Many molecular epidemiology studies focusing on high prevalent diseases, such as metabolic disorders and cancer, investigate metabolic and hormonal markers. In general, sampling for these markers can occur at any time-point during the day or after an overnight fast. However, environmental factors, such as light exposure and food intake might affect the levels of these markers, since they provide input for the internal time-keeping system. When diurnal variation is larger than the inter-individual variation, time of day should be taken into account. Importantly, heterogeneity in diurnal variation and disturbance of circadian rhythms among a study population might increasingly occur as a result of our increasing 24/7 economy and related variation in exposure to environmental factors (such as light and food).

The aim of the present study was to determine whether a set of often used biomarkers shows diurnal variation in a setting resembling large molecular epidemiology studies, i.e., non-fasted and limited control possibilities for other environmental influences.

We show that markers for which diurnal variation is not an issue are adrenocorticotropic hormone, follicle stimulating hormone, estradiol and high-density lipoprotein. For all other tested markers diurnal variation was observed in at least one gender (cholesterol, cortisol, dehydroepiandrosterone sulfate, free fatty acids, low-density lipoprotein, luteinizing hormone, prolactin, progesterone, testosterone, triglycerides, total triiodothyronine and thyroid-stimulating hormone) or could not reliably be detected (human growth hormone).

Thus, studies investigating these markers should take diurnal variation into account, for which we provide some options. Furthermore, our study indicates the need for investigating diurnal variation (in literature or experimentally) before setting up studies measuring markers in routine and controlled settings, especially since time-of-day likely matters for many more markers than the ones investigated in the present study.