High cortisol levels and sleep: what the data shows

Elevated evening cortisol is one of the better-documented drivers of poor sleep quality. A 2021 study of 2,672 adults found that higher late-day salivary cortisol correlated with significantly shorter sleep duration and worse subjective sleep quality. The mechanism isn't mysterious - cortisol is an arousal hormone, and when it's circulating at the wrong time, sleep simply doesn't come easily. Here's what the evidence actually looks like.

What the evidence actually shows

The relationship between cortisol and sleep isn't speculative. It's been measured directly, repeatedly, across populations. The question is how strong the effect is and whether it's causal or correlational - and the honest answer is: probably both, depending on which direction you're looking at it.

Vgontzas et al. (2001) found that insomnia patients had significantly higher 24-hour cortisol secretion compared to good sleepers, with the most pronounced difference in the evening and first half of the night - exactly when cortisol should be at its lowest. The effect wasn't trivial. Mean cortisol AUC (area under the curve) was roughly 20% higher in the insomnia group.

Backhaus et al. (2007) demonstrated this in a controlled setting: experimentally induced stress in the evening elevated cortisol and reduced slow-wave sleep (SWS) by a statistically significant margin (p < 0.05), with SWS reductions of around 15-20 minutes per night. That matters because SWS is where most physical recovery happens.

Hirotsu et al. (2015) reviewed the bidirectional nature of the relationship: poor sleep raises cortisol the following day, which then makes the next night's sleep worse. It's a loop, not a one-way street. I think this is the bit most people miss - they assume cortisol is the cause and sleep is the victim, but the causality runs both ways.

The biology: why cortisol and sleep are fundamentally opposed

Cortisol is a glucocorticoid - it's produced by the adrenal glands in response to signals from the hypothalamic-pituitary-adrenal (HPA) axis. Under normal conditions, it follows a clear diurnal rhythm: it peaks within 30-45 minutes of waking (the cortisol awakening response, or CAR), then declines steadily throughout the day, reaching its lowest point around midnight.

Sleep, particularly the transition into it, requires the opposite physiological state. Core body temperature needs to drop. Melatonin needs to rise. Arousal systems need to quiet down. Cortisol actively opposes all three of these. It raises core body temperature, suppresses melatonin secretion, and activates the locus coeruleus - the brain's primary noradrenaline hub, which drives alertness.

There's also a direct architectural effect. Cortisol preferentially suppresses slow-wave sleep and, at high enough levels, REM sleep. Steiger (2003) showed that exogenous cortisol administration reduced SWS in healthy volunteers in a dose-dependent manner. This isn't a subtle effect - it's measurable within a single night.

For a deeper look at the physiology of this, I've written more on sleep, cortisol, and why you're wired at 11pm: the physiology of stress-disrupted sleep.

What drives cortisol too high in the evening

It's worth being specific here, because "stress" as a catch-all explanation isn't particularly useful. Several distinct mechanisms elevate evening cortisol in otherwise healthy people.

Psychological stress and rumination

Perceived stress - particularly the kind that involves unresolved cognitive load, like work deadlines, financial worry, or relationship conflict - activates the HPA axis. Adam et al. (2006) found that daily stressor exposure was associated with a flattened diurnal cortisol slope, meaning the evening drop-off was blunted. That's the pattern associated with worse health outcomes and worse sleep.

Light exposure at the wrong time

Blue-spectrum light in the evening suppresses melatonin and, through downstream HPA signalling, can sustain cortisol at higher-than-appropriate levels. This is often underappreciated. The screen you're staring at at 10pm isn't just keeping you mentally stimulated - it's directly interfering with the hormonal conditions required for sleep onset.

Caffeine timing

Caffeine has a half-life of roughly 5-7 hours in most adults, though there's meaningful individual variation based on CYP1A2 genotype. A 200mg coffee at 3pm could still be contributing to cortisol-adjacent arousal at 9pm. Drake et al. (2013) showed that caffeine consumed 6 hours before bedtime still significantly reduced total sleep time (by about 41 minutes on objective actigraphy measures).

Training timing and overtraining

High-intensity exercise raises cortisol acutely. That's normal and appropriate. The problem is when training happens too late in the evening, or when cumulative training volume is high enough to chronically elevate baseline cortisol. The evidence on exact cut-off times for evening exercise is mixed, but the general signal is that vigorous training within 2-3 hours of bed tends to delay sleep onset in most people.

Glycine, taurine, and the supporting evidence

I want to be careful here. Neither glycine nor taurine has a registered health claim in the UK or EU for sleep or cortisol. What I can tell you is what the existing research shows, and I'll be honest about where it falls short.

Glycine - at 3g doses - has been studied in small but well-designed Japanese RCTs. Bannai et al. (2012) found that 3g glycine before bed reduced subjective daytime sleepiness and improved sleep quality scores in 11 subjects with self-reported poor sleep. The proposed mechanism involves glycine acting as an inhibitory neurotransmitter in the spinal cord and lowering core body temperature via peripheral vasodilation - which, as I mentioned, is exactly what sleep onset requires. The sample size is small. I wouldn't overstate this. But the mechanism is biologically plausible and the effect direction is consistent across the available trials. Research is ongoing and large-scale human trials are limited.

Taurine has been studied in the context of GABAergic signalling - it appears to act as a partial GABA-A receptor agonist in some in vitro and animal models. Jia et al. (2008) demonstrated taurine-induced hyperpolarisation in hippocampal neurons via GABA-A receptors. Whether this translates meaningfully to human sleep architecture at supplemental doses is genuinely unclear. The human data on this is thin and I'd be overstating it to claim otherwise. Research is ongoing and large-scale human trials are limited.

The Kojo Daily Formula includes 2000mg glycine and 2000mg taurine - below the 3g glycine dose used in Bannai et al., but within a range that reflects the broader available literature.

Vitamin C, oxidative stress, and the cortisol connection

This one is more interesting than it sounds. The adrenal glands have among the highest concentrations of vitamin C in the body - it's involved in cortisol biosynthesis, but also in its regulation. Brody et al. (2002) ran a double-blind RCT (n=91) in which 3000mg/day vitamin C significantly attenuated cortisol and blood pressure responses to a psychological stressor (the Trier Social Stress Test), compared to placebo. The cortisol attenuation was statistically significant (p < 0.05).

I'm not going to claim vitamin C is a cortisol-lowering supplement - the evidence doesn't support that as a blanket statement, and the doses in that study are higher than what's practical long-term. But the relationship between oxidative stress, HPA reactivity, and vitamin C status is real. Vitamin C contributes to the protection of cells from oxidative stress, and contributes to the reduction of tiredness and fatigue - these are authorised claims, and they're relevant context here.

What actually moves the needle: the behavioural evidence

Supplements aside - and I say this as someone who makes supplements - the interventions with the strongest evidence for normalising cortisol rhythm are behavioural. Not because supplements don't matter, but because the effect sizes from lifestyle changes are larger.

• Sleep consistency: Irregular sleep timing (varying by more than 60-90 minutes across the week) blunts the cortisol awakening response and disrupts the diurnal slope. Phillips et al. (2017) modelled this computationally and showed that circadian misalignment - even without sleep deprivation - degrades HPA rhythm independently.
• Morning light: Natural light within 30-60 minutes of waking anchors the circadian clock, sharpens the CAR, and helps ensure cortisol is high when it should be (morning) rather than when it shouldn't (evening).
• Magnesium: There's reasonable evidence that magnesium deficiency is associated with HPA hyperactivity. Held et al. (2002) showed that magnesium supplementation increased slow-wave sleep and reduced cortisol in a small crossover trial. If you want the full picture on forms and dosing, I've written a detailed piece on magnesium supplements uk.
• Reducing evening cognitive load: Easier said than done, I know. But the evidence for journalling, structured worry time, and even simple to-do list writing before bed reducing pre-sleep cognitive arousal is reasonably solid.

What the clinical picture looks like when cortisol is chronically elevated

I want to be clear: I'm not talking about Cushing's syndrome here. That's a clinical condition involving pathologically elevated cortisol and it needs medical management. What I'm talking about is the more common pattern - subclinical HPA hyperactivity in people under chronic psychological or physiological stress.

The sleep consequences of chronically elevated cortisol are fairly consistent across the literature: longer sleep onset latency, more frequent nocturnal awakenings, reduced SWS, and - over time - reduced total sleep time. Meerlo et al. (2015) reviewed the animal and human evidence and concluded that chronic stress-induced HPA activation produces lasting changes in sleep architecture, some of which persist even after the stressor is removed. That's the part that should give people pause. It's not just about tonight's sleep - it's about what repeated nights of stress-disrupted sleep do to the system over months.

Frequently asked questions

My honest take

I got interested in this topic because I recognised the pattern in myself. Years of early starts, late-night work, and treating sleep as the thing that happened when everything else was done. My sleep wasn't catastrophically bad - but it wasn't good either. I'd lie awake for 45 minutes most nights. I'd wake at 3am with a to-do list assembling itself in my head. I felt tired in the afternoon and oddly alert at 10pm.

Reading the primary literature on cortisol and sleep was clarifying, not because it gave me a supplement to take, but because it helped me understand what was actually happening biologically. The evening alertness wasn't a character flaw or a caffeine problem. It was a predictable consequence of a dysregulated HPA rhythm.

The things that helped me most were, in order: consistent wake time (non-negotiable, even weekends), morning light within 30 minutes of waking, cutting caffeine after 1pm, and genuinely reducing cognitive load in the evening - which meant not checking email after 8pm, even when I wanted to. None of that is glamorous. None of it is something I can put in a formula.

Where I think supplementation plays a supporting role - and I mean supporting, not primary - is in filling genuine nutritional gaps that affect HPA function. Magnesium is the obvious one. Glycine before bed has enough mechanistic plausibility and enough small-trial signal that I include it. But I'd be doing you a disservice if I suggested any supplement was going to fix a cortisol rhythm that's being disrupted by chronic stress, poor sleep timing, and blue light at midnight.

The honest version is: sort the basics first. Then consider whether there's a nutritional gap worth addressing. In that order.

This article is for informational purposes only and does not constitute medical advice. Consult your healthcare provider before starting any supplement regimen.