Bleeding outcomes of inpatients receiving therapeutic plasma exchange: A propensity-matched analysis of the National Inpatient Sample

Abstract

Background: Although therapeutic plasma exchange (TPE) is associated with hemostatic abnormalities, its impact on bleeding outcomes is unknown. Therefore, the main study objective was to determine bleeding outcomes of inpatients treated with TPE.

Study design and methods: In a cross-sectional analysis of the National Inpatient Sample (NIS), discharges were identified with 10 common TPE-treated conditions. A 1:3 propensity-matched analysis of TPE- to non-TPE-treated discharges was performed. The primary outcome was major bleeding and secondary outcomes were packed red blood cell (PRBC) transfusion, mortality, disposition, hospital length of stay (LOS), and charges. Multivariable regression analyses were used to examine the association between TPE and study outcomes.

Results: The study population was 15,964 discharges, of which 3991 were TPE- treated. The prevalence of major bleeding was low (5.4%). When compared to non-TPE discharges, TPE had a significant and positive association with major bleeding (OR = 1.37, 95% CI: 1.16-1.63, p = .0003). TPE was also associated with PRBC transfusion (OR = 1.66, 95% CI: 1.42-1.94, p < .0001), in-hospital mortality (OR = 1.45, 95% CI: 1.10-1.90, p = .0008), hospital length of stay (12.45 [95% CI: 11.95-12.97] vs. 7.38 [95% CI: 7.12-7.65] days, p < .0001) and total charges, ($125,123 [95% CI: $119,220-$131,317] vs. $61,953 [95% CI: $59,391-$64,625], p < .0001), and disposition to non-self-care (OR = 1.29, 95% CI: 1.19-1.39, p < .0001).

Discussion: The use of TPE in the inpatient setting is positively associated with bleeding; however, with low prevalence. Future studies should address risk factors that predispose patients to TPE-associated bleeding.

Keywords: National Inpatient Sample; bleeding; plasmapheresis.