T-REx: resistance training limited lean-mass loss on tirzepatide

A 40-week randomized trial in Perth found that adding supervised resistance training to tirzepatide reduced the loss of lean soft tissue and reversed a decline in strength, while mean weight loss was similar in both groups. The T-REx study, led by Daniel J. Green, Andrew Haynes, and senior author Bu B. Yeap at the University of Western Australia, was published in Nature Medicine on Nov. 9, 2026. The authors say that, to their knowledge, it is the first randomized trial to test resistance exercise together with tirzepatide for overweight and obesity.

How the trial was run

The paper describes a two-arm, single-center study registered as ACTRN12623001236684. One hundred twenty adults aged 50 to 70 years — 60 men and 60 postmenopausal women — entered the trial (61 assigned to tirzepatide plus usual activities, 59 to tirzepatide plus resistance exercise). One additional person was randomized but withdrew before any study injections or training and was not among the 120. Participants had obesity, or overweight with at least one weight-related comorbidity, and did not have diabetes or cardiovascular disease. They were also required to have been relatively inactive at entry.

Everyone received open-label tirzepatide for 40 weeks. People in the training arm attended three one-hour, supervised, center-based sessions each week; the comparison arm continued usual activities and came in weekly for injections and the same assessment schedule. Body composition was measured by DXA at baseline and every eight weeks through week 40. Co-primary endpoints were changes in fat mass and in total lean soft tissue mass. Regional strength — bench press, lat pulldown, and knee extension, each as a one-repetition maximum — was a key secondary outcome. Analyses were by intention to treat, using linear mixed models.

The published abstract states: “Tirzepatide was safe, and most participants tolerated up-titration to 15 mg weekly.” Mean adherence to the thrice-weekly sessions was 91% (range 64 to 100%). Of the 120 participants, 111 (93%) completed the trial. Gastrointestinal adverse events, including nausea in about half of participants, were self-reported at similar frequencies in both arms. The authors reported six serious adverse events, five in the usual-activity arm and one in the training arm, all assessed as unrelated to tirzepatide or, in the training arm, to the exercise program.

Weight, fat, and lean soft tissue

Mean change in body weight from baseline to week 40 was −22.61 kg in the tirzepatide-only arm and −22.28 kg in the tirzepatide-plus-training arm (P=0.694). Mean fat-mass change, a co-primary outcome, was −17.19 kg versus −17.98 kg (group × time P=0.700). The between-group difference in fat-mass change was −0.77 kg. In the Results, that interval is reported as a 97.5% confidence interval of −2.14 to 0.60 kg (Bonferroni-adjusted P=0.416), the correction the Methods apply to the two co-primary outcomes. The abstract describes the same interval as a 95% CI.

For total lean soft tissue mass, the other co-primary outcome, mean change from baseline was −4.67 kg without the training program and −3.57 kg with it. The between-group difference was 1.09 kg (Results: 97.5% CI 0.43 to 1.75 kg; Bonferroni-adjusted P less than 0.001). The abstract again labels that same interval as a 95% CI. A University of Western Australia release dated Nov. 10 said people in the training arm lost 25% less lean tissue, including muscle, than people who did not train. That 25% figure is the university’s description, not a percentage printed in the paper’s abstract. The lean-mass losses were not eliminated in either arm; they were smaller with training, and the paper says the trajectory appeared to plateau after week 24.

For appendicular lean soft tissue, the published abstract reports mean changes of −3.55 kg in the tirzepatide-only arm and −2.92 kg in the training arm (P less than 0.001). The Results text also gives a between-group difference of 0.63 kg (95% CI 0.30 to 0.96). In a later sentence describing week 40, the paper states that appendicular lean soft tissue was 3.52 kg lower than baseline in the tirzepatide-only arm and 2.94 kg lower in the training arm. Those two pairs are close but not identical; we report both as they appear.

Strength moved in opposite directions

Mean strength fell in the usual-activity arm and rose in the training arm, all P less than 0.001 versus tirzepatide alone, as mean change from baseline:

• Bench press: −4.09 kg (−8.4%) versus +4.68 kg (+10.7%)
• Lat pulldown: −4.42 kg (−6.5%) versus +8.58 kg (+15.7%)
• Knee extension: −4.42 kg (−9.0%) versus +5.88 kg (+17.5%)

The paper also reports net between-arm differences of 8.59 kg, 12.97 kg, and 9.78 kg (18.8%, 23.4%, and 21.8%). Blood pressure, fasting glucose, insulin, HbA1c, total cholesterol, and triglycerides improved similarly in both groups; resistance training did not add a detected benefit on those measures.

What the authors and Green said

In the university release, Green said resistance exercise “should be part of the prescription” when people were using a weight-loss drug such as tirzepatide. The paper’s own closing line is that resistance training “should be considered when prescribing incretin mimetics as anti-obesity therapy.”

Speaking to RACGP newsGP, Green said: “Even though we saw an attenuation of the muscle loss, we still saw some muscle loss, but the strength increased in the resistance training group, and it decreased quite a lot in the drug-only group.” He added: “So, as much as the muscle mass benefits were good to see, the strength benefits are more important for outcomes like frailty, which we were concerned for.” Those comments are attributed to newsGP. We are not using two errors in that article: it described tirzepatide as “sold as Wegovy,” which is a semaglutide brand, and it rounded the training arm’s mean weight change to 22.8 kg. The paper’s figure is 22.28 kg.

An earlier oral abstract from the same group was listed for Nov. 3, 2026, at the ESA-SRB-NZSE meeting. Some of that abstract’s mean changes differ slightly from the Nature Medicine paper; the numbers above are from the journal article.

What is not known

The study does not show what happens after week 40, or in people with type 2 diabetes or established cardiovascular disease, who were excluded. It also does not compare tirzepatide with a placebo, and it cannot say whether the smaller lean-mass loss will change long-term outcomes such as falls or frailty. The authors note that whether this supervised, center-based program can be delivered in routine practice still needs to be established.

Limits

The authors list several constraints. The trial was open-label for the exercise intervention, done at one center, and lasted 40 weeks. It enrolled middle-aged and older adults without diabetes or cardiovascular disease, who were not taking cardiovascular medicines and who were initially sedentary, so the findings may not apply to everyone treated with tirzepatide. There was no placebo arm without tirzepatide, and no specific diet program beyond instructions to maintain protein, fluid, and fiber intake. Strength testers could not be fully blinded. The authors used DXA rather than MRI to quantify lean soft tissue, and they say further work would be needed to know how the lean-mass differences affect long-term outcomes, and whether the training model can be delivered in routine practice. Eli Lilly supplied tirzepatide through an investigator-initiated program but, the paper says, had no other role in conducting the trial.