Non-linear associations of risk factors with mild hypoglycemia among Chinese patients with type 2 diabetes

Screening of Individuals with Type 2 Diabetes on Anti-Diabetic Agents for Probable Hypoglycaemia Using the Stanford Hypoglycemia Questionnaire (SHQ) in Outpatient Settings: A Cross-Sectional Study from Outpatient Diabetes Care Centres in North India

Introduction

The study was aimed at identifying the incidence of unreported probable hypoglycaemia in individuals with type 2 diabetes (T2DM) on anti-diabetic medications, using the screening Stanford Hypoglycemia Questionnaire (SHQ) in real-world situations.

Methods

It was a multicentre cross-sectional study on consecutive individuals attending 10 diabetes care centres in Lucknow, Uttar Pradesh, India. The inclusion criteria were as follows: known individuals with T2DM, literate, age greater than or equal to 18 years, on at least one anti-diabetic agent for more than a month and not engaged in regular self-monitoring of blood glucose (SMBG).

Results

This study was conducted from August 2017 to April 2018, involving 1198 participants. The mean age of the individuals enrolled was 53.45 years (±10.83), with males comprising 55.3% of the population. It was found that 63.6% of patients were on sulphonylurea (SU), 14.5% were on pioglitazone, 92.2% on metformin, 62.3% on Dipeptidyl peptidase (DPP4i) and 12.8% on Sodium-glucose cotransporter (SGLT2i). The mean SHQ score was 1.81 (±1.59). Probable hypoglycaemia was mild in 57.59%, moderate in 14.69% and severe in 1.41%. Those with diabetic neuropathy (P = <0.001), retinopathy (P = <0.001) and nephropathy (P = <0.001) had significantly higher SHQ scores. Insulin or SU use was associated with a significantly higher SHQ score. Concomitant statin use was associated with a lower incidence of mild, moderate and severe hypoglycaemia (P = 0.01). On multivariate analysis, we found that age, sex, systolic blood pressure (SBP), insulin use and fasting blood sugar were the most important factors associated with an increased risk of hypoglycaemia with an R2 cut-off of 0.7.

Conclusion

SHQ was discovered to be a simple and cost-effective screening tool for outpatient detection of hypoglycaemia in an Indian setting, and it can add value to management.

Prediction of Hypoglycemia in Diabetic Patients During Colonoscopy Preparation

OBJECTIVE

To explore the clinical outcomes and establish a predictive model of hypoglycemia during colonoscopy preparation for diabetic patients.

METHODS

Three-hundred ninety-four patients with diabetes who received colonoscopy were retrospectively enrolled in this study and assigned to hypoglycemia or non-hypoglycemia groups. Information about clinical characteristics and outcomes during colonoscopy preparation was collected and compared between the two groups. Logistic regression analysis was applied to identify the risk factors of hypoglycemia. These risk factors were used to construct a hypoglycemia predictive model verified by the receiver operating characteristic (ROC) curve and Hosmer-Lemeshow goodness fit test.

RESULTS

Among 394 participants, 66 (16.8%) underwent a total of 88 hypoglycemia attacks during the bowel preparation. Grade 1 hypoglycemia (≤3.9 mmol/L) comprised 90.9% (80/88) of all hypoglycemia attacks and grade 2 hypoglycemia accounted for 9.1% (8/88), signifying that grade 1 hypoglycemia is the most common type. No severe hypoglycemia was identified. The incidence of nocturnal hypoglycemia was 15.9%. Logistic regression analyses revealed that the main risk factors of hypoglycemia during colonoscopy preparation were postprandial C-peptide, serum triglyceride, gender, type of diabetes mellitus, and insulin injection frequencies. The area under the ROC curve of the hypoglycemia prediction model was 0.777 (95% CI: 0.720-0.833).

CONCLUSION

Diabetic patients are prone to develop mild to moderate hypoglycemia during colonoscopy preparation. This study proposes a predictive model that could provide a reference for identifying patients with a high risk of hypoglycemia during colonoscopy preparation.

Development and validation of a prediction model of perioperative hypoglycemia risk in patients with type 2 diabetes undergoing elective surgery

AIM

To develop and validate a prediction model to evaluate the perioperative hypoglycemia risk in hospitalized type 2 diabetes mellitus (T2DM) patients undergoing elective surgery.

METHODS

We retrospectively analyzed the electronic medical records of 1410 T2DM patients who had been hospitalized and undergone elective surgery. Regression analysis was used to develop a predictive model for perioperative hypoglycemia risk. The receiver operating characteristic (ROC) curve and the Hosmer-Lemeshow test were used to verify the model.

RESULTS

Our study showed an incidence of 10.7% for level 1 hypoglycemia and 1.8% for level 2 severe hypoglycemia during the perioperative period. A perioperative hypoglycemic risk prediction model was developed that was mainly composed of four predictors: duration of diabetes ≥ 10 year, body mass index (BMI) < 18.5 kg/m², standard deviation of blood glucose (SDBG) ≥ 3.0 mmol/L, and preoperative hypoglycemic regimen of insulin subcutaneous. Based on this model, patients were categorized into three groups: low, medium, and high risk. Internal validation of the prediction model showed high discrimination (ROC statistic = 0.715) and good calibration (no significant differences between predicted and observed risk: Pearson χ² goodness-of-fit P = 0.765).

CONCLUSIONS

The perioperative hypoglycemic risk prediction model categorizes the risk of hypoglycemia using only four predictors and shows good reliability and validity. The model serves as a favorable tool for clinicians to predict hypoglycemic risk and guide future interventions to reduce hypoglycemia risk.

Exploratory study on classification of diabetes mellitus through a combined Random Forest Classifier

BACKGROUND

Diabetes Mellitus (DM) has become the third chronic non-communicable disease that hits patients after tumors, cardiovascular and cerebrovascular diseases, and has become one of the major public health problems in the world. Therefore, it is of great importance to identify individuals at high risk for DM in order to establish prevention strategies for DM.

METHODS

Aiming at the problem of high-dimensional feature space and high feature redundancy of medical data, as well as the problem of data imbalance often faced. This study explored different supervised classifiers, combined with SVM-SMOTE and two feature dimensionality reduction methods (Logistic stepwise regression and LAASO) to classify the diabetes survey sample data with unbalanced categories and complex related factors. Analysis and discussion of the classification results of 4 supervised classifiers based on 4 data processing methods. Five indicators including Accuracy, Precision, Recall, F1-Score and AUC are selected as the key indicators to evaluate the performance of the classification model.

RESULTS

According to the result, Random Forest Classifier combining SVM-SMOTE resampling technology and LASSO feature screening method (Accuracy = 0.890, Precision = 0.869, Recall = 0.919, F1-Score = 0.893, AUC = 0.948) proved the best way to tell those at high risk of DM. Besides, the combined algorithm helps enhance the classification performance for prediction of high-risk people of DM. Also, age, region, heart rate, hypertension, hyperlipidemia and BMI are the top six most critical characteristic variables affecting diabetes.

CONCLUSIONS

The Random Forest Classifier combining with SVM-SMOTE and LASSO feature reduction method perform best in identifying high-risk people of DM from individuals. And the combined method proposed in the study would be a good tool for early screening of DM.

Safety and effectiveness of ipragliflozin in elderly versus non-elderly Japanese type 2 diabetes mellitus patients: 12 month interim results of the STELLA-LONG TERM study

Objective: STELLA-LONG TERM is an ongoing post-marketing surveillance study examining the safety and effectiveness of ipragliflozin in real-world clinical practice in Japan. This interim report of STELLA-LONG TERM examined the safety and effectiveness of ipragliflozin in non-elderly and elderly Japanese patients with type 2 diabetes mellitus (T2DM) using data up to 12 months. Methods: Data from T2DM patients who were first prescribed ipragliflozin between July 2014 and October 2015 and whose 12 month data were locked by January 2018 were analyzed and compared between non-elderly (<65 years) and elderly patients (≥65 years). Results: The safety and efficacy analysis sets included 11,051 and 8788 patients, respectively. Elderly patients accounted for 28.6% (n = 3157) of the study population. The mean body mass index was 29.9 kg/m² and 26.8 kg/m², the percentage of patients with glycated hemoglobin (HbA1c) <8.0% was 50.1% and 59.5%, and the percentage of patients with complications was 83.2% and 87.3% in the non-elderly and elderly groups, respectively. Mean HbA1c and body weight decreased significantly from baseline to 12 months in both age groups, regardless of baseline HbA1c and body weight (all p < .05). The incidence of adverse drug reactions (ADRs) was 14.8% and 14.2% and that of serious ADRs was 0.8% and 1.4% in non-elderly and elderly patients, respectively (p = .002 for serious ADRs). Conclusion: The incidence of serious ADRs was higher in elderly patients than non-elderly patients. Ipragliflozin was effective in both non-elderly and elderly patients with T2DM in the real-world clinical setting.

Artificial neural network approach to predict surgical site infection after free-flap reconstruction in patients receiving surgery for head and neck cancer

Background

The aim of this study was to develop an effective surgical site infection (SSI) prediction model in patients receiving free-flap reconstruction after surgery for head and neck cancer using artificial neural network (ANN), and to compare its predictive power with that of conventional logistic regression (LR).

Materials and methods

There were 1,836 patients with 1,854 free-flap reconstructions and 438 postoperative SSIs in the dataset for analysis. They were randomly assigned tin ratio of 7:3 into a training set and a test set. Based on comprehensive characteristics of patients and diseases in the absence or presence of operative data, prediction of SSI was performed at two time points (pre-operatively and post-operatively) with a feed-forward ANN and the LR models. In addition to the calculated accuracy, sensitivity, and specificity, the predictive performance of ANN and LR were assessed based on area under the curve (AUC) measures of receiver operator characteristic curves and Brier score.

Results

ANN had a significantly higher AUC (0.892) of post-operative prediction and AUC (0.808) of pre-operative prediction than LR (both P<0.0001). In addition, there was significant higher AUC of post-operative prediction than pre-operative prediction by ANN (p<0.0001). With the highest AUC and the lowest Brier score (0.090), the post-operative prediction by ANN had the highest overall predictive performance.

Conclusion

The post-operative prediction by ANN had the highest overall performance in predicting SSI after free-flap reconstruction in patients receiving surgery for head and neck cancer.

Real-world evidence for the safety of ipragliflozin in elderly Japanese patients with type 2 diabetes mellitus (STELLA-ELDER): final results of a post-marketing surveillance study

OBJECTIVE

To investigate the real-world safety of ipragliflozin in elderly Japanese patients with type 2 diabetes mellitus (T2DM).

RESEARCH DESIGN AND METHODS

Japanese patients (≥65 years old) who were first prescribed ipragliflozin within 3 months after its launch in April 2014 were registered in this post-marketing surveillance (PMS). Final data collection was in July 2015. Survey items included demographics, treatments, adverse drug reactions (ADRs), vital signs, and laboratory variables.

RESULTS

The PMS included 8505 patients (4181 males/4324 females). The mean age and diabetes duration were 72.3 years and 10.6 years, respectively. In 84.3% of patients, ipragliflozin was prescribed at 50 mg/day, which was continued unchanged. Overall, 16.91% of patients experienced 1880 ADRs, and 165 ADRs were classified as serious in 127 patients (1.49%). ADRs of special interest included skin complications, volume depletion, polyuria/pollakiuria, genital infection, urinary tract infection, renal disorders, hypoglycemia, cerebrovascular disease, cardiovascular disease, malignant tumor, fracture, and ketone body-related events.

CONCLUSIONS

This 1-year PMS revealed probable ADRs in elderly Japanese patients with T2DM prescribed ipragliflozin in real-world settings, with no new safety concerns. The risk factors for ADRs varied but could be rationalized. The results should help physicians to identify possible treatment-emergent ADRs in ipragliflozin-treated patients.