Reference: Obesity (Silver Spring). 2025 Jun 10 early online
Practice Point: Real-world weight loss with GLP-1s may not match trial (and media) outcomes, so manage expectations accordingly.
EBM Pearl: Selection bias tends to inflate the magnitude of effect of an intervention. Recognizing selection bias in its various forms is critical to accurate application of evidence to patient care.
"Ozempic" has become a household name and a topic of frequent discussion in pop culture. GLP-1 receptor agonists like semaglutide and tirzepatide have been described as modern-day miracle drugs and have even been compared to penicillin in terms of impact. That might be a stretch, and a new study in Obesity suggests as much. Real-world data demonstrate that these injectables may not be quite as effective for long-term weight loss as early trials indicated—largely due to tolerability issues that lead to high discontinuation rates and difficulty maintaining higher doses.
A retrospective cohort study published this month in Obesity analyzed electronic medical record data from nearly 8,000 adults (mean age 51, mean BMI 40) with overweight or obesity but without type 2 diabetes who were dispensed semaglutide (78%) or tirzepatide (12%) for weight loss. Most were privately insured, all had access to the medications, and a quarter lived in the least deprived neighborhoods per Area Deprivation Index. The primary outcome was weight loss stratified by time to discontinuation, categorized as early discontinuation (0–3 months), late discontinuation (3–12 months), or continued use at 12 months.
Early discontinuation occurred in 22% with semaglutide and 16% with tirzepatide. Late discontinuation added another 31% and 34%, respectively, leaving only about half of patients still on therapy at 12 months. Tolerability appeared to limit dose escalation, with 80% remaining on low doses throughout treatment. High maintenance doses were considered 1.7–2.4 mg for semaglutide and 10–15 mg for tirzepatide; everything else was considered low dose. While the reasons for discontinuation were not captured in these data, nausea, vomiting, diarrhea, constipation, and dyspepsia/gastritis are all fairly common with GLP-1s and tend to be dose-dependent.
Weight loss with either semaglutide or tirzepatide followed a predictable pattern: the longer the duration of use and the higher the dose, the greater the weight loss. Patients who discontinued early lost on average 3.6% of baseline weight, compared to 6.8% with late discontinuation and 11.9% for those still on the medication at 12 months. Patients who stayed on high-dose therapy for 12 months lost 13.7% with semaglutide and 18% with tirzepatide. In a subgroup of patients with prediabetes, A1C dropped by a mean of 0.4%.
These real-world numbers are lower than the weight loss demonstrated in two of the most prominent clinical trials, which, importantly, were of significantly longer duration. The STEP-1 trial of semaglutide reported an average 14.9% weight loss at 68 weeks, and the SURMOUNT-1 trial of tirzepatide reported 20.9% weight loss at 72 weeks for the highest dose. If we extrapolate the trend in real-world data to 72 weeks, we would expect increased rates of discontinuation and much lower weight loss than was reported in the clinical trials at 72 weeks. Even so, very few other medications come close to the weight loss demonstrated with GLP-1s: Real-world weight loss with off-label use of metformin, topiramate, bupropion, and zonisamide is typically in the 2%–5% range—which is a good reminder of why GLP-1s were thought to be “miracle drugs” in the first place.
Why do we think the clinical trials of GLP-1s overestimated the magnitude of effect? In addition to ideal trial conditions offering better support, tightly controlled dosing, more frequent follow-up, etc., one of the biggest reasons is selection bias. In both the STEP-1 and SURMOUNT-1 trials, adherence bias was likely a key player. However, any one of the many forms of selection bias can contribute to overestimation of treatment benefits and underestimation of harms, including:
- Adherence (healthy user) bias: Participants who adhere to trial interventions are likely to adhere to other health-related recommendations, like not smoking, exercising, following a healthy diet, and having regular check-ups and cancer screenings. These baseline healthy habits translate to better health outcomes overall regardless of the intervention.
- Volunteer (self-selection) bias: Volunteers motivated enough to seek out or respond to an enrollment invitation may be more likely to be more motivated in general. Volunteerism in many cases represents increased access to the internet or transportation as well as higher socioeconomic status, which is generally associated with better health outcomes regardless of the intervention.
- Loss to follow-up/attrition bias: Attrition bias is the flip side of adherence bias. Participants who don’t adhere to study protocols or drop out are systematically different than those who remain, and tend to have worse health outcomes regardless of the intervention. When people who drop-out or switch protocols are excluded from analysis, which is what happens in a per-protocol analysis, poorer outcomes are systematically removed and the prognostic equivalence of the study groups is compromised, almost always favoring the intervention and thus overestimating treatment effect. Differential loss to follow-up between treatment groups is also a problem.
- Referral bias: When patients are recruited from tertiary care centers, they tend to be sicker and have more complex health conditions and therefore have a greater potential for improvement than patients not sick enough to be referred out from primary care.
So, back to the real world. What do we tell our patients about GLP-1s for weight loss? The truth: These drugs can result in substantial weight loss but only if patients can tolerate and stick with them, which, according to this study, is about a 50/50 chance at 1 year. We think it’s safe to say that Ozempic is no penicillin, but when it comes to weight loss, only time will tell whether it’s a true miracle drug or a fleeting pop culture trend.
For more information, see the topic Anti-Obesity Medications for Adults in DynaMed.
DynaMed EBM Focus Editorial Team
This EBM Focus was written by Katharine DeGeorge, MD, MS, Senior Deputy Editor at DynaMed and Associate Professor of Family Medicine at the University of Virginia. Edited by Alan Ehrlich, MD, FAAFP, Executive Editor at DynaMed and Associate Professor in Family Medicine at the University of Massachusetts Medical School; Dan Randall, MD, MPH, FACP, Senior Deputy Editor at DynaMed; Gayle Sulik, PhD, Senior Medical Editor and Team Lead for Palliative Care at DynaMed; McKenzie Ferguson, PharmD, BCPS, Senior Clinical Writer at DynaMed; Rich Lamkin, MPH, MPAS, PA-C, Clinical Writer at DynaMed; Matthew Lavoie, BA, Senior Medical Copyeditor at DynaMed; Hannah Ekeh, MA, Senior Associate Editor II at DynaMed; and Jennifer Wallace, BA, Senior Associate Editor at DynaMed.