Evening cortisol levels are prognostic for progression-free survival in a prospective pilot study of head and neck cancer patients

Abstract

Introduction: Cortisol rhythm disruptions predict early mortality in renal, colorectal, lung, and metastatic breast cancer. In head and neck cancer (HNC), various cortisol indices are known to correlate with adverse psychological and biological (e.g., inflammatory) outcomes, but links to mortality have yet to be demonstrated. We hypothesize that the prognostic value of diurnal cortisol aberrations will hold in HNC. Prior work leads us to predict that flattened or elevated diurnal cortisol profiles will be associated with elevations of serum inflammatory and tumor-promoting cytokines in this population, and that these immune markers would themselves predict poor progression-free survival.

Method: We prospectively recruited a pilot sample of HNC patients (N=40) at a multidisciplinary HNC clinic. Most patients presented with late-stage oral/oropharyngeal cancer, were older than 50, male, and subsequently received combined-modality (surgery and/or radiotherapy with or without chemotherapy) treatment with curative intent. Saliva was collected twice daily for six days to assess diurnal slope, mean, waking, and evening cortisol levels. Serum was assayed for an exploratory panel of inflammatory and tumor-promoting cytokines. Two years post study-entry, disease progression and survivorship status were abstracted from medical records. Bivariate correlations, linear regressions, and Cox Proportional Hazards models tested hypotheses.

Results: Elevations of evening cortisol and diurnal mean levels were each associated with shorter progression-free survival (evening: Hazard Ratio [HR]=1.848, 95% Confidence Interval [CI]=1.057-3.230, p=.031; diurnal mean: HR=2.662, 95% CI=1.115-6.355, p=.027). Bivariate correlations revealed that higher levels of the serum inflammatory marker interferon (IFN)-γ were linked with elevated evening (r=.405, p=.014) and mean (r=.459, p=.004) cortisol. Higher expression of IFN-γ also predicted poorer progression-free survival (HR=4.671, 95% CI=1.409-15.484, p=.012).

Discussion: Elevated evening and diurnal mean cortisol were both prognostic; suggesting cortisol secretion is both dysregulated and elevated among patients who subsequently experienced accelerated disease progression. These exploratory data from 40 HNC patients mirror relationships between cortisol and survival identified among patients with numerous other tumor types. This pilot study highlights the need for research on effects of cortisol rhythm disruption among HNC patients. Future research in larger samples should also examine the role of inflammatory and tumor-promoting factors-both systemically and within the tumor microenvironment-as potential mediators of cortisol rhythm disruption.

Keywords: circadian rhythm disruption; cortisol; head and neck cancer; interferon-gamma; progression-free survival.

Figure 1

Mean diurnal (A) and evening (B) cortisol were both significantly associated with progression-free survival. Data were split at the median and Kaplan-Meier curves generated for illustrative, not analytic, purposes.

Figure 2

Raw mean and 95% CI cortisol based on a median split of the diurnal cortisol slope within our sample. As noted, prior studies found diurnal cortisol slope prognostic for cancer mortality, a finding not replicated in this small (n=40) sample of HNC patients. Rather, we found that only the evening and overall mean cortisol levels predicted survival. To inform future research, we felt it important to describe interrelationships between cortisol summary variables. It’s worth noting that evening cortisol appears to distinguish steep from flattened slopes more distinctly than do waking values. The restricted CI for evening cortisol among those with steep, healthier cortisol is also interesting. This figure demonstrates the magnitude of the difference in evening cortisol level associated with differences in cortisol slope.