Elevated cortisol at night is one of the more plausible biological explanations for lying awake with a racing mind. The relationship is bidirectional and well-documented: high cortisol disrupts sleep architecture, and poor sleep may push cortisol higher the following day. One 2024 study found that total sleep deprivation significantly altered cortisol responses to acute stressors in healthy individuals — the feedback loop is real, and it matters.
evidence-actually-shows">What the evidence actually shows
The cortisol-sleep relationship is not subtle. Messa et al. (2024) examined healthy adults after total sleep deprivation and found measurable changes in both autonomic nervous system activity and cortisol responses to stressors — the kind of physiological shift that, if repeated chronically, could plausibly worsen the very symptoms people associate with high cortisol: anxiety, poor concentration, difficulty winding down at night.
Liu et al. (2022) looked at dementia caregivers — a group under sustained psychological load — and found that cortisol total output mediated the relationship between poor daily sleep and anxious and depressive symptoms. That mediation pathway is important. It suggests cortisol is not just a correlate of bad sleep; it may be part of the mechanism connecting sleep disruption to mood deterioration.
Then there is the clinical extreme. Ferrari et al. (2025) studied patients with Cushing's syndrome — a condition characterised by pathologically high cortisol — and found that chronotherapy restoring a more normal cortisol rhythm also led to measurable improvements in sleep quality and quality of life scores. That is not directly applicable to someone who is simply stressed and sleeping badly. But it does confirm, in a clinical setting, that cortisol rhythm and sleep quality are tightly coupled.
My honest read: the evidence is consistent but much of it is observational. Causality is hard to establish cleanly. What I am confident saying is that the cortisol-sleep relationship is real, bidirectional, and worth taking seriously — without overstating what any individual can do about it.
The biology: what is actually happening at night
Cortisol follows a circadian rhythm. Under normal conditions, it is at its lowest point around midnight and in the early hours of sleep, then rises steeply in the hour before waking — the so-called cortisol awakening response. That morning rise serves a genuine purpose: it mobilises energy, primes alertness, and prepares the body for the demands of the day.
The problem arises when that rhythm is disrupted. Psychological stress, irregular light exposure, late-night screen use, and poor sleep itself can all shift the curve. Cortisol that should be low at 11pm may remain elevated. And elevated cortisol at night is directly antagonistic to sleep onset for a straightforward reason: it activates the hypothalamic-pituitary-adrenal (HPA) axis, which in turn promotes arousal rather than the quieting of the nervous system that sleep requires.
Swaab et al. (2005) described how chronic HPA axis dysregulation — sustained elevated cortisol output — is associated with structural and functional changes in brain regions involved in stress regulation, including the hippocampus. The hippocampus normally acts as a brake on cortisol secretion via negative feedback. When that feedback weakens, cortisol may remain elevated for longer, making the problem self-perpetuating.
Separately, cortisol suppresses melatonin production. Melatonin is not a sedative in the pharmacological sense, but it is a key timing signal — it tells the brain that darkness has arrived and sleep should begin. If cortisol is high when it should be low, that signal is blunted. You may not feel sleepy at the right time, or you may fall asleep but wake in the early hours as cortisol begins its pre-dawn rise earlier than it should.
Late-night screen use is worth flagging here. Lissak (2019) reviewed evidence linking screen time to adverse physiological and psychological outcomes, including disrupted sleep — a finding consistent with the light-cortisol-melatonin pathway described above. The mechanism is not purely cortisol, but the overlap is relevant.
Recognising high cortisol symptoms at night
This is where I want to be careful. "High cortisol" has become a catch-all phrase online for a range of symptoms that could have many causes. I am not going to list symptoms and imply that ticking three boxes means your cortisol is elevated — it does not work like that without measurement.
That said, the research does describe a recognisable pattern. People with disrupted cortisol rhythms — whether measured via saliva, blood, or hair samples — tend to report:
- Difficulty falling asleep despite feeling physically tired
- Waking between 2am and 4am with a racing mind
- Feeling alert and wired in the evening when they should be winding down
- Fatigue during the day that does not resolve with rest
- Heightened anxiety or irritability, particularly in the afternoon and evening
Mlili et al. (2022) used hair cortisol concentration — a measure of longer-term cortisol exposure — as a biomarker in perinatal depression research. Hair cortisol captures weeks to months of output rather than a single snapshot, which is arguably more clinically meaningful than a one-off blood test for someone experiencing chronic stress.
Mao et al. (2022) found that cortisol mediated and moderated the relationship between sleep quality and depressive symptoms in their study population — suggesting that the symptoms people experience are not simply "stress" or simply "bad sleep" in isolation, but a combined picture where cortisol sits at the intersection.
The adolescent data: why early patterns matter
I find the adolescent research particularly interesting because it shows how early these patterns can establish themselves. Kuhlman et al. (2020) followed adolescents prospectively and found that sleep problems in early adolescence were linked to later depressive symptoms, with the cortisol awakening response acting as a mediating pathway. In plain terms: poor sleep in teenagers may alter the cortisol awakening response in ways that increase vulnerability to low mood over time.
This is not a reason to panic about teenage sleep habits. But it does suggest that cortisol rhythm is not a fixed biological trait — it is shaped by behaviour, and those behavioural patterns can have downstream consequences that persist.
The bidirectional problem: lack of sleep cortisol and the feedback loop
One of the more frustrating aspects of this topic is that the relationship runs in both directions. High cortisol disrupts sleep. Poor sleep may push cortisol higher. And then elevated cortisol the following night makes sleep harder again.
Messa et al. (2024) demonstrated this in healthy adults — people without pre-existing conditions, under controlled laboratory conditions. After total sleep deprivation, their cortisol stress responses were altered. That is a single night. Chronic partial sleep restriction, which is far more common in real life, likely produces a less dramatic but more sustained version of the same effect.
The practical implication is that trying to address cortisol symptoms without also addressing sleep — or trying to address sleep without acknowledging the cortisol component — may mean you are only solving half the problem. They need to be approached together.
What the evidence supports for sleep and cortisol management
I want to be honest about the limits here. There is no supplement with strong, replicated human RCT data showing it directly normalises cortisol rhythm in otherwise healthy adults. Anyone claiming otherwise is overstating the evidence. What does exist is a body of research on behavioural interventions, and some preliminary ingredient-level data worth understanding.
Behavioural approaches with the strongest evidence base
Consistent sleep and wake times — even at weekends — have the most reliable effect on circadian rhythm stability, which in turn supports a more predictable cortisol curve. Light exposure matters too: bright light in the morning may help anchor the cortisol awakening response to the right time, while reducing artificial light in the evening may support the natural evening decline.
Exercise is more nuanced. Braschler et al. (2025) reviewed the physiology of endurance exercise and noted the significant cortisol response that high-intensity training provokes. For sleep and cortisol, timing of exercise matters — intense training close to bedtime may delay the cortisol decline that sleep requires.
Ingredients with research interest — and their honest limits
Glycine has been studied in small human trials for sleep-related outcomes. The human data is preliminary, large-scale RCTs are limited, and I would not claim it normalises cortisol — but the research is ongoing and worth watching. The KōJō Daily Formula includes 2,000mg of crystalline glycine, which is consistent with doses used in the available research.
Taurine has been examined in animal and some human studies for its potential effects on the nervous system and stress responses. Research is ongoing, large-scale human trials are limited, and I would not overstate what the current data shows.
Aged Garlic Extract, Olive Leaf Extract, Grape Seed Extract, and Pine Bark Extract each have research interest relating to oxidative stress pathways. The human data on their specific relevance to cortisol or sleep is thin, and I would be overstating it to claim otherwise. Research is ongoing.
Vitamin C contributes to the reduction of tiredness and fatigue, and contributes to the protection of cells from oxidative stress — both registered claims with an established evidence base. Whether that translates to meaningful support for someone experiencing high cortisol symptoms is a reasonable question, but the direct cortisol-specific data in healthy adults is limited.
If you are thinking about a once daily supplement as part of a broader approach, the honest framing is this: no supplement replaces the behavioural foundations. But some ingredients have genuine research interest, and transparency about what that research actually shows is the minimum standard.
Frequently asked questions
Can high cortisol cause waking up at 3am or 4am?
Early morning waking is consistent with the pattern seen in cortisol rhythm disruption. Cortisol naturally begins rising in the pre-dawn hours, and if that rise starts earlier than it should — or is more pronounced — it may pull you out of sleep. The evidence linking disrupted cortisol rhythm to sleep architecture changes is documented in Ferrari et al. (2025), though that study focused on a clinical population with pathologically elevated cortisol.
How do I know if my cortisol is actually high, rather than just stressed?
Symptoms alone are not diagnostic. Salivary cortisol tests taken at multiple points across the day give a more useful picture of rhythm than a single blood draw. Hair cortisol, as used in research by Mlili et al. (2022), reflects longer-term exposure. If you suspect a genuine clinical issue, a GP referral for formal testing is the right first step — not a home kit.
Does poor sleep make cortisol worse the next day?
The evidence suggests it may. Messa et al. (2024) found that total sleep deprivation altered cortisol stress responses in healthy adults under controlled conditions. Whether a single night of poor sleep has the same effect as total deprivation is less clear, but the directional relationship — poor sleep pushing cortisol reactivity higher — is biologically plausible and supported by the available data.
Can anxiety and high cortisol symptoms overlap enough to be confused?
Yes, and that overlap is not accidental. Liu et al. (2022) found that cortisol total output mediated the relationship between sleep disruption and anxious symptoms in their study population. The physiological experience of elevated HPA axis activity — racing thoughts, heightened alertness, difficulty settling — overlaps substantially with what people describe as anxiety. Distinguishing them clinically requires measurement, not symptom-matching.
Is there a specific time of day when cortisol symptoms are worst for sleep?
The evening is the most critical window. Cortisol should be declining steadily from the afternoon onward, reaching its lowest point around midnight. Anything that delays that decline — high psychological stress, bright light exposure, intense exercise, or stimulants — may extend the period during which cortisol is antagonistic to sleep onset. The timing data is consistent across the circadian literature, including work by Swaab et al. (2005).
Do teenagers experience high cortisol sleep symptoms differently to adults?
The adolescent data suggests the consequences may be more lasting. Kuhlman et al. (2020) found that sleep problems in adolescence were prospectively linked to later depressive symptoms via the cortisol awakening response — suggesting that disrupted sleep during development may shape cortisol patterns in ways that persist into adulthood. The mechanisms are not fully established, but the prospective design of that study makes the finding harder to dismiss.
My honest take
I started reading this literature because I kept waking up at 3am during a particularly pressured period of building KōJō. I wanted to know whether that was "just stress" or whether there was something more specific happening biologically. The answer, as far as I can tell, is both — and the distinction matters less than I thought it did.
What the evidence actually shows is a feedback loop that is genuinely difficult to interrupt once it is established. Cortisol disrupts sleep. Poor sleep may make the cortisol response to stress more reactive the following day. And then the cycle continues. That is not catastrophising — it is what the data describes.
What I am less certain about is how much any specific supplement intervention moves the needle for someone in otherwise good health. The behavioural evidence — sleep timing, morning light, managing evening stimulation — is considerably stronger than the supplement evidence. I think that is worth saying plainly, even though it is not the most commercially convenient thing for me to write.
The ingredient research I find most interesting is around glycine and taurine — not because the human data is definitive, but because the mechanistic rationale is plausible and the safety profile is well established. I included both in the formula at meaningful doses for that reason. But I would not want anyone to read that as a claim that taking them will fix disrupted cortisol. The honest position is: the research is ongoing, the human data is limited, and behavioural foundations come first.
If your sleep has been poor for months and the pattern described above resonates, it is worth speaking to a GP. Cortisol rhythm disruption at a clinical level — not just stress-related — is diagnosable and treatable. That conversation is more important than any supplement decision.
This article is for informational purposes only and does not constitute medical advice. Consult your healthcare provider before starting any supplement regimen.
References (9 studies)
- Messa et al. (2024) — The effect of total sleep deprivation on autonomic nervous system and cortisol responses to acute stressors in healthy individuals. PMID 38991306.
- Liu et al. (2022) — Daily Sleep and Anxious and Depressive Symptoms Among Dementia Caregivers — The Mediation of Cortisol Total Output. PMID 35595515.
- Ferrari et al. (2025) — Chronotherapy With Once-Daily Osilodrostat Improves Cortisol Rhythm, Quality of Life, and Sleep in Cushing's Syndrome. PMID 40172910.
- Swaab et al. (2005) — The stress system in the human brain in depression and neurodegeneration. PMID 15996533.
- Lissak (2019) — Adverse physiological and psychological effects of screen time on children and adolescents: Literature review. PMID 29499467.
- Mlili et al. (2022) — Hair Cortisol Concentration as a Biomarker of Symptoms of Depression in the Perinatal Period. PMID 35297354.
- Mao et al. (2022) — The Mediating and Moderating Roles of Life Skills and Cortisol in the Relationship Between Sleep Quality and Depressive Symptoms. PMID 35903627.
- Kuhlman et al. (2020) — Sleep problems in adolescence are prospectively linked to later depressive symptoms via the cortisol awakening response. PMID 31387652.
- Braschler et al. (2025) — Physiology and Pathophysiology of Marathon Running: A narrative Review. PMID 39871014.
