Weight regain and cardiometabolic effects after withdrawal of semaglutide: The STEP 1 trial extension

Abstract Aim To explore changes in body weight and cardiometabolic risk factors after treatment withdrawal in the STEP 1 trial extension. Materials and Methods STEP 1 (NCT03548935) randomized 1961 adults with a body mass index ≥ 30 kg/m2 (or ≥ 27 kg/m2 with ≥ 1 weight‐related co‐morbidity) without diabetes to 68 weeks of once‐weekly subcutaneous semaglutide 2.4 mg (including 16 weeks of dose escalation) or placebo, as an adjunct to lifestyle intervention. At week 68, treatments (including lifestyle intervention) were discontinued. An off‐treatment extension assessed for a further year a representative subset of participants who had completed 68 weeks of treatment. This subset comprised all eligible participants from any site in Canada, Germany and the UK, and sites in the United States and Japan with the highest main phase recruitment. All analyses in the extension were exploratory. Results Extension analyses included 327 participants. From week 0 to week 68, mean weight loss was 17.3% (SD: 9.3%) with semaglutide and 2.0% (SD: 6.1%) with placebo. Following treatment withdrawal, semaglutide and placebo participants regained 11.6 (SD: 7.7) and 1.9 (SD: 4.8) percentage points of lost weight, respectively, by week 120, resulting in net losses of 5.6% (SD: 8.9%) and 0.1% (SD: 5.8%), respectively, from week 0 to week 120. Cardiometabolic improvements seen from week 0 to week 68 with semaglutide reverted towards baseline at week 120 for most variables. Conclusions One year after withdrawal of once‐weekly subcutaneous semaglutide 2.4 mg and lifestyle intervention, participants regained two‐thirds of their prior weight loss, with similar changes in cardiometabolic variables. Findings confirm the chronicity of obesity and suggest ongoing treatment is required to maintain improvements in weight and health.

participants who had completed 68 weeks of treatment. This subset comprised all eligible participants from any site in Canada, Germany and the UK, and sites in the United States and Japan with the highest main phase recruitment. All analyses in the extension were exploratory.
Conclusions: One year after withdrawal of once-weekly subcutaneous semaglutide 2.4 mg and lifestyle intervention, participants regained two-thirds of their prior weight loss, with similar changes in cardiometabolic variables. Findings confirm the chronicity of obesity and suggest ongoing treatment is required to maintain improvements in weight and health.

K E Y W O R D S
antiobesity drug, clinical trial, GLP-1 analogue, obesity therapy, phase III study, weight control

| INTRODUCTION
Obesity is a highly prevalent, complex, chronic disease 1-4 associated with cardiometabolic complications, including type 2 diabetes, hypertension, dyslipidaemia and cardiovascular disease. [4][5][6] Obesity also leads to a wide range of other health problems, is associated with substantial socioeconomic burden and was estimated to cause 5 million deaths globally in 2019. [2][3][4]7 Weight loss activates compensatory biological changes that prevent the maintenance of long-term weight loss 8,9 ; weight regain is common. 8,10 In people with obesity, pharmacotherapy is indicated as an adjunct to lifestyle intervention for chronic weight management and can help with achieving and maintaining weight loss. 2,11,12 Long-term obesity pharmacotherapy may be required for weight maintenance, as cessation of pharmacological treatment is frequently followed by weight regain, even with continued lifestyle intervention. [13][14][15] The present observational study examined changes in body weight and cardiometabolic risk factors over 52 weeks following cessation of once-weekly subcutaneous (s.c.) semaglutide 2.4 mg (or placebo) and lifestyle intervention in participants who had completed 68 weeks of initial treatment in the randomized, placebo-controlled Semaglutide Treatment Effect in People with obesity (STEP) 1 trial. 16 Semaglutide is a glucagon-like peptide-1 (GLP-1) analogue recently approved by the US Food and Drug Administration, Health Canada and the UK Medicines and Healthcare products Regulatory Agency for chronic weight management, as an adjunct to lifestyle intervention, at a once-weekly s.c. dose of 2.4 mg in adults with overweight (with ≥ 1 weight-related condition) or obesity. [17][18][19] The efficacy and safety of semaglutide 2.4 mg for weight management have been investigated in the global phase III STEP programme. 20 In the main phase of the STEP 1 trial, 68 weeks of treatment with semaglutide plus lifestyle intervention in adults with overweight/ obesity produced clinically meaningful reductions in body weight, as well as improvements in cardiometabolic risk factors, and physical functioning. 16 A prior trial (STEP 4) assessed the effects of semaglutide withdrawal, with participants initially receiving 20 weeks of semaglutide 2.4 mg treatment during a run-in period, followed by randomization to continued semaglutide treatment or withdrawal (switch to placebo) for an additional 48 weeks, with lifestyle intervention throughout. 13 The present STEP 1 extension study complements STEP 4 by exploring post-treatment changes in body weight and cardiometabolic risk factors following a longer (68-week) initial treatment period with semaglutide or placebo, and in the absence of active lifestyle intervention support during the 1-year off-treatment follow-up period.

| MATERIALS AND METHODS
2.1 | Trial design and participants STEP 1 (NCT03548935) was a randomized, double-blind, placebocontrolled trial conducted at 129 sites across 16 countries. The design and eligibility criteria have previously been published. 16 Participants were adults (aged ≥ 18 years) with a body mass index (BMI) of 30 kg/m 2 or higher, or of 27 kg/m 2 or higher with at least one weight-related co-morbidity (hypertension, dyslipidaemia, obstructive sleep apnoea or cardiovascular disease), and a history of at least one self-reported unsuccessful dietary effort to lose weight. Key exclusion criteria were type 1 or 2 diabetes and obesity pharmacotherapy 90 days or less before enrolment. Participants were randomized to 68 weeks of treatment with once-weekly s.c. semaglutide 2.4 mg (n = 1306) or placebo (n = 655) (2:1), plus lifestyle intervention. The lifestyle intervention consisted of counselling every 4 weeks on diet (500 kcal deficit per day relative to total estimated energy expenditure at randomization) and physical activity (150 minutes per week).
Semaglutide was initiated at 0.25 mg, with escalation every 4 weeks until the 2.4 mg target dose was reached ( Figure S1). At week 68 (the end of the treatment period), participants were withdrawn from treatment (including lifestyle intervention) and followed for 7 weeks until week 75.
The STEP 1 extension followed a subset of participants for an additional 45 weeks (a total of 52 weeks off-treatment) until the end-of-trial visit at week 120 ( Figure S1). The extension was offered in five selected countries (Canada, Germany, Japan, the UK and the United States) that were representative of the global trial population and aimed to include approximately 300 participants. In Canada, Germany and the UK, all sites with subjects interested in participating in the extension were included (Canada: six sites; Germany: 13 sites; UK: 10 sites). In the United States and Japan, the sites included in the extension (United States: five sites; Japan: three sites) were objectively selected based on those with the highest recruitment in the STEP 1 main phase (using an assumed on-treatment completion rate

| Procedures
The first two visits of the off-treatment observational extension phase Participants unable to attend site visits because of the COVID-19 pandemic were offered telephone visits as an alternative. Analyses were limited to data recorded during onsite visits; data obtained from telephone visits were self-reported and therefore not evaluated.

| Outcomes
The extension addressed two exploratory objectives: (1)  Post hoc analyses explored changes in body weight from baseline to week 120 and in a variety of participant subgroups (from baseline to week 68 and week 120, and from week 68 to week 120), and the proportion of participants with 5% or higher weight loss from baseline at week 120.

| Statistical analysis
No power calculations were performed to determine the sample size.
All extension phase analyses were exploratory and performed in the extension analysis set (ExAS), which included all participants eligible for the extension who attended at least one visit on week 75, 80, 104 or 120.
Endpoints addressing the first exploratory objective were analysed descriptively using observed data from the in-trial period (the time from randomization to last contact with a trial site) for the ExAS. No estimands were specified.
F I G U R E 1 Change from baseline in body weight by week for A, All participants in the ExAS, B, Participants in the semaglutide arm, grouped by categorical weight loss from week 0 to week 68, C, Participants not using obesity pharmacotherapy during the extension † , and D, Participants in the semaglutide arm for the full ExAS and the subgroup with prediabetes resolution at week 68 and subsequent reversion by week 120 ‡ . † Participants who did not use obesity pharmacotherapies (investigator-assessed) during the extension phase. ‡ Participants who shifted from prediabetes at baseline to normoglycaemia at week 68 to prediabetes at week 120. Glycaemic category was determined from HbA1c assessments, as per American Diabetes Association HbA1c criteria. 21 Normoglycaemia was defined by HbA1c < 5.7% (< 39 mmol/mol); prediabetes was defined by HbA1c 5.7%-6.4% (39-47 mmol/mol). Data are observed mean changes from baseline (± standard error) for the ExAS from the in-trial period. The dashed vertical line at week 68 indicates the end of the main phase and start of the off-treatment extension phase. Numbers shown in the lower panels are participants contributing to the mean. ExAS, extension analysis set Endpoints addressing the second exploratory objective were analysed descriptively and statistically using data from the in-trial period for the ExAS. Results for the ExAS were compared with results previously reported for the full analysis set (FAS), 16  Percentage changes in body weight and selected cardiometabolic risk factors were calculated relative to the baseline value of the variable.

| Study participants
From September 2019 to April 2020, 336 participants from the main phase were screened for the extension and 333 entered the extension, including 232 participants who received semaglutide during the main phase (referred to as the "semaglutide arm" hereafter) and 101 who received placebo during the main phase ("placebo arm" hereafter). In total, 327 participants (98.2%) were included in the ExAS

| Change in body weight and BMI
During the main treatment phase (from baseline [week 0] to week 68), semaglutide reduced body weight more than placebo ( Figure 1A and Tables 2 and S1); observed mean weight loss was 17.3% (standard deviation [SD]: 9.3%) with semaglutide versus 2.0% (SD: 6.1%) with placebo (Table S1). After treatment withdrawal, body weight regain was observed in both the semaglutide and placebo arms ( Figure 1A and Tables 2 and S1). Participants regained a mean of 11.6 percentage points (SD: 7.7) of body weight in the semaglutide arm versus 1.9 percentage points (SD: 4.8) in the placebo arm (Table S1). The net mean body weight loss over the full duration of the main treatment phase and off-treatment extension phase (from week 0 to week 120) was  Changes in BMI during the main treatment phase and extension phase were consistent with changes in body weight and are reported in Tables 2, S1 and S2.

| Change in body weight in participant subgroups
Changes from baseline in body weight at week 120 differed according to the weight lost by week 68. Subgroups with greater weight losses from week 0 to week 68 tended to have numerically greater weight regains from week 68 to week 120, but maintained numerically greater net weight losses from week 0 to week 120 ( Figure 1B and Table S3). Participants in the semaglutide arm with weight losses (i.e. < 5%, ≥ 5%-< 10%, ≥ 10%-< 15%, ≥ 15%-< 20% and ≥ 20%) from week 0 to week 68 had net weight changes from week 0 to week 120 of 4.2% (SD: 6.2%), -0.7% (SD: 5.9%), -1.8% (SD: 5.0%), -5.5% (SD: 7.2%) and -12.1% (SD: 8.9%), respectively. In absolute terms, respective changes from week 0 to week 120 for these subgroups  (Table S4). In the semaglutide arm, mean changes from week 0 to week 120 were numerically greater for women versus men; higher versus lower baseline BMI categories; and participants with normoglycaemia versus prediabetes at baseline. Mean changes in body weight from week 0 to week 120 in the semaglutide arm were similar across subgroups defined by baseline age tertiles (Table S4).
Changes in body weight from week 0 to week 68 and from week 68 to week 120 for these subgroups are reported in Table S4.
When obesity pharmacotherapy users were excluded, changes in body weight ( Figure 1C) were similar to the full ExAS ( Figure 1A).
Among the subgroup not using obesity pharmacotherapy, mean changes in body weight from week 0 to week 120 were -5.6% (SD: 8.9%) and -0.3% (SD: 5.3%) in the semaglutide and placebo arms, respectively.

| Change in cardiometabolic risk factors
During treatment, greater improvements in cardiovascular risk factors were observed from week 0 to week 68 with semaglutide than placebo (Tables 2 and S1), which tended to be slightly larger in the ExAS than in the FAS (Table S2). After treatment withdrawal, mean SBP and F I G U R E 3 Change in glycaemic category from baseline (week 0) to weeks 68 and 120. Data are observed proportions (%) of participants for the extension analysis set from the in-trial period. Proportions are based on participants with an observation at the visit. Glycaemic category was determined from HbA1c assessments, as per American Diabetes Association HbA1c criteria. 21 Normoglycaemia was defined by HbA1c < 5.7% (< 39 mmol/mol); prediabetes was defined by HbA1c 5.7%-6.4% (39-47 mmol/mol); diabetes was defined by HbA1c ≥ 6.5% (≥ 48 mmol/mol) DBP increased in both treatment arms, reverting to baseline levels by week 120 (Figure 2 and Tables 2 and S1). CRP and lipids increased from week 68 to week 120 in the semaglutide arm, but remained improved relative to the placebo arm for CRP, HDL cholesterol, VLDL cholesterol and triglycerides at week 120 (Figures 2 and S3 and   Tables 2 and S1). Improvements were observed in some cardiovascular risk factors from baseline to week 120 (Figures 2 and S3 and During treatment, a greater decrease in HbA1c was observed from week 0 to week 68 with semaglutide than placebo (Tables 2 and   S1), which was slightly larger in the ExAS than in the FAS (Table S2).
After treatment withdrawal, increases in mean HbA1c were observed in both treatment arms. Although the magnitude of increase from week 68 to week 120 was greater in the semaglutide arm than in the placebo arm (Table S1) Table 2).
In the semaglutide group, greater magnitudes of weight loss from week 0 to week 68 tended to be associated with the most favourable changes in cardiometabolic risk factors from baseline at week 120 (Tables S5 and S6).

| Change in glycaemic category
Among participants with prediabetes at baseline, numerically more participants reverted to normoglycaemia at week 68 with semaglutide than placebo (93.6% vs. 41.5%; Figure 3). After treatment withdrawal, improvements deteriorated in both arms. Although the deterioration was larger in the semaglutide arm, a relative improvement was maintained versus the placebo arm at week 120, with reversion to normoglycaemia in 43.3% versus 34.0% of participants with baseline prediabetes in the semaglutide and placebo arms, respectively.
Changes in body weight in the subgroup of participants who shifted from prediabetes at baseline to normoglycaemia at week 68 and then reverted to prediabetes by week 120 are shown in Figure 1D and Table S4.
Few participants with normoglycaemia at baseline had progression to prediabetes at week 68 or week 120, in either of the semaglutide or placebo arms (Figure 3).

| DISCUSSION
In STEP 1, 68 weeks of treatment with once-weekly s.c. semaglutide 2.4 mg plus lifestyle intervention provided significant, clinically relevant reductions in body weight versus placebo among adults with overweight/obesity. 16 After withdrawal of semaglutide and structured lifestyle intervention, participants regained a mean of two-thirds of their prior weight loss in the 1-year off-treatment extension phase; weight regain continued until the end of follow-up (week 120). However, some treatment effects were sustained and weight remained 5.6% below baseline in the semaglutide arm. Almost half of the participants in the semaglutide arm (48.2%) still had clinically meaningful weight loss of 5% or more from baseline at week 120, although this proportion represented a substantial fall from that originally achieved at the end of 68 weeks of treatment (86.4%). 16 Subgroup analyses suggested that participants in the semaglutide arm with greater weight loss during the 68-week treatment period tended to have greater regain in body weight after semaglutide withdrawal, but ultimately retained greater weight loss at week 120 versus subgroups who lost less weight during the 68-week treatment period.
Weight regain has previously been observed following withdrawal of obesity pharmacotherapies, including orlistat and lorcaserin-and also semaglutide, as shown in STEP 4-despite continued lifestyle intervention. [13][14][15] Taken together, these findings and those of the present study confirm the chronicity of obesity and highlight the importance of maintaining long-term pharmacological treatment for weight management in people with obesity.
In the STEP 1 extension, weight regain was comparatively rapid following semaglutide withdrawal compared with that seen in other obesity pharmacotherapy withdrawal trials, including after semaglutide withdrawal in STEP 4. [13][14][15] This may well relate to participants who received semaglutide in STEP 1 having achieved greater weight loss prior to withdrawal than in other trials, and thus having greater potential for regain, driven by physiological and behavioural factors. 22 patients with overweight/obesity and type 2 diabetes in the STEP 2 trial, 24 relative to those seen during the main treatment period of STEP 1, which excluded patients with type 2 diabetes. 16 We observed residual benefits in some cardiovascular risk factors at week 120 compared with baseline. Previously, a post hoc analysis of the Look AHEAD trial explored the effects of weight regain on cardiometabolic risk factors among people with overweight/obesity and type 2 diabetes who received intensive lifestyle intervention. 25 In that analysis, large initial weight losses followed by partial or full regain by year 4 were associated with a sustained benefit on HbA1c at year 4 relative to that seen in participants with smaller or no initial weight loss. 25  The key limitations of the extension were the relatively small sample size compared with the main STEP 1 trial population and the selection of sites based on those with the highest recruitment in the STEP 1 main phase, which could introduce an element of selection bias.

PEER REVIEW
The peer review history for this article is available at https://publons. com/publon/10.1111/dom.14725.

DATA AVAILABILITY STATEMENT
Data will be shared with bona fide researchers who submit a research proposal approved by the independent review board. Individual participant data will be shared in data sets in a de-identified and anonymized format. Data will be made available after research completion and approval of the product and product use in the European Union and the USA. Information about data access request proposals can be found at novonordisk-trials.com.