Prediction of Outcome by Measuring Serum Lactate Concentration-Time Data in Patient with Acute Paraquat Poisoning

Abstract

The standard prognostic assay for acute paraquat poisoning—plasma paraquat concentration—is frequently unavailable in resource-limited settings. While serum lactate and time-lagged composite models have been proposed as accessible alternatives, they lack robust external validation. To externally validate an established arterial lactate-time prognostic model using venous serum lactate, and evaluate its utility for predicting in-hospital mortality. Methods: We conducted a hospital-based observational study of 46 adults presenting with acute paraquat poisoning at a single center in Central India. Model discrimination was assessed using the Area Under the Receiver Operating Characteristic Curve (AUC) with bootstrap 95% confidence intervals (CIs). The cohort experienced a 78.26% in-hospital mortality rate. Admission serum lactate alone demonstrated excellent discrimination (AUC 0.963; 95% CI: 0.896–1.000). The externally validated lactate-time model yielded a lower point estimate (AUC 0.861; 95% CI: 0.718–0.966). An in-sample optimal serum lactate cut-off of 3.0 mmol/L predicted mortality with a sensitivity of 0.944. Time lag did not independently predict mortality in an exploratory multivariable model (p = 0.153). In this setting, complex composite models did not demonstrate a statistically significant discriminative advantage over unadjusted serum lac­tate. While a 3.0 mmol/L triage threshold shows promise, its in-sample derivation requires future multicenter validation.