do people with major depression have dysregulated cortisol?
“major depression has people have cortisol dysregulation”
a guest, a psychologist and neuroscientist said on The Peter Attia Drive, 2026-10-26, at 54:32. hear it in context →
first, the part that holds
on average, people with major depression do show more cortisol and more hpa-axis activity than people without it. four decades of studies point the same way, and the gap is largest in older inpatients and in melancholic or psychotic depression.
what we found when we went and looked
we read the largest meta-analysis on this question: 361 studies, 18,454 participants, in psychosomatic medicine (2011). depressed groups had higher cortisol than controls (d = 0.60), but when only methodologically stronger studies were counted, the gap shrank to d = 0.33. a 14-study meta-analysis of the dexamethasone suppression test found cortisol nonsuppression in 41% of people with nonpsychotic depression and 64% with psychotic depression, so a large share of people with depression do not show it on that test. a 20-study meta-analysis of salivary cortisol found a small morning difference and concluded that salivary cortisol cannot tell people with depression apart from people without it. so the group difference is real, but "people with major depression have dysregulated cortisol" reads as true of the typical person with the diagnosis, and the record supports it for a subgroup and as an average.
assertion by assertion
| assertion | verdict | what the record says |
|---|---|---|
| people with major depression have dysregulated cortisol (read as a feature of people with the diagnosis generally) | overstated | the record shows a group-average shift and a subgroup effect, not a trait most individuals share: dexamethasone nonsuppression was 41% in nonpsychotic depression (64% psychotic), and the cortisol gap shrank to d = 0.33 in stronger studies. 'dysregulation' can also mean flattened daily rhythms or a blunted response, which these meta-analyses measure only partly [1][2][3] |
| on average, depressed groups have higher cortisol than nondepressed groups | confirmed | d = 0.60 (95% ci 0.54-0.66) across 361 studies; acth also higher (d = 0.28); crh not higher (d = 0.02). a 2022 meta-analysis of 17 serum/plasma studies found smd 1.18, but heterogeneity was high (i² = 85%) [1][4] |
| the effect holds at that size in rigorous studies | overstated | limited to studies meeting methodological standards, d fell to 0.33 (95% ci 0.21-0.45). by our own normal-curve calculation, that leaves about 87% overlap between the depressed and nondepressed distributions [1] |
| cortisol dysregulation is concentrated in particular kinds of depression | confirmed | larger gaps in older hospitalized samples than in younger outpatients, and in melancholic and psychotic features; dexamethasone nonsuppression was 64% in psychotic vs 41% in nonpsychotic depression [1][3] |
| cortisol can tell people with depression apart from people without it | false | for salivary cortisol: morning difference of 2.58 nmol/l. ranges overlapped heavily (0-29 depressed vs 1-23 controls), and the authors concluded it 'is unable to discriminate between persons with and without depression'. the speaker did not make this claim; we checked it because it is the popular reading [2] |
both directions
for the speaker: the 2011 meta-analysis is large (361 studies, 671 effect sizes) and points one way: higher cortisol and acth in depressed groups. the effect survives methodological filtering at a smaller size. a 2022 serum/plasma meta-analysis found a larger pooled difference (smd 1.18). dexamethasone nonsuppression shows up in a large minority of depressed patients and in most people with psychotic depression. read broadly, as an hpa-axis abnormality that is common in depression, the statement has a real basis.
against: the sentence reads as a property of people with depression, and the evidence shows an average difference whose size depends on who is sampled and how. stronger studies shrink it to d = 0.33. most people with nonpsychotic depression suppress normally on the dexamethasone test. salivary cortisol, the cheapest and most common measure, could not discriminate cases from controls, and trial-sequential analysis suggested the small difference may be random error. crh, the hormone at the top of the axis, was not elevated. nothing in this record shows that cortisol dysregulation causes depression, rather than coming with it or following from it.
what this cannot say
the speaker gave no source and did not define "dysregulation", which can mean high levels, nonsuppression, flattened rhythm or blunted reactivity. the meta-analyses we read mostly measure levels and suppression. all of these studies are cross-sectional case-control designs that mix samples, assays, times of day and medication status. the dexamethasone figures come from studies largely from the 1980s and 90s. we read abstracts, not full texts. the 87% overlap figure is our own calculation from d = 0.33, assuming normal distributions with equal variance, and is not reported by the authors.
sources, each one fetched and read
- meta-analysis, 'Depression and hypothalamic-pituitary-adrenal activation: a quantitative summary of four decades of research', Psychosomatic Medicine 2011;73(2):114-26, PMID 21257974 What we read there: depressed individuals tended to display increased cortisol (d = 0.60; 95% CI 0.54-0.66) and ACTH (d = 0.28) but not CRH (d = 0.02); the cortisol effect fell to d = 0.33 (0.21-0.45) under methodological standards; larger differences in older hospitalized samples and in melancholic and psychotic depression Fetched 2026-10-27 (HTTP 203). https://pubmed.ncbi.nlm.nih.gov/21257974/
- systematic review and meta-analysis, 'Salivary cortisol in depressed patients versus control persons', Psychoneuroendocrinology 2010;35(9):1275-86, PMID 20447770 What we read there: 20 case-control studies, 1,354 patients and 1,052 controls; morning salivary cortisol higher by 2.58 nmol/l; reference intervals 0-29 vs 1-23 nmol/l; 'salivary cortisol is unable to discriminate between persons with and without depression' Fetched 2026-10-27 (HTTP 203). https://pubmed.ncbi.nlm.nih.gov/20447770/
- meta-analysis, 'DST studies in psychotic depression: a meta-analysis', American Journal of Psychiatry 1997;154(11):1497, PMID 9356556 What we read there: 14 studies; cortisol nonsuppression on the dexamethasone suppression test in 64% of psychotic vs 41% of nonpsychotic depressed patients Fetched 2026-10-27 (HTTP 203). https://pubmed.ncbi.nlm.nih.gov/9356556/
- systematic review and meta-analysis, 'A systematic review and meta-analysis of serum and plasma cortisol levels in depressed patients versus control', Indian Journal of Psychiatry 2022, PMC9707655 What we read there: 17 studies; pooled smd of cortisol in depressed vs control 1.18 (95% CI 0.84-1.52), heterogeneity I² = 85% Fetched 2026-10-27 (HTTP 200). https://pmc.ncbi.nlm.nih.gov/articles/PMC9707655
speakers are described by role, not named. one sentence is quoted under fair use, with attribution and a link to the original at the timestamp; we do not re-host audio or video. nothing on this page is a reason to start, stop or change a medication. if you have the file behind anything marked unverifiable, the research board is where people can answer. adam