ABSTRACT
Objective
This study assessed the impact of sex and age on weight loss outcomes in patients receiving tirzepatide in real‐world clinical practice.
Methods
We conducted a retrospective cohort study of adults with overweight or obesity who initiated tirzepatide at Mayo Clinic between June 2022 and May 2024 and completed ≥ 12 months of continuous therapy. Primary outcome was total body weight loss percentage (TBWL%) at 15 months, stratified by sex and age groups (≤ 45, 46–59, ≥ 60 years).
Results
Among 1039 patients (57% women; mean age 56 ± 11 years), women achieved significantly greater TBWL% than men after 15 months of treatment (15.1% vs. 10.7%, p < 0.001). Patients aged ≤ 45 years had greater weight loss than those ≥ 60 years (15.1% vs. 12.3%, p = 0.024). In multivariable analyses, greater weight loss was independently associated with female sex, absence of type 2 diabetes, no prior obesity medication use, no concomitant weight gain‐promoting medications, and higher tirzepatide dosage. Age was not an independent predictor.
Conclusions
Sex, but not age, predicted weight loss with tirzepatide among individuals using the medication continuously for ≥ 12 months. These findings underscore the importance of incorporating sex‐specific considerations into personalized obesity treatment strategies.
Keywords: age, obesity, sex differences, tirzepatide, weight loss
1. Introduction
Obesity is a highly prevalent and chronic disease that affects over 890 million people globally [1]. It is closely associated with a range of comorbidities, including hypertension, cardiovascular disease, type 2 diabetes, dyslipidemia, metabolic associated steatotic liver disease (MASLD), and obstructive sleep apnea (OSA) [2]. Over the past years, the prevalence of obesity has steadily risen, contributing to increased morbidity and mortality rates [3]. By 2030, the US obesity prevalence is projected to reach nearly 49% [4]. In addition to its severe health consequences, obesity also places a substantial economic burden on countries worldwide. It has been estimated that obesity costs up to 2.24% of a nation's gross domestic product (GDP), and the World Obesity Federation further predicts that by 2035, the global economic impact of obesity could escalate to a staggering $4 trillion in US dollars [5, 6].
A wide range of interventions are available to address obesity, including lifestyle modification, pharmacologic therapies, endoscopic bariatric procedures, and bariatric surgery. Pharmacotherapy is currently indicated as an adjunct to lifestyle modification in individuals with a body mass index (BMI) ≥ 30 kg/m2 or individuals with a BMI ≥ 27 kg/m2 in the presence of adiposity‐related comorbidities. Tirzepatide, a dual glucagon‐like peptide‐1 (GLP‐1) and glucose‐dependent insulinotropic polypeptide (GIP) receptor agonist, is among the newest and most effective obesity medications. At its highest dose of 15 mg, obesity clinical trials have demonstrated that tirzepatide induces a total body weight loss (TBWL) of approximately 20% over 2 years of continuous treatment, compared with approximately 3% weight loss in the placebo group [7].
Despite tirzepatide's proven efficacy, real‐world data evaluating the influence of sex and age on weight loss outcomes remain limited, with most evidence derived from post hoc analyses of the pivotal phase 3 randomized‐controlled trials of tirzepatide in obesity and diabetes (SURMOUNT and SURPASS programs, respectively). In these trials, women consistently achieve approximately 30% greater weight loss than men with tirzepatide, irrespective of diabetes status [8]. As with female sex, younger age has also been identified as an independent predictor of greater weight reduction in the SURPASS program trials [9]. Building on this evidence, this study aims to investigate whether sex and age predict weight loss outcomes in response to long‐term, continuous tirzepatide use in a real‐world setting.
2. Methods
2.1. Study Design and Participants
This retrospective analysis of electronic health records (EHR) from the Mayo Clinic Health System evaluated the effect of sex and age on weight loss outcomes with continuous use of tirzepatide for at least 12 months among adults with overweight or obesity. The Mayo Clinic Institutional Review Board approved the study (IRB #17‐001068). Adults who initiated weekly subcutaneous tirzepatide injections for weight management between June 3, 2022, and May 17, 2024, were identified from a large clinical database (Figure 1). The cohort comprised patients receiving either the diabetes‐ or obesity‐labeled formulation of tirzepatide, generally according to their type 2 diabetes status. Eligible participants included individuals with overweight (BMI ≥ 27 kg/m2) in the presence of adiposity‐related comorbidities or those with obesity (BMI ≥ 30 kg/m2) regardless of the presence of adiposity‐related comorbidities. Per protocol design, only individuals with at least 12 months of continuous tirzepatide use were included. This is because this study was not designed to estimate the overall effectiveness of tirzepatide; rather, it aimed to evaluate how sex and age relate to variation in weight loss response among patients maintained on long‐term, continuous therapy. Patients were excluded if they had undergone prior bariatric surgery (surgical or endoscopic), concomitantly used other obesity medications (including compounded products), exhibited inconsistent medication adherence (defined as a proportion of days covered < 80% or a therapy gap of ≥ 45 days without medication), were pregnant, had active malignancy, or carried medical conditions that could significantly impact weight loss (e.g., Prader–Willi syndrome). Individuals without documented baseline or follow‐up weight measures were also excluded. The study adhered to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines.
FIGURE 1.
Study flowchart.
2.2. Data Collection
Weight measurements were obtained from the EHR at baseline (±14 days from tirzepatide initiation) and at subsequent follow‐up visits at months 3 (±30 days), 6 (±30 days), 9 (±30 days), 12 (±45 days), and 15 (±45 days). All weights were clinic‐measured by trained personnel using calibrated scales during routine clinical visits. Weight values were recorded in the EHR in kilograms. Collected data included demographic characteristics (age, sex, race), anthropometric measurements (weight, BMI), and vital signs (systolic [SBP] and diastolic blood pressure [DBP]). Total body weight loss percentage (TBWL%) was calculated as [(weight at predefined time point − baseline weight)/baseline weight × 100]. TBWL% is coded such that negative values indicate greater weight loss. Laboratory data comprised glycated hemoglobin A1c (HbA1c), fasting glucose, triglycerides, total cholesterol, low‐density lipoprotein (LDL) cholesterol, high‐density lipoprotein (HDL) cholesterol, aspartate aminotransferase (AST), and alanine aminotransferase (ALT). Medical history data included the following adiposity‐related comorbidities: type 2 diabetes, dyslipidemia, hypertension, gastroesophageal reflux disease (GERD), MASLD, OSA, anxiety, and depression. Comorbidities were identified based on documented diagnoses in the EHR, as coded by treating clinicians, and corroborated with health history questionnaires collected during routine clinical visits. Tirzepatide initiation date and adherence were determined via detailed EHR chart review, incorporating evidence of prescription refill activity (including submitted and missed refills), patient‐reported use and extended interruptions, and clinician documentation from follow‐up visits. The maximum tirzepatide dose achieved, defined as the dose achieved at last follow‐up rather than a time‐specific exposure, was recorded and categorized as low (2.5, 5, or 7.5 mg) or high (10, 12.5, or 15 mg). History of previous obesity medication use was documented and included any US Food and Drug Administration (FDA)–approved weight loss drugs at therapeutic doses (liraglutide, semaglutide, tirzepatide, phentermine, orlistat, phentermine/topiramate, bupropion‐naltrexone). Previous GLP‐1 receptor agonist (GLP‐1 RA) exposure was defined as use of GLP‐1–based therapies indicated for weight management, regardless of diabetes or obesity labeling. Concomitant use of weight gain‐promoting medications was also collected and included the use of systemic corticosteroids, opioids, β blockers, diabetes medications (insulin, sulfonylureas, metiglinides, and thiazolidinediones), anticonvulsants (e.g., pregabalin), antipsychotics (e.g., olanzapine, quetiapine), and antidepressants (e.g., amitriptyline). All EHR‐derived data were rigorously validated through a comprehensive manual chart review to ensure the accuracy and completeness of weight measurements and medication records. Additionally, quality checks were performed, including cross‐checking recorded values against clinician notes and visit summaries, as well as identifying and reconciling inconsistent entries.
2.3. Study Endpoints
The primary endpoint was TBWL% at 15 months after tirzepatide initiation, assessed for the overall cohort stratified by sex (female vs. male) and age groups. Age groups were stratified based on life stage categories as young adults (≤ 45 years), middle‐aged adults (46–59 years), and older adults (≥ 60 years), reflecting meaningful life stages and aligning with the cohort's age distribution to support clinically interpretable comparisons with adequate sample sizes. Secondary endpoints included TBWL% at 3, 6, 9, and 12 months and the proportion of individuals achieving weight loss thresholds (≥ 5%, ≥ 10%, ≥ 15%, and ≥ 20%) at each time point by sex and age groups. Furthermore, we explored predictors of weight loss across all time points.
2.4. Statistical Analysis
Data normality was assessed using the Shapiro–Wilk test. Continuous variables, including TBWL%, are presented as mean (standard deviation [SD]), and categorical variables as number (percentage [%]). Differences in TBWL% between groups were assessed at each time point (3, 6, 9, 12, and 15 months) using independent two‐sample t‐tests for sex comparisons and one‐way ANOVA for age group comparisons, with Bonferroni‐adjusted pairwise comparisons. Predictors of weight loss outcomes were evaluated using linear mixed‐effects models (LMM), with a random intercept for each patient to account for correlation among repeated measurements. The multivariable model included a time × sex interaction and adjusted for age, baseline BMI, type 2 diabetes, anxiety, depression, prior obesity medication use, concomitant weight gain‐promoting medication use, and maximum tirzepatide dose achieved. Interactions were tested in a stepwise fashion, with retention in the final model if p < 0.05 and model performance (as measured by R 2) improved by ≥ 0.005. In addition, given the differences in type 2 diabetes prevalence between sexes and a statistically significant three‐way interaction among sex, type 2 diabetes status, and time, we performed analyses stratified by diabetes status to determine whether weight loss trajectories differed according to the combined effects of sex and diabetes status over time. Chi‐square tests were used to compare the proportions of participants achieving clinically meaningful weight loss thresholds (≥ 5%, ≥ 10%, ≥ 15%, and ≥ 20%) at each time point between sexes and age groups. All statistical tests were two‐tailed with significance set at p < 0.05. Analyses were performed using BlueSky Statistics (v10.3.7) and R software (v4.4.2), with figures generated using GraphPad Prism (v10.4.2).
3. Results
A total of 24,230 tirzepatide prescriptions were issued between June 2022 and May 2024 within the Mayo Clinic Health System. Of these, 23,191 patients met at least one exclusion criterion. The most common reasons for exclusion were less than 12 months of use (n = 19,288), insufficient data (n = 1865), and patients who never initiated the medication (n = 741). Additional exclusions included history of bariatric surgery (n = 405), concomitant use of other obesity medications (n = 371), active malignancy (n = 221), inconsistent use (n = 138), missing baseline weight (n = 87), no follow‐up weight (n = 42), baseline BMI < 27 kg/m2 (n = 14), pregnancy (n = 8), medical conditions that interfere with outcomes (e.g., Prader–Willi syndrome, uncontrolled hypothyroidism; n = 7), and use of compounded tirzepatide (n = 4). The remaining 1039 patients were included in the analysis (Figure 1).
3.1. Baseline Characteristics
Baseline characteristics of the 1039 patients in this study are summarized for the entire cohort and stratified by sex in Table 1. Nearly 60% (n = 594) of participants were women. The cohort was predominantly middle‐aged, with a mean age of 56 years (SD 11). Age differed significantly between groups, with men being older (58 years, SD 11) compared to women (54 years, SD 12) (p < 0.001). Most participants were White (90%, n = 936), with a similar racial distribution across sexes. Men had significantly higher baseline weight (121 vs. 104 kg, p < 0.001), but BMI did not differ significantly. The average BMI for the overall cohort was 38 kg/m2 (SD 7.4); most participants (34%, n = 353) had obesity class III (≥ 40 kg/m2). The most prevalent adiposity‐related comorbidities were dyslipidemia (78%, n = 808), type 2 diabetes (74%, n = 774), and hypertension (72%, n = 752). Men were more likely to have dyslipidemia (87%, n = 386 vs. 71%, n = 422), diabetes (82%, n = 363 vs. 69%, n = 411), hypertension (83%, n = 368 vs. 65%, n = 384), and OSA (53%, n = 237 vs. 33%, n = 194), whereas women had higher rates of depression (37%, n = 221 vs. 23%, n = 101) and anxiety (33%, n = 195 vs. 22%, n = 99). Baseline cardiometabolic parameters revealed sex‐specific differences: women had higher total cholesterol, LDL‐cholesterol, and HDL‐cholesterol, while men had higher SBP, HbA1c, triglycerides, ALT, and AST. Nearly half of the cohort had a history of previous use of obesity pharmacotherapy. The highest dose of tirzepatide achieved varied, with the most frequently achieved dose being 15 mg subcutaneous weekly, used by over a third of participants. Approximately 70% of patients were on a high dose (≥ 10 mg weekly), with no significant difference in dose‐intensity patterns by sex.
TABLE 1.
Comparison of demographic and clinical characteristics stratified by sex among individuals using tirzepatide for ≥ 12 months.
| Variable | Full cohort (n = 1039) | Women (n = 594) | Men (n = 445) | p |
|---|---|---|---|---|
| Age, years (SD) | 55.6 (11.5) | 53.9 (11.5) | 57.8 (11.0) | < 0.001 |
| ≤ 45 years, n (%) | 198 (19%) | 134 (23%) | 64 (14%) | |
| 46–59 years, n (%) | 447 (43%) | 282 (48%) | 165 (37%) | |
| ≥ 60 years, n (%) | 394 (38%) | 178 (30%) | 216 (49%) | |
| Race, n (%) | 0.063 | |||
| White | 936 (90%) | 537 (90%) | 399 (90%) | |
| Black | 41 (4%) | 26 (4%) | 15 (3%) | |
| Asian | 31 (3%) | 20 (3%) | 11 (2%) | |
| Other | 31 (3%) | 11 (2%) | 20 (5%) | |
| Baseline body composition | ||||
| Weight, kg (SD) | 111.4 (24.7) | 104.2 (22.8) | 120.7 (24.2) | < 0.001 |
| BMI, kg/m2 (SD) | 38.2 (7.4) | 38.6 (7.8) | 37.7 (6.9) | 0.059 |
| Obesity category, n (%) | 0.24 | |||
| Overweight (≥ 27 kg/m2) | 103 (10%) | 59 (10%) | 44 (10%) | |
| Obesity class I (≥ 30 kg/m2) | 315 (30%) | 175 (29%) | 140 (31%) | |
| Obesity class II (≥ 35 kg/m2) | 267 (26%) | 143 (24%) | 124 (28%) | |
| Obesity class III (≥ 40 kg/m2) | 354 (34%) | 217 (37%) | 137 (31%) | |
| Adiposity‐related comorbidity, n (%) | ||||
| Dyslipidemia | 808 (78%) | 422 (71%) | 386 (87%) | < 0.001 |
| Type 2 diabetes | 774 (74%) | 411 (69%) | 363 (82%) | < 0.001 |
| Hypertension | 752 (72%) | 384 (65%) | 368 (83%) | < 0.001 |
| OSA | 431 (41%) | 194 (33%) | 237 (53%) | < 0.001 |
| GERD | 348 (33%) | 204 (34%) | 144 (32%) | 0.50 |
| Depression | 322 (31%) | 221 (37%) | 101 (23%) | < 0.001 |
| Anxiety | 294 (28%) | 195 (33%) | 99 (22%) | < 0.001 |
| MASLD | 194 (19%) | 114 (19%) | 80 (18%) | 0.62 |
| Baseline laboratories and vital signs | ||||
| SBP, mmHg (SD) | 127.5 (14.1) | 126.2 (14.0) | 129.2 (13.9) | < 0.001 |
| DBP, mmHg (SD) | 78.4 (9.1) | 78.4 (9.1) | 78.5 (9.0) | 0.83 |
| Fasting glucose, mg/dL (SD) | 147.1 (56.9) | 148.3 (61.7) | 145.4 (49.4) | 0.61 |
| HbA1c, % (SD) | 7.3 (1.5) | 7.2 (1.5) | 7.4 (1.5) | 0.041 |
| Triglycerides, mg/dL (SD) | 168.6 (98.7) | 160.2 (83.4) | 179.0 (114.0) | 0.007 |
| Total cholesterol, mg/dL (SD) | 163.5 (43.6) | 176.8 (44.3) | 147.2 (36.5) | < 0.001 |
| LDL‐cholesterol, mg/dL (SD) | 87.9 (36.6) | 98.0 (38.0) | 75.6 (30.7) | < 0.001 |
| HDL‐cholesterol, mg/dL (SD) | 46.3 (13.2) | 50.2 (13.6) | 41.6 (11.0) | < 0.001 |
| AST, U/L (SD) | 27.7 (14.1) | 25.9 (12.3) | 30.1 (15.9) | < 0.001 |
| ALT, U/L (SD) | 33.8 (22.8) | 29.3 (17.5) | 39.5 (27.0) | < 0.001 |
| Previous obesity medication, yes, n (%) | 362 (35%) | 205 (34%) | 157 (35%) | 0.80 |
| Concomitant weight gain‐promoting medication use, yes, n (%) | 323 (31.1%) | 149 (25.1%) | 174 (39.1%) | < 0.001 |
| Prior GLP‐1 RA exposure, n (%) | 456 (43.9%) | 252 (42.4%) | 204 (45.8%) | 0.27 |
| Tirzepatide dosing, n (%) | 0.91 | |||
| 2.5 mg weekly SQ | 31 (3%) | 18 (3%) | 13 (3%) | |
| 5 mg weekly SQ | 92 (9%) | 55 (10%) | 37 (8%) | |
| 7.5 mg weekly SQ | 183 (18%) | 109 (18%) | 74 (17%) | |
| 10 mg weekly SQ | 178 (17%) | 103 (17%) | 75 (17%) | |
| 12.5 mg weekly SQ | 187 (18%) | 101 (17%) | 86 (19%) | |
| 15 mg weekly SQ | 368 (35%) | 208 (35%) | 160 (36%) | |
| Tirzepatide low versus high dose, n (%) | 0.33 | |||
| Low dose (≤ 7.5 mg) | 306 (30%) | 182 (31%) | 124 (28%) | |
| High dose (≥ 10 mg) | 733 (70%) | 412 (69%) | 321 (72%) | |
Note: p values are reported for all comparisons, with bold values indicating statistical significance (p < 0.05). Continuous variables are reported as means and standard deviations (SD) with t‐test p values, while qualitative variables are shown as counts and percentages with Fisher's exact test p values.
Abbreviations: ALT, alanine aminotransferase; AST, aspartate aminotransferase; DBP, diastolic blood pressure; GERD, gastroesophageal reflux disease; GLP‐1 RA, glucagon‐like peptide‐1 receptor agonist.; HbA1c, hemoglobin A1C; HDL, high‐density lipoprotein; LDL, low‐density lipoprotein; MASLD, metabolic associated liver disease; OSA, obstructive sleep apnea; SBP, systolic blood pressure.
Table 2 summarizes data by age groups. Most patients were aged 46–59 years (43%, n = 447). Among the remaining patients, 19%, n = 198 were ≤ 45 years and 38%, n = 394 were ≥ 60 years. Women made up the majority of patients in the first two age groups, with 68% (n = 134) and 63% (n = 282), respectively; however, in the oldest group, most were men (55%, n = 216). There was a similar racial distribution across age groups, with most identified as White. There was a stepwise decline in both baseline weight and BMI across age groups (p < 0.001), with the youngest having the highest weight (119.6 kg, SD = 26.7) and BMI (41.0 kg/m2, SD = 8.0), followed by those aged 46–59 years (111.0 kg, SD = 26.1; BMI 38.1 kg/m2, SD = 7.7) and the ≥ 60 years group (107.6 kg, SD = 20.9; BMI 36.8 kg/m2, SD = 6.3). The youngest had a higher proportion of class III obesity at baseline (51%, n = 100 and 34%, n = 151 respectively), while the oldest (≥ 60 years) was mainly composed of individuals with class I obesity (≥ 30 kg/m2) (34%, n = 135) (p < 0.001). There was a stepwise increase in the baseline prevalence of dyslipidemia, type 2 diabetes, hypertension, OSA, and GERD across age groups, with the oldest having the highest rates (all p < 0.001). Anxiety and depression showed a decreasing trend with age, being more common in the youngest. Baseline cardiometabolic parameters varied across age groups. Total cholesterol, LDL‐cholesterol, ALT, and DBP showed a decline with age, while SBP increased (all p < 0.05). Fasting glucose, HbA1c, triglycerides, and AST did not differ significantly across groups. Regarding tirzepatide dosing, the most frequently achieved dose was 15 mg weekly, used by approximately one‐third of participants in all age groups. Nearly 70% of patients were on a high dose (≥ 10 mg weekly), with no significant difference in dose intensity by age groups (p = 0.111).
TABLE 2.
Comparison of demographic and clinical characteristics stratified by age groups among individuals using tirzepatide for ≥ 12 months.
| Variable | ≤ 45 years (n = 198) | 46–59 years (n = 447) | ≥ 60 years (n = 394) | p |
|---|---|---|---|---|
| Sex, n (%) | < 0.001 | |||
| Women | 134 (68%) | 282 (63%) | 178 (45%) | |
| Men | 64 (32%) | 165 (37%) | 216 (55%) | |
| Age, years (SD) | 38.3 (6.0) | 53.3 (3.9) | 66.8 (5.4) | < 0.001 |
| Race, n (%) | 0.039 | |||
| White | 174 (88%) | 396 (88%) | 366 (93%) | |
| Black | 8 (4%) | 21 (5%) | 12 (3%) | |
| Asian | 10 (5%) | 18 (4%) | 3 (1%) | |
| Other | 6 (3%) | 12 (3%) | 13 (3%) | |
| Baseline body composition | ||||
| Weight, kg (SD) | 119.6 (26.7) | 111.0 (26.1) | 107.6 (20.9) | < 0.001 |
| BMI kg/m2 (SD) | 41.0 (8.0) | 38.1 (7.7) | 36.8 (6.3) | < 0.001 |
| Obesity category, n (%) | < 0.001 | |||
| Overweight (≥ 27 kg/m2) | 11 (6%) | 43 (10%) | 49 (12%) | |
| Obesity class I (≥ 30 kg/m2) | 39 (20%) | 141 (31%) | 135 (35%) | |
| Obesity class II (≥ 35 kg/m2) | 48 (24%) | 112 (25%) | 107 (27%) | |
| Obesity class III (≥ 40 kg/m2) | 100 (50%) | 151 (34%) | 103 (26%) | |
| Adiposity‐related comorbidity, n (%) | ||||
| Dyslipidemia | 106 (53%) | 345 (77%) | 357 (91%) | < 0.001 |
| Type 2 diabetes | 123 (62%) | 325 (73%) | 326 (83%) | < 0.001 |
| Hypertension | 102 (51%) | 319 (71%) | 331 (84%) | < 0.001 |
| OSA | 51 (26%) | 180 (40%) | 200 (51%) | < 0.001 |
| GERD | 53 (27%) | 140 (31%) | 155 (39%) | 0.004 |
| Depression | 68 (34%) | 144 (32%) | 110 (28%) | 0.21 |
| Anxiety | 59 (30%) | 125 (28%) | 110 (28%) | 0.87 |
| MASLD | 35 (18%) | 93 (21%) | 66 (17%) | 0.30 |
| Baseline laboratories and vital signs | ||||
| SBP, mmHg (SD) | 124.9 (13.7) | 126.6 (13.2) | 129.7 (14.9) | < 0.001 |
| DBP, mmHg (SD) | 80.7 (9.6) | 79.5 (8.6) | 76.2 (8.8) | < 0.001 |
| Fasting glucose, mg/dL (SD) | 155.7 (70.2) | 148.5 (60.8) | 142.5 (45.2) | 0.24 |
| HbA1c, % (SD) | 7.1 (1.7) | 7.4 (1.6) | 7.3 (1.4) | 0.24 |
| Triglycerides, mg/dL (SD) | 166.0 (93.0) | 167.4 (102.0) | 170.9 (98.0) | 0.85 |
| Total cholesterol, mg/dL (SD) | 168.8 (39.9) | 167.9 (42.9) | 156.9 (45.0) | 0.001 |
| LDL‐cholesterol, mg/dL (SD) | 95.5 (33.5) | 91.2 (35.4) | 81.5 (38.2) | < 0.001 |
| HDL‐cholesterol, mg/dL (SD) | 43.7 (11.9) | 47.4 (13.9) | 46.4 (13.0) | 0.018 |
| AST, U/L (SD) | 28.2 (16.3) | 29.0 (15.5) | 26.1 (11.0) | 0.056 |
| ALT, U/L (SD) | 36.5 (30.0) | 36.0 (22.9) | 29.8 (16.9) | 0.002 |
| Previous obesity medication, yes, n (%) | 67 (33.8%) | 162 (36.2%) | 133 (33.8%) | 0.71 |
| Concomitant weight gain‐promoting medication use, yes, n (%) | 47 (23.7%) | 123 (27.5%) | 153 (38.8%) | < 0.001 |
| Prior GLP‐1 RA exposure, n (%) | 75 (37.9%) | 201 (45.0%) | 180 (45.7%) | 0.16 |
| Tirzepatide dosing, n (%) | 0.090 | |||
| 2.5 mg weekly SQ | 8 (4%) | 12 (3%) | 11 (3%) | |
| 5 mg weekly SQ | 15 (8%) | 31 (7%) | 46 (12%) | |
| 7.5 mg weekly SQ | 35 (18%) | 75 (17%) | 73 (19%) | |
| 10 mg weekly SQ | 29 (15%) | 76 (17%) | 73 (19%) | |
| 12.5 mg weekly SQ | 47 (24%) | 77 (17%) | 63 (16%) | |
| 15 mg weekly SQ | 64 (31%) | 176 (39%) | 128 (31%) | |
| Tirzepatide low versus high dose, n (%) | 0.111 | |||
| Low dose (≤ 7.5 mg) | 58 (29%) | 118 (26%) | 130 (33%) | |
| High dose (≥ 10 mg) | 140 (71%) | 329 (74%) | 264 (67%) | |
Note: p values are reported for all comparisons, with bold values indicating statistical significance (p < 0.05). Continuous variables are reported as means and standard deviations (SD) with t‐test p‐values, while qualitative variables are shown as counts and percentages with Fisher's exact test p‐value.
Abbreviations: ALT, alanine aminotransferase; AST, aspartate aminotransferase; BMI, body mass index; DBP, Diastolic Blood Pressure; GERD, gastroesophageal reflux disease; GLP‐1 RA, glucagon‐like peptide‐1 receptor agonist.; HbA1c, hemoglobin A1C; HDL, high‐density lipoprotein; LDL, low‐density lipoprotein; MASLD, metabolic associated liver disease; OSA, obstructive sleep apnea; SBP, systolic blood pressure.
3.2. Weight Loss Outcomes by Sex
In unadjusted analyses, women achieved significantly greater TBWL% at 3, 6, 9, 12, and 15 months (p < 0.05 for all time points) (Figure 2A). Across follow‐up, the proportion of patients achieving clinically meaningful TBWL% thresholds rose steadily over time for both sexes. Females were more likely than males to reach each threshold at nearly every time point, with the gap widening at higher loss thresholds (≥ 10%, ≥ 15%, and ≥ 20%) (Figure 2B and Table S1). Women were significantly more likely to achieve clinically meaningful weight loss thresholds: ≥ 5% at 3, 12, and 15 months; ≥ 10% and ≥ 15% TBWL at all assessed time points; and ≥ 20% from 6 months onward (Figure 2B and Table S1). At 12 and 15 months, compared with men, women were approximately 2‐fold more likely to achieve ≥ 15% TBWL and 2.6‐fold more likely to achieve ≥ 20% TBWL. Using a LMM that accounted for missing data and other covariates with potential influence on weight loss, the sex differences remained significant. The estimated marginal mean (standard error [SE]) for women versus men were 5.2% (0.39) versus 4.8% (0.36), p = 0.42 at 3 months; 9.0% (0.39) versus 7.5% (0.45), p < 0.001 at 6 months; 11.6% (0.39) versus 8.7% (0.45), p < 0.001 at 9 months; 13.7% (0.39) versus 9.9% (0.41), p < 0.001 at 12 months; and 14.8% (0.37) versus 11.0% (0.45), p < 0.001 at 15 months (Table S2).
FIGURE 2.
Weight loss outcomes by sex and age groups among individuals using tirzepatide for ≥ 12 months. (A) Total body weight loss percentage (TBWL%) trajectories for women and men at 3, 6, 9, 12, and 15 months. Plotted data represent group means and standard errors. p values derived from Independent two‐sample t‐tests. (B) Proportion of women and men achieving ≥ 5%, ≥ 10%, ≥ 15%, and ≥ 20% TBWL at 3, 6, 9, 12, and 15 months. Chi‐square tests were used to compare the proportions between groups (Table S1). (C) TBWL% trajectories for age groups at 3, 6, 9, 12, and 15 months. Plotted data represent group means and standard errors. p values derived from one‐way ANOVA. (D) Proportion of age groups achieving ≥ 5%, ≥ 10%, ≥ 15%, and ≥ 20% TBWL at 3, 6, 9, 12, and 15 months. Chi‐square tests were used to compare the proportions between groups (Table S1).
3.3. Weight Loss Outcomes by Age Groups
In unadjusted analyses, younger individuals achieved consistently and significantly greater TBWL% at 3, 6, 9, 12, and 15 months than older individuals (Figure 2C). Statistically significant differences were observed only between individuals aged ≤ 45 years and those aged ≥ 60 years, while TBWL% did not differ significantly between the ≤ 45 and 46–59 years groups or between the 46–59 and ≥ 60 years groups (Table S3). Across age groups, the proportion of patients achieving each TBWL% threshold increased steadily from 3 to 15 months (Figure 2D and Table S1). Early differences by age were modest, with younger patients (≤ 45 years) more frequently reaching ≥ 5% TBWL at 3 months and showing higher attainment of ≥ 10% TBWL at 6 and 9 months. For higher thresholds (≥ 15% and ≥ 20%), younger patients generally had greater attainment throughout follow‐up, with significant age‐group differences at 6 months for ≥ 15% and at 15 months for ≥ 20%. By 12 months, age‐group differences attenuated for ≥ 10% and ≥ 15%, suggesting convergence over time, though older adults (≥ 60 years) remained less likely to reach the most substantial weight loss, particularly ≥ 20% at later follow‐up. In the LMM that accounted for missingness and other covariates potentially influencing weight loss, TBWL% did not differ significantly across age groups (p = 0.35; Table S4).
3.4. Predictors of Weight Loss Outcomes
Table 3 summarizes predictors of TBWL% from the LMM, noting that more negative values reflect greater weight loss. In univariate models (Table 3A), female sex and age ≤ 45 years old were associated with significantly greater weight loss. Several clinical factors were associated with significantly less weight loss, including type 2 diabetes, prior obesity medication use, and concomitant weight gain‐promoting medication use. BMI and age as continuous variables, anxiety, depression, self‐reported positive dietary changes, and tirzepatide dose intensity were not significant in univariate analyses.
TABLE 3.
Predictors of total body weight loss among individuals using tirzepatide for ≥ 12 months.
| (A) Univariate linear mixed‐effects models for predictors of total body weight loss | |||
|---|---|---|---|
| Variable | Estimate | 95% CI | p |
| Time | |||
| 3 months | Reference | ||
| 6 months | −3.33 | −3.75, −2.93 | < 0.001 |
| 9 months | −5.34 | −5.76, −4.93 | < 0.001 |
| 12 months | −7.24 | −7.65, −6.83 | < 0.001 |
| 15 months | −8.19 | −8.60, −7.78 | < 0.001 |
| Sex, female | −2.96 | −3.83, −2.09 | < 0.001 |
| Age, years | 0.06 | 0.02, 0.10 | 0.002 |
| Age ≤ 45 years old, yes | −1.55 | −2.69, −0.41 | 0.008 |
| Baseline BMI, kg/m2 | 0.02 | −0.04, 0.08 | 0.50 |
| Type 2 diabetes, yes | 4.13 | 3.13, 5.13 | < 0.001 |
| Anxiety, yes | 0.35 | −0.62, 1.32 | 0.48 |
| Depression, yes | 0.21 | −0.74, 1.16 | 0.66 |
| Previous obesity medication use, yes | 2.56 | 1.65, 3.47 | < 0.001 |
| Concomitant weight gain‐promoting medication use, yes | 3.58 | 2.66, 4.50 | < 0.001 |
| Dietary change, yes a | −0.003 | −1.02, 1.01 | > 0.99 |
| Maximum dose achieved ≥ 10 mg, yes a | −0.69 | −1.66, 0.28 | 0.16 |
| (B) Multivariable linear mixed‐effects model for predictors of total body weight loss | |||
|---|---|---|---|
| Variable | Estimate | 95% CI | p |
| Time*sex interaction | < 0.001 | ||
| Female versus male | |||
| 3 months | −0.41 | −1.41, 0.58 | 0.42 |
| 6 months | −1.53 | −2.53, −0.54 | 0.002 |
| 9 months | −2.91 | −3.91, −1.91 | < 0.001 |
| 12 months | −3.40 | −4.37, −2.42 | < 0.001 |
| 15 months | −3.78 | −4.78, −2.79 | < 0.001 |
| Age, years | 0.02 | −0.02, 0.05 | 0.40 |
| BMI, kg/m2 | 0.02 | −0.04, 0.08 | 0.55 |
| Type 2 diabetes, yes | 2.73 | 1.67, 3.79 | < 0.001 |
| Previous obesity medication use, yes | 2.05 | 1.18, 2.92 | < 0.001 |
| Concomitant weight gain‐promoting medication use, yes | 1.76 | 0.79, 2.73 | < 0.001 |
| Maximum dose achieved ≥ 10 mg, yes a | −1.19 | −2.11, −0.27 | 0.012 |
Note: p values are reported for all comparisons, with bold values indicating statistical significance (p < 0.05). Since TBWL% values are negative in this dataset, positive estimates indicate less weight loss, whereas negative estimates indicate greater weight loss.
These variables were documented at any time during follow‐up, one response per patient, so are not true time‐varying covariates.
In the multivariable model (Table 3B), the association between sex and weight loss varied over time (time × sex interaction p < 0.001): differences in TBWL% were not significant at 3 months but became progressively larger in women than men at 6, 9, 12, and 15 months. After adjustment, age and baseline BMI were not associated with TBWL%, whereas type 2 diabetes, prior obesity medication use, and concomitant use of weight gain‐promoting medications remained independently associated with less weight loss (all p < 0.001). Higher maximum dose exposure (≥ 10 mg) was independently associated with greater weight loss (p = 0.012).
Univariate analyses (Figure S1) were followed by multivariable logistic mixed‐effects models (Table 4), which showed that the likelihood of achieving clinically meaningful weight loss thresholds increased over follow‐up time. Female sex was independently associated with higher odds of achieving ≥ 5%, ≥ 10%, and ≥ 15% TBWL (with a similar but nonsignificant trend for ≥ 20%). Across thresholds, type 2 diabetes was consistently associated with lower odds of achieving weight loss targets, and both prior obesity medication use and concomitant weight gain‐promoting medication use were associated with reduced likelihood of meeting ≥ 5% and ≥ 10% TBWL. Higher maximum dose exposure (≥ 10 mg) was associated with greater likelihood of achieving ≥ 5% TBWL but was not a significant predictor at higher thresholds. Age and baseline BMI were not independently associated with achieving any weight loss threshold.
TABLE 4.
Multivariable logistic mixed effects regression models for predictors of total body weight loss across time among individuals using tirzepatide for ≥ 12 months.
| Variable | Odds ratio | 95% CI | p |
|---|---|---|---|
| ≥ 5% | |||
| Time | |||
| 3 months | 0.02 | 0.01–0.03 | < 0.001 |
| 6 months | 0.13 | 0.08–0.21 | < 0.001 |
| 9 months | 0.29 | 0.18–0.45 | < 0.001 |
| 12 months | 0.61 | 0.39–0.95 | 0.029 |
| 15 months | Reference | ||
| Sex, female | 2.01 | 1.09–3.69 | 0.024 |
| Age, each year | 0.98 | 0.96–1.01 | 0.16 |
| BMI, each kg/m2 | 0.97 | 0.93–1.01 | 0.096 |
| Type 2 diabetes, yes | 0.14 | 0.06–0.30 | < 0.001 |
| Previous obesity medication use, yes | 0.24 | 0.13–0.46 | < 0.001 |
| Concomitant weight gain‐promoting medication use, yes | 0.23 | 0.11–0.46 | < 0.001 |
| Maximum dose achieved ≥ 10 mg, yes a | 2.03 | 1.05–3.93 | 0.034 |
| ≥ 10% | |||
| Time | |||
| 3 months | < 0.01 | < 0.01–< 0.01 | < 0.001 |
| 6 months | 0.05 | 0.03–0.09 | < 0.001 |
| 9 months | 0.26 | 0.17–0.40 | < 0.001 |
| 12 months | 0.59 | 0.39–0.89 | 0.012 |
| 15 months | Reference | ||
| Sex, female | 13.61 | 5.79–31.97 | < 0.001 |
| Age, each year | 0.99 | 0.96–1.02 | 0.52 |
| BMI, each kg/m2 | 0.97 | 0.93–1.02 | 0.29 |
| Type 2 diabetes, yes | 0.10 | 0.04–0.27 | < 0.001 |
| Previous obesity medication use, yes | 0.18 | 0.08–0.40 | < 0.001 |
| Concomitant weight gain‐promoting medication use, yes | 0.26 | 0.11–0.62 | 0.002 |
| Maximum dose achieved ≥ 10 mg, yes a | 1.96 | 0.88–4.37 | 0.099 |
| ≥ 15% | |||
| Time | |||
| 3 months | < 0.01 | < 0.01–< 0.01 | < 0.001 |
| 6 months | < 0.01 | < 0.01–< 0.01 | < 0.001 |
| 9 months | 0.02 | 0.01–0.05 | < 0.001 |
| 12 months | 0.28 | 0.15–0.53 | < 0.001 |
| 15 months | Reference | ||
| Sex, female | 5.94 | 1.83–19.25 | 0.003 |
| Age, each year | 0.99 | 0.94–1.03 | 0.55 |
| BMI, each kg/m2 | 0.99 | 0.92–1.06 | 0.76 |
| Type 2 diabetes, yes | 0.14 | 0.03–0.63 | 0.011 |
| Previous obesity medication use, yes | 0.36 | 0.11–1.17 | 0.090 |
| Concomitant weight gain‐promoting medication use, yes | 0.51 | 0.15–1.70 | 0.27 |
| Maximum dose achieved ≥ 10 mg, yes a | 1.72 | 0.51–5.84 | 0.39 |
| ≥ 20% | |||
| Time | |||
| 3 months | < 0.01 | < 0.01–< 0.01 | < 0.001 |
| 6 months | < 0.01 | < 0.01–< 0.01 | < 0.001 |
| 9 months | 0.02 | 0.01–0.06 | < 0.001 |
| 12 months | 0.17 | 0.08–0.38 | < 0.001 |
| 15 months | Reference | ||
| Sex, female | 3.47 | 0.99–12.23 | 0.053 |
| Age, each year | 0.99 | 0.94–1.04 | 0.72 |
| BMI, each kg/m2 | 1.00 | 0.92–1.08 | 0.97 |
| Type 2 diabetes, yes | 0.35 | 0.09–1.33 | 0.12 |
| Previous obesity medication use, yes | 0.42 | 0.12–1.50 | 0.18 |
| Concomitant weight gain‐promoting medication use, yes | 0.59 | 0.14–2.48 | 0.47 |
| Maximum dose achieved ≥ 10 mg, yes a | 1.55 | 0.41–5.84 | 0.51 |
Note: Odds ratios > 1 indicate greater likelihood of achieving greater weight loss, while odds ratios < 1 reflect lower likelihood of greater weight loss.
Abbreviation: BMI, body mass index.
These variables were documented at any time during follow‐up, one response per patient, so are no true time‐varying covariates.
Given the differences in type 2 diabetes rate between sexes, sensitivity analyses were performed to address potential confounding. These analyses identified a significant interaction between sex, type 2 diabetes status, and time (p < 0.001), indicating that weight loss trajectories differed according to the combined effects of sex and type 2 diabetes status. Stratified analyses showed that women experienced greater weight loss than men in both the type 2 diabetes and non‐type 2 diabetes groups, although these differences were more pronounced among participants without type 2 diabetes (Tables S5 and S6).
Similarly, given the potential confounding effect of prior obesity medication use, particularly previous GLP‐1 RA exposure, additional sensitivity analyses were restricted to GLP‐1 RA–naïve individuals and showed that women continued to experience greater weight loss than men, a difference that was also more pronounced among participants without type 2 diabetes, supporting the robustness of the primary findings (Tables S7 and S8).
4. Discussion
This study is one of the first to evaluate the influence of sex and age on weight loss outcomes in a real‐world cohort of patients maintained on long‐term, continuous therapy with tirzepatide. After 15 months of treatment, women experienced approximately 41% greater TBWL than men. Furthermore, women were significantly more likely to achieve clinically meaningful weight loss thresholds. At 12 and 15 months, women were approximately twice as likely as men to achieve ≥ 15% TBWL and about 2.6 times as likely to achieve ≥ 20% TBWL. When stratified by age, TBWL% appeared to decrease in a stepwise manner across the three age groups, with differences primarily driven by comparisons between those ≤ 45 years and those ≥ 60 years. At 15 months, individuals aged ≤ 45 years were approximately 1.6 times more likely to achieve ≥ 20% TBWL than older participants. Notably, all age groups achieved clinically meaningful weight loss. The apparent age gradient observed in unadjusted analyses may be attributable to confounding factors, such as higher baseline BMI and a lower rate of type 2 diabetes among younger adults, given that after adjustment, age was not independently associated with TBWL%. In contrast, sex in conjunction with diabetes status, previous use of obesity medications, and concomitant use of weight gain‐promoting medications emerged as strong and consistent predictors of weight loss outcomes with continuous, long‐term tirzepatide use.
The greater weight loss observed in women aligns with prior research demonstrating sex differences in response to GLP‐1–based medications for obesity treatment [10, 11]. Our findings showed greater differences than analyses from the SURMOUNT program, which reported approximately 30% greater weight loss in women compared to men after 72 weeks of treatment (~16 months), whereas the difference in our cohort was 41% after 15 months [8]. Furthermore, in our cohort and after 15 months of treatment, the proportion of women achieving ≥ 15% TBWL was approximately twofold higher than that of men (49% vs. 25%). By comparison, post hoc analyses of the SURMOUNT trials reported more modest sex differences: after 72 weeks of treatment (~16 months), the proportion of participants achieving ≥ 15% TBWL was ~30% higher in women than men in SURMOUNT‐1 (84% vs. 66%) and ~50% higher in SURMOUNT‐2 (62% vs. 41%) [12].
This study demonstrates that while all age groups achieved clinically significant weight loss, notable differences were observed between the youngest and oldest adult groups, highlighting a potential age‐related dimorphism in response to treatment. However, it is important to note that our regression analyses revealed no significant association between age and weight loss outcomes, suggesting that other factors such as sex, diabetes status, and prior obesity medication use may be driving these differences. A prior study derived from the SURPASS program showed that younger age was an independent predictor of ≥ 15% TBWL [9]. Despite this, post hoc analyses from the SURMOUNT program found no differences in body composition outcomes when comparing age groups (< 50, 50 to < 65, or ≥ 65 years) [13].
The significant differences in weight loss outcomes between sexes likely reflect the influence of multiple biological and behavioral factors. Although further studies are needed to confirm and identify the underlying mechanisms driving these sex‐based differences, literature suggests that these differences may be explained, at least in part, by estrogen‐driven effects, which interact with GLP‐1 analogues through important biochemical pathways [14]. Preclinical data from rodent models have shown that estrogen can enhance the appetite‐suppressing effects of GLP‐1 RAs, increase energy expenditure, and reduce food‐motivated behavior. Notably, these studies suggest that estrogen amplifies the reward‐reducing properties of GLP‐1 RAs; for example, the satiating effects of GLP‐1 are diminished following ovariectomy in female rodents but are restored with estrogen replacement [14]. Clinically, although these medications have been shown to be effective across all reproductive stages in women, the role of estrogen across the reproductive stages in women remains to be fully elucidated [15]. Emerging observational data have shown an association between menopause hormone therapy use in postmenopausal women with overweight or obesity and superior weight loss outcomes with tirzepatide and semaglutide, reinforcing the preclinical evidence and indicating that estrogen may modulate the effects of GLP‐1–based therapies. Given the small cohort sizes and methodological limitations inherent to the retrospective nature of these studies, future research is warranted to systematically evaluate the role of hormone therapy as a potential modifier of weight loss response to GLP‐1 analogues [16, 17].
Additional contributors, particularly sex‐related differences in behavior and context, remained unexplored. These may include differential engagement in lifestyle modifications, treatment expectations, and motivation. Although sex differences in adherence could also influence outcomes, our cohort was restricted to individuals maintained on long‐term, continuous tirzepatide therapy, which reduces the likelihood that discontinuation drove the observed differences. Notably, prior data on GLP‐1 RA discontinuation patterns suggest that female sex may be associated with a higher risk of discontinuation [7, 18]. Finally, while sex‐based differences in weight loss with GLP‐1–based medications have been described, corresponding differences in cardiometabolic outcomes have been less consistent across studies [14, 19, 20, 21].
The consistent weight loss achieved with tirzepatide across all age groups underscores its broad clinical applicability for obesity treatment. Beyond its impact on weight loss, tirzepatide may also support healthier aging by reducing obesity‐related health risks and improving quality of life [22]. Although age, as a continuous variable, was not significantly associated with TBWL%, an attenuation in weight loss was observed in older age groups. This may potentially reflect age‐related factors such as decline in metabolism, reduced resting energy expenditure, higher rates of sarcopenia, and cumulative allostatic load [23, 24, 25]. Additionally, older adults often have a higher comorbidity burden, such as type 2 diabetes, which may contribute to a blunted weight loss response [26]. Importantly, the 12% TBWL observed in the oldest group remains clinically meaningful, but it raises concerns about potential exacerbation of age‐related losses in muscle and bone mass [27, 28]. This highlights the importance of pairing pharmacologic therapy with resistance and aerobic exercise, as well as adequate macro‐ and micronutrient intake, to mitigate the risk of sarcopenia, frailty, and functional decline [29, 30].
In addition to age and sex, our analysis identified other key predictors of weight loss response. Consistent with published data, history of type 2 diabetes, prior use of obesity medications, and concomitant use of weight gain‐promoting medications were associated with inferior weight loss outcomes, while higher doses of tirzepatide were associated with greater weight loss [26, 31, 32, 33]. The reduced weight loss observed in patients with type 2 diabetes may be attributed to the metabolic abnormalities inherent to the disease, such as insulin resistance, or the concomitant use of diabetes medications that promote weight gain [26, 34]. Overall, the concomitant use of weight gain‐promoting medications prescribed for comorbidities or treatment‐related complications may further blunt weight loss response, underscoring the need to systematically evaluate and minimize such therapies when clinically feasible [35]. Furthermore, the diminished response among those with previous obesity medication use may be partly explained by patients reaching a weight loss plateau before starting tirzepatide or metabolic adaptations that limit further loss, or it may potentially reflect a history of challenges with medication adherence, tolerance, or biological resistance to pharmacotherapy [36]. Although baseline BMI initially appeared to influence weight loss response based on univariate logistic analyses, it did not remain a strong predictive value in multivariable models, suggesting that other factors may play a more critical role in determining individual responses [37].
This study has several strengths, including a large sample size and the use of real‐world outcomes in a more heterogeneous patient population compared to randomized controlled trials. It only includes patients maintained on long‐term (at least 12 months), continuous therapy with tirzepatide, allowing minimization of certain confounders, such as sex differences in medication persistence or discontinuation. Importantly, missing weight measurements at various time points were appropriately addressed using linear mixed‐effects models, which are well suited for handling incomplete longitudinal data and enable a robust analysis of weight loss trajectories over time. However, this is an observational study, not designed to infer causality. Results may be influenced by residual confounding, and other limitations should be considered when interpreting the findings. First, the nature of the study design inherently limits the precision of medication start and discontinuation. Additionally, although all patients received lifestyle intervention advice, the absence of detailed records on diet and physical activity limited our ability to assess the contribution of lifestyle behaviors. Specifically, while we considered self‐reported changes in eating behaviors and choices, we did not consider dietitian consultation as it may reflect biased referral patterns rather than true dietary modification, with referral more common among patients experiencing suboptimal weight loss. Future prospective studies with standardized dietary and lifestyle assessments will be required to fully account for the impact of eating behavior on weight loss outcomes, as women may demonstrate greater adherence to lifestyle modifications than men. Notably, a recent systematic review and meta‐analysis, which aimed to quantify the heterogeneity of treatment effects (HTE) of GLP‐1–based therapies and included only randomized controlled trials with more rigorous control of adherence to lifestyle interventions, reported similar results: the efficacy of GLP‐1 RAs was consistently greater among women than men [38]. These findings indicate that observed sex differences in weight loss response to GLP‐1 RAs are unlikely to be solely attributable to differential adherence to lifestyle interventions. Furthermore, dosing information reflects the dose at last follow‐up rather than time‐specific exposure, warranting systematic longitudinal dose capture to evaluate dose–response relationships and determine whether observed differences persist within comparable dosing strata. A potential overestimation of treatment effect may have been introduced by including only patients who completed at least 12 months of tirzepatide treatment, as individuals who discontinued therapy early due to side effects or lack of efficacy were excluded. However, this study was not designed to assess the effectiveness of tirzepatide, but instead to examine the influence of sex and age on the response to long‐term, continuous therapy with tirzepatide. Although several factors that may influence weight loss response were identified, including comorbidities such as type 2 diabetes, anxiety, and depression, these conditions may not be fully or accurately captured in the EHR. This limitation is particularly relevant for mental health diagnoses, which may have been established outside the study institution, reflecting an inherent constraint of EHR‐based research. Lastly, the study population was predominantly White, which limits the generalizability of our findings to more diverse populations.
5. Conclusion
In this large real‐world analysis of over 1000 patients treated with tirzepatide for obesity, sex emerged as a strong and consistent predictor of weight loss outcomes. Women achieved significantly greater TBWL than men at all time points. Although younger individuals experienced modestly greater weight loss than older adults, age was not an independent predictor in multivariable analyses. Instead, the most influential predictors were sex, diabetes status, prior obesity medication use, concomitant use of a weight gain‐promoting medication, and tirzepatide dose achieved. These findings align with prior clinical trial data and highlight the importance of considering patient‐specific factors when tailoring obesity treatment strategies.
The observed sex‐based differences, potentially influenced by hormonal, behavioral, or pharmacodynamic factors, should be interpreted as associations and merit further investigation to better inform individualized treatment. Tirzepatide demonstrated consistent and clinically meaningful efficacy across all age groups, underscoring its utility as a powerful pharmacologic option for weight management.
Future prospective studies including more diverse populations and comprehensive lifestyle data are needed to validate and expand upon these findings. Overall, this study adds to the growing body of evidence supporting tirzepatide's effectiveness and highlights key predictors of response that can guide more personalized, equitable, and effective obesity care.
Author Contributions
R.C., D.B., R.J.R.G., M.A.E., J.V., E.T., N.S., S.F., and M.D.H.A. contributed to the study design. R.C., D.B., R.J.R.G., M.A.E., J.V., E.T., N.S., S.F., A.V., and A.W.M. collected the data. R.C., A.V., K.C.C., S.J.W., and M.D.H.A. performed statistical analyses. R.C., D.B., R.J.R.G., M.A.E., J.V., E.T., N.S., S.F., A.V., A.W.M., K.C.C., S.J.W., and M.D.H.A. had full access to all study data. All authors participated in data interpretation, manuscript writing, and critical review of the manuscript and approved the final version of the manuscript for submission. M.D.H.A. is the guarantor of this work and, as such, takes responsibility for the integrity of the data and the accuracy of the analysis.
Funding
This work was supported by Mayo Clinic Center for Women's Health Research. The funding source had no role in the design of this study, execution, analysis, interpretation of the data, or decision to submit results. The Mayo Clinic Center for Women's Health Research provided protected research time for M.D.H.A to complete this research project.
Conflicts of Interest
A.A. has research technologies licensed by Gila Therapeutics and Phenomix Sciences from the University of Florida and Mayo Clinic. He has received consultant fees in the past 5 years from Rhythm Pharmaceuticals, Gila Therapeutics, Amgen, General Mills, Regeneron, Boehringer Ingelheim, Novo Nordisk, Currax, Nestlé, Phenomix Sciences, Bausch Health, and RareDiseases. He receives research funding from Vivus Pharmaceuticals, Novo Nordisk, Apollo Endosurgery, Satiogen Pharmaceuticals, Spatz Medical, Rhythm Pharmaceuticals, Regeneron, and Boehringer Ingelheim. S.S.F. is a consultant for Era Women's Health Platform, delivers CME lectures for PriMed, AiCME, MedAll, and Medscape, and serves on the scientific advisory board for Weight Watchers. C.L.S. is an advisor for Bayer Pharmaceutics. M.D.H.A. has served as a consultant for Novo Nordisk, Hoffman‐La Roche, and Verge Genomics. M.D.H.A. receives research funding from the Mayo Clinic Center for Women's Health Research, Eli Lilly, Endogenex, and Phenomix Sciences. The other authors declare no conflicts of interest.
Supporting information
Table S1: Weight loss distribution by sex and age groups among individuals using tirzepatide for ≥ 12 months.
Table S2: Estimated total body weight loss percentage by sex among individuals using tirzepatide for ≥ 12 months with linear mixed model.
Table S3: Pairwise comparisons for total body weight loss percentage across all time points between age groups among individuals using tirzepatide for ≥ 12 months.
Table S4: Estimated total body weight loss percentage by age group among individuals using tirzepatide for ≥ 12 months with linear mixed model.
Table S5:. Predictors of total body weight loss among individuals using tirzepatide for ≥ 12 months stratified by type 2 diabetes status.
Table S6: Multivariable logistic mixed effects regression models for predictors of total body weight loss across time among individuals using tirzepatide for ≥ 12 months stratified by type 2 diabetes status.
Table S7: Predictors of total body weight loss among GLP‐1 RA‐naïve individuals using tirzepatide for ≥ 12 months stratified by type 2 diabetes.
Table S8: Multivariable logistic mixed effects regression models for predictors of total body weight loss across time among GLP‐1 RA‐naïve individuals using tirzepatide for ≥ 12 months stratified by type 2 diabetes.
Figure S1: Univariate logistic model for predictors of total body weight loss among individuals using tirzepatide for ≥ 12 months. The forest plot depicts the univariate logistic regression models for categorical TBWL%. Bars indicate the 95% confidence intervals (CIs); filled circles denote statistically significant associations (p < 0.05), while unfilled circles represent nonsignificant results. The high dose tirzepatide group includes individuals who were using tirzepatide ≥ 10 mg weekly at the last follow‐up.
Data Availability Statement
Deidentified individual participant data and dictionary will be made available with publication upon request.
References
- 1. “Obesity and Overweight,” World Health Organization. accessed December 8, 2025, https://www.who.int/news‐room/fact‐sheets/detail/obesity‐and‐overweight.
- 2. Pantalone K. M., Hobbs T. M., Chagin K. M., et al., “Prevalence and Recognition of Obesity and Its Associated Comorbidities: Cross‐Sectional Analysis of Electronic Health Record Data From a Large US Integrated Health System,” BMJ Open 7 (2017): e017583. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Wang Y., Beydoun M. A., Min J., Xue H., Kaminsky L. A., and Cheskin L. J., “Has the Prevalence of Overweight, Obesity and Central Obesity Levelled Off in the United States? Trends, Patterns, Disparities, and Future Projections for the Obesity Epidemic,” International Journal of Epidemiology 49 (2020): 810–823. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Ward Z. J., Bleich S. N., Cradock A. L., et al., “Projected US State‐Level Prevalence of Adult Obesity and Severe Obesity,” New England Journal of Medicine 381 (2019): 2440–2450. [DOI] [PubMed] [Google Scholar]
- 5. Li Y., Du X., Wu Y., et al., “Estimates and Projections in the Economic Impacts of Fifteen Dietary Risk Factors for Two Hundred Four Countries and Territories From 2020 to 2050: A Health‐Augmented Macroeconomic Modeling Study,” American Journal of Clinical Nutrition 121 (2025): 1099–1108. [DOI] [PubMed] [Google Scholar]
- 6. 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 (2022): e009773. [DOI] [PMC free article] [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 (2022): 205–216. [DOI] [PubMed] [Google Scholar]
- 8. Chao A. M. T. R., Mojdami D., Forrester T., Chigutsa F. B., and Malik R., “Body Weight Reduction Associated With Tirzepatide by Sex: A Subgroup Analysis of the SURMOUNT Clinical Trials,” 2024, Presented at Endocrine Society, Boston, MA.
- 9. Małecki M. T., Batterham R. L., Sattar N., et al., “Predictors of ≥15% Weight Reduction and Associated Changes in Cardiometabolic Risk Factors With Tirzepatide in Adults With Type 2 Diabetes in SURPASS 1–4,” Diabetes Care 46 (2023): 2292–2299. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Powell W., Song X., Mohamed Y., et al., “Medications and Conditions Associated With Weight Loss in Patients Prescribed Semaglutide Based on Real‐World Data,” Obesity (Silver Spring) 31 (2023): 2482–2492. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Pearson E. R., Del Prato S., Pavo I., et al., “Predictors of Initial and Sustained Glycemic and Weight Response to Tirzepatide: A Post Hoc Analysis of SURPASS‐4,” Diabetes 74 (2025): 1850–1862. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. García‐Pérez L. E., Chao A., Taylor R., et al., “Body Weight Reduction With Tirzepatide by Sex: A Subgroup Analysis of the SURMOUNT Clinical Trials,” Diabetologia 67, no. S1 (2024): 361–362. [Google Scholar]
- 13. Look M., Dunn J. P., Kushner R. F., et al., “Body Composition Changes During Weight Reduction With Tirzepatide in the SURMOUNT‐1 Study of Adults With Obesity or Overweight,” Diabetes, Obesity & Metabolism 27 (2025): 2720–2729. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Börchers S. and Skibicka K. P., “GLP‐1 and Its Analogs: Does Sex Matter?,” Endocrinology 166 (2025): bqae165. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Tchang B. G., Mihai A. C., Stefanski A., et al., “Body Weight Reduction in Women Treated With Tirzepatide by Reproductive Stage: A Post Hoc Analysis From the SURMOUNT Program,” Obesity (Silver Spring) 33 (2025): 851–860. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Castaneda R., Bechenati D., Tama E., et al., “The Role of Menopause Hormone Therapy in Modulating Tirzepatide‐Associated Weight Loss in Postmenopausal Women With Overweight or Obesity: A Retrospective Cohort Study,” Lancet Obstetrics, Gynaecology, & Women's Health 2 (2026): e118–e128. [Google Scholar]
- 17. Hurtado M. D., Tama E., Fansa S., et al., “Weight Loss Response to Semaglutide in Postmenopausal Women With and Without Hormone Therapy Use,” Menopause 31 (2024): 266–274. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. 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 (2025): e2457349. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Hanlon P., Butterly E., Wei L., et al., “Age and Sex Differences in Efficacy of Treatments for Type 2 Diabetes: A Network Meta‐Analysis,” JAMA 333 (2025): 1062–1073. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Shamim M. A., Patil A. N., Amin U., et al., “Glucagon‐Like Peptide‐1 Receptor Agonists in Adolescents With Overweight or Obesity With or Without Type 2 Diabetes Multimorbidity—A Systematic Review and Network Meta‐Analysis,” Diabetes, Obesity & Metabolism 26 (2024): 4302–4317. [DOI] [PubMed] [Google Scholar]
- 21. American Diabetes Association Professional Practice Committee , “13. Older Adults: Standards of Care in Diabetes—2025,” Diabetes Care 48, no. 1S1 (2025): S266–S282. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Gibble T. H., Cao D., Forrester T., Fraseur Brumm J., and Chao A. M., “Tirzepatide and Health‐Related Quality of Life in Adults With Obesity or Overweight: Results From the SURMOUNT‐3 Phase 3 Randomized Trial,” Diabetes, Obesity & Metabolism 27 (2025): 4268–4279. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Bosy‐Westphal A., Eichhorn C., Kutzner D., Illner K., Heller M., and Müller M. J., “The Age‐Related Decline in Resting Energy Expenditure in Humans Is due to the Loss of Fat‐Free Mass and to Alterations in Its Metabolically Active Components,” Journal of Nutrition 133 (2003): 2356–2362. [DOI] [PubMed] [Google Scholar]
- 24. Manini T. M., “Energy Expenditure and Aging,” Ageing Research Reviews 9 (2010): 1–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Tremblay A. and Chaput J.‐P., “Obesity: The Allostatic Load of Weight Loss Dieting,” Physiology & Behavior 106 (2012): 16–21. [DOI] [PubMed] [Google Scholar]
- 26. Castaneda R., Sepulveda D., Rivera Gutierrez R., et al., “Leveraging the Individualized Metabolic Surgery Score to Predict Weight Loss With Tirzepatide in Adults With Type 2 Diabetes and Obesity,” Diabetology 7 (2026): 10. [Google Scholar]
- 27. Jiao R., Lin C., Cai X., et al., “Characterizing Body Composition Modifying Effects of a Glucagon‐Like Peptide 1 Receptor‐Based Agonist: A Meta‐Analysis,” Diabetes, Obesity & Metabolism 27 (2025): 259–267. [DOI] [PubMed] [Google Scholar]
- 28. McCarthy D. and Berg A., “Weight Loss Strategies and the Risk of Skeletal Muscle Mass Loss,” Nutrients 13 (2021): 2473. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Hashimoto Y., Takahashi F., Okamura T., Hamaguchi M., and Fukui M., “Diet, Exercise, and Pharmacotherapy for Sarcopenia in People With Diabetes,” Metabolism 144 (2023): 155585. [DOI] [PubMed] [Google Scholar]
- 30. Mehrtash F., Dushay J., and Manson J. E., “Integrating Diet and Physical Activity When Prescribing GLP‐1s—Lifestyle Factors Remain Crucial,” JAMA Internal Medicine 185 (2025): 1151. [DOI] [PubMed] [Google Scholar]
- 31. Moiz A., Levett J. Y., Filion K. B., Peri K., Reynier P., and Eisenberg M. J., “Long‐Term Efficacy and Safety of Once‐Weekly Semaglutide for Weight Loss in Patients Without Diabetes: A Systematic Review and Meta‐Analysis of Randomized Controlled Trials,” American Journal of Cardiology 222 (2024): 121–130. [DOI] [PubMed] [Google Scholar]
- 32. Ghusn W., Fansa S., Anazco D., et al., “The Association Between Previous Use of Anti‐Obesity Medication and Semaglutide Weight Loss Outcomes,” Diabetes, Obesity & Metabolism 26 (2024): 2167–2175. [DOI] [PubMed] [Google Scholar]
- 33. Ghusn W., Anazco D., Fansa S., et al., “Weight Loss Outcomes With Semaglutide Based on Diabetes Severity Using the Individualized Metabolic Surgery Score,” EClinicalMedicine 72 (2024): 102625. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Várkonyi T. T., Pósa A., Pávó N., and Pavo I., “Perspectives on Weight Control in Diabetes ‐ Tirzepatide,” Diabetes Research and Clinical Practice 202 (2023): 110770. [DOI] [PubMed] [Google Scholar]
- 35. Nadolsky K., Garvey W. T., Agarwal M., et al., “American Association of Clinical Endocrinology Consensus Statement: Algorithm for the Evaluation and Treatment of Adults With Obesity/Adiposity‐Based Chronic Disease–2025 Update,” Endocrine Practice 31 (2025): 1351–1394. [DOI] [PubMed] [Google Scholar]
- 36. Kadrić S. I., Ćesić A. K., and Dujić T., “Pharmacogenetics of New Classes of Antidiabetic Drugs,” Bosnian Journal of Basic Medical Sciences 21 (2021): 659–671. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Ma J., Liu M., Wang R., Du L., and Ji L., “Efficacy and Safety of Tirzepatide in People With Type 2 Diabetes by Baseline Body Mass Index: An Exploratory Subgroup Analysis of SURPASS‐AP‐Combo,” Diabetes, Obesity & Metabolism 26 (2024): 1454–1463. [DOI] [PubMed] [Google Scholar]
- 38. Alexander G. C., Xiao X., Dilek S., et al., “Heterogeneity of Treatment Effects of Glucagon‐Like Peptide‐1 Receptor Agonists for Weight Loss in Adults: A Systematic Review and Meta‐Analysis,” JAMA Internal Medicine 186 (2026): 567–577. [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
Table S1: Weight loss distribution by sex and age groups among individuals using tirzepatide for ≥ 12 months.
Table S2: Estimated total body weight loss percentage by sex among individuals using tirzepatide for ≥ 12 months with linear mixed model.
Table S3: Pairwise comparisons for total body weight loss percentage across all time points between age groups among individuals using tirzepatide for ≥ 12 months.
Table S4: Estimated total body weight loss percentage by age group among individuals using tirzepatide for ≥ 12 months with linear mixed model.
Table S5:. Predictors of total body weight loss among individuals using tirzepatide for ≥ 12 months stratified by type 2 diabetes status.
Table S6: Multivariable logistic mixed effects regression models for predictors of total body weight loss across time among individuals using tirzepatide for ≥ 12 months stratified by type 2 diabetes status.
Table S7: Predictors of total body weight loss among GLP‐1 RA‐naïve individuals using tirzepatide for ≥ 12 months stratified by type 2 diabetes.
Table S8: Multivariable logistic mixed effects regression models for predictors of total body weight loss across time among GLP‐1 RA‐naïve individuals using tirzepatide for ≥ 12 months stratified by type 2 diabetes.
Figure S1: Univariate logistic model for predictors of total body weight loss among individuals using tirzepatide for ≥ 12 months. The forest plot depicts the univariate logistic regression models for categorical TBWL%. Bars indicate the 95% confidence intervals (CIs); filled circles denote statistically significant associations (p < 0.05), while unfilled circles represent nonsignificant results. The high dose tirzepatide group includes individuals who were using tirzepatide ≥ 10 mg weekly at the last follow‐up.
Data Availability Statement
Deidentified individual participant data and dictionary will be made available with publication upon request.