A statistical model to identify determinants of glycemic control in patients with type 2 diabetes with different pharmacotherapeutic profiles

Effects of Positive Airway Pressure Therapy on Glycemic Variability in Patients With Type 2 Diabetes and OSA: A Randomized Controlled Trial

BACKGROUND

Glycemic variability is associated with increased risk for cardiovascular disease in patients with type 2 diabetes independent of glycosylated hemoglobin A1c (HbA1c) levels. Given the conflicting evidence on the effect of positive airway pressure (PAP) therapy for OSA on HbA1c, elucidating its effect on glycemic variability has value.

RESEARCH QUESTION

Does the use of PAP therapy for OSA improve glycemic variability in patients with type 2 diabetes?

STUDY DESIGN AND METHODS

A randomized controlled trial was conducted in 184 patients with type 2 diabetes and moderate-to-severe OSA. Participants received either 3 months of PAP therapy with lifestyle counseling or lifestyle counseling alone. End points included the SD of glucose levels along with other metrics derived from continuous glucose monitoring and self-monitoring of blood glucose.

RESULTS

No differences were noted in either primary or secondary continuous glucose monitoring end points between the two groups. Average use of PAP therapy was 5.4 h/night (SD, 1.6). Exploratory analyses by sex showed significant differences in the primary and secondary outcomes. In female participants, PAP therapy was associated with improvement in the SD of glucose levels, with a mean difference in change between intervention and control groups of 3.5 mg/dL (P = .02). PAP therapy was also associated with lower post-dinner and bedtime glucose levels: 20.1 mg/dL (P < .01) and 34.6 mg/dL (P < .01), respectively.

INTERPRETATION

PAP therapy did not improve glycemic control or variability in patients with moderate-to-severe OSA and type 2 diabetes. Exploratory analyses suggested that PAP therapy may improve glucose variability in female participants. Post-dinner and bedtime glucose levels were higher in those who did not receive PAP therapy.

TRIAL REGISTRATION

ClinicalTrials.gov; No.: NCT02454153; URL: www.

CLINICALTRIALS

gov.

Integrated statistical and machine learning analysis provides insight into key influencing symptoms for distinguishing early-onset type 2 diabetes

Background

Being able to predict with confidence the early onset of type 2 diabetes from a suite of signs and symptoms (features) displayed by potential sufferers is desirable to commence treatment promptly. Late or inconclusive diagnosis can result in more serious health consequences for sufferers and higher costs for health care services in the long run.

Methods

A novel integrated methodology is proposed involving correlation, statistical analysis, machine learning, multi-K-fold cross-validation, and confusion matrices to provide a reliable classification of diabetes-positive and -negative individuals from a substantial suite of features. The method also identifies the relative influence of each feature on the diabetes diagnosis and highlights the most important ones. Ten statistical and machine learning methods are utilized to conduct the analysis.

Results

A published data set involving 520 individuals (Sylthet Diabetes Hospital, Bangladesh) is modeled revealing that a support vector classifier generates the most accurate early-onset type 2 diabetes status predictions with just 11 misclassifications (2.1% error). Polydipsia and polyuria are among the most influential features, whereas obesity and age are assigned low weights by the prediction models.

Conclusion

The proposed methodology can rapidly predict early-onset type 2 diabetes with high confidence while providing valuable insight into the key influential features involved in such predictions.