On-treatment decline in MELD score predicts one-month transplant-free survival in rodenticidal hepatotoxicity patients treated with low-volume plasma exchange.

Background and aim: Plasma exchange (PLEX) improves survival in patients with rodenticidal hepatotoxicity. However, predictors of treatment response are unknown. We aimed at assessing predictors of response to PLEX treatment in these patients.

Methods: Patients with rodenticidal hepatotoxicity from 2014 to 2023 managed in our department were included in this study. Kochi criteria (model for end-stage liver disease [MELD] score $\ge$ 36 or international normalized ratio [INR] $\ge$ 6 with hepatic encephalopathy [HE]) derived specifically for rodenticidal hepatotoxicity (PubMed IDentifier [PMID]: 26310868) were used to assess need for liver transplantation. We analyzed predictors of survival at one month. ∆Bilirubin, ∆MELD score and ∆INR were calculated as percentage change of the parameter after third PLEX session (or after last PLEX if < 3 PLEX sessions done) from baseline pre-PLEX value.

Results: Of 200 patients with rodenticidal hepatotoxicity, 114 patients were treated with low-volume PLEX (PLEX-LV). No patient had liver transplantation. Of 78 patients who fulfilled Kochi criteria, 32 patients were PLEX-LV eligible and underwent PLEX-LV (M: 10; age: 20.5, 7-70 years; median, range; acute liver failure: 24). Twenty-two (69%; acute liver failure: 14) of the 32 patients were alive at one month. Presence of HE (p = 0.03) and ∆MELD (p < 0.001) were significant predictors on univariate analysis, while ∆MELD (aOR = 0.88, 95% CI: 0.79-0.98, p = 0.01) was the only significant independent predictor of one-month transplant-free survival. Area under receiver operating characteristic (ROC) for ∆MELD was 0.93 (95% CI:0.85-1.00) and a decrease of $\ge$ 20% in MELD score while on PLEX-LV had 90% sensitivity and 90% specificity in predicting one-month survival.

Conclusions: Decline in MELD while on PLEX-LV independently predicted one-month transplant-free survival in rodenticidal hepatotoxicity patients. This may help guide decision on stopping PLEX-LV in patients predicted to respond to treatment and to consider alternate treatment options in non-responders.