The 'biological age' claims, explained

In September 2026 a paper reported that people taking rentosertib in a small lung-disease trial had a lower predicted “biological age” on a set of blood tests. Headlines around the world said the medicine had made people years younger. The paper itself is considerably more cautious, and its authors say so.

What was actually published

On 7 September 2026, Nature Biotechnology published an analysis of blood samples already collected during the GENESIS-IPF phase 2a trial. This was not a new trial. It was a re-analysis of stored samples from 42 of the 71 people who took part.

The researchers measured 2,841 proteins in those samples. They then ran the results through six proteomic ageing clocks: ProtAge, two OrganAge clocks, PAC, ipfP3GPT and PAOPAC. Predictions were compared against a reference set of 55,319 blood-protein profiles from UK Biobank.

All six clocks gave a lower predicted biological age for people taking rentosertib than for those on placebo. The largest effect was about 3 to 4 years, rising to about 6 years on one clock. It appeared at week 4, in the group taking 30 mg twice daily. The authors propose a “senomorphic” mechanism, meaning the medicine may quieten the behaviour of ageing cells rather than remove them.

What a proteomic ageing clock is

Blood contains thousands of proteins. Their levels shift with age, in patterns that are broadly similar across large populations.

A proteomic ageing clock is a statistical model trained on those patterns. Feed it a person’s protein measurements and it returns a number in years: the age of a typical person with that protein profile. The gap between that number and the person’s actual age is what gets called “biological age”.

A clock is a prediction, not a measurement of ageing itself. Illness moves protein levels too. Treating an illness moves them back. A clock cannot tell the difference on its own.

Why the headlines went further than the paper

Several outlets reported that the medicine had made people biologically younger by up to six years. Six caveats sit between the data and that claim, and most are in the paper.

The paper says the clocks cannot separate the two explanations. It concedes that the clocks “cannot deconvolute aging- and disease-specific effects”. In plain terms: a person whose lungs are less inflamed will look younger to a clock, whether or not anything about their ageing has changed.

Most comparisons were not statistically significant. Of 54 comparisons made, 21 reached the significance threshold the authors set. The effect also plateaued by week 12, rather than continuing to grow.

Lung function explained only a small part of it. Change in FVC, or forced vital capacity, a measure of how much air you can breathe out, accounted for a median of about 6% of the variation in predicted age. The authors present this as evidence the effect is not simply a lung effect. It equally means most of the signal is unexplained.

The biggest single contributor was a fibrosis protein. LTBP2, which is associated with scarring in the lung, contributed more than any other protein to the change in predicted age. That is what you would expect from a medicine acting on fibrosis.

The clocks were mostly marked by their own makers. Four of the five teams behind the clocks used are co-authors on the paper. Longevity.Technology reported on 8 September 2026 that only the PAC clock was applied by a team independent of the work.

The first author runs the company. Alex Zhavoronkov is the first author of the paper and the chief executive of Insilico Medicine, which owns rentosertib and funded the trial.

What the people involved said

Michael Levitt, a Nobel laureate in chemistry who has sat on Insilico’s scientific advisory board since 2014, said the work “cannot yet separate slower aging from a treated lung”.

Zhavoronkov himself described the analysis as a pilot and as early and exploratory work, and said it “will take us a decade to prove that it’s anti-aging”. Elsewhere the company’s promotional material has been less restrained, including a claim about billions of life years that has no basis in this paper.

How the story was covered

The New York Times ran it on 7 September under a hedged headline about early data, attributing the ageing claim to the company throughout. The Wall Street Journal covered it the same day in sceptical terms, framing it as the co-chief executive’s hopes rather than a finding. Bloomberg’s report, carried by the Japan Times on 8 September, noted the work had been presented at a Nature conference at the Sorbonne.

USA Today covered it on 9 September, quoting Zhavoronkov saying there is no drug known to extend the life of already healthy people.

In the UK, one consumer outlet covered the story: International Business Times UK, on 9 September. Despite a headline about reversing biological age, the article itself reported the 21-of-54 figure and named LTBP2 as the largest contributor. No UK national newspaper or broadcaster covered it.

What this analysis does and does not show

It shows that in 42 people with a scarring lung disease, taking an experimental medicine for up to 12 weeks, six statistical models of blood proteins produced younger-looking predictions than placebo did.

It is an exploratory analysis of a small, short, single-country trial that was designed to measure side effects. It is not evidence that rentosertib slows human ageing. No medicine is licensed anywhere in the world for slowing or reversing ageing, and ageing is not an approved indication for any regulator, including the MHRA.

Further detail on the trial itself, including the liver findings and the 22.5% early-withdrawal rate, is on the phase 2a trial page and the safety page.