Risk factors analysis and prediction model construction of SGLT2 inhibitor-associated euglycemic diabetic ketoacidosis
10.12206/j.issn.2097-2024.202403056
- VernacularTitle:钠-葡萄糖协同转运蛋白2抑制剂相关非高血糖性酮症酸中毒的影响因素分析及预测模型建立
- Author:
Wenhui HUANG
1
;
Xiufen CHEN
1
;
Jianming CHEN
1
;
Yana HONG
1
;
Jingjing CAI
1
;
Jinshan CHEN
1
Author Information
1. Department of Pharmacy, b. Department of Endocrinology and Rheumatology, The 909th Hospital/Dongnan Hospital Affiliated to Xiamen University, Zhangzhou 363000, China.
- Publication Type:Originalarticles
- Keywords:
sodium-dependent glucose transporters 2（SGLT2）;
SGLT2 inhibitor;
euglycemic diabetic ketoacidosis(euDKA);
type 2 diabetes;
prediction model
- From:
Journal of Pharmaceutical Practice and Service
2026;44(5):247-252
- CountryChina
- Language:Chinese
-
Abstract:
Objective To explore risk factors of sodium-dependent glucose transporters 2 （SGLT2） inhibitor-associated euglycemic diabetic ketoacidosis （euDKA） and to construct a risk prediction model. Methods A retrospective analysis was performed on the clinical data of type 2 diabetes patients treated with SGLT2 inhibitors in Dongnan Hospital of Xiamen University from January 2020 to December 2023, including age, gender and course of diabetes. The risk factors of SGLT2 inhibitor-associated euDKA were analyzed by univariate analysis and multivariate Logistic regression, and a prediction model was established. According to the receiver's operating characteristic （ROC） curve, the area under the curve （AUC） and the optimal critical value of the prediction model were determined. The prediction model was subjected to both internal and external validation. Results A total of 119 patients with type 2 diabetes treated with SGLT2 inhibitors were included in this study. Among them, there were 98 cases without euDKA （non-euDKA group）and 21 cases with euDKA （euDKA group）. Multivariate Logistic regression analysis showed the DKA history (OR=114.153), appetite or diet decreased three days before admission (OR=21.774), elevated neutrophil count (OR=2.056) and pre-hospital adjustment of hypoglycemic agents (OR=45.745) were independent factors to increase risks of euDKA associated with SGLT2 inhibitors （P<0.05）. Surgical history before admission was an independent factor to reduce this risk （OR=0.007, P<0.05）. By establishing the calculation formula of the prediction model = neutrophil count+6.571 （DKA history）−6.874 （surgical history before admission）+4.273 （appetite or diet decreased three days before admission）+5.302 （pre-hospital adjustment of hypoglycemic drugs）, the ROC curve was drawn. The AUC of the ROC of the prediction model was 0.982 （95%CI: 0.961-1.000, P<0.001）, with accuracy of 94.96%, sensitivity of 0.905, specificity of 0.959 and a critical value of 7.405. The AUC of ROC curve after the model’s ten-fold cross validation was 0.930. And the accuracy of the external validation of the prediction model was 85.29%. Conclusion The DKA history, appetite or diet decreased three days before admission, elevated neutrophil count and pre-hospital adjustment of hypoglycemic agents increased the risk of SGLT2 inhibitor-associated euDKA, while the surgical history before admission reduced this risk. The risk prediction model constructed on this basis could better predict the risk of SGLT2 inhibitor-associated euDKA.
