Eli Lilly releases its first preliminary findings showing that tirzepatide slowed epigenetic aging clocks, an effect previously reported for Novo Nordisk’s semaglutide using a different method
Eli Lilly releases its first preliminary findings showing that tirzepatide slowed epigenetic aging clocks, an effect previously reported for Novo Nordisk’s semaglutide using a different method
At the ARDD2026 aging conference, held at Harvard on October 1–3, Eli Lilly released its own findings: in 71 trial participants receiving tirzepatide (Mounjaro, Zepbound), a drug for diabetes and obesity, epigenetic clocks indicated a slower pace of aging over 72 weeks of treatment. These clocks estimate biological age from chemical marks on DNA. At the same conference, Novo Nordisk presented a separate estimate: a model based on data from 19 117 patients in routine clinical care suggested that its drug semaglutide (Ozempic, Wegovy) could increase life expectancy by an average of 1,9 years. Both companies describe their findings as preliminary.
The SURMOUNT-5 trial compares tirzepatide with semaglutide in people with obesity. Within that trial, Lilly collected paired blood samples and used them to calculate fifteen epigenetic clock measures. Two laboratories independently analyzed DNA methylation. Their results were almost identical: all fifteen clocks indicated that biological aging had advanced less than the elapsed calendar time. The difference was statistically significant for two clocks, with median gains of 1,1 and 0,7 years. The DunedinPACE measure of aging rate fell from 1,13 to 1,08. This still indicates more than one year of biological aging per calendar year, but the rate was closer to that benchmark than before treatment. Lilly itself describes the findings as preliminary because the analysis lacks a control group. A separate eight-month trial of tirzepatide using the same clocks will provide that comparison, with results expected in 2027. Protein analysis showed fewer signs of chronic inflammation, a pattern Lilly also sees with calorie restriction and after weight-loss surgery.
Novo Nordisk had previously found a similar slowing of aging clocks using a different method: in a September analysis of five semaglutide trials, proteomic clocks, which estimate age from blood proteins, indicated a reduction of 2–4 years in the heart’s biological age. At ARDD2026, the company presented a different type of calculation. Researchers applied the cardiovascular benefits observed in the SELECT trial of semaglutide to data from patients in routine clinical care in the United Kingdom. They estimated an average gain in life expectancy of 1,9 years, with larger gains for people who started treatment at younger ages. Novo describes the model as preliminary and unpublished.
Steve Horvath, who developed the epigenetic clock method in 2013, called the effect “quite strong” and suggested that GLP-1 drugs could be geroprotectors, medicines that slow biological aging. He added that the benefits for people with obesity are clear, while he has yet to see evidence of benefits for lean people. He is not ready to take the drugs himself because of the risk of muscle loss. Harvard’s Vadim Gladyshev, who took part in the Novo analysis, described the evidence gap differently: he considers this a real anti-aging effect in an unhealthy population, while the effect in healthy people remains unknown.
Speaking from the stage, Alex Zhavoronkov, the organizer of ARDD and founder of Insilico Medicine, asked Novo Nordisk scientific director Alejandro Aguayo-Orozco whether a sensible person should start taking semaglutide. Aguayo-Orozco declined to answer: “I’m not a doctor. I can’t answer that question.” Zhavoronkov had previously disclosed that he injects both tirzepatide and semaglutide despite not having obesity. The discussion concerns the world’s two best-selling drugs: tirzepatide generated more than $36 billion in annual revenue for Lilly.
Resolving this uncertainty will require a controlled trial in healthy people who are not overweight. Horvath summed it up briefly: “I guess we’ll have to wait.”