Changes in aging clocks and health were inconsistently aligned in a review of 26 studies
Changes in aging clocks and health were inconsistently aligned in a review of 26 studies
On October 24, Experimental Gerontology published a systematic review of human trials that measured both epigenetic clocks, which use chemical marks on DNA, and health outcomes. The authors compared changes following drugs, exercise, diets, surgery, and other interventions.
Tracking disease development and changes in mortality takes years. Researchers are looking for a measure that can provide an earlier assessment of interventions intended to slow aging. Some algorithms were trained to estimate chronological age, while others were trained to estimate mortality risk or the pace of age-related changes in the body. A single blood sample can therefore yield several different estimates.
The authors distinguish four tests. Does the baseline clock value predict future disease or death? Does it change after an intervention? Do changes in the clock align with changes in health measures? To assess this alignment, researchers compare average changes within a group and examine whether people with larger favorable changes in the clock also experience greater health improvements.
The fourth test asks whether a change in the clock can reliably predict how much a treatment will improve health. This relationship is assessed by comparing treatment effects across several independent trials in which participants are randomly assigned to groups.
The review included 26 studies: 13 with a comparison group and 13 that assessed participants before and after an intervention. In 15, the results were mixed: different pairings of clocks and health measures gave different answers. In some pairings, both measures improved; in others, a change could be established with confidence for only one measure. Some measures changed in opposite directions. Agreement could not be assessed in another six studies, and in four, changes in both measures could not be confidently distinguished from random variation.
The authors found improvement in both measures across all assessed pairings in one study. Among 20 participants assessed before and one year after weight-loss surgery, the difference between age estimated by one algorithm and chronological age decreased, while measures of weight and glucose metabolism improved. This before-and-after comparison had no control group. None of the studies with a comparison group found agreement across all assessed pairings, and none of the 26 studies formally tested whether clocks could predict treatment effects across trials.
One possible reason for the discrepancies is blood cell composition. Different types of immune cells in the blood have different patterns of DNA marks. The authors explain that a drug or surgery can change the proportions of these cells in a sample and thereby change the clock estimate, even if the marks within the cells remain unchanged. Muscle, brain, or kidney function may change through different mechanisms and on different timescales. Assessing the relationship therefore requires accounting for sample composition, the organ being assessed, and the timing of the response.
“A reduction in epigenetic age should not be called ‘age reversal’ or ‘clinical rejuvenation.’”
For the next trial, the authors suggest selecting one clock algorithm and a specific treatment outcome in advance, such as improved memory or physical performance. Both measures should be assessed repeatedly in the same participants, and the changes should be compared with those in a control group. Repeated measurements would help establish the sequence of changes in the clock and bodily functions. Independent trials would then need to test how accurately changes in the selected clock predict benefits under those conditions.