Two-year effects of semaglutide in adults with overweight or obesity: the STEP 5 trial

Abstract

The STEP 5 trial assessed the efficacy and safety of once-weekly subcutaneous semaglutide 2.4 mg versus placebo (both plus behavioral intervention) for long-term treatment of adults with obesity, or overweight with at least one weight-related comorbidity, without diabetes. The co-primary endpoints were the percentage change in body weight and achievement of weight loss of ≥5% at week 104. Efficacy was assessed among all randomized participants regardless of treatment discontinuation or rescue intervention. From 5 October 2018 to 1 February 2019, 304 participants were randomly assigned to semaglutide 2.4 mg (n = 152) or placebo (n = 152), 92.8% of whom completed the trial (attended the end-of-trial safety visit). Most participants were female (236 (77.6%)) and white (283 (93.1%)), with a mean (s.d.) age of 47.3 (11.0) years, body mass index of 38.5 (6.9) kg m^-2 and weight of 106.0 (22.0) kg. The mean change in body weight from baseline to week 104 was -15.2% in the semaglutide group (n = 152) versus -2.6% with placebo (n = 152), for an estimated treatment difference of -12.6 %-points (95% confidence interval, -15.3 to -9.8; P < 0.0001). More participants in the semaglutide group than in the placebo group achieved weight loss ≥5% from baseline at week 104 (77.1% versus 34.4%; P < 0.0001). Gastrointestinal adverse events, mostly mild-to-moderate, were reported more often with semaglutide than with placebo (82.2% versus 53.9%). In summary, in adults with overweight (with at least one weight-related comorbidity) or obesity, semaglutide treatment led to substantial, sustained weight loss over 104 weeks versus placebo. NCT03693430.

Fig. 1. Flow chart of trial participants in the STEP 5 clinical trial.

s.c., subcutaneous.

Fig. 2. Comparison of body weight parameters for semaglutide versus placebo (co-primary endpoints; treatment policy estimand).

a, Observed mean percentage change from baseline in body weight over time for participants in the full analysis set during the in-trial observation period (error bars are standard error of the mean; numbers below the panels are the number of participants contributing to the mean) and estimated treatment difference for the percentage change from baseline to week 104 in body weight based on the treatment policy estimand. b, Observed proportions of participants and OR for achieving weight loss of at least 5% from baseline at week 104 in the full analysis set during the in-trial observation period, based on the treatment policy estimand. *Estimated means in percent are from the primary analysis. The in-trial observation period was the time from random assignment to last contact with a trial site, regardless of treatment discontinuation or rescue intervention. The treatment policy estimand assesses treatment effect regardless of treatment discontinuation or rescue intervention; see Extended Data Fig. 6 for corresponding data for the trial product estimand (which assesses treatment effect assuming all participants adhered to treatment and did not receive rescue intervention). The change in body weight analysis was conducted with the use of the analysis-of-covariance method, with randomized treatment as a factor and baseline body weight as a covariate. The achievement of at least 5% weight loss analysis was conducted with the use of logistic regression, with the same factor and covariate. A multiple imputation approach was used for missing data. The results were accompanied by two-sided 95% CIs and corresponding P values (significance defined as P < 0.05). As co-primary endpoints, the analyses were controlled for multiple comparisons.

Extended Data Fig. 1. Trial design for STEP 5 clinical study.

s.c., subcutaneous.

Extended Data Fig. 2. Body weight (kg) by week.

Observed mean body weight (kg) over time for participants in the full analysis set during the in-trial observation period (from randomization to last contact with trial site, regardless of treatment discontinuation or rescue intervention). Error bars are standard error of the mean. Numbers below the panels are the number of participants contributing to the mean.

Extended Data Fig. 3. Cumulative distribution plot of change from baseline to week 104 in body weight.

(a, b) Cumulative distribution plot of observed percentage change from baseline over time in body weight for participants in the full analysis set during the in-trial observation period* (a) and on-treatment observation period^† (b). *From randomization to last contact with trial site, regardless of treatment discontinuation or rescue intervention. ^†During treatment with trial product (any dose of trial medication administered within the previous 2 weeks (that is, any period of temporary treatment interruption with trial product was excluded)).

Extended Data Fig. 4. Comparison of change from baseline by week for selected cardiometabolic endpoints for semaglutide versus placebo.

(a-d) Observed mean percentage change from baseline over time for participants in the full analysis set during the in-trial observation period in waist circumference (a), systolic blood pressure (b), diastolic blood pressure (c), and HbA_1c (d). Error bars are standard error of the mean; numbers below the panels are the number of participants contributing to the mean.

Extended Data Fig. 5. Shift from baseline to week 104 in glycemic status.

(a-d) Observed data for participants in the full analysis set treated with semaglutide 2.4 mg (a, c) or placebo (b, d) during the in-trial period. As illustrated by the gray shading, the week 104 bars present results at this time point among the subgroups of participants with baseline prediabetes (a and b) or baseline normoglycemia (c and d). Glycemic category was determined by investigators on the basis of available information (for example, medical records, concomitant medication, and blood glucose variables) and in accordance with American Diabetes Association criteria, which for prediabetes includes fasting plasma glucose levels of 100 mg/dL (5.6 mmol/L) to 125 mg/dL (6.9 mmol/L) or HbA_1c levels of 5.7–6.4% (39–47 mmol/L), and for type 2 diabetes includes fasting plasma glucose levels of ≥126 mg/dL (7.0 mmol/L) or HbA_1c levels ≥6.5% (48 mmol/L). *Number of participants in the full analysis set. ^†Number of participants with prediabetes (a and b) or normoglycemia (c and d) at baseline and evaluable data at week 104.

Extended Data Fig. 6. Comparison of body weight parameters for semaglutide versus placebo (trial product estimand).

(a) Observed mean percentage change from baseline in body weight over time for participants in the full analysis set during the on-treatment observation period (error bars are standard error of the mean; numbers below the panels are the number of participants contributing to the mean) and estimated treatment difference for the percentage change from baseline to week 104 in body weight based on the trial product estimand. (b) Observed proportions of participants and odds ratio for achieving weight loss of at least 5% from baseline at week 104 in the full analysis set during the on-treatment observation period, based on the trial product estimand. *Estimated means in percent. A time point is considered as on treatment if any dose of trial product has been administered within the previous 14 days. The trial product estimand assesses treatment effect assuming all participants adhered to treatment and did not receive rescue intervention. CI, confidence interval; ETD, estimated treatment difference.

Extended Data Fig. 7. Prevalence and duration of gastrointestinal events by severity.

(a-d) The proportion of participants receiving semaglutide or placebo who reported nausea (a), diarrhea (b), constipation (c), or vomiting (d) events classed as mild, moderate, or severe over the course of the treatment period. Data are from the on-treatment observation period (during treatment with trial product [any dose of trial medication administered within the previous 49 days (that is, any period of temporary treatment interruption with trial product was excluded)). Adverse events were classified by severity as mild (easily tolerated, causing minimal discomfort, and not interfering with everyday activities), moderate (causes sufficient discomfort and interferes with normal everyday activities), or severe (prevents normal everyday activities).