Table 2.
Summary of studies that examined cortisol and cancer- and treatment-related symptoms (CTRS).
| Authors | Fatigue | Sleep | Depressed mood | Cognition | Patient population | Stage of cancer trajectory/assessment times | Cortisol markers | Outcome measures | Study design | Found association between cortisol and CRTS | Results |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Abercrombie et al. (159) | X | Breast cancer (n=17) and healthy controls (n=31) | metatastic | 3-day salivary cortisol, 4 times/day: diurnal cortisol slope, mean cortisol levels | Cognition: RAVLT | cross-sectional | no | No associations between cortisol slope and cognitive functioning among patients. | |||
| Alexander et al. (163) | X | X | X | Breast cancer (n=200) | after primary treatment | 24-h urinary cortisol | Fatigue: FACT-F, BFS, FCS, SCID for CRF Depression: HADS |
cross-sectional | yes | No differences in urinary cortisol between fatigued and non-fatigued patients. | |
| Bower et al. (36) | X | Breast cancer, fatigued (n=13) and non-fatigued (n=16) | 1 to 5 years after diagnosis | 2-day salivary cortisol, 4 times/day: diurnal cortisol slope, mean cortisol levels, AUC | Fatigue: RAND SF-36 – energy/fatigue subscale | cross-sectional | yes | Fatigued survivors had significantly flatter cortisol slope than non-fatigued survivors, less rapid decline in in evening hours. Survivors with highest fatigue had flattest cortisol slopes. |
|||
| Castonguay et al. (164) | X | X | Breast cancer (n=145) | T1: ≤20 weeks after primary treatment T2: 3 months later T3: 6 months later T4: 9 months later T5: 12 months later |
3 non-consecutive days (T1) or 2 nonconsecutive days (T2-T5) salivary cortisol: AUC |
Fatigue: BFI Depression: CES-D |
longitudinal | no | Fatigue did not significantly predict intra-individual changes in physical activity or cortisol. Depressive symptoms significantly predicted physical activity but not cortisol levels. |
||
| Chang and Lin (154) | X | X | X | Lung cancer (n=40) and healthy controls (n=40) | Newly diagnosed | Salivary cortisol 3 times/day: diurnal cortisol slope and levels | Fatigue: BFI Sleep: PSQI Depression: HADS |
cross-sectional | yes | Cortisol slope and fatigue were significant predictors of PSQI score. Flatter cortisol slope (and higher fatigue) predicted more severe sleep disturbance. Anxiety and depression were not influenced by cortisol rhythms. |
|
| Cuneo et al. (39) | X | Ovarian cancer (n=30) | At least 5 years post-diagnosis | 3-day salivary cortisol, 3 times/day: diurnal cortisol slope | Fatigue: POMS-SF fatigue subscale | cross-sectional | yes | Flatter diurnal cortisol slopes were associated with significantly higher fatigue after controlling for age and cancer stage. | |||
| Giese-Davis et al. (165) | X | Breast cancer, nondepressed (n=45) and depressed (n=45) | Metastatic T1: assessed 1 week before T2: day of, and T3: day after Trier Social Stress Test |
2-day salivary cortisol, 5 times/day: baseline diurnal cortisol slope, mean waking cortisol, mean wake + 30 rise | Depression: PANAS | longitudinal + manipulation | yes | Compared with nondepressed patients, depressed patients had lower 2-day average baseline waking rise in log cortisol level. No other differences in cortisol between groups at any time. |
|||
| Ho et al. (166) | X | X | Breast cancer (n=181) | Non-metastatic or no recurrence | 2-day salivary cortisol, 4 times/day | Sleep: 10-point scale of sleep quality, time of awakening, total sleep hours | cross-sectional | yes | Controlling for initial cortisol level, flatter diurnal cortisol slope associated with later time of awakening, poorer sleep quality, shorter total sleep hours. Depression and anxiety not correlated with slope. |
||
| Hoyt et al. (167) | X | X | Prostate cancer (n=66) | T1: Localized and treated in prior 2 years T2: 4 months later |
3-day salivary cortisol, 4 times/day: diurnal cortisol slope, AUC and CAR |
Sleep: PSQI Depression: CES-D |
longitudinal | yes | Mediation models examining sleep at T1 on depression at T2 with BMI and age as covariates: Flatter cortisol slope and less overall cortisol output related to higher depressive symptoms. No indirect effect for CAR. Reverse mediation models - i.e., cortisol –> sleep quality –> depression was non-significant or small indirect effects |
||
| Hsiao et al. (168) | X | X | Breast cancer (n=62) | T1: Aged 40 and below who completed treatment T2: 2nd month T3: 5th month T4: 8th month |
1-day salivary cortisol, 6 times/day: diurnal cortisol slope | Sleep: MOS sleep scale Depression: BDI-II |
longitudinal | mixed | Diurnal cortisol slopes were steeper compared with baseline. Significant decreases in depression, but no significant changes in sleep variables across the 8 months. Habitually later time of awakening over the 8 months predicted flatter cortisol slope. Habitual time of going to bed, sleep problem index, and depression not associated with cortisol patterns during 8 month follow up. |
||
| Huang et al. (169) | X | X | Hepatocellular cancer (n=75) and healthy controls (n=33) | Outpatients not under surgical treatment | 3-day salivary cortisol, 5 times/day: mean daily cortisol, mean cortisol levels at each sampling time point, peak cortisol levels, cortisol slope, AUC | Sleep: PSQI | cross-sectional | yes | Patients with higher daily cortisol levels tended to report poorer sleep quality. This was not case with healthy controls. Poor sleepers among patients with least severe liver disease showed slight increase in cortisol level at bedtime. Bedtime cortisol level of poor sleepers with least severe liver disease was higher than that of healthy poor sleepers. (Note: all patients with more severe disease had poor sleep). |
||
| Jehn et al. (170) | X | Cancer patients, with (n=31) and without depression (n=83) | Advanced metastatic cancer | Plasma cortisol at 8am and 8pm: cortisol VAR | Depression: HADS-D | cross-sectional | yes | Relative cortisol VAR decreased in depressed patients compared with non-depressed patients. Cortisol VAR independently negatively associated with depression. |
|||
| Kuhlman et al. (171) | X | Breast cancer (n=135) | T1: Recently diagnosed T2: 6 months after primary treatment |
72-h salivary cortisol, 4 times/day: CAR, diurnal cortisol slope, AUC |
Depression: CES-D | longitudinal | yes | No cortisol indices were associated with depressive symptoms at T1. After controlling for T1 depression, CAR predicted depressive symptoms at 6 months. When history of major depression was included as covariate, association between CAR and increases in depressive symptoms trended in the same direction but were not significant. |
|||
| Lutgendorf et al. (172) | X | Ovarian cancer (n=112) and tumors of low malignant potential (n=25) | Awaiting surgery | 3-d salivary cortisol, 4 times/day: CAR, diurnal cortisol slope, AUC |
Depression: CES-D | cross-sectional | yes | Elevations in total depression and vegetative depression were related to higher evening cortisol. | |||
| Mormont et al. (173) | X | X | X | Colorectal cancer (n=200) | Metastatic before chronotherapy | 2-day serum cortisol at 8am and 4pm: amplitude |
Fatigue: EORTC QLQ-C30 v2. Depression: HADS |
cross-sectional | not tested | Examined associations between cortisol circadian rhythms and rest-activity rhythms only. Cortisol circadian rhythms were positively correlated with r24 but not I<O or mean activity. | |
| Palesh et al. (174) | X | Breast cancer (n=99) | Metastatic | 2-day salivary cortisol, 5 times/day: cortisol slope, AUC |
Sleep: actigraph-assessed time in bed, SOL, SE, nocturnal wake episodes, WASO | cross-sectional | yes | Longer nocturnal wake episodes associated with flatter diurnal cortisol slope. No significant relationships between 2-day mean of waking cortisol or cortisol rise and other measures of sleep. |
|||
| 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 | 2-day serum cortisol at 8am and 4pm: amplitude | Fatigue: EORTC QLQ-C30 v2. Depression: HADS |
cross-sectional | not tested | Examined associations between r24 status and cortisol only. High r24 patients had higher serum cortisol ratios between 8am and 4pm than low r24 patients. | ||
| Sannes et al. (175) | X | Endometrial cancer (n=82) | Nonmetastatic, before surgery | 3-day salivary cortisol, 4 times/day: diurnal cortisol slope, intradindividual variability | Depression: SIGH-AD | cross-sectional | yes | Depressive symptoms unrelated to cortisol slope. After controlling for presence of poorer prognosis cancer subtypes, greater depressive symptoms were significantly related to greater cortisol intraindividual variability. |
|||
| Schmidt et al. (176) | X | Breast cancer (n=265) | T0: pre-adjuvant treatment T1: week 7 T2: post-intervention week 13 |
1-day salivary cortisol, 5 times/day: CAR, AUC, diurnal cortisol slope | Fatigue: FAQ | longitudinal | yes |
Cross-sectional results:
Higher evening cortisol levels associated with higher physical fatigue levels T0 and T1. Larger AUC associated with higher physical fatigue levels. At T0, physical fatigue associated with flatter cortisol slope and higher CAR, but not at T1 and T2. Longitudinal results: Significant positive associations of change in evening cortisol level and AUC with change in physical fatigue, but no association with change in morning cortisol, CAR, or slope. Changes in affective or cognitive fatigue not associated with changes in cortisol parameters. |
|||
| Schrepf et al. (160) | X | X | Ovarian cancer (n=117) | T1: Prior to surgery T2: 6 months T3: 1 year |
3-day salivary cortisol, 3 times/day: mean cortisol, diurnal cortisol slope | Fatigue: POMS-SF Sleep: how many hours of sleep? Depression: CES-D (vegetative symptoms) |
longitudinal | yes | At 6 months, reductions in nocturnal cortisol secretion and more normalized diurnal cortisol rhythm, maintained at 1 year. Reductions in nocturnal cortisol associated with declines in fatigue, and marginally with vegetative depression. |
||
| Sephton et al. (177) | X | Breast cancer (n=72) | Metastatic | 3-day salivary cortisol, 4 times/day: diurnal mean cortisol level, diurnal cortisol slope |
Depression: CES-D | cross-sectional | yes | Depression scores were uncorrelated with mean cortisol levels. Patients with greater depression had higher morning cortisol and accentuated diurnal cortisol rhythms. |
|||
| Tell et al. (178) | X | X | X | Breast cancer (n=130) | Recently diagnosed, after surgery | 2-day salivary cortisol, 5 times/day: wake-up cortisol, CAR, linear and quadratic slope from wake-up to bedtime | Fatigue: MFSI-SF Sleep: PSQI |
cross-sectional | yes | Significant associations between ongoing fatigue and sleep quality and cortisol parameters: Women reporting greater ongoing fatigue had higher cortisol levels upon awakening Elevation in ongoing fatigue associated with a less pronounced CAR, slower decline over the day. Poor sleep quality associated with linear slope (flatter), but not associated with cortisol upon awakening, CAR or quadratic change. Reduction in sleep quality predicted slower cortisol decline in linear slope. |
|
| Vedhara et al. (179) | X | Breast cancer (n=85) and healthy controls (n=59) | Newly diagnosed | 2-day salivary cortisol, 4 times/day: AUC, diurnal cortisol, early morning peak | Depression: HADS | cross-sectional | yes | No associations between cortisol variables and depressed mood or distress among cancer patients. | |||
| Weinrib et al. (38) | X | X | Ovarian cancer (n=100), benign disease (n=77), healthy women (n=33) | Suspected ovarian cancer | 3-day salivary cortisol, 3 times/day: diurnal cortisol variability, nocturnal cortisol | Fatigue: POMS fatigue subscale Depression: CES-D |
cross-sectional | yes | Ovarian cancer patients had significantly elevated nocturnal cortisol and diminished cortisol variability compared with women with benign disease and healthy women. Among cancer patients, higher nocturnal cortisol and less cortisol variability associated with greater functional disability, fatigue, and vegetative depression. |
||
| Zeitzer et al. (115) | X | Breast cancer (n=97) and healthy controls (n=24) | Advanced | 28-h plasma cortisol at 20-60 minute intervals: Diurnal variation, phase, amplitude, mesor, phase angles |
Sleep: polysomnography | cross-sectional | yes | The circadian pattern of cortisol (timing, timing relative to sleep, or amplitude) was indistinguishable between patients and controls. There was an aberrant spike of cortisol during the sleep of a subset of women, during which there was an eightfold increase in the amount of objectively measured wake time. This cortisol aberration was associated with shorter disease-free interval. |
AUC, Area Under the Curve; BDI, Beck Depression Inventory; BFI, Brief Fatigue Inventory; BFS, Bidimensional Fatigue Scale; CAR, Cortisol Awakening Response; CES-D, Center for Epidemiologic Studies Depression Scale; cortisol VAR, relative diurnal variation of cortisol; EORTC QLQ-C30 v.2, The European Organization for Research and Treatment of Cancer Core Quality of Life Questionnaire version 2; FACT-F, Functional Assessment of Cancer Therapy: Fatigue; FAQ, Fatigue Assessment Questionnaire; FCS, Fatigue Catastrophising Scale; HADS, Hospital Anxiety and Depression Scale; HADS-D, Hospital Anxiety and Depression Scale – Depression; MFSI-SF, Multidimensional Fatigue Symptom Inventory – Short Form; MOS, Medical Outcomes Study; PANAS, The Positive and Negative Affect Schedule; POMS, Profile Of Mood States; PSQI, Pittsburgh Sleep Quality Index; r24, autocorrelation coefficient at 24 hours; RAVLT, Rey Auditory Verbal Learning Test; SCID for CRF, Structured Clinical Interview for the Diagnostic and Statistical Manual – IV to diagnose Cancer-Related Fatigue; SE, sleep efficiency; SF-36, 36-item Short-Form Survey; SIGH-AD, Structured Interview Guide for the Hamilton Depression Inventory; SOL, sleep onset latency; WASO, wake after sleep onset.