A one-week reduced-carbohydrate diet to mitigate iatrogenic peripheral hyperinsulinemia does not improve insulin sensitivity or endothelial function in a randomized, crossover trial in patients with type 1 diabetes

Abstract

Background: Iatrogenic peripheral hyperinsulinemia, resulting from peripheral insulin administration in type 1 diabetes, may increase insulin resistance and impair endothelial function. We hypothesized that lowering iatrogenic peripheral hyperinsulinemia via a one-week, reduced-carbohydrate diet (RCD) would improve insulin sensitivity and endothelial function compared with an isocaloric standard carbohydrate diet (SCD).

Methods: In this randomized, single-blinded, crossover trial, we studied 12 adults with type 1 diabetes. Participants completed both a one-week RCD and a one-week SCD, separated by a three-week washout. After each intervention, we measured insulin sensitivity using a hyperinsulinemic-euglycemic clamp and assessed endothelial function via brachial-artery flow-mediated vasodilation (FMD).

Results: The RCD reduced total daily insulin doses by 16% compared with the SCD. Despite this reduction, insulin sensitivity did not improve (median glucose infusion rates: RCD 8.1 mg/kg FFM/min [IQR 6.7-10.1] vs. SCD 8.6 mg/kg FFM/min [7.0-11.0], p = 0.47). Similarly, endothelial function did not differ significantly (FMD after RCD 7.50% [3.25-15.5] vs. SCD 9.81% [4.96-14.3], p = 0.91). Although higher insulin doses correlated with lower insulin sensitivity under both conditions, lowering insulin dose through the RCD alone did not yield measurable improvements.

Conclusions: Although a one-week RCD significantly lowered insulin requirements, it failed to enhance insulin sensitivity or endothelial function in adults with type 1 diabetes. These findings underscore the complex and dynamic relationship between insulin exposure and cardiometabolic health. Similar basal overnight insulin delivery may have masked potential benefits by the time of testing, highlighting the need for further studies to refine strategies aimed at mitigating hyperinsulinemia's adverse effects.

Trial registration: ClinicalTrials.gov NCT04118374.

Keywords: Endothelial dysfunction; Endothelial function; Hyperinsulinemia; Insulin resistance; Insulin sensitivity; Reduced-carbohydrate diet; Type 1 diabetes.

Fig. 1

Study design and diet macronutrients. A Macronutrient composition of isocaloric, standard carbohydrate diet (SCD) B Macronutrient composition of isocaloric reduced carbohydrate diet (RCD). Percentages indicate percent of caloric intake from each macronutrient. The study design anticipated total daily dose of insulin (TDD_insulin) would be 0.25 units/kg/day lower on RCD than SCD. C Schematic of crossover study design. T1DM A = participants undergoing treatment order (A) T1DM B = participants undergoing treatment order (B) CGM = continuous glucose monitoring. Solid shaded squares at the end of each diet intervention indicates the study team supplied all food in the final 24 h prior to each study. D Schematic for research visit procedures. FMD = flow mediated dilation. AID = automated insulin delivery. SQ = subcutaneous

Fig. 2

Insulin dose and glycemic control. Total daily dose (TDD) of insulin (A-B) and continuous glucose monitor (CGM) data (C-D) during the week and 24 h preceding each experiment. Column scatter plots show the medians of mean daily insulin dose or mean CGM glucose levels during reduced carbohydrate diet (RCD) and standard carbohydrate diet (SCD) interventions, along with the interquartile range

Fig. 3

Endothelial function following interventions. Endothelium-dependent flow-mediated dilation (FMD, A–C) and endothelium-independent nitroglycerin-mediated dilation (NMD, D–F) after the reduced-carbohydrate diet (RCD) and standard-carbohydrate diet (SCD) interventions. Plots A and D show individual FMD and NMD data, intervention medians, and the interquartile ranges. Plots B and E show within-participant changes in FMD and NMD between interventions. Plots C and F show within-participant differences in FMD and NMD between interventions (SCD minus RCD) for each individual, along with median and interquartile ranges for these differences. Per protocol, two participants did not receive nitroglycerin following the SCD because of low systolic blood pressure

Fig. 4

Hormone and glucose parameters. Plasma concentrations of basal insulin lispro (A), insulin lispro during the clamp (B), glucose (D), C-peptide (E), basal estradiol (F), glucagon (G), cortisol (H), epinephrine (I), and norepinephrine (J), as well as insulin clearance (C) during hyperinsulinemic, euglycemic clamp studies. Median concentrations are shown as circles for the reduced-carbohydrate diet (RCD) and squares for the standard-carbohydrate diet (SCD). Insulin clearance was calculated by dividing the intravenous insulin infusion rate by steady-state insulin concentrations during the last 30 min of the clamp (C). Error bars represent the interquartile range

Fig. 5

Metabolite responses. Blood concentrations of non-esterified fatty acids (A), glycerol (B), β-hydroxybutyrate (C), β-hydroxybutyrate during the basal period (D), lactate (E), and alanine (F). Median concentrations are shown as circles for reduced carbohydrate diet (RCD) studies and as squares for standard carbohydrate diet (SCD) studies. Error bars represent the interquartile range

Fig. 6

Insulin sensitivity and insulin dose relationship. Glucose infusion rate (GIR) during the final 30 min of hyperinsulinemic, euglycemic clamp studies parameterizing insulin sensitivity following reduced-carbohydrate diet (RCD) and standard-carbohydrate diet (SCD) interventions. Plot A shows individual GIR data, intervention medians, and the interquartile ranges. Plot B shows within-participant changes in GIR between interventions. Plot C shows within-participant changes in GIR between interventions (SCD minus RCD) for each individual, along with median and interquartile ranges for these differences. Plot D shows the relationship between the total daily dose of insulin (averaged over the seven days of each intervention) and GIR during the RCD and SCD interventions. Points marked “R” represent the RCD study and points marked “S” represent the SCD study, with lines connecting each participant’s two studies