Albumin-corrected anion gap and 28-/90-day mortality in cirrhotic patients with invasive fungal infection: a retrospective cohort study

BACKGROUND: The prognostic role of the albumin-corrected anion gap (ACAG) in cirrhotic patients with invasive fungal infection (IFI) remains unclear. This retrospective cohort study aimed to investigate the association between ACAG and 28-/90-day mortality in this high-risk population. METHODS: A retrospective cohort study was conducted at Beijing Ditan Hospital, including patients with cirrhosis complicated by IFI who were admitted between January 2009 and December 2021. ACAG was calculated using initial anion gap and albumin measurements within 24 h of admission. Patients were stratified into high and normal ACAG groups by an ACAG cutoff of 20 mmol/L. The primary outcomes were 28-day and 90-day all-cause mortality. Cox proportional hazards regression models and Kaplan-Meier survival analysis were employed to assess the association between ACAG levels and mortality outcomes. Subgroup analyses were performed to explore the consistency of the association. The prognostic value of ACAG was assessed by receiver operating characteristic (ROC) curve analysis. RESULTS: A total of 313 cirrhotic patients with IFI were enrolled. Patients with high ACAG (≥ 20 mmol/L) had significantly higher 28-day (36.44% vs. 12.82%, P < 0.001) and 90-day mortality (55.93% vs. 25.13%, P < 0.001) compared to the normal ACAG group. After adjusting covariates (Model3), elevated ACAG levels remained an independent risk factor for both 28-day (adjusted HR = 1.06, 95% CI: 1.01-1.11, P = 0.03) and 90-day mortality (adjusted HR = 1.05, 95% CI: 1.01-1.1, P = 0.018). When ACAG was analyzed as a categorical variable, the high ACAG group exhibited a significantly higher mortality rate compared with the normal ACAG group. The results remained consistent across all subgroups. The ACAG demonstrated good predictive ability, with an AUC of 0.721 (95% CI: 0.653-0.790) for 28-day mortality. The AUC value of ACAG for predicting 90-day mortality was 0.699. Furthermore, the predictive performance for both 28-day and 90-day mortality was improved by combining ACAG with the MELD score, yielding AUCs of 0.803 (95% CI: 0.741-0.865) and 0.750 (95% CI: 0.693-0.807), respectively. CONCLUSION: ACAG was an independent predictor of 28-/90-day mortality in cirrhotic patients with IFI, and may serve as a valuable and readily available prognostic biomarker to complement existing risk assessment tools.