Characteristics of Cancer-Related Fatigue and Concomitant Sleep Disturbance in Cancer Patients

Abstract

Context:

Cancer patients often experience cancer-related fatigue (CRF) and sleep disturbances due to cancer and cancer treatment, and symptoms can persist long after treatment. Despite these common occurrences, few studies simultaneously characterize CRF and sleep architecture among cancer patients.

Objectives:

The objective was to characterize CRF and the sleep architecture of patients seen in a CRF clinic and sleep clinic at the University of Texas MD Anderson Cancer Center.

Methods:

CRF Clinic medical records were retrospectively reviewed from September 1, 2006, to September 30, 2010, for self-reported performance status, fatigue, pain, sleep disturbance, depression, anxiety, and sleepiness (n = 219). Polysomnography results were recorded for those referred for additional sleep consultation (n = 39).

Results:

Notably, patients often reported fatigue, sleep disturbance, excessive daytime sleepiness, and a majority of patients referred for a sleep consultation were diagnosed with obstructive sleep apnea (n = 33).

Conclusion:

The results highlight the promise of an interdisciplinary collaboration between dedicated a CRF clinic and sleep clinic to conduct effective assessments to identify treatable CRF and sleep disorders.

INTRODUCTION

Cancer-related fatigue (CRF) is among the most prevalent, distressing, and often overlooked side effects of cancer and cancer treatment in adults, and they can persist for years after treatment completion.¹ Although estimates of CRF may vary widely according to patient age, treatment type, cancer type, disease stage, and time since treatment completion, virtually all cancer patients are expected to experience some form of fatigue,² and up to 75% will experience sleep disturbances³—both of which can persist for years after treatment completion. These figures are in stark contrast to 0.89% prevalence rate for chronic fatigue syndrome,⁴ and 7.1% prevalence rate for a sleep disorder nationally.⁵ Sleep disturbances in turn can result in worsening daytime fatigue, neurocognitive deficits, mood disorders and lowering pain thresholds.^{6, 7} Detection and evaluation of sleep disturbances in the context of CRF is especially important, for it may uncover an underlying sleep disorder.

CRF is persistent physical, emotional, and/or cognitive tiredness or exhaustion related to cancer or cancer treatment disproportional to recent activity that causes distress and interferes with usual functioning.⁸ In comparison, sleep disorders affect or disrupt sleep-related behavior. CRF and sleep disorders are distinct yet interrelated conditions⁹ sharing common biological mechanisms.¹⁰ Whereas cancer and cancer treatment can lead to CRF and sleep disorders, daytime sleepiness and sleep disturbances can influence perceptions of fatigue.¹¹ The high prevalence and many possible etiologies of CRF and sleep disorders among cancer patients call for multidisciplinary approaches to the evaluation and treatment of fatigue and sleep disturbances.¹¹

Patients often report symptom clusters, which are defined as the experience of multiple co-occurring and interrelated symptoms that are stable across time.¹² For example, the most commonly identified clusters among cancer survivors include symptoms of fatigue, sleep disturbance, pain, depression, and anxiety¹³ with a fatigue, sleep, and pain symptom cluster persistent before, during, and after cancer treatment.¹⁴ However, limitations in past studies characterizing symptoms among cancer survivors include patients without physical or psycholoigical diagnoses or impariments ^{15, 16}, lacking measures of sleep¹⁷ or the use of actigraph-measured sleep variables,¹⁶ cancer-specific samples,^{14, 17} and PSG measured sleep without other subjective measures of QoL.¹⁸ Given the clinical importance of disentangling interrelated symptoms to inform supportive cancer care, a multidisciplinary approach holistlcally considers the patients experience. Thus, our primary aim was to describe patients with CRF and review findings of a sleep evaluation and polysomnography (PSG) among those with sleep disturbance, while also describing other important co-morbid symptoms.

METHODS

Sample, Context, and Procedures

We retrospectively reviewed the medical records of all patients seen in the University of Texas MD Anderson Cancer Center Cancer-Related Fatigue clinic from September 1, 2006 to September 30, 2010. Information gathered from clinical records included age, sex, race/ethnicity, cancer diagnosis, comorbidities, medication use, as well as on-site self-reported questionnaire and PSG scores. Inclusion criteria were 18 years or older with a diagnosis of CRF, and patients evaluated at the CRF clinic evaluated for daytime and nighttime sleep quality.

The CRF Clinic was created in 1998 to improve the quality of life for cancer patients by alleviating fatigue; educating patients, families, and healthcare workers about CRF; evaluate CRF, and conduct research on innovative CRF treatments. The Sleep Center opened in September 2006 and strives to understand and mitigate the effects of cancer and cancer treatment on patients with disordered sleep and CRF.

Patients in the sample were referred to the Sleep Clinic for PSG if patients noted sleep disturbances during evaluation at the CRF clinic. PSG were conducted at the Sleep Center, which is accredited by the American Academy of Sleep Medicine. PSG were recorded with standard 12 channel polysomnography equipment (Nihon Kohden, Foothills Ranch, CA) and manually scored according to The AASM Manual for the Scoring of Sleep and Associated Events: Rules, Terminology and Technical Specifications 2007 by a registered polysomnographic technologist and interpreted under the medical direction of the director of the Sleep Clinic.

The study was approved by the MDACC Institutional Review Board.

Measures

Higher self-reported questionnaire scores reflect greater endorsement of the measured construct unless otherwise specified.

The Brief Fatigue Inventory (BFI) is a validated 10-item instrument developed among cancer patients to assess fatigue. Patients were asked to rate their fatigue for the last 24 hours on a 11-point scale ranging from 0 (“no fatigue”) to 10 (“fatigue as bad as you can imagine”).¹⁹

The Brief Pain Inventory (BPI) is a 7-item instrument that was used to measure overall self-reported pain. Patients were asked to rate their pain for the last 24 hours on an 10-point Likert scale ranging from 0 (“No pain”) to 10 (“Pain as bad as you can imagine”).²⁰

The Brief Sleep Disturbance Scale (BSDS) is 5-item instrument that was used to measure self-reported sleep quality during the past 24 hours. Patients were asked to rate how true sleep quality related statements (e.g., “I feel rested and refreshed on awakening”, and “I have difficulty falling asleep at night”) were on 10-point Likert scale ranging from 0 (“Not at all true of me”) to 10 (“Completely true of me”).²¹

The Beck Depression Inventory-II (BDI-II) is a 21-item instrument evaluating symptoms of depression. Patients were asked to rate symptoms on a 3-point Likert scale ranging from 0 (“Absent”) to 3 (“Most severe”).²² The BDI-II has demonstrated adequate reliability among cancer patients.²³

The Beck Anxiety Inventory is a 21-item instrument that was used to measure severity of anxiety symptoms in the past month. Patients were asked to rate on symptoms of anxiety on a 4-point Likert scale ranging from 0 (“Not at all”) to 3 (“Severe”).²⁴

The Epworth Sleepiness Scale is an 8-item instrument that was used to measure general level of daytime sleepiness “in recent times.” Patients were asked to rate daytime behaviors related to sleepiness on a 4-point Likert scale ranging from 0 (“Would never doze”) to 3 (“High chance of dozing”).²⁵ The ESS has demonstrated adequate reliability among cancer patients.²⁶

The Zubrod scale is a one-item performance scale measure of a of a patient’s ability to complete activities of daily living without assistance. The scale ranges from 0 (fully functional and asymptomatic) to 4 (bedridden).²⁷

Measures from PSG included total sleep time (TST), total recording time, wake after sleep onset (i.e., time spent awake after sleep onset; WASO), sleep onset latency (i.e., time taken to fall asleep; SOL), sleep efficiency (i.e., percentage of time in bed actually spent asleep), Apnea Hypopnea Index (i.e., number of apneas/hypopneas per hour sleep; AHI), and periodic limb movement with arousal index (i.e., repetitive leg movement during sleep). In addition, percentage of time spent in the four sleep stages (Stage 1, 2, 3, and rapid eye movement) were recorded. In-depth descriptions of the sleep stages²⁸ and other measures are described elsewhere.^{29, 30}

Data analysis

Descriptive statistics were used to examine the distribution of demographic characteristics (age, sex, race), cancer and comorbidities (cancer type, comorbid diseases), medication use, health-related self-report questionnaire scores (Zubrod score, fatigue, pain sleep, depression, anxiety), PSG scores, and sleep disorder diagnoses. We provide descriptive statistics for the full sample (n = 219) and then only those who received PSG (n = 39).

RESULTS

The overall sample consisted of 219 patients (Table 1), and 39 patients underwent polysomnography (Table 2). The majority of patients in both groups were not undergoing active treatment for their cancers. Overall, the patients reported adequate functioning. More than one fifth of the patients (21%) had a Zubrod performance score of at least 2, and the average score suggested that patients required occasional assistance for daily activities. On average, the fatigue scores were within the moderate range, the pain scores were within the moderate-severe range, the sleep disturbance scores were within the moderate-severe range, and the daytime sleepiness scores were within the excessive range. Both anxiety and depression severity were within the low-mild range.

Table 1.

Patient characteristics, comorbidities, and self-report surveys for all patients (n = 219).

Variable	n (%) / Mean ± SD
Mean age, years	57.7 ± 11.9
Sex
Male	75 (34)
Female	144 (66)
Race
White	169 (77)
Black	26 (12)
Hispanic/Latino	16 (7)
Asian/Pacific Islander	6 (3)
Other	1 (<1)
BMI, kg/m2	29.3 ± 6.9
No evidence of disease,	155 (71)
Cancer type
Breast	79 (36)
Lymphoma	18 (8)
Head and neck	11 (5)
Myeloma	10 (5)
Lung	10 (5)
Melanoma	7 (3)
Colon/rectal	7 (3)
Ovarian	7 (3)
Comorbidities
Hypertension	98 (45)
Depression	80 (37)
Hypothyroidism	43 (20)
Cardiovascular disease	40 (18)
Diabetes	33 (15)
Stroke	6 (3)
Renal insufficiency	5 (2)
COPD	8 (4)
Prescribed medication use
Wake stimulating agents	18 (8)
Antidepressants	17 (8)
Non benzodiazepine sedative hypnotics	15 (7)
Benzodiazepines	6 (3)
Pain medication	6 (3)
Anxiolytics	1 (<1)
Self-reported measures
Zubrod score	1.13 ± 0.56
0	16 (7.3)
1	133 (60.7)
2	37 (16.9)
3	2 (0.9)
Missing	31 (14.1)
Brief Fatigue Inventory score	6.09 ± 1.99
Mild (<4)	76 (34.7)
Moderate (4–6.9)	56 (25.6)
Severe (7–10)	86 (39.3)
Missing	1 (0.5)
Brief Pain Inventory score	4.59 ± 3.21
No/Mild (0–4)	100 (45.7)
Severe (5–10)	117 (56.4)
Missing	2 (0.9)
Brief Sleep Disturbance Scale score	23.21 ± 8.88
None (0–21.9)	95 (46.3)
Mild (22–29.9)	58 (27.4)
Moderate (30–34.9)	30 (12.6)
Severe (≥35)	30 (13.7)
Missing	6 (2.7)
Beck Depression Inventory-II score	15.81 ± 9.38
Minimal (<13)	105 (47.9)
Mild/Moderate (14–28)	85 (38.8)
Severe (≥29)	27 (12.3)
Missing	2 (0.9)
Beck Anxiety Inventory score	13.10 ± 9.04
No/Minimal risk (<16)	141 (64.4)
Clinical risk (≥16)	70 (32.0)
Missing	8 (3.7)

Abbreviations: COPD, chronic obstructive pulmonary disease.

Table 2.

Patient characteristics, comorbidities, self-report surveys, polysomnographic characteristics, and sleep disorders for patients who underwent polysomnography (n=39).

Variable	n (%) / Mean (SD)
Mean age, years	58.0 ± 10.3
Sex
Male	12 (31)
Female	27 (69)
Race
White	27 (69)
Black	7 (18)
Hispanic/Latino	4 (10)
Asian/Pacific Islander	1 (3)
Other	0
BMI, kg/m2	32.1 ± 2.2
No evidence of disease	33 (85)
Cancer type
Breast	17 (44)
Lymphoma	4 (10)
Head and neck	2 (5)
Myeloma	2 (5)
Lung	0
Melanoma	2 (5)
Colon/rectal	0
Ovarian	0
Comorbidities
Hypertension	18 (46)
Depression	16 (41)
Hypothyroidism	2 (5)
Cardiovascular disease	7 (18)
Diabetes	10 (26)
Stroke	1 (3)
Renal insufficiency	1 (3)
COPD	1 (3)
Prescribed medication use
Wake stimulating agents	3 (8)
Antidepressants	1 (3)
Non benzodiazepine sedative hypnotics	4 (10)
Benzodiazepines	0
Pain medication	2 (5)
Anxiolytics	0
Self-reported measures
Zubrod score	1.03 ± 0.63
0	7 (17.9)
1	24 (61.5)
2	8 (20.5)
3	0
Brief Fatigue Inventory score	6.42 ± 1.73
Mild (<4)	9 (23.1)
Moderate (4–6.9)	11 (28.2)
Severe (7–10)	19 (48.7)
Brief Pain Inventory score	5.69 ± 3.04
No/Mild (0–4)	12 (30.8)
Severe (5–10)	27 (69.2)
Brief Sleep Disturbance Scale score	24.45 ± 8.53
None (0–21.9)	14 (36.8)
Mild (22–29.9)	10 (26.3)
Moderate (30–34.9)	8 (21.1)
Severe (≥35)	6 (15.8)
Beck Depression Inventory-II score	17.31 ± 9.46
Minimal (<13)	16 (41.0)
Mild/Moderate (14–28)	18 (46.1)
Severe (≥29)	5 (12.8)
Beck Anxiety Inventory score	13.69 ± 9.80
No/Minimal risk (<16)	24 (61.5)
Clinical risk (≥16)	12 (30.8)
Missing	3 (7.7)
Epworth Sleepiness Scale score	12.85 ± 6.21
Minimal (≤10)	13 (33.3)
Excessive (>10)	23 (59.0)
Missing	3 (7.7)
Total sleep time, hours	6.50 ± 1.02
Total recording time, hours	7.71 ± 0.63
WASO, minutes	58.41 ± 41.74
SOL, minutes	11.28 ± 12.83
Sleep efficiency	86.47 (9.32)
Sleep stage
W	13.22 (9.27)
N1	11.51 (8.70)
N2	59.05 (12.07)
N3	3.93 (5.14)
REM	13.21 (6.31)
Apnea-Hypopnea Index score	13.52 ± 16.79
PLM index with arousal score	4.94 ± 8.33
PLM index without arousal score	22.35 ± 43.35
Sleep disorders
OSA	33 (85)
Hypersomnolence	3 (8)
PLM	1 (3)
Dyssomnia	1 (3)
Nocturnal hypoxemia	1 (3)

Abbreviations: COPD, chronic obstructive pulmonary disease; PSG, polysomnography; WASO, wake after sleep onset; SOL, sleep onset latency; W, waking; REM, rapid eye movement; PLM, periodic limb movement.

Patients who underwent PSG self-reported physical and mental health functioning scores within the same range as that in the entire sample (Table 2). PSG revealed disturbed sleep and abnormal sleep architecture. The average (± SD) total sleep time was 6.50 ± 1.02 hours with elevated WASO durations (58.41 ± 41.74 minutes) and SOL (11.28 ± 12.83 minutes) despite acceptable sleep efficiency (86.47% ± 9.32%). PSG also revealed increased Stage 1 (N1; 12%) and 2 (N2; 59%) sleep with commensurate decreases in Stage 3 (N3; 4%) and rapid eye movement (13%) sleep. Obstructive sleep apnea (OSA) was identified in the majority of cases (n = 33), and on average, severity levels were within the severe range (>30/hour) based on their respiratory disturbance index scores. The average Periodic Limb Movement (PLM) index score with and without arousal score was 4.84 ± 8.33 and 22.35 ± 43.35, respectively, and PLM was noted in one case (3%). Hypersomnolence (i.e., excessive daytime sleepiness) was noted in three cases (8%), dyssomnia (i.e., difficulty falling asleep) was noted in one case (3%), and nocturnal hypoxemia (i.e., temporary drop in oxygen saturation during sleep) was noted in one case (3%).

DISCUSSION

The aim of the study was to characterize sleep disturbances in patients with CRF. Indeed, CRF places significant symptom burden on cancer patients and survivors, and our patients experienced moderate levels of fatigue. Further, PSG results for those with sleep disturbance revealed OSA and other treatable underlying sleep disorders. Our findings highlight the importance of screening cancer patients with fatigue for sleep disturbances given their high risk of comorbid sleep disorders.

The sleep architecture in patients with underlying sleep disorders are abnormal, and likely contribute to worsening daytime symptoms and potentially exacerbating fatigue. Compared with the general population,³¹ these patients experienced 20% and 70% less N3 and rapid eye movement sleep, respectively. This is likely due to the high incidence of OSA among those undergoing PSG and, consistent with previous studies, higher than normal prevalence of OSA among breast³² and head and neck³³ cancer patients. Following National Comprehensive Cancer Network guidelines,⁸ periodic screening for fatigue and sleep disturbances may help patients receive preventive or early treatment for fatigue and sleep disorders, and detect OSA, which has been associated with greater tumor aggressiveness³⁴ and colorectal neoplasias.³⁵ Indeed, a disrupted circadian rhythm has been offered as an important contributor to CRF through dysregulated cytokine production.³⁶

As one of the only dedicated CRF clinics in a comprehensive cancer center, screening for sleep disturbances are integrated into screening algorithms to mitigate oversight of sleep problems commonly experienced among cancer patients.³⁷ Prompt referral to boarded sleep specialists with polysomnography in those indicated helped to identify sleep disorders among cancer patients. Building upon an effective multimodal and integrative approach to CRF treatment,³⁸ referrals for additional sleep disorder screening using PSG, when clinically indicated, can help identify conditions that negatively impact patient-related outcomes. For example, early detection and treatment for sleep disorders may decrease hospital lengths of stay³⁹ and decrease post-operative complications.⁴⁰

Review of descriptive statistics revealed elevated levels of sleep disturbance, pain, fatigue, and depression among cancer patients seen at the CRF Clinic. The globally elevated scores are consistent with prior symptom cluster research suggesting persistent fatigue, sleep disturbance, and pain symptoms among cancer survivors even long after cancer treatment.¹⁴ Notably, more patients who received PSG, compared to those who did not, reported severe fatigue (PSG: 48.7%, no PSG: 39.3%), severe pain (PSG: 69.2%, no PSG: 56.4%), and modestly more severe sleep disturbances (PSG: 15.8%, no PSG: 13.7%). The high percentage of self-reported severe pain among those who ultimately received a PSG may highlight the link between pain and OSA.⁴¹ However, this assertion requires empirical testing among the cancer patient population.

These data should be interpreted considering the following limitations. First, the retrospective descriptive and cross-sectional nature of the data precluded us from inferring temporal relationships between cancer development and OSA. Indeed, chemotherapy can increase risk of OSA among head and neck patients and cause fatigue and sleep disturbances.⁴² However, future research should examine psychosocial and sociodemographic predictors of sleep disturbance and CRF among cancer survivors. Second, given the descriptive nature of the study, we did not empirically test a hypothesis regarding relationships between symptoms. While the existence of symptom clusters are well-documented, it is unclear if symptom clusters are more appropriately characterized as typologies or along a severity continuum. For example, advanced statistical methods (e.g., latent class/profile analysis) have identified various symptom clusters among cancer patients, yet results generally represent high, moderate, and low symptom severity.¹⁶ Second, our sample only included patients referred to the CRF and Sleep clinic, and do not include potentially many patients with sleep disturbances without reported or detected CRF. Lastly our sample includes patients from one cancer center and may not generalize to the patient population.

Clinical Implications

Greater collaboration between CRF and sleep disorder clinics can benefit cancer patients in numerous ways. First, patients can benefit from early diagnosis of sleep disorders. Given the high prevalence of CRF, clinicians knowledgeable about the importance of screening for sleep disorders may be able to refer patients with underlying sleep disturbances to a sleep clinic earlier. MDACC CRF Clinic physicians and staff are highly trained in detecting sleep disorders as a result of the longstanding collaboration with the Sleep Center. Our data suggested that sleep disorders were often confirmed via additional PSG assessment at the Sleep Center. Further, OSA was the most commonly diagnosed sleep disorder suggesting clinicians be aware of the hallmark symptoms (e.g., snoring, waking with loss of breath) as well as patients with a history of radiation therapy, head and neck cancer, or squamous cell pathology.³³ While patients with any self-reported sleep disturbance were referred for PSG, clinicians should use well-validated screeners for OSA such as the STOP-BANG.³⁹

Second, patients can benefit from a multidisciplinary approach to CRF and sleep disorder treatment. The close ties between the CRF and Sleep Clinic allowed efforts to be coordinated across the different levels of care across nurses, counselors, and physicians. For example, when sleep disturbances are present upon initial evaluation at the CRF Clinic, patients are referred to the sleep clinic for additional evaluation to determine the need for a primary sleep disorder or symptom management. Prioritizing the treatment of symptom clusters, patients are also referred to rehabilitation services for physical therapy, and integrative care for pain management. Thus, the addition of a CRF and sleep specialist to the treatment team can help differentiate symptoms of fatigue and sleepiness to better guide treatment.

CONCLUSION

CRF and sleep disturbances are among the most prevalent and distressing symptoms related to cancer and cancer treatment. Our study of fatigue clinic referrals with additional PSG conducted in a sleep clinic revealed a high prevalence of both OSA and other sleep disorders. Given the significant effect of CRF and sleep disorders on patient-related cancer outcomes, the use of collaborative interdisciplinary teams to conduct effective assessments and treatments of CRF and sleep disturbances is recommended.⁴³

Table 3.

Description of cancer-related quality of life and polysomnography measures.

Instrument	Assessment	Time interval	Cut-off Scores
Self-reported measures
Zubrod27	Performance status for patients with cancer	Current	0: Normal activity 1: Symptomatic, ambulatory, and cares for self 2: Ambulatory (>50%) and in need of occasional assistance 3: Ambulatory (<50%) and in need of nursing care 4: Bedridden
Brief Fatigue Inventory19	Fatigue	Past 24 hours	<7: Non-severe fatigue 7–10: Severe fatigue
Brief Pain Inventory20	Perceived pain	Past 24 hours or past week	0–4: No/Mild 5–10: Severe
Brief Sleep Disturbance Scale21	Sleep disturbance	Past 24 hours	0–21.9: None 22–29.9: Mild 30–34.9: Moderate ≥35: Severe
Beck Depression Inventory-II22	Depression	Past 2 weeks	<13: Minimal 14–28: Mild/Moderate ≥29: Severe
Beck Anxiety Inventory24	Anxiety	Past month	<16: No/Minimal risk ≥16: Clinical risk
Epworth25	Daytime sleepiness	Recent times	<10: None/minimal daytime sleepiness ≥10:Severe daytime sleepiness.
Objective measure	Assessment	Time interval	General population norms31 (95% CI)
Polysomnography	Sleep architecture	Past night	TST: 394.6 m (288.4–400.8) WASO: 48.2 m (43.8–52.6) SOL: 15.4 m (14.2–16.7) SE: 85.7% (84.8–86.6) AHI: 2.9 (2.6–3.1) PLM: 2.5 (2.1–2.9) Time spent: Stage 1: 7.9% (7.3–8.5) Stage 2: 51.4% (50.2–52.6) Stage 3: 20.4% (19.0–21.8) REM: 19.0% (18.5–19.6)

Note: TST = total sleep time (minutes); WASO = wake after sleep onset (minutes); SOL = sleep onset latency (minutes); SE = sleep efficiency; REM = rapid eye movement AHI = apnea hypopnea index (apnea/hypopnea events/hour); PLM = periodic limb movement (PLM events/hour).