Reducing HbA1c in Type 2 Diabetes Using Digital Twin Technology-Enabled Precision Nutrition: A Retrospective Analysis

Introduction The objective of this study was to examine changes in hemoglobin A1c (HbA1c), anti-diabetic medication use, insulin resistance, and other ambulatory glucose profile metrics between baseline and after 90 days of participation in the Twin Precision Nutrition (TPN) Program enabled by Digital Twin Technology. Methods This was a retrospective study of patients with type 2 diabetes who participated in the TPN Program and had at least 3 months of follow-up. The TPN machine learning algorithm used daily continuous glucose monitor (CGM) and food intake data to provide guidelines that would enable individual patients to avoid foods that cause blood glucose spikes and to replace them with foods that do not produce spikes. Physicians with access to daily CGM data titrated medications and monitored patient conditions. Results Of the 89 patients who initially enrolled in the TPN Program, 64 patients remained in the program and adhered to it for at least 90 days; all analyses were performed on these 64 patients. At the 90-day follow-up assessment, mean (± standard deviation) HbA1c had decreased from 8.8 ± 2.2% at baseline by 1.9 to 6.9 ± 1.1%, mean weight had decreased from 79.0 ± 16.2 kg at baseline to 74.2 ± 14.7 kg, and mean fasting blood glucose had fallen from 151.2 ± 45.0 mg/dl at baseline to 129.1 ± 36.7 mg/dl. Homeostatic model assessment of insulin resistance (HOMA-IR) had decreased by 56.9% from 7.4 ± 3.5 to 3.2 ± 2.8. At the 90-day follow-up assessment, all 12 patients who were on insulin had stopped taking this medication; 38 of the 56 patients taking metformin had stopped metformin; 26 of the 28 patients on dipeptidyl peptidase-4 (DPP-4) inhibitors discontinued DPP-4 inhibitors; all 13 patients on alpha-glucosidase inhibitors discontinued these inhibitors; all 34 patients on sulfonylureas were able to stop taking these medications; two patients stopped taking pioglitazone; all ten patients on sodium-glucose cotransporter-2 (SGLT2) inhibitors stopped taking SGLT2 inhibitors; and one patient stopped taking glucagon-like peptide-1 analogues. Conclusion The results provide evidence that daily precision nutrition guidance based on CGM, food intake data, and machine learning algorithms can benefit patients with type 2 diabetes. Adherence for 3 months to the TPN Program resulted in patients achieving a 1.9 percentage point decrease in HbA1c, a 6.1% drop in weight, a 56.9% reduction in HOMA-IR, a significant decline in glucose time below range, and, in