Association Between Platelet Transfusion and ICU Mortality Within 28 Days Among Patients With Intracranial Hemorrhage Complicated by Thrombocytopenia: A Propensity Score-Matched Cohort Study.
Qiu D, Wang Z, et al. • Medical science monitor : international medical journal of experimental and clinical research • 2026
Platelet transfusion was not independently associated with ICU mortality within 28 days among neuro-ICU patients with intracranial hemorrhage complicated by thrombocytopenia.
Key Findings
Results
After propensity score matching, ICU mortality within 28 days did not significantly differ between patients who received platelet transfusions and those who did not.
936 neuro-ICU patients with intracranial hemorrhage complicated by thrombocytopenia were analyzed, including 173 who received platelet transfusions and 763 who did not
After propensity score matching, 113 matched pairs were obtained
Mortality difference between transfusion and non-transfusion groups was not significant (P=0.292)
Survival time also did not significantly differ between groups (P=0.576)
Results
Cox regression and Kaplan-Meier analyses both showed no significant association between platelet transfusion status and 28-day ICU mortality.
Cox proportional hazards models were used to assess associations of platelet transfusion status with ICU mortality
All P-values from Cox regression and Kaplan-Meier analyses were greater than 0.05
Restricted cubic spline analysis was also employed to assess dose-response relationships
Results
Exploratory subgroup analyses suggested heterogeneity in the association between platelet transfusion and mortality among patients with pneumonia.
A significant subgroup finding was observed in patients with pneumonia (P=0.007)
The interaction P-value for the pneumonia subgroup was 0.010, suggesting effect modification
These findings were described as exploratory and requiring further validation
Results
Exploratory subgroup analyses suggested heterogeneity in the association between platelet transfusion and mortality among patients with heart failure.
A borderline finding was observed in the heart failure subgroup (P=0.074)
The interaction P-value for the heart failure subgroup was 0.043
Authors cautioned that these subgroup findings should be interpreted cautiously
Results
Higher platelet transfusion volume was associated with higher mortality in unadjusted analyses, but this association was not significant after multivariable adjustment.
The association between transfusion volume and mortality was observed in unadjusted analyses
After multivariable adjustment, the association between higher transfusion volume and higher mortality was no longer statistically significant
This suggests confounding by indication may explain the unadjusted relationship between transfusion volume and mortality
Methods
The study population consisted of neuro-ICU patients with both traumatic and non-traumatic intracranial hemorrhage complicated by thrombocytopenia.
Total cohort comprised 936 patients admitted to the neuro-intensive care unit
Thrombocytopenia was described as common among patients with traumatic or non-traumatic intracranial hemorrhage admitted to the neuro-ICU
Cox proportional hazards models, Kaplan-Meier analysis, restricted cubic spline analysis, and exploratory subgroup analyses were used as the analytical framework
What This Means
This research suggests that giving platelet transfusions to patients in the neuro-intensive care unit who have bleeding in the brain and low platelet counts does not appear to reduce their risk of dying within 28 days. The study followed 936 such patients, comparing those who received platelet transfusions to those who did not. After using a statistical technique called propensity score matching to create more comparable groups (113 pairs of patients), there was no meaningful difference in survival rates or time between the transfusion and non-transfusion groups.
The study also found that while patients who received larger volumes of platelets appeared to have higher death rates when looking at raw numbers, this relationship disappeared after accounting for other factors that could influence outcomes. This suggests that sicker patients were more likely to receive transfusions in the first place, making it look like transfusions were harmful when they were not independently causing worse outcomes. Two specific patient subgroups — those with pneumonia and those with heart failure — showed signs that platelet transfusion might have different effects in their cases, but the authors emphasize these findings were exploratory and not definitive.
This research matters because platelet transfusions are a common but resource-intensive intervention, and clinicians need to know whether they actually help patients with brain bleeding and low platelet counts. This study suggests that routine platelet transfusion in this population may not improve short-term survival, though the potential differences seen in patients with pneumonia or heart failure warrant further dedicated investigation before any conclusions can be drawn about those specific groups.
Qiu D, Wang Z, Hu X, Wang L, Jiang S, Zhang S, et al.. (2026). Association Between Platelet Transfusion and ICU Mortality Within 28 Days Among Patients With Intracranial Hemorrhage Complicated by Thrombocytopenia: A Propensity Score-Matched Cohort Study.. Medical science monitor : international medical journal of experimental and clinical research. https://doi.org/10.12659/MSM.953547