Beyond the first prescription
What a study of older adults can—and cannot—tell us about sustained GLP-1 use.
Starting a medication and continuing it are not the same experience. A large health-record study in Israel followed this distinction among older adults with obesity, asking whether people with more prescription dispensations had different metabolic outcomes from those with only one or two.
The findings are encouraging, but their meaning depends on the comparison. This was not a randomized trial of a GLP-1 medicine against placebo. Everyone included had started treatment.
What the researchers compared
Zaritsky and colleagues examined Clalit Health Services records for 13,387 adults aged 65 or older who had obesity and did not have diabetes at baseline. All had paired body-mass-index measurements; 69% were women. Minimal exposure meant one or two prescription dispensations. Sustained exposure meant three or more.
Three fills do not establish years of uninterrupted treatment, a particular dose, or whether every dispensed dose was taken. That definition is essential to reading the headline.
The observed differences
The primary outcome was movement to a lower BMI category. This occurred in 55% of the sustained-exposure group and 30% of the minimal-exposure group: a 25-percentage-point observed difference. The median relative BMI changes were −8% and −2%, respectively.
Among 7,392 people with paired HbA1c measurements, blood-sugar category improvement occurred in 34% versus 20%. Diagnosed diabetes was recorded in 2.5% of the sustained-exposure group and 5.3% of the minimal-exposure group, an observed difference of 2.8 percentage points. The abstract does not supply a follow-up duration, so that difference must not become an annual prevention estimate or a number needed to treat.
Why association is not a treatment effect
The statistical analysis adjusted for age, sex, socioeconomic status and other medical conditions. Even so, people who continue a medicine may differ from those who stop in ways that influence the outcome: access, tolerability, illness, health behavior and clinical follow-up.
The authors explicitly describe changes observed after treatment initiation, not a proven causal effect. Nor does this study establish lifespan extension or resolve the balance of benefits and harms for an individual.
The useful question
For a healthspan platform, this is a reason to study the whole treatment pathway rather than celebrate the prescription alone. Which patients can continue treatment? What changes in function as well as weight? What happens after discontinuation? Those are questions this study helps sharpen, not questions it fully answers.
The result deserves attention precisely because its limits remain visible. Real-world observations and randomized trials can inform one another without becoming interchangeable.
Read the research: Zaritsky et al. — primary PubMed record
Study publication: October 22, 2026. Source edition: October 25, 2026. Review scope: primary abstract; not a full-text appraisal.