Cortisol too high, adrenal fatigue: what actually holds up

An expert panel around Stalder, convened by the International Society of Psychoneuroendocrinology, reviewed the methodological evidence and set binding standards.

The validity of the awakening response depends critically on whether the first saliva sample is really taken at the moment of waking. The consensus requires objective control of sampling adherence, standardised instruction, consideration of covariates, defined sampling protocols and predefined analysis strategies.

For you this means: the awakening response is a research instrument with strict conditions. A test kit sent home without any control over the sampling time does not meet those conditions.

A group around Hellhammer, the first author then at the Department of Psychobiology of the University of Trier, measured the cortisol rise after waking on six consecutive days and analysed it with structural equation models.

The rise on a single day is determined to a large extent by situational factors and only to a small extent by stable personal characteristics. For a reliable stable measure, two to six measurement days are needed depending on the metric.

For you this means: saliva is not the problem. A single measurement day is the problem. What you measure on that one morning is to a large extent that one morning.

A team around Perogamvros compared women on oestrogen or the pill, people with a non-functional variant of the binding globulin, and healthy controls before and after 20 mg hydrocortisone, given intravenously and orally.

After intravenous administration, the area under the curve and half-life of total serum cortisol were clearly higher in the pill group and clearly lower in the deficiency group than in controls. Free cortisol and the salivary markers, by contrast, did not differ between the deficiency group and controls. In fairness the second part belongs with it: for free cortisol too, the area under the curve and half-life were higher in the pill group than in controls. The authors therefore record that women on oestrogens may be more exposed to glucocorticoids.

For you this means: if you take the pill and your total cortisol in blood looks high, that may be down to the transport protein and not to the adrenal gland. In women this is one of the most common reasons for an apparently abnormal value. That is not yet an all-clear. It is a reason not to let the value stand on its own.

Elena Valassi summarised, in a review, the data of the European Registry on Cushing's Syndrome, with 1,564 included patients, of whom 1,045 had the pituitary form, 385 the adrenal form and 89 ectopic ACTH production.

The most common features were weight gain at 83 percent, high blood pressure at 79 percent, skin changes at 76 percent and muscle weakness at 70 percent. Diabetes mellitus was at 32 percent, depression at 35 percent. On average two years lay between symptom onset and diagnosis, with a high number of specialists consulted.

For you this means: Cushing's syndrome rarely looks like just tired. It comes as a pattern. Weight, blood pressure, skin, loss of strength, often together with menstrual cycle changes and declining libido.

A group around Galm searched five databases up to August 2018 and pooled 139 studies with 14,140 participants in a hierarchical meta-analysis, comparing seven methods in adults.

The overnight 1 mg dexamethasone suppression test reached a sensitivity of 98.6 percent at a specificity of 90.6 percent. Late-night salivary cortisol came to 95.8 percent sensitivity and 93.4 percent specificity, free cortisol in 24-hour urine to 94.0 and 93.0 percent.

For you this means: saliva as a material is not the problem. Salivary cortisol performs very well for the Cushing question when it is collected late in the evening and measured against a validated cut-off.

A European guideline group around Fassnacht revised the international guideline on incidental adrenal findings in 2023, with four predefined clinical questions and updated systematic reviews.

Everyone with an incidental adrenal finding needs a hormonal assessment including metanephrines and a 1 mg dexamethasone suppression test, cut-off serum cortisol 50 nanomol per litre. For values above this cut-off without overt Cushing's syndrome, the guideline introduces the term mild autonomous cortisol secretion and describes an increased risk of comorbidity and mortality for this group.

For you this means: there is a real zone between normal and Cushing. It has a name, a cut-off and a follow-up plan. And it is something entirely different from a staged model of exhaustion.

Husebye and colleagues summarised causes, clinical picture, diagnostics and treatment of adrenal insufficiency in an invited Lancet review.

The leading symptoms are unintended weight loss, loss of appetite, a drop in blood pressure on standing, deep exhaustion, muscle and abdominal pain and sodium deficiency; in the primary form additionally skin darkening and salt craving. About 50 percent of those affected experience an adrenal crisis after diagnosis. Despite modern replacement therapy, quality of life and ability to work are reduced and mortality is raised.

For you this means: when exhaustion comes with weight loss, circulatory weakness on standing and salt craving, that is a different path from any stress interpretation. This combination belongs in medical assessment and not in a saliva profile.

A GRADE-based guideline group around Bornstein, co-sponsored by the European Society of Endocrinology and the American Association for Clinical Chemistry, drew on two commissioned systematic reviews.

The reference method is the short corticotropin test with 250 micrograms. Where that is not possible, morning plasma ACTH plus cortisol serve as screening. To clarify the cause, a validated test for 21-hydroxylase antibodies follows. Training on dose adjustment during illness, an emergency card and an emergency preparation for injection are explicitly recommended.

For you this means: genuine adrenal insufficiency is not established through a saliva profile but through a stimulation test. And anyone who has it needs an emergency card and emergency medication, not just a prescription.

A group around Broersen searched seven databases up to February 2014 and included 74 papers with a total of 3,753 adult cortisone users who had been tested for adrenal insufficiency.

By route of use, the proportion with adrenal insufficiency ranged from 4.2 percent with nasal use to 52.2 percent after injection into a joint. By dose it ranged from 2.4 percent at a low dose to 21.5 percent at a high dose. In asthma it rose with duration from 1.4 percent under 28 days to 27.4 percent beyond one year.

For you this means: it is not only tablets that count. Sprays, creams and joint injections can dampen your own axis as well. Anyone with persistent unspecific symptoms after cortisone treatment belongs tested, and before adrenal fatigue is even considered.

Cadegiani and Kater of the Adrenal and Hypertension Unit at the Escola Paulista de Medicina, Universidade Federal de São Paulo, searched PubMed, MEDLINE and Cochrane up to April 2016 with ten search terms around adrenal, fatigue, burnout and cortisol.

Of 3,470 papers found, 58 met the criteria, 33 of them in healthy people and 25 in symptomatic people. The measures tested most often were direct waking cortisol, the cortisol awakening response and the salivary cortisol rhythm. The result: almost consistently contradictory findings, regardless of the validation and quality of the tests used. As a limitation the authors name, among other things, a methodology of cortisol assessment that is not robust and not recognised in endocrinology.

For you this means: it is not that nobody looked. People looked, and the findings produce no picture. That is the difference between unresearched and researched without a robust result.

Miller, Chen and Zhou brought together the contradictory literature on chronic stress and the function of the axis in humans in a meta-analysis.

A large part of the differences is explained by features of the stressor and of the person. The time factor is particularly important: hormone activity is raised at the onset of stress and decreases over time. Stressors that threaten physical integrity, involve trauma and are uncontrollable go along with a high, flat daily profile. In people with post-traumatic stress disorder, activity is lower.

For you this means: whether your cortisol comes out high or low depends on when it is measured and what kind of strain is involved. A four-stage model does not capture that.

Chida and Steptoe, the first author then at the Psychobiology Group, Department of Epidemiology and Public Health, University College London, analysed 147 analysable studies from 62 articles and separated the rise after waking from the total amount across the waking period.

The rise was positively associated with work stress and general life stress and negatively with fatigue, burnout and exhaustion. The total amount across the waking period was positively associated with general life stress and negatively with post-traumatic stress.

For you this means: this is precisely why the staged model sounds so plausible. Findings really do exist in both directions. But they depend on the kind of strain, not on a progression from stage one to stage four.

Danhof-Pont, van Veen and Zitman searched MEDLINE and EMBASE for studies comparing biomarkers between people with and without burnout, and ran meta-analyses wherever that was methodologically possible.

31 studies covering 38 biomarkers went in. The meta-analyses showed no differences for the cortisol awakening response, for the awakening response after dexamethasone, for blood cortisol and for blood pressure. The authors conclude: no usable biomarker for burnout, largely because the studies compare so poorly with one another.

For you this means: for the state most people mean when they say adrenal fatigue, there is so far no laboratory value that could confirm or rule it out.

Stetler and Miller pooled 671 effect sizes from 361 studies with 18,454 people and compared cortisol, ACTH and CRH between depressed and non-depressed people.

Depressed people showed raised cortisol with an effect size of d equal to 0.60 and raised ACTH with d equal to 0.28, but no raised CRH. Restricted to studies with minimum methodological standards, the cortisol effect nearly halved to d equal to 0.33. The differences were largest in older inpatients with a melancholic or psychotic presentation.

For you this means: if cortisol goes in any direction at all in severe exhaustion, then in depression it tends to go up. That is the opposite of what the exhaustion model expects.

A group around Tak analysed 85 studies on chronic fatigue syndrome, fibromyalgia and irritable bowel syndrome and additionally ran a meta-regression of possible influencing factors.

Across all three conditions, basal cortisol tended to be lower, without statistical significance. Considered separately, significantly lower basal cortisol was found in chronic fatigue syndrome, with a standardised mean difference of minus 0.14 and an interval reaching zero. In fibromyalgia it appeared only in the subgroup of women, and in irritable bowel syndrome not at all. In the meta-regression, only the type of disorder and female sex remained as independent predictors.

For you this means: even where exhaustion is at its most severe, no consistently low cortisol is found. The effect is small and its interval touches zero.

A group around Stalder analysed 124 subsamples from 66 independent studies with a total of 10,289 people on hair cortisol.

Strained groups showed 22 percent higher hair cortisol values overall. Considered separately, the elevation appeared only when the strain was still ongoing at the time of the examination, then at plus 43 percent, while no meaningful difference remained for past or ended strain. In anxiety disorders including post-traumatic stress disorder, values were 17 percent lower. For self-reported stress and low mood, no consistent associations were found.

For you this means two things. If any method responds to ongoing strain, it is this one. And even there the laboratory result does not reliably match what you subjectively feel.

A group around Debono prospectively examined 220 people at high risk of adrenal insufficiency. A saliva sample was taken at home on waking, after which the ACTH stimulation test followed in the clinic. Salivary cortisone was measured by mass spectrometry, funded by the National Institute for Health Research.

The frequency of adrenal insufficiency on the stimulation test was 44 percent. The area under the ROC curve for salivary cortisone was 0.95. Cut-offs with at least 95 percent sensitivity and specificity gave a negative predictive value of 96 percent and a positive one of 95 percent. In 70 percent of participants the saliva value delivered the same information as the stimulation test, and 83 percent preferred collection at home.

For you this means: a home test can be very good. Two qualifications belong with it. What was measured was cortisone and not cortisol. And the question was one of underactivity, not one of adrenal fatigue.

Leproult and colleagues at the University of Chicago measured plasma cortisol profiles over 32 hours in healthy young men, under normal sleep, under partial sleep deprivation and under total sleep deprivation.

Changes appeared only in the evening after the deprivation night. After partial sleep deprivation, cortisol values between 6 and 11 pm the following day were 37 percent higher, after total sleep deprivation 45 percent higher. The onset of the evening quiescent period was delayed by at least an hour.

For you this means: a single bad night shifts the curve measurably upwards the next evening. That is the fastest adjustable screw on the axis, and it costs nothing.

Spiegel, Leproult and Van Cauter had eleven young men spend six nights with only four hours in bed and compared that with six recovery nights with twelve hours in bed.

Under sleep debt, glucose tolerance was poorer and the thyrotropin level lower. Evening cortisol values were raised and sympathetic nervous system activity was increased.

For you this means: it is not only cortisol that shifts. Glucose handling and thyroid regulation go along with it. That may explain why chronic sleep loss can feel like a hormonal problem without any gland being diseased.

A group around Ken-Dror pooled 22 studies, of them 16 prospective cohorts with 385 participants and 6 randomised controlled trials with 252 participants, at a mean age of 41 to 62 years.

Under CPAP, plasma cortisol fell by a standardised mean difference of minus 0.28 in the cohort studies and by minus 0.39 in the randomised trials. Systolic blood pressure fell by 5.4 mmHg, diastolic by 3.3 mmHg. The authors explicitly place untreated sleep apnoea among the sources of false positive findings when Cushing is suspected.

For you this means: if you snore, pauses in breathing have been observed or you wake up feeling wrecked, that belongs looked for before cortisol is interpreted. Treatment may then favourably influence both.

Pascoe, Thompson and Ski analysed 42 randomised controlled trials comparing approaches involving yoga postures, with or without mindfulness-based stress reduction, against an active control group.

Compared with the active control, lower evening cortisol, lower waking cortisol, lower ambulatory systolic blood pressure, lower resting heart rate, altered high-frequency heart rate variability as well as lower fasting blood sugar, cholesterol and LDL were seen. The authors explicitly stress the heterogeneity of the included approaches.

For you this means: the comparison against an active control is decisive, because it largely excludes pure attention effects. What these approaches have in common is regularity, not one particular technique.

Hill and colleagues had twelve active, moderately trained men exercise for 30 minutes each on separate days at 40, 60 and 80 percent of maximum oxygen uptake, plus a 30-minute rest control, with time of day, diet and prior load controlled.

The cortisol change from before to after the session was minus 6.6 percent in the rest control, plus 5.7 percent at 40 percent intensity, plus 39.9 percent at 60 percent and plus 83.1 percent at 80 percent. ACTH rose significantly only at 80 percent. After correction for the fall in plasma volume, the low intensity even lowered cortisol.

For you this means: exercise is not a one-way street. A calm endurance pace appears to lower cortisol, hard sessions raised it clearly in this study. Twelve men, a single acute bout, no statement about long-term effects and none about women.