Abstract
BACKGROUND: Platelets are frequently transfused, but supply and potential harms highlight the importance of appropriate use. STUDY DESIGN AND METHODS: Our systematic review (SR) followed a predefined protocol. Eligible studies included SRs, randomized controlled trials (RCTs), and matched cohort observational studies between 1946 and March 2025. Populations included were hypoproliferative thrombocytopenia, periprocedural prophylaxis, cardiovascular surgery, consumptive thrombocytopenia, and intracranial hemorrhage. The intervention was restrictive versus liberal platelet transfusion strategies on outcomes of mortality and bleeding. Duplicate screening and data extraction occurred. Meta-analysis used Mantel-Haenszel method of random effects model. RESULTS: Twenty-one RCTs, 24 observational studies, and 20 SRs were included. The evidence quality varied. For hypoproliferative thrombocytopenia, 11 RCTs were analyzed, with 9 RCTs at moderate risk of bias (ROB). Two RCTs were identified for dengue, with high ROB for bleeding. One RCT was identified each in cardiovascular surgery, intracranial hemorrhage, and periprocedural prophylaxis. Meta-analyses indicated no significant effect for outcomes of mortality or bleeding by strategy, but confidence intervals (CIs) were wide. Effect estimates were 1.32 [0.93, 1.86] for all-cause mortality in hypoproliferative thrombocytopenia, 0.80 [0.38, 1.70] in cardiovascular surgery, and 0.69 [0.47, 1.03] in critically ill neonates or dengue patients. DISCUSSION: A consistent lack of benefit with liberal platelet transfusion was observed across analyzed populations, although wide confidence intervals do not exclude clinically meaningful impacts. Important research gaps are highlighted in areas where the RCT data is limited.