Two-Year Treatment With Metformin During Puberty Does Not Preserve β-Cell Function in Youth With Obesity

Abstract

Context: Youth-onset type 2 diabetes is a disease of pubertal onset, associated with additional burden of pubertal insulin resistance on the β-cell.

Objective: Evaluate the impact of metformin treatment during puberty, a critical window of cardiometabolic change, on insulin sensitivity (Si) and compensatory β-cell response in youth with obesity.

Setting: Pediatric academic hospital clinical translational research center.

Participants: Healthy youth in early puberty [Tanner stage (T) 2-3] with normoglycemia and obesity (n = 44).

Intervention: Double-blinded placebo-control trial of metformin during puberty (until T5).

Main outcome measures: Insulin sensitivity (Si), insulin response [acute insulin response to glucose (AIRg)], and disposition index (DI), estimated from frequently sampled intravenous glucose tolerance testing; body fat (dual X-ray absorptiometry); and other laboratory parameters, collected at baseline, T4, and T5. Placebo-subtracted treatment effect was calculated using linear mixed models.

Results: At T5, metformin treatment, adjusting for sex, race, and baseline value, was associated with improved BMI z-score (-0.44 ± 0.16, P = 0.02), percentage body fat (%body fat; -3.4 ± 1.2%, P = 0.06), and waist circumference (-11.3 ± 3.2cm, P = 0.003). There were no significant treatment effects at T5 on Si or secretion: Si (0.85 ± 0.87 × 10-4/min-1/μIU/mL, P = 0.34), AIRg (-259 ± 386 μIU/mL, P = 0.51), or DI (508 ± 802 × 10-4/min-1, P = 0.53). High baseline DI predicted longitudinal decline in DI.

Conclusions: Two years of metformin treatment in obese youth during puberty improved BMI and body fat, but not Si or β-cell function. Of note, high DI in early puberty may be predictive of later decline in DI. Further studies are needed to develop strategies for preservation of β-cell function in youth at risk for type 2 diabetes.

Trial registration: ClinicalTrials.gov NCT01775813.

Keywords: beta-cell function; insulin sensitivity; metformin; obesity; puberty.

Figure 1.

Trajectories of S_i, AIR_g, leptin, BMIz, leptin, hsCRP, and waist circumference over time in the 2 treatment groups. *P < 0.01, **P < 0.001.

Figure 2.

(A) Relationship of S_i and β-cell response by treatment group. The black dot represents the mean for the combined groups at baseline, with the surrounding egg representing the SD around the mean. The dotted lines represent change from baseline (●   ) to T4 (▲) and from T4 to T5 (■). A change that remains on or goes above the baseline (ie, remains on or to the right of the black curve) would represent maintenance or improvement in β-cell function in relation to S_i. At T4, it appears that β-cell function was reasonably well-maintained in the metformin group, while it fell in the placebo group to the point that it is just at the edge of the baseline error. However, by T5 β-cell function had fallen from baseline in both groups, although remains within the baseline error margin. Importantly, there is significant overlap in the error surrounding the treatment group changes, showing the extreme variability of response. (B) Relationship between baseline DI and change in DI in the metformin (dotted line) and placebo (dashed line) groups, demonstrating that a higher baseline DI was associated with a greater fall in DI during the study. There was no significant treatment group difference in the relationship between baseline DI and ΔDI.