The association between sleep and physical function in the longitudinal TLC cohort study of women

Background: Sleep disturbances, such as insomnia, are common in postmenopausal women and may negatively impact physical functioning. These disturbances are more frequent in breast cancer survivors than in the general female population. Limited data exist on how sleep affects physical function in older breast cancer patients. We analyzed whether sleep quality influences physical function at a one-year follow-up, and vice versa. Methods: Data were from the Thinking and Living with Cancer study, a multicenter, longitudinal study involving women aged 60 and older, newly diagnosed with breast cancer and frequency-matched non-cancer controls. Participants were followed annually for up to five years. Sleep quality was assessed with the PSQI (range 0-21; higher scores indicate poorer sleep quality); physical functioning was measured by the SF-12 physical component score (PCS) and the Timed Up and Go (TUG). A 3-5 point SF-12 PCS decline was considered clinically meaningful. Random intercept cross-lag panel models (RI-CLPM) analyzed associations using data from 767 women (339 controls and 428 survivors), adjusting for age, race, comorbidities, and recruitment site. Results: Mean age of participants was 69.5 years (70.0 in survivors vs. 69.3 in controls). Disrupted sleep was more common among survivors at baseline (PSQI above 5: 43.9% vs. 29.0%; p less than 0.001) and remained more prevalent during follow-up. Survivors showed a clinically meaningful decline in PCS (-2.8 points in year 1; -3.0 points over two years), whereas controls had a small, gradual, non-clinically meaningful decline (≈ -0.5 points/year). There was no clear trajectory for TUG, and the RI-CLPM revealed no significant associations between PSQI and TUG. In survivors, each 1-point higher PSQI predicted 0.4-0.5 points (P less than 0.001) lower SF-12 PCS at the subsequent annual visit during the first three years, decreasing to 0.3 points by year 4 (P = 0.03). Among controls, effects were small initially (-0.25 points, p = 0.10) but from years 2 to 5 a 1-point PSQI increase predicted 0.4-0.65 points lower SF-12 PCS (p = 0.01 to less than 0.001). For both survivors and controls, a 10-point worsening in a woman's PSQI relative to her usual level would be expected to correspond to about a 4-point lower SF-12 PCS one year later, assuming a 0.4-point decrease in PCS per 1-point increase in PSQI. Conclusion: Sleep disruption is associated with subsequent declines in physical function among survivors and controls, which can reach clinically meaningful magnitudes as sleep quality worsens. Given the high prevalence of sleep problems among breast cancer survivors, routine monitoring for sleep disturbances and timely intervention may help preserve function in the growing population of older cancer survivors, particularly in the early years following treatment. This work was conducted with contributions from the entire TLC-Age research team.