Diurnal cortisol rhythm as a predictor of lung cancer survival

Abstract

Background

Poorly coordinated diurnal cortisol and circadian rest-activity rhythms predict earlier mortality in metastatic breast and colorectal cancer, respectively. We examined the prognostic value of the diurnal cortisol rhythm in lung cancer.

Methods

Lung cancer patients (n = 62, 34 female) were within 5 years of diagnosis and had primarily non small-cell lung cancer, with disease stage ranging from early to advanced. Saliva collected over two days allowed calculation of the diurnal cortisol slope and the cortisol awakening response (CAR). Lymphocyte numbers and subsets were measured by flow cytometry. Survival data were obtained for 57 patients. Cox Proportional Hazards analyses were used to test the prognostic value of the diurnal cortisol rhythm on survival calculated both from study entry and from initial diagnosis.

Results

The diurnal cortisol slope predicted subsequent survival over three years. Early mortality occurred among patients with higher slopes, or relatively “flat” rhythms indicating lack of normal diurnal variation (Cox Proportional Hazards p = .009). Cortisol slope also predicted survival time from initial diagnosis (p = .012). Flattened profiles were linked with male gender (t = 2.04, df = 59, p = .046) and low total and cytotoxic T cell lymphocyte counts (r = −.39 and −.30, p = .004 and .035, respectively). After adjustment for possible confounding factors, diurnal slope remained a significant, independent predictor of survival.

Conclusions

Flattening of the diurnal cortisol rhythm predicts early lung cancer death. Data contribute to growing evidence that circadian disruption accelerates tumor progression.

Introduction

Among patients with advancing cancer, dysregulation of circadian physiology is often conspicuous and extensive. Disrupted circadian cycles are evident in endocrine, immune, metabolic, and cellular systems (Mazzoccoli et al., 2010, Hrushesky, 1985, Mormont et al., 2000). Dysregulation of circadian hypothalamic–pituitary-adrenal (HPA) rhythms has been linked specifically with advancing cancer and accelerated tumor growth rates (Mormont and Levi, 1997, Touitou et al., 1995, Hrushesky et al., 1998, Sephton and Spiegel, 2003, Eismann et al., 2010).

Human cortisol peaks 30–45 min after first awakening and drops to a nadir during sleep (Clow et al., 2010). From 30% to 70% of patients with advanced cancer display idiosyncratic rhythm abnormalities including unsynchronized peaks and troughs, consistently high or low levels and erratic circadian fluctuations. Viewed in aggregate, the diurnal cortisol profiles of advanced cancer patients often appear “flattened” (Mormont and Lévi, 1997; Sephton et al., 2000). The diurnal cortisol slope captures multiple types of rhythm deviation and is measurable in saliva (Turner-Cobb et al., 2000, Kraemer et al., 2006, Hellhammer et al., 2009). Elevated diurnal cortisol slope predicts early metastatic breast cancer mortality, independent of other prognostic factors (Sephton et al., 2000). Flattened slopes were associated with low natural killer cell counts and cytotoxicity, sleep disruption, and marital dissolution; none of which explained the cortisol-survival relationship (Sephton et al., 2000). Circadian rest/activity rhythms are also irregular in patients with advancing cancer (Mormont and Lévi, 1997), and are measurable by wrist-worn actigraphy (Ancoli-Israel et al., 2003). Poor circadian coordination of rest/activity rhythm predicts early mortality over four years in metastatic colorectal cancer patients (Mormont et al., 2000), a finding recently replicated in large multi-site study (Innominato et al., 2009). Among cancer patients, circadian disruption may be a preexisting cause, a correlate of distress and physiological burden, a direct effect of cancer on the brain, and/or a mediator of psychosocial effects on tumor progression (Sephton and Spiegel 2003; Eismann et al., 2010; Chida et al., 2008). A few clinical prospective studies have provided a fascinating demonstration of prognostic value for circadian rhythms in human cancers. These require replication and testing in other cancer types.

As compared with the comorbidities of other cancers, lung cancer-related distress, depression, fatigue, and sleep disruption are extreme (Parker et al., 2008). All these symptoms have been linked with circadian disruption in lung cancer, suggesting that this disease may convey uniquely high circadian disruption-mediated risk (Hrushesky et al., 2009). Lung cancer patients demonstrate markedly poor integration of circadian neuroendocrine-immune function evidenced by high evening cortisol, suppressed melatonin, and changes in the circadian patterns of distribution of peripheral lymphocytes (Mazzoccoli et al., 2003, Mazzoccoli et al., 2005). Non-small cell lung cancer generally has a better prognosis than small-cell disease; however, five-year survival rates for both are poor (49–75% for early stage, and < 1% for advanced disease).

We examined the prognostic value of the diurnal salivary cortisol rhythm in lung cancer. Secondary analyses explored associations with potential explanatory and/or confounding factors including traditional prognostic indicators, depression, fatigue, and sleep quality. Additional biological factors were chosen based on their relevance in cancer studies of the effects of the stress, sleep, and circadian rhythms (Cole and Sood, 2012, Antoni et al., 2006, Eismann et al., 2010, Mazzoccoli et al., 2003, Mazzoccoli et al., 2005). These included serum cortisol, sympathetic activation measured by overnight urinary catecholamines, lymphocyte counts and lymphocyte subsets.