Skip to main content
PMC home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2023 Feb 1.
Published in final edited form as: Cancer Epidemiol. 2021 Nov 17;76:102057. doi: 10.1016/j.canep.2021.102057

Sleep problems and risk of cancer incidence and mortality in an older cohort: The Cardiovascular Health Study (CHS)

Arthur Sillah 1,2, Nathaniel F Watson 3, Ulrike Peters 1,2, Mary L Biggs 1, F Javier Nieto 5, Christopher I Li 1,2, David Gozal 4, Timothy Thornton 1, Sonnah Barrie 6, Amanda I Phipps 1,2
PMCID: PMC8792277  NIHMSID: NIHMS1753684  PMID: 34798387

Abstract

Background:

Sleep problems (SP) can indicate underlying sleep disorders, such as obstructive sleep apnea, which may adversely impact cancer risk and mortality.

Methods:

We assessed the association of baseline and longitudinal sleep apnea and insomnia symptoms with incident cancer (N=3,930) and cancer mortality (N=4,580) in the Cardiovascular Health Study. We used Cox proportional hazards regression to calculate adjusted hazard ratios (HR) and 95% confidence intervals (CI) to evaluate the associations.

Results:

Overall, 885 incident cancers and 804 cancer deaths were identified over a median follow-up of 12 and 14 years, respectively. Compared to participants who reported no sleep apnea symptoms, the risk of incident cancer was inversely associated [(HR (95%CI)] with snoring [0.84 (0.71, 0.99)]. We noted an elevated prostate cancer incidence for apnea [2.34 (1.32, 4.15)] and snoring [1.69 (1.11, 2.57)]. We also noted an elevated HR for lymphatic or hematopoietic cancers [daytime sleepiness: 1.81 (1.06, 3.08)]. We found an inverse relationship for cancer mortality with respect to snoring [0.73 (0.62, 0.8)] and apnea [(0.69 (0.51, 0.94))]. We noted a significant inverse relationship between difficulty falling asleep and colorectal cancer death [0.32 (0.15, 0.69)] and snoring with lung cancer death [0.56 (0.35, 0.89)].

Conclusions:

The relationship between SP and cancer risk and mortality was heterogeneous. Larger prospective studies addressing more cancer sites, molecular type-specific associations, and better longitudinal SP assessments are needed for improved delineation of SP-cancer risk dyad.

Keywords: Sleep apnea symptoms, Insomnia symptoms, cancer incidence, cancer mortality, circadian rhythm

Introduction

Sleep problems are common, with an estimated 50–70 million U.S. adults suffering from chronic sleep and wakefulness problems1 and roughly 35% averaging less than the recommended 7–8 hours of nightly sleep.1 Even in the absence of a formal diagnosis, sleep problems (SP) are frequently indicative of an underlying sleep disorder, such as obstructive sleep apnea (OSA)2 and insomnia.3 Sleep problems have been linked to numerous adverse health outcomes, including type 2 diabetes and hypertension,4 cardiovascular disease,5 mortality,6 substantial increases in healthcare utilization,7 and in increased motor vehicle accidents.8 Together, these effects impose substantial deleterious social, economic, and quality of life consequences for older adults.9

Evidence suggests a bidirectional relationship between different SP and circadian rhythm disruptions, which could be impacted by certain environmental inputs, such as shiftwork or other occupations with irregular sleep-wake cycles.10,11 Regardless of the directionality, SP resulting in/from circadian rhythm disruptions are a risk factor for cancer development and progression. Animal and in vitro studies highlight the tumor suppressor role of the circadian system through its activities on the cell cycle, cellular proliferation, regulation of apoptosis, and homeostasis of angiogenesis.12–14

Most epidemiologic studies of SP in relation to human cancer have focused on the impact of shift work and sleep duration. Numerous empirical findings showing the adverse impact of shift work on cancer development and progression have led the International Agency for Research on Cancer (IARC) to classify shift work involving significant circadian disruptions as a probable carcinogen.15,16 Evidence of this carcinogenic effect is strongest for breast cancer,17 but also suggests an increased risk of incident prostate and colorectal cancers.17 Furthermore, short sleep duration (typically defined as <7 hours) has been linked with increased risks of cancer overall,18,19 gastric cancer20, and breast cancer21 and poor overall cancer and breast cancer survival, particularly when in combination with snoring.22,23 With respect to other aspects of SP, some recent studies have provided suggestive evidence for an association between OSA and cancer incidence24–27 and mortality.28,29 A recent meta-analysis of eight studies found an overall increased risk of cancer for individuals with insomnia compared to those without insomnia, and reported that risk was mostly driven by studies conducted in women.30

Beyond this prior work, several methodologic challenges in evaluating the relationship of sleep disturbances with cancer have imposed limitations to epidemiologic studies. For instance, sleep patterns are likely to change over time and with age, and wide domains of sleep disordered symptoms indicative of OSA and insomnia have not been simultaneously assessed with respect to cancer incidence and mortality in a cohort of older individuals. To advance our understanding of various components of sleep quality and sleep disorder symptoms, we assessed several dimensions of SP and their potential impact on cancer incidence and mortality, using longitudinally collected sleep data from the Cardiovascular Health Study (CHS).

Methods

Study Participants and settings

The CHS is an observational cohort study of men and women aged ≥65 years recruited from a random sample of the Health Care Financing Administration Medicare eligibility list of 4 US communities: Forsyth Country, North Carolina; Sacramento County, California; Washington County, Maryland; and Pittsburgh, Pennsylvania.31 Participants forming part of the original cohort were enrolled between 1989–1990 (94% White, 5% African Americans, 1% other ethnic group; N=5,201), and an additional supplemental cohort was enrolled between 1992–1993 (predominantly African American (AA), N=687).31 For this study, we excluded participants missing SP and covariate data at baseline (n=1,308); for analyses of cancer incidence, we also excluded participants who had a history of cancer at the time of recruitment (N=650). The excluded participants did not differ from the analytic sample with respect to age, sex, and racial distribution.

Sleep problems ascertainment

Longitudinal sleep data was ascertained through self-reported sleep symptoms collected at baseline (1989–90) and follow-up assessments in 1991–92, 1993–94, and 1994–95 for the original CHS cohort. For the supplemental cohort, these data were collected at baseline (1992–93) and the follow-up assessment in 1994–95. The same set of sleep questions was asked at each of the visits, and have been previously validated with objectively measured sleep data from a subset of participants enrolled in the Sleep Heart Health Study (n= 1,240) who underwent an overnight polysomnography study.32,33 Details of the validation study are described elsewhere.33 Participants reported the following sleep apnea symptoms (SAS):

  1. Are you usually sleepy in the daytime? (Daytime sleepiness)

  2. Has your spouse/roommate complained about your loud snoring? (Snoring)

  3. Has anyone observed you while sleeping to have episodes where you stop breathing for a while & then snore? (Observed apnea)

In addition to daytime sleepiness, participants also reported the following insomnia symptoms (IS):

  1. Do you usually have trouble falling asleep? (Problem falling asleep)

  2. Do you usually wake up several times at night? (Problem maintaining sleep)

  3. Do you usually wake up far too early? (Early morning wakefulness)

Other variables

We included demographic attributes [i.e., sex, age, race (white, black, other), ethnicity (Hispanic, non-Hispanic) marital status (partnered, unpartnered), the number of years of education completed)], and behavioral factors [i.e., smoking (current, former and never), weekly alcohol consumption, and physical activity], collected by questionnaires at baseline and follow-up visits. Alcohol consumption per week was computed from the reported usual frequency of consumption of beer, wine, and liquor, and the usual number of drinks consumed on each occasion,31 and categorized into <1, 1–6, 7–13, and 14+ alcoholic beverages per week. Physical activity (in kilocalories per week) was collected using a modified Minnesota Leisure-Time Activities questionnaire31,34 and categorized into quartiles. Body mass index (BMI) was calculated by dividing study staff-measured weight in kilograms by height in meters squared and grouped into standard categories (normal < 25 kg/m2, overweight 25–29 kg/m2, obese ≥30 kg/m2). Participants with diabetes were identified based on fasting blood glucose >=126 mg/dL, a non-fasting blood glucose >=200 mg/dL, or use of a diabetes medication. Participants were also asked about their disease history (diabetes, hypertension and coronary heart disease) and history of cardiovascular events were verified from hospital and physician records, as previously described.35 In particular, hypertension was ascertained based on measured systolic blood pressure (BP)>=140 OR diastolic BP >=90 OR self-reported physician diagnosis of hypertension plust taking a hypertension medication. 35

Cancer incidence ascertainment

Cancer incidence (overall and site-specific) was ascertained through 5 population-based cancer registries serving the 4 CHS regions; linkage to these data is complete through 2005.36 The completeness of the cancer ascertainment is 93–100% for the CHS registries, according to the North American Association of Central Cancer Registries (NAACCR).37

Cancer-specific mortality

Standardized protocols for the identification of cardiovascular events and deaths were implemented during follow-up.31 Deaths among CHS participants have been adjudicated through 2015 by a study-wide Event Review Committee made up of physicians representing the 4 study sites.38,39 The committee reviewed medical records, death certificates, and other information to adjudicate the underlying cause of death.

Statistical analysis

All analytical procedures were conducted using Stata 14.0 (College Station, Texas),40 with statistical significance considered at a 2-sided alpha value of 0.05. In descriptive analyses, we examined the distribution of participants’ baseline characteristics overall and stratified by the number of SAS. We calculated means (standard deviation) for normally distributed continuous variables and median (interquartile range) for non-normally distributed continuous variables and with percentages for categorical measures. We tabulated the overall and site-specific counts for incident cancers and subsequent cancer mortality along with their tumor attributes according to tumor summary stage.

We used Cox proportional hazards regression to evaluate the association between SP and cancer incidence and cancer mortality overall and for cancer sites with at least 50 incident cases and deaths. We calculated the hazard ratios (HR) and 95% confidence intervals (CI) for the associations. For the cancer incidence outcome analysis, 3930 participants free of cancer at baseline were included in the analysis. Participants contributed person-time to the analyses until cancer incidence, all-cause death, or end of follow up (31 December 2005) whichever came first. With respect to the cancer mortality outcome, 4580 participants were included regardless of their baseline cancer status. Participants contributed person-time to the analyses until cancer mortality or at the time of the last attended visit, other cause-specific death, or end of follow up (June 2015) whichever came first.

For both analyses, we modeled SP in several ways: 1) using baseline values only, 2) allowing SP to be time-varying by incorporating follow-up data as time-dependent, and 3) allowing SP variables to be time-varying by modeling a cumulative average of previous values of the SP variable (coded as 0,1) each year to examine chronic effects of SP. We also allowed BMI to vary for the time-varying analysis given its strong correlation with both SP and cancer outcomes.41 Each missing value of the time-varying SP and BMI was replaced by the last observed value of that variable.

In addition to running a separate model for the SAS, we also assessed the joint effects of the baseline values coded as 0 for no sleep apnea symptoms, 1 for having any symptom, and 2 for two or three symptoms. We treated the IS variables similarly by combining symptoms based on the number of the symptoms reported.

Finally, we ran several sensitivity analyses. We censored the first 2-years and 5-years of follow-up for cancer incidence and cancer mortality analyses, respectively, to reduce the possible influence of reverse causation (i.e. undiagnosed cancer or worsening cancer contributing to SP). In addition, we calculated a modified version of the 8-item validated STOP-BANG for assessment of OSA risk with scores ranging from 0–8 and categorized into low risk (0–2), intermediate risk (3–4), and high risk (5–8).42 The STOP-BANG uses information on whether a patient snores (S), experiences daytime tiredness (daytime sleepiness, T), breathing cessation during sleep (observed apnea, O), has high blood pressure (P), body mass index (B) >35 kg/m2, older than 50 years of age (A), > 40cm neck circumference (N), and male gender (G). Since neck circumference was only assessed for years 1994–95, we upweighted baseline BMI by assigning 0 for <25 kg/m2, 1 for 25–34 and 2 for ≥ 35 kg/m2. We also computed a STOP-BANG score for the supplementary cohort which was enrolled 1-year prior to neck circumference ascertainment. We opted to use the modified STOP-BANG (which had excellent agreement with the original STOP-BANG, kappa=0.76) to make use of most of the data in our sensitivity analysis. Furthermore, because of the strong relationship between BMI and SP and cancer outcomes, we also assessed the possibility of BMI as an effect modifier of the relationship of SP with cancer incidence and cancer mortality by presenting HR estimates stratified by BMI categories. In addition, we presented cancer stage-specific estimates. Lastly, because most prior studies of SP and cancer have primarily focused on breast and prostate cancers, we presented a sex-stratified analysis.

Models were adjusted for a priori selected confounders including: gender, enrollment wave (1989–90 or 2992–93), baseline age (continuous), smoking, BMI (continuous), diabetes, physical activity levels (continuous), and alcohol consumption (continuous). A formal test for proportional hazards assumptions was performed by evaluating the nonzero slope of the scaled Schoenfeld residuals on ranked failure times.43

Results

Patient characteristics

The mean age of the study population at enrollment was 73 years, 57% were female, 83% were white, 70% were partnered, 20% were at least a college graduate, 60% were overweight or obese, and 12% were current smokers. The distribution was similar across the SAS groups except that, among those with 2+ SAS symptoms, 35% were female, 81% were partnered, 70% were overweight or obese (Table 1).

Table 1:

Selected baseline characteristics overall and according to composite sleep apnea symptoms (SAS) in CHS

SAS composite group
Characteristics Total (n=4580)* 0 symptoms (n=2864) Any 1 symptom (n=1235) 2+ symptoms (n=481)
Age at baseline, mean (SD) 72.6 (5.5) 72.5 (5.4) 72.8 (5.8) 72.6 (5.4)
age group
64–69 35.6 36. 1 34.9 34.7
70–74 32.1 32.2 31.7 32.6
75–79 19.8 19.5 20.6 19.8
80–84 9.1 9.1 8.4 11.0
85+ 3.4 3.2 4.45 1.9
Partnered living 70.2 67.9 71.5 80.5
Female 56.6 62.4 55.7 34.9
Education
<HS 29.9 27.1 34.5 34.1
HS or GED 27.7 29.2 25.8 25.2
Vocational 8.3 8.7 7.9 7.5
Some college 14.0 14 14 13.5
College degree 10.2 10.4 9.8 10.2
Grad or professional deg 9.8 10.7 7.9 9.6
Race
White 83.4 85.1 81.4 78.6
African American/Black 16.2 14.6 18.4 20.6
other 0.4 0.4 0.24 0.8
Hispanic 1.2 1.0 1.3 1.5
BMI, kg/m 2
Mean (SD) 26.7 (4.7) 26.2 (4.5)) 27.8 (5.0) 27.7 (4.8)
<25 39.3 43.4 33.4 29.9
25–29 41.1 39.2 45.0 42.8
30+ 19.6 17.5 21.6 27.2
Smoking
Never 46.6 49.5 43.7 36.2
former 42.0 39.2 44.8 51.4
current 11.5 11.3 11.5 12.5
Alcohol beverages per week
median (IQR) 0.0 (0, 1.25) 0.0 (0, 1.25) 0.0 (0, 1.02) 0.019 (0, 1.5)
none 51.3 50.4 54.7 48.4
<1 18.3 18.8 16.7 20.0
1–6 16.9 17.6 15 17.7
7–13 5.8 6.11 5.6 4.6
14+ 7.6 7.16 8.1 9.4
Physical Activity (kcal/week)
median (IQR) 1080 (373.1, 2305.6) 1132.5 (405.0, 2376.9) 916.3 (307.5, 2160) 990 (292.5, 2265)
quartiles
<405 24.3 22.5 27.6 27.0
408.8–1110 25.0 24.8 25.6 24.5
1113.8–2415 25.3 26.5 23.1 23.9
2420–14805 25.4 26.2 23.7 24.5
Baseline Coronary Heart Disease 19.0 17.7 20.4 23.5
Baseline Hypertension 66.0 64.4 69.3 67.4
Baseline Diabetes 16.1 13.5 19.9 21.8
Open in a new tab
*

Includes prevalent cancer

Categorical variables are in percentages and continuous variables in mean (SD) or Median (IQR) SAS: Sleep Apnea Symptoms

Prevalence of sleep problems

The prevalence of SAS was 17% for daytime sleepiness, 9% for observed apnea, and 24% for snoring; 63% reported none of the three SAS. The prevalence of IS was 21.9% for difficulty falling asleep, 63% for problems staying asleep, and 32% for early morning awakenings; 38% reported no IS (including daytime sleepiness).

Cancer characteristics

Overall, 885 first incident cancers and 804 cancer deaths were identified over a median follow-up of 12 and 14 years, respectively. The median interval between SP data collection and cancer incidence ranged from 1.6 years for cervical cancer to 9 years for lymphatic or hematopoietic cancer; for cancer mortality, follow-up ranged from 2.9 years for cervical to almost 13 years for uterine cancer death. The most common cancer sites were prostate and lung (13%), colorectal cancer (12%), lymphatic or hematopoietic cancers (9%), and breast cancer (8%). The most common cancer sites for cancer death were lung (17%), lymph and colorectal cancer (11%), prostate (8%), pancreas (6%), urinary bladder, stomach, and breast cancer (4%). Among those diagnosed with cancer during follow-up, 41% were diagnosed with early-stage disease. Conversely, among participants for whom cancer stage information was available (~50%) and who died during follow-up, the largest proportion were diagnosed at a late stage (18%) (Table 2)

Table 2:

Selected characteristics of incident cancers and cancer deaths in CHS

Cancer group
Incidence (n=885) Death (n=804)
Median time at risk (yrs.) n % Median time at risk (yrs.) n %
Overall 12.0 13.5
Common cancer sites
breast 4.9 73 8.3 11.2 29 3.6
buccal 7.0 12 1.4 11.1 11 1.4
cervix 1.6 1 0.1 2.9 1 0.1
colorectal 7.0 102 11.5 9.2 84 10.5
digest 5.6 8 0.9 9.5 10 1.2
esophagus 6.8 9 1.0 7.7 11 1.4
kidney 6.4 27 3.1 11.8 17 2.1
liver 5.6 7 0.8 5.9 8 1.0
lung 6.0 113 12.8 8.8 138 17.2
lymphatic or hematopoietic 9.1 78 8.8 11.2 87 10.8
melanoma 5.0 2 0.2 8.1 2 0.3
ovary 8.4 11 1.2 12.1 15 1.9
pancreas 4.9 28 3.2 7.9 47 5.9
prostate 3.4 114 12.9 8.8 66 8.2
stomach 7.7 20 2.3 8.9 22 2.7
bladder 4.8 25 2.8 12.2 30 3.7
uterus 8.6 25 2.8 12.8 16 2.0
Cancer stage
localized 5.9 360 40.7 11.1 106 13.2
regional 6.3 188 21.2 9.0 115 14.3
distant 7.6 174 19.7 8.6 144 17.9
unknown 6.1 76 8.6 7.2 48 6.0
Open in a new tab
*

Cancer stage does not sum up to 100% due to missing stage information

Association between SP and cancer incidence

Table 3 presents the adjusted HRs and 95% CIs [reported as HR (95% CIs)] for cancer incidence in relation to baseline, time-dependent, cumulative averages, and baseline cumulative number of SAS and insomnia symptoms respectively. Compared to participants who reported no SAS, the risk of incident cancer was inversely associated with snoring [baseline: 0.84 (0.71, 0.99), time-dependent: 0.76 (0.65, 0.89)]. There were no observed significant associations with cancer incidence for daytime sleepiness and apneas. The total number of baseline SAS showed an inverse association for increasing number of symptoms compared to those reporting no SAS [any symptom: 0.81 (0.69, 0.95); 2+ symptoms 0.77 (0.61, 0.97)]. We found no evidence of association for symptoms of insomnia, whether considered individually or cumulatively. Of the specific cancer sites evaluated, significantly elevated cancer incidence was noted for prostate cancer for time-dependent analyses of apnea [2.34 (1.32, 4.15)] and baseline snoring [1.69 (1.11, 2.57)]; the observed association with snoring was increased when modeling this exposure as cumulative average [2.17 (1.22, 3.86)]. For prostate cancer, we also found a doseresponse relationship for baseline cumulative SAS symptoms compared to those reporting no symptoms [any symptom: HR=1.30 (0.84, 2.01); 2+ symptoms HR=2.22 (1.30, 3.79)]. We also noted a significantly elevated HR for lymphatic or hematopoietic cancers [baseline daytime sleepiness: 1.81 (1.06, 3.08)]; with respect to the IS symptoms, a significant inverse relationship was noted for problem staying asleep (cumulative average 0.54 (0.31, 0.92)).

Table 3:

Association of SP with cancer incidence overall and with specific cancer sites in CHS, n=3930

Cancer overall (n=885) Breast (n=73) Colorectal (102) Lung (113) Lymphatic or hematopoietic (78) Prostate (114)
Sleep apnea symptoms (SAS): (Primary exposure) Daytime Sleepiness HR (95% CI) HR (95% CI) HR (95% CI) HR (95% CI) HR (95% CI) HR (95% CI)
Baseline 0.93 (0.77, 1.12) 0.99 (0.44, 1.88) 0.98 (0.58, 1.67) 1.38 (0.86, 2.19) 1.81 (1.06, 3.08)* 0.95 (0.55, 1.62)
Time dependent 0.84 (0.69, 1.01) 0.81 (0.44, 1.49) 0.85 (0.52, 1.38) 1.05 (0.67, 1.62) 1.21 (0.77, 1.92) 1.20 (0.78, 1.86)
Cumulative average 0.78 (0.59, 1.04) 0.77 (0.31, 1.90) 0.75 (0.36, 1.57) 1.33 (0.68, 2.58) 1.57 (0.80, 3.08) 0.99 (0.51, 1.93)
Stopped breathing (Apnea)
Baseline 0.80 (0.61, 1.05) 0.63 (0.15, 2.72) 0.64 (0.26, 1.61) 0.65 (0.28, 1.52) 0.74 (0.31, 1.73) 1.28 (0.71, 2.29)
Time dependent 0.98 (0.72, 1.36) 0.42 (0.12, 1.45) 0.25 (0.058, 1.06) 0.54 (0.20, 1.44) 0.71 (0.26, 1.94) 2.34 (1.32, 4.15)**
Cumulative average 0.92 (0.60, 1.42) 0.25 (0.027, 2.41) 0.37 (0.09, 1.46) 0.39 (0.08, 1.78) 0.49 (0.12, 1.93) 2.25 (1.03, 4.88)
Snoring
Baseline 0.84 (0.71, 0.99)* 1.06 (0.56, 2.01) 0.78 (0.45, 1.35) 0.50 (0.30, 0.83)* 1.06 (0.62, 1.80) 1.69 (1.11, 2.57)*
Time dependent 0.76 (0.65, 0.89)** 0.78 (0.48, 1.27) 1.03 (0.70, 1.52) 0.80 (0.55, 1.15) 0.75 (0.49, 1.13) 1.28 (0.83, 199)
Cumulative average 0.80 (0.62, 1.04) 0.57 (0.22, 1.44) 1.01 (0.052, 1.94) 0.56 (0.28, 1.11) 0.98 (0.49, 1.98) 2.17 (1.22,3.86)**
Combined baseline SAS (Reference: 0 symptoms)
1 0.81 (0.69, 0.95)** 0.99 (0.57, 1.71) 1.01 (0.66, 1.56) 0.81 (0.53, 1.25) 1.13 (0.69, 1.85) 1.30 (0.84, 2.01)
2+ 0.77 (0.61, 0.97)* 0.80 (0.23, 2.71) 0.46 (0.20, 1.09) 0.46 (0.22, 0.96)* 1.27 (0.63, 2.57) 2.22 (1.30, 3.79)**
Insomnia Symptoms (IS) Difficulty falling asleep
Baseline 0.92 (0.77, 1.10) 0.87 (0.49, 1.53) 0.81 (0.49, 1.33) 0.74 (0.44, 1.24) 1.06 (0.56, 2.02) 0.49 (0.27, 0.92)
Time dependent 0.90 (0.74, 1.10) 0.83 (0.27, 1.46) 0.74 (0.44, 1.26) 0.64 (0.37, 1.11) 1.40 (0.86, 2.25) 0.20 (0.093, 0.43)**
Cumulative average 0.92 (0.72, 1.19) 0.94 (0.45, 1.95) 0.78 (0.40, 1.50) 0.54 (0.27, 1.06) 1.21 (0.62, 2.37) 0.15 (0.064, 0.37)**
Problems staying asleep
Baseline 1.07 (0.92, 1.10) 1.00 (0.61, 1.64) 1.00 (0.65, 1.53) 1.35(0.90, 2.02) 0.81 (0.49, 1.33) 1.08 (0.72, 1.62)
Time dependent 1.07 (0.92, 1.24) 0.77 (0.48, 1.24) 0.85 (0.58, 1.25) 1.04 (0.73, 1.48) 0.77 (0.52, 1.13) 1.42 (0.97, 2.11)*
Cumulative average 0.98 (0.79, 1.20) 0.86 (0.46, 1.59) 0.85 (0.50, 1.44) 1.20 (0.71, 2.00) 0.54 (0.31, 0.92)* 1.31 (0.79, 2.17)
Early morning Awakenings
Baseline 1.11 (0.96, 1.29) 0.82 (0.48, 1.38) 1.42 (0.93, 2.16) 1.44 (0.96, 2.18) 1.17 (0.69, 1.97) 1.01 (0.65, 1.55)
Time dependent 0.92 (0.78, 1.08) 1.08 (0.67, 1.71) 1.25 (0.84, 1.86) 0.87 (0.57, 1.31) 0.84 (0.55, 1.28) 1.03 (0.70, 1.52)
Cumulative average 1.12 (0.89, 1.45) 0.94 (0.43, 2.04) 1.49 (0.78. 2.84) 1.19 (0.61, 2.32) 1.19 (0.58, 2.43) 1.40 (0.76, 2.56)
Combined baseline IS (Reference: 0 symptoms)
 1 1.06 (0.89, 1.26) 0.85 (0.49, 1.50) 0.96 (0.58, 1.61) 1.28 (0.78, 2.09) 1.03 (0.56, 1.88) 1.36 (0.86, 2.16)
 2 0.96 (0.78, 1.17) 0.51 (0.24, 1.06) 0.92 (0.52, 1.65) 1.12 (0.74, 1.96) 1.44 (0.77, 2.68) 0.80 (0.45, 1.44)
 3+ 1.19 (0.89,1.26) 0.82 (0.42, 1.62) 1.33 (0.74, 1.36) 1.72 (0.99, 3.00) 1.19 (0.57, 2.50) 0.85 (0.44, 1.61)
Open in a new tab
*

p-value < 0.05

**

p-value < 0.01

Adjusted for: gender, study phase, baseline age, smoking, body mass index, diabetes, physical activity levels, and alcohol consumption

Association between SP and cancer mortality

Table 4 presents the adjusted HRs and 95% CIs [reported as HR (95% CIs)] for cancer mortality in relation to baseline, time-dependent, cumulative averages, and baseline cumulative number of SAS and insomnia symptoms respectively. We found significant inverse relationship for cancer mortality with respect to snoring [time-dependent: 0.73 (0.62, 0.86); cumulative average: 0.67 (0.58, 0.97)) and baseline apnea (0.69 (0.51, 0.94)]. The total number of baseline SAS showed an inverse association for increasing number of symptoms compared to those reporting no SAS [any symptom: 0.90 (0.76, 1.06); 2+ symptoms 0.75 (0.58, 0.97)]. None of the IS was significantly associated with cancer mortality. We noted a significant inverse relationship between difficulty falling asleep and colorectal cancer death [baseline: 0.32 (0.15, 0.69), time dependent: 0.41 (0.17, 0.98) and cumulative average: 0.28 (0.09, 0.84)] and baseline snoring with lung cancer death (0.56 (0.35, 0.89)). We also found an inverse dose response relationship between lung cancer mortality and baseline cumulative SAS compared to those reporting no symptoms [any symptom: HR= 0.84 (0.56, 1.24); 2+ symptoms HR= 0.35 (0.16, 0.77)].

Table 4:

Association of SP with cancer death overall and site-specific in CHS, n=4580

Cancer Overall (n=804) Colorectal (n=84) Lung (138) Lymphatic or hematopoietic (n=87) Prostate (n=66)
Sleep apnea symptoms (SAS): (Primary exposure) HR (95% CI) HR (95% CI) HR (95% CI) HR (95% CI) HR (95% CI)
Daytime Sleepiness
Baseline 1.08 (0.89, 1.31) 1.37 (0.80, 2.35) 1.11 (0.70, 1.77) 1.69 (0.99, 2.85) 1.42 (0.77, 2.61)
Time dependent 1.01 (0.83, 1.23) 1.03 (0.57, 1.86) 1.40 (0.90, 2.18) 1.02 (0.57, 1.83 0.93 (0.45, 1.94)
Cumulative average 1.11 (0.82, 1.49) 1.29 (0.53, 3.13) 1.45 (0.71, 2.97) 2.08 (0.97, 4.49) 1.91 (0.71, 5.13)
Stopped breathing (Apnea)
Baseline 0.69 (0.51, 0.94)* 0.87 (0.34, 2.20) 0.58 (0.25, 1.36) 0.61 (0.24, 1.52) 0.71 (0.27, 1.86)
Time dependent 0.99 (0.70, 1.41) 1.05 (0.37, 2.96) 0.64 (0.23, 1.78) 0.74 (0.23, 2.39) 1.93 (0.67, 5.58)
Cumulative average 0.94 (0.58, 1.55) 1.22 (0.29, 5.18) 0.36 (0.07, 2.02) 0.56 (0.11, 2.79) 1.68 (0.35, 8.10)
Snoring
Baseline 0.86 (0.72, 1.03) 0.71 (0.34, 1.20) 0.56 (0.35, 0.89)* 0.88 (0.51, 1.49) 1.50 (0.84, 2.69)
Time dependent 0.73 (0.62, 0.86)** 0.88 (0.53, 1.46) 0.76 (0.51, 1.12) 0.66 (0.41, 1.09) 0.83 (0.45, 1.53)
Cumulative average 0.67 (0.50, 0.90)** 0.72 (0.29, 1.83) 0.48 (0.22, 1.02) 0.57 (0.24, 1.38) 1.43 (0.51, 4.09)
Combined SAS (Reference: 0 symptoms)
 1 0.90 (0.76, 1.06) 0.82 (0.49, 1.37) 0.84 (0.56, 1.24) 1.04 (0.64, 1.70) 1.56 (0.90, 2.69)
 2+ 0.75 (0.58, 0.97)* 0.89 (0.43, 1.83) 0.35 (0.16, 0.77)** 0.93 (0.46, 1.89) 1.46 (0.64, 3.30)
Insomnia Symptoms Difficulty falling asleep
Baseline 0.90 (0.74, 1.08) 0.32 (0.15, 0.69)* 0.80 (0.49, 1.30) 0.91 (0.50, 1.65) 0.63 (0.32, 1.23)
Time dependent 0.89 (0.72, 1.10) 0.41 (0.17, 0.98)* 0.69 (0.40, 1.21) 1.04 (0.56, 1.94) 0.44 (0.17, 1.12)
Cumulative average 0.97 (0.74, 1.28) 0.28 (0.09, 0.84)* 0.70 (0.35, 1.39) 0.90 (0.39, 2.09) 0.46 (0.15, 1.41)
Problems staying Asleep
Baseline 1.02 (0.88, 1.20) 1.13 (0.69, 1.83) 1.18 (0.82, 1.69) 1.05 (0.65, 1.68) 0.97 (0.58, 1.62)
Time dependent 1.01 (0.86, 1.20) 1.01 (0.61, 1.67) 0.97 (0.65, 1.42) 1.06 (0.64, 1.74) 1.45 (0.76, 2.79)
Cumulative average 0.89 (0.71, 1.11) 1.03 (0.51, 2.07) 1.16 (0.66, 2.03) 0.73 (0.38, 1.39) 1.25 (0.52, 2.99)
Early morning Awakenings
Baseline 1.13 (0.96, 1.33) 1.47 (0.69, 1.82) 1.08 (0.72, 1.62) 0.88 (0.54, 1.42) 1.09 (0.65, 1.83)
Time dependent 1.00 (0.85, 1.19) 0.93 (0.54, 1.59) 1.03 (0.68, 1.56) 1.26 (0.77, 2.06) 0.95 (0.50, 1.80)
Cumulative average 1.23 (0.17, 1.64) 1.32 (0.53, 3.30) 1.25 (0.61, 2.61) 1.60 (0.70, 3.68) 1.20 (0.40, 3.67)
Combined IS (Reference: 0 symptoms)
 1 1.05 (0.88, 1.26) 1.25 (0.72, 2.19) 1.08 (0.70, 1.66) 1.39 (0.80, 2.42) 0.64 (0.34, 1.19)
 2 0.97 (0.79, 1.20) 1.34 (0.73, 2.48) 0.89 (0.54, 1.46) 1.12 (0.59, 2.14) 0.66 (0.33, 1.32)
 3+ 1.14 (0.91, 1.42) 0.75 (0.34, 1.70) 1.29 (0.77, 2.13) 1.26 (0.61, 2.59) 0.76 (0.36, 1.59)
Open in a new tab
*

p-value < 0.05

**

p-value < 0.01

Adjusted for: gender, study phase, baseline age, smoking, body mass index, diabetes, physical activity levels, and alcohol consumption

Sensitivity analysis

The overall patterns observed in the main results did not differ for cancer incidence overall and cancer mortality when we imposed 2- and 5-years outcome censoring and stratified by sex (Supplemental Table 1), when we stratified by BMI and cancer stage (Supplemental Table 2) or when we used a modified STOP-BANG measure instead of combined SAS (Supplemental Table 3).

Discussion

Several notable positive and inverse relationships between SP and cancer-site specific incidence and cancer-specific mortality emerged in the present study, even if many of the correlations examined did not achieve statistical significance. We found an inverse association between cancer incidence overall and snoring, and a null association with daytime sleepiness and apneas. In addition, we found an inverse dose response with combined SAS and no evidence of association for symptoms of insomnia. Of all the site-specific cancer incidence considered, we noted an association of snoring, apnea, and a dose-response for cumulative SAS with elevated prostate cancer incidence. We also found an association between daytime sleepiness and elevated lymphatic or hematopoietic cancers incidence, and an inverse relationship with problem staying asleep for these cancers. With respect to cancer mortality, we found an inverse relationship with snoring, apnea, and combined SAS. Of the cancer sites considered for site-specific mortality, colorectal cancer and lung cancer were inversely associated with difficulty falling asleep and snoring, respectively. We also found an inverse dose-response relationship between lung cancer mortality and baseline cumulative SAS compared to those reporting no symptoms.

Not all cancers are equally susceptible to the physiologic effects of SP, as noted in the heterogeneity across the cancer sites in our present study, and several other studies.27,44–46 Furthermore, even same organ cancers may respond divergently to SP, as illustrated by Marhuenda et al in their study of several lung cancer lines exposed to a model of sleep apnea.47 Worthy of note, is the elevated prostate cancer incidence with respect to snoring, apnea, and cumulative SAS. Given the age range of the study population and the fact that about 75% of prostate cancer are diagnosed in men who are 65 or older48 it may be suggestive of SP impacting increased incidence of prostate cancer in this population. This finding is inconsistent with previous observational studies of SP and prostate cancers.49–51 For instance, Markt et al. found no association between SP and prostate cancer incidence in 12,976 men with 785 cases of incident prostate cancer over 13 years.50 A more recent study by Tan et al. in 2,322 Swedish men aged 50 years or older followed for 40 years also found no significant association between SP and prostate cancer incidence.52 Furthermore, a recent meta-analysis consisting of 18 large studies found no association between rotating or night-shift work and prostate cancer.53 Conversely, Chung, et al. revealed a 35% higher risk of prostate cancer in patients aged ≥65 years with sleep disorders compared to their non-sleep disorder counterparts.54 These studies illustrate the inconsistency to date in the findings in the context of associations between sleep problems and cancer.

The inverse and null relationships identified herein are inconsistent with the previously reported adverse effects of SAS on cancer development and progression, even if some of those adverse effects have not been consistently detected across the studies.55 It is also noteworthy that numerous mechanisms may be operationally involved in the effect of SP on cancer development and progression, particularly given the many downstream implications of disruptions in the 24-hour circadian rhythm.56,57 Animal and in vitro studies provide evidence for the potential role of the circadian system in tumor suppression through cell cycle freeze, inhibition of cellular proliferation, promotion of apoptosis, and anti-angiogenesis.12–14 Another specific aspect of SP, apnea (i.e., recurrent cessations of breathing during sleep and a symptom of OSA), has been associated with accelerated tumor progression and metastatic potential through enhanced oncogenic pathways in the presence of intermittent hypoxia.25,58–61 Additionally, human studies have provided suggestive evidence for an association between OSA and cancer incidence24,27,44 and cancer mortality in adults.28,29 With respect to insomnia symptoms, a recent meta-analysis of eight studies suggested an overall increased risk of cancer incidence for individuals with insomnia symptoms in comparison to those without insomnia symptoms.30 Consistent with our study findings, another meta-analysis consisting of 6 studies found a null association between insomnia symptoms and cancer mortality.62

Furthermore, in considering the differences between the a priori anticipated effects of SP and cancer incidence and cancer mortality and current findings, we should remark that the population under study is ≥65 years, whereas some studies have suggested a significant association could be mostly restricted to younger people.63,64 For instance, Christensen et al. found no evidence of the relationship between SAS and incidence of cancer overall, but observed an elevated risk in persons younger than 50 years.63 Moreover, in a murine model of SP and lung cancer, old age was protective rather than deleterious, suggesting that the type of cancer and age may impact the nature of epidemiologic associations with SP.65 These assumptions are consistent with the inverse association between SP and lung cancer and colorectal cancer mortality, respectively, as observed in our study. Therefore, the mostly null associations in this older cohort do not necessarily rule out the presence of significant associations in younger age groups.66

Our study examining the association between SP and cancer incidence and cancer mortality in the CHS cohort should be interpreted in the context of some key study limitations.

First, some cancer diagnoses may have been missed if participants relocated out of state; however, this is less of a concern than in younger study populations given that older individuals such as those in our study population will be less likely to relocate. Furthermore, any migration would be expected to be non-differential with respect to SP and cancer.

In addition, we did not have data on other potential confounding variables such as occupation (e.g., shiftwork) which could both impact SP 67 and cancer incidence and cancer mortality risk. 68,69 However, given the older age of this cohort, a low prevalence of shift work would be expected. Also, we did not adjust for additional comorbidities including chronic obstructive pulmonary disease (COPD) or chronic kidney disease (CKD) which might contribute to nocturnal hypoxemia and sleep disordered breathing70,71 and cancer incidence and mortality.72,73 However, our a priori selection process (using a directed acyclic graphs74) focused on the minimum set of measured confounders sufficient to adjust for the associations between SP and cancer incidence and mortality.

Furthermore, the physiologic insults of sleep apnea could not only vary across cancer sites, but also molecular subtypes of a given cancer site. As mentioned above, Marhuenda et. al noted that tumor cell growth varied according to the presence of a representative oncogenic mutation on different cell lines of the most prevalent histological subtypes of non–small cell lung cancer (adenocarcinoma and squamous cell carcinoma) in response to intermittent hypoxia (IH) mimicking OSA.47 In particular, they found significant differences in HIF-1α activation in two of the four histologic lung cancer cell lines that were exposed to IH compared with normoxia.69,7047 As such, while colorectal cancer and lung cancer-specific mortality were inversely associated with difficulty falling asleep and combined SAS respectively in this present study, our findings could have been impacted by the heterogeneous distribution of histologic cancer cell types that are less susceptible to the impact of SP. However, the lack of histologically defined cancer sites precludes us from teasing out the potential heterogeneous relationship across molecular subtypes of a given cancer site. Second, different cancers have different treatments, and given that data on treatment are unavailable, we could not account for this factor in our analysis of cancer mortality. However, analysis of cancer mortality stratified by cancer stage did not differ from the overall results, which suggests that treatment is unlikely to fully explain the results, since we would expect late-stage cancers to be less impacted by treatment. Moreover, the lack of this treatment information could also have biased our estimates to the null.

In addition, we do not have information on obstructive sleep apnea or insomnia prevalence in this group which could also bias the results to the null. Third, sleep patterns change over time and with age, which remains a challenging exposure to capture. For instance, some of the sleep problems among participants had resolved, but with no documented reasons for such resolution. In fact, the SP is self-report, and is subject to misclassification. However, given that the measures were validated in a subset of the participants, misclassification is unlikely to significantly impact the results. Additionally, given that 70% of the cohort is partnered, SAS such as snoring, apnea should be viewed as very reliable symptoms. Moreover, when we restricted the analysis to the 70% of the partnered cohort, the estimates did not change by much. Furthermore, we did not have information on sleep duration, which is a key sleep quality metric that could vary substantially across participants with SAS or insomnia symptoms, while also pointing out that an individual with short sleep duration may not have a SAS or insomnia symptoms.75,76 Individuals with both a SAS and short sleep duration might be experiencing a more severe underlying SP, and therefore subsequent worse cancer outcomes.22 Despite these limitations, the reported SP were pre-diagnostic (which reduces the likelihood that our results were impacted by reverse causality), and we were able to assess the impact of baseline and longitudinal SP on cancer incidence and mortality which is lacking in the current literature.

A key strength of our study is the relatively large sample size, which allowed us to evaluate cancer incidence and cancer mortality, both overall and site-specific. Furthermore, the study design ensured temporality between the SP and the study endpoints with several covariates (including demographic and lifestyle factors) which allowed us to adjust for confounders selected a priori. Furthermore, the cancer incidence data was ascertained through a cancer registry (the gold standard in cancer research) while the cancer mortality data was also thoroughly assessed through half-yearly adjudication leveraging hospital diagnostic and procedure codes with associated text fields, death certificate diagnoses, and other information on the study criteria for underlying cause-of-death categories.

Conclusions

In conclusion, our study showed mixed results for the association between SP modeled at baseline and longitudinally and cancer incidence and mortality. We noted an inverse association between SP and lung and colorectal cancer mortality. Our results also showed an elevated risk of prostate cancer with respect to SAS. Given the high prevalence in men 65 years and older, and the ease with which SAS can be ascertained and managed, this finding may suggest an opportunity for reducing the burden of prostate cancer in older men if the association is causal. Future larger community based prospective studies addressing more cancer sites, molecular type-specific associations, and improved SP self-report documentation over time are needed.

Supplementary Material

Supp.Materials

Highlights.

  • Sleep apnea symptoms (SAS) was associated with increased prostate cancer incidence.

  • Daytime sleepiness was associated with lymphatic or hematopoietic cancers incidence.

  • Colorectal cancer and lung cancer were inversely associated with snoring.

  • Cumulative SAS was inversely associated with lung cancer mortality.

  • These associations in an older cohort may not translate to younger age groups.

Funding and Support:

This research was supported by contracts HHSN268201200036C, HHSN268200800007C, HHSN268201800001C, N01HC55222, N01HC85079, N01HC85080, N01HC85081, N01HC85082, N01HC85083, N01HC85086, 75N92021D00006, and grants U01HL080295 and U01HL130114 from the National Heart, Lung, and Blood Institute (NHLBI), with additional contribution from the National Institute of Neurological Disorders and Stroke (NINDS). Additional support was provided by R01AG023629 from the National Institute on Aging (NIA). A full list of principal CHS investigators and institutions can be found at CHS-NHLBI.org. Additionally Funding from NIH T32CA094880

Footnotes

Disclaimer:

The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Conflict of Interest: NONE

References

  • 1.Control C for D, Prevention (CDC). Unhealthy sleep-related behaviors–12 States, 2009. MMWRMorbidity Mortal Wkly Rep 2011;60(8):233–238. doi:mm6008a2 [pii] [PubMed] [Google Scholar]
  • 2.Young T, Palta M, Dempsey J, Peppard PE, Nieto FJ, Hla KM. Burden of sleep apnea: rationale, design, and major findings of the Wisconsin Sleep Cohort study. WMJ Off Publ State Med Soc Wis 2009;108(5):246–249. [PMC free article] [PubMed] [Google Scholar]
  • 3.Krystal AD. The effect of insomnia definitions, terminology, and classifications on clinical practice. J Am Geriatr Soc 2005;53(S7):S258–S263. [DOI] [PubMed] [Google Scholar]
  • 4.McArthur D Incidence of Hypertension and Type 2 Diabetes Among Obstructive Sleep Apnea Patients Published online 2016.
  • 5.Hla KM, Young T, Hagen EW, et al. Coronary heart disease incidence in sleep disordered breathing: the Wisconsin Sleep Cohort Study. Sleep 2015;38(5):677–684., doi: 10.5665/sleep.4654 [doi] [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Grandner MA, Hale L, Moore M, Patel NP. Mortality associated with short sleep duration: the evidence, the possible mechanisms, and the future. Sleep Med Rev 2010;14(3):191–203. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Tarasiuk A, Reuveni H. The economic impact of obstructive sleep apnea. Curr Opin Pulm Med 2013;19(6):639–644. doi: 10.1097/MCP.0b013e3283659e1e [doi] [DOI] [PubMed] [Google Scholar]
  • 8.Hägg SA. Role of sleep disturbances in occupational accidents among women. Scand J Work Environ Health 2015;41(4):368. [DOI] [PubMed] [Google Scholar]
  • 9.Ancoli‐Israel S, Cooke JR. Prevalence and comorbidity of insomnia and effect on functioning in elderly populations. J Am Geriatr Soc 2005;53(S7):S264–S271. [DOI] [PubMed] [Google Scholar]
  • 10.Lange T, Dimitrov S, Born J. Effects of sleep and circadian rhythm on the human immune system. Ann N Y Acad Sci 2010;1193(1):48–59. [DOI] [PubMed] [Google Scholar]
  • 11.James SM, Honn KA, Gaddameedhi S, Van Dongen HP. Shift work: disrupted circadian rhythms and sleepimplications for health and well-being. Curr Sleep Med Rep 2017;3(2):104–112. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Kelleher FC, Rao A, Maguire A. Circadian molecular clocks and cancer. Cancer Lett 2014;342(1):9–18. doi: 10.1016/j.canlet.2013.09.040 [doi] [DOI] [PubMed] [Google Scholar]
  • 13.Filipski E, Delaunay F, King VM, et al. Effects of chronic jet lag on tumor progression in mice. Cancer Res 2004;64(21):7879–7885. doi:64/21/7879 [pii] [DOI] [PubMed] [Google Scholar]
  • 14.Sainz RM, Mayo JC, Tan D, León J, Manchester L, Reiter RJ. Melatonin reduces prostate cancer cell growth leading to neuroendocrine differentiation via a receptor and PKA independent mechanism. The Prostate 2005;63(1):29–43. [DOI] [PubMed] [Google Scholar]
  • 15.Straif K, Baan R, Grosse Y, et al. Carcinog Shift-Work Paint Fire-Fight Published online 2007.
  • 16.Stellman SD. Int Agency Res Cancer IARC Monogr Eval Carcinog Risk Chem HumansSaccharin Rep DrMorris FCranmer Vol22 Some Non-Nutr Sweeten Agents 208 Pages No Illus Published online 1981.
  • 17.Fritschi L Shift work and cancer. BMJ 2009;339:b2653. doi: 10.1136/bmj.b2653 [doi] [DOI] [PubMed] [Google Scholar]
  • 18.Hurley S, Goldberg D, Bernstein L, Reynolds P. Sleep duration and cancer risk in women. Cancer Causes Control 2015;26(7):1037–1045. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Ruesten AV, Weikert C, Fietze I, Boeing H. Association of sleep duration with chronic diseases in the European Prospective Investigation into Cancer and Nutrition (EPIC)-Potsdam study. PloS One 2012;7(1):e30972. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Gu F, Xiao Q, Chu LW, et al. Sleep duration and cancer in the NIH-AARP diet and health study cohort. PloS One 2016;11(9):e0161561. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Xiao Q, Signorello LB, Brinton LA, Cohen SS, Blot WJ, Matthews CE. Sleep duration and breast cancer risk among black and white women. Sleep Med 2016;20:25–29. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Phipps AI, Bhatti P, Neuhouser ML, et al. Pre-diagnostic Sleep Duration and Sleep Quality in Relation to Subsequent Cancer Survival. J Clin Sleep Med JCSM Off Publ Am Acad Sleep Med 2016;12(4):495–503. doi: 10.5664/jcsm.5674 [doi] [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Thompson CL, Li L. Association of sleep duration and breast cancer OncotypeDX recurrence score. Breast Cancer Res Treat 2012;134(3):1291–1295. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Martínez-García MÁ, Campos-Rodriguez F, Barbé F. Cancer and OSA: current evidence from human studies. Chest 2016;150(2):451–463. [DOI] [PubMed] [Google Scholar]
  • 25.Perini S, Martinez D, Montanari CC, Fiori CZ. Enhanced expression of melanoma progression markers in mouse model of sleep apnea. Rev Port Pneumol Engl Ed 2016;22(4):209–213. [DOI] [PubMed] [Google Scholar]
  • 26.Gozal D, Farré R, Nieto FJ. Putative links between sleep apnea and cancer: from hypotheses to evolving evidence. CHEST J 2015;148(5):1140–1147. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Gozal D, Ham SA, Mokhlesi B. Sleep Apnea and Cancer: Analysis of a Nationwide Population Sample. Sleep 2016;39(8):1493–1500. doi: 10.5665/sleep.6004 [doi] [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Marshall NS, Wong KK, Cullen SR, Knuiman MW, Grunstein RR. Sleep apnea and 20-year follow-up for all-cause mortality, stroke, and cancer incidence and mortality in the Busselton Health Study cohort. J Clin Sleep Med 2014;10(4):355–362. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Nieto FJ, Peppard PE, Young T, Finn L, Hla KM, Farré R. Sleep-disordered breathing and cancer mortality: results from the Wisconsin Sleep Cohort Study. Am J Respir Crit Care Med 2012;186(2):190–194. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Shi T, Min M, Sun C, Zhang Y, Liang M, Sun Y. Does insomnia predict a high risk of cancer? A systematic review and meta‐analysis of cohort studies. J Sleep Res 2020;29(1):e12876. [DOI] [PubMed] [Google Scholar]
  • 31.Fried LP, Borhani NO, Enright P, et al. The cardiovascular health study: design and rationale. Ann Epidemiol 1991;1(3):263–276. [DOI] [PubMed] [Google Scholar]
  • 32.Quan SF, Howard BV, Iber C, et al. The sleep heart health study: design, rationale, and methods. Sleep 1997;20(12):1077–1085. [PubMed] [Google Scholar]
  • 33.Strand LB, Carnethon M, Biggs ML, et al. Sleep Disturbances and Glucose Metabolism in Older Adults: The Cardiovascular Health Study. Diabetes Care 2015;38(11):2050–2058. doi: 10.2337/dc15-0137 [doi] [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Geffken DF, Cushman M, Burke GL, Polak JF, Sakkinen PA, Tracy RP. Association between physical activity and markers of inflammation in a healthy elderly population. Am J Epidemiol 2001;153(3):242–250. [DOI] [PubMed] [Google Scholar]
  • 35.Mittelmark M, Psaty BM, Rautaharju PM, et al. Prevalence of cardiovascular diseases among older adults: the Cardiovascular Health Study. Am J Epidemiol 1993;137(3):311–317. [DOI] [PubMed] [Google Scholar]
  • 36.Greenlee H, Unger JM, LeBlanc M, Ramsey S, Hershman DL. Association between Body Mass Index and Cancer Survival in a Pooled Analysis of 22 Clinical Trials. Cancer Epidemiol Biomark Prev Publ Am Assoc Cancer Res Cosponsored Am Soc Prev Oncol 2017;26(1):21–29. doi: 10.1158/1055-9965.EPI-15-1336 [doi] [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Sieh W, Edwards KL, Fitzpatrick AL, et al. Genetic susceptibility to prostate cancer: prostate-specific antigen and its interaction with the androgen receptor (United States). Cancer Causes Control 2006;17(2):187–197. [DOI] [PubMed] [Google Scholar]
  • 38.Psaty BM, Kuller LH, Bild D, et al. Methods of assessing prevalent cardiovascular disease in the Cardiovascular Health Study. Ann Epidemiol 1995;5(4):270–277. [DOI] [PubMed] [Google Scholar]
  • 39.Ives DG, Fitzpatrick AL, Bild DE, et al. Surveillance and ascertainment of cardiovascular events: the Cardiovascular Health Study. Ann Epidemiol 1995;5(4):278–285. [DOI] [PubMed] [Google Scholar]
  • 40.StataCorp. Stata Statistical Software: Release 14 College Station, TX: StataCorp LP. Published online 2015. [Google Scholar]
  • 41.Almendros I, Martinez-Garcia MA, Farré R, Gozal D. Obesity, sleep apnea, and cancer. Int J Obes Published online 2020:1–15. [DOI] [PubMed]
  • 42.Chung F, Abdullah H, Liao P. STOP BANG Questionnaire. Chest 2016;149(4):631–638. [DOI] [PubMed] [Google Scholar]
  • 43.Therneau TM, Grambsch PM. The Cox model. In: Modeling survival data: extending the Cox model Springer; 2000:39–77. [Google Scholar]
  • 44.Sillah A, Watson NF, Schwartz SM, Gozal D, Phipps AI. Sleep apnea and subsequent cancer incidence. Cancer Causes Control Published online 2018:1–8. [DOI] [PMC free article] [PubMed]
  • 45.Fang H-F, Miao N-F, Chen C-D, Sithole T, Chung M-H. Risk of Cancer in Patients with Insomnia, Parasomnia, and Obstructive Sleep Apnea: A Nationwide Nested Case-Control Study. J Cancer 2015;6(11):1140. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46.Sillah A, Watson NF, Gozal D, Phipps AI. Obstructive sleep apnea severity and subsequent risk for cancer incidence. Prev Med Rep 2019;15:100886. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47.Marhuenda E, Campillo N, Gabasa M, et al. Effects of Sustained and Intermittent Hypoxia on Human Lung Cancer Cells. Am J Respir Cell Mol Biol 2019;61(4):540–544. [DOI] [PubMed] [Google Scholar]
  • 48.Quinn M, Babb P. Patterns and trends in prostate cancer incidence, survival, prevalence and mortality. Part I: international comparisons. BJU Int 2002;90(2):162–173. [DOI] [PubMed] [Google Scholar]
  • 49.Dickerman BA, Markt SC, Koskenvuo M, et al. Sleep disruption, chronotype, shift work, and prostate cancer risk and mortality: a 30-year prospective cohort study of Finnish twins. Cancer Causes Control 2016;27(11):1361–1370. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50.Markt SC, Flynn-Evans EE, Valdimarsdottir UA, et al. Sleep Duration and Disruption and Prostate Cancer Risk: a 23-Year Prospective Study. Cancer Epidemiol Biomark Prev Publ Am Assoc Cancer Res Cosponsored Am Soc Prev Oncol 2016;25(2):302–308. doi: 10.1158/1055-9965.EPI-14-1274 [doi] [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 51.Demark-Wahnefried W, Platz EA, Ligibel JA, et al. The role of obesity in cancer survival and recurrence. Cancer Epidemiol Biomark Prev Publ Am Assoc Cancer Res Cosponsored Am Soc Prev Oncol 2012;21(8):1244–1259. doi: 10.1158/1055-9965.EPI-12-0485 [doi] [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 52.Tan X, Cedernaes J, Forsberg LA, Schiöth HB, Benedict C. Self‐reported sleep disturbances and prostate cancer morbidity and mortality in Swedish men: A longitudinal study over 40 years. J Sleep Res 2018;27(6):e12708. [DOI] [PubMed] [Google Scholar]
  • 53.Rivera-Izquierdo M, Martínez-Ruiz V, Castillo-Ruiz EM, Manzaneda-Navío M, Pérez-Gómez B, Jiménez-Moleón JJ. Shift Work and Prostate Cancer: An Updated Systematic Review and Meta-Analysis. Int J Environ Res Public Health 2020;17(4):1345. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 54.Chung W-S, Lin C-L. Sleep disorders associated with risk of prostate cancer: a population-based cohort study. BMC Cancer 2019;19(1):146. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 55.Gozal D, Almendros I, Phipps AI, Campos-Rodriguez F, Martínez-García MA, Farré R. Sleep Apnoea Adverse Effects on Cancer: True, False, or Too Many Confounders? Int J Mol Sci 2020;21(22):8779. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 56.Cherrie JW, Crawford JO, Davis A, et al. A review of the impact of shift-work on cancer: summary of the evidence for practitioners. Policy Pract Health Saf Published online 2017:1–7.
  • 57.DiTacchio L, DiTacchio KA, Panda S. Relevance of Circadian Rhythm in Cancer. In: Murine Models, Energy Balance, and Cancer Springer; 2015:1–19. [Google Scholar]
  • 58.Li L, Ren F, Cao J, Chen B. Relevant Mechanism of Intermittent Hypoxia-Induced Melanoma Lung Metastases in a Murine Model of Sleep Apnea. CHEST J 2016;149(4_S):A556–A556. [Google Scholar]
  • 59.Martinez-Garcia MA, Campos-Rodriguez F, Nagore E, et al. Sleep-Disordered Breathing is Independently Associated with Increased Aggressiveness of Cutaneous Melanoma. A multicentre observational study in 443 patients. Chest Published online 2018. [DOI] [PubMed]
  • 60.Toth K, Chintala S, Rustum YM. Constitutive expression of HIF-alpha plays a major role in generation of clear-cell phenotype in human primary and metastatic renal carcinoma. Appl Immunohistochem Mol Morphol AIMM 2014;22(9):642–647. doi: 10.1097/PAI.0000000000000012 [doi] [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 61.Vilaseca A, Campillo N, Torres M, et al. Intermittent hypoxia increases kidney tumor vascularization in a murine model of sleep apnea. PloS One 2017;12(6):e0179444. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 62.Ge L, Guyatt G, Tian J, et al. Insomnia and risk of mortality from all-cause, cardiovascular disease, and cancer: systematic review and meta-analysis of prospective cohort studies. Sleep Med Rev 2019;48:101215. [DOI] [PubMed] [Google Scholar]
  • 63.Christensen AS, Clark A, Salo P, et al. Symptoms of sleep disordered breathing and risk of cancer: a prospective cohort study. Sleep 2013;36(10):1429–1435. doi: 10.5665/sleep.3030 [doi] [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 64.Campos-Rodriguez F, Martinez-Garcia MA, Martinez M, et al. Association between obstructive sleep apnea and cancer incidence in a large multicenter Spanish cohort. Am J Respir Crit Care Med 2013;187(1):99–105. [DOI] [PubMed] [Google Scholar]
  • 65.Torres M, Campillo N, Nonaka PN, et al. Aging Reduces Intermittent Hypoxia-induced Lung Carcinoma Growth in a Mouse Model of Sleep Apnea. Am J Respir Crit Care Med 2018;(ja). [DOI] [PMC free article] [PubMed]
  • 66.Martinez-Garcia MA, Gozal D. Osa and Cpap Treament in the Very Elderly: The Challenge of the Unknown. Sleep Published online 2021. [DOI] [PubMed]
  • 67.Paciorek M, Korczyoski P, Bielicki P, Byśkiniewicz K, Zielioski J, Chazan R. Obstructive sleep apnea in shift workers. Sleep Med 2011;12(3):274–277. [DOI] [PubMed] [Google Scholar]
  • 68.Su F, Huang D, Wang H, Yang Z. Associations of shift work and night work with risk of all-cause, cardiovascular and cancer mortality: a meta-analysis of cohort studies. Sleep Med Published online 2021. [DOI] [PubMed]
  • 69.Purdue MP, Hutchings SJ, Rushton L, Silverman DT. The proportion of cancer attributable to occupational exposures. Ann Epidemiol 2015;25(3):188–192. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 70.Maung SC, El Sara A, Chapman C, Cohen D, Cukor D. Sleep disorders and chronic kidney disease. World J Nephrol 2016;5(3):224. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 71.Omachi TA, Blanc PD, Claman DM, et al. Disturbed sleep among COPD patients is longitudinally associated with mortality and adverse COPD outcomes. Sleep Med 2012;13(5):476–483. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 72.Na SY, Sung JY, Chang JH, et al. Chronic kidney disease in cancer patients: an independent predictor of cancer-specific mortality. Am J Nephrol 2011;33(2):121–130. [DOI] [PubMed] [Google Scholar]
  • 73.Wong G, Hayen A, Chapman JR, et al. Association of CKD and cancer risk in older people. J Am Soc Nephrol 2009;20(6):1341–1350. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 74.Tennant PW, Murray EJ, Arnold KF, et al. Use of directed acyclic graphs (DAGs) to identify confounders in applied health research: review and recommendations. Int J Epidemiol 2021;50(2):620–632. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 75.Dongen HPV, Vitellaro KM, Dinges DF. Individual differences in adult human sleep and wakefulness: Leitmotif for a research agenda. Sleep 2005;28(4):479–498. [DOI] [PubMed] [Google Scholar]
  • 76.Kripke DF, Garfinkel L, Wingard DL, Klauber MR, Marler MR. Mortality associated with sleep duration and insomnia. Arch Gen Psychiatry 2002;59(2):131–136. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supp.Materials
NIHMS1753684-supplement-Supp_Materials.docx (20.6KB, docx)

RESOURCES