A randomized controlled trial of pharmacist-led therapeutic carbohydrate and energy restriction in type 2 diabetes

Abstract

Type 2 diabetes can be treated, and sometimes reversed, with dietary interventions; however, strategies to implement these interventions while addressing medication changes are lacking. We conducted a 12-week pragmatic, community-based parallel-group randomized controlled trial (ClinicalTrials.gov: NCT03181165) evaluating the effect of a low-carbohydrate (<50 g), energy-restricted diet (~850-1100 kcal/day; Pharm-TCR; n = 98) compared to treatment-as-usual (TAU; n = 90), delivered by community pharmacists, on glucose-lowering medication use, cardiometabolic health, and health-related quality of life. The Pharm-TCR intervention was effective in reducing the need for glucose-lowering medications through complete discontinuation of medications (35.7%; n = 35 vs. 0%; n = 0 in TAU; p < 0.0001) and reduced medication effect score compared to TAU. These reductions occurred concurrently with clinically meaningful improvements in hemoglobin A1C, anthropometrics, blood pressure, and triglycerides (all p < 0.0001). These data indicate community pharmacists are a viable and innovative option for implementing short-term nutritional interventions for people with type 2 diabetes, particularly when medication management is a safety concern.

Fig. 1. Trial CONSORT flow diagram.

Pharm-TCR Pharmacist-led therapeutic carbohydrate restriction, TAU Treatment-as-usual, ITT Intention-to-treat. Created with BioRender.com.

Fig. 2. Descriptive analysis of weekly data collected in the Pharm-TCR group.

Data are weekly effect estimates for changes from baseline (Week 0; gray line) with confidence intervals in the Pharm-TCR group for (a) medication effect score (MES); (b) body mass index; (c) waist circumference; (d) body fat percentage; (e) systolic blood pressure; and (f) diastolic blood pressure. Values are effect estimates for adjusted mean change from baseline in a–f. Bias-adjusted and accelerated confidence intervals derived from non-parametric bootstrap analysis are presented in panel a. Error bars for panels b–f represent 95% confidence intervals. Data are based on participants for which baseline data were collected (n = 92) except for waist circumference (n = 90) and body fat percentage (n = 91). Source data are provided as a source data file.