Becoming active in midlife was linked to lower odds of frailty later

On 19 October 2026 The Journal of Frailty & Aging published a 22-year analysis from the ATTICA study, a long-running population cohort in the greater Athens area of Greece. Researchers from the National and Kapodistrian University of Athens and Harokopio University followed 574 adults who joined in 2001–2002, when they were 48 years old on average (range 32 to 73) and free of cardiovascular disease, cancer and conditions that ruled out regular physical activity. Half were men.

Physical activity was measured by questionnaire three times, at the start, after 5 years and after 10 years. Each time a participant counted as active if they reported at least 600 MET-minutes a week, the threshold the authors linked to the minimum in World Health Organization recommendations; for moderate activity it corresponds to about 150 minutes a week. That gave four patterns: inactive throughout (41% of participants), active at first and then inactive (12%), inactive at first and then active (28%), and active throughout (19%).

In 2024, 22 years after enrolment, a geriatrician and a physiotherapist interviewed participants using the FRAIL scale, which scores fatigue, difficulty climbing stairs, difficulty walking a few hundred metres, five or more chronic illnesses and unplanned weight loss. Of the 574, 165 (29%) were classed as pre-frail and 34 (6%) as frail.

Is it too late to start in your 40s or 50s?

In this cohort, people whose activity rose from below to at least 600 MET-minutes a week did markedly better than those who stayed inactive. Compared with people who stayed inactive, those who became active had 79% lower odds of being pre-frail or frail two decades later (odds ratio 0.21, 95% confidence interval 0.12 to 0.36). Those who were active throughout had 80% lower odds (0.20, 0.11 to 0.37). The two groups looked almost the same. None of the participants who stayed active at all three checks was frail at the end. People who were active at first and then stopped did no better than those who were never active (1.28, 0.69 to 2.38).

The analysis was adjusted for age, sex, education, body mass index, high blood pressure, diabetes and high cholesterol at the start. A single activity measurement at enrolment, on its own, was not significantly linked to frailty 22 years later (0.741, 0.496 to 1.107). In this analysis, it was the ten-year activity trajectories, rather than the baseline measurement alone, that were statistically linked to the outcome.

The authors also estimated that about half of the pre-frailty and frailty in the cohort (49.5%, 36.3% to 59.9%) could be attributed to inactive or declining activity, more than the 20.8% attributed to obesity, high blood pressure, diabetes and high cholesterol together. As they note themselves, such a figure only holds if the link is causal, and this study cannot show that.

What does this study not show?

It is an observational cohort, not a trial. Nobody was assigned to exercise, so people who became active may have differed in ways the adjustments did not capture. Activity was self-reported, and so was frailty, although every case was checked in a personal interview. Frailty status was not formally measured at the start, so the study reports pre-frailty and frailty at the end rather than new cases.

Of 2,444 eligible participants, 1,416 had complete 20-year data; after 24 deaths and 818 people who could not be reached or withdrew consent, 574 completed the final assessment. A weighted re-analysis that accounted for dropout gave similar results (0.20 for becoming active, 0.24 for staying active), but healthier volunteers may still be over-represented. Most of the outcome cases were pre-frail rather than frail, because frailty itself was rare in this group. The sample came from urban communities in Attica, the greater Athens area, and the study measured frailty, not lifespan. It was funded by the Hellenic Cardiology Society and the Hellenic Atherosclerosis Society.

Why does frailty matter for a long life?

Frailty is the state in which a minor illness, a fall or an operation knocks a person down for much longer than expected. It is one of the conditions that decide whether the last decades of life are lived independently. In a meta-analysis cited by the authors, eight of ten cohorts measured activity only once, at the start, and earlier long-term trajectory studies enrolled people with a mean age of 58 to 66. This cohort started at 48 on average and repeated the measurement, which is closer to the question a person in midlife actually has: whether changing habits now still counts.

The answer it offers is cautious but consistent with earlier trajectory studies in older groups: people whose activity rose in midlife had almost the same adjusted odds of pre-frailty or frailty as those who had always been active (0.21 versus 0.20). That is an association in one cohort, and it does not prove that starting exercise prevents frailty. If you are planning to become more active and have heart disease, joint problems or another chronic condition, discuss with your doctor how to begin.

This article is for information only and does not replace a consultation with a doctor.

Sources

• Sgouropoulou D, et al. Midlife physical activity trajectories are associated with reduced physical pre-frailty/frailty in older age: insights from the prospective ATTICA sub-study 2002-2024. The Journal of Frailty & Aging, 2026. PMID 42895852 · DOI 10.1016/j.tjfa.2026.100217

• Bull FC, et al. World Health Organization 2020 guidelines on physical activity and sedentary behaviour. British Journal of Sports Medicine, 2020. PMID 33239350