Machine learning algorithms to predict intraoperative hemorrhage in surgical patients: a modeling study of real-world data in Shanghai, China

Safety of Bronchoalveolar Lavage in Hematological Patients with Thrombocytopenia. A Retrospective Cohort Study

Background

Hospitalized hematological patients often require bronchoalveolar lavage (BAL). Scarce evidence exists regarding the potential risks in patients with very severe thrombocytopenia (VST).

Methods

This retrospective-cohort study included adult hematological in-patients with VST, defined as platelets<20x103/μL, undergoing BAL during 2012-2021. Mechanically ventilated patients or those with known active bleeding were excluded. Primary outcomes included major bleeding halting the BAL or deemed significant by the treating physician, need for any respiratory support other than low flow O2, or death within 24 hours. Any other bleedings were recorded as secondary outcomes.

Results

Of the 507 patients included in the final analysis, the 281 patients with VST had lower hemoglobin (Md=0.3, p=0.003), longer prothrombin-time (Md=0.7s, p=0.025), higher chances of preprocedural platelet transfusion (RR 3.68, 95%CI [2.86,4.73]), and only one primary-outcome event (death of septic shock 21h postprocedurally) - compared with 3 (1.3%) events (two bleedings halting procedure and one need for non-invasive-ventilation) in patients with platelets ≥20x103/μL (p=0.219). The risk of minor spontaneously resolved bleeding was higher (RR=3.217, 95% CI [0.919,11.262]) in patients with VST (4.3% vs 1.3%, p=0.051). No association was found between the complications recorded and preprocedural platelets, age, aPTT, P.T., hematological status, or platelet transfusion.

Conclusions

This data suggests BAL to be safe even when platelet counts are <20x103/μL.