Abstract Objective This study aimed to estimate the effect of extracorporeal treatments (ECTRs) on 28-day survival in patients with acute poisoning. Methods This multi-center retrospective cohort, from June 2015 to December 2023, was analyzed using a target trial emulation with time zero defined as presentation to the emergency department (ED). The treatment strategy was receipt of ECTR (hemoperfusion, hemodialysis, continuous kidney replacement therapy, and/or or plasma exchange) versus a non-ECTR strategy. Eligible participants were patients with acute poisoning presenting to the ED with documented toxin exposure and ECTR status. Toxin exposure was stratified as single versus multiple (≥ 2 toxins). Single-toxin cases were further grouped into five categories: pesticides, biotoxins, medications, chemicals, and alcohols and drugs. This study applied inverse probability of treatment weighting (IPTW) to emulate randomization. Within each stratum, a logistic regression propensity score model predicted receipt of ECTR using baseline covariates available at or before time zero: demographics, physiological status, presentation and exposure history, pre-decision interventions and supports, and comorbidities. We estimated IPTW-adjusted 28-day survival risks under the ECTR and no-ECTR strategies, then derived the risk difference (RD) and relative risk (RR) with 95% confidence intervals using a nonparametric bootstrap of the weighted risks. Results The study cohort included 5,269 patients with acute poisonings: pesticides (n = 1,795, including 611 receiving ECTR), medications (n = 1,198, including 111 receiving ECTR), chemicals (n = 838, including 11 receiving ECTR), biotoxins (n = 606, including 12 receiving ECTR), alcohols and drugs (n = 118, including 11 receiving ECTR), and multiple toxins (n = 714, including 152 receiving ECTR). Multiple-toxin exposures showed no significant benefit from ECTR (RD: −0.0576, 95% CI: −0.1176 to 0.0025; RR: 0.939, 95% CI: 0.877 to 1.005). In patients with pesticide poisoning, ECTR was associated with reduced survival (RD: −0.0575, 95% CI: −0.1044 to − 0.0107; RR: 0.930, 95% CI: 0.877 to 0.987). Patients with biotoxin poisoning showed a modest survival benefit with ECTR (RD: 0.0104, 95% CI: 0.0019 to 0.0189; RR: 1.011, 95% CI: 1.002 to 1.019). For medication poisonings, ECTR did not present a meaningful effect on survival (RD: 0.0096, 95% CI: −0.0129 to 0.0321; RR: 1.010, 95% CI: 0.987 to 1.033). Chemical exposures revealed large uncertainty (RD: −0.1694, 95% CI: −0.4066 to 0.0677; RR: 0.827, 95% CI: 0.596 to 1.148). The alcohols and drugs group was excluded due to insufficient mortality events (n = 2). Conclusions This study showed that the effect of ECTR on survival after acute poisoning is toxin-dependent, with heterogeneity across toxin categories. This abstract is funded by: National Natural Science Foundation of China, grant number No. 82372214
Sun et al. (Fri,) studied this question.