FitProven
Evidence Breakdown
Based on 7 studies
THE BIGGER QUESTION
Whether this one statement holds is settled above. What to actually do about it is a wider question, weighed across every claim that bears on it.
Evidence map
For & against, at a glance
Crossover Trial
Neutral
Lovallo WR et al. · 2005Psychosomatic MedicineDouble-blind crossover trial in 96 healthy adults (48 men, 48 women, mean age ~28) run over four weeks. In each week participants abstained from dietary caffeine for five days while taking capsules providing 0, 300, or 600 mg/day, then on day six received either placebo or 250 mg caffeine at 09:00, 13:00 and 18:00 while salivary cortisol was sampled eight times from 07:30 to 19:00. After abstinence, caffeine produced a robust rise in cortisol across the test day (p < .0001); with 300 mg/day of prior daily intake tolerance was incomplete and cortisol remained elevated after the afternoon dose, while at 600 mg/day tolerance was more complete. The cortisol response was an acute, dose-linked elevation that was blunted — but not abolished — by habitual consumption; no body-composition or morphological outcomes were measured.
0.64
Double-blind crossover trial in 96 healthy adults (48 men, 48 women, mean age ~28) run over four weeks. In each week participants abstained from dietary caffeine for five days while taking capsules providing 0, 300, or 600 mg/day, then on day six received either placebo or 250 mg caffeine at 09:00, 13:00 and 18:00 while salivary cortisol was sampled eight times from 07:30 to 19:00. After abstinence, caffeine produced a robust rise in cortisol across the test day (p < .0001); with 300 mg/day of prior daily intake tolerance was incomplete and cortisol remained elevated after the afternoon dose, while at 600 mg/day tolerance was more complete. The cortisol response was an acute, dose-linked elevation that was blunted — but not abolished — by habitual consumption; no body-composition or morphological outcomes were measured.
Design Crossover Trial (0.75) × quality 0.85 = impact 0.64
View sourceMeta-Analysis
Neutral
Nakamura Y et al. · 2016StressSystematic review and meta-analysis of 13 studies (357 participants) examining the effect of caloric restriction on circulating cortisol. Pooled across all restriction types, caloric restriction significantly raised serum cortisol, but the effect was driven almost entirely by outright fasting; very-low-calorie diets and less severe low-calorie diets produced no significant increase. Meta-regression showed a negative association between cortisol and the duration of restriction, indicating the elevation is transient — cortisol rises in the initial period and falls back toward baseline after several weeks of continued restriction. What varied with the cortisol response was the *severity* of energy deprivation (fasting vs VLCD vs LCD). The review did not compare diets of differing food composition, so it cannot speak to whether what you eat matters — only to how hard you restrict.
0.60
Systematic review and meta-analysis of 13 studies (357 participants) examining the effect of caloric restriction on circulating cortisol. Pooled across all restriction types, caloric restriction significantly raised serum cortisol, but the effect was driven almost entirely by outright fasting; very-low-calorie diets and less severe low-calorie diets produced no significant increase. Meta-regression showed a negative association between cortisol and the duration of restriction, indicating the elevation is transient — cortisol rises in the initial period and falls back toward baseline after several weeks of continued restriction. What varied with the cortisol response was the *severity* of energy deprivation (fasting vs VLCD vs LCD). The review did not compare diets of differing food composition, so it cannot speak to whether what you eat matters — only to how hard you restrict.
Design Meta-Analysis (1.0) × quality 0.60 = impact 0.60
View sourceProspective Cohort
Neutral
Tsutsumi T et al. · 2024Journal of the Endocrine SocietyProspective 24-week observational study of 90 patients (72 analysed) newly started on oral prednisolone for conditions such as microscopic polyangiitis and IgA nephropathy, using objective facial measurements at five anatomical sites plus patient and physician assessment to define moon-like facies. Cumulative incidence of moon-like facies reached 37.6% by 24 weeks. The dominant predictor was dose: an initial prednisolone dose of at least 30 mg/day carried an odds ratio of 63.91 (95% CI 5.82–701.81), with female sex a secondary risk factor (OR 6.66). Patients who developed moon facies also had far higher rates of glucocorticoid-induced diabetes (48.0% vs 9.1%) and hypertension (36.8% vs 7.7%). The facial change was produced by sustained pharmacologic glucocorticoid doses roughly five- to ten-fold above normal daily cortisol output.
0.33
Prospective 24-week observational study of 90 patients (72 analysed) newly started on oral prednisolone for conditions such as microscopic polyangiitis and IgA nephropathy, using objective facial measurements at five anatomical sites plus patient and physician assessment to define moon-like facies. Cumulative incidence of moon-like facies reached 37.6% by 24 weeks. The dominant predictor was dose: an initial prednisolone dose of at least 30 mg/day carried an odds ratio of 63.91 (95% CI 5.82–701.81), with female sex a secondary risk factor (OR 6.66). Patients who developed moon facies also had far higher rates of glucocorticoid-induced diabetes (48.0% vs 9.1%) and hypertension (36.8% vs 7.7%). The facial change was produced by sustained pharmacologic glucocorticoid doses roughly five- to ten-fold above normal daily cortisol output.
Design Prospective Cohort (0.6) × quality 0.55 = impact 0.33
View sourceDouble-Blind RCT
Con
Soltani H et al. · 2019NutrientsRandomized, double-blind, controlled 8-week fully-fed feeding trial in 43 overweight and obese women, comparing a whole-food diet based on the Dietary Guidelines for Americans (56% carbohydrate, 18% protein, 26% fat; whole grains, low-fat dairy, vegetables, fruit, polyunsaturated fat) against a Typical American Diet (52% carbohydrate, 15% protein, 34% fat; added sugars, refined grains, saturated fat). There were no statistically significant differences between the two diet groups in salivary cortisol at baseline or after eight weeks. A secondary exploratory analysis found that within the whole-food group, the women who increased carbohydrate intake the most showed a blunted salivary cortisol response to an acute stress test (roughly 0.35–0.51 nmol/L lower at 30–120 minutes post-stressor).
0.47
Randomized, double-blind, controlled 8-week fully-fed feeding trial in 43 overweight and obese women, comparing a whole-food diet based on the Dietary Guidelines for Americans (56% carbohydrate, 18% protein, 26% fat; whole grains, low-fat dairy, vegetables, fruit, polyunsaturated fat) against a Typical American Diet (52% carbohydrate, 15% protein, 34% fat; added sugars, refined grains, saturated fat). There were no statistically significant differences between the two diet groups in salivary cortisol at baseline or after eight weeks. A secondary exploratory analysis found that within the whole-food group, the women who increased carbohydrate intake the most showed a blunted salivary cortisol response to an acute stress test (roughly 0.35–0.51 nmol/L lower at 30–120 minutes post-stressor).
Design Double-Blind RCT (0.85) × quality 0.55 = impact 0.47
View sourceCrossover Trial
Neutral
Lovallo WR et al. · 2005Psychosomatic MedicineDouble-blind crossover trial in 96 healthy adults (48 men, 48 women, mean age ~28) run over four weeks. In each week participants abstained from dietary caffeine for five days while taking capsules providing 0, 300, or 600 mg/day, then on day six received either placebo or 250 mg caffeine at 09:00, 13:00 and 18:00 while salivary cortisol was sampled eight times from 07:30 to 19:00. After abstinence, caffeine produced a robust rise in cortisol across the test day (p < .0001); with 300 mg/day of prior daily intake tolerance was incomplete and cortisol remained elevated after the afternoon dose, while at 600 mg/day tolerance was more complete. The cortisol response was an acute, dose-linked elevation that was blunted — but not abolished — by habitual consumption; no body-composition or morphological outcomes were measured.
0.64
Double-blind crossover trial in 96 healthy adults (48 men, 48 women, mean age ~28) run over four weeks. In each week participants abstained from dietary caffeine for five days while taking capsules providing 0, 300, or 600 mg/day, then on day six received either placebo or 250 mg caffeine at 09:00, 13:00 and 18:00 while salivary cortisol was sampled eight times from 07:30 to 19:00. After abstinence, caffeine produced a robust rise in cortisol across the test day (p < .0001); with 300 mg/day of prior daily intake tolerance was incomplete and cortisol remained elevated after the afternoon dose, while at 600 mg/day tolerance was more complete. The cortisol response was an acute, dose-linked elevation that was blunted — but not abolished — by habitual consumption; no body-composition or morphological outcomes were measured.
Design Crossover Trial (0.75) × quality 0.85 = impact 0.64
View sourceMeta-Analysis
Neutral
Nakamura Y et al. · 2016StressSystematic review and meta-analysis of 13 studies (357 participants) examining the effect of caloric restriction on circulating cortisol. Pooled across all restriction types, caloric restriction significantly raised serum cortisol, but the effect was driven almost entirely by outright fasting; very-low-calorie diets and less severe low-calorie diets produced no significant increase. Meta-regression showed a negative association between cortisol and the duration of restriction, indicating the elevation is transient — cortisol rises in the initial period and falls back toward baseline after several weeks of continued restriction. What varied with the cortisol response was the *severity* of energy deprivation (fasting vs VLCD vs LCD). The review did not compare diets of differing food composition, so it cannot speak to whether what you eat matters — only to how hard you restrict.
0.60
Systematic review and meta-analysis of 13 studies (357 participants) examining the effect of caloric restriction on circulating cortisol. Pooled across all restriction types, caloric restriction significantly raised serum cortisol, but the effect was driven almost entirely by outright fasting; very-low-calorie diets and less severe low-calorie diets produced no significant increase. Meta-regression showed a negative association between cortisol and the duration of restriction, indicating the elevation is transient — cortisol rises in the initial period and falls back toward baseline after several weeks of continued restriction. What varied with the cortisol response was the *severity* of energy deprivation (fasting vs VLCD vs LCD). The review did not compare diets of differing food composition, so it cannot speak to whether what you eat matters — only to how hard you restrict.
Design Meta-Analysis (1.0) × quality 0.60 = impact 0.60
View sourceProspective Cohort
Neutral
Tsutsumi T et al. · 2024Journal of the Endocrine SocietyProspective 24-week observational study of 90 patients (72 analysed) newly started on oral prednisolone for conditions such as microscopic polyangiitis and IgA nephropathy, using objective facial measurements at five anatomical sites plus patient and physician assessment to define moon-like facies. Cumulative incidence of moon-like facies reached 37.6% by 24 weeks. The dominant predictor was dose: an initial prednisolone dose of at least 30 mg/day carried an odds ratio of 63.91 (95% CI 5.82–701.81), with female sex a secondary risk factor (OR 6.66). Patients who developed moon facies also had far higher rates of glucocorticoid-induced diabetes (48.0% vs 9.1%) and hypertension (36.8% vs 7.7%). The facial change was produced by sustained pharmacologic glucocorticoid doses roughly five- to ten-fold above normal daily cortisol output.
0.33
Prospective 24-week observational study of 90 patients (72 analysed) newly started on oral prednisolone for conditions such as microscopic polyangiitis and IgA nephropathy, using objective facial measurements at five anatomical sites plus patient and physician assessment to define moon-like facies. Cumulative incidence of moon-like facies reached 37.6% by 24 weeks. The dominant predictor was dose: an initial prednisolone dose of at least 30 mg/day carried an odds ratio of 63.91 (95% CI 5.82–701.81), with female sex a secondary risk factor (OR 6.66). Patients who developed moon facies also had far higher rates of glucocorticoid-induced diabetes (48.0% vs 9.1%) and hypertension (36.8% vs 7.7%). The facial change was produced by sustained pharmacologic glucocorticoid doses roughly five- to ten-fold above normal daily cortisol output.
Design Prospective Cohort (0.6) × quality 0.55 = impact 0.33
View sourceDouble-Blind RCT
Con
Soltani H et al. · 2019NutrientsRandomized, double-blind, controlled 8-week fully-fed feeding trial in 43 overweight and obese women, comparing a whole-food diet based on the Dietary Guidelines for Americans (56% carbohydrate, 18% protein, 26% fat; whole grains, low-fat dairy, vegetables, fruit, polyunsaturated fat) against a Typical American Diet (52% carbohydrate, 15% protein, 34% fat; added sugars, refined grains, saturated fat). There were no statistically significant differences between the two diet groups in salivary cortisol at baseline or after eight weeks. A secondary exploratory analysis found that within the whole-food group, the women who increased carbohydrate intake the most showed a blunted salivary cortisol response to an acute stress test (roughly 0.35–0.51 nmol/L lower at 30–120 minutes post-stressor).
0.47
Randomized, double-blind, controlled 8-week fully-fed feeding trial in 43 overweight and obese women, comparing a whole-food diet based on the Dietary Guidelines for Americans (56% carbohydrate, 18% protein, 26% fat; whole grains, low-fat dairy, vegetables, fruit, polyunsaturated fat) against a Typical American Diet (52% carbohydrate, 15% protein, 34% fat; added sugars, refined grains, saturated fat). There were no statistically significant differences between the two diet groups in salivary cortisol at baseline or after eight weeks. A secondary exploratory analysis found that within the whole-food group, the women who increased carbohydrate intake the most showed a blunted salivary cortisol response to an acute stress test (roughly 0.35–0.51 nmol/L lower at 30–120 minutes post-stressor).
Design Double-Blind RCT (0.85) × quality 0.55 = impact 0.47
View sourceShowing the 4 strongest of 7 studies. Tap any node to expand its detail.
Evidence
AGAINST (2)
AGAINST Double-Blind RCTn=43CONFLICTED0.55 Soltani H, Keim NL et al. (2019)
No significant between-diet difference in salivary cortisol at baseline or 8 weeks; in a secondary exploratory analysis, women within the whole-food arm who raised carbohydrate intake most showed a blunted cortisol response to acute stress (~0.35-0.51 nmol/L lower at 30-120 min post-stressor)
Randomized, double-blind, controlled 8-week fully-fed feeding trial in 43 overweight and obese women, comparing a whole-food diet based on the Dietary Guidelines for Americans (56% carbohydrate, 18% protein, 26% fat; whole grains, low-fat dairy, vegetables, fruit, polyunsaturated fat) against a Typical American Diet (52% carbohydrate, 15% protein, 34% fat; added sugars, refined grains, saturated fat). There were no statistically significant differences between the two diet groups in salivary cortisol at baseline or after eight weeks. A secondary exploratory analysis found that within the whole-food group, the women who increased carbohydrate intake the most showed a blunted salivary cortisol response to an acute stress test (roughly 0.35–0.51 nmol/L lower at 30–120 minutes post-stressor).
Weighted 0.55 — n=43 analysed (22 vs 21), but the cortisol-responsiveness finding that gets cited rests on only ~14 per group who completed both stress tasks. Fully-fed 8-week feeding design is a real strength (intake was controlled, not self-reported). Against that: the headline positive is a SECONDARY EXPLORATORY analysis while the primary comparison was null; the two diets differed in several macronutrients at once, so the carbohydrate attribution is not cleanly isolated; women only, single site. Part-funded by the National Dairy Council and Campbell Soup Company -- an industry-funded positive finding, which warrants caution.
Funding: USDA Agricultural Research Service (2032-51530-022-00-D); National Dairy Council; Campbell Soup Company — the funder profits from this result.
Nutrients
AGAINST Narrative Review0.55 Clutter WE (2011)
Cushing's syndrome was confirmed in 0 of 369 weight-loss-programme patients and 0 of 201 diabetes-clinic patients, despite 23% screening abnormal on cortisol testing in each series
Clinical review of the diagnostic yield of screening for Cushing's syndrome in people with obesity. It reports two screening series with the same result: among 369 patients enrolled in a weight-loss programme, 23% had an abnormal cortisol screening test yet definitive evaluation confirmed Cushing's syndrome in none of them; among 201 consecutive diabetes-clinic patients, 23% again screened abnormal and again none proved to have the syndrome. The review concludes that in unselected obese patients the prevalence of Cushing's syndrome is very low and that the features shared with common obesity — weight gain, central adiposity, a rounded face, hypertension, depression — do not discriminate, whereas the catabolic signs (thin skin, easy bruising, wide purple striae, facial plethora, proximal muscle weakness, unexplained osteoporosis) do.
Weighted 0.55 — A narrative clinical review, so no pooled sample size and no systematic search — but it is independent, declares no conflicts, and reports two concrete screening series whose results agree with each other and bear directly on the claim. Rated at the top of the narrative-review band on directness; still a review, so the reader is taking the author's selection on trust.
Missouri Medicine
NEUTRAL (5)
NEUTRAL Crossover Trialn=960.85 Lovallo WR, Whitsett TL et al. (2005)
after abstinence, caffeine produced a robust rise in salivary cortisol across the test day (p<0.0001); tolerance was incomplete at 300 mg/day and more complete at 600 mg/day
Double-blind crossover trial in 96 healthy adults (48 men, 48 women, mean age ~28) run over four weeks. In each week participants abstained from dietary caffeine for five days while taking capsules providing 0, 300, or 600 mg/day, then on day six received either placebo or 250 mg caffeine at 09:00, 13:00 and 18:00 while salivary cortisol was sampled eight times from 07:30 to 19:00. After abstinence, caffeine produced a robust rise in cortisol across the test day (p < .0001); with 300 mg/day of prior daily intake tolerance was incomplete and cortisol remained elevated after the afternoon dose, while at 600 mg/day tolerance was more complete. The cortisol response was an acute, dose-linked elevation that was blunted — but not abolished — by habitual consumption; no body-composition or morphological outcomes were measured.
Weighted 0.85 — 96 adults analysed (48 men, 48 women) in a double-blind crossover with eight salivary cortisol samples per test day across four weeks — well powered and tightly controlled for a mechanistic trial, and independently funded (NIH/VA). The authors limit generality to younger adults and to a short 5-day tolerance-induction period. No body-composition or morphological outcomes were measured, so it is indirect evidence for any claim about those.
Funding: National Heart, Lung, and Blood Institute (NHLBI) and the US Department of Veterans Affairs
Psychosomatic Medicine
NEUTRAL Cross-Sectionaln=25270.80 Jackson SE, Kirschbaum C et al. (2017)
Hair cortisol correlated with body weight r = 0.102, BMI r = 0.101 and waist circumference r = 0.082 (all p <= 0.001) -- roughly 1% of shared variance -- and was higher in those whose obesity persisted across four years
Cross-sectional analysis of 2,527 men and women aged 54–87 in the English Longitudinal Study of Ageing, measuring hair cortisol concentration (an index of cumulative cortisol exposure over roughly the preceding two months) against adiposity. Hair cortisol was positively but very weakly correlated with body weight (r = 0.102), BMI (r = 0.101) and waist circumference (r = 0.082, all p ≤ .001), and was higher in participants with obesity and in those whose obesity persisted across four years. The correlations correspond to roughly 1% of shared variance; the authors explicitly note causality is undetermined and that persistent obesity could itself raise cortisol rather than the reverse.
Weighted 0.80 — Large (n=2527) population-based ELSA sample using a validated cumulative-exposure biomarker, independently funded with nothing to disclose -- strong execution for its design class. Its weaknesses are inherent to the design rather than the conduct: cross-sectional, so the direction of causation is undetermined (persistent obesity could itself raise cortisol), and the correlations are near-trivial in magnitude.
Funding: Cancer Research UK (C1418/A14133); ELSA cohort supported by NIA/NIH (R01 AG017644)
Obesity
NEUTRAL Meta-Analysisn=3570.60 Nakamura Y, Walker BR et al. (2016)
Caloric restriction raised circulating cortisol overall, but the effect was driven almost entirely by outright fasting; very-low-calorie and low-calorie diets produced no significant rise, and the elevation was transient
Systematic review and meta-analysis of 13 studies (357 participants) examining the effect of caloric restriction on circulating cortisol. Pooled across all restriction types, caloric restriction significantly raised serum cortisol, but the effect was driven almost entirely by outright fasting; very-low-calorie diets and less severe low-calorie diets produced no significant increase. Meta-regression showed a negative association between cortisol and the duration of restriction, indicating the elevation is transient — cortisol rises in the initial period and falls back toward baseline after several weeks of continued restriction. What varied with the cortisol response was the severity of energy deprivation (fasting vs VLCD vs LCD). The review did not compare diets of differing food composition, so it cannot speak to whether what you eat matters — only to how hard you restrict.
Weighted 0.60 — 13 studies and only 357 participants — a small pooled sample for a meta-analysis. Restriction protocols, durations and cortisol assays are heterogeneous, and the decisive subgroup findings (fasting vs VLCD vs LCD) each rest on a handful of studies. Independently funded.
Funding: British Heart Foundation (grant RG/11/4/28734)
Stress
NEUTRAL Prospective Cohortn=720.55 Tsutsumi T, Nakagomi D et al. (2024)
Cumulative incidence of moon-like facies 37.6% at 24 weeks; initial prednisolone >=30 mg/day OR 63.91 (95% CI 5.82-701.81); female sex OR 6.66
Prospective 24-week observational study of 90 patients (72 analysed) newly started on oral prednisolone for conditions such as microscopic polyangiitis and IgA nephropathy, using objective facial measurements at five anatomical sites plus patient and physician assessment to define moon-like facies. Cumulative incidence of moon-like facies reached 37.6% by 24 weeks. The dominant predictor was dose: an initial prednisolone dose of at least 30 mg/day carried an odds ratio of 63.91 (95% CI 5.82–701.81), with female sex a secondary risk factor (OR 6.66). Patients who developed moon facies also had far higher rates of glucocorticoid-induced diabetes (48.0% vs 9.1%) and hypertension (36.8% vs 7.7%). The facial change was produced by sustained pharmacologic glucocorticoid doses roughly five- to ten-fold above normal daily cortisol output.
Weighted 0.55 — 72 of 90 analysed over 24 weeks, single centre, with objective facial measurements at five sites rather than impression alone. The headline dose effect is directionally clear but extremely imprecise (95% CI 5.82-701.81), so the magnitude is not established. No funding statement; authors declare nothing to disclose.
Journal of the Endocrine Society
NEUTRAL Cross-Sectionaln=310.35 Rockall AG, Sohaib SA et al. (2003)
Visceral-to-subcutaneous fat ratio markedly elevated in Cushing's syndrome (men 1.175 +/- 0.59 vs 0.77 +/- 0.39; women 0.845 +/- 0.53 vs 0.38 +/- 0.19), abolishing the normal male-female difference in visceral fat
Observational CT study of abdominal fat distribution in 31 patients with active Cushing's syndrome (7 male, 24 female), with visceral and subcutaneous fat areas compared against published reference values for non-Cushingoid subjects. The visceral-to-subcutaneous fat ratio was markedly elevated in Cushing's patients of both sexes (men 1.175 ± 0.59 vs 0.77 ± 0.39; women 0.845 ± 0.53 vs 0.38 ± 0.19), and the normal male-female difference in visceral fat was abolished. Pathological cortisol excess therefore drives fat preferentially into the visceral compartment; the study describes patients with frank endogenous hypercortisolism, not healthy people with normal-range cortisol.
Weighted 0.35 — n=31 patients with active Cushing's syndrome, but the comparison group is published reference values from the literature rather than concurrently recruited controls, which is a serious bias source. Male subgroup is only n=7 despite carrying a headline finding. Single centre, retrospective. Severe indirectness for any claim about normal-range cortisol in healthy people: this is frank endogenous hypercortisolism. No funding statement retrievable.
European Journal of Endocrinology