ABSTRACT
Aims
To describe obesity treatments in real‐world settings after discontinuation of semaglutide or tirzepatide and variability in weight change post‐discontinuation.
Materials and Methods
This retrospective cohort study used electronic health records from 1 January 2021, to 30 June 2025, in a large health system in Ohio and Florida. Adults with overweight or obesity who initiated injectable semaglutide or tirzepatide for obesity or T2D between 2021 and 2023 and discontinued the medication within 3–12 months were included. Main outcomes were reinitiation of the index medication or receipt of an alternative treatment, body weight changes from baseline to index medication discontinuation and from discontinuation to 1‐year post discontinuation.
Results
A total of 7938 patients (mean [SD] age, 55.7 [13.4] years; 5061 [63.8%] female) were identified. During 1‐year post‐discontinuation, 19.6% restarted the index medication and 35.2% received an alternative obesity treatment, including starting another medication (27.4%), lifestyle modification visit (13.7%) and metabolic and bariatric surgery (0.6%). Mean percentage weight change from baseline to discontinuation was −8.4% [95% CI, −8.7%, −8.1%] when treating obesity and −4.4% [95% CI, −4.7%, −4.2%] when treating T2D. Mean percentage weight change from discontinuation to 1 year later was 0.5% [95% CI, 0.0%, 1.0%] when treating obesity and −1.3% [95% CI, −1.6%, −1.0%] when treating T2D; however, there was considerable individual‐level variability.
Conclusion
In this large sample of patients who discontinued semaglutide or tirzepatide, reinitiation of the original medication or receipt of alternative obesity treatment was common. The average weight change 1‐year post‐discontinuation was relatively small; however, there was considerable individual‐level variability.
Keywords: glucagon‐like peptide‐1 receptor agonist, lifestyle modification, long‐term weight outcomes, metabolic and bariatric surgery, obesity pharmacotherapy, treatment discontinuation
1. Introduction
Obesity represents a major public health and economic challenge, impacting 40% of US adults and is associated with a 3–11‐fold increase in the risk of major health complications, such as type 2 diabetes (T2D), metabolic dysfunction‐associated steatotic liver disease and obstructive sleep apnea [1, 2, 3, 4, 5]. Novel long‐acting glucagon‐like peptide‐1 receptor agonist [GLP‐1 RA] semaglutide and dual GLP‐1 RA/glucose‐dependent insulinotropic polypeptide [GIP] receptor agonist tirzepatide can reduce body weight by 10%–21% [6, 7, 8]. Treatment with semaglutide also prevents cardiovascular events, particularly in patients with established atherosclerotic cardiovascular disease [9] and slows progression of kidney disease in those with T2D and chronic kidney disease [10]. Tirzepatide also lowers the risk of death from heart failure [11]. As a result of their clinical efficacy, the demand for these medications has skyrocketed, with one in five US women ages 50–64 reporting to have taken a GLP‐1 RA or dual agonist medication [12].
Despite their popularity, discontinuation rates of these medications at 1‐year range from 47% to 65% [13, 14]. Discontinuation generally leads patients to regain weight and reduces the other health benefits [15, 16, 17, 18]. For example, in an extension to the randomised controlled trial (RCT) for semaglutide (STEP 1), after discontinuation patients on average regained two‐thirds of their lost weight and cardiometabolic improvements reversed [15]. Similarly, in the SURMOUNT‐4 trial, after discontinuing tirzepatide, only 17% of the participants maintained at least 80% of the weight loss achieved [16].
Data are limited about what additional weight management efforts patients (with or without T2D) pursue in clinical practice after discontinuing semaglutide or tirzepatide, and whether long‐term weight changes post‐discontinuation follow the patterns reported in RCTs. The primary objective of this study was to describe the obesity treatment that patients with overweight or obesity pursue after discontinuing injectable semaglutide or tirzepatide. We also characterised variability in weight change 1 year later.
2. Methods
2.1. Study Design and Setting
Data for this retrospective cohort study were obtained from the Cleveland Clinic electronic health record (EHR) in Ohio and Florida, including linked Surescripts pharmacy dispensation records [19], from 1 January 2021, to 30 June 2025. The Surescripts prescription data service captures prescriptions paid for through insurance benefits, cash, coupons, or other methods, from nearly all major pharmacies and pharmacy benefit managers in the US [20].
This study was approved by the Cleveland Clinic Institutional Review Board as minimal‐risk research using data collected for routine clinical practice, for which the requirement for informed consent was waived. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guidelines were followed.
2.2. Study Participants
We identified adult patients (aged ≥ 18 years) who initiated pharmacotherapy for obesity or T2D under medical supervision and filled a prescription for injectable semaglutide or tirzepatide (index medication) from 1 January 2021, to 31 December 2023, and then discontinued the medication between 3 and 12 months after treatment initiation. Discontinuation was defined as a gap of greater than 90 days between exhaustion of the previous supply and the next dispense or between exhaustion of the last supply and the end of the first year, which has been commonly used gap period [21, 22]. Patients were required to have a body mass index (BMI; calculated as the weight in kilograms divided by the height in metres squared) of at least 30.0 recorded on the date of treatment initiation (index date) or during the last available primary care visit before the index date, or had a BMI of at least 27.0 and one or more weight‐related comorbidity (i.e., hypertension, dyslipidaemia, obstructive sleep apnea, or cardiovascular disease, including coronary artery disease, atrial fibrillation, heart failure and stroke). To ensure that these were initial fills, we required that they were preceded by at least 6 months of no prescription fills for FDA‐approved obesity medications (OMs), including semaglutide, liraglutide, tirzepatide, phentermine‐topiramate, naltrexone‐bupropion, phentermine, or orlistat. The presence of T2D was defined by either (1) an HbA1c level of at least 6.5% (to convert to proportion of total haemoglobin, multiply by 0.01), (2) presence of a diagnostic code for T2D (International Classification of Diseases, Ninth Revision, code 250.X0 or 250.X2 or International Statistical Classification of Diseases, Tenth Revision, code E11.X) plus prescription for a T2D medication (other than semaglutide or tirzepatide prescribed on or before the index date), or (3) T2D noted in the patient's problem list. Patients who underwent bariatric surgery within 2 years before the initial medication fill and those with incomplete data for prescription fill coverage days were excluded (Figure S1).
This study captured injectable semaglutide and tirzepatide under the brand names approved by the FDA for obesity (Wegovy and Zepbound) and for T2D (Ozempic and Mounjaro), including all starting and maintenance doses. Receipt of other OMs approved by the FDA (including liraglutide, phentermine‐topiramate, naltrexone‐bupropion, orlistat and phentermine) and other medications commonly used (off‐label) for obesity (including topiramate, bupropion, pramlintide, zonisamide, oral semaglutide and dulaglutide) was also captured.
2.3. Study Variables
The primary outcome measures in this study were (1) reinitiation of the index medication (definitions provided in Table S1) or (2) receipt of an alternative obesity treatment, including either (a) initiation of another medication, (b) at least one lifestyle modification visit with a healthcare professional or (c) undergoing metabolic and bariatric surgery (completed operations), all captured within 1 year after discontinuation. To be counted under initiation of another medication, we required that it not be filled during the period from the index date to 1 month before the index medication's discontinuation date. Since the insurance preauthorisation process for metabolic and bariatric surgery could take 6–9 months [23], we also captured treatment reinitiation and alternative obesity treatment uptake at 2 years. We also captured body weight changes (1) from baseline to index medication discontinuation, and (2) from discontinuation to 1 year later, among patients who had corresponding weight measurements recorded during routine clinical encounters, as secondary outcomes (Figure S1 and Table S1).
The end of the follow‐up period for outcome ascertainment was 30 June 2025. Patients were censored at the latest available follow‐up in the EHR, if they became pregnant, were diagnosed with cancer, or died, or (administratively) 1 or 2 years after the index medication's discontinuation date, as applicable.
The index medication's indication was categorised based on the brand name and presence of T2D (Table S1). Sociodemographic variables, including patients' age, sex, race, ethnicity, primary insurance type and Area Deprivation Index (ADI) quartile based on census block group neighbourhood‐level data [24] were recorded from the primary care visit closest to the index date. Data on sex and race and ethnicity were based on fixed categories recorded in the EHR. Race was categorised into Asian, Black, White and other groups (including American Indian or Alaska Native, multiracial, Native Hawaiian or Other Pacific Islander and other). Ethnicity was grouped into Hispanic and non‐Hispanic categories. Data on race and ethnicity were included because they could be associated with both exposure and study outcomes. Area Deprivation Index percentiles were based on a nationwide ranking from 1 to 100, where an ADI with a ranking of 1 indicates the lowest level of disadvantage [24]. Age‐adjusted Charlson Comorbidity Index was also captured from the EHR [25].
2.4. Statistical Analysis
Means and SDs were used to summarise normally distributed data, and medians and IQRs were used for data that were not normally distributed. Pearson χ 2 test, Fisher's exact test, Wilcoxon rank sum test and Kruskal–Wallis rank sum test were used for standard group comparisons [26].
We calculated cumulative incidence curves and 95% CIs for index medication's reinitiation and alternative treatments initiation after discontinuation using the Kaplan–Meier estimator (one minus the product‐limit). Since the obesity treatment categories were not mutually exclusive, patients could have contributed to each of these events. Within the Kaplan–Meier plots for reinitiation, time zero was the date of medication supply exhaustion and the event date was the earliest occurrence of index medication's reinitiation or alternative treatments initiation (e.g., first prescription fill of another medication), respectively.
We also calculated mean weight changes and 95% CIs from baseline to the index medication's discontinuation and from discontinuation to 1 year later. To account for potential differential surveillance, we conducted sensitivity analyses where we used inverse probability weights for missing weight measurements to estimate the mean weight changes under full surveillance. Inverse probability weights were estimated using a logistic regression model which included treatment indication, index medication, age, sex, race, ethnicity, primary insurance type, ADI quartile, age‐adjusted Charlson comorbidity index, year of treatment initiation, baseline body weight and baseline BMI as covariates. Mean weight change was then estimated as weighted mean and weighted SD with 95% CI calculated using a robust variance estimator to account for the two‐step estimation process. We used stabilised weights in this process (for weight diagnostics see Figures S5 and S6 and Tables S15 and S16). We also conducted additional sensitivity analyses, reporting study outcomes stratified by the discontinuation timing (3–< 6 and 6–12 months) and by baseline BMI (27.0–< 35.0 and ≥ 35.0), as well as the baseline characteristics of patients with and without valid discontinuation weight measurements.
In order to construct the weight trajectory graph from baseline to 1 year after treatment discontinuation, we summarised the mean percent change from baseline to discontinuation and then in 1‐month increments between 3 and 12 months after discontinuation by treatment indication across all patients who had a weight measurement in a given month; these data are presented unadjusted. All statistical testing was 2 tailed with α = 0.05 used to determine statistical significance. All analyses were conducted in R, version 4.4.0.
3. Results
This study included 7938 patients who filled an initial prescription for injectable semaglutide (6184) or tirzepatide (1754) from 1 January 2021, to 31 December 2023, and discontinued it between 3 and 12 months later. Mean (SD) age was 55.7 (13.4) years, mean baseline weight was 112.9 (26.2) kg and mean baseline BMI was 39.4 (8.0). A total of 5061 patients (63.8%) were female; 77 patients (1.0%) were Asian, 1447 (18.2%) were Black, 5948 (74.9%) were White and 281 (3.5%) were of other race; 560 (7.1%) had Hispanic ethnicity. The majority were privately insured (5077 [64.0%]) and initiated the therapy in 2023 (4737 [59.7%]) (Table 1).
TABLE 1.
Characteristics of patients who initiated injectable semaglutide or tirzepatide during 2021–2023 and discontinued it between 3 and 12 months later, N = 7938.
| Characteristic | Overall, no. (%) (N = 7938) | Obesity indication, no. (%) (n = 3321) | T2D indication, no. (%) (n = 4617) |
|---|---|---|---|
| Index medication | |||
| Semaglutide | 6184 (77.9%) | 2520 (75.9%) | 3664 (79.4%) |
| Tirzepatide | 1754 (22.1%) | 801 (24.1%) | 953 (20.6%) |
| Age, mean (SD), years | 55.7 (13.4) | 51.6 (13.7) | 58.6 (12.3) |
| Sex | |||
| Male | 2877 (36.2%) | 857 (25.8%) | 2020 (43.8%) |
| Female | 5061 (63.8%) | 2464 (74.2%) | 2597 (56.2%) |
| Race | |||
| Asian | 77 (1.0%) | 14 (0.4%) | 63 (1.4%) |
| Black | 1447 (18.2%) | 506 (15.2%) | 941 (20.4%) |
| Other a | 281 (3.5%) | 124 (3.7%) | 157 (3.4%) |
| White | 5948 (74.9%) | 2607 (78.5%) | 3341 (72.4%) |
| Not reported | 185 (2.3%) | 70 (2.1%) | 115 (2.5%) |
| Ethnicity | |||
| Hispanic | 560 (7.1%) | 203 (6.1%) | 357 (7.7%) |
| Non‐Hispanic | 7164 (90.2%) | 3019 (90.9%) | 4145 (89.8%) |
| Not reported | 214 (2.7%) | 99 (3.0%) | 115 (2.5%) |
| Primary insurance | |||
| Private | 5077 (64.0%) | 2415 (72.7%) | 2662 (57.7%) |
| Traditional Medicare | 248 (3.1%) | 93 (2.8%) | 155 (3.4%) |
| Medicare Advantage | 1612 (20.3%) | 439 (13.2%) | 1173 (25.4%) |
| Medicaid | 810 (10.2%) | 289 (8.7%) | 521 (11.3%) |
| Self‐pay/other | 167 (2.1%) | 74 (2.2%) | 93 (2.0%) |
| Unknown | 24 (0.3%) | 11 (0.3%) | 13 (0.3%) |
| ADI quartile b | |||
| 1–25 | 803 (10.1%) | 408 (12.3%) | 395 (8.6%) |
| 26–50 | 1970 (24.8%) | 950 (28.6%) | 1020 (22.1%) |
| 51–75 | 2646 (33.3%) | 1114 (33.5%) | 1532 (33.2%) |
| 76–100 | 2402 (30.3%) | 797 (24.0%) | 1605 (34.8%) |
| Unknown | 117 (1.5%) | 52 (1.6%) | 65 (1.4%) |
| Baseline weight, kg, mean (SD) | 112.9 (26.2) | 112.5 (26.1) | 113.2 (26.2) |
| Baseline BMI, mean (SD) | 39.4 (8.0) | 39.7 (7.9) | 39.2 (8.1) |
| Baseline glycated haemoglobin, %, mean (SD) | 7.2 (2.9) | 5.6 (0.5) | 7.9 (3.2) |
| Unknown | 2019 (25.4%) | 1474 (44.4%) | 545 (11.8%) |
| Charlson comorbidity index, median (IQR) | 4.0 (2.0, 6.0) | 2.0 (1.0, 4.0) | 5.0 (4.0, 8.0) |
| Year of treatment initiation | |||
| 2021 | 678 (8.5%) | 207 (6.2%) | 471 (10.2%) |
| 2022 | 2523 (31.8%) | 1178 (35.5%) | 1345 (29.1%) |
| 2023 | 4737 (59.7%) | 1936 (58.3%) | 2801 (60.7%) |
Note: Index medication's discontinuation at 1 year was defined as a gap of greater than 90 days between exhaustion of previous supply and next dispense or between exhaustion of last supply and end of 1‐year follow‐up.
Abbreviations: ADI, area deprivation index; BMI, body mass index; IQR, interquartile range; SD, standard deviation.
Other race included American Indian or Alaska Native, multiracial, Pacific Islander and other.
ADI quartiles were structured by ranking the ADI from low to high nationally, where an ADI with a ranking of 1 indicates the lowest level of disadvantage.
Overall, 3664 patients (46.2% of the study cohort) received semaglutide indicated for T2D, 953 (12.0%) received tirzepatide for T2D, 2520 (31.7%) received semaglutide for obesity and 801 (10.1%) received tirzepatide for obesity. Distribution of discontinuation timing in the study cohort is presented in Table S2.
3.1. Treatments After the Index Medication's Discontinuation
Overall, 1557 patients (19.6% of the study cohort) reinitiated their index medication within 1 year of treatment discontinuation, including 472 (14.2%) who received the original therapy for obesity and 1085 (23.5%) for T2D. Among these patients, the mean (SD) days to reinitiate the index medication was 176.7 (73.1), including 185.6 (75.9) for obesity and 172.8 (71.6) for T2D indication. Figure 1 (panel A) presents the cumulative incidence of index medication reinitiation at 1 year by its indication.
FIGURE 1.
Cumulative incidence of index medication reinitiation (A), receipt of another medication (B), bariatric surgery (C), or lifestyle modification visit with a healthcare professional (D) during 1 year after index medication's discontinuation by treatment indication. Since the obesity treatment categories were not mutually exclusive, patients could have contributed to each of these events. Index medication's indication was categorised based on the brand name and presence of T2D. T2D, type 2 diabetes.
During 1‐year post‐discontinuation, 2794 patients (35.2% of the study cohort) received an alternative obesity treatment, including starting another medication (2175 [27.4%]), lifestyle modification visit with a healthcare professional (1088 [13.7%]) and metabolic and bariatric surgery (50 [0.6%]). Panels B–D in Figure 1 present the cumulative incidences of alternative treatments receipt at 1 year by indication, and Tables S3 and S4 present granular data on other medications receipt at 1‐ and 2 years after discontinuation, respectively.
In the overall cohort, 282 (16.1% of patients originally on tirzepatide) switched to injectable semaglutide and 1148 (18.6% of patients originally on semaglutide) switched to tirzepatide (Figure 2). The most commonly used other medications at 1 year after discontinuation in patients whose original therapy was for obesity were tirzepatide (in 390 patients out of 2520 whose index medication was semaglutide for obesity), phentermine (184 out of 3321 whose index medication was for obesity), topiramate (139 out of 3321 whose index medication was for obesity), bupropion (97 out of 3321) and semaglutide (88 out of 3321). At the time of discontinuation, 27.0% (680 out of 2520) of patients taking injectable semaglutide for obesity reached a dosage of ≥ 1.7 mg and 88.5% (709 out of 801) of patients taking tirzepatide for obesity reached a dosage of ≥ 5 mg (Tables S5 and S6).
FIGURE 2.
Sankey diagram of post‐discontinuation surgical and pharmacological treatments pursued by the study cohort during 1 year after index medication's discontinuation. Index medication's indication was categorised based on the brand name and presence of T2D. T1 indicates first, T2, second and T3, third treatment after the index medication's discontinuation. Bar heights are proportionate to the number of patients in each subgroup. OM, obesity medication; T2D, type 2 diabetes.
Among patients with at least one lifestyle modification visit, 381 (56.4%) of those treated for obesity and 186 (45.1%) of those treated for T2D had 2 or more visits within the year after discontinuation. These included nutritional counselling visits, individual and group education with an exercise physiologist, shared medical appointments on weight management, diabetes patient education with a weight management component and other structured visits focused on weight management. The results of sensitivity analyses (stratified by discontinuation timing and baseline BMI) supported the main findings (Tables S9 and S10).
3.2. Weight Change
Figure 3 presents the mean percent change in weight by treatment indication from initiation of medication to discontinuation and then in the year after discontinuation. A total of 6634 patients (83.6% of the study cohort) had baseline and discontinuation weight measurement or baseline and at least 1 post‐discontinuation weight measurement (with no censoring events) and contributed to this analysis. The median patient in the study cohort had weight measurements recorded in 7 separate months (IQR, 4–13) from index medication fill date through 1‐year after discontinuation.
FIGURE 3.
Mean percentage weight change by treatment indication from initiation of the medication through 1‐year after discontinuation. A total of 6634 patients contributed to this analysis; they had baseline and discontinuation weight measurement or baseline and at least 1 post‐discontinuation weight measurement, and no censoring events. The median patient in the study cohort had weight measurements recorded in 7 separate months (IQR, 4–13) from index date through 1‐year after discontinuation. In order to construct the weight trajectory graph from baseline to 1 year after treatment discontinuation, we summarised the mean percent change from baseline to discontinuation and then in 1‐month increments between 3 and 12 months after discontinuation by treatment indication across all patients who had a weight measurement in a given month; these data are presented unadjusted. Shaded areas indicate 95% CI.
Among patients with recorded weight measurement at the time of index medication's discontinuation and no censoring events prior to discontinuation (n = 5046, 63.6% of the study cohort), the mean (SD) percentage weight change from baseline to discontinuation when treating obesity was −8.4% (7.6%) [95% CI, −8.7%, −8.1%] and when treating T2D was −4.4% (6.1%) [95% CI, −4.7%, −4.2%]. Table S7 presents the percent weight change from baseline to discontinuation, stratified by medication and indication, and Table S11 shows the baseline characteristics of patients who did and did not have a weight measurement at the index medication's discontinuation time.
Among patients with recorded weight measurement both at discontinuation and 1 year later, and no censoring events (n = 3810, 48.0% of the study cohort), the mean (SD) percentage weight change in the year after discontinuation when treating obesity was 0.5% (9.9%) [95% CI, 0.0%, 1.0%] and when treating T2D was −1.3% (7.8%) [95% CI, −1.6%, −1.0%]. Table S8 presents the weight change from discontinuation to 1‐year post‐discontinuation, stratified by medication and indication. Our sensitivity analyses, including the estimates for mean weight changes using inverse probability weights, supported the main results (Tables S12–S14).
As shown in Figure 4, there was considerable individual‐level variation in weight change following discontinuation. Among patients who started the treatment for obesity, 677 (44.6%) either had additional weight reduction or no change in weight, while 842 (55.4%) gained weight in the year after discontinuation. Among patients who started the treatment for T2D, 1291 (56.4%) either had additional weight reduction or no change in weight, while 1000 (43.6%) gained weight. Figures S3 and S4 present the individual‐level variation in weight change after discontinuation by indication.
FIGURE 4.
Individual‐level variation in percent weight change from index medication's discontinuation time to 1‐year post‐discontinuation by treatment status following discontinuation. A total of 3810 patients contributed to this analysis; they had both discontinuation and 1‐year post‐discontinuation weight measurements and no censoring events. We identified adult patients who initiated injectable semaglutide or tirzepatide from 1 January 2021, to 31 December 2023, and then discontinued the medication between 3 and 12 months after treatment initiation. Index medication's discontinuation at 1 year was defined as a gap of greater than 90 days between exhaustion of previous supply and next dispense or between exhaustion of last supply and end of 1‐year follow‐up.
4. Discussion
In this large cohort of patients with overweight or obesity who initiated treatment with injectable semaglutide or tirzepatide for obesity or T2D and discontinued the medication between 3 and 12 months after initiation, 19.6% restarted the same medication and 35.2% received an alternative treatment in the following year. Patients treated for T2D (vs. obesity) restarted their original medication at higher rates, while those treated for obesity had higher rates of lifestyle modification visits. The mean percentage weight change in the year following discontinuation was 0.5% in patients who started the treatment for obesity and −1.3% for T2D, with considerable individual‐level variation.
To our knowledge, this is the first study to provide granular data on obesity treatments uptake, as well as weight trajectories and variability in weight change 1‐year after discontinuation of injectable semaglutide or tirzepatide. Previously, Rodriguez and colleagues found that 47% of patients with T2D and 36% of those without T2D who discontinued a GLP‐1 RA medication within the first year of initiation subsequently reinitiated a GLP‐1 RA medication (including the index medication or another GLP‐1 RA) within 1 year [14]. In our study, too, more patients with T2D indication restarted their original medication within 1 year (23.5%) than those treated for obesity (14.2%). This is not surprising, as the insurance coverage for these medications currently is much broader for the T2D indication compared to the obesity indication [14, 27, 28]. Furthermore, a recent study reported that while cost or insurance‐related issues were the most common reason for discontinuation of semaglutide or tirzepatide for obesity during the first year, patients discontinuing due to cost were more likely to do so later in treatment, whereas those discontinuing due to side effects tended to do so earlier [21]. The uptake of metabolic and bariatric surgery in this cohort was at 0.6% during the first year after discontinuation and 1.3% at 2 years, which is comparable to what has been observed in the general population of patients eligible for surgery [29, 30].
Many patients in our study who started the therapy for obesity discontinued before reaching adequate therapeutic dosages. This is in line with what has been observed in other US‐based large studies [22, 31], and may have contributed to the weight loss achieved in this cohort [32]. Future studies should examine why patients deviate from the recommended dose‐escalation schedules, including whether manufacturers' approach of pricing lower dosages of these medications more cheaply (compared to higher dosages) in direct‐to‐consumer (self‐pay) channels contributes to it [33].
A recent systematic review study reported that discontinuation of OMs (without alternative treatment) results in an estimated average weight regain at a rate of 0.4 kg/month [18]. Our real‐world findings indicate that many patients either reinitiate the original therapy or pursue an alternative treatment after discontinuation of semaglutide or tirzepatide, which may have contributed to the relatively small observed mean changes from discontinuation to 1 year later in this cohort. In the year after discontinuation, more patients treated for obesity regained weight compared to those treated for T2D. This is likely a result of more robust access to weight‐reducing medications when a patient has T2D, compared to those who do not. There was also considerable variability in weight change in the year following initial treatment discontinuation. Future studies should examine the comparative effectiveness of treatments for obesity in patients who discontinue their novel GLP‐1 RA or dual receptor agonist medication, as currently no evidence‐based guidelines exist for this commonly occurring situation in clinical practice.
5. Strengths and Limitations
Strengths of this study include its large and contemporaneous sample, integration of prescription dispensation data from Surescripts with clinical information [19, 20] and capturing a broad spectrum of treatments for obesity after initial treatment's discontinuation. Our study population was diverse (including 18% Black, 7% Hispanic, 30% residing in an area in the most disadvantaged ADI quartile) and we did not limit our cohort to only those receiving care from obesity specialists.
This study also has some limitations. We included adult patients from a single large integrated health system in Ohio and Florida. Patient characteristics and health care delivery patterns vary across the US, which may limit the generalisability of our findings. Similarly, while the estimates for mean weight changes using inverse probability weights were similar to the main results, the results of this study may not apply to patients who do not seek care where they initiate their original treatment. Shortages of semaglutide and tirzepatide during the study period could have contributed to the index medication's temporary discontinuation. Some of the observed weight reduction may be associated with other interventions (e.g., unstructured lifestyle or dietary) that we were unable to capture in our datasets. This study did not capture the use of compounded medications, however in a randomly selected subset of Cleveland Clinic patients who started injectable semaglutide or tirzepatide for obesity from 1 January 2022, to 31 December 2023, only 2.4% discontinued the brand medication due to switching to a compounded agent [21]. This study did not capture specific reasons for discontinuation of the medication therapy (such as due to inadequate weight reduction, adverse effects, or coverage issues), and the impact of these factors could not be examined in this study. Finally, our analysis assumes that the existence of missing weights is independent of a patient's weight outcomes. If this assumption is wrong, our secondary outcomes may be biased, with the amount of bias being a function of the percent of missing outcomes and the outcome difference between patients with observed versus missing outcomes.
6. Conclusion
In this cohort study of adults with overweight or obesity who initiated treatment with injectable semaglutide or tirzepatide and discontinued the index medication between 3 and 12 months after initiation, 19.6% restarted the index medication and 35.2% received an alternative treatment in the year after initial treatment discontinuation. The average weight change 1 year after index medication discontinuation was relatively small; however, there was considerable individual‐level variability. Future studies should examine the comparative effectiveness of treatments for obesity in patients who discontinue their novel GLP‐1 RA or dual receptor agonist medication.
Funding
This work was supported by Cleveland Clinic—Jack, Joseph and Morton Mandel Accelerator Grants, Grant Number: MAG0030.
Conflicts of Interest
Dr. Hamlet Gasoyan reported receiving grant funding from the National Cancer Institute outside the submitted work. Dr. Christopher B. Boyer reports consulting fees from Janssen Pharmaceuticals outside the submitted work. Dr. W. Scott Butsch is currently employed by Novo Nordisk A/S; he also reported advisory board fees from Eli Lilly, Boehringer Ingelheim and Abbott and research funding from Eli Lilly and Co outside the submitted work. Dr. Ali Aminian reported receiving research grants from Amgen, Ethicon and Medtronic outside the submitted work. He serves as a consultant for Amylyx, Eli Lilly, Ethicon and Medtronic. Dr. Marcio L. Griebeler reported research funding from Novo Nordisk A/S outside the submitted work. Dr. Michael B. Rothberg reported receiving consulting fees from the Blue Cross Blue Shield Association outside the submitted work. The other authors declare no conflicts of interest.
Supporting information
Appendix S1: Supporting Information.
Data Availability Statement
The dataset generated during the current study is not publicly available to preserve patient confidentiality. However, the dataset is available from the corresponding author to academic investigators following receipt of a signed data‐sharing agreement and after reviewing the study protocol (approved by a local institutional review board or research ethics committee), statistical analysis plan and publication plan. The Cleveland Clinic Institutional Review Board and the Law Department must approve the request before sharing the deidentified data. The corresponding author will respond to these requests within 2 months of receipt.
References
- 1. Yao Z., Tchang B. G., Albert M., Blumenthal R. S., Nasir K., and Blaha M. J., “Associations Between Class I, II, or III Obesity and Health Outcomes,” NEJM Evidence 4, no. 4 (2025): EVIDoa2400229, 10.1056/EVIDoa2400229. [DOI] [PubMed] [Google Scholar]
- 2. Bhaskaran K., Douglas I., Forbes H., dos‐Santos‐Silva I., Leon D. A., and Smeeth L., “Body‐Mass Index and Risk of 22 Specific Cancers: A Population‐Based Cohort Study of 5·24 Million UK Adults,” Lancet 384, no. 9945 (2014): 755–765, 10.1016/S0140-6736(14)60892-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Guh D. P., Zhang W., Bansback N., Amarsi Z., Birmingham C. L., and Anis A. H., “The Incidence of Co‐Morbidities Related to Obesity and Overweight: A Systematic Review and Meta‐Analysis,” BMC Public Health 9, no. 1 (2009): 88, 10.1186/1471-2458-9-88. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Emmerich S. D., Fryar C. D., Stierman B., and Ogden C. L., Obesity and Severe Obesity Prevalence in Adults: United States, August 2021–August 2023 (National Center for Health Statistics, 2025), https://www.cdc.gov/nchs/products/databriefs/db508.htm. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Okunogbe A., Nugent R., Spencer G., Powis J., Ralston J., and Wilding J., “Economic Impacts of Overweight and Obesity: Current and Future Estimates for 161 Countries,” BMJ Global Health 7, no. 9 (2022): e009773, 10.1136/bmjgh-2022-009773. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Wilding J. P. H., Batterham R. L., Calanna S., et al., “Once‐Weekly Semaglutide in Adults With Overweight or Obesity,” New England Journal of Medicine 384, no. 11 (2021): 989–1002, 10.1056/NEJMoa2032183. [DOI] [PubMed] [Google Scholar]
- 7. Jastreboff A. M., Aronne L. J., Ahmad N. N., et al., “Tirzepatide Once Weekly for the Treatment of Obesity,” New England Journal of Medicine 387, no. 3 (2022): 205–216, 10.1056/NEJMoa2206038. [DOI] [PubMed] [Google Scholar]
- 8. Davies M., Færch L., Jeppesen O. K., et al., “Semaglutide 2·4 mg Once a Week in Adults With Overweight or Obesity, and Type 2 Diabetes (STEP 2): A Randomised, Double‐Blind, Double‐Dummy, Placebo‐Controlled, Phase 3 Trial,” Lancet 397, no. 10278 (2021): 971–984, 10.1016/S0140-6736(21)00213-0. [DOI] [PubMed] [Google Scholar]
- 9. Lincoff A. M., Brown‐Frandsen K., Colhoun H. M., et al., “Semaglutide and Cardiovascular Outcomes in Obesity Without Diabetes,” New England Journal of Medicine 389, no. 24 (2023): 2221–2232, 10.1056/NEJMoa2307563. [DOI] [PubMed] [Google Scholar]
- 10. Perkovic V., Tuttle K. R., Rossing P., et al., “Effects of Semaglutide on Chronic Kidney Disease in Patients With Type 2 Diabetes,” New England Journal of Medicine 391, no. 2 (2024): 109–121, 10.1056/NEJMoa2403347. [DOI] [PubMed] [Google Scholar]
- 11. Packer M., Zile M. R., Kramer C. M., et al., “Tirzepatide for Heart Failure With Preserved Ejection Fraction and Obesity,” New England Journal of Medicine 392, no. 5 (2025): 427–437, 10.1056/NEJMoa2410027. [DOI] [PubMed] [Google Scholar]
- 12. Bozick R., Donofry S. D., and Rancaño K. M., New Weight Loss Drugs: GLP‐1 Agonist Use and Side Effects in the United States (RAND, 2025), https://www.rand.org/pubs/research_reports/RRA4153‐1.html#:~:text=The%20KFF%20survey%20findings%20show,the%20RAND%20American%20Life%20Panel. [PMC free article] [PubMed] [Google Scholar]
- 13. Gasoyan H., Pfoh E. R., Schulte R., Le P., Butsch W. S., and Rothberg M. B., “One‐Year Weight Reduction With Semaglutide or Liraglutide in Clinical Practice,” JAMA Network Open 7, no. 9 (2024): e2433326, 10.1001/jamanetworkopen.2024.33326. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Rodriguez P. J., Zhang V., Gratzl S., et al., “Discontinuation and Reinitiation of Dual‐Labeled GLP‐1 Receptor Agonists Among US Adults With Overweight or Obesity,” JAMA Network Open 8, no. 1 (2025): e2457349, 10.1001/jamanetworkopen.2024.57349. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Wilding J. P. H., Batterham R. L., Davies M., et al., “Weight Regain and Cardiometabolic Effects After Withdrawal of Semaglutide: The STEP 1 Trial Extension,” Diabetes, Obesity & Metabolism 24, no. 8 (2022): 1553–1564, 10.1111/dom.14725. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Aronne L. J., Sattar N., Horn D. B., et al., “Continued Treatment With Tirzepatide for Maintenance of Weight Reduction in Adults With Obesity: The SURMOUNT‐4 Randomized Clinical Trial,” Journal of the American Medical Association 331, no. 1 (2024): 38–48, 10.1001/jama.2023.24945. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Rubino D., Abrahamsson N., Davies M., et al., “Effect of Continued Weekly Subcutaneous Semaglutide vs Placebo on Weight Loss Maintenance in Adults With Overweight or Obesity,” Journal of the American Medical Association 325, no. 14 (2021): 1414–1425, 10.1001/jama.2021.3224. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. West S., Scragg J., Aveyard P., et al., “Weight Regain After Cessation of Medication for Weight Management: Systematic Review and Meta‐Analysis,” BMJ 392 (2026): e085304, 10.1136/bmj-2025-085304. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Blecker S., Adhikari S., Zhang H., et al., “Validation of EHR Medication Fill Data Obtained Through Electronic Linkage With Pharmacies,” JMCP 27, no. 10 (2021): 1482–1487, 10.18553/jmcp.2021.27.10.1482. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Surescripts , “Surescripts Dispensed Prescription Data Service Now Covers Nearly 240 Million Patients to Advance Population Health,” 2017, https://surescripts.com/news‐center/press‐releases/detail/surescripts‐dispensed‐prescription‐data‐service‐now‐covers‐nearly‐240‐million‐patients‐to‐advance‐population‐health.
- 21. Gasoyan H., Butsch W. S., Casacchia N. J., et al., “Reasons for Discontinuation of Obesity Pharmacotherapy With Semaglutide or Tirzepatide in Clinical Practice,” Obesity (Silver Spring, Md.) 33 (2025): 70058, 10.1002/oby.70058. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Samuels J. M., Ye F., Irlmeier R., Silver H., Srivastava G., and Spann M., “Real‐World Titration, Persistence & Weight Loss of Semaglutide and Tirzepatide in an Academic Obesity Clinic,” Diabetes, Obesity & Metabolism 27, no. 11 (2025): 6200–6209, 10.1111/dom.70004. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Gasoyan H., Soans R., Ibrahim J. K., Aaronson W. E., and Sarwer D. B., “Do Insurance‐Mandated Precertification Criteria and Insurance Plan Type Determine the Utilization of Bariatric Surgery Among Individuals With Private Insurance?,” Medical Care 58, no. 11 (2020): 952–957, 10.1097/MLR.0000000000001358. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Kind A. J. H. and Buckingham W. R., “Making Neighborhood‐Disadvantage Metrics Accessible—The Neighborhood Atlas,” New England Journal of Medicine 378, no. 26 (2018): 2456–2458, 10.1056/NEJMp1802313. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Austin S. R., Wong Y. N., Uzzo R. G., Beck J. R., and Egleston B. L., “Why Summary Comorbidity Measures Such as the Charlson Comorbidity Index and Elixhauser Score Work,” Medical Care 53, no. 9 (2015): e65–e72, 10.1097/MLR.0b013e318297429c. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Du Prel J. B., Röhrig B., Hommel G., and Blettner M., “Choosing Statistical Tests,” Deutsches Ärzteblatt International 107, no. 19 (2010): 343–348, 10.3238/arztebl.2010.0343. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Ippolito B. and Levy J. F., “Expanding Medicare Coverage of Anti‐Obesity Medicines Could Increase Annual Spending by $3.1 Billion to $6.1 Billion: Article Examines Spending Implications if Medicare Were to Cover Antiobesity Medicines,” Health Affairs (Project Hope) 43, no. 9 (2024): 1254–1262, 10.1377/hlthaff.2024.00356. [DOI] [PubMed] [Google Scholar]
- 28. Gasoyan H. and Sarwer D. B., “Addressing Insurance‐Related Barriers to Novel Antiobesity Medications: Lessons To Be Learned From Bariatric Surgery,” Obesity 30, no. 12 (2022): 2338–2339, 10.1002/oby.23556. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Adekunle O. A., Le P., Gupta D. Y., et al., “Socio‐Demographic and Clinical Factors Associated With the Receipt of Anti‐Obesity Medication Prescriptions and Metabolic and Bariatric Surgery Among Eligible All of US Participants,” Diabetes, Obesity & Metabolism 27, no. 9 (2025): 4978–4988, 10.1111/dom.16544. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Gasoyan H., Tajeu G., Halpern M. T., and Sarwer D. B., “Reasons for Underutilization of Bariatric Surgery: The Role of Insurance Benefit Design,” Surgery for Obesity and Related Diseases 15, no. 1 (2019): 146–151, 10.1016/j.soard.2018.10.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Xu Y., Carrero J. J., Chang A. R., et al., “Titration and Discontinuation of Semaglutide for Weight Management in Commercially Insured US Adults,” Obesity 33, no. 7 (2025): 1243–1248, 10.1002/oby.24315. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Gasoyan H., Butsch W. S., Schulte R., et al., “Changes in Weight and Glycemic Control Following Obesity Treatment With Semaglutide or Tirzepatide by Discontinuation Status,” Obesity 33, no. 9 (2025): 1657–1667, 10.1002/oby.24331. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Gasoyan H. and Rothberg M. B., “Coverage Is the Key to Realizing the Promise of Semaglutide,” Journal of General Internal Medicine (2026): s11606‐026‐10195‐y, 10.1007/s11606-026-10195-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Appendix S1: Supporting Information.
Data Availability Statement
The dataset generated during the current study is not publicly available to preserve patient confidentiality. However, the dataset is available from the corresponding author to academic investigators following receipt of a signed data‐sharing agreement and after reviewing the study protocol (approved by a local institutional review board or research ethics committee), statistical analysis plan and publication plan. The Cleveland Clinic Institutional Review Board and the Law Department must approve the request before sharing the deidentified data. The corresponding author will respond to these requests within 2 months of receipt.