Associations between salivary cortisol and sociodemographic, sampling and health factors

Background Cortisol levels are increasingly often assessed in large-scale psychosomatic research. Although determinants of different salivary cortisol indicators have been described, they have not yet been systematically studied within the same study with a large sample size. Sociodemographic, health and sampling-related determinants of salivary cortisol levels were examined in a sample without potential disturbances because of psychopathology. Methods Using 491 respondents (mean age = 43.0 years, 59.5% female) without lifetime psychiatric disorders from the Netherlands Study of Depression and Anxiety (NESDA), sociodemographic, sampling and health determinants of salivary cortisol levels were examined. Respondents collected seven salivary cortisol samples providing information about 1-h awakening cortisol, diurnal slope, evening cortisol and a dexamethasone (0.5 mg) suppression test (DST). Results Higher overall morning cortisol values were found for smokers, physically active persons, persons without cardiovascular disease, sampling on a working day or in a month with less daylight. In addition, the cortisol awakening response was significantly flattened for males, persons with cardiovascular disease, those with late awakening times and those with longer sleep duration. Diurnal slope was steeper in men, physically active persons, late awakeners, working persons, and season with less daylight. A higher evening cortisol level was associated with older age, smoking and season with more daylight. Cortisol suppression after dexamethasone ingestion was found to be less pronounced in smokers, less active persons and sampling on a weekday. Conclusion Sociodemographic variables (sex, age), sampling factors (awakening time, working day, sampling month, sleep duration) and health indicators (smoking, physical activity, cardiovascular disease) were shown to influence different features of salivary cortisol levels. Smoking had the most consistent effect on all cortisol variables. These factors should be considered in psychoneuroendocrinology research. SA Vreeburg, BP Kruijtzer, J van Pelt, R van Dyck, RH de Rijk, WJG Hoogendijk, JH Smit, FG Zitman, BWJH Penninx Psychoneuroendocrinology 2009; 34: 1109-1120 INTRODUCTION The hypothalamic—pituitary—adrenal (HPA) axis is hypothesized to be one of the key biological mechanisms underlying several stress-related disorders, including somatic and psychiatric. Three decades ago, new techniques were developed to measure cortisol in saliva which provided opportunities for psychoneuroendocrinology research within large epidemiological studies. Sampling of cortisol in saliva returns the active, unbound form of cortisol and is considered to be minimally intrusive on HPA axis regulation. Commonly, the cortisol awakening response (CAR) is target of research since it reflects the natural response of the HPA axis to awakening, usually consisting of a 50—100% increase, peaking around 30 min after awakening. More basal indicators include (‘unchallenged’) cortisol levels during the day or evening. The dexamethasone suppression test is considered a measure of the negative feedback system of the HPA axis. Although determinants of different salivary cortisol indicators have been described, they have not yet been systematically studied within the same study with a large sample size. Moreover, studies vary in the extent to which they take other covariates into account when examining HPA axis indicators, which might partly explain discrepancies in earlier findings. In the literature, there is some indication that sociodemographic indicators, sampling factors as well as (somatic) health indicators may all influence salivary cortisol measures. For instance, for women most studies reported a higher CAR compared to men, although there are exceptions showing no difference in CAR. In addition, lower as well as similar daytime cortisol levels or slope in women compared to men have been found. Although older age has been associated with a lower CAR independent of awakening time, various other studies found no age differences in CAR or total amount of cortisol secretion during the day. Lower socioeconomic status (SES) has been found to be related to a higher CAR and daytime cortisol in univariable analyses, but Cohen et al. found that this association disappeared after considering smoking status in the analysis. Consequently, the absence of adjustment for other SES differences, such as smoking status, can contribute to inconsistent findings across studies. The role of sampling factors has been mostly investigated in relation to the CAR. A higher CAR was associated with sampling on working days, early awakening, and longer as well as shorter sleep duration, however, several studies found no effects on the CAR for working day, awakening time, or sleep duration. Also, the effect of month of sampling on salivary cortisol indicators still remains unclear since some studies reported a higher CAR when more light is present, but Polk et al. did not find a higher CAR in May compared to December. Among the health indicators that have been shown to affect salivary cortisol levels are good health, which has been linked to higher CAR and physical exercise, which has been associated with higher daytime cortisol levels after sufficient exercise. For smokers, studies reported a higher CAR, a slightly attenuated CAR, or no difference in CAR. The use of alcohol has been associated with a larger CAR and a flatter diurnal slope, but no alcohol use and CAR associations have been described as well. Also Body mass index (BMI) was found to be inconsistently associated with lower cortisol levels at 45 min postawakening, a higher CAR, or no association with CAR, evening cortisol or diurnal slope. Clearly, there is a need to examine the wide range of potential determinants of salivary cortisol together, to ascertain the relative importance of different mechanisms influencing salivary cortisol levels. This is also important for future research since it indicates which factors to control for when examining certain specific associations with cortisol levels. The present study uses detailed information of 491 healthy subjects participating in the Netherlands Study of Depression and Anxiety (NESDA) to examine which sociodemographic, sampling and health characteristics most importantly affect the main salivary cortisol measures (1-h awakening cortisol, diurnal slope, evening level and results on the dexamethasone suppression test). METHODS Study sample Data are from the Netherlands Study of Depression and Anxiety, a large cohort study on the course of depressive and anxiety disorders in adults (aged 18—65 years). In total, 2981 respondents were recruited from the community, general practice and specialized mental health care. The study sample included persons with psychopathology as well as controls without a psychiatric diagnosis. All respondents provided written informed consent before participating in the study. For objectives and methods of NESDA see Penninx et al. (2008). Controls (n = 676) were recruited from the community (n = 140) and general practice (n = 536) and were more often male (38.5 versus 32.2%, p = 0.002), had a higher mean education (in years: 12.7 versus 12.0, p < 0.001), were less often sleeping ≤ 6h (16.0 versus 31.8%, p < 0.001), less often currently smoking (27.5 versus 41.8%, p < 0.001), had a lower mean BMI (25.1 versus 25.7, p = 0.01), and were less likely to have a chronic disease (32.5 versus 41.1%, p < 0.001) compared to the remainder of the NESDA sample (n = 2305), but did not differ in age, race, use of alcohol, or level of physical activity. To obtain an indication of the main determinants of salivary cortisol variables unbiased by potential psychopathology effects, the present study only included controls from the NESDA cohort. Controls were defined as having no prior lifetime history of anxiety disorder (panic disorder, generalized anxiety disorder, or social phobia) or depressive disorder (major depressive disorder (MDD) or dysthymia) as assessed by the DSM-IV Composite Interview Diagnostic Instrument (CIDI) (WHO version 2.1). These criteria fitted 676 of all NESDA respondents. In addition, we excluded eight subjects taking antidepressants, 12 pregnant or breastfeeding women and 29 participants on corticosteroids, leaving an initial 627 subjects.