Lactate-albumin ratio as an early predictor of mortality in critically ill children: A prospective cohort study.

Early risk stratification in critically ill children relies on severity scores such as PRISM-III and PELOD-2, requiring multiple variables. The lactate-albumin ratio (LAR), a two-parameter marker, shows prognostic value in adult and paediatric critical illness. This study evaluated serial LAR for predicting PICU mortality and compared its performance at admission against PRISM-III and PELOD-2. This prospective cohort study enrolled 335 of 563 screened children aged 1 month to less than 13 years admitted to a tertiary-care PICU in South India over 12 months. Serum lactate and albumin were measured at 0, 12, 24, and 48 h. Discriminatory performance was assessed using ROC curve analysis with DeLong's test. Firth-penalized multivariable logistic regression identified independent predictors. Of 335 children, 42 (12.5%) were non-survivors. LAR at admission discriminated mortality with an area under the curve (AUC) of 0.859 (95% CI 0.806-0.907), exceeding PRISM-III (AUC 0.751) and PELOD-2 (AUC 0.759) at the same timepoint (p < 0.05 for both). LAR remained independently associated with mortality after adjustment for PRISM-III, sepsis, and vasoactive support (adjusted odds ratio 3.77, 95% CI 2.22-6.40, p < 0.001), and outperformed lactate and albumin. LAR remained persistently elevated in non-survivors through 48 h.Conclusions: LAR at admission discriminated PICU mortality better than PRISM-III and PELOD-2 and remained an independent predictor after adjustment. Serial LAR changes may be a useful adjunct in predicting prognosis, though their clinical utility requires further prospective study. Derived from two routine investigations, LAR may serve as a practical addition to severity scoring, pending external validation. What is Known • Mortality risk in critically ill children is estimated with multivariable severity scores such as PRISM-III and PELOD-2. • Lactate and albumin refl ect tissue perfusion and systemic infl ammation, and their ratio has been associated with mortality in critical illness. What is New • We hypothesised that the lactate-albumin ratio, from two routine blood tests, may predict PICU mortality at least as well as multivariable severity scores. • We also hypothesised that serial measurement over the fi rst 48 hours may add prognostic information beyond the admission value.