GLP-1 Agonists (Semaglutide, Tirzepatide)
Evidence BWhat the evidence actually shows
The GLP-1 receptor agonists — semaglutide, and the dual agonist tirzepatide — are the first weight-loss drugs that actually work at scale without the safety record reading like a horror novel. The landmark cardiovascular outcomes trial, SELECT, randomised over 17,000 people with established cardiovascular disease and obesity (but not diabetes) to semaglutide or placebo, and found roughly a 20 percent relative reduction in the primary composite endpoint of cardiovascular death, non-fatal myocardial infarction, and non-fatal stroke.
That is a real, well-conducted, hard-endpoint result, which is exactly why it deserves to be taken seriously and exactly why it deserves careful reading. A 20 percent relative reduction on a composite is not the same as a 20 percent reduction in your chance of dying, and the all-cause mortality curve, while pointing the right way, did not reach the kind of decisive separation that would let anyone quote a life-expectancy figure with a straight face.
How big the effect really is, in years
Here is the uncomfortable answer a marketer would never give: nobody knows, because no trial has run long enough with all-cause mortality as its primary endpoint. What we have is a composite cardiovascular benefit in a high-risk population, over a few years of follow-up. Extrapolating that to a life-expectancy gain in years requires assumptions the data has not earned yet.
The relative-versus-absolute distinction is everything. A 20 percent relative reduction in a group whose baseline event rate is already high translates to a meaningful absolute benefit; the same relative reduction in a low-risk 35-year-old translates to almost nothing. Any single "GLP-1 adds X years" number ignores that the benefit is a function of your baseline risk, which is precisely what an actuarial model is supposed to compute rather than assume.
The catch: it works largely through channels we already count
The mechanistic point is the one that matters for this calculator. Most of what SELECT and the diabetes trials show is consistent with the drug lowering weight, blood pressure, glucose, and inflammation — the very risk factors that already drive mortality through the pathways our BMI and metabolic inputs represent. If a drug helps you mainly by moving the risk factors we already measure, then its benefit is largely captured the moment your inputs change.
There may be a residual, weight-independent cardiovascular effect on top of that — the SELECT benefit appeared somewhat earlier and larger than weight loss alone easily explains, which is scientifically interesting. But "somewhat more than weight loss explains" is not a quantity we can responsibly convert into calculator years today. And the standard caveats apply: real gastrointestinal side effects, unknown effects of decades-long use, and muscle loss that matters for exactly the older population most interested in longevity.
Why we do NOT give it a year-effect
This intervention has no relatedModifierId, and that is a deliberate design choice, not an oversight. If we handed GLP-1 use its own positive year-effect and also credited the weight loss it produces through the BMI modifier, we would be counting the same benefit twice — inflating your actuarial age on the strength of a single mechanism wearing two hats. Double-counting is the most common way these calculators quietly lie.
So the honest treatment is: take the drug's effect where it shows up, in your weight and metabolic inputs, and let the BMI modifier price it. The drug is directory content here — informative, evidence-graded, worth understanding — but it does not get a separate lever, because it does not do separate work. When a long-term all-cause mortality trial reports out, we will revisit whether there is a residual effect worth its own weight. Until then, restraint is the correct answer. Whether the drug is right for you is a clinician conversation.
Sources
Reviewed 2026-07-06 by Dmytro Dubina, Actuary · MSc Probability & Statistics · 20+ years in insurance. Population statistics, not medical advice.