DOI: 10.2337/db23-44-lb ISSN: 0012-1797

44-LB: The Effect of Substituting Water for Diet Beverages on CGM Metrics in People with Type 2 Diabetes (T2D)—The Study of Drinks with Artificial Sweeteners (SODAS)

  • Endocrinology, Diabetes and Metabolism
  • Internal Medicine

Objectives: Diet beverages (DB) are the largest contributor to artificial sweetener intake in the U.S. diet. People with T2D are the highest consumers of DB, tending to consume them in place of added sugar, especially sugar-sweetened beverages. However, there is little evidence on the health impact of habitual DB consumption by people with T2D. To inform this topic we conducted a randomized trial.

Methods: The SODAS study is a two-arm parallel trial with a run-in period of 2 weeks and 24 week intervention that randomized 181 adults with T2D (HbA1c 6.5 - ≤ 8.5%), age 35+ years, who were usual consumers of commercial DB to receive and consume either: 1) A commercial DB of choice (24 oz. daily); or 2) Unflavored sparkling or flat bottled water of choice (H2O) (24 oz. daily). Outcomes include measures of clinical diabetes control and mechanisms that may alter clinical risk. Here we report preliminary secondary outcome results from continuous glucose monitor (CGM) metrics. Participants wore a blinded, Abbott Freestyle Libre CGM for 3, 2-week periods (run-in, week 12, week 24). We conducted an intent to treat analysis to assess the between-arm difference in change of CGM metric from baseline-week 12- week 24 via linear mixed regression models.

Results: 179 participants will have complete data when study finishes collection in late March 2023, and 166 had provided complete data at the time of analysis. Participants were a mean age of 58 years, and 55% were men. Time in range (TIR, % of readings 70-180 mg/dL) decreased in both arms and at week 24 the mean TIR for DB was 71.3 (SE 2.8) v. 64.3 (SE 2.8) for H2O, difference of means 7.0 (95% CI -0.7, 14.8), P=0.08). Glycemic variability (%CV) was 27.0 (SE 0.8) for DB at week 24 v. 28.7 (SE 0.8) for H2O, difference in means 1.6 (95% CI -3.8, 0.5), P=0.13.

Conclusions: Preliminarily, in people with T2D who habitually consumed DB, there were no significant differences in CGM metrics when substituting H20 intake for DB.


A. Odegaard: None. A. Bantle: None. E. R. Seaquist: Consultant; Zucara Therapeutics. A. Reikes: Other Relationship; GlaxoSmithKline plc. M. A. Pereira: None.


National Institute of Diabetes and Digestive and Kidney Diseases (R01DK117028)

More from our Archive