Associations of accelerometer-derived ‘weekend warrior’ physical activity pattern with incident sarcopenia and falls

BACKGROUND: To examine associations between accelerometer-derived active weekend warrior (WW) pattern [most moderate to vigorous-intensity physical activity (MVPA) achieved over 1-2 days] vs. MVPA spread more evenly with risks of incident probable sarcopenia, sarcopenia and falls. METHODS AND FINDINGS: This prospective cohort study comprises three substudies (probable sarcopenia, confirmed sarcopenia and falls). Objective physical activity data were collected from Axivity AX3 wrist-worn triaxial accelerometers on their dominant wrist for 7 consecutive days. Three MVPA patterns were compared: active WW (≥150 min/week and ≥50% of total MVPA over 1-2 days), active regular (≥150 min/week but not meeting active WW), and inactive (<150 min/week). Over 7.8 years median follow-up, there were 1785 (8.4%) incident (probable) sarcopenia, 1855 (8.7%) incident sarcopenia and 4166 (4.9%) incident falls. Compared to inactive participants, the effect sizes for active WW pattern at the guideline-based threshold with ≥50% of total MVPA over 1-2 days were: probable sarcopenia {hazard ratio [HR], 0.79 [95% confidence interval (CI), 0.70-0.89], P = 1.5 × 10-4}, sarcopenia [HR, 0.74 (95% CI, 0.63-0.87), P = 6.6 × 10-4], falls [HR, 0.79 (95% CI, 0.72-0.85), P = 7.4 × 10-8] and recurrent falls [HR, 0.82 (95% CI, 0.75-0.90), P = 9.1 × 10-5]. The active WW pattern was associated with a 24%-32% reduction in the risk of (probable) sarcopenia, while active regular pattern was associated with a 12%-23% reduction in the risk. Regardless of the thresholds defining the active WW group, both activity patterns (active regular and active WW) showed a lower risk of falls. CONCLUSIONS: Weekly physical activity concentrated in 1-2 days had a similarly low risk of sarcopenia and falls as a regular active pattern. Even resistance training 1-2 days per week can serve as an effective public-health strategy for preventing sarcopenia and falls.