Preoperative fasting glucose value can predict acute kidney injury in non-cardiac surgical patients without diabetes but not in patients with diabetes

TL;DR

A preoperative fasting glucose level of 5.39 mmol/L can predict acute kidney injury after non-cardiac surgery in non-diabetic patients, but not in diabetic patients.

Contribution

This study identifies a specific glucose threshold for predicting AKI in non-diabetic surgical patients, highlighting a difference in risk prediction between diabetic and non-diabetic groups.

Findings

A non-linear association exists between preoperative glucose and AKI in non-diabetic patients.

The optimal cut-point for predicting AKI is 5.39 mmol/L in non-diabetic patients.

Preoperative glucose levels are not significantly associated with AKI in diabetic patients.

Abstract

Postoperative acute kidney injury (AKI) is a common and costly complication after non-cardiac surgery. Patients with or without diabetes could develop hyperglycemia before surgery, and preoperative hyperglycemia was closely associated with postoperative poor outcomes, but the association between preoperative fasting blood glucose level and postoperative AKI is still unclear.

Data from patients undergoing non-cardiac surgery in Chongqing University Cancer Hospital from January 1, 2017, to May 31, 2023, were collected, preoperative glucose value and perioperative variables were extracted, the primary exposure of interest was preoperative glucose value, and the outcome was postoperative AKI.

Data from 39,986 patients were included in the final analysis, 741(1.9%) patients developed AKI, 134(5.6%) in the cohort with DM, and 607(1.6%) in the cohort without DM(OR 1.312, 95% CI 1.028–1.675, P = 0.029). A significant non-linear association between preoperative glucose and AKI exists in the cohort without DM after covariable adjustment (P = 0.000), and every 1 mmol/L increment of preoperative glucose level increased OR by 15% (adjusted OR 1.150, 95% CI 1.078–1.227, P = 0.000), the optimal cut-point of preoperative fasting glucose level to predict AKI was 5.39 mmol/L (adjusted OR 1.802, 95%CI 1.513–2.146, P = 0.000). However, in the cohort with DM, the relation between preoperative glucose and postoperative AKI was not significant after adjusting by covariables (P = 0.437). No significance exists between both cohorts in the risk of AKI over the range of preoperative glucose values.

A preoperative fasting glucose value of 5.39 mmol/L can predict postoperative acute kidney injury after non-cardiac surgery in patients without diagnosed diabetes, but it is not related to AKI in patients with the diagnosis.

The online version contains supplementary material available at 10.1186/s13741-024-00398-4.

Figures