Predicting post one-year durability of glucose-lowering monotherapies in patients with newly-diagnosed type 2 diabetes mellitus - A MASTERMIND precision medicine approach (UKPDS 87)

Abstract

Aims: Predicting likely durability of glucose-lowering therapies for people with type 2 diabetes (T2D) could help inform individualised therapeutic choices.

Methods: We used data from UKPDS patients with newly-diagnosed T2D randomised to first-line glucose-lowering monotherapy with chlorpropamide-glibenclamide-basal insulin or metformin. In 2339 participants who achieved one-year HbA_1c values <7.5% (<59 mmol/mol)-we assessed relationships between one-year characteristics and time to monotherapy-failure (HbA_1c ≥ 7.5% or requiring second-line therapy). Model validation was performed using bootstrap sampling.

Results: Follow-up was median (IQR) 11.0 (8.0-14.0) years. Monotherapy-failure occurred in 72%-82%-75% and 79% for those randomised to chlorpropamide-glibenclamide-basal insulin or metformin respectively-after median 4.5 (3.0-6.6)-3.7 (2.6-5.6)-4.2 (2.7-6.5) and 3.8 (2.6- 5.2) years. Time-to-monotherapy-failure was predicted primarily by HbA_1c and BMI values-with other risk factors varying by type of monotherapy-with predictions to within ±2.5 years for 55%-60%-56% and 57% of the chlorpropamide-glibenclamide-basal insulin and metformin monotherapy cohorts respectively.

Conclusions: Post one-year glycaemic durability can be predicted robustly in individuals with newly-diagnosed T2D who achieve HbA_{1c values} < 7.5% one year after commencing traditional monotherapies. Such information could be used to help guide glycaemic management for individual patients.

Keywords: Durability; Glucose-lowering agents; Modelling; Monotherapy failure; Precision medicine.