1. Publication
- Title: Prehospital Resuscitation with Type O Whole Blood for Trauma and Hemorrhage
- Acronym: TOWAR (Type O Whole blood and Assessment of age during prehospital Resuscitation)
- Year & Journal: New England Journal of Medicine, published online May 18, 2026
- Citation: Sperry JL, Guyette FX, Cotton BA, et al; TOWAR Study Group. Prehospital Resuscitation with Type O Whole Blood for Trauma and Hemorrhage. N Engl J Med. 2026. doi:10.1056/NEJMoa2602167
2. Context & Rationale
Background: Major traumatic hemorrhage remains a leading cause of potentially preventable early trauma death. Modern resuscitation has moved from crystalloid-heavy care toward earlier hemostatic resuscitation with blood products. Prehospital transfusion has biologic and temporal plausibility: it occurs before definitive hemorrhage control, before trauma-induced coagulopathy is fully established, and before shock becomes irreversible. Early observational studies suggested whole blood was safe and had better outcomes than components, but no large randomized US trial had directly tested this.
Research Question/Hypothesis: In prehospital trauma patients with hemorrhagic shock, does low-titer group O whole blood (LTOWB) improve 30-day mortality compared with blood components? A secondary question examined whether whole blood storage age (15-21 vs 1-14 days) affected outcomes.
Why This Matters: Prehospital whole blood requires substantial blood-bank infrastructure, donor screening, cold-chain management, and governance; a demonstrated mortality benefit would justify widespread civilian air-medical and military far-forward adoption of this resource-intensive approach.
3. Design & Methods
- Study Type: Pragmatic, multicenter, cluster-randomized, open-label, phase 3 trial
- Setting & Centers: 44 air medical bases feeding 10 (per one source) trauma centers, USA; conducted May 2022 – June 2025
- Population:
- Inclusion: Patients transported from scene of injury or referring ED with ≥1 episode of hypotension (SBP <91) plus tachycardia (HR >107), or any SBP <71
- Exclusions: Not fully detailed in available trial text
- Intervention: Up to 2 units of cold-stored low-titer group O whole blood
- Comparator: Blood components (as indicated)
- Randomization: Cluster-randomized by air medical base (2:1 allocation ratio); 1020 eligible patients (715 whole blood, 305 components)
- Blinding: Open-label
- Statistical Power & Follow-Up: 1020 patients from 40 sites required to detect a 10-percentage-point absolute reduction in 30-day mortality (26%→16%), 80% power, two-sided α=0.05, ICC=0.02. Primary outcome: 30-day all-cause mortality.
4. Key Results
1020 eligible; 695 (whole blood) and 298 (components) in primary analysis.
Outcome | Whole Blood | Components | Effect Size | 95% CI | p-value | Clinical Notes |
30-day mortality (primary) | 25.9% | 20.5% | Adjusted OR 1.24 | 0.87–1.76 | 0.24 | No significant difference; numerically higher with whole blood |
Storage-age substudy: 15-21d vs 1-14d whole blood | 27.1% (n=210) | 26.4% (n=443) | Adjusted OR 0.99 | 0.74–1.32 | — | No association between blood age and mortality |
Subgroup: severe prehospital hypotension | 34.5% | 24.1% | OR 1.59 | 1.02–2.47 | — | Only subgroup showing a significant difference — numerically worse with whole blood |
Secondary outcomes (3h, 6h, 24h mortality; in-hospital mortality; units transfused; MOF; nosocomial infection; ARDS) | Comparable | Comparable | Not reported | Not reported | — | All secondary outcomes similar between groups |
Crossover: ~30% of whole-blood-arm patients received components; ~10% of component-arm patients received whole blood — a substantial contamination given the cluster (not individual-patient) randomization design.
5. Internal Validity Assessment
- Randomization & Allocation: Cluster-randomized by air medical base (not individual patient) — introduces baseline imbalance risk and reduces effective sample size relative to individual randomization (ICC=0.02 accounted for in power calculation).
- Protocol Adherence & Separation: Substantial crossover (30% whole-blood→components; 10% components→whole-blood) meaningfully attenuates the intervention contrast.
- Blinding & Detection Bias: Open-label; mortality is a hard, objective endpoint, limiting detection bias despite lack of blinding.
- Missing Data & Sensitivity Analyses: 695/715 (97.2%) and 298/305 (97.7%) included in primary analysis — high follow-up completeness.
- Overall Internal Validity Conclusion: Moderate-to-strong per independent CCR assessment — randomization, objective outcome, high follow-up, and prespecified analysis support credibility, but cluster design, open-label care, substantial crossover, pre-randomization transfusion variability, and baseline imbalance (whole-blood arm appeared somewhat sicker per independent commentary) limit confidence in excluding smaller or phenotype-specific effects.
6. External Validity Assessment
- Population Representativeness: Severely injured patients selected for prehospital transfusion in mature US air-medical trauma systems; median Injury Severity Score 25, most injuries blunt, most transported from scene. Comparison with a national trauma registry suggested TOWAR patients were younger and more often male than the overall trauma population, though closer to the subset meeting TOWAR's physiologic criteria.
- Practice Context: Requires blood-bank infrastructure, donor screening, cold-chain management, and governance for whole-blood programs — not universally available, particularly outside mature US/European air-medical systems.
- Overall External Validity Conclusion: Moderate-to-good for mature air-medical trauma systems with existing or planned whole-blood capability; less directly informative for systems without such infrastructure given the resource question is somewhat moot without it.
7. Strengths & Limitations
Strengths:
- Among the largest prehospital blood-transfusion trials ever conducted in the US
- High primary-outcome follow-up completeness (>97%)
- Directly addresses both the whole-blood-vs-components question and the storage-age question in one trial
- Pragmatic, real-world air-medical delivery model
Limitations:
- Cluster (not individual-patient) randomization introduces baseline imbalance risk
- Substantial crossover (30% vs 10%) attenuates intervention contrast
- Open-label design
- In-hospital blood product administration was not part of the trial protocol, limiting the ability to assess whole-course resuscitation strategy
- Enrolled fewer than 1000 patients per arm combined; wide confidence intervals (0.87-1.76) span both potential benefit and harm
8. Interpretation & Practice Impact
- Clinical Implications: Does not support routine, mandatory adoption of prehospital whole blood over components for hemorrhagic shock — both approaches appear similarly effective, with 30-day mortality rates (20.5-25.9%) both comparing favorably to the ~one-third mortality estimated among severely injured patients receiving no prehospital blood transfusion at all.
- Mechanistic Coherence: The lack of a storage-age effect (1-14 vs 15-21 days) is reassuring for whole-blood shelf-life logistics, supporting use throughout the product's full approved shelf life.
- Systems-Level Takeaway: Per the investigators' own framing, results support flexibility — emergency teams can use whichever blood product (whole blood or components) is most readily accessible in their setting, rather than one being clearly superior.
9. Controversies & Subsequent Evidence
- Editorial Commentary: An accompanying NEJM editorial (Rowell, Rogers) frames TOWAR and the UK's SWiFT trial together as providing "consistent evidence from two randomized trials" that prehospital whole blood does not clearly outperform components — a notable convergence of independent trials in different health systems.
- Ongoing Debate: Independent commentary (First10EM, tactical medicine sources) cautions against the conclusion "whole blood doesn't work," noting the trial enrolled fewer than 1000 total patients, wide CIs spanning meaningful benefit and harm, and substantial crossover — arguing the trial is underpowered to fully exclude a true effect, rather than definitively disproving one.
- Guideline Integration: Considered alongside SWiFT as complementary, concordant negative trials; supports guideline language permitting either whole blood or components based on local availability rather than mandating one.
10. Summary & Executive Takeaway
Summary: TOWAR cluster-randomized 1020 severely injured US trauma patients (by air medical base) to prehospital whole blood or blood components. 30-day mortality was 25.9% vs 20.5% (adjusted OR 1.24, 95% CI 0.87-1.76, P=0.24) — not significantly different, though numerically favoring components. Blood storage age (1-14 vs 15-21 days) showed no association with outcomes.
Overall Takeaway: Prehospital whole blood and blood components appear similarly effective for hemorrhagic shock resuscitation in mature air-medical trauma systems — both far better than no prehospital transfusion — supporting flexible use of whichever product is most readily available rather than mandating whole blood adoption, though wide confidence intervals and substantial crossover mean this should not be over-read as definitively ruling out a true difference.
11. Bibliography
- Smith JE, Cardigan R, Sanderson E, et al; SWiFT Trial Group. Prehospital Whole Blood in Traumatic Hemorrhage — a Randomized Controlled Trial. N Engl J Med. 2026. doi:10.1056/NEJMoa2516043
- Rowell SE, Rogers SO. Prehospital Whole Blood for Traumatic Hemorrhage — Consistent Evidence from Two Randomized Trials [editorial]. N Engl J Med. 2026;394(23):2372-2373.
- Guyette FX, Zenati M, Triulzi DJ, et al. Prehospital low titer group O whole blood is feasible and safe: pilot trial. J Trauma Acute Care Surg. 2022;92(5):839-847.