Table 3.

Summary of studies that examined rest-activity and cancer- and treatment-related symptoms.

Authors	Fatigue	Sleep	Depressed mood	Cognition	Patient population	Stage of cancer trajectory	Rest-activity markers	Outcome measures	Study design	Found association between rest/activity and CRTS	Results
Ancoli-Israel et al. (196)	X	X	X		Breast cancer (n=85)	Newly diagnosed stage I-III scheduled to receive chemotherapy	72-h actigraphy: F statistic, acrophase	Fatigue: MFSI-SF Depression: CES-D Sleep: PSQI, FOSQ	cross-sectional	no	No significant correlations between rhythm variables and fatigue, sleep, depression, or functional outcome of sleep.
Ancoli-Israel et al. (14)	X	X	X		Breast cancer (n=68) and cancer-free controls (n=60)	T1: Newly diagnosed stage I-III scheduled to receive chemotherapy T2: end of cycle 4 T3: 1 year post-chemotherapy	72-h actigraphy: R-squared	Fatigue - MFSI-SF; Sleep - total sleep time, total nap time; PSQI global sleep quality score; Depression - CES-D	longitudinal	not tested	T1: Patients longer total daytime nap time, worse sleep quality, more fatigue, more depression, and more disrupted rest-activity rhythms than controls; T2: Patients worse sleep, increased fatigue, more depression, more disrupted rest-activity rhythms compared to T1 and to controls. T3: Patients’ fatigue, depression returned to T1 levels, but still worse than controls; nap time and rest-activity rhythms did not differ from controls.
Ancoli-Israel et al. (197)	X	X	X	X	Breast cancer (n=69) and matched controls (n=64)	T1: Newly diagnosed stage I-III scheduled to receive chemotherapy T2: end of cycle 4 T3: 1 year post-chemotherapy	72-h actigraphy: R-squared	Fatigue: MFSI-SF Depression: CES-D Sleep: PSQI, TST, WASO, %sleep, NAPTIME Cognition: Neuropsychological test battery, PAOFI	longitudinal	yes	No significant group-by-time interaction in objective cognition. Significant group-by-time interaction in self-reported cognition: significant decrements from Baseline to Cycle 4 and 1 year among breast cancer patients but no marked changes in controls. Decreases in objective cognitive functioning in cancer patients predicted by less robust circadian activity rhythms, worsening sleep quality and increases in nap time compared to baseline.
Berger et al. (185)	X	X			Colon and rectal cancer (n=14)	Stage II-III during chemotherapy	7-day actigraphy: mesor, amplitude, acrophase, circadian quotient, r24	Fatigue: Piper Fatigue Scale Sleep: PSQI	cross-sectional	not tested	Disturbed sleep, low daytime activity, and impaired rest-activity rhythms during the first week after chemotherapy cycles 1–3. Rhythm measures were 78%–83% (mesor) and 66%–72% (amplitude) of values obtained in healthy young adults. Rhythm consistency from day to day (r24) was 0.33-0.34 during the first week after chemotherapy 1–3.
Berger et al. (198)	X		X		Breast cancer (n=190) randomized to behavioral therapy sleep intervention or healthy eating control group	T1: Newly diagnosed stage I-IIIA 3-4 weeks before chemotherapy T2: During cycle 3 T3: 30 days after last cycle	7-day actigraphy: mesor, amplitude, peak activity, acrophase, circadian quotient, r24	Fatigue – Piper Fatigue Scale, Daily Fatigue Intensity (Item #7 from Piper) Depression - HADS	longitudinal intervention	yes	Rhythm parameters more disrupted compared to healthy adults. More robust rhythms associated with lower fatigue, depressive symptoms, BMI and higher performance status.
Berger et al. (199)	X	X			Breast cancer (n=130)	Stage I-IIIA during 48 hours before chemotherapy	≤48-h actigraphy: mesor, amplitude, acrophase, goodness of fit	Fatigue: Piper Fatigue scale Sleep: PSQI	cross-sectional	yes	Worse sleep quality and more impaired components of sleep quality, sleep latency, and habitual sleep efficiency correlated with lower mesor. Longer subjective sleep latency was the only component associated with lower amplitude.
Cash et al. (200)			X		Head and neck cancer (n=55)	Newly diagnosed before chemoradiation	6-day actigraphy: r24; I<O (nighttime restfulness); acrophase	Depression: PHQ-9 depressive symptoms	cross-sectional	yes	Cognitive and affective depression symptoms associated with rest-activity rhythm disruption. Overall and somatic depression symptoms associated with phase shifts, shifting from morning to evening. Rest/activity rhythm disruption and lower nighttime restfulness, but not acrophase, associated with 2-year overall survival.
Chang and Lin (193)	X	X	X		Lung cancer (n=82)	T1: Newly diagnosed T2: 6 weeks T3: 12 weeks T4: 24 weeks T5: 48 weeks after start of treatment	72-h actigraphy: I<O	Fatigue: BFI Sleep: PSQI Depression: HADS	longitudinal	not tested	Compared with baseline, sleep–wake rhythms improved significantly after starting treatment. Fatigue worsened significantly; depressive symptoms improved.
Chen et al. (192)		X			Lung cancer (n=106)	Post-treatment	72-h actigraphy: r24, I<O	Sleep: Actigraph-assessed TST, SE, SOL	cross-sectional	yes	Significant positive correlations of the TST with I<O, SOL with UP activity mean. Negative correlations of SE with UP activity mean, SOL with I<O, SOL with 24-h light-activity minutes. Lower I<O, lower 24-hour light-activity minutes, and higher value for the UP activity mean had poorer objective sleep quality.
Du-Quiton et al. (201)		X	X		Non-small cell lung cancer inpatients (n=42) and outpatients (n=42)	Advanced stage before or beginning of chemotherapy	3 to 7-day actigraphy during first chemotherapy cycle (inpatients) or prior to chemotherapy (outpatients): mesor, amplitude, acrophase, circadian quotient, rhythm quotient, peak activity, r24, day-night activity balance, night-day sleep balance	Depression: HADS	cross-sectional	yes	Outpatients: - more robust daily activity patterns (mean daily activity, daily amplitude, peak of activity) and longer, more consolidated nighttime sleep (night-day sleep balance) compared with inpatients. - more disrupted daily sleep-activity rhythms associated with worse depression and/or anxiety. - severely depressed outpatients - lower activity levels during daytime and more active during night than those with lower depression scores. - inactivity during the day and daytime napping associated with depression; - higher circadian amplitudes of activity, and higher peak daily activity associated with less depressed mood. - better night-day sleep balance, more nighttime sleep and less daytime sleep were associated with lower depression scores. Inpatients: No associations between circadian rhythms and depression.
Grutsch et al. (191)	X	X		X	Lung cancer (n=84)	Advanced stage in period prior to and during first treatment	4 to 7-day actigraphy: mesor, amplitude, acrophase, circadian rhythm robustness and day-to-day stability, peak activity, r24	Fatigue, Sleep, Cognition: EORTC fatigue, sleep, & cognitive subscales	cross-sectional	yes	Insomnia severity correlates negatively with 24-hour autocorrelation, day-to-day stability. Outpatient fatigue levels associated with diminished robustness of circadian quotient, rhythm quotient, night-day balance of time spent asleep. More robust day-night activity/sleep measurement differences, the less fatigue these patients experience during each day.
Innominato et al. (202)	X	X			Cohort 1: Metastatic colorectal cancer (n=237) Cohort 2: Histologically proven advanced or metastatic cancer requiring medical treatment (n=31)	Metastatic or advanced post-treatment	72-h actigraphy: I<O	Cohort 1 – Fatigue, Sleep: EORTC QLQ-C30 v2 Cohort 2 – Fatigue, Sleep: MDASI	cross-sectional	yes	Cohort #1: Significantly lower I<O associated with greater fatigue and sleep trouble. Greater circadian disruption associated with more severe fatigue and sleep problems. Cohort #2: Significantly lower I<O associated with greater fatigue, but not sleep disturbance. Greater circadian disruption associated with higher fatigue.
Levin et al. (190)	X	X	X		Non-small cell lung cancer (n=33)	Stage IIIA, IIIB-IV before or during chemotherapy	4 to 7-day actigraphy: circadian amplitude; circadian fragmentation/ amplitude of ultradian rhythms; circadian quotient; peak activity; I<O, rhythm quotient	Fatigue: EORTC QLQ-    C30 Sleep: PSQI Depression: EORTC QLQ-C30, QLI Cognition: EORTC QLQ-C30	cross-sectional (part of intervention study)	not tested	No correlation between performance scores and any actigraphy data item. Patients reported poor sleep quality as well as fatigue 20-points below population-based surveys.
Li et al. (151)		X			Breast cancer (n=180)	T1: Stage I-III awaiting chemotherapy T2: During first cycle T3: At last cycle of chemotherapy	Percent rhythm, F-statistic, amplitude, mesor, acrophase	Sleep: PSQI, actigraphically-assessed nighttime sleep duration, sleep efficiency, nighttime total wake time	longitudinal	not tested	Sleep efficiency significantly lower than at baseline, but higher than beginning of chemotherapy. Percent rhythm, F-statistic, and mesor at end of chemotherapy were significantly lower than baseline. Did not examine associations between CTRS and rest-wake activity.
Liu et al. (122)	X				Breast cancer (n=148) and healthy controls (n=61)	T1: Stage I-III before chemotherapy T2: After 4 cycles of chemotherapy	72-h actigraphy: amplitude, acrophase, mesor, up-mesor, down-mesor, F-statistic	Fatigue: MFSI-SF	longitudinal	yes	Increases in fatigue significantly associated with greater disruptions in amplitude, mesor, and F-statistic over time.
Ma et al. (203)		X			Cancer (n=68)	Advanced stage, at least 3 months post-surgery	72-h actigraphy: r24, I<O	Sleep: PSQI; 3 d sleep log; actigraphy-assessed TIB, TST, SE, WASO, SOL, waking episodes	cross-sectional	yes	Actigraphy-assessed sleep parameters (TST, SE, SOL, WASO but not TIB, waking episodes) correlated with rhythm parameters (r24 and I<O). Rest/activity rhythms of patients with poor sleep quality (PSQI>5) were much less regular than those with good sleep quality. r24 significantly predicted sleep quality (negative association).
Miaskowski et al. (204)	X	X			Breast, prostate, lung, or brain cancer (n=185)	Before radiation therapy	24-h actigraphy: mesor, amplitude, acrophase, r24	Fatigue: LFS Sleep: PSQI, GSDS	cross-sectional	yes	Significant correlations between PSQI global score and SOL, TST, sleep period time, mesor, and circadian quotient. Significant correlation between fatigue subscores and acrophase.
Mormont and Waterhouse (205)*	X		X		Colorectal cancer (n=200)	Metastatic, before chronotherapy	3 to 5-day actigraphy: mean activity level, r24, I<O	Fatigue: EORTC QLQ-C30 v2. Depression: HADS	cross-sectional	yes	Fatigue was associated with all activity scores. Depression was associated with dampened rhythm parameters (r24 and I<O)
Ortiz-Tudela et al. (188)	X				Cancer (n=49)	T1: Advanced stage, before chronotherapy T2: During chronotherapy T3: Right after chronotherapy T4: Late after chronotherapy	13-day actigraphy spanning the four time points: I<O, r24, interdaily stability, intradaily variability, relative amplitude	Fatigue: NCI CTC-AE v3.0	longitudinal	yes	Circadian disruption (i.e., I<O ≤ 97.5%) during or after chemotherapy associated with significantly higher risk of clinically significant fatigue.
Palesh et al. (206)		X			Colorectal cancer with sleep problems (n=155) and no sleep problems (n=82)	Metastatic, before treatment	72-h actigraphy: I<O, clock time of lowest activity using cosinor analysis, average activity counts	Sleep: TST, SE, sleep latency, WASO	cross-sectional	yes	Patients with sleep complaints had worse circadian function (i.e., lower I<O) compared to those without sleep problems (96.4% vs 98.1%) with clinical cut off of 97.5%.
Rich et al. (59)	X		X		Colorectal cancer, normal (high r24, n=40) and dampened 24-h rest-activity patterns (low r24, n=40)	Metastatic, before chronotherapy	3-day actigraphy: r24 (high = top quartile, low = bottom quartile), I<O, mean activity	Fatigue: EORTC QLQ-C30 v2. Depression: HADS	cross-sectional	yes	High r24 patients had significantly fewer fatigue symptoms than low r24 patients, but no differences in anxiety and depression.
Roscoe et al. (207)	X		X		Breast cancer (n=78) undergoing chemotherapy or radiation or both (but not overlapping) randomized to either paroxetine or placebo	T1: 72 h after second on-study treatment T2: 72 h after fourth on-study treatment	Actigraphy: r24, mean activity	Fatigue - FSCL, MAF Depression: CES-D, HDI, POMS	longitudinal	yes	r24, mean activity significantly correlated with percent sleep and mostly significantly correlated with fatigue, mood and depression. Change scores in r24 and mean activity over time were in general significantly correlated with changes in fatigue, mood, depression (unrelated to paroxetine).
Sultan et al. (183)	X			X	Breast cancer (n=25)	T1: Newly diagnosed during cycle 1 chemotherapy T2: During cycle 2 chemotherapy T3: During cycle 3 chemotherapy	72-h actigraphy: mesor, amplitude, acrophase; rhythm quotient, circadian quotient, peak activity, I<O, r24	Fatigue, Sleep, Cognition: Hindi QLQ-C30	longitudinal	no	Significant decrease on all functional and symptom scales that include fatigue, insomnia and emotional domains from cycle 1 to cycle 6.

*Note that this study overlaps with Mormont et al. (173) so have omitted one from this table.

CES-D, Center for Epidemiologic Studies Depression Scale; EORTC QLQ-C30 v.2, The European Organization for Research and Treatment of Cancer Core Quality of Life Questionnaire version 2; FOSQ, Functional Outcomes of Sleep Questionnaire; FSCL, the Fatigue Symptom Checklist; GSDS, General Sleep Disturbance Scale; HADS, Hospital Anxiety and Depression Scale; HDI, Hamilton Depression Inventory; I<O, actigraphic dichotomy index; LFS, the Lee Fatigue Scale; MAF, the Multidimensional Assessment of Fatigue; MDASI, MD Anderson Symptom Inventory; MFSI-SF, Multidimensional Fatigue Symptom Inventory – Short Form; NCI CTC-AE v3.0, National Cancer Institute Common Terminology Criteria for Adverse Events version 3.0; PAOFI, Patient Assessment of Own Functioning Inventory; PHQ-9, Patient Health Questionnaire-9; POMS, Profile of Mood States; PSQI, Pittsburgh Sleep Quality Index; QLI, Ferrans and Powers Quality of Life Index; r24, autocorrelation coefficient at 24 hours; SE, sleep efficiency; SOL, sleep onset latency; TIB, time in bed; TST, total sleep time; WASO, wake after sleep onset.