Tirzepatide and Longevity: What the Death and Aging Data Show

No study has shown that tirzepatide extends life, and none has tried. What exists is narrower. In one four-year trial that enrolled 13,299 people who had both type 2 diabetes and heart disease, 8.5% of those on tirzepatide died against 10.1% on dulaglutide, an older drug in the same family.^[1] That death result was a secondary, exploratory finding, and the trial’s main goal was only to show tirzepatide was not worse.^[3] In people with obesity but no diabetes, no trial has reported deaths or heart attacks yet; the one designed to do so finishes in October 2027.^[11] Nobody has published a tirzepatide result on biological age, frailty or lifespan.

The verdict, outcome by outcome

SURPASS-CVOT: the one death signal, and how to read it

SURPASS-CVOT took 13,299 adults, all with type 2 diabetes and all with atherosclerotic heart or vascular disease, and assigned them at random to weekly tirzepatide or weekly dulaglutide 1.5 mg. Their mean age was 64.^[3] The comparison was deliberately hard: dulaglutide had already been shown to lower cardiovascular events, so tirzepatide was being tested against an active drug, not a placebo.^[3]

The main result was a draw with a lean. The combined count of cardiovascular deaths, heart attacks and strokes reached 12.2% in the tirzepatide arm and 13.1% in the dulaglutide arm, a hazard ratio of 0.92. That met the bar for noninferiority. It missed the bar for superiority.^[3] On that basis, in 2026 FDA gave Mounjaro a second use: cutting cardiovascular events in high-risk people who have type 2 diabetes. The label states that superiority to dulaglutide was not established.^[1]

The death numbers sit further down the same label, in Table 10. Over a median of 210.1 weeks, about four years, death from any cause was recorded in 567 of 6,647 people on Mounjaro (8.5%) and 670 of 6,647 on dulaglutide (10.1%). The hazard ratio was 0.84, with a confidence interval of 0.75 to 0.94.^[1] Vital status was known for 99.7% of participants, so the count is not distorted by people lost to follow-up.^[1]

“Not controlled for family-wise type I error rate”^[1]

— MOUNJARO prescribing information, Table 10 footnote

That footnote is attached to the all-cause death row, and it decides how much weight the row can carry. The trial tested its primary endpoint under strict statistical rules. Death from any cause was not part of that tested sequence, so a favorable result could be partly chance. It is a signal worth taking seriously. It is not a proven survival benefit. A post hoc analysis that folded all-cause death into a six-part heart and kidney composite also favored tirzepatide (23.7% vs 27.4%), but post hoc is the weaker grade of evidence again.^[4] The kidney side of that analysis is covered in tirzepatide vs dulaglutide on the kidneys.

Two limits travel with the number. Everyone in the trial had diabetes and established vascular disease, a group with a high baseline death rate. And the comparison is against another incretin drug, so the result says tirzepatide may beat dulaglutide on survival, not how either compares with taking nothing.

SUMMIT: fewer heart failure events, no survival answer

SUMMIT enrolled 731 people whose heart failure was the preserved-ejection-fraction type (HFpEF) and whose BMI was 30 or more, and followed them for a median of 104 weeks. Cardiovascular death or a worsening heart failure event occurred in 9.9% on tirzepatide and 15.3% on placebo, a hazard ratio of 0.62. Symptoms and physical limitation scores improved more on the drug.^[5]

The benefit came from heart failure events: 8.0% against 14.2%. Cardiovascular deaths went the other way, 8 on tirzepatide and 5 on placebo, a difference far too small to interpret either way.^[5] The trial’s posted results list 19 deaths from any cause on tirzepatide and 15 on placebo.^[6] A trial of this size cannot detect a mortality effect, and it should not be quoted as showing one. The Zepbound label carries no heart failure indication.^[2]

SURMOUNT-OSA: a disease that shortens life, a surrogate that improved

Two 52-week trials enrolled people with obesity whose sleep apnea was moderate or severe. In both, tirzepatide cut breathing interruptions by about 25 to 29 events an hour, against about 5 on placebo. Oxygen dips, blood pressure and C-reactive protein also fell.^[7] Sleep apnea is tied to serious cardiovascular complications, and Zepbound is now approved to treat it.^[7][2] The trials measured breathing, not heart attacks or deaths. Whether treating apnea this way lengthens life is an inference, not a result. More on the indication is in Zepbound for sleep apnea.

Obesity without diabetes: the trial that has not reported

Most people asking about tirzepatide and longevity do not have diabetes. For them, the hard-outcome evidence is empty. SURMOUNT-MMO randomized 15,374 adults. Entry required a BMI of at least 27 plus established cardiovascular disease or risk factors. Anyone with diabetes is excluded. Its primary endpoint is the first of all-cause death, heart attack, stroke, coronary revascularization or a heart failure event, over up to five years. It is active, no longer recruiting, has no posted results, and lists an estimated primary completion of October 2027.^[11]

Until then, the best data are surrogates. In SURMOUNT-1, people without diabetes on 15 mg ended 72 weeks 20.9% lighter, against 3.1% on placebo.^[8] In the 1,032 who also had prediabetes, treatment continued for 176 weeks, weight loss held near 20%, and new type 2 diabetes was diagnosed in 1.3% against 13.3% on placebo.^[9] A post hoc analysis of those three-year data used the Framingham risk equation to estimate that the top dose lowered predicted 10-year cardiovascular risk. Its authors say that estimate still has to be confirmed by event data from SURMOUNT-MMO.^[12] A risk score is a forecast. It is not a death count.

Lean mass: about a quarter of the weight lost

Muscle enters the longevity question because strength and mobility decide how independent someone stays in later life. The tirzepatide measurement comes from a DXA scan substudy of SURMOUNT-1: 160 of the 2,539 participants, scanned at the start and at week 72.^[10]

Lean tissue made up roughly a quarter of what came off, and the split was the same in the placebo group.^[10] The Zepbound label puts it as greater fat loss than lean loss.^[2] That proportion is typical of weight loss in general, which lean mass on a drug versus dieting sets against diet trials. What the scan does not show is strength or function, and the substudy was small and ran for 72 weeks.

Bone, falls and frailty: thin, conflicting, mostly observational

No randomized trial has used fractures or bone density as a tirzepatide endpoint. What exists points in different directions:

• A pooled look at trial participants 65 and older. Across the SURMOUNT, SURMOUNT-OSA and SUMMIT trials, a post hoc analysis published under Eli Lilly’s copyright found no clinically meaningful difference between tirzepatide and placebo in falls or fractures among older adults.^[13] These trials were not built to count fractures.
• A records study comparing tirzepatide with other GLP-1 drugs. In 66,329 matched pairs followed about 14 months, tirzepatide starters had a higher rate of new osteoporosis or fragility fracture, a hazard ratio of 1.44.^[15]
• A records study in adults 65 and older with diabetes. Against DPP-4 inhibitors, tirzepatide users had fewer femoral fractures (0.2% vs 0.4%) and fewer falls (3.6% vs 5.7%) over one year.^[16]
• A bone scan study of 255 semaglutide or tirzepatide users. Among those without diabetes, hip bone density fell slightly faster than in matched non-users, 1% against 0.6% a year, and more weight loss went with more bone loss.^[14]

Records studies can show associations, not causes, and these ones disagree. The honest reading is that the bone question is open. Frailty itself, meaning the combination of weakness, slowness and low reserve that predicts decline, has never been a tirzepatide trial outcome. The weight-loss trials also enrolled few older people: 9% of tirzepatide participants in the two pooled weight studies were 65 or older, and 0.5% were 75 or older.^[2] The broader picture for this age group is in GLP-1 drugs after 65 and GLP-1s, bone density and fracture risk.

Biological age and epigenetic clocks: no tirzepatide data yet

A PubMed search in October 2026 found no published human study of tirzepatide on an epigenetic clock, DNA methylation age or any other aging-clock measure. The only clock data in this drug class come from semaglutide, in a 32-week trial of people with HIV, and they carry heavy caveats; they are taken apart in does semaglutide slow aging. A 2026 review of obesity drugs in longevity medicine concluded that the early epigenetic and protein-marker findings across the class “do not establish slowed ageing.”^[20]

Trials are now registered. None has reported.

Two features of these trials limit what they can say. Every one of them enrolls people with overweight or obesity, so a positive result would not tell anyone whether the drug slows aging in a lean person. And an aging clock is a surrogate. Clocks shift with inflammation and metabolic state, both of which change with weight loss, so a moved clock may be reporting fat loss rather than slower aging. A clock result would still need hard outcomes behind it, the same problem seen with inflammation markers in tirzepatide and inflammation.

What has not been measured

• Lifespan. No human study, and no published animal lifespan study, has tested tirzepatide.
• Death in people without diabetes. SURMOUNT-MMO is the first trial built to count it and has not reported.^[11]
• All-cause death as a tested primary endpoint. The only favorable death result is exploratory.^[1]
• Aging in people without obesity. No trial, completed or registered, enrolls lean adults to test tirzepatide for aging.^[17][18]
• Anything past about four years. SURPASS-CVOT followed people for a median of 210.1 weeks and the longest obesity data run 176 weeks; SURMOUNT-MMO runs up to five years.^[1][9][11]
• Frailty, strength and fractures as trial endpoints. Only scans, post hoc pooling and records studies exist.^[10][13]

How far the tirzepatide safety record reaches is set out in tirzepatide long-term side effects. How tirzepatide fits with semaglutide, metformin and the rest of the longevity claims is on our hub, GLP-1s as longevity drugs, with the head-to-head in metformin vs a GLP-1 for longevity.

What this means for someone considering it

For a person with type 2 diabetes and heart disease, tirzepatide has a cardiovascular indication and an exploratory signal of fewer deaths than dulaglutide over four years.^[1] For a person with obesity, it reliably removes weight and improves risk markers, and the trial that will say whether that prevents deaths is due in 2027.^[8][11] For a healthy person without obesity taking it to slow aging, there is no evidence at all, plus the known side effects and lean-mass loss of a drug tested in a different population. Anyone selling it as an anti-aging drug is ahead of the data.