Sex-divergent trends in cardiorespiratory endurance after the COVID-19 pandemic: a 21-year interrupted time series analysis of Chinese university students.

INTRODUCTION

Studies of COVID-19 and physical fitness in Chinese university students rarely extend beyond 2022, and none has formally tested whether post-pandemic trends differ by sex.

METHODS

Using 21 years of routinely collected fitness data (2005-2025), we examined indicator-specific change and sex differences in population-level trends in a repeated cross-sectional design. The cardiorespiratory analytic sample comprised 297,565 valid records from 91,686 full-time undergraduates at one comprehensive university in Zhejiang Province, China. Six indicators were classified by data availability and suitability for interrupted time series (ITS) analysis; cardiorespiratory endurance (1000-m run for men, 800-m for women) was pre-specified as the primary outcome and body weight was analysed within the same framework as a negative control. The unit of analysis was the annual grade-standardised mean, giving 18 time points per sex. Single-breakpoint models estimated the immediate level change at 2020 (β 2) and the change in annual trend thereafter (β 3); a sex × trend interaction tested whether post-2020 trends differed between women and men; Benjamini-Hochberg correction was applied across five confirmatory tests.

RESULTS

Run times in 2020 rose above counterfactual projections in both sexes (women +9.60 s, 95% CI 2.58-16.62, p = 0.011, q = 0.019; men +13.02 s, 95% CI 4.76-21.28, p = 0.005, q = 0.019), although this immediate change was attenuated and no longer significant when fourth-year students, tested in May 2020 rather than December, were excluded. Post-2020 trends differed by sex (interaction p = 0.030, q = 0.037) and this difference was robust to that exclusion. In men the post-2020 trend did not differ significantly from the pre-pandemic trend (β 3 = -0.47 s/year, p = 0.609), with the 2025 annual mean remaining about 14.9 s above the counterfactual projection. In women the trend was significantly steeper (β 3 = +2.25 s/year, p = 0.011, q = 0.019), widening the gap to about 26.0 s by 2025. Body weight rose steadily across the whole period but showed no change attributable to the pandemic (level change p = 0.31 and 0.70; trend change p = 0.57 and 0.92 in women and men), which is consistent with the cardiorespiratory result reflecting something other than an artefact of the model.

DISCUSSION: Cardiorespiratory endurance was the only indicator analysed by ITS that showed a sustained pandemic-associated change, and its subsequent trend diverged by sex. These are aggregate population trends rather than individual trajectories, but they suggest that post-pandemic aerobic programming may need to be sex-specific, and that continuity of aerobic activity deserves explicit attention in emergency preparedness planning.