Saturation effect of anion gap in predicting long-term mortality in critically ill patients with chronic obstructive pulmonary disease (COPD): A retrospective cohort study based on the MIMIC-IV database

Clinical Snapshot

Key Findings

• P Value: P < 0.05

• Effect Size: Adjusted hazard ratio 1.03 per 1 mmol/L increase in anion gap

• Primary Outcome: 365-day mortality: 36.24% of critically ill COPD patients

• Nnt Or Sensitivity: Quartile 4 vs Q1: HR 1.49 (1.19-1.87); saturation threshold at 14 mmol/L

• Confidence Interval: 95% CI: 1.01-1.05

Clinical Application

Highly feasible - anion gap is routine laboratory test available in all Australian ICUs Relevant to Australian intensive care practice; aligns with ANZICS guidelines for COPD management and prognostication Critically ill COPD patients requiring ICU admission

Abstract

The relationship between anion gap (AG) and long-term mortality in intensive care unit has been widely reported, but whether this association exists in critically ill patients with chronic obstructive pulmonary disease (COPD) is still unknown. The data of this study were collected from the Medical Information Mart for Intensive Care-IV. First of all, we used the Cox regression analysis and Kaplan-Meier curves to measure the relationship between AG and 365-day mortality for critically ill patients with COPD. Next, a restricted cubic spline was used to analyze the relationship between AG and mortality. Finally, age, sex, weight, hypertension, type 2 diabetes mellitus, heart failure, myocardial infarction, and chronic kidney disease were considered for subgroup analysis. A total of 2594 eligible subjects were sampled, of which 36.24% died within 365 days of intensive care unit admission. Cox regression analysis, after adjusting for confounders, demonstrated a significant association between AG and 365-day mortality in patients with COPD (hazard ratio = 1.03, 95% confidence interval: 1.01-1.05. P < .05). Stratifying AG into quartiles revealed higher levels of AG associated with an increased risk of death (Q1: 1.00, Q2: 1.34 [1.09-1.66], Q3: 1.44 [1.17-1.78], and Q4: 1.49 [1.19-1.87]). Additionally, restricted cubic spline analysis indicated a nonlinear relationship, with a critical value of AG at 14 mmol/L. Subgroup analysis highlighted AG as a significant predictor of long-term mortality in COPD patients across different subgroups, with an interaction effect observed in the subgroup with type 2 diabetes mellitus. In critically ill patients with COPD, there was a significant positive association between AG and 365-day mortality. In addition, there is a saturation effect at AG of 14 mmol/L.

References

1. 1.Xiu, X., Yang, Z., & Chen, X. (2026). Saturation effect of anion gap in predicting long-term mortality in critically ill patients with chronic obstructive pulmonary disease (COPD): A retrospective cohort study based on the MIMIC-IV database. Medicine, 105(16), e41364. https://doi.org/10.1097/MD.0000000000041364