Early Prehospital Tranexamic Acid Following Injury Is Associated With a 30-day Survival Benefit

A Secondary Analysis of a Randomized Clinical Trial

Shimena R. Li, MD; Francis Guyette, MD, MPH; Joshua Brown, MD, MSc; Mazen Zenati, MD, MPH, PhD; Katherine M. Reitz, MD, MSc; Brian Eastridge, MD; Raminder Nirula, MD, MPH; Gary A. Vercruysse, MD; Terence O'Keeffe, MD; Bellal Joseph, MD; Matthew D. Neal, MD; Brian S. Zuckerbraun, MD; Jason L. Sperry, MD, MPH


Annals of Surgery. 2021;274(3):419-426. 

In This Article


For the current secondary analysis, 476 patients met all inclusion and no exclusion criteria (EARLY, N = 238; DELAYED, N = 238). (Figure 1). Overall, patients were moderately injured with median Injury Severity Score (ISS) of 14 (IQR, 6–37), and overall mortality of 11.5%. The majority of injuries were blunt (83.6%) and were most commonly due to motor vehicle collision (57.9%). Penetrating mechanism comprised 16.4% of injuries and were largely due to firearms (47.4%) or stabbings (42.3%). Randomization of the primary STAAMP trial was excellent, with no significant differences in baseline patient characteristics between treatment arms.[18]

Figure 1.

Patient selection flowchart.

Overall, EARLY and DELAYED patients had similar demographics, injury characteristics, and shock severity. DELAYED patients had greater resuscitation requirements including higher prehospital crystalloid volume and more frequent prehospital packed red blood cell (PRBC) transfusion. Additionally, DELAYED patients were more likely to be transferred from a referral hospital and were more frequently intubated before arrival to a participating trauma center (Table 1). Compliance with prehospital TXA dosing was high with 95.8% of EARLY prehospital TXA and 97.6% of DELAYED prehospital TXA patients receiving the full prehospital dose.

We first evaluated the association of prehospital TXA treatment and timing of administration on 30-day mortality. The prehospital TXA and time of administration interaction was assessed utilizing cox regression clustered by trial site which was statistically significant (P = 0.03). EARLY prehospital TXA was associated with significantly reduced 30-day mortality compared to placebo (3.5% vs 10.9%; difference −7.5; 95% CI, −14.0% to −0.9%; P = 0.03) with no significant difference in the DELAYED group (15.2% vs 16.4%; difference −1.2; 95% CI −10.5% to 8.5%; P = 0.81). Next, we performed 24-hour and 30-day Kaplan-Meier survival analysis comparing prehospital TXA versus placebo across EARLY and DELAYED subgroups (Figure 2). There was no statistically significant difference in 24-hour mortality across treatment arms in both EARLY (log-rank chi-square test, 2.8; P = 0.09) and DELAYED groups (log-rank chi-square test, 0.01; P = 0.94). However, patients that received EARLY prehospital TXA demonstrated significant separation in 30-day mortality (log-rank chi-square test, 4.99; P = 0.03) with no significant separation found for DELAYED patients (log-rank chi-square test, 0.04; P = 0.83).

Figure 2.

Unadjusted Kaplan-Meier survival analysis for 24-hour and 30-day survival comparing prehospital TXA and placebo patients for EARLY (A, C) and DELAYED (B, D) Subgroups. P values were calculated using log-rank testing.

Multivariate analysis of 30-day survival using a Cox proportional-hazard model to adjust for clinically relevant and significant covariates verified that EARLY prehospital TXA was independently associated with a survival benefit (HR 0.35, 95% CI 0.19–0.65, P = 0.001) when compared to placebo, representing a 65% lower hazard for 30-day mortality (Table 2). No independent survival benefit was found in DELAYED prehospital TXA patients (HR 1.00, 95% CI 0.63–1.60, P = 0.99).

We then compared incidence of MOF, NI, and ALI across treatment arms in both EARLY and DELAYED groups. EARLY prehospital TXA was associated with decreased incidence of MOF compared to placebo (P = 0.02) (Table 3). No significant difference in MOF incidence was observed between treatment arms in the DELAYED group (P = 0.51). Incidence of NI and ALI were not significantly different across randomization groups in either EARLY or DELAYED subgroups.

We also assessed 6-hour and 24-hour blood and blood component transfusion requirements across treatment arms in EARLY and DELAYED subgroups. EARLY prehospital TXA was associated with a lower incidence of requiring any blood transfusion in the first 6 hours (31.4% vs 46.7%, P = 0.02) and decreased volume of 6-hour PRBC, platelet, and plasma transfusion when compared to placebo (Table 4). EARLY prehospital TXA was also associated with a decreased 24-hour volume of PRBC, platelet, and plasma transfusion. Importantly, EARLY prehospital TXA was associated with reduced incidence of the need for massive transfusion (≥10 U PRBC in first 24 hours; 2.5% vs 9.2%, P = 0.03) without significant difference between treatment arms in the DELAYED group (7.9% TXA vs 10.7% placebo, P = 0.46). No significant differences were observed between DELAYED prehospital TXA and placebo for 6- and 24-hour transfusion.

We then examined measures of coagulopathy including presenting INR and TEG parameters on arrival to the definitive trauma center across treatment arms in EARLY and DELAYED subgroups. There was no significant difference in INR between prehospital TXA and placebo in either EARLY or DELAYED subgroups. TEG comparison demonstrated that EARLY prehospital TXA was associated with a decreased α angle compared to placebo (P = 0.04) (Table 4). In the DELAYED subgroup, prehospital TXA was associated with a significantly decreased lysis at 30 minutes (LY30) compared to placebo (P < 0.01).

Lastly, we evaluated measures of safety for EARLY and DELAYED prehospital TXA by comparing rates of venous thromboembolism (VTE) and seizures across treatment arms. There were no significant differences in DVT, PE, overall VTE or seizures between prehospital TXA and placebo across both EARLY and DELAYED subgroups (Table 5).