Safety of lumbar puncture in adult patients with acute lymphoblastic leukemia and thrombocytopenia: A retrospective study

Abstract Introduction: Acute lymphoblastic leukemia (ALL) requires lumbar punctures (LPs) for central nervous system (CNS) diagnosis and treatment. Guidelines suggest a platelet threshold of ≥50×10⁹/L; however, patients often present with thrombocytopenia. Evidence on the safety of performing LPs with low platelet counts in adults is limited. Objective: To evaluate the association between thrombocytopenia and adverse events (AEs) following LP in adults with ALL. Methods: Retrospective study conducted at the Instituto Nacional de Cancerología (Mexico, 2017–2023). We included 185 adult ALL patients, totaling 442 LPs. LPs were categorized by pre-procedure platelet count: ≥100×10⁹/L, 50–99×10⁹/L, and <50×10⁹/L. Hemorrhagic events, traumatic LPs, headache, and other AEs were assessed. Statistical tests compared groups. Results: Median age was 33.7 years; 68.6% were young adults. 93.5% had B-cell ALL, and 17.8% had CNS involvement. 72.4% had thrombocytopenia at diagnosis, with 30.8% having <20×10⁹/L. Overall, 13 hemorrhagic events (2.9%) were identified, with no significant difference between platelet groups (*p*=0.7891). A total of 46 LP-associated AEs (10.4%) were recorded, primarily headache (9.2%), puncture-site bleeding, subdural hematoma, and seizures. AE incidence was higher in thrombocytopenic groups (*p*=0.003), though grade III–IV events were rare (*n*=2) and not significantly different (*p*=0.8074). Traumatic LP rates were similar across platelet categories (*p*=0.9473). Prophylactic transfusion practices varied. Among patients with <50×10⁹/L, 79 transfusions were administered, with no significant differences in headache or puncture-site bleeding between transfused and non-transfused patients (*p*>0.4). One subdural hematoma occurred in a transfused patient. Conclusions: In our study, thrombocytopenia <50×10⁹/L was not associated with increased risk of severe bleeding or traumatic LP. However, the optimal safety threshold remains uncertain, with emerging literature suggesting ≥30×10⁹/L. Operator skill may be a key factor in procedural safety. Multicenter prospective studies are needed to establish optimal thresholds.