Association of the Albumin-Bilirubin score with 7-day incident delirium risk following bloodstream infection in critically ill adults: evidence from a propensity-weighted cohort.

backgroundDelirium is a frequent complication in patients with bloodstream infection (BSI) and is associated with poor outcomes. The albumin–bilirubin (ALBI) score, an evidence-based index of hepatic function and systemic inflammation, has shown prognostic value in sepsis. However, its association with the risk of incident delirium in patients with BSI remains unclear. We aimed to evaluate the utility of the ALBI score in stratifying the risk of incident delirium in this population.

methodsWe conducted a retrospective cohort study using the Medical Information Mart for Intensive Care IV database (MIMIC-IV, v3.1) involving adult patients admitted to the intensive care unit (ICU) with a first episode of BSI. The primary exposure was the baseline ALBI score, and the primary outcome was incident delirium within 7 days. To rigorously control for confounding, we adopted a doubly robust estimation strategy. Inverse probability of treatment weighting (IPTW) based on covariate-balancing propensity scores (CBPS) was applied to maximize the balance of baseline characteristics between groups. Subsequently, multivariable Fine-Gray proportional subdistribution hazard models were employed to estimate subdistribution hazard ratios (sHRs), accounting for death as a competing risk and adjusting for residual confounding using covariates identified via a hybrid feature selection strategy. Additionally, restricted cubic spline (RCS) analysis was performed to characterize dose–response relationships, and subgroup analyses were conducted to evaluate consistency across clinical strata.

resultsA total of 651 patients were included and stratified into low- and high-ALBI groups based on a median cutoff of -1.51. The high-ALBI group had a significantly higher 7-day cumulative incidence of delirium compared to the low-ALBI group (56.4% vs. 44.3%, Gray’s test P = 0.004). In the fully adjusted Fine-Gray model, each 1-unit increase in the ALBI score was independently associated with an increased risk of delirium (sHR 1.80, 95% CI 1.38–2.33, P < 0.001). Categorical analysis confirmed that patients in the high-ALBI group had nearly double the risk (sHR 1.98, 95% CI 1.38–2.83, P < 0.001). RCS analysis revealed a non-linear, J-shaped dose–response relationship. Subgroup analyses demonstrated that these associations remained largely consistent across various clinical strata, with no significant interactions identified (P for interaction > 0.05).

conclusionsHigher baseline ALBI scores are independently associated with an increased risk of 7-day incident delirium in critically ill adults with BSI. Derived from routinely available laboratory tests, the ALBI score may serve as a complementary objective marker for early risk stratification identifying patients who may benefit from preventive interventions, rather than a standalone predictive tool. CLINICAL TRIAL NUMBER: Not applicable.