The Effect of Tranexamic Acid on Functional Outcomes: An Exploratory Analysis of the CRASH-2 Randomized Controlled Trial

Study objective: Tranexamic acid improves survival in severely injured adults. However, its effectiveness on overall functional outcome is unknown. We hypothesized that tranexamic acid improves overall functional outcome compared with placebo in severely injured adults and conduct an exploratory analysis of the Clinical Randomization of an Antifibrinolytic in Significant Haemorrhage (CRASH-2) data to investigate this hypothesis.

Methods: We included injured adults from the CRASH-2 trial who were randomized 3 hours or less from injury. The primary outcome measure was functional status at hospital discharge or on day 28 if the subject was still in the hospital. Functional status was measured with the modified Oxford Handicap Scale, a 6-category ordinal functional outcome scale. We conducted 3 separate analyses using 3 different outcome measures to evaluate the effectiveness of tranexamic acid versus placebo on functional outcomes, including the mean utility-weighted modified Oxford Handicap Scale score (overall functional outcome), the area under the curve (based on functional outcome and rate of recovery), and a sliding dichotomy analysis (favorable versus unfavorable functional outcome) stratified by baseline mortality risk (stratified analysis).

Results: There were 13,432 patients (6,679 randomized to placebo and 6,753 to tranexamic acid) included in the study cohort. The mean utility-weighted modified Oxford Handicap Scale score was 0.66 (SD 0.33) for patients randomized to tranexamic acid compared with a mean of 0.64 (SD 0.34) for those randomized to placebo (mean difference 0.02; 95% confidence interval [CI] 0.01 to 0.03). The area under the curve analysis demonstrated that patients randomized to tranexamic acid had a higher 28-day mean utility-weighted modified Oxford Handicap Scale score compared with those randomized to placebo (mean score 0.55 [SD 0.30] versus 0.53 [SD 0.31]; mean difference 0.02 [95% CI 0.01 to 0.03]). The sliding dichotomy analysis demonstrated heterogeneity of treatment effects across risk groups. The overall proportion of patients with favorable functional outcomes was higher in the tranexamic acid group (5,360/6,753 [79.4%]; 95% CI 78.4% to 80.3%) compared with the placebo group (5,174/6,679 [77.5%]; 95% CI 76.5% to 78.5%; difference 1.9% [95% CI 0.5% to 3.3%]; number needed to treat=52). When each risk group was tested separately, only the lowest-risk group (<6% baseline mortality risk) demonstrated a statistically significant effect of tranexamic acid toward favorable functional outcomes (tranexamic acid versus placebo adjusted odds ratio 0.78; 95% CI 0.67 to 0.90). There were no differences between tranexamic acid and placebo in the other risk groups.

Conclusion: Across 3 exploratory analyses, severely injured adult patients randomized within 3 hours from injury demonstrated better functional outcomes with tranexamic acid compared with placebo. When heterogeneity of treatment effects across risk groups was evaluated, only the lowest-risk group demonstrated a significant effect of tranexamic acid toward favorable outcomes. Given the overall safety and cost-effectiveness of tranexamic acid use in injured adults, our results further support the use of tranexamic acid for this population. Future trauma trials that evaluate tranexamic acid use should also consider functional status as an important outcome.