Relationships Among Physical Activity, Sleep, and Cancer-related Fatigue: Results From the International ColoCare Study

Abstract

Background: Risk factors for cancer-related fatigue are understudied in colorectal cancer.

Purpose: This study aimed to address this critical gap in the literature by (a) describing changes in colorectal cancer-related fatigue and health behavior (physical activity, sleep problems) and (b) examining if physical activity and sleep problems predict fatigue trajectories from baseline (approximately at the time of diagnosis), to 6- and 12 months after enrollment.

Methods: Patients participating in the international ColoCare Study completed self-report measures at baseline (approximately time of diagnosis), 6-, and 12 months assessing physical activity using the International Physical Activity Questionnaire (IPAQ) and fatigue and sleep using the European Organization for Research and Treatment of Cancer Quality of Life Questionnaire (EORTC-QLQ-C30). Mixed-effect models examined changes in physical activity, sleep problems, and fatigue. Cross-lagged panel models examined bidirectional relationships between physical activity or sleep and fatigue across time.

Results: Colorectal cancer patients (n = 649) had a mean age of 61 ± 13 years. Most were male (59%), non-Hispanic White (91%), diagnosed with Stages III-IV (56%) colon cancer (58%), and treated with surgery (98%). Within-person cross-lagged models indicated higher physical activity at Month 6 was associated with higher fatigue at Month 12 (β = 0.26, p = .016). When stratified by cancer stage (I-II vs. III-IV), the relationship between physical activity at Month 6 and fatigue at Month 12 existed only for patients with advanced cancer (Stages III and IV, β = 0.43, p = .035). Cross-lagged associations for sleep and fatigue from baseline to Month 6 were only observed in patients with Stages III or IV cancer, however, there was a clear cross-sectional association between sleep problems and fatigue at baseline and Month 6.

Conclusions: Within-person and cross-lagged association models suggest fatiguability may become increasingly problematic for patients with advanced colorectal cancer the first year after diagnosis. In addition, sleep problems were consistently associated with higher fatigue in the first year, regardless of cancer stage.

Trial registration: The international ColoCare Study was registered on clinicaltrials.gov, NCT02328677, in December 2014.

Keywords: Colorectal cancer; Fatigue; Physical activity; Sleep; Survivors.

Fig. 1.

Three-wave random intercept cross-lagged panel model: (a) cross-sectional paths, (b) autoregressive paths, (c) cross-lagged paths, (d) correlation between stable traits of behavior variable (physical activity level and sleep) and fatigue level at the between-person level. Squares denote observed variables; circles represent “latent” variables.

Fig. 2.

Mixed models to examine changes of outcomes across 1 year. (a) fatigue (n = 649) with quadratic associations (turning point is Month 5.2), (b) physical activity (n = 543) with quadratic associations (turning point is Month 6.6), and (c) sleep problems (n = 649) with linear associations. Refer to Table 3 for model estimates.

Fig. 3.

Three-wave random intercept cross-lagged panel model with standardized effect. Bidirectional association between physical activity (PA) and fatigue controlling for center location, age at diagnosis, sex, race, cancer site, cancer stage, cancer treatment, and BMI among patients with valid fatigue measures in at least one time point (n = 649). Statistically significant lines are solid, whereas non-significant lines are dotted. *p < .05. **p < .01. †p < .10.

Fig. 4.

Multigroup three-wave random intercept cross-lagged panel model with standardized effect. Bidirectional association between physical activity (PA) and fatigue controlling for center location, age at diagnosis, sex, race, cancer site, cancer treatment, and BMI. (a) Group 1: patients diagnosed with mild CRC cancer (Stages I–II; n = 288); and (b) Group 2: patients with advanced cancer (Stage III or IV; n = 361). Statistically significant lines are solid, whereas non-significant lines are dotted. *p < .05. **p < .01. †p < .10.

Fig. 5.

Three-wave random intercept cross-lagged panel model with standardized effect. Bidirectional association between sleep (SP) and fatigue among patients with valid fatigue measures in at least one time point (n = 649) controlling for center location, age at diagnosis, sex, race, cancer site, cancer stage, cancer treatment, and BMI. Statistically significant lines are solid, whereas non-significant lines are dotted. *p < .05. **p < .01. †p < .10.

Fig. 6.

Multigroup three-wave random intercept cross-lagged panel model with standardized effect. Bidirectional association between sleep (SP) and fatigue controlling for center location, age at diagnosis, sex, race, cancer site, cancer treatment, and BMI. (a) Group 1: patients diagnosed with mild CRC cancer (Stages I–II; n = 288); and (b) Group 2: patients with advanced cancer (Stage III or IV; n = 361). Statistically significant lines are solid, whereas non-significant lines are dotted. *p < .05. **p < .01. †p < .10.