The association between acute kidney injury and dysglycaemia in critically ill patients with and without diabetes mellitus: a retrospective single-center study

Abstract

Background: Critically ill patients in the intensive care unit (ICU) often experience dysglycaemia. However, studies investigating the link between acute kidney injury (AKI) and dysglycaemia, especially in those with and without diabetes mellitus (DM), are limited.

Methods: We used the Medical Information Mart for Intensive Care IV database to investigate the association between AKI within 7 days of admission and subsequent dysglycaemia. The primary outcome was the occurrence of dysglycaemia (both hypoglycemia and hyperglycemia) after 7 days of ICU admission. Logistic regression analyzed the relationship between AKI and dysglycaemia, while a Cox proportional hazards model estimated the long-term mortality risk linked to the AKI combined with dysglycaemia.

Results: A cohort of 20,008 critically ill patients were included. The AKI group demonstrated a higher prevalence of dysglycaemia, compared to the non-AKI group. AKI patients had an increased risk of dysglycaemia (adjusted odds ratio [aOR] 1.53, 95% confidence interval [CI] 1.41-1.65), hypoglycemia (aOR 1.56, 95% CI 1.41-1.73), and hyperglycemia (aOR 1.53, 95% CI 1.41-1.66). In subgroup analysis, compared to DM patients, AKI showed higher risk of dysglycaemia in non-DM patients (aOR: 1.93 vs. 1.33, P_int<0.01). Additionally, the AKI with dysglycaemia group exhibited a higher risk of long-term mortality compared to the non-AKI without dysglycaemia group. Dysglycaemia also mediated the relationship between AKI and long-term mortality.

Conclusion: AKI was associated with a higher risk of dysglycaemia, especially in non-DM patients, and the combination of AKI and dysglycaemia was linked to higher long-term mortality. Further research is needed to develop optimal glycemic control strategies for AKI patients.

Keywords: Acute kidney injury; diabetes mellitus; dysglycaemia; hyperglycemia; hypoglycemia; long-term mortality.

Graphical abstract

Figure 1.

Study flow chart. Abbreviations: AKI: acute kidney injury; ICU: intensive care unit; MIMIC-IV: Medical Information Mart for Intensive Care IV.

Figure 2.

Risk for dysglycaemia among by AKI stages. Abbreviations: AKI: acute kidney injury. Note: Non-AKI group was selected as reference in the multivariable logistic regression.

Figure 3.

The association between AKI and dysglycaemia in logistic regression with sensitivity analysis and IPTW models. (A) The association between AKI and dysglycaemia; (B) the association between AKI and hypoglycemia; (C) the association between AKI and hyperglycemia. Abbreviations: AKI: acute kidney injury; CI: confidence interval; ICU: intensive care unit; IPTW: inverse probability treatment weighting; LOS: length of stay; OR: odds ratio.

Figure 4.

Subgroup analysis. (A) Subgroup analysis for dysglycaemia; (B) subgroup analysis for hyperglycemia; (C) subgroup analysis for hypoglycemia. The statistical significance was defined as p for interaction <0.05. Abbreviations: AKI: acute kidney injury; AMI: acute myocardial infarction; CHF: congestive heart failure; CI: confidence interval; DM: diabetes mellitus; HTN: hypertension; OR: odds ratio.

Figure 5.

Mediation analysis. (A) Mediating effect of dysglycaemia and long-term mortality associated with AKI; (B) mediating effect of hypoglycemia and long-term mortality associated with AKI; (C) mediating effect of hyperglycemia and long-term mortality associated with AKI. Path A, the effect of AKI level on dysglycaemia; path B, the effect of dysglycaemia on long-term mortality; path C, the total effect of AKI on long-term mortality; path C’, the direct effect of AKI on long-term mortality after controlling dysglycaemia. Abbreviations: ACME: average causal mediating effect; ADE: average direct effect; AKI: acute kidney injury; TE: total effect.