Cancer incidence, mortality, and immunotherapy outcomes in relation to sleep problems: Results from Cardiovascular Health Study and a cancer immunotherapy cohort

Background: Sleep problems (SP) affect a large proportion of adults: an estimated 50-70 million U.S. adults suffer from sleep problems. Among numerous potential health consequences, sleep problems may be adversely associated with cancer risk and cancer mortality. There has been evidence from human studies and numerous animal studies linking sleep problems to cancer development and proliferation in recent years. The mechanisms underlying such cancer-related effects likely reflect the adverse downstream consequences of disruptions to the 24-hour circadian rhythm or physiologic insults of sleep disorders such as sleep apnea. In light of the high prevalence of sleep problems in the population, the potential connection of these problems to cancer occurrence and mortality is a pressing public health concern. Beyond its impact on cancer risk, sleep disturbances are a prominent concern of cancer patients, with up to 80% reporting disturbed sleep. One particular concern in this patient population is the fact that disturbed and insufficient sleep adversely affects immune health. The presence of a strong T-cell response in cancer, indicating activation of the adaptive immune system, has been consistently associated with better patient outcomes. Existing and emerging immunotherapies are attempting to harness the T-cell response to treat several forms of cancer by targeting immune-suppressive proteins, such as programmed cell death-1 receptor (PD-1), its ligand PD-L, and cytotoxic T lymphocyte antigen 4 (CTLA-4) using immune checkpoint inhibitors (ICIs). Accordingly, disruptions to the circadian rhythm could plausibly have implications for the effect of these therapies. Therefore, we sought to quantify the associations of SP with cancer incidence, mortality, aggressiveness, and ICI outcomes in the context of the longitudinal Cardiovascular Health Study (CHS) and a pilot study of cancer patients initiating ICIs (The Lifestyle Attributes and Sleep in Immunotherapy Response (LASIR) Study). Methods: We assessed the association of self-reported SP with incident cancer (N=3930, excluding prevalent cancers) and cancer mortality (N=4580) among Cardiovascular Health Study (CHS) participants, a population-based study of adults aged >=65 years recruited from four US communities. Participants reported sleep apnea symptoms (SAS) and insomnia symptoms at each visit between 1989–1994. Cancer incidence was ascertained through linkage with state cancer registries through 2005; cancer-specific death was adjudicated through 2015. We used Cox proportional hazards regression to calculate hazard ratios (HR) and 95% confidence intervals (CI) for associations of baseline and longitudinal SP with subsequent cancer incidence and cancer mortality, adjusting for a priori selected confounders, including gender, study phase, age, smoking, body mass index, diabetes, physical activity levels, and alcohol consumption. For the LASIR cohort, 33 participants consented to the study, of whom 32 initiated an ICI treatment. We collected questionnaire data on primary SP (sleep apnea risk) and secondary SP, including insomnia and general sleep patterns. We extracted patient attributes on the date corresponding to the patient’s last visit at SCCA prior to or at ICI initiation from the Electronic health records (EHR) six months post-ICI initiation. We also extracted information on their cancer attributes at diagnosis, prior cancer treatments information, vital status, the type, dates, number of ICI infusions, and response assessment indicators. Our primary analyses included the associations of SP with tumor aggressiveness at diagnosis and ICI tolerance, indicating a likely favorable response to treatment. Tumor aggressiveness was defined by M-stage (M0 vs. M1). Six or more infusions was considered ICI tolerability. We ran a Poisson regression with robust standard errors to assess the association between SP and tumor aggressiveness, adjusting for age, gender, body mass index, and reported prevalence ratio (PR) and associated 95% CI for the associations. We used logistic regression models to assess the association between sleep apnea and insomnia risk with the number of ICI infusions received, adjusting for male gender, age at ICI, and prior cancer treatment. We calculated the odds ratio (OR) and 95% confidence intervals (CI) for associations with this outcome. Results: The mean age (SD) of the CHS study population was 73 years, 57% were female, and 83% were white. Overall, 885 first incident cancers and 804 cancer deaths were identified over a median follow-up of 12 and 14 years, respectively. Briefly, compared to participants who reported no SAS, the risk of incident cancer was inversely associated [(HR(95%CI)] with snoring [baseline: 0.84 (0.71, 0.99), time-dependent: 0.76 (0.65, 0.89)]. We noted an elevated cancer incidence for prostate cancer for time-dependent analyses of apnea [2.34 (1.32, 4.15)], baseline snoring [1.69 (1.11, 2.57)] and cumulative average snoring [2.17(1.22, 3.86)]. We found a significantly elevated HR for lymphatic or hematopoietic cancers [baseline snoring: 1.81 (1.06, 3.08)]. We also noted an inverse relationship for cancer mortality with respect to snoring [time-dependent: 0.73 (0.62, 0.86); cumulative average: (0.67 (0.50, 0.90)) and baseline apnea (0.69 (0.51, 0.94)]. We found 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)]. The mean age of the LASIR cohort was 61 years, 61% were male, 85% were white, 64% were partnered, 70% had a college degree, 79% had a BMI of at least 25 kg/m2, 3% were current smokers, 18% reported sufficient weekly physical activity (150 min/week at moderate equivalent) and 18% high stress. The prevalence of low, intermediate, and high-risk OSA risk was 36%, 42%, and 21%, respectively. Of the secondary SP considered, 58% of participants reported clinically significant insomnia, 72% experienced an average or restless night sleep, 30% reported taking longer than 15 min to fall asleep, 46% had ideal night sleep, and 36% reported an evening chronotype. We did not find a significant association [(PR(95%CI)] between intermediate or high-risk OSA and metastatic cancer compared to low-risk OSA [1.01 (0.28, 3.67)] Of the secondary sleep attributes considered, patients who reported taking more than 15 minutes to fall asleep were 3.6 times more likely to be diagnosed with metastatic cancer compared to those reporting shorter sleep latency [95% CI (1.74, 7.35)]. Additionally, patients reporting a morning chronotype were less likely to be diagnosed with metastatic cancer compared to those reporting evening chronotypes [0.23 (0.09, 0.58)]. Similarly, we did not find any significant association [(OR (95%CI)] between intermediate or high-risk OSA and six or more infusions compared to low-risk OSA [0.27 (0.02, 3.41)]. Similarly, we found no significant association between insomnia and six or more infusions [0.23 (0.03, 1.60)]. Conclusion: Despite the mostly null results in the CHS study, there were a few notable elevated and inverse relationships between SP and cancer-site specific incidence and cancer mortality. The results add to the growing evidence suggesting the physiologic effect of sleep problems on cancer is heterogeneous across cancer sites. Therefore, future larger community based prospective studies addressing more cancer site and molecular type-specific associations and improved SP self-report documentation over time are needed. Additionally, the immunotherapy cohort gives insights into the potential burden and impact of SP on tumor aggressiveness and ICI treatment response. The results could inform larger-scale observational studies with an ultimate goal of informing clinical trials focused on finding effective sleep quality improvement interventions in ICI cancer treatment populations.