Abstract
Background: Platelet consumptive disorders (immune thrombocytopenia [ITP], thrombotic thrombocytopenic purpura [TTP], heparin-induced thrombocytopenia [HIT], and disseminated intravascular coagulation [DIC]) are characterized by accelerated platelet destruction and severe thrombocytopenia. Platelet transfusions (PLT-T) remain controversial in these conditions due to concerns that exogenous platelets may fuel microthrombosis or worsen the underlying disorder. While guidelines discourage PLT-T in TTP and HIT, real-world use persists for life-threatening bleeding or peri-procedural settings. This study addresses these gaps by evaluating associations between PLT-T and thrombotic/mortality outcomes.
Methods:
We conducted a retrospective cohort study of all admitted adult patients (aged 18 years and older) from 2018 to 2020 who had primary or secondary diagnoses of platelet consumptive disorders—including ITP, TTP, HIT, and other primary thrombocytopenias—identified using validated ICD-10-CM codes. Patient records were drawn from the Healthcare Cost and Utilization Project all-payer inpatient healthcare database. Elective admissions were excluded. The exposure was PLT-T administration (ICD-10-PCS codes). Primary outcomes were arterial/venous thrombotic events and in-hospital mortality. Secondary outcomes included length of stay (LOS) and inflation-adjusted hospitalization costs. Multivariable linear and logistic regression (as applicable) was used to estimate adjusted odds ratios (aORs) for primary outcomes, adjusting for demographics, comorbidities, disorder subtype, hematologic malignancy, illness severity (mechanical ventilation, sepsis, shock), and treatments (plasma exchange). Sensitivity analyses were done with stratifying by disorder subtype and hematologic malignancy. All analyses were done using SPSS v26. Significance was set to α of 0.05.Results:
There were a total of 41,678 hospitalizations with consumptive thrombocytopenias during the study period. (ITP:71.8%, HIT:16.3%, TTP:9.5%, Other:2.8%). Of them, 11.3% received PLT-T. The PLT-T cohort had significantly higher baseline acuity (ventilation:12.0% vs 8.7%, shock:11.8% vs 9.4%, hematologic malignancy:10.3% vs 6.5%; all p<0.001).
Thrombotic Events: PLT-T was associated with lower adjusted odds of thrombosis (aOR 0.764, p<0.001). Factors increasing thrombosis risk included HIT (aOR 3.210, p<0.001), shock (aOR 1.546, p<0.001), and being on mechanical ventilation (aOR 1.593, p<0.001).
In-Hospital Mortality: PLT-T was associated with higher adjusted odds of mortality (aOR 1.411, p<0.001). Mortality was high with TTP (aOR 2.422, p=0.005), sepsis (aOR 1.843, p<0.001), shock (aOR 3.307, p<0.001), and mechanical ventilation (aOR 8.643, p<0.001).
Secondary Outcomes: PLT-T was associated with longer mean LOS (10.04 vs 8.18 days, p<0.001) and higher mean costs ($179,597 vs $120,825, p<0.001). Thrombotic events also significantly increased LOS (11.92 vs 7.53 days, p<0.001).
Sensitivity Analyses: The association between PLT-T and reduced thrombosis persisted when stratified by presence/absence of hematologic malignancy (No malignancy: aOR 0.777, p<0.001; Malignancy: aOR 0.603, p=0.002).
Conclusion: In consumptive thrombocytopenias, PLT-T was linked to fewer thrombotic events but higher mortality, likely due to confounding by indication (i.e., sicker patients receiving transfusions). These findings reinforce guideline recommendations to avoid PLT-T in high-thrombosis-risk conditions (e.g., TTP/HIT) and highlight the need for judicious use in acutely ill populations.
