secondary outcomes
University of Alberta Hospital (UAH), the Royal Alexandra First INR measured in the ED † INR-ED
Hospital (RAH) (level one and two trauma centers respectively) Massive hemorrhage protocol activation in the ED MHP-ED
in Edmonton, Alberta, and data from the Alberta Trauma Reg- Need for surgery and/or angiography Surg/angio
istry. The study period was January 2015 to August 2019. We
enrolled consecutive patients of any age over the weight of Discharge status (alive or dead) Discharge
50kg with blunt or penetrating trauma and transported by Mortality within 14 days of injury from Mortality14
hemorrhage/multiorgan failure
STARS who received at least one unit of packed red blood cells
(pRBC) by STARS. In total, we collected data from 65 trauma Primary predictors
patients (supplemental material Figure S1). SBP <90 with EMS N/A
SBP <90 after >250mL IV crystalloid administration N/A
The primary outcome variable was the first ED fibrinogen SBP <90 after first unit of HEMS pRBC N/A
level, transformed into a dichotomous variable, in which val- Secondary predictors
ues below 1.6g/L were classified as the low fibrinogen group, ED arrival SBP <90 EDSBP<90
and those at or above 1.6g/L were classified as the high fi- ED arrival SBP EDSBP ‡
brinogen group. This value was selected because provincial ED arrival Shock Index EDSI
massive hemorrhage protocols guide fibrinogen replacement Time of injury to arrival in the ED ITED
below this level. Secondary outcome variables and predictor
variables are specified in Appendix 1. Injury type (blunt or penetrating) InjType
Age, y Age
The analysis focused on the effects of prediction of first ED *Transformed into a categorical variable with the following levels (<1;
fibrinogen levels, 14-day mortality, and prediction of second- 1–1.49; 1.5–1.99; >2).
†
ary variables. Direct multivariable logistic regression was used ‡ Converted to dichotomous variable.
Transformed into categorical variable with the following levels (0/
for all analyses except for 14-day mortality (where multino- PEA–59; 60–74; 75–90; >90).
mial logistic regression was used). All analyses were conducted ED = emergency department; HEMS = helicopter EMS; INR = in-
using jamovi, a graphical user interface for R (Supplemental ternational normalized ratio; pRBC = packed red blood cells; SBP =
Material, Detailed Methods and Results). systolic blood pressure; N/A = not applicable.
Patients with a fibrinogen level <1.6g/L had a 21.6 (2.4–193.9)
Results
times greater odds of mortality from hemorrhage or multi-
See Table 1 for patient characteristics for persistent hypoten- organ failure, P=.006. There was no association with death
sion post pRBC transfusion. Patients meeting any SBP <90 post from other causes, P=.149, OR 3.1 (0.7–14.3).
fluid administration (CRC) had an average fibrinogen level of
1.6 (SD 0.9), 1.5g/L, while patients that did not meet CRC Secondary Analysis
(had SBP >90) had an average fibrinogen level of 2.41 (SD CRC emerged as a significant predictor of in-hospital death,
1.0), 2.2g/L. See Table 2 for patient characteristics grouped by P=.04, with those meeting code red criteria having 8.3 (1.2–
ongoing hypotension post pRBC transfusion. 59.0) times greater odds of death.

Primary Analysis No significant predictors emerged for INR-ED >1.5. When
For the first fibrinogen level only, CRCpost emerged as a sig- INR-ED was treated as a continuous variable, CRCpost was a
nificant predictor of low fibrinogen levels (P=.04). Patients significant predictor of INR-ED, P=.012, standardized β=0.71
with an SBP <90 after pRBC infusion had 6.6 (1.1–40.2) times (0.16–1.25). The INR level for SBP >90 after the first unit of
greater odds of fibrinogen <1.6g/L. Fibrinogen level for the pRBC (CRCpost NO code red remains) was mean 1.2 (SD
cases with SBP >90 after first unit of pRBC (CRCpost NO 0.2), 1.1g/L, while for SBP <90 after the first unit of pRBC
code red remains) was mean 2.4 (SD 0.9), median 2.1g/L. (CRCpost YES code red remains) was mean 1.6 (SD 0.6),
For SBP <90 after first unit of pRBC (CRCpost YES code red 1.4g/L.
remains), mean fibrinogen was 1.5 (SD 0.8), median 1.3g/L.
There was no association of injury type (blunt vs penetrating) For MHP-ED and Surgery/Angiography, no primary predic-
with fibrinogen level, P=.45. tors were significant.

Outcomes for CRCpost YES were alive: 20/34 (58.8%); mor- Discussion
tality within 14 days of injury from hemorrhage/multiorgan
failure (Mortality14): 7/34 (20.5%); other mortality (Mortal- In our cohort of consecutive trauma patients transported by
ityOther): 7/34 (20.5%). For CRCpost NO, outcomes were: HEMS and transfused at least one unit of pRBC, we observed
alive: 29/31 (93.5%); Mortality14: 1/31 (3%); Mortality- a significant association between patients with persistent hy-
Other: 1/31 (3%). The odds of being discharged from hospital potension meeting CRC (SBP<90) after transfusion and low
alive were 6.8 (2.0–23.5) times higher with a fibrinogen level fibrinogen levels (<1.6g/L). These patients were 7.4 times as
≥1.6g/L, P=.003. likely to have a first fibrinogen level below 1.6g/L compared to

HEMS Predictors of Low Fibrinogen | 27