The cortisol awakening response: Fact or fiction?

Abstract

There has been significant discussion in recent years whether the increase in cortisol release that accompanies waking is dependent on the waking process itself or instead reflects a continuation of an underlying circadian rhythm. Establishing the origin or indeed existence of the so-called cortisol awakening response is important as disturbances in post-awakening cortisol secretion are associated with a range of stress-related disorders. The study reviewed in this article adopted an innovative in vivo microdialysis approach to measure tissue-free cortisol levels in 201 healthy volunteers before and after awakening in a home setting (Klaas et al., 2025). Rather surprisingly, the rate of increase in cortisol secretion did not change when participants awoke compared with the preceding hour when participants were asleep. However, considerable between-subject variability was observed, which was partly explained by sleep duration and the timing of waking relative to the previous morning. These findings highlight the complexity of the cortisol awakening response and summon caution in the interpretation of cortisol measurements based solely on post-awakening responses.

Reviewed article: (Klaas et al., 2025)

Introduction

The cortisol awakening response (CAR) is a widely recognised neuroendocrine phenomenon involving a rise in cortisol secretion during the first hour of awakening. A post-awakening surge in cortisol release is hypothesised to prepare individuals for anticipated energy (or stress) demands of the forthcoming day (Stalder et al., 2025) and is often assumed to be a discrete rather than emergent signal of underlying circadian fluctuations in cortisol release (Wilhelm et al., 2007). CAR has attracted interest as a biomarker of stress reactivity in depression, post-traumatic stress disorder, and other stress-related disorders, with tantalising evidence that elevated cortisol responses to waking may be related to positive health outcomes (Caulfield and Cavigelli, 2020). However, methodological limitations have frequently restricted the validity and robustness of prior studies despite published guidelines on how to acquire reliable CAR data (Stalder et al., 2022). For example, studies involving the collection of saliva samples do not permit the reliable assessment of pre-awakening cortisol measurements. Repeated blood sampling in a controlled laboratory setting overcomes this limitation and has the advantage of allowing accurate awakening times to be assessed (Anderson et al., 2025), but this approach may affect the quality and duration of sleep compared with a home setting. Klaas and colleagues circumvented these constraints using the technique of in vivo microdialysis to evaluate free cortisol levels in interstitial fluid samples, collected continuously from participants in their own homes, before and after waking (Klaas et al., 2025). By continuously sampling cortisol in a naturalistic environment, this study permitted the assessment of CAR as a putatively distinct post-awakening response over and above normal circadian fluctuations in cortisol secretion.

Experimental approach

Overall, 214 male and female healthy volunteers, aged 18–68 years were recruited across four countries to an observational clinical trial (ULTRADIAN), as previously described (Upton et al., 2023). A small number of participants were excluded who either had missing values in the hour before and after waking (n = 3) or where wake times had not been accurately recorded (n = 10). A linear microdialysis probe was inserted subcutaneously in abdominal tissue with 20-min samples collected automatically over a 24-h period using a portable device secured around the waist. The system was considered safe, well tolerated, and allowed free ambulation and relatively normal daily activities with participants self-reporting sleep and wake times. Adrenal steroids, including cortisol, were analysed using ultrasensitive liquid chromatography coupled with tandem mass spectroscopy. Notably, for the purposes of validation, blood plasma and tissue-free cortisol levels correlated strongly with one another in a small subset of participants (Upton et al., 2023).

Main findings

Dynamic cortisol profiles are shown for each participant in Figure 1. In general, cortisol levels began increasing well before the onset of awakening and peaked roughly within the first hour of being awake. However, crucially, at a population level, the rate of change of cortisol increase was no different between the first hour of awakening and the preceding hour. These findings demonstrate that waking per se is not accompanied by a distinct acceleration in cortisol release. Rather, the best predictor of increased release was the level of cortisol reached in the hour preceding awakening. Nevertheless, as evident in Figure 1, substantial between-subject variability was present with some individuals showing a steeper rise in cortisol release than others. To investigate the origin of this variability, the authors considered how sleep duration and wake time variation (relative to the preceding morning) affected the results. The cohort was split into participants with short (mean 369 min) and long (mean 548 min) sleep durations and participants considered ‘aligned’ (less than 1-h variation in wake time) and ‘misaligned’ (greater than 1-h variation in wake time) sleepers. Remarkably, for long sleepers, the maximal rate of cortisol release occurred 97 min before waking. Short sleepers, by contrast, showed a maximum increase in cortisol release 12 min after waking. This same pattern emerged for misaligned and aligned sleepers with the maximum rate of cortisol increase occurring 68 min before waking and 12 min after waking, respectively.

Figure 1.

Subcutaneous levels of cortisol (nmol/L) measured using in vivo microdialysis in 201 healthy volunteers (depicted by the individual blue lines) aligned by wake time (0 h).

The dashed lines depict the 5th, 25th, 75th, and 95th percentiles (bottom to top). The red-shaded area extends from 1 h before waking to 1 h after waking. Reproduced from the work by Klaas et al. (2025) and published by the Royal Society under the terms of the Creative Commons Attribution Licence http://creativecommons.org/licences/by/4.0/.

Significance of the study

This study is important because it challenges the long-standing assertion that CAR is a distinctive post-awakening response superimposed on an endogenous cortisol rhythm. While numerous studies have reported an increase in cortisol secretion after waking (for review, see Clow et al., (2004)), only a small number of studies have employed quantitative methods to compare the rate of change in cortisol release before and after awakening. Furthermore, this study is notable in allowing participants to go about their everyday lives while adrenal steroids were measured using a relatively nonintrusive sampling procedure.

The absence of a measurable increase in the rate of cortisol secretion during the first hour of waking is an important null finding that calls into the question the existence of CAR as a distinct event. The assumed purpose of CAR is that of an endocrine response to the stress of waking up. However, this study clearly shows there is nothing remarkable about waking per se and that cortisol secretion during initial waking appears to be more tightly regulated by intrinsic cortisol rhythmicity. This view is supported by recent evidence that humans deprived of sleep show a significantly blunted cortisol response in the morning (Vargas and Lopez-Duran, 2020). Thus, circadian processes appear to influence CAR only when awakening is anticipated. Indeed, cortisol levels do not increase when participants are forcibly awoken during the night (Dettenborn et al., 2007), suggesting the transition from sleep to awakening is not the primary driver of increased cortisol secretion.

This study is also significant in revealing an extraordinary level of intersubject variability in cortisol dynamics. A long duration of sleep (~9 h) predicted maximal cortisol secretion well before the onset of awakening, whereas for short sleepers (~6 h), the maxima occurred after waking. Variability in cortisol dynamics were also explained by irregularities in the expected wake time of participants relative to the previous morning. These findings highlight the complexity and difficulty of interpreting CAR as a dynamic response so close to the circadian zenith. They also encourage further research to understand the consequences of cortisol variability for day-to-day behavioural and cognitive functions and more especially longer-term health impacts.

Although this study mitigated several important limitations in how CAR is normally assessed, several caveats are worth noting. First, it is likely the case that interstitial levels of cortisol lagged saliva and plasma levels making it more difficult to compare the present findings with prior studies. In addition, the relationship of this lag with actual wakening times is unknown. Second, as the sampling technique averaged cortisol levels over 20-min intervals, a more precise temporal alignment of cortisol levels before and after waking was not possible. Third, as awakening times were self-reported by participants, there may have been some discrepancies with actual wake times. However, as these limitations do not affect the rate of change in cortisol secretion measured in the same way before and after waking, the conclusion that awakening is not associated with an increased rate of cortisol secretion is both justified and valid.

In summary, this important study challenges the concept of an invariant CAR in humans. The reported findings are expected to stimulate research on the origins of individual differences in cortisol dynamics and how changes in the amplitude and rate of cortisol secretion during the awakening period predict general health-related outcomes.