Does a Liberal Transfusion Strategy Reduce Unfavorable Neurological Outcomes? TRAIN Trial Results Analyzed in JAMA

Purpose of the TRAIN Trial Analysis

A secondary analysis published in JAMA Network Open, “evaluated the effects of different hemoglobin (Hb) thresholds for guiding red blood cell transfusion (RBCT) on the neurological outcome among patients with moderate to severe traumatic brain injury (TBI) included in the TRAIN trial. [The authors] hypothesized that, as in the main trail, patients randomized to the liberal strategy group would experience a reduction in the occurrence of unfavorable neurological outcome.”

This paper began by describing the results of the TRAIN (Transfusion Strategies in Acute Brain Injured Patients) trial as the authors noted that the findings, “demonstrated a significant reduction in the risk of unfavorable neurological outcome with a liberal strategy (Hb <9 g/dL) compared with a restrictive strategy (Hb <7 g/dL); however, because the study population included individuals with acute brain injury, the findings cannot be considered specific to TBI,” the researchers wrote. Thus, the authors of this manuscript performed a, “preplanned secondary analysis of the TRAIN trial, a prospective, multicenter, phase III, pragmatic, open-label randomized clinical trial conducted across 72 ICUs in 22 countries between September 2017 and December 2022.” They noted that the, “original study included adult patients (≥18 years of age) with acute brain injury, Glasgow Coma Scale (GCS) score of 13 or less (overall score range, 3-15, with 3 indicating deep unconsciousness or coma and 15 full alertness and orientation; higher scores indicate better neurological function; lower scores, more severe impairment), Hb levels of 9 g/dL or lower within the first 10 days of admission, and an expected ICU stay of at least 72 hours. Participants were randomized in a 1:1 ratio to receive either a restrictive transfusion strategy (RBCT triggered by Hb levels ≤7 g/dL) or a liberal transfusion strategy (RBCT triggered by Hb levels ≤9 g/dL) across a 28-day period.”

The primary outcome was described as, “the occurrence of unfavorable neurological outcome at 180 days after randomization. Neurological outcome was assessed using the Glasgow Outcome Scale Extended (GOS-E) (range, 1-8, with 1 being death and higher scores indicating better outcomes), with scores dichotomized as unfavorable for GOS-E scores ranging from 1 to 5 and favorable for GOS-E scores from 6 to 8. [Secondary outcomes] included 28-day mortality, the distribution of GOS-E scores in both the restrictive and liberal transfusion groups, ICU and hospital lengths of stay, and the incidence of organ failure, as assessed daily through the SOFA score, calculated from 6 baseline variables (range, 0-24, with higher scores indicating greater severity of disease).”

Findings

The paper explained that the, “median (IQR) Hb concentration at randomization was 8.4 (7.8-8.7) g/dL in the liberal group and 8.4 (7.8-8.8) g/dL in the restrictive group (P = .96). During the 28-day study period, daily median minimum Hb concentrations differed between the two groups (P < .001). In the liberal group, 215 patients (89.6 percent) required transfusion during the 28-day trial period, compared with 126 patients (51.2 percent) in the restrictive group (P < .001). The liberal group received a median (IQR) of 2 (1-3) units of RBCs per patient, while the restrictive group received 0 (0-1) units (P < .001).” Additionally, the authors noted that the, “primary outcome was available in 475 patients (97.7 percent), including 236 (98.3 percent) in the liberal group and 239 (97.2 percent) in the restrictive group. At 180 days after randomization, unfavorable neurological outcome occurred in 138 patients (58.5 percent) in the liberal group and in 161 patients (67.4 percent) in the restrictive group (RR, 0.86 [95 percent CI, 0.75-1.00]; P = .047; absolute risk difference, −8.77 [95 percent CI, −17.35 to −0.20] percentage points; fragility index of 1). The median (IQR) GOS-E score at 180 days was 3 (2-7) in the liberal group compared with 4 (1-6) in the restrictive group. The distribution of GOS-E scores shifted toward a larger proportion of patients distributed in higher GOS-E subscores in the liberal group compared with the restrictive group (OR, 1.38 [95 percent CI, 1.01-1.90]; P = .046).” The study found, “[n]o significant difference in 28-day mortality was observed between the liberal and restrictive groups (42 of 240 [17.5 percent] vs 51 of 244 [20.9 percent]; RR, 0.84 [95 percent CI, 0.58-1.21]; P = .34).”

Conclusion

The researchers concluded that, “[i]n this preplanned secondary analysis of patients with TBI enrolled in the TRAIN randomized clinical trial, a liberal transfusion strategy (Hb threshold of 9 g/dL) was associated with an absolute 9 percent lower risk of unfavorable neurological outcome at 180 days compared with a restrictive strategy (Hb threshold of 7 g/dL). The incidence of severe adverse events, including cerebral ischemia and infection, was comparable between the two groups. [The] results on transfusion threshold are in line with a single-center study on patients with TBI, in which a liberal strategy (Hb threshold of 9 g/dL) was associated with better neurological outcomes compared with the restrictive strategy (Hb threshold of 7 g/dL). [These] results should be interpreted with caution in light of the inherent uncertainty of the estimate. Our findings contribute to the growing body of evidence on transfusion thresholds in TBI.” Limitations of the analysis acknowledged by the authors included: “the open-label design may have introduced bias; some patients received transfusions before randomization, which could have confounded [the] results; a higher rate of unfavorable neurological outcomes [was observed] compared with prior studies, potentially reflecting [the] inclusion of only patients with anemia and [the] use of a stricter dichotomization of GOS-E categories; this secondary analysis may have been underpowered to detect differences of the primary outcome; and the decision to withdraw life-sustaining therapies was not specifically collected.”

Citation: Zernini, I.S., Faso, C., Bittencourt Rynkowski, C. et al. “Restrictive vs. Liberal Transfusion Strategy in Traumatic Brain Injury: A Secondary Analysis of the TRAIN Trial.” JAMA Network Open. 2026.