Frailty Should Not Preclude Semaglutide? My Read of the SELECT Frailty Analysis
Andres Zuleta, MD · Physician read
If you want the patient and family version, I wrote that for ThriveMed here: Frail and on a weight-loss drug? What a 17,604-person heart study found.
Prefer it shorter? Here is the 49-second version on YouTube Shorts.
Many of us have hesitated to start GLP-1 drugs in frail patients, out of concern that weight loss would erode what little reserve they have. On September 16, 2026, JAMA Cardiology published a post hoc analysis of SELECT by frailty status. Here is what it shows, what it can't, and how I would use it.
1. The trial
SELECT randomized adults 45 or older with established cardiovascular disease and a BMI of 27 or higher, without diabetes, to semaglutide 2.4 mg weekly or placebo. In the parent trial, the primary MACE outcome occurred in 6.5% vs 8.0% (HR 0.80).
In this analysis, Ostrominski and colleagues applied a 31-item Rockwood cumulative deficit frailty index to all 17,604 participants (mean age 61.6; 72.3% male): 31% not frail (FI up to 0.210), 47% more frail (0.211 to 0.310) and 22% most frail (0.311 or higher).
2. The results
MACE: incidence rose with frailty. The semaglutide effect "appeared consistent": HR 0.84 (95% CI 0.65 to 1.07) not frail, 0.70 (0.59 to 0.82) more frail, 0.92 (0.76 to 1.10) most frail; P = .09 for interaction (continuous FI, P = .30).
Secondary outcomes: the heart failure composite, all-cause hospitalization and all-cause death were also reduced regardless of frailty (P for interaction .82, .71 and .28).
Frailty trajectory: from baseline to week 104, frailty category was more likely to improve (OR 2.46, 1.80 to 3.37) and less likely to worsen (OR 0.47, 0.34 to 0.65).
Quality of life: the EQ-5D benefit appeared larger with higher frailty (P = .02 for interaction).
Tolerability: hazard ratios for adverse events leading to permanent discontinuation appeared lower with higher frailty (P < .001 for interaction).
3. The drawbacks
Post hoc. The authors label the findings hypothesis-generating. The first author told Healio the findings "should be cautiously interpreted."
Subgroup precision. Only the middle group's MACE confidence interval excludes 1. The consistency claim rests on the lack of detectable heterogeneity, not on three independently significant effects.
Construct. A deficit-accumulation index is not the physical frailty phenotype. It says nothing directly about grip, gait speed or appendicular lean mass. Bruemmer (Cleveland Clinic, not an author) noted to Medscape that the study was not designed to assess whether GLP-1 related muscle loss increases frailty or fall risk.
Population. No diabetes, mostly men, mean age about 62: generalizability to the oldest and frailest is limited.
Industry ties. SELECT was funded by Novo Nordisk; four authors of this analysis are company employees.
Tolerability context. In the parent trial, discontinuation for adverse events was 16.6% vs 8.2%.
4. The possibilities
The Importance statement is candid: whether frailty changes the benefit-risk balance of GLP-1 receptor agonists "is uncertain." The authors' Key Points conclude that "frailty status should not preclude consideration of semaglutide in eligible people with cardiovascular disease and overweight or obesity."
This matters for the future because it could mean frail patients stop being excluded by default from a therapy with proven cardiovascular benefit, while we build trials that include older, frailer adults and measure muscle and function directly.
How I would pair it with strength and protein
My opinion, not a finding of the trial:
Measure function at baseline. Grip, chair rise or gait speed take minutes and give you a trend line. ICFSR guidelines recommend screening older adults with a validated instrument.
Prescribe resistance training. It is the first-line frailty intervention in ICFSR guidelines (strong recommendation). In Cochrane data (121 trials), progressive resistance training had a large effect on strength (SMD 0.84) and on chair rise (SMD -0.94 favoring training).
Protect protein intake. PROT-AGE suggests at least 1.0 to 1.2 g/kg/day in adults over 65, more with acute or chronic disease, with the exception of severe kidney disease (eGFR under 30, not on dialysis).
Combine. In Lundgren et al. (liraglutide, not semaglutide), exercise plus the GLP-1 drug reduced body-fat percentage about twice as much as either alone.
Be proactive. More on proactive care at thrivemed.ai.
References
Ostrominski JW, Plutzky J, Scirica BM, Hovingh GK, Jeppesen OK, Lincoff AM, Lingvay I, Hoffmann Morville T, Quiroga M, Sofer Y, Urina-Triana M, Aroda VR; SELECT Trial Investigators. Efficacy and Safety of Semaglutide According to Frailty Status: A Post Hoc Analysis of the SELECT Randomized Clinical Trial. JAMA Cardiol, published online September 16, 2026:e263566. doi.org/10.1001/jamacardio.2026.3566 (free abstract: PubMed 42747817)
Lincoff AM, et al. N Engl J Med 2023;389:2221-2232. doi.org/10.1056/NEJMoa2307563
Dent E, et al. ICFSR International Clinical Practice Guidelines. J Nutr Health Aging 2019. PMID 31641726.
Liu CJ, Latham NK. Cochrane Database Syst Rev 2009. PMID 19588334.
Bauer J, et al. (PROT-AGE). J Am Med Dir Assoc 2013. PMID 23867520.
Lundgren JR, et al. N Engl J Med 2021. PMID 33951361.
Healio, September 17, 2026 (Ostrominski interview); Medscape, September 22, 2026 (Bruemmer comment).
Educational only, not medical advice.