Therapeutic-dose heparin reduces 28-day mortality versus prophylactic dosing in hospitalized COVID-19 (mainly non-critically ill) but increases major bleeding; intermediate dosing offers no mortality benefit.

Background

Thrombosis and systemic inflammation are common in hospitalized COVID-19. Trials comparing higher- versus lower-dose anticoagulation have yielded inconsistent results. The WHO REACT Working Group conducted a prospective meta-analysis of randomized trials to clarify effects on mortality, respiratory progression, thromboembolism, and major bleeding.

Patients

• 22 randomized trials; 11,733 hospitalized patients with suspected or confirmed COVID-19 across 21 countries.
• Predominantly heparin-based regimens (low–molecular-weight heparins or unfractionated heparin); one trial of rivaroxaban.
• Severity at randomization varied; most in the therapeutic vs prophylactic comparison required no or low-flow oxygen.

Intervention

• Higher-dose anticoagulation with heparins:
  • Therapeutic-dose regimens.
  • Intermediate-dose regimens (for intermediate vs prophylactic and therapeutic vs intermediate comparisons).

Control

• Lower-dose anticoagulation:
  • Prophylactic-dose heparin (primary comparator).
  • Intermediate-dose heparin (for therapeutic vs intermediate comparisons).

Outcome

• Primary: 28-day all-cause mortality.
• Secondary:
  • Progression to invasive mechanical ventilation (IMV) or death (among those not ventilated at randomization).
  • Arterial or venous thromboembolic events by 28 days.
  • Major bleeding by 28 days (ISTH criteria).

Study Design

Systematic review and prospectively planned meta-analysis of randomized trials (fixed-effects, inverse-variance methods). Risk of bias (RoB 2) generally low to some concerns; GRADE certainty high for several key outcomes.

Level of Evidence

• Level I: Systematic review/meta-analysis of randomized controlled trials.
• GRADE: High certainty for therapeutic vs prophylactic mortality; moderate to high for other outcomes (some imprecision and outcome-assessment subjectivity).

Follow up period

28 days after randomization.

Results

Primary outcome: 28-day mortality

• Therapeutic vs prophylactic heparin: OR 0.77 (95% CI, 0.64–0.93); events 252/3189 vs 304/3108 (7.9% vs 9.8%).
  • Absolute risk reduction (ARR): ~1.9%.
  • NNT: ~54 to prevent one death.
• Intermediate vs prophylactic: OR 0.95 (95% CI, 0.76–1.19); events 205/1939 vs 215/1958 (10.6% vs 11.0%).
  • ARR: ~0.4% (no meaningful difference).
  • NNT: ~244 (not clinically meaningful given imprecision).
• Therapeutic vs intermediate: OR 1.21 (95% CI, 0.93–1.58); events 204/888 vs 194/915 (23.0% vs 21.2%).
  • Absolute risk increase (ARI): ~1.8%.
  • NNH: ~56 (estimate imprecise).

Subgroups: Effects broadly consistent across levels of respiratory support; data in patients on IMV/ECMO were limited.

Secondary outcomes

Progression to IMV or death (by 28 days)

• Therapeutic vs prophylactic: OR 0.80 (95% CI, 0.68–0.94); events 327/3011 vs 384/2937 (10.9% vs 13.1%).
  • ARR: ~2.2%; NNT: ~45.
• Intermediate vs prophylactic: OR 0.84 (95% CI, 0.68–1.05); events 190/1678 vs 222/1695 (11.3% vs 13.1%).
  • ARR: ~1.8%; NNT: ~57.
• Therapeutic vs intermediate: OR 1.30 (95% CI, 1.00–1.71); events 222/619 vs 196/623 (35.9% vs 31.5%).
  • ARI: ~4.4%; NNH: ~23.

Thromboembolic events (by 28 days)

• Therapeutic vs prophylactic: OR 0.48 (95% CI, 0.36–0.64); events 74/3187 vs 148/3102 (2.3% vs 4.8%).
  • ARR: ~2.5%; NNT: ~41.
• Intermediate vs prophylactic: OR 0.67 (95% CI, 0.45–1.00); events 46/1916 vs 75/1931 (2.4% vs 3.9%).
  • ARR: ~1.5%; NNT: ~68.
• Therapeutic vs intermediate: OR 0.63 (95% CI, 0.43–0.93); events 47/880 vs 74/907 (5.3% vs 8.2%).
  • ARR: ~2.8%; NNT: ~36.

Major bleeding (by 28 days)

• Therapeutic vs prophylactic: OR 1.90 (95% CI, 1.19–3.05); events 55/3189 vs 26/3109 (1.7% vs 0.8%).
  • ARI: ~0.9%; NNH: ~112.
• Intermediate vs prophylactic: OR 1.22 (95% CI, 0.66–2.25); events 24/1848 vs 20/1865 (1.3% vs 1.1%).
  • ARI: ~0.2%; NNH: ~435.
• Therapeutic vs intermediate: OR 1.21 (95% CI, 0.66–2.20); events 26/881 vs 20/912 (3.0% vs 2.2%).
  • ARI: ~0.8%; NNH: ~132.

Limitations

• Few data on direct oral anticoagulants; findings largely reflect heparins.
• Outcome ascertainment for bleeding/thromboembolism may involve subjectivity; limited data on event severity.
• Not all eligible trials contributed data; some outcomes/subgroups unavailable, reducing precision.
• Heterogeneity in dosing within comparator arms and disease severity across comparisons complicates interpretation.
• Summary (trial-level) data precluded assessment of interactions across multiple patient-level factors.
• Trials conducted largely earlier in the pandemic; generalizability to current practice may be reduced.

Funding

No direct funding. The World Health Organization supported administrative coordination and data submission processes.

Citation

WHO REACT Working Group. Anticoagulation Among Patients Hospitalized for COVID-19: A Systematic Review and Prospective Meta-analysis. Annals of Internal Medicine. 2025;178:59–69. doi:10.7326/ANNALS-24-00800. Published online 24 December 2024.