Clinical context enhances interpretation of extracorporeal treatment criteria in lithium poisoning

ABSTRACT Background Lithium poisoning presents as acute, acute-on-chronic, or chronic exposure, phenotypes that differ in clinical expression and outcome. The Extracorporeal Treatments in Poisoning Workgroup criteria guide extracorporeal treatment, but whether their prognostic performance varies by poisoning type is uncertain. We assessed these criteria at initial evaluation and whether poisoning type improved risk stratification. Methods We conducted a retrospective multicenter cohort study of adult lithium-poisoning episodes between 2015 and 2022 in 26 tertiary hospitals in Spain, Ireland, and Switzerland. Poisoning was defined by compatible exposure plus serum lithium ≥ 1.5 mEq/L and classified as acute, acute-on-chronic, or chronic. The primary outcome was syndrome of irreversible lithium-effectuated neurotoxicity (SILENT) at 60 days or in-hospital mortality. Sequential logistic-regression models evaluated any criterion, poisoning type, and age; discrimination used the area under the receiver-operating-characteristic curve (AUC). Results Among 644 episodes (136 acute, 196 acute-on-chronic, 312 chronic), the primary outcome occurred in 69 (11%). Any criterion was associated with the outcome after adjustment for poisoning type and age (odds ratio 1.94; 95% CI 1.10–3.44), as were acute-on-chronic poisoning (3.21; 1.27–8.11) and age (per year 1.05; 1.03–1.07). AUC rose from 0.61 to 0.72 with poisoning type and age added. In an exploratory model the clinical criterion remained associated with the outcome, whereas the analytical criterion did not. Conclusions Any Workgroup criterion was associated with adverse outcome. Discrimination improved when poisoning type and age were incorporated, and acute-on-chronic poisoning carried the highest adjusted risk. These findings support interpreting extracorporeal-treatment criteria within the clinical context rather than as serum lithium thresholds alone.