The Association of Lifestyle and Mood with Long-Term Levels of Cortisol: A Systematic Review

Figures & data

Figure 1. PRISMA flow diagram of study selection.

Table 1. Summary of Studies Investigating Long-term Cortisol and Lifestyle Factors

Author (Year/Region)	Study type	Hypothesised intervention mechanism	Population and Study Size	Type of samples (indices if specified)	Time-frame cortisol was measured	Lifestyle factors and measures	Time-frame lifestyle was measured	Results
Chen et al. (2015/United States)	Observational	N/A	Caregivers of children with disabilities, n=85 (female=75, male=10)	Hair (HCC)	3 months	BMI (measured by researcher) Smoking (Self-identification of being smoker or not)	One-time assessment	HCC was not correlated with obesity status. No association between smoking status and HCC was found.
Faresjö et al. (2013/Sweden, Greece)	Observational	NA	General population, n=591 (female=321, male=270)	Hair (HCC)	3 months	Smoking (Self-identification of being daily smoker or not)	One-time assessment	No association between smoking status and HCC was found.
Garcia-Leon et al. (2018/Spain)	Observational	N/A	Pregnant women (n=62)	Hair (HCC)	3 trimester for pregnant participants	BMI (self-reported) Physical activity (self-reported regular/non-regular participant of physical activity)	One-time assessment	HCC was not correlated with BMI in pregnant women. Physical activity was significantly correlated with HCC in general sample.
Geng et al. (2016/China)	Observational	N/A	Participant were addicted to methamfphetamine, n=51 (female=51)	Hair (HCC)	1 month	BMI (measure not described)	One-time assessment	No significant association between HCC and BMI was found.
Gerber, Kalak et al.(2013/Switzerland)	Observational	N/A	General population, n=46 (female=26, male=20)	Hair (HCC)	3 months	Vigorous physical activity: assessed with an accelerometer worn around the hip. All activities with >5,724 counts/min were defined as vigorous.	7 consecutive days	Higher HCC was significantly associated with greater vigorous physical activity.
Heinze et al. (2016/UK)	Observational	N/A	n=30 were mental health help-seekers and n=28 were healthy controls, (female=52, male=6)	Hair (HCC)	6 months	BMI (measure not described) Smoking and alcohol consumption (Self-reported: Alcohol, Smoking and Substance Involvement Screening Test)	BMI: one-time assessment Smoking and alcohol consumption: over the past 3 months	No significant correlations between HCC and BMI. No significant associations between smoking and HCC either over the past 3 month or the 3 month before were found. No significant association between alcohol consumption and HCC over the past 6 month.
Larsen et al. (2016/Denmark)	Observational	N/A	Parents of children at risk of obesity, n=532 (female=301, male=231)	Hair (HCC)	1-2 months	BMI (measured by researcher)	One-time assessment	Higher HCC was significantly associated with higher BMI.
Manenscijn et al. (2017/The Netherlands)	Observational	N/A	n=89 were shift workers and n=33 were controls, (m=122)	Hair (HCC)	3 months	BMI (measured by researcher)	One-time assessment	In shift workers there was a significant correlation between elevated levels of HCC and increased BMI.
Qi et al. (2014/China)	Observational	N/A	Healthy women, n=39	Hair (HCC)	1 month	BMI (self-reported)	One-time assessment	No significant correlations between HCC and BMI.
Sumra and Schillaci (2015/Canada)	Observational	N/A	Healthy women, n=31	Urine, hair (HCC)	1 month for urinary cortisol; 3 months for HCC	Physical activity: self-reported frequency	frequency of exercise per week	Neither hair, nor mean urinary cortisol was correlated with any other primary variables including physical activity (correlation coefficient not reported).
Wang et al. (2016/United States)	Observational	N/A	Female veterans, n=82	Hair (HCC)	3 months	BMI (measured by researcher) Sleep quality (the Pittsburgh Sleep Quality Index)	BMI: One-time assessment Sleep quality: not described	No significant correlations between HCC and BMI. No significant relationship between self-reported sleep quality and HCC.
Batista et al. (2015/Brazil)	6-week Yoga intervention 50 minute sessions twice a week for 6 weeks. Postures included snake posutre, boat posture, yoga mudra, Karmasana, self massage and deep relaxation.	Yoga practice would reduce stress, as reflected by cortisol	General population n=22 (female=15, male=7)	Saliva (daily rhythm of cortisol production, morning and night-time values)	Saliva samples were collected at 7.30am, 9am, noon, 12pm for 6 weeks (4 samples before and after yoga over 6 weeks).	NA	NA	Over long-term, Yoga induced higher cortisol production in the morning and lower production in the evening.
Brown et al. (2002/United States)	Diet intervention Crossover design of diet treatment for 11 menstrual cycles in total. Baseline + diet (polyunsaturated fat, PUFA) +washout + diet (soy protein + high saturated fat); control (high saturated fat). Each treatment lasting two menstrual cycles, with a washout time of two menstrual cycles between each treatment.	N/A	Healthy women n=26	Blood	11 menstrual cycles in total. Blood samples were collected for 2 menstrual cycles for each treatment. 2-day consecutive blood sample for mid-follicular phase (day 7&8 or 8&9) and mid-luteal phase (day 21&22 or 22&23) after menses during the second cycle of each diet treatment period and the baseline month.	Meals were provided to the subjects each day. The subject’s free meal night food records were obtained each week during all diet treatments. Caloric intake was individualized for each subject. The standardised study diets were consumed with 37% of kilocalories as fat, 48% as carbohydrate, and 15% as protein. Only the type of fat and amount of soy protein were altered for the study diets.	NA	A high polyunsaturated fat diet may decrease cortisol concentrations by 81.8 nmol/l (p = 0.088) compared with a high saturated fat diet. Compared with the control diet, no significant changes of cortisol were found for the soy diet.
Chen et al. (2017/China (Taiwan))	20-week Yoga intervention Prospective randomized control trial with a longitudinal, repeated-measures design. Intervention was yoga (70 minutes, twice per week for 20 weeks in groups of 10-12 women).	Yoga practice would reduce stress, as reflected by cortisol	Pregnant women n=101 (female=101)	Saliva	Saliva samples were collected 10 minutes before yoga and immediately after and at same time points for control group at 16, 20, 24, 28, 32 and 36 weeks gestational age (12 samples per participant)	NA	NA	No significant long-term effect of yoga on salivary cortisol levels.
Coiro et al. (2007/Italy)	8-week Alcohol abstinence Participants exercised on an electrically braked cycle ergometer. An initial load of 50 W was increased by 50 W every 3 minutes until subjective exhaustion, for about 15 minutes. Alcoholic patients were tested at 4, 6 and 8 weeks of alcohol abstinence.	Not described	Alcoholics n=20 (male=20) and healthy controls	Blood	Blood samples were taken at 30 minutes before exercise, initiation of exercise, 10, 20, 30, 40, 50, 60 minutes. Alcoholic participants provided 8 samples at 4, 6, 8 weeks after abstinence. Healthy controls were only assessed once.	Abstinence from alcohol was monitored with daily measurement of ethanol levels in 24-hour urine samples.	8 weeks	Serum cortisol levels significantly rose during exercise in the healthy controls and peaked after 30 minutes after the exercise task. However, the cortisol levels among abstinent alcohol patients did not significantly change during the exercise task, after 4-weeks of alcohol abstinence. At 8 weeks post-abstinence, the cortisol responses to exercise task for the abstinent alcoholic patients were significantly higher than 2 weeks previously and was no longer distinguishable from non-alcoholic participants.
Diment et al. (2012/UK)	8-week Supplemented diet intervention Normal (habitual) diet plus mixed nutritional supplement.		Healthy male soldiers n=30	Saliva	8 weeks Baseline, pre-field exercise at the end of week 6; post-field exercise at week 8.	A 24-hour dietary record from each participant on the day before their visits with study staff was acquired.	NA	No intervention effect on plasma cortisol was found.
Goldberg et al. (2014/Canada)	7-week mindfulness smoking cessation intervention: mindfulness training or freedom from smoking enhanced Participants were randomised to 1 of 2 smoking cessation interventions: mindfulness training for smoking versus freedom from smoking enhanced. Both groups had 8 meetings (over 7 weeks) and made quit attempts 1 month after the 1^st meeting.	Negative affect and increased cortisol output under stress were related to increased risk of smoking relapse and that stress reduction interventions designed to cultivate emotion regulation skills and target social-emotional processes could aid smoking cessation and may be marked by lower levels of cortisol in hair.	Smokers n=135 (female=55)	Hair (HCC)	A 3cm hair sample was taken from each participant at 1-month post-quit study visit, to determine the HCC (1) during the month before the intervention, and (2) during the month post-quit	NA	NA	Participants’ HCC was lower at 1-month post-quit compared to the month before the intervention. Intervention participants who stayed abstinent at the 1-month post-quit study visit had significantly lower levels of HCC compared with that in participants who had relapsed post-quit. However, differences were not found between two intervention groups for either decreases in HCC or the prevalence of smoking abstinence at post-quit.
Lovallo et al. (2005/United States)	4-week crossover caffeine stimulation intervention On each week, participants were instructed to abstain for 5 days from dietary caffeine but took caffeine capsules, followed by a test day, and a crossover day to buffer sudden changes in intake between study weeks. On the test day, each participant was administrated caffeine capsules three times and underwent a stress challenge between the first two administrations. In order to investigate the effect of different levels of caffeine intake on cortisol responses to caffeine challenges, three different maintenance doses were administrated to participants, to mimic a range of consumption commonly found in the US diet: none (0 mg/day), moderate (300 mg/day), high (600 mg/day).	NA	General population n=96 (female=48, male=48)	Saliva	Eight saliva samples were collected from each participant on the test day each week, before and after the stress challenge and each administration of caffeine capsules.	NA	NA	Taking caffeine capsules on the test day after five days of abstinence resulted in a significant elevation of cortisol throughout the entire day. A significant elevation of cortisol was only seen in the group that had moderate caffeine intake of 300mg, but not 600mg, for five days. The elevation of cortisol in the 300mg group lasted only for about 6 hours on the test day.
Mangine et al. (2017/United States)	8-week resistance training Participants randomly assigned to high intensity low volume training or high volume moderate intensity training groups. Both groups required at least 28 training sessions in 8 weeks. Food intake was monitored via 3-day food diaries collected weekly.	Not described	Healthy resistance-trained men n=33	Blood (AUC)	Blood samples were collected at week 1 and 8, before exercise and immediately after exercise.	NA	NA	At week 1 of training, greater cortisol levels in response to exercise in the high-volume moderate-intensity training group were observed, compared to the high-intensity low-volume group. The difference in cortisol levels between groups was still significant at week 8 after training.
Miller et al. (2015/Australia)	5-week sleep restriction therapy Sleep restriction therapy comprised of one 40-minute face to face session which set a minimum time in bed and go to bed threshold time. Minimum time in bed and a ‘go-to-bed’ threshold time were set for each participant at the beginning of the intervention. All participants were instructed to not go to bed until after the threshold time and to only go to bed if they felt ‘sleepy tired’. Up to 5 weekly phone calls were made to participants to review the minimum time in bed, which was modified weekly for each individual.	The sleep-restriction therapy would reduce early night plasma cortisol	People with insomnia n=11 (female=9, male=12)	Blood (mean plasma cortisol concentration)	Blood samples were taken intravenously every half hour from 22.00 to 06.00 on both nights of data collection at baseline and after the intervention.	Other sleep-related variables measured: Sleep onset latency; wake-time after sleep-onset; Insomnia Severity Index; Sleep Quality; Number of Awakenings; Total Sleep Time; Time in Bed	NA	Cortisol results from six patients showed that cortisol levels in the early morning were significantly higher post-intervention compared to baseline but no changes were observed in the pre-sleep phase or early part of the night at post-intervention.
Papacosta et al. (2013/Greece)	5-week exercise intervention Week 1 participants trained as normal. Weeks 2-3 participants trained at a higher intensity. Weeks 4-5 training was tapered to half of the normal training intensity (week 1 intensity).	Not described	Male judo athletes n=11	Saliva	2 saliva samples were collected at the end of each week during week 1-3, and everyday during week 4-5.	NA	NA	Evening cortisol fell below baseline at the beginning of the tapering week and then returned to baseline by the end of the taper training.
Sinha et al. (2011/United States)	Each participant who were part of a 1- month in-hospital abstinence intervention underwent the stress and alcohol cue tasks twice: at baseline and at 1-month-abstinence.	Not deescribed	Alcoholics n=72 (female=41, male=31)	Blood	Blood samples were taken from participants each time before and after the tasks, at baseline and follow-up at week 5.	NA	NA	Although there were no significant differences at baseline, alcoholic participants showed a lack of cortisol response to both stress and alcohol cues compared with non-alcoholic controls at 1-month-abstinence.
Tam et al. (2014/United States, Australia)	24-week exercise and Diet Intervention Randomised study with four groups: control (weight maintenance diet based on the American Heart Association Step 1 diet); calorie restriction 25%; calorie restriction 12.5% + exercise (supervised aerobics); low calorie diet until achieved 15% weight loss.		Overweight individuals n=46 (female=20, male=26)	Saliva (morning cortisol and diurnal cortisol) Morning cortisol was defined as the mean cortisol concentration at 08:00 and 08:30. Diurnal cortisol was calculated as the mean of the 8 cortisol measures across the day.	6 months (8 samples at baseline and end of intervention) Saliva samples collected at 8 time points over the day: 8:00, 8:30, 11:00, 11:30, 12:30, 1300,16:00 and 16:30.	Body composition (measured using dual x-ray absorptiometry and abdominal fat distribution by multi-slice computed tomography)	NA	At baseline and 24 weeks post caloric restriction, there were no associations between morning/diurnal cortisol levels and body composition (fat mass, visceral adipose tissue), in the whole group, or when groups were examined separately (data not shown).
Tanskanen et al. (2011/Finland)	7-week Exercise (military training) Participants divided into 3 categories according to levels of physical activity before military service, and then randomly selected 19 from each category.	Not described	Healthy male soldiers n=57	Blood	3 blood samples were taken at baseline and 3 at the end of intervention. Blood drawn from an antecubital vein after an overnight fast, 2 hours after breakfast, and immediately after exercise.	NA	NA	They reported that basal cortisol decreased from week 4 to week 7 of the training. Basal serum cortisol levels after 8 weeks of training were significantly higher in participants who were overreaching, compared to non-overreaching participants.
Timon et al. (2013/Spain)	8-week strength training Strength training program during 8 weeks with sessions 3 times a week.	Not described	Healthy women n=20	Urine	The first morning urine was collected from all subjects in three different moments across the menstrual cycle. Urine samples were taken the 1^st –2^nd day (menstruation phase), the 11^th –12^th day (follicular phase) and the 21^st–22^nd day (luteal phase). These samples were collected both the month before training and the month after training.	NA	NA	Significant changes in cortisol in urine following the training were not observed.
Torres et al. (2008/Australia)	4-week Diet (crossover design: 2-week control and 4-week diet, followed by 2-week control and 4-week diet) (mood) Randomised cross over. Participants received two (of 3) 4 week diets (OD diet, high in potassium, magnesium and fibre and low in sodium and sat fat; LNAHK diet was lower in sodium and calcium; HC, high calcium), each preceeded by a 2 week control diet. Subjects were instructed to follow the diets. During the CD and all test diet periods, a maximum of four Caffeine-containing drinks (e.g. cola drinks, coffee and tea) and two standard alcoholic drinks (10 g alcohol) were permitted daily.		General population n=94 (female=38, male=56)	Saliva (AUC)	10 weeks (3 samples at baseline and end of study) Saliva samples were collected at the end of each dietary period (4 weeks of diets, 2 weeks of control). A saliva sample was collected at approximately 1200, 1600 and 2000 hours			Higher cortisol levels were weakly associated with greater vigour, lower fatigue, and higher levels of urinary potassium and magnesium. However, the overall effects of interventions on cortisol were not reported.

Table 2. Summary of Studies Investigating Long-term Cortisol and Mood

Author (Year/Region)	Study type	Hypothesised intervention mechanism	Population and Study Size	Type of samples (indices if specified)	Time-frame cortisol was measured	Mood factors and measures	Time-frame mood was measured	Results
Booij et al. (2015/The Netherlands)	Observational	N/A	Adult with and without depression (n=15/15), n=30 (female=22, male=8)	Saliva (daily slope of cortisol)	3 times a day (in the morning, six hours later in the afternoon, six hours later in the evening) for 30 days	Depression (the Beck Depression Inventory; depressed individuals were identified using the DSM-IV diagnose of Major Depressive Disorder)	Unclear	Cortisol in saliva was higher, and the daily cortisol slope steeper in the depressed participants compared to the non-depressed ones.
Braig et al. (2016/Germany)	Observational	N/A	Women shortly after delivery, n=768	Hair (HCC^a)	3 months	Chronic stress (the screening scale of the Tirer Inventory of Chronic Stress) Anxiety (the Pregnancy-Related Anxiety Questionnaire; the Hospital Anxiety and Depression scale) Depression (the Hospital Anxiety and Depression scale)	Over the past 3 months	No association between chronic stress and HCC. No association between anxiety and HCC. No association between depression and HCC.
Byun et al. (2013/United States)	Observational	N/A	Caregivers of stroke survivors, n=63 (female=42, male=21)	Saliva (indices not specified)	Over 4 weeks (T1=within 2 weeks post-stroke, T2=6 weeks post-stroke)	Perceived stress (Perceived Stress Scale) Depression (the Patient Health Questionnaire)	Over the past month	Salivary cortisol on waking was not correlated with perceived stress but salivary cortisol in the evening had a positive correlation with higher levels of perceived stress. Salivary cortisol was not significantly correlated with depression at the beginning of the study but elevated cortisol levels in saliva in the evening were significantly correlated with greater self-reported depressive symptoms four weeks later.
Caparros-Gonzalez et al. (2017/Spain)	Observational	N/A	Pregnant women, n=44	Hair (HCC)	9 months (over trimester 1, 2, and 3)	Depression (the Edinburgh Postnatal Depression Scale)	At postpartum but time-frame unclear	Higher HCC during 1^st and 3^rd trimester was prospectively associated with greater postpartum depression.
Chen et al. (2015/United States)	Observational	N/A	Caregivers of children with disabilities, n=85 (female=75, male=10)	Hair (HCC)	3 months	Perceived stress (Perceived Stress Scale)	Over the past month	No association between perceived stress and HCC.
Cieslak et al. (2011/United States)	Observational	N/A	People with PTSD^f, N=30 (F=17, M=13)	Saliva (AUCg, AUCi)	Saliva samples from participants four times a day (upon awakening, 1h, 4h, 12h after waking) at 1 week, 1 month, and 3 months after a traffic accident (i.e., trauma).	PTSD (the Impact of Event Scale)	At 1 week, 1 month, and 3 months after the accident	The impact of trauma at 1-month post-accident was negatively associated with AUCi (reflecting changes in daytime cortisol secretions) at 3 months and the impact of trauma at 3 months was positively associated with AUCg (reflecting total daytime cortisol release) at 1-month post-accident.
Faresjö et al. (2013/Sweden, Greece)	Observational	N/A	General population, n=591 (female=321, male=270)	Hair (HCC)	3 months	Perceived stress (Perceived Stress Scale) Anxiety (the Hospital Anxiety and Depression scale) Depression (the Hospital Anxiety and Depression scale)	Perceives stress: over the past month Anxiety and depression: unclear	No association between perceived stress and HCC. No association between anxiety and HCC. Higher HCC was significantly correlated with greater depression.
Geng et al. (2016/China)	Observational	N/A	Participant were addicted to methamphetamine, n=51 (female=51)	Hair (HCC)	1 month	Perceived stress (Perceived Stress Scale) Implicit perceived stress (Implicit Association Test)	Over the past month	No association between self-reported perceived stress measured by Perceived Stress Scale and HCC. Higher HCC was significantly associated with greater implicit perceived stress.
Gerber et al. (2013/Switzerland)	Observational	N/A	General population, n=46 (female=26, male=20)	Hair (HCC)	3 months	Perceived stress (Perceived Stress Scale) Depression (the Beck Depression Inventory)	Perceived stress: over the past month Depression: unclear	Higher HCC was significantly correlated with lower perceived stress, but the correlation was weak. Lower HCC was significantly correlated with greater depression.
Giesbrecht et al. (2013/Canada)	Observational	N/A	Pregnant women, n=85	Saliva (Cortisol awaking levels, CAR, diurnal slope)	3 trimesters of pregnancy (saliva samples were collected 5 times per day over 2 consecutive days at approximately 14, 20, 32 weeks of gestation)	Distress (the Profile of Mood States)	At each sampling point over the same time frame.	Higher salivary cortisol was significantly associated with distress.
Grassi-Oliveira et al. (2012/Brazil)	Observational	N/A	Crack-cocaine-dependent women, n=23	Hair (HCC)	3 months	Mood impact of stressors (semi-structured interview with the Psykel Life Events Scale)	Over the past 90 days	Higher HCC was significantly correlated with the number of stressful events exposure in the previous 90 days and 30 days.
Heinze et al. (2016/UK)	Observational	N/A	n=30 were mental health help-seekers and n=28 were healthy controls, (female=52, male=6)	Hair (HCC)	6 months	Perceived stress (Perceived Stress Scale) Anxiety (the Overall Anxiety Severity and Impairment Scale) Depression (the Quick Inventory of Depression Symptoms; the Kessler Psychological Distress Scale)	Perceive stress: Over the past month Anxiety: 3 months within the 6 months when cortisol was measured Depression: over the past 7-30 days.	No association between perceived stress and HCC. No association between anxiety and HCC. No association between depression and HCC.
Hoffman et al. (2017/United States)	Observational	N/A	Pregnant women, n=92	Hair (HCC)	3 trimesters (at 16, 28, 40 weeks of gestation)	Perceived stress (Perceived stress scale) Anxiety (the State-Trait Anxiety Inventory) Depression (the Center for Epidemiologic Studies – Depression Scale)	3 trimesters (at 16, 22, 28, 34, 40 weeks of gestation for the past month at each time)	Higher HCC was prospectively correlated with higher levels of perceived stress at 40 weeks of gestation. Higher HCC at 2-trimester was significantly correlated with perceived stress at 16 weeks of gestation. Greater anxiety measured at 16 and 28 weeks of pregnancy was correlated with higher HCC in both the second and the third trimester. Higher HCC during the 1^st trimester was prospectively associated with greater depression at 40 weeks of gestation, higher HCC during the 2^nd-trimester was significantly associated with greater depression at 16 and 28 weeks of gestation, higher HCC at the 3^rd-trimester was significantly associated with depression at 16 and 28 weeks of gestation.
Kramer et al. (2009/Canada)	Observational	N/A	Pregnant women, n=651	Hair (HCC)	9 months	Perceived stress (Perceived Stress Scale) Anxiety (a 4-item scale) Depression (the Center for Epidemiologic Studies – Depression Scale)	Perceived stress: at 24-26 weeks of gestation for the past month. Anxiety: at 24-26 weeks of gestation Depression: at 24-26 weeks of gestation	No association between perceived stress and HCC. No association between anxiety and HCC. No association between depression and HCC.
O’Brien et al. (2013/United States)	Observational	N/A	General population, N=135 (F=88, M=47)	Hair (HCC)	3 months	Perceived stress (Perceived stress scale) Mood impact of stressors (chaos in the neighbourhood in the past year, City Stress Index; chaos/order of the home environment, Chaos, Hubbub, and Order Scale)	Perceives stress: Over the past 3 months Mood impact of stressors: not specified	No association between perceived stress and HCC. No association between HCC and chaos either in the neighbours or in the home.
Qi et al. (2014/China)	Observational	N/A	Healthy women, n=39	Hair (HCC)	1 month	Work stress (the Effort-Reward Imbalance Scale)	Over the past month	Neither the effort nor the reward scores were correlated with HCC. However, the effort-reward imbalance (1.83*effort-score/reward-score) at work was positively correlated with HCC.
Schalinski et al. (2015/Germany)	Observational	N/A	Women with stress-related disorders, n=45	Hair (HCC)	3 months	Depression (the clinical-administrated Hamilton Rating Scale) PTSD (the Clinician-Administered PTSD Scale)	Unclear	No association between depression and HCC. No association between PTSD and HCC.
Stoppelbein et al. (2015/United States)	Observational	N/A	Mothers of children diagnosed with cancer, n=27	Saliva (AUCg)	Saliva samples were collected between 3-5 pm on the first day, and between 6-8 am and 3-5 pm on the following day. This saliva collection protocol was repeated monthly for 12 months.	PTSD (the Posttraumatic Stress Disorder Checklist)	Over 12 months	Higher levels of cortisol were significantly associated with one type of self-reported PTSD symptoms: numbing symptoms
Sumra and Schillaci (2015/Canada)	Observational	N/A	Healthy women, n=31	Urine, hair (HCC)	1 month for urinary cortisol; 3 months for HCC	Perceived stress (perceived stress scale)	Over the past month	No association between perceived stress and HCC.
Ullmann et al. (2017/Germany)	Observational	N/A	General population, N=40 (F=22, M=18)	Hair (HCC)	3 months	Mood impact of stressors (mental burdening events, the Recalled Parental Rearing Behaviour Questionnaire)	Over the past 3 months	Those with mental burdening in the previous 3 months had higher cortisol compared with those without mentally burdening events.
Wang et al. (2016/United States)	Observational	N/A	Female veterans, n=82	Hair (HCC)	3 months	Depression (the Center for Epidemiologic Studies – Depression Scale)	Unclear	No association between depression and HCC.
Wikenius et al. (2016/Norway)	Observational	N/A	Pregnant women, n=181	Hair (HCC)	1-3 month (Second trimester)	Depression (the Edinburgh Postnatal Depression Scale)	Over the past week	No association between depression and HCC.
Bergen-Cico et al. (2014/United States)	4-week mindfulness intervention Weekly in person 90 minute mindfulness based meditation group sessions for 4 consecutive weeks. Participants were given a book and audio CD with guided practice.	Engagement in the stress-reduction intervention would reduce diurnal cortisol output and CAR among veterans with PTSD	Veterans n=40 (female=4, male=36)	Saliva(AUCg, AUCi, CAR)	Saliva samples were collected at waking, 45min after waking, 8hrs, 12hrs, and bedtime each day for 2 consecutive days, before treatment at baseline and after intervention.	Depression (Patient Health Questionnaire) PTSD (the Military PTSD checklist).	At baseline only	Significant changes in cortisol levels which was associated with engagement and number of sessions completed of the intervention. Participants who completed all four sessions had a significant decrease in the CAR and increase in AUC, compared with the control group. Changes in either PTSD or stress were not described.
Brown et al. (2016/United States)	8-week mindfulness intervention Participants attended weekly classes of 1.5-2 hours in length.	The intervention was to reduce psychological and neuroendocrine markers of stress.	Adult caregivers n=38 (female=32,male=6)	Saliva (diurnal cortisol)	Saliva samples were collected at 6 time point a day at 3 study periods (pre- and post-intervention, 3 month follow up).	Perceived stress (Perceived Stress Scale) Mood (Profile of Mood States)	NA	The mindfulness intervention significantly improved participants’ perceived stress levels and mood. However, the improvements in mood, perceived stress, and burden among participants at 3-month follow-up were not significantly different between the mindfulness intervention and social support group. Neither conditions showed changes in diurnal cortisol response curve across the study time-frame.
Daubenmier et al. (2011/United States)	4-month mindfulness intervention Twenty-four participants were asked to attend a weekly 2.5h class for nine weeks and a 7-hour silent day of guided meditation practice after class 6.	Mindfulness training would enhance awareness of and responsiveness to bodily sensations and reduce psychological stress, emotional eating, and cortisol secretion, which would lead to reductions in abdominal fat	Stressed women n=47	Blood; Saliva(CAR, cortisol slope)	Saliva sampled were collected upon awakening, 30 minutes after and before bed, on four days before and after the intervention. Blood samples were collected pre- and post-intervention.	Mindfulness (the Kentucky Inventory of Mindfulness Skills) Stress (Wheaton Chronic Stress Inventory; Perceived Stress Scale) Anxiety (The State-Trait Anxiety Scale)	NA	Participants in the intervention group showed significant improvements in mindfulness, anxiety and external-based eating after the intervention, compared to control participants. However, differences between the intervention and control groups were not significant for emotional eating, the average cortisol awakening response, overall weight, abdominal fat, or fat distribution over time. Neither groups showed significant changes in the cortisol slope or morning serum cortisol concentrations. Improvement of mindfulness, chronic stress, and Subgroup of obese participants in the intervention group showed significant reductions in the cortisol awakening response and maintained body weight while obese participants in the control group had a stable cortisol awakening response and gained weight.
Holt-Lundstad et al. (2008/United States)	4-week support enhancement intervention Twenty couples (n=40) in the intervention group were asked to attend two 1-hour group sessions for couple contact enhancement techniques. These couples were also instructed to practice the techniques for 30 minutes, 3 times per week for 4 weeks. A behaviour monitoring control group was designed where subjects were instructed not to change anything about their normal behaviours with their spouse and to simply keep a diary of their physical affection and mood.	Warm touch enhancement may improve mood and stress and result in reduced cortisol levels.	n=36 couples (female=36, male=36)	Saliva (AUC)	Saliva samples were collected five times throughout a day at pre- and post-intervention.	Unclear	Unclear	No significant intervention effect on salivary cortisol.
Huang et al. (2012/UK)	5-week stress reduction acupuncture intervention The intervention group received weekly acupuncture sessions for 5 weeks. The relaxation-only group received weekly relaxation (without acupuncture) sessions for 5 weeks. The waitlist control group did not receive any treatment for 5 weeks.	Traditional Chinese acupuncture to treat chronic stress could change diurnal salivary cortisol profile in healthy individuals.	Stressed individuals n=18 (female=15, male=3)	Saliva (CAR, daily profile)	Saliva samples were collected four times a day on 9 days pre-, post-intervention and at follow-up after 2 weeks.	Perceived stress (Perceived Stress Scale)	NA	Perceived stress was reduced in all groups however the differences between pre- and post-intervention and between- and within-groups were not significant. Non-significant increased CAR was seen for both the intervention and relaxation-only groups. No significant changes in the decline in cortisol or the daily average cortisol concentrations were found between groups.
Jacobs et al. (2013/United States)	8-week mindfulness meditation intervention The intervention group met twice per day for 1-hour sessions for meditation practice.	Whether self-reported mindfulness, as well as changes in mindfulness following the intervention, were related to resting cortisol. Direction of the intervention effect on cortisol not hypothesised.	General population n=60 (female=32, male=28)	Saliva (p.m. cortisol^e)	Saliva samples were collected for 3 consecutive days at pre-retreat (2 weeks after arrival) and post-retreat (10 weeks after arrival). Samples were collected in the afternoon (2 hr after lunch) and at bedtime (right before sleep).	Self-reported mindfulness (the 37-item Five Factor Mindfulness Questionnaire)	NA	Self-reported mindfulness significantly increased after the intervention. No significant changes were found in cortisol levels. Changes in mindfulness were inversely associated with changes in resting p.m. cortisol levels.
Leivadi et al. (1999/United States)	5-week massage intervention Participants randomised to massage received a 30-minute session twice a week for 5 weeks. The relaxation control group received a 30-minute session twice a week for 5 weeks.	Massage therapy would improve mood, decrease cortisol and other stressors such as anxiety and neck, shoulder, and back pain.	General population n=30 (massage group=15, relaxation group=15)	Saliva	Saliva samples were collected twice a day pre- and post-intervention.	Anxiety (State Anxiety Inventory) Mood (Profile Of Mood States)	NA	Both groups reported improvements in mood and anxiety after sessions. Changes in mood, anxiety, or cortisol levels were not observed over time. Participants reported reduced n neck, shoulder, and back pain after the treatment sessions.F.
Limm et al. (2011/Germany)	One-year work-site stress-management intervention The intervention consisted of two full-day group sessions followed by two refresher sessions within the next 8 months. At each group session, participants were asked to remember and share individual stressful situations at work. Several tools to cope with difficult and stressful situations were introduced by the trainer at the group session.	Hypothesised intervention effects on cortisol not described.	General population n=174 (female=22, male=152)	Saliva	Saliva samples to measure cortisol were collected seven times throughout a day both at pre-intervention and at 1-year.	Perceived stress reactivity (Stress Reactivity Scale) Effort-reward imbalance (the Effort-Reward Imbalance Scale) Anxiety and depression (the Hospital Anxiety and Depression Scale)	NA	Perceived stress reactivity reduced in both groups but a significantly greater reduction was seen in the intervention group at 1-year follow-up. No intervention effect on cortisol was observed. Improvements were seen for anxiety, depression, effort-reward imbalance and while greater in the intervention group compared to the control group, they were not significant.
Lynch et al. (2011/UK)	8-week mindfulness-based coping intervention Participants attended one 1.5h meeting per week for eight consecutive weeks.	Whether the intervention reduce stress which may be reflected by the biomarkers of stress. Direction of the intervention effect on cortisol not hypothesised.	General population n=60	Saliva	Saliva samples were collected 8 times throughout the day for 2 consecutive days at pre- and post-intervention.	Mindfulness (Five Factor Mindfulness Questionnaire) Perceived stress (Perceived Stress Scale) Anxiety and depression (the Hospital Anxiety and Depression Scale)	NA	Mindfulness intervention significantly improved perceived stress, anxiety, depression among participants, along with increased self-reported mindfulness. However, cortisol levels did not significantly change after the intervention.
Marcus et al. (2003/United States)	8-week mindfulness-based stress reduction intervention Participants attended a mindfulness meditation session for 2.5 hours per week for eight consecutive weeks. They were also asked to practice mindfulness outside the sessions for 45-60 minutes for 6 days per week.	Whether the intervention reduce stress which may be reflected by the biomarkers of stress. Direction of the intervention effect on cortisol not hypothesised.	People in theraputic community n=21 (female=18, male=3)	Saliva (AUC)	Saliva samples were collected at 0, 30, 45 and 60 minutes after awakening on the first morning pre intervention and post intervention.	Perceived stress (Perceived Stress Scale)	NA	A non-significant decrease in self-reported perceived stress after intervention among participants was found. The awakening salivary cortisol levels were significantly lower following the intervention.
Sannes et al. (2015/United States)	10-week cognitive behavioural stress management The intervention involved 10 two-hour sessions in groups of 4-10 on a weekly basis, which included education on awareness of the effects of stress, cognitive reframing, cognitive coping skills training etc.	Women in the intervention group would demonstrate lower intraindividual cortisol variability (i.e., the degree to which an individual’s cortisol output may be erratic on a given day).	Healthy women at risk of breast cancer n=91	Saliva (cortisol slope, AUCg, CAR, individual cortisol variability)	Saliva samples were collected 5 times a day for 2 consecutive days at pre- and post-intervention, and at 1-month follow-up.	Unclear	Unclear	Lower levels of intraindividual cortisol variability among intervention participants were found, compared to the waitlist control group. Improvements towards a steeper cortisol slope and less overall cortisol output (area under the curve with respect to ground, AUCg) were also observed in the intervention group. Significant changes in other cortisol indices were not found.
Tsiouli et al. (2014/Greece)	8-week stress management Intervention group were trained in progressive muscle relaxation and diaphragmatic breathing for 37 minutes. They were asked to perform these techniques twice a day for 8 weeks, and asked to adapt a healthy lifestyle.	The intervention could reduce perceived and parenting stress, increase internal locus of control, promote healthy lifestyle, and normalise cortisol levels. However the direction of changes in cortisol was not specified.	Caregivers of children and adolescents with type 1 diabetes n=80	Saliva (diurnal cortisol)	Samples collected at 5 time points throughout a day pre- and post-intervention.	Lifestyle (questionnaire including quality of sleep, eating habit, smoking habits, drug use, physical activity etc.) Perceived stress (Perceived Stress Scale) Parental Stress (Parenting Stress Index-Short Form)	NA	No statistically significant improvement was found in cortisol levels in either the intervention or the control group. Stress was significantly reduced in the intervention group. Improvement in lifestyle factors such as sleep, stress-related symptoms, and eat habits were improved in both groups.

Figure 2. Random effect meta-analysis for the correlation between long-term levels of cortisol and BMI.

Figure 3. Random effect meta-analysis for the correlation between long-term levels of cortisol and physical activity.

Figure 4. Random effect meta-analysis for the correlation between long-term levels of cortisol and perceived stress.

Figure 5. Random effect meta-analysis for the correlation between long-term levels of cortisol and mood impact of stressors.

Figure 6. Random effect meta-analysis for the correlation between long-term levels of cortisol and other indices of stress.

Figure 7. Random effect meta-analysis for the correlation between long-term levels of cortisoland anxiety.

Figure 8. Random effect meta-analysis for the correlation between long-term levels of cortisol and depression.

Data Availability Statement

All available data including the data extraction tables for all included studies, and raw data for meta-analysis are available here: https://osf.io/36fnt/