Patients bring us two biological-age results from two companies and ask which is right. A study from the University of Southern California, published in npj Aging and described by the university last week, offers the best answer so far: they are measuring different things. The researchers took blood from 3,227 participants in the national Health and Retirement Study, ran five widely used epigenetic clocks — Horvath, Hannum, PhenoAge, GrimAge and DunedinPACE — and then looked at which genes were active in people each clock called "old." The clocks shared some themes, in immunity and metabolism. But they mapped mostly to different biology: energy balance for one, cellular growth for another, immune activation and inflammatory signaling for others. In the authors' words, there were more unique than common processes underlying the clocks.
What that means for the kit on your counter. A consumer biological-age test picks one clock, or a proprietary blend, and reports a single number. Today's study says that number reflects one slice of aging biology, chosen by the vendor, and that a different vendor's slice can disagree without either being wrong. The work is publicly funded, the authors declare no competing interests, and it adds a new tool of its own — gene-activity scores that, in several cases, predicted frailty, walking speed and death better than the clocks they were built from. It does not report how far apart two clocks typically land on one person. It explains why they land apart at all.
Loree’s takeWhen two patients' kits say 47 and 51 for the same person, the usual response is to find a third kit. The right response is this paper: the clocks are instruments pointed at different organs of the aging process, and none of them is pointed at the patient's actual risk the way a blood pressure, a waist and a walking speed are. I use those. They cost nothing, they are validated, and no two of them disagree about what to do next.
Ben’s takeThe biological-age business depends on a single number that moves, because a number that moves is a reason to retest. A paper saying the number is one of several, each tracking different biology, is bad for the business and good for the customer. It is also a reminder of who funds what: this work came from the National Institute on Aging. The clocks in the kits came from companies that sell the kits.
Why it matters to a recovering optimizer. Biological age is a research idea that became a product before the research settled. Today's study is the research continuing: five clocks, five partial views, more differences than overlaps. If you own a result, treat it as a photograph of one wall of a house. If you are about to buy one, know that the house has more walls, and the fundamentals improve all of them.
Treat a biological-age number as one instrument's view of one process. Then measure what predicts risk and costs nothing: blood pressure, waist, walking speed.