The impact of platelet transfusions on mortality in critically ill patients with thrombocytopenia
| Reference . | N . | Population . | Study design . | Results . | Study quality* . |
|---|---|---|---|---|---|
| Adults | |||||
| 12 | 350 | Dengue fever | Prospective cohort | In patients with PLT counts <50 × 109/L, 2 transfused patients died vs 1 nontransfused patient; deaths were not related to bleeding | Low |
| 11 | 147 | Surgical ICU | Prospective cohort | Transfusion was not associated with an increased risk of death in univariate analysis | Low |
| 10 | 72 | Surgical ICU | Case-control study | In patients with PLTs <50 × 109/L, there was no difference in mortality between transfused and nontransfused patients (50% and 45%, respectively) | Low |
| Neonates | |||||
| 26 | 1389 | Neonates | Retrospective cohort | Thirty-three percent mortality in transfused group vs 3% mortality in the nontransfused group (P = .0001). OR for death (P < .0001): 10.4 with 1 PLT transfusion; 9.4 with 2-4 PLT transfusions; 29.9 with >4 transfusions. | Low |
| 23 | 494 | Neonates | Retrospective cohort | Risk of death increased with additional transfusions (OR = 1.14 per transfusion): 2% in nontransfused; 11% with 1-2 transfusions; 20% with 3-10 transfusions; 35% with >10 transfusions | Moderate |
| 25 | 284 | ELBW neonates | Retrospective cohort | Overall mortality was 23% in transfused neonates. Risk of death increased with additional transfusions: 9% in nontransfused; 20% for 1-5 transfusions; 29% for >5 transfusions. | Low |
| 24 | 273 | Neonates | Retrospective cohort | Unadjusted mortality increased with additional transfusions: 0% in nontransfused; 4% for 1 transfusion; 14% for 2-5 transfusions; 24% for 6-10 transfusions; 36% for 11-20 transfusions; 50% for >20 transfusions | Moderate |
| 14 | 194 | Neonates | Prospective cohort | Thirty-three percent mortality in 31 neonates (≥5 transfusions) vs 2% in 53 nontransfused patients | Moderate |
| 27 | 182 | Neonates | Prospective observational | Sixty-one (33.5%) of 182 thrombocytopenic neonates died, of whom 96.5% received PLTs. Bleeding was not the primary cause of death for any patient. | Low |
| 16 | 164 | Preterm neonates | Case control (94 cases, 70 controls) | Mortality was 48.3% in transfused neonates and 18.2% in nontransfused neonates | Low |
| 8 | 152 | Preterm neonates | RCT | Sixteen deaths (20.5%) in the intervention group (n = 78) vs 11 deaths (14.9%) in the control group (n = 74) (statistical test of significance not provided) | Low |
| 18 | 61 | Neonates | Retrospective cohort | Unadjusted mortality did not correlate with number of transfusions: 42.8% with 1 transfusion; 15.3% with 2-4 transfusions; 28.5% with >4 transfusions | Low |
| 17 | 45 | Neonates | Retrospective cohort | Mortality was 48.8% in patients given ≥20 PLT transfusions | Low |
| 21 | 44 | Preterm neonates | Retrospective review | Five deaths in the transfused group (n = 25) vs no deaths in the nontransfused group (n = 19) in neonates with PLTs <50 × 109/L | Moderate |
| 19 | NR | Neonates | Prospective cohort | No difference in mortality in patients transfused based on count vs mass (0.9% vs 0.4%, respectively; P = .10); no deaths ascribed to bleeding | Low |
| Children | |||||
| 28 | 138 | Medical and surgical ICU | Prospective cohort | Transfusion was not a significant contributor to mortality in adjusted analysis. Unadjusted OR for death: 3.8 (95% CI, 1.25-11.5; P = .01) transfused vs nontransfused. “Transfusion” was not specified but assumed to be PLT transfusion. | Low |
| Reference . | N . | Population . | Study design . | Results . | Study quality* . |
|---|---|---|---|---|---|
| Adults | |||||
| 12 | 350 | Dengue fever | Prospective cohort | In patients with PLT counts <50 × 109/L, 2 transfused patients died vs 1 nontransfused patient; deaths were not related to bleeding | Low |
| 11 | 147 | Surgical ICU | Prospective cohort | Transfusion was not associated with an increased risk of death in univariate analysis | Low |
| 10 | 72 | Surgical ICU | Case-control study | In patients with PLTs <50 × 109/L, there was no difference in mortality between transfused and nontransfused patients (50% and 45%, respectively) | Low |
| Neonates | |||||
| 26 | 1389 | Neonates | Retrospective cohort | Thirty-three percent mortality in transfused group vs 3% mortality in the nontransfused group (P = .0001). OR for death (P < .0001): 10.4 with 1 PLT transfusion; 9.4 with 2-4 PLT transfusions; 29.9 with >4 transfusions. | Low |
| 23 | 494 | Neonates | Retrospective cohort | Risk of death increased with additional transfusions (OR = 1.14 per transfusion): 2% in nontransfused; 11% with 1-2 transfusions; 20% with 3-10 transfusions; 35% with >10 transfusions | Moderate |
| 25 | 284 | ELBW neonates | Retrospective cohort | Overall mortality was 23% in transfused neonates. Risk of death increased with additional transfusions: 9% in nontransfused; 20% for 1-5 transfusions; 29% for >5 transfusions. | Low |
| 24 | 273 | Neonates | Retrospective cohort | Unadjusted mortality increased with additional transfusions: 0% in nontransfused; 4% for 1 transfusion; 14% for 2-5 transfusions; 24% for 6-10 transfusions; 36% for 11-20 transfusions; 50% for >20 transfusions | Moderate |
| 14 | 194 | Neonates | Prospective cohort | Thirty-three percent mortality in 31 neonates (≥5 transfusions) vs 2% in 53 nontransfused patients | Moderate |
| 27 | 182 | Neonates | Prospective observational | Sixty-one (33.5%) of 182 thrombocytopenic neonates died, of whom 96.5% received PLTs. Bleeding was not the primary cause of death for any patient. | Low |
| 16 | 164 | Preterm neonates | Case control (94 cases, 70 controls) | Mortality was 48.3% in transfused neonates and 18.2% in nontransfused neonates | Low |
| 8 | 152 | Preterm neonates | RCT | Sixteen deaths (20.5%) in the intervention group (n = 78) vs 11 deaths (14.9%) in the control group (n = 74) (statistical test of significance not provided) | Low |
| 18 | 61 | Neonates | Retrospective cohort | Unadjusted mortality did not correlate with number of transfusions: 42.8% with 1 transfusion; 15.3% with 2-4 transfusions; 28.5% with >4 transfusions | Low |
| 17 | 45 | Neonates | Retrospective cohort | Mortality was 48.8% in patients given ≥20 PLT transfusions | Low |
| 21 | 44 | Preterm neonates | Retrospective review | Five deaths in the transfused group (n = 25) vs no deaths in the nontransfused group (n = 19) in neonates with PLTs <50 × 109/L | Moderate |
| 19 | NR | Neonates | Prospective cohort | No difference in mortality in patients transfused based on count vs mass (0.9% vs 0.4%, respectively; P = .10); no deaths ascribed to bleeding | Low |
| Children | |||||
| 28 | 138 | Medical and surgical ICU | Prospective cohort | Transfusion was not a significant contributor to mortality in adjusted analysis. Unadjusted OR for death: 3.8 (95% CI, 1.25-11.5; P = .01) transfused vs nontransfused. “Transfusion” was not specified but assumed to be PLT transfusion. | Low |
CI, confidence interval; ELBW, extremely low birth weight (<1000 g); NR, not reported; OR, odds ratio. Other abbreviations are explained in Table 1.
Study quality included applicability to the research question.