The Prediction Effect of HbA1c on Nosocomial Infection in Diabetic Patients was Analyzed Based on Decision Curve and Dose Response.

OBJECTIVE

To analyze the predictive efficacy of HbA1c on nosocomial infection in diabetic patients.

METHODS

566 patients with diabetes who received treatment in our hospital from January 2021 to January 2023 were selected as the study objects. All patients received relevant treatment in the hospital. Patients with nosocomial infection during treatment were included in the occurrence group, and those without nosocomial infection were included in the non-occurrence group. The level of HbA1c and other laboratory indicators before admission were compared between the two groups of patients [gender, hypertension, age, body mass index (BMI), length of stay, primary caregiver, duration of disease, diabetes complications, antibiotic use, fasting blood glucose (FBG), invasive treatment, hemoglobin (HGB) and insulin resistance index (HO) MA-IR), to analyze the relationship between each index and the occurrence of hospital infection in diabetic patients, and to test the predictive value of HbA1c level in the occurrence of hospital infection in diabetic patients.

RESULTS

Among 566 patients with diabetes admitted to our hospital, 139 patients had nosocomial infection, accounting for 24.56%, and 427 patients did not have nosocomial infection, accounting for 75.44%. There were no differences in gender, hypertension, BMI, main caregiver, or HGB between the two groups (P > .05). Age, hospital stay, course of disease, FBG, HbA1c and HOMA-IR in the occurrence group were higher than those in the non-occurrence group, and the proportion of diabetes complications, antibiotic use and invasive treatment was significantly higher than that in the non-occurrence group, with statistical significance (P < .05). Logistics regression analysis showed that old age, long hospital stay, long course of disease, diabetes complications, antibiotic use, high level of FBG, high level of HbA1c, invasive treatment and high level of HOMA-IR were all risk factors for nosocomial infection in diabetic patients (OR > 1, P < .05). The ROC curve showed that the AUC of FBG and HbA1c in predicting the occurrence of hospital infection in diabetic patients was 0.764 and 0.875, respectively, and the predictive energy of HbA1c was higher than that of FBG.

CONCLUSION

HbA1c level is correlated with the occurrence of hospital infection events in diabetic patients, and the correlation intensity with the occurrence of hospital infection events in diabetic patients presents a nonlinear dose-response relationship. Detection of HbA1c levels in diabetic patients is conducive to predicting the probability of hospital infection events, and strict control of HbA1c levels in diabetic patients is conducive to improving patient prognosis.