Wearable Accelerometers Prove Highly Feasible in AKG Longevity Trial
A 120-person AKG supplementation trial shows accelerometers are practical in aging research, and reveals who is most likely to drop out.
Summary
Wearable activity trackers are rarely used in trials of treatments aimed at slowing biological aging. Researchers in Singapore tested whether they work in the ABLE trial, which gave 120 healthy adults aged 40 to 60 either 1 gram a day of alpha-ketoglutarate or a placebo for six months. Participants wore a wrist accelerometer for one week at each of four visits. Wear was very high: a median of about 24 hours a day, with over 80% logging at least four valid days per visit. Supplement adherence exceeded 90%. Only 5% of participants reported device problems. Overall, 17% dropped out. People who wore the device longer, and men, were less likely to drop out. Wear behavior may therefore flag who needs extra support to stay in a trial.
Detailed Summary
Trials of gerotherapeutics, interventions aimed at slowing biological aging, run for many months and increasingly rely on wearable sensors to capture digital biomarkers of aging. Yet detailed reports on whether these devices are practical, and whether participants stick with them, are scarce, especially for busy middle-aged adults. Missing data and dropout can undermine an otherwise well-designed trial, so operational lessons matter for the whole field.
This exploratory analysis used data from ABLE, a single-site, double-blind, randomized, placebo-controlled trial in Singapore (NCT05706389). It enrolled 120 healthy adults aged 40 to 60 (mean age about 50.6, 50% women) who were biologically older by DNA methylation age. Of these, 59 received 1 g/day of alpha-ketoglutarate (AKG) and 61 received placebo for six months, with a three-month follow-up. At baseline, 3, 6, and 9 months, participants wore an ActiGraph wGT3X-BT on the non-dominant wrist for seven days, day and night, removing it only for water activities. The device was mailed back, and no reminder messages were sent during monitoring.
Adherence was strong. All 120 agreed to wear the device at baseline, and one participant lost it and withdrew. Median daily wear was 1,400 minutes (IQR 1,245 to 1,440), roughly 24 hours. Weekend wear was slightly higher than weekday wear (1,440 vs 1,340 minutes, p<0.001, small effect size r=0.105). More than 80% of participants achieved at least four valid days (16 or more hours of wear) per visit. Supplement adherence by pill count was 100% (IQR 95 to 100) in both arms, and proportion of days covered was 93% for AKG and 95% for placebo. Supplement adherence was unrelated to wear time.
Operational problems were minimal. Six participants raised concerns affecting 8 of 432 participant-visits (about 5% of participants). One refused to wear the device because of its appearance, two lost devices, and three reported comfort issues: skin irritation, a rash, or strap hygiene. Group differences in wear were statistically detectable at some visits, such as valid days at baseline (p=0.026) and end of intervention (p=0.001), but effect sizes were small (r=0.082 to 0.134).
Overall dropout was 17% (20 of 120), with similar completion of about 83% in both arms (log-rank p=0.92). Work or schedule conflicts were the leading reason (n=10), followed by loss to follow-up (n=4), product-related issues (n=2), and protocol non-compliance (n=2). In adjusted Cox models, each extra minute of wear time was linked to a 0.2% lower dropout hazard (HR 0.998, 95% CI 0.997 to 0.999, p<0.001), and men had a markedly lower hazard than women (HR 0.208, 95% CI 0.065 to 0.663, p=0.008). A model excluding age, which violated proportional hazards, gave similar estimates, as did logistic regression (wear time OR 0.998, p=0.029; male OR 0.174, p=0.005).
The results suggest wrist accelerometers are a low-burden, feasible tool for aging-focused trials. Wear behavior could serve as an early-warning signal for dropout risk. Caveats: this is an exploratory analysis in a single-site cohort of healthy, motivated adults, only 20 dropouts underlie the models (hence wide confidence intervals, especially for sex), and wear time is an association with engagement, not proof of cause. The paper does not report any effect of AKG itself on biological age.
Key Findings
- Median daily accelerometer wear was 1,400 minutes (IQR 1,245–1,440), about 24 hours, across all four visits in 120 adults aged 40–60
- More than 80% of participants reached at least four valid days (16 or more hours) per visit, with no reminder messages sent
- Supplement adherence was 100% (IQR 95–100) by pill count; proportion of days covered was 93% (AKG) and 95% (placebo)
- Only 6 participants (5%) raised device concerns, affecting 8 of 432 participant-visits; 1 refused wear because of appearance and 2 lost devices
- Overall dropout was 17% (20/120), with no difference between AKG and placebo (log-rank p=0.92; about 83% completion in both)
- Each additional minute of wear time was associated with a 0.2% lower dropout hazard (HR 0.998, 95% CI 0.997–0.999, p<0.001)
- Men had a much lower dropout hazard than women (HR 0.208, 95% CI 0.065–0.663, p=0.008); logistic sensitivity analysis agreed (OR 0.174, p=0.005)
Methodology
ABLE is a single-site, parallel, double-blind, randomized, placebo-controlled trial at Alexandra Hospital, Singapore, enrolling 120 healthy adults aged 40–60 who were biologically older by DNA methylation age (AKG n=59, placebo n=61). Participants took 1 g/day AKG or placebo for six months, with visits at baseline, 3, 6, and 9 months. An ActiGraph wGT3X-BT was worn on the non-dominant wrist for seven days per visit, with a valid day defined as at least 16 hours of wear and adherence as at least four valid days. Group comparisons used t-tests or Wilcoxon tests, and dropout was analyzed with Kaplan-Meier curves and Cox models using study visit as the time scale, with wear time as a time-varying covariate, cluster-robust errors, and logistic regression as a sensitivity check.
Study Limitations
This was an exploratory analysis from a single site in Singapore in healthy, probably highly motivated middle-aged adults, so findings may not generalize to other populations or to older or less healthy people. Only 20 participants dropped out, giving wide confidence intervals, particularly for the sex effect, and wear time is an association with engagement rather than a proven cause of retention. The full conflict-of-interest statement was not included in the supplied text; the funders reportedly had no role in study design, analysis, or reporting.
Enjoyed this summary?
Get the latest longevity research delivered to your inbox every week.
Enter your email to subscribe: