Description

Obesity is now epidemic, and as a consequence, the use of weight loss medications and surgery to manage obesity is increasing. Weight loss surgery is associated with significant bone loss, concerning during the adolescent years of peak bone accrual. With the increasing use of weight loss medications, particularly glucagon-like-peptide 1 receptor agonists (GLP-1 RAs), in adolescents, it is essential to determine whether weight loss following use of these medications is associated with detrimental effects on skeletal health. Data from the literature are conflicting following use of semaglutide (a commonly used GLP-1 RA) in adults with obesity. Whether GLP-1 RA use preserves bone anabolic activity during adolescence merits investigation and is the focus of the proposal. If the investigators' hypotheses prove correct and use of GLP-1 RAs preserves bone accrual and skeletal health in youth with obesity, the results of this study may favor GLP-1 RA treatment over surgery following lifestyle management. Multiple mechanisms contribute to bone loss after surgery in youth including mechanical unloading of bone from weight loss, loss of lean mass, and changes in hormones that stimulate anabolic bone activity and/or are anti-resorptive. Weight loss following use of GLP1-RAs should similarly lead to skeletal unloading, reductions in lean mass, and changes in hormones. However, GLP-1 RAs have direct bone anabolic effects, and anti-resorptive effects as demonstrated in both rodent and human studies. Thus, GLP-1 RAs might mitigate deleterious effects of weight loss on skeletal health through a direct impact on bone formation and resorption. DXA-based BMD measurements have limitations in obesity and during weight loss, being susceptible to artifactual changes from a reduction in soft tissue thickness after severe weight loss. The study will therefore use advanced 3D imaging techniques to overcome limitations of DXA in the context of the marked soft tissue changes following weight loss. The investigators' overall hypothesis is that despite marked weight loss in adolescents with obesity receiving GLP-1 RAs (specifically semaglutide, an FDA approved, commonly used GLP-1 RA in youth) over 24 months, study participants will demonstrate preservation of areal and volumetric BMD, bone geometry, structure and estimated strength, and improvements in estimated fracture risk.

Aim 1: To determine to what extent GLP-1 RA therapy alters bone density, geometry, structure, strength and load-to-strength ratio prospectively over 24 months in adolescents and young adults ages 12-21 years with obesity compared to controls of similar weight followed with lifestyle management. The investigators hypothesize that following GLP-1 RA (semaglutide) therapy vs. lifestyle management:

Hypothesis 1A: Volumetric BMD of the distal radius (a non-weight bearing site) (primary endpoint) and distal tibia (a weight-bearing site) by HRpQCT, and areal BMD of the hip and spine by DXA will be preserved.

Hypothesis 1B: Cortical and trabecular geometry and structure at the distal radius and tibia (by HR-pQCT), and bone strength estimates (by μFEA) will be preserved; load-to-strength ratio will improve.

Aim 2: To investigate novel physiologic mechanisms mediating maintenance of skeletal integrity following use of semaglutide.

Hypothesis 2: The investigators hypothesize that preservation of skeletal health despite reductions in weight, lean mass, ghrelin, insulin, oxytocin and estrone and increases in sclerostin is associated with increases in bone formation (as assessed by P1NP, a marker of bone formation) that equal or exceed those of bone resorption (as assessed by CTX, a marker of bone resorption), indicative of balanced bone turnover and net skeletal stability.

Adolescence is a critical time for bone accrual and the use of both GLP-1 RAs and MBS is increasing in youth. The study will provide novel data needed to establish whether use of GLP-1 RAs prevents the impairment in skeletal health observed following surgery in youth. Clarifying these mechanisms will identify optimal weight loss strategies in youth with obesity following lifestyle intervention.

Design: This is a non-randomized two-group parallel observational pragmatic trial. The study will recruit 120 adolescents and young adults with obesity 12-21 years old, 60 of whom are being started clinically on semaglutide therapy and 60 who will be followed with 'usual' care. Participants will be matched for BMI, sex, self-described race, age, and pubertal stage. Normally menstruating females will be studied during the first 10 days of their cycles. The baseline visit will occur before starting semaglutide therapy (for the active arm), and will be followed by visits 6, 12, 18, and 24-months after starting therapy. Controls will be followed at the same frequency. Participants will be counseled regarding lifestyle measures per protocol ('usual' care).

Analytical Plan:

Data generated will be longitudinal over 12 and 24-months. There are two study groups for the longitudinal component: those undergoing semaglutide therapy, and adolescents with obesity followed with usual care.

Analysis of Treatment Group Comparability: Demographic and baseline characteristics will be summarized by treatment group (semaglutide vs. usual care) using descriptive statistics and will be compared using a t-test or Chi-square test depending on data types.

Analysis of Aim 1:

The study is powered for analysis of the primary endpoint i.e. baseline to 24-month change in the HR-pQCT measure of total vBMD (distal radius). This analysis will include all randomized subjects according to treatment (intent to treat). The investigators will test for equivalency of the between-group treatment effects by examining if the upper limit of the 90% confidence interval of the semaglutide versus 'usual care' difference in the group-specific HR-pQCT mean change is above -30 (i.e., no less than 10% below 300 mgHA/cm3) and the lower limit is below 30 (i.e., no more than 10% above 300 mgHA/cm3). The same analysis approach as that of the primary endpoint analysis will be applied to secondary endpoints. Investigators will not adjust Type-1 error for the inference of these multiple secondary endpoints.

Method to analyze longitudinal data: Although the primary analysis endpoints are 24-month change, for longitudinal data collected at multiple time points (baseline, 12 and 24 months), as parallel analyses, investigators will utilize all available repeated measures in longitudinal general linear mixed effects models with the treatment difference at 24-month as the primary contrast of interest. The subject level intercept will be considered as random. The closest pattern of time dependency will be identified by means of exploratory longitudinal plots before fitting the model. The model will include group and time as the main effects and group x time as the interaction, and the above-mentioned equivalency test method will be applied to the model-based estimate of the treatment difference at 24-month. SAS Proc Mixed procedure with exchangeable- or more appropriate correlation structure will be used. This analysis will include all data collected on all subjects irrespective of whether the subject completed all 24 months of follow up, and follows the Institute of Medicine (IOM) suggestion for analysis of data with missing observations. We will adjust for confounders as necessary (the study will match participants for age, sex, race, and BMI). For Aim 2, the investigators will estimate within-group correlations (Pearson or Spearman as appropriate based on distribution) of 24-month change in weight, lean mass, and hormones with change in P1NP and CTX; and then examine if these correlations differ between the two treatment groups.