Abstract
Context
Biological stress is related to cardiovascular disease in adults. The associations of stress with cardiovascular and metabolic diseases may originate in childhood.
Objective
This work aims to examine the associations of hair cortisol concentrations at age 6 years with cardiometabolic risk factors at ages 6 and 10 years.
Methods
Cortisol concentrations were measured in hair of 6-year-old children (n = 2598) participating in the Generation R Study, a population-based prospective cohort study in Rotterdam, the Netherlands. Main outcome measures included blood pressure, heart rate, concentrations of insulin, glucose, lipids, and C-reactive protein in blood at ages 6 and 10 years.
Results
Higher hair cortisol concentrations at age 6 years were associated with higher systolic blood pressure at age 10 years (difference 0.17 SD score; 95% CI, 0.03-0.31). The association attenuated into nonsignificance after adjustment for childhood body mass index (BMI) at age 6 years. Higher hair cortisol concentrations at age 6 years were associated with an increase in total and low-density lipoprotein cholesterol between ages 6 and 10 years but not with those measurements at age 6 or 10 years. Hair cortisol concentrations were not associated with other cardiometabolic risk factors at age 6 or 10 years.
Conclusion
Hair cortisol concentrations were not independent of BMI associated with cardiometabolic risk factors at 6 or 10 years. The associations of biological stress with cardiometabolic risk factors may develop at later ages.
Keywords: hair cortisol, child, cardiometabolic risk, blood pressure, heart rate, lipids, cholesterol, insulin, glucose, C-reactive protein
Stress is associated with cardiometabolic disease in adults (1, 2). Results from a study among 136 637 individuals showed that stress-related disorders were robustly associated with multiple types of cardiovascular disease, such as hypertensive diseases and heart failure (3). Similarly, a prospective cohort study, among 10 308 men and women, reported that those with chronic work stress were twice as likely to develop metabolic syndrome (4). It has also been suggested that long-term exposure to elevated cortisol concentrations may lead to long-term physiological alterations compromising the anatomy and function of the cardiovascular and metabolic systems (5, 6). Cortisol concentrations measured in saliva, serum, and urine are subject to situational and intraindividual fluctuations (7). Hair cortisol concentrations reflect long-term cumulative cortisol concentrations and are therefore a useful biomarker of long-term systemic cortisol exposure, which is mainly determined by hypothalamic-pituitary-adrenal-axis activity (7, 8). A recent review among 11 cross-sectional studies shows positive associations of hair cortisol with adverse cardiometabolic outcomes including higher systolic blood pressure, diabetes, metabolic syndrome, and adiposity (1). In addition, hair cortisol concentrations have been shown to be associated with an increased risk of having cardiovascular diseases, such as coronary heart disease, stroke, and peripheral arterial disease, in elderly (9). The associations of chronic stress with adverse cardiometabolic outcomes may originate in early life. It is well known that adverse exposures in early life are associated with cardiovascular risk factors development from childhood onward (10-12). Also, cardiometabolic risk factors tend to track from childhood into adulthood (13-16). We previously reported associations of hair cortisol concentrations with childhood body mass index (BMI) and fat mass distribution at ages 6 and 10 years (17, 18). Thus far, studies in children have not reported associations between hair cortisol concentrations and cardiometabolic risk factors (19-21). These previous studies had small sample sizes and, mostly, a cross-sectional design.
We hypothesized that chronic exposure to higher cortisol concentrations is associated with an adverse cardiometabolic risk profile already in school-aged children, and thereby predispose individuals to later-life cardiovascular disease. We examined, in a population-based prospective cohort study among 2598 children, the associations of hair cortisol concentrations at age 6 years with blood pressure, heart rate, lipid profile, glucose metabolism, and C-reactive protein concentrations at ages 6 and 10 years, and explored the potential mediating role of childhood BMI.
Materials and Methods
Study Design
This study was embedded in the Generation R Study, a population-based prospective cohort study from early pregnancy onward in Rotterdam, the Netherlands (22). Written informed consent was provided for all children. The medical ethics committee of Erasmus MC approved the study (MEC 198.782/2001/31). This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline. In total, 2984 children had information on hair cortisol concentrations at age 6 years. Twins (N = 58) and children without any measurement of cardiometabolic risk factors at ages 6 and 10 years (N = 23) were excluded. Since the highest values of hair cortisol concentrations may be incorrect because of external factors such as glucocorticoid use, we excluded the extreme values of cortisol (N = 305) using the Tukey definition of outliers (Q1-1.5 * interquartile range [IQR] and Q3 + 1.5 * IQR) (23). There were no substantial differences in the cardiometabolic risk factors between the population of analysis and the excluded outliers (data not shown). The population for analysis consisted of 2598 children. The same selection procedure was followed for the cortisone analyses (N = 2605). The flowchart of participants is given in Supplementary Fig. 1 (24).
Hair Cortisol Concentration Measurements
As described previously, in children aged 6 years, a hair strand of approximately 100 hairs was cut from the posterior vertex using small surgical scissors, as close to the scalp as possible (25). Details on collection, sample preparation, extraction, and analysis using the liquid chromatography–tandem mass spectrometry method are provided in the Supplementary Methods (24). To reduce variability and account for right skewedness of the distribution, cortisol and cortisone concentrations outliers defined by the Tukey definition of outliers (Q1-1.5 * IQR and Q3 + 1.5 * IQR) were excluded, after which values were either divided in quintiles, or natural log-transformed and further standardized by the IQR to ease the interpretation of effect sizes (23).
Cardiometabolic Risk Factors
Outcome assessments were performed at ages 6 and 10 years (22). Blood pressure and heart rate were measured at the right brachial artery 4 times with 1-minute intervals, using the validated automatic sphygmanometer Datascope Accutor Plus (26). We calculated the mean value for systolic and diastolic blood pressure and heart rate using the last 3 measurements for each participant. Thirty-minute fasting venous blood samples were collected to measure serum concentrations of insulin, total cholesterol, high-density lipoprotein (HDL) cholesterol, triglycerides, C-reactive protein at ages 6 and 10 years, and glucose only at age 10 years (22). Because the blood samples were collected at different time points during the day, it was not possible to have fasting samples. Participants were asked to stop eating and drinking 30 minutes before the blood draw. Unfortunately, we do not have information on the exact time between the last meal and samples, nor on the nutrient composition of the meals. Glucose, total cholesterol, HDL cholesterol, triglycerides, and C-reactive protein concentrations were measured using the c702 module on the Cobas 8000 analyzer. Insulin was measured by electrochemiluminescence immunoassay on the e601 module (Roche) (27). Low-density lipoprotein (LDL) cholesterol was calculated according to the Friedewald formula (28, 29). We defined children with a clustering of cardiometabolic risk factors as being at risk for the metabolic syndrome phenotype, in line with other studies (30, 31). Clustering of cardiometabolic risk factors was defined as having 3 or more out of the following 4 adverse risk factors: android fat mass percentage above the 75th percentile; systolic or diastolic blood pressure above the 75th percentile; HDL cholesterol below the 25th percentile or triglycerides above the 75th percentile; and insulin above the 75th percentile of our study population. We measured total body fat mass and fat mass in the abdomen (android fat mass) using a dual-energy x-ray absorptiometry scanner (iDXA, GE140 Lunar, 2008, enCORE software v.12.6) according to standard procedures (32). We calculated android fat mass percentage as android fat mass divided by total body fat mass. For the clustering of cardiometabolic risk factors at age 10 years, we also had visceral fat mass obtained by magnetic resonance imaging scans available, as described previously (22). Because the distribution of insulin and triglycerides concentrations was skewed, we used their natural logged values. Since C-reactive protein was not normally distributed and transformation did not yield an acceptable distribution, we categorized C-reactive protein concentrations into less than 3 mg/L (normal levels) or greater than or equal to 3 mg/L (high levels) in line with previous studies (33, 34). To enable comparison of effect sizes of different measures, we constructed SDS ([observed value – mean] / SD) for all variables.
Covariates
Information on child sex was obtained from midwife/obstetric records. Maternal height was assessed at the first visit. Information about maternal weight just before pregnancy was obtained by questionnaire. Maternal prepregnancy BMI was calculated as weight in kilograms divided by height in meters squared. Information on maternal education, family income, child ethnicity, and television watching time was obtained by questionnaires. Hair color was partially coded through parent report and was completed by 2 raters using photographs made at the research center. We calculated BMI at age 6 years from height and weight, both measured without shoes and heavy clothing. Parents completed a questionnaire about their child on factors that can potentially influence hair cortisol concentrations, such as hair washing frequency, time since last wash, hair product use, and use of and administration route of glucocorticoid medications at age 6 years. We tested whether birth weight was a confounder in the associations of hair cortisol concentrations and cardiometabolic risk factors but birth weight did not change the effect estimates more than 10% and thus was not included in the final confounder model.
Statistical Analysis
First, we examined differences in participant characteristics between hair cortisol concentration quintiles with analysis of variance tests for continuous variables and chi-squared tests for categorical variables. For nonresponse analyses, we compared participants and nonparticipants using chi-square tests, t tests, and Mann-Whitney tests. Second, we used linear regression models to assess the associations of hair cortisol concentrations at age 6 years in quintiles with cardiometabolic risk factors at ages 6 and 10 years, and the change in cardiometabolic risk factor SD scores between these ages (systolic blood pressure, diastolic blood pressure, heart rate, total cholesterol, HDL and LDL cholesterol, triglycerides, insulin, glucose). Third, we used logistic regression models to assess the associations of hair cortisol concentrations at age 6 years in quintiles with the odds of increased C-reactive protein concentrations (≥ 3 mg/L) and the odds of having clustered cardiometabolic risk factors at ages 6 and 10 years. Only cases with complete data on cardiometabolic outcomes were used for the analyses with clustered cardiometabolic risk factors. Tests for trend across quintiles were performed by analyzing cortisol quintiles as a continuous variable. Fourth, we performed linear regression models to assess the associations of continuous hair cortisol concentrations (the natural log-transformed hair cortisol measures further standardized with IQR) with all cardiometabolic outcomes. The basic models were adjusted for child sex, age at cortisol measurement, and age at assessment of cardiometabolic outcomes. The confounder models were additionally adjusted for maternal prepregnancy BMI, maternal education, family income, child ethnicity, hair color, and average duration of television watching per day. We performed an additional model to assess whether any significant association in the confounder model was explained by childhood BMI. We visualized potential covariates by drawing a directed acyclic graph and included the covariates in the models that were associated with exposure and outcome at age 6 years and changed the effect estimates more than 10% (Supplementary Fig. 2) (24). We tested if there was an interaction of cortisol with sex by adding an interaction term to the basic model. After taking multiple testing into account, the interaction was only significant for insulin and triglyceride concentrations at 6 years (P < .01). For these associations, we performed sex-stratified analyses. As a sensitivity analysis, we only included children without any glucocorticoid use in the 3 months prior to the hair sample collection (N = 2296). Also, we repeated all analyses for cortisone, the less active form of cortisol (N = 2605). Considering 3 groups of outcomes (blood pressure and heart rate, lipids, and glucose metabolism), multiple testing adjustment would lead to a P value cutoff of less than .017. We depicted both significance levels (.05 and .017) in the tables and figures. Missing data of covariates were multiple-imputed using a Markov chain Monte Carlo approach (35). Five imputed data sets were created and analyzed together. All statistical analyses were performed using the Statistical package of Social Sciences version 24.0 for Windows (IBM Corp, released 2016; IBM SPSS Statistics for Windows, version 24.0).
Results
Participant Characteristics
As compared to children in the lower cortisol quintiles, children in the upper cortisol quintiles more often had a mother who had a higher prepregnancy BMI, was lower educated, and had a lower family income. Also, these children more often had non-European ethnicity, a higher BMI and systolic blood pressure at ages 6 and 10 years, brown or black hair color, and a higher average duration of television watching at age 6 years (Tables 1 and 2) (24). Nonresponse analyses showed that, compared to mothers of participants, mothers of nonparticipants more often had a higher BMI, a lower family income, and lower education. Nonparticipants more often were boys, had non-European ethnicity, a higher BMI, brown or dark hair, watched more television, and used more often glucocorticoids in the 3 months prior to hair sampling (Supplementary Table 2) (24).
Table 1.
Family and birth characteristics (N = 2598)
| Hair cortisol concentrations | |||||||
|---|---|---|---|---|---|---|---|
| Total groupa | Quintile 1a | Quintile 2a | Quintile 3a | Quintile 4 a | Quintile 5a | P b | |
| 0.131-6.764 | 0.131-0.744 pg/mg | 0.745-1.173 pg/mg | 1.174-1.831 pg/mg | 1.832-2.925 pg/mg | 2.926-6.764 pg/mg | ||
| (N = 2598) | (N = 519) | (N = 520) | (N = 520) | (N = 520) | (N = 519) | ||
| Family characteristics | |||||||
| Prepregnancy BMI, median (95% range) | 22.6 (18.2-35.1) | 22.1 (18.8-33.9) | 22.7 (18.1-34.6) | 22.6 (18.1-35.1) | 22.3 (18.1-35.9) | 23.0 (17.5-36.2) | .039 |
| Maternal education, % | < .001 | ||||||
| Primary school | 98 (4.5) | 8 (1.7) | 15 (3.4) | 22 (5.0) | 34 (7.9) | 19 (4.6) | |
| Secondary school | 815 (37.2) | 151 (32.7) | 150 (34.0) | 167 (38.2) | 169 (39.1) | 178 (42.8) | |
| High education | 1275 (58.3) | 303 (65.6) | 276 (62.6) | 248 (56.8) | 229 (53.0) | 219 (52.6) | |
| Family income, % | < .001 | ||||||
| Low (< €1600/mo) | 327 (15.8) | 39 (8.8) | 51 (12.3) | 58 (14.1) | 90 (22.1) | 89 (22.6) | |
| Medium (€1600-4000/mo) | 981 (47.4) | 202 (45.5) | 186 (44.7) | 212 (51.7) | 197 (48.3) | 184 (46.8) | |
| High (> €4000/mo) | 763 (36.8) | 203 (45.7) | 179 (43.0) | 140 (34.1) | 121 (29.7) | 120 (30.5) | |
| Birth characteristics | |||||||
| Sex, No., % | .054 | ||||||
| Boys | 1237 (47.6) | 223 (43.0) | 248 (47.7) | 247 (47.5) | 247 (47.5) | 272 (52.4) | |
| Girls | 1361 (52.4) | 296 (57.0) | 272 (52.3) | 273 (52.5) | 273 (52.5) | 247 (47.6) | |
| Ethnicity, % | < .001 | ||||||
| European | 1644 (65.0) | 419 (82.2) | 355 (70.9) | 302 (59.2) | 283 (55.6) | 285 (57.1) | |
| Non-European | 885 (35.0) | 91 (17.8) | 146 (29.1) | 208 (40.8) | 226 (44.4) | 214 (42.9) |
Abbreviation: BMI: body mass index.
aValues are means (SD), medians (95% range), or numbers of participants (valid %).
bP values for differences in participant characteristics between cortisol quintiles were tested using one-way analysis of variance tests for continuous variables and chi-square tests for categorical variables.
Table 2.
Child characteristics (N = 2598)
| Hair cortisol concentrations | |||||||
|---|---|---|---|---|---|---|---|
| Total groupa | Quintile 1a | Quintile 2a | Quintile 3a | Quintile 4a | Quintile 5a | P b | |
| 0.131-6.764 | 0.131-0.744 pg/mg | 0.745-1.173 pg/mg | 1.174-1.831 pg/mg | 1.832-2.925 pg/mg | 2.926-6.764 pg/mg | ||
| (N = 2598) | (N = 519) | (N = 520) | (N = 520) | (N = 520) | (N = 519) | ||
| Child characteristics at age 6 y | |||||||
| Age at measurements, median (95% range), y | 5.9 (5.7-8.1) | 5.9 (5.6-8.1) | 5.9 (5.7-8.1) | 5.9 (5.7-8.2) | 5.9 (5.7-8.2) | 5.9 (5.7-8.1) | .013 |
| Body mass index, median (95% range) | 15.8 (13.6-21.2) | 15.7 (13.7-19.0) | 15.7 (13.6-21.0) | 16.0 (13.7-20.9) | 15.9 (13.7-21.5) | 16.1 (13.6-22.4) | < .001 |
| Hair cortisol concentrations, median (95% range), pg/mg | 1.46 (0.33-5.62) | 0.56 (0.23-0.74) | 0.96 (0.75-1.17) | 1.46 (1.18, 1.81) | 2.28 (1.86-2.88) | 3.98 (2.98-6.60) | < .001 |
| Hair cortisone concentrations, median (95% range), pg/mg | 7.50 (2.63-29.00) | 4.65 (2.00-8.78) | 6.19 (2.95, 10.66) | 8.25 (3.31, 14.56) | 11.52 (3.60, 23.18) | 16.00 (3.65-44.58) | < .001 |
| Systolic blood pressure, mean (SD), mm Hg | 102.6 (8.4) | 101.7 (8.1) | 102.1 (8.2) | 103.1 (8.2) | 103.0 (8.2) | 103.2 (9.0) | .007 |
| Diastolic blood pressure, mean (SD), mm Hg | 60.5 (6.7) | 60.1 (6.4) | 60.1 (6.6) | 60.4 (7.1) | 60.4 (6.6) | 61.2 (7.0) | .072 |
| Heart rate, mean (SD), beats/min | 82.7 (9.8) | 82.0 (9.0) | 83.6 (10.1) | 83.1 (10.4) | 82.4 (9.6) | 82.7 (9.9) | .091 |
| Insulin, median (95% range), pmol/L | 116.90 (18.58-405.28) | 110.70 (24.60-401.89) | 130.30 (15.27-411.94) | 93.49 (13.30-388.15) | 128.60 (18.26-422.35) | 115.30 (19.53-455.04) | .014 |
| Total cholesterol, mean (SD), mmol/L | 4.25 (0.65) | 4.27 (0.63) | 4.28 (0.62) | 4.28 (0.66) | 4.23 (0.63) | 4.22 (0.68) | .571 |
| HDL cholesterol, mean (SD), mmol/L | 1.38 (0.32) | 1.38 (0.33) | 1.37 (0.32) | 1.39 (0.33) | 1.35 (0.30) | 1.39 (0.30) | .640 |
| LDL cholesterol, mean (SD), mmol/L | 2.37 (0.56) | 2.37 (0.55) | 2.41 (0.54) | 2.39 (0.59) | 2.37 (0.54) | 2.35 (0.58) | .690 |
| Triglycerides, median (95% range), mmol/L | 0.99 (0.41-2.35) | 0.97 (0.41-2.64) | 0.97 (0.40-2.35) | 1.02 (0.42-2.14) | 0.99 (0.41-2.37) | 1.00 (0.38-2.30) | .973 |
| C-reactive protein, No., % | .025 | ||||||
| < 3 mg/L | 1591 (89.2) | 334 (92.0) | 313 (90.2) | 301 (85.8) | 319 (91.1) | 324 (86.9) | |
| ≥ 3 mg/L | 193 (10.8) | 29 (8.0) | 34 (9.8) | 50 (14.2) | 31 (8.9) | 49 (13.1) | |
| Prevalence cardiometabolic clustering, No., %c | 218 (10.8) | 34 (8.4) | 44 (11.2) | 47 (11.6) | 45 (11.1) | 48 (11.7) | .529 |
| Glucocorticoid use in 3 mo prior to hair sample collection, No., % | .622 | ||||||
| No | 2296 (93.0) | 452 (92.2) | 469 (94.6) | 460 (92.4) | 455 (93.0) | 460 (92.7) | |
| Yes | 173 (7.0) | 38 (7.8) | 27 (5.4) | 38 (7.6) | 34 (7.0) | 36 (7.3) | |
| Hair color, No., % | < .001 | ||||||
| Red | 78 (3.0) | 20 (3.9) | 16 (3.1) | 20 (3.9) | 11 (2.1) | 11 (2.1) | |
| Blond | 1381 (53.2) | 376 (72.4) | 281 (54.0) | 241 (46.4) | 246 (47.4) | 237 (45.7) | |
| Brown | 857 (33.0) | 111 (21.4) | 175 (33.7) | 207 (39.9) | 180 (34.7) | 184 (35.5) | |
| Black | 280 (10.8) | 12 (2.3) | 48 (9.2) | 51 (9.8) | 82 (15.8) | 87 (16.8) | |
| Television watching time, No., % | < .001 | ||||||
| < 2 h/d | 1631 (81.7) | 381 (88.6) | 338 (83.5) | 318 (79.5) | 303 (78.1) | 291 (78.0) | |
| ≥ 2 h/d | 365 (18.3) | 49 (11.4) | 67 (16.5) | 82 (20.5) | 85 (21.9) | 82 (22.0) | |
| Child characteristics at age 10 y | |||||||
| Age at measurements, median (95% range), years | 9.7 (9.3-10.6) | 9.7 (9.3-10.7) | 9.7 (9.3-10.4) | 9.7 (9.4-10.6) | 9.8 (9.2-10.7) | 9.7 (9.3-11.0) | .155 |
| Body mass index, median (95% range) | 16.9 (14.0-24.8) | 16.5 (13.9-22.4) | 16.7 (14.2-23.9) | 17.0 (14.0-23.5) | 17.0 (14.1-25.4) | 17.2 (13.8-26.6) | < .001 |
| Systolic blood pressure, mean (SD), mm Hg | 103.2 (8.1) | 102.4 (7.8) | 102.7 (7.4) | 103.0 (8.3) | 103.9 (8.4) | 104.2 (8.4) | .005 |
| Diastolic blood pressure, mean (SD), mm Hg | 58.6 (6.6) | 58.4 (6.8) | 58.8 (6.3) | 58.4 (6.7) | 58.5 (6.8) | 59.0 (6.6) | .698 |
| Heart rate, mean (SD), beats/min | 74.3 (10.1) | 74.3 (10.0) | 74.3 (9.8) | 74.2 (10.4) | 74.8 (10.6) | 74.2 (10.0) | .928 |
| Insulin, median (95% range), pmol/L | 193.2 (33.37-709.27) | 177.10 (37.21-613.68) | 189.50 (30.72-618.40) | 201.40 (37.03-698.83) | 202.00 (28.19-718.82) | 199.55 (30.81-777.55) | .341 |
| Glucose, mean (SD), mmol/L | 5.38 (0.92) | 5.42 (0.91) | 5.30 (0.84) | 5.47 (0.99) | 5.35 (0.92) | 5.38 (0.94) | .227 |
| Total cholesterol, mean (SD), mmol/L | 4.30 (0.65) | 4.28 (0.64) | 4.32 (0.64) | 4.29 (0.68) | 4.28 (0.63) | 4.34 (0.66) | .790 |
| HDL cholesterol, mean (SD), mmol/L | 1.47 (0.33) | 1.48 (0.33) | 1.49 (0.35) | 1.46 (0.31) | 1.46 (0.32) | 1.47 (0.33) | .831 |
| LDL cholesterol, mean (SD), mmol/L | 2.33 (0.57) | 2.31 (0.57) | 2.32 (0.56) | 2.34 (0.59) | 2.33 (0.53) | 2.36 (0.58) | .835 |
| Triglycerides, median (95% range), mmol/L | 0.93 (0.42-2.56) | 0.95 (0.39-2.63) | 0.93 (0.43-2.84) | 0.93 (0.37-2.43) | 0.95 (0.45-2.38) | 0.93 (0.44-2.65) | .910 |
| C-reactive protein, No., % | .140 | ||||||
| < 3 mg/L | 1303 (93.9) | 272 (96.1) | 266 (94.7) | 254 (94.8) | 267 (92.1) | 244 (91.7) | |
| ≥ 3 mg/L | 85 (3.3) | 11 (3.9) | 15 (5.3) | 14 (5.2) | 23 (7.9) | 22 (8.3) | |
| Prevalence cardiometabolic clustering, No., %c | 172 (13.3) | 26 (10.0) | 35 (13.4) | 40 (15.7) | 31 (11.7) | 40 (15.7) | .219 |
Abbreviations: HDL, high-density lipoprotein; LDL, low-density lipoprotein; pg/mg, picogram per milligram.
aValues are means (SD), medians (95% range), or numbers of participants (valid %).
bP values for differences in participant characteristics between cortisol quintiles were tested using one-way analysis of variance tests for continuous variables and chi-square tests for categorical variables.
cClustering of cardiometabolic risk factors was defined as having 3 or more out of the following 4 adverse risk factors: android fat mass percentage above the 75th percentile; systolic or diastolic blood pressure above the 75th percentile; HDL cholesterol below the 25th percentile or triglycerides above the 75th percentile; and insulin above the 75th percentile of our study population. We used android fat mass as a percentage of total body fat mass as a proxy for waist circumference because this was not available. For the clustering of cardiometabolic risk factors at 10 years, we also had visceral fat mass obtained by magnetic resonance imaging scans available, as described previously (22). (N = 2022 at 6 years and N = 1.298 at 10 years).
Cardiovascular Risk Factors
Results from the basic models showed that, as compared to the lowest quintile of hair cortisol concentrations at age 6 years, children in the highest quintile had a higher systolic and diastolic blood pressure and heart rate at age 6 years (see Supplementary Table 2) (24). When we adjusted these models for potential confounders, these associations attenuated into nonsignificance (Table 3) (24). As compared to the lowest quintile of hair cortisol concentrations at age 6 years, children in the highest quintiles had a higher systolic blood pressure at age 10 years in the basic models (see Supplementary Table 2) (24). This association remained significant after adjustment for confounders (difference 0.15 SDS; 95% CI 0.00-0.29; and 0.17 SDS; 95% CI, 0.03-0.31) for the fourth and fifth quintile, respectively) but attenuated into nonsignificance after additional adjustment for childhood BMI at age 6 years (see Table 3 and Supplementary Table 3) (24). Associations for continuous cortisol measures showed similar results (Supplementary Table 4) (24). The tests for trend across the quintiles were not significant.
Table 3.
Association of hair cortisol quintiles at age 6 years with blood pressure and heart rate at ages 6 years and 10 years and with the change between ages 6 and 10 years, confounder models
| Cardiovascular risk factors at age 6 y | |||
|---|---|---|---|
| Difference (95% CI) in SDS | |||
| Hair cortisol Q at age 6 y | Systolic blood pressurea | Diastolic blood pressurea | Heart ratea |
| (N = 2466) | (N = 2466) | (N = 2465) | |
| Q1 | Reference | Reference | Reference |
| Q2 | 0.02 (–0.11 to 0.14) | –0.02 (–0.14 to 0.11) | 0.17 (0.05 to 0.29) e |
| Q3 | 0.11 (–0.02 to 0.23) | –0.01 (–0.13 to 0.11) | 0.09 (–0.03 to 0.21) |
| Q4 | 0.09 (–0.04 to 0.22) | –0.02 (–0.14 to 0.10) | 0.01 (–0.11 to 0.12) |
| Q5 | 0.09 (–0.04 to 0.21) | 0.09 (–0.03 to 0.22) | 0.05 (–0.07 to 0.17) |
| Cardiovascular risk factors at age 10 y | |||
| Difference (95% CI) in SDS | |||
| Hair cortisol Q at age 6 y | Systolic blood pressure b | Diastolic blood pressure b | Heart rate b |
| (N = 1938) | (N = 1938) | (N = 1891) | |
| Q1 | Reference | Reference | Reference |
| Q2 | 0.03 (–0.11 to 0.17) | 0.04 (–0.10 to 0.19) | –0.01 (–0.15 to 0.14) |
| Q3 | 0.04 (–0.11 to 0.18) | –0.04 (–0.19 to 0.11) | –0.04 (–0.19 to 0.11) |
| Q4 | 0.15 (0.00 to 0.29) d | –0.04 (–0.19 to 0.11) | 0.01 (–0.13 to 0.16) |
| Q5 | 0.17 (0.03 to 0.31) d | 0.05 (–0.11 to 0.19) | –0.01 (–0.16 to 0.13) |
| Cardiovascular risk factors between ages 6 and 10 y | |||
| Difference (95% CI) in SDS | |||
| Hair cortisol Q at age 6 y | Change in systolic blood pressure c | Change in diastolic blood pressure c | Change in heart rate c |
| (N = 1829) | (N = 1828) | (N = 1789) | |
| Q1 | Reference | Reference | Reference |
| Q2 | 0.02 (–0.12 to 0.16) | 0.14 (–0.02 to 0.30) | –0.08 (–0.22 to 0.07) |
| Q3 | –0.06 (–0.20 to 0.08) | 0.01 (–0.15 to 0.18) | –0.13 (–0.27v0.02) |
| Q4 | 0.06 (–0.09 to 0.20) | –0.00 (–0.17 to 0.16) | 0.02 (–0.13 to 0.16) |
| Q5 | 0.05 (–0.10 to 0.19) | –0.05 (–0.22 to 0.11) | –0.07 (–0.21 to 0.07) |
Confounder models are adjusted for child sex, child age at cortisol assessment, child age at assessment of cardiometabolic outcomes, maternal prepregnancy BMI, maternal education, family income, child ethnicity, television watching time, and hair color.
Abbreviations: Q, quintile; SDS, SD score.
aValues are linear regression coefficients (95% CI) and reflect the change in blood pressure and heart rate at age 6 years in SDS for the cortisol quintiles compared to the first quintile.
bValues are linear regression coefficients (95% CI) and reflect the change in blood pressure and heart rate at age 10 years in SDS for the cortisol quintiles compared to the first quintile.
cValues are linear regression coefficients (95% CI) and reflect the change in blood pressure and heart rate between ages 6 and 10 years in SDS for the cortisol quintiles compared to the first quintile.
dP less than .05.
eP less than .017.
Metabolic Risk Factors
Hair cortisol concentrations were not associated with lipid and glucose metabolism biomarkers in the basic and main models (Supplementary Tables 4 and 5, respectively) (24). In the sex-stratified analyses, higher hair cortisol concentrations were associated with lower triglycerides and insulin concentrations among boys only at age 6 years (differences –0.11 SDS; 95% CI, –0.21 to –0.01; and –0.09 SDS; 95% CI, –0.18 to 0.00, respectively) (Supplementary Table 6) (24). As compared to the lowest quintile of hair cortisol concentrations at age 6 years, children in the highest quintile had a higher increase in total cholesterol and LDL cholesterol concentrations from ages 6 to 10 years (differences 0.19 SDS; 95% CI, 0.05-0.34; and 0.15 SDS; 95% CI, 0.00-0.29, respectively), but no difference in change of other metabolic risk factors (Table 4) (24). The association for change in total cholesterol was independent of childhood BMI at age 6 years (see Supplementary Table 3) (24). The associations for continuous cortisol measures showed similar results (Supplementary Table 4) (24).
Table 4.
Association of hair cortisol quintiles at 6 years with lipids, insulin and glucose at ages 6 years and 10 years and with the change between 6 and 10 years, confounder models
| Metabolic risk factors at age 6 y | ||||||
|---|---|---|---|---|---|---|
| Difference (95% CI) in SDS | ||||||
| Hair cortisol Q at age 6 y | Total cholesterola | HDLa | LDLa | Triglyceridesa | Insulina | |
| (N = 1781) | (N = 1781) | (N = 1782) | (N = 1773) | (N = 1766) | ||
| Q1 | Reference | Reference | Reference | Reference | Reference | |
| Q2 | 0.01 (–0.14, 0.16) | –0.04 (–0.19, 0.11) | 0.08 (–0.06, 0.23) | –0.01 (–0.16, 0.14) | 0.10 (–0.05, 0.24) | |
| Q3 | –0.02 (–0.17, 0.13) | 0.00 (–0.15, 0.15) | 0.01 (–0.13, 0.16) | 0.02 (–0.13, 0.16) | –0.10 (–0.25, 0.04) | |
| Q4 | –0.08 (–0.23, 0.07) | –0.11 (–0.26, 0.04) | 0.01 (–0.14, 0.16) | 0.02 (–0.13, 0.17) | 0.13 (–0.02, 0.27) | |
| Q5 | –0.10 (–0.25, 0.05) | –0.01 (–0.16, 0.14) | –0.04 (–0.19, 0.10) | –0.03 (–0.18, 0.12) | –0.04 (–0.19, 0.11) | |
| Metabolic risk factors at age 6 y | ||||||
| Difference (95% CI) in SDS | ||||||
| Hair cortisol Q at age 6 y | Total cholesterol b | HDL b | LDL b | Triglycerides b | Insulin b | Glucose b |
| (N = 1388) | (N = 1387) | (N = 1377) | (N = 1380) | (N = 1382) | (N = 1387) | |
| Q1 | Reference | Reference | Reference | Reference | Reference | Reference |
| Q2 | 0.05 (–0.11 to 0.22) | 0.05 (–0.11 to 0.20) | 0.02 (–0.14 to 0.18) | 0.01 (–0.15 to 0.18) | –0.03 (–0.20 to 0.14) | –0.12 (–0.28 to 0.05) |
| Q3 | 0.01 (–0.16 to 0.17) | –0.03 (–0.19 to 0.13) | 0.05 (–0.11 to 0.21) | –0.07 (–0.24 to 0.09) | 0.08 (–0.09 to 0.26) | 0.06 (–0.11 to 0.22) |
| Q4 | –0.01 (–0.18 to 0.15) | –0.01 (–0.16 to 0.15) | 0.02 (–0.14 to 0.18) | –0.06 (–0.23 to 0.10) | 0.05 (–0.12 to 0.22) | –0.05 (–0.21 to 0.12) |
| Q5 | 0.10 (–0.07 to 0.27) | 0.04 (–0.12 to 0.20) | 0.11 (–0.06 to 0.27) | –0.05 (–0.22 to 0.12) | 0.03 (–0.15 to 0.20) | –0.02 (–0.18 to 0.15) |
| Metabolic risk factors change between ages 6 and 10 y | ||||||
| Difference (95% CI) in SDS | ||||||
| Hair cortisol Q at age 6 y | Change in total cholesterol c | Change in HDL c | Change in LDL c | Change in triglycerides c | Change in insulin c | |
| (N = 1051) | (N = 1050) | (N = 1043) | (N = 1041) | (N = 1036) | ||
| Q1 | Reference | Reference | Reference | Reference | Reference | |
| Q2 | 0.09 (–0.05 to 0.23) | 0.03 (–0.11 to 0.18) | –0.01 (–0.15 to 0.13) | 0.08 (–0.14 to 0.30) | –0.14 (–0.39 to 0.10) | |
| Q3 | 0.10 (–0.05 to 0.25) | 0.07 (–0.08 to 0.22) | 0.06 (–0.08 to 0.20) | –0.01 (–0.24 to 0.21) | 0.15 (–0.10 to 0.40) | |
| Q4 | 0.12 (–0.03 to 0.26) | 0.13 (–0.02 to 0.28) | 0.06 (–0.09 to 0.20) | –0.10 (–0.32 to 0.13) | –0.08 (–0.33 to 0.17) | |
| Q5 | 0.19 (0.05 to 0.34) e | 0.02 (–0.13 to 0.16) | 0.15 (0.00 to 0.29) d | 0.04 (–0.18 to 0.27) | 0.08 (–0.16 to 0.33) |
Confounder models are adjusted for child sex, child age at cortisol assessment, child age at assessment of cardiometabolic outcomes, maternal prepregnancy body mass index, maternal education, family income, child ethnicity, television watching time, and hair color.
Abbreviations: HDL, high-density lipoprotein; LDL, low-density lipoprotein; Q, quintile; SDS, SD score.
aValues are linear regression coefficients (95% CI) and reflect the change in lipids and insulin concentrations at age 6 years in SDS for the cortisol quintiles compared to the first quintile.
bValues are linear regression coefficients (95% CI) and reflect the change in lipids, insulin, and glucose concentrations at age 10 years in SDS for the cortisol quintiles compared to the first quintile.
cValues are linear regression coefficients (95% CI) and reflect the change in the delta of lipids and insulin concentrations between ages 6 and 10 years in SDS for the cortisol quintiles compared to the first quintile.
dP less than .05.
eP less than .017.
Increased C-Reactive Protein Concentrations and Clustering of Cardiovascular Risk Factors
Results from the basic models showed that, as compared to the lowest quintile of hair cortisol concentrations at age 6 years, children in the highest quintile had a higher risk of increased C-reactive protein at age 6 years (odds ratio [OR]: 1.76; 95% CI, 1.08-2.86) and a higher risk of increased C-reactive protein (OR: 2.23; 95% CI, 1.05-4.70) and cardiometabolic clustering (OR 1.73; 95% CI, 1.01-2.97) at age 10 years (Supplementary Table 7) (24). When we further adjusted the models for potential confounders, these associations attenuated into nonsignificance (Table 5) (24). The associations for continuous cortisol measures showed similar results (Supplementary Table 8) (24).
Table 5.
Association of hair cortisol quintiles at age 6 years with risk of increased C-reactive protein and risk of cardiometabolic clustering at ages 6 and 10 years, confounder models
| Odds ratio (95% CI) for outcomes at age 6 y | ||
|---|---|---|
| Hair cortisol Q at age 6 y | Risk of C-reactive protein ≥ 3 mg/La | Risk of cardiometabolic clusteringb |
| (N = 1784) | (N = 2022) | |
| Q1 | Reference | Reference |
| Q2 | 1.19 (0.70 to 2.01) | 1.39 (0.86 to 2.27) |
| Q3 | 1.69 (1.04 to 2.77) c | 1.32 (0.81 to 2.14) |
| Q4 | 0.98 (0.57 to 1.68) | 1.21 (0.74 to 1.98) |
| Q5 | 1.54 (0.94 to 2.54) | 1.29 (0.79 to 2.11) |
| Odds ratio (95% CI) for outcomes at age 10 y | ||
| Hair cortisol Q at age 6 y | Risk of C-reactive protein ≥ 3 mg/La | Risk of cardiometabolic clustering b |
| (N = 1389) | (N = 1299) | |
| Q1 | Reference | Reference |
| Q2 | 1.34 (0.59 to 3.02) | 1.34 (0.75 to 2.39) |
| Q3 | 1.14 (0.50 to 2.61) | 1.50 (0.86 to 2.64) |
| Q4 | 1.76 (0.82 to 3.79) | 1.00 (0.55 to 1.82) |
| Q5 | 1.91 (0.88 to 4.13) | 1.32 (0.74 to 2.35) |
Confounder models are adjusted for child sex, child age at cortisol assessment, child age at assessment of cardiometabolic outcomes, maternal prepregnancy body mass index, maternal education, family income, child ethnicity, television watching time, and hair color.
Abbreviation: Q, quintile.
aValues are odds ratios (95% CI) and represent the risk of childhood high C-reactive protein concentrations (≥ 3 mg/L) at ages 6 and 10 years for the cortisol quintiles compared to the first quintile.
bValues are odds ratios (95% CI) and reflect the odds of cardi-metabolic clustering at ages 6 and 10 years defined as having 3 or more out of the following 4 adverse risk factors: android fat mass percentage above the 75th percentile; systolic or diastolic blood pressure above the 75th percentile; high-density lipoprotein cholesterol below the 25th percentile or triglycerides above the 75th percentile; and insulin above the 75th percentile of our study population for the cortisol quintiles compared to the first quintile.
cP less than .05.
dP less than .017.
Sensitivity Analyses
In the confounder models, excluding children with all types of glucocorticoid use in the 3 months prior to hair sample collection (N = 173), we observed similar but slightly stronger results for systolic blood pressure at age 10 years (differences 0.20 SDS; 95% CI, 0.05-0.34) (Supplementary Table 9) (24). When we further adjusted the model for childhood BMI, the association attenuated into nonsignificance (Supplementary Table 10) (24). In these analyses, results from the confounder model showed that children in the highest quintile of hair cortisol concentrations at age 6 years, compared to those in the lowest quintile, had a higher risk of increased C-reactive protein at age 6 years (OR: 1.83; 95% CI: 1.06-3.13) and age 10 years (OR: 2.53; 95% CI, 1.11-5.77) (Supplementary Table 11), independent of childhood BMI at age 6 years (see Supplementary Table 10) (24). Hair cortisone concentrations at age 6 years were not associated with any of the cardiometabolic outcomes at age 6 or 10 years (results not shown).
Discussion
In this population-based, prospective cohort study among 2598 children, we observed that hair cortisol concentrations at age 6 years were not consistently associated with cardiometabolic risk factors at ages 6 and 10 years. The association of higher hair cortisol concentrations at age 6 years with higher systolic blood pressure at age 10 years was explained by childhood BMI.
Interpretation of Main Findings
A meta-analysis in 2832 adults from 11 studies showed that higher hair cortisol concentrations were associated with higher systolic blood pressure, but not with diastolic blood pressure (36). Also, a review including 20 studies investigating the relationships between various cortisol measures and cardiometabolic parameters in adults reported that 3 out of 6 studies found positive associations between cortisol measures and systolic blood pressure and reported inconclusive results for the other outcomes (37). Previous studies in children did not find associations between hair cortisol concentrations and blood pressure, heart rate, lipids, C-reactive protein, or glucose metabolism (19-21). These studies in children had smaller sample sizes and most of them had a cross-sectional design. Results of studies into the associations of salivary, serum, or urinary cortisol with cardiometabolic risk factors in childhood were not consistent (38-42).
In the present study, we observed that higher hair cortisol concentrations at age 6 years were associated with a higher systolic blood pressure at age 10 years. This finding is in line with the findings of 3 cross-sectional studies in children that showed a positive association between serum cortisol concentrations and systolic blood pressure but not, or less clearly, with diastolic blood pressure (40-42). However, we did not find an association between hair cortisol concentrations at age 6 years and systolic blood pressure at age 6 years. Thus, it may be that higher cortisol concentrations lead to increased systolic blood pressure later in childhood, which is known to track into adulthood (13). Additional adjustment of the association between hair cortisol concentrations at age 6 years and systolic blood pressure at age 10 years for childhood BMI at age 6 years resulted in attenuation of this association. This is in contrast with the findings of the 3 cross-sectional studies mentioned earlier, which showed that the association remained after adjustment for BMI or total body fat mass (40-42). Childhood BMI can be either an intermediate or a confounder in the association of hair cortisol concentration at age 6 years and systolic blood pressure at age 10 years. We know from previous studies in our cohort that higher hair cortisol concentrations are associated with higher childhood BMI at ages 6 and 10 years (17, 18). A bidirectional association between cortisol and adiposity may be present that should be further explored in future studies (43). Future studies are also needed to obtain further insight into the role of BMI in the association of hair cortisol concentrations with blood pressure.
We observed an association between higher hair cortisol concentrations and the increase in total cholesterol and LDL concentrations between ages 6 and 10 years, but not with any of the lipid concentrations at age 6 or 10 years. Studies that used different types of samples to measure cortisol did not find an association with lipid concentrations in children (20, 21, 38, 39, 42). Studies in adults are not consistent about the association between cortisol and lipids, but most provide evidence for a positive association between cortisol and total cholesterol and LDL (44-48). It may be that the association between higher hair cortisol and higher total cholesterol and LDL concentrations becomes more apparent at later ages.
In sex-stratified analyses, we observed that higher hair cortisol concentrations were associated with lower triglyceride and insulin concentrations among boys at age 6 years, independent of childhood BMI, and higher concentrations of triglycerides and insulin among girls at age 6 years. These findings were significant only among boys and not among girls, which may be explained by a higher variability in hair cortisol concentrations among boys. In our study and similarly to previous studies, hair cortisol concentrations were significantly higher among boys than girls (36, 49). It has been hypothesized that sex differences in reactivity to psychological stress might contribute to the sex differences in morbidity and mortality rates of cardiovascular diseases. However, studies in adults have not reported differences in the associations between hair cortisol concentrations and cardiometabolic risk factors after stratification on sex (36, 37, 50, 51). The sex-specific associations of cortisol concentrations with cardiovascular risk factors and disease need further study.
Metabolic syndrome shares many characteristics of Cushing syndrome, caused by the endogenous overproduction of cortisol, such as impaired glucose tolerance, dyslipidemia, abdominal fat distribution, and hypertension (52). Therefore, it has been suggested that altered activity of the hypothalamus-pituitary-adrenal axis leading to the hypersecretion of glucocorticoids may play an important role in the development of metabolic syndrome (52-55). However, we did not find clear evidence for an association of higher cortisol concentrations and characteristics of metabolic syndrome in childhood.
Strengths and Limitations
One of the strengths of this study was the prospective data collection from early pregnancy onward. We had a large sample size and detailed measurements of hair cortisol concentrations and childhood cardiometabolic risk factors. A limitation of our study is the lack of hair cortisol measurements at age 10 years. Therefore, we do not know how cortisol concentrations develop over time. To prevent contamination of data caused by hair cortisol outliers, we excluded cortisol values using the Tukey definition of outliers (23). Excluding these values would have affected the effect estimates if cardiometabolic risk factors were different for the excluded children and the population of analysis. However, there were no substantial differences in the characteristics of these groups. Also, the hair cortisol concentration values in the population of analysis were all within the liquid chromatography–tandem mass spectrometry–based reference interval for children aged 6, provided by a recent study that aimed to establish age-adjusted reference intervals for hair cortisol in children (56). We used nonfasting venous blood samples to measure the serum concentrations of the cardiometabolic risk factors. The blood samples were collected at different time points during the day, depending on the time of the study visit. Since glucose and insulin concentrations change easily during the day and in response to carbohydrate intake, this may have caused nondifferential misclassification. We think the effect of this potential misclassification will be minor. A previous study reported that insulin resistance or sensitivity in semifasted blood samples are moderately correlated with fasting values (57). Also, studies have reported that nonfasting lipid concentrations can predict increased risks of cardiovascular events later in life (58, 59). Overall, results should be interpreted with caution and this study should be replicated using fasting samples. Even though the analyses were adjusted for a large number of potential confounding factors, residual cofounding may still be a concern, as in any observational study. Owing to the observational design of the study, we cannot establish causality of the observed associations.
Conclusion
Our results suggest that hair cortisol concentrations at age 6 years are not consistently associated with cardiometabolic risk factors at ages 6 and 10 years. The association between hair cortisol concentrations and systolic blood pressure was explained by childhood BMI. The associations of stress with cardiometabolic risk factors may develop at later ages.
Acknowledgments
The Generation R Study is conducted by the Erasmus MC, University Medical Center Rotterdam, in close collaboration with the School of Law and Faculty of Social Sciences of the Erasmus University Rotterdam; the Municipal Health Service Rotterdam area, Rotterdam; the Rotterdam Homecare Foundation, Rotterdam; and the Stichting Trombosedienst and Artsenlaboratorium Rijnmond (STAR), Rotterdam. We gratefully acknowledge the contribution of participating parents, children, general practitioners, hospitals, midwives, and pharmacies in Rotterdam.
Financial Support: This work was supported by the European Union’s Horizon 2020 Research and Innovation Programme (Nos. 848158, EarlyCause; 733206, LifeCycle; and 874739 LongITools). The general design of the Generation R Study is supported by the Erasmus Medical Center, University Medical Center, Rotterdam, the Netherlands; Organization for Health Research and Development (ZonMw); the Netherlands Organization for Scientific Research (NWO); the Ministry of Health, Welfare and Sport; and the Ministry of Youth and Families. Dr Jaddoe has received funding from the European Research Council (grant No. ERC-2014-CoG-648916). Dr van den Akker and Dr van Rossum are supported by the Elisabeth Foundation. Dr van Rossum received a Vidi Grant (No. 91716453) from the Netherlands Organization of Scientific Research NWO.
Glossary
Abbreviations
- BMI
body mass index
- CRP
C-reactive protein
- DBP
diastolic blood pressure
- HDL
high-density lipoprotein
- IQR
interquartile range
- LDL
low-density lipoprotein
- OR
odds ratio
- SDS
SD score
Additional Information
Disclosures: The authors have no conflicts of interest relevant to this article to disclose.
Data Availability
The datasets generated during and/or analyzed during the current study are not publicly available but are available from the corresponding author on reasonable request.
References
- 1.Iob E, Steptoe A. Cardiovascular disease and hair cortisol: a novel biomarker of chronic stress. Curr Cardiol Rep. 2019;21(10):116. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Brotman DJ, Golden SH, Wittstein IS. The cardiovascular toll of stress. Lancet. 2007;370(9592):1089-1100. [DOI] [PubMed] [Google Scholar]
- 3.Song H, Fang F, Arnberg FK, et al. Stress related disorders and risk of cardiovascular disease: population based, sibling controlled cohort study. BMJ. 2019;365:l1255. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Chandola T, Brunner E, Marmot M. Chronic stress at work and the metabolic syndrome: prospective study. BMJ. 2006;332(7540):521-525. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.McEwen BS. Protective and damaging effects of stress mediators. N Engl J Med. 1998;338(3):171-179. [DOI] [PubMed] [Google Scholar]
- 6.Walker BR. Glucocorticoids and cardiovascular disease. Eur J Endocrinol. 2007;157(5):545-559. [DOI] [PubMed] [Google Scholar]
- 7.Stalder T, Kirschbaum C. Analysis of cortisol in hair—state of the art and future directions. Brain Behav Immun. 2012;26(7):1019-1029. [DOI] [PubMed] [Google Scholar]
- 8.Russell E, Koren G, Rieder M, Van Uum S. Hair cortisol as a biological marker of chronic stress: current status, future directions and unanswered questions. Psychoneuroendocrinology. 2012;37(5):589-601. [DOI] [PubMed] [Google Scholar]
- 9.Manenschijn L, Schaap L, van Schoor NM, et al. High long-term cortisol levels, measured in scalp hair, are associated with a history of cardiovascular disease. J Clin Endocrinol Metab. 2013;98(5):2078-2083. [DOI] [PubMed] [Google Scholar]
- 10.Gillman MW. The first months of life: a critical period for development of obesity. Am J Clin Nutr. 2008;87(6):1587-1589. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Osmond C, Barker DJ. Fetal, infant, and childhood growth are predictors of coronary heart disease, diabetes, and hypertension in adult men and women. Environ Health Perspect. 2000;108(Suppl 3):545-553. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Gluckman PD, Hanson MA, Cooper C, Thornburg KL. Effect of in utero and early-life conditions on adult health and disease. N Engl J Med. 2008;359(1):61-73. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Chen X, Wang Y. Tracking of blood pressure from childhood to adulthood: a systematic review and meta-regression analysis. Circulation. 2008;117(25):3171-3180. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Clarke WR, Schrott HG, Leaverton PE, Connor WE, Lauer RM. Tracking of blood lipids and blood pressures in school age children: the Muscatine study. Circulation. 1978;58(4):626-634. [DOI] [PubMed] [Google Scholar]
- 15.Joshi SM, Katre PA, Kumaran K, et al. Tracking of cardiovascular risk factors from childhood to young adulthood—the Pune Children’s Study. Int J Cardiol. 2014;175(1):176-178. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Juonala M, Viikari JSA, Rönnemaa T, Taittonen L, Marniemi J, Raitakari OT. Childhood C-reactive protein in predicting CRP and carotid intima-media thickness in adulthood: the Cardiovascular Risk in Young Finns Study. Arterioscler Thromb Vasc Biol. 2006;26(8):1883-1888. [DOI] [PubMed] [Google Scholar]
- 17.Noppe G, van den Akker EL, de Rijke YB, Koper JW, Jaddoe VW, van Rossum EF. Long-term glucocorticoid concentrations as a risk factor for childhood obesity and adverse body-fat distribution. Int J Obes (Lond). 2016;40(10):1503-1509. [DOI] [PubMed] [Google Scholar]
- 18.Vehmeijer FOL, Santos S, Gaillard R, et al. Associations of hair cortisol concentrations with general and organ fat measures in childhood. J Clin Endocrinol Metab. 2021;106(2):e551-e561. [DOI] [PubMed] [Google Scholar]
- 19.Gerber M, Endes K, Brand S, et al. In 6- to 8-year-old children, hair cortisol is associated with body mass index and somatic complaints, but not with stress, health-related quality of life, blood pressure, retinal vessel diameters, and cardiorespiratory fitness. Psychoneuroendocrinology. 2017;76:1-10. [DOI] [PubMed] [Google Scholar]
- 20.Genitsaridi SM, Karampatsou S, Papageorgiou I, et al. Hair cortisol concentrations in overweight and obese children and adolescents. Horm Res Paediatr. 2019;92(4):229-236. [DOI] [PubMed] [Google Scholar]
- 21.Petimar J, Rifas-Shiman SL, Hivert MF, Fleisch AF, Tiemeier H, Oken E. Prenatal and childhood predictors of hair cortisol concentration in mid-childhood and early adolescence. PloS One. 2020;15(2):e0228769. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Kooijman MN, Kruithof CJ, van Duijn CM, et al. The Generation R Study: design and cohort update 2017. Eur J Epidemiol. 2016;31(12):1243-1264. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Tukey JW.Exploratory Data Analysis. Reading, PA: Addison-Wesley; 1977. [Google Scholar]
- 24.Vehmeijer FOL, Santos S, de Rijke YB, et al. Supplemental data for “Associations of hair cortisol concentrations with cardio-metabolic risk factors in childhood.” https://figshare.com/s/491222eeb214ee6472b4. Deposited January 1, 2021.
- 25.Noppe G, de Rijke YB, Dorst K, van den Akker EL, van Rossum EF. LC-MS/MS-based method for long-term steroid profiling in human scalp hair. Clin Endocrinol (Oxf). 2015;83(2):162-166. [DOI] [PubMed] [Google Scholar]
- 26.Wong SN, Tz Sung RY, Leung LC. Validation of three oscillometric blood pressure devices against auscultatory mercury sphygmomanometer in children. Blood Press Monit. 2006;11(5):281-291. [DOI] [PubMed] [Google Scholar]
- 27.Kruithof CJ, Kooijman MN, van Duijn CM, et al. The Generation R Study: Biobank update 2015. Eur J Epidemiol. 2014;29(12):911-927. [DOI] [PubMed] [Google Scholar]
- 28.Onyenekwu CP, Hoffmann M, Smit F, Matsha TE, Erasmus RT. Comparison of LDL-cholesterol estimate using the Friedewald formula and the newly proposed de Cordova formula with a directly measured LDL-cholesterol in a healthy South African population. Ann Clin Biochem. 2014;51(Pt 6):672-679. [DOI] [PubMed] [Google Scholar]
- 29.Friedewald WT, Levy RI, Fredrickson DS. Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge. Clin Chem. 1972;18(6):499-502. [PubMed] [Google Scholar]
- 30.Jaddoe VW, de Jonge LL, Hofman A, Franco OH, Steegers EA, Gaillard R. First trimester fetal growth restriction and cardiovascular risk factors in school age children: population based cohort study. BMJ. 2014;348:g14. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Steinberger J, Daniels SR, Eckel RH, et al. ; American Heart Association Atherosclerosis, Hypertension, and Obesity in the Young Committee of the Council on Cardiovascular Disease in the Young; Council on Cardiovascular Nursing; and Council on Nutrition, Physical Activity, and Metabolism . Progress and challenges in metabolic syndrome in children and adolescents: a scientific statement from the American Heart Association Atherosclerosis, Hypertension, and Obesity in the Young Committee of the Council on Cardiovascular Disease in the Young; Council on Cardiovascular Nursing; and Council on Nutrition, Physical Activity, and Metabolism. Circulation. 2009;119(4):628-647. [DOI] [PubMed] [Google Scholar]
- 32.Gishti O, Gaillard R, Manniesing R, et al. Fetal and infant growth patterns associated with total and abdominal fat distribution in school-age children. J Clin Endocrinol Metab. 2014;99(7):2557-2566. [DOI] [PubMed] [Google Scholar]
- 33.Toemen L, Gishti O, Vogelezang S, et al. Cross-sectional population associations between detailed adiposity measures and C-reactive protein levels at age 6 years: the Generation R Study. Int J Obes (Lond). 2015;39(7):1101-1108. [DOI] [PubMed] [Google Scholar]
- 34.Silva CCV, Vehmeijer FOL, El Marroun H, Felix JF, Jaddoe VWV, Santos S. Maternal psychological distress during pregnancy and childhood cardio-metabolic risk factors. Nutr Metab Cardiovasc Dis. 2019;29(6):572-579. [DOI] [PubMed] [Google Scholar]
- 35.Sterne JA, White IR, Carlin JB, et al. Multiple imputation for missing data in epidemiological and clinical research: potential and pitfalls. BMJ. 2009;338:b2393. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Stalder T, Steudte-Schmiedgen S, Alexander N, et al. Stress-related and basic determinants of hair cortisol in humans: a meta-analysis. Psychoneuroendocrinology. 2017;77:261-274. [DOI] [PubMed] [Google Scholar]
- 37.Abraham SB, Rubino D, Sinaii N, Ramsey S, Nieman LK. Cortisol, obesity, and the metabolic syndrome: a cross-sectional study of obese subjects and review of the literature. Obesity (Silver Spring). 2013;21(1):E105-E117. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Strait RB, Slattery MJ, Carrel AL, Eickhoff J, Allen DB. Salivary cortisol does not correlate with metabolic syndrome markers or subjective stress in overweight children. J Child Obes. 2018;3(2):8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Reinehr T, Kulle A, Wolters B, et al. Relationships between 24-hour urinary free cortisol concentrations and metabolic syndrome in obese children. J Clin Endocrinol Metab. 2014;99(7):2391-2399. [DOI] [PubMed] [Google Scholar]
- 40.Soriano-Rodríguez P, Osiniri I, Grau-Cabrera P, et al. Physiological concentrations of serum cortisol are related to vascular risk markers in prepubertal children. Pediatr Res. 2010;68(5):452-455. [DOI] [PubMed] [Google Scholar]
- 41.Weigensberg MJ, Toledo-Corral CM, Goran MI. Association between the metabolic syndrome and serum cortisol in overweight Latino youth. J Clin Endocrinol Metab. 2008;93(4):1372-1378. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Guzzetti C, Pilia S, Ibba A, Loche S. Correlation between cortisol and components of the metabolic syndrome in obese children and adolescents. J Endocrinol Invest. 2014;37(1):51-56. [DOI] [PubMed] [Google Scholar]
- 43.Walker BR. Activation of the hypothalamic-pituitary-adrenal axis in obesity: cause or consequence? Growth Horm IGF Res. 2001;11(Suppl A):S91-S95. [DOI] [PubMed] [Google Scholar]
- 44.Stalder T, Kirschbaum C, Alexander N, et al. Cortisol in hair and the metabolic syndrome. J Clin Endocrinol Metab. 2013;98(6):2573-2580. [DOI] [PubMed] [Google Scholar]
- 45.Rosmond R, Dallman MF, Björntorp P. Stress-related cortisol secretion in men: relationships with abdominal obesity and endocrine, metabolic and hemodynamic abnormalities. J Clin Endocrinol Metab. 1998;83(6):1853-1859. [DOI] [PubMed] [Google Scholar]
- 46.Fraser R, Ingram MC, Anderson NH, Morrison C, Davies E, Connell JM. Cortisol effects on body mass, blood pressure, and cholesterol in the general population. Hypertension. 1999;33(6):1364-1368. [DOI] [PubMed] [Google Scholar]
- 47.Roy MP, Kirschbaum C, Steptoe A. Psychological, cardiovascular, and metabolic correlates of individual differences in cortisol stress recovery in young men. Psychoneuroendocrinology. 2001;26(4):375-391. [DOI] [PubMed] [Google Scholar]
- 48.Maduka IC, Neboh EE, Ufelle SA. The relationship between serum cortisol, adrenaline, blood glucose and lipid profile of undergraduate students under examination stress. Afr Health Sci. 2015;15(1):131-136. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 49.Gray NA, Dhana A, Van Der Vyver L, Van Wyk J, Khumalo NP, Stein DJ. Determinants of hair cortisol concentration in children: a systematic review. Psychoneuroendocrinology. 2018;87:204-214. [DOI] [PubMed] [Google Scholar]
- 50.Almeida DM, Piazza JR, Stawski RS. Interindividual differences and intraindividual variability in the cortisol awakening response: an examination of age and gender. Psychol Aging. 2009;24(4):819-827. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 51.Van Cauter E, Leproult R, Kupfer DJ. Effects of gender and age on the levels and circadian rhythmicity of plasma cortisol. J Clin Endocrinol Metab. 1996;81(7):2468-2473. [DOI] [PubMed] [Google Scholar]
- 52.Anagnostis P, Athyros VG, Tziomalos K, Karagiannis A, Mikhailidis DP. Clinical review: the pathogenetic role of cortisol in the metabolic syndrome: a hypothesis. J Clin Endocrinol Metab. 2009;94(8):2692-2701. [DOI] [PubMed] [Google Scholar]
- 53.Guignat L, Bertherat J. The diagnosis of Cushing’s syndrome: an Endocrine Society clinical practice guideline: commentary from a European perspective. Eur J Endocrinol. 2010;163(1):9-13. [DOI] [PubMed] [Google Scholar]
- 54.Isidori AM, Graziadio C, Paragliola RM, et al. ; ABC Study Group . The hypertension of Cushing’s syndrome: controversies in the pathophysiology and focus on cardiovascular complications. J Hypertens. 2015;33(1):44-60. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 55.Pivonello R, Faggiano A, Lombardi G, Colao A. The metabolic syndrome and cardiovascular risk in Cushing’s syndrome. Endocrinol Metab Clin North Am. 2005;34(2):327-339, viii. [DOI] [PubMed] [Google Scholar]
- 56.de Kruijff I, Noppe G, Kieviet N, et al. LC-MS/MS-based reference intervals for hair cortisol in healthy children. Psychoneuroendocrinology. 2020;112:104539. [DOI] [PubMed] [Google Scholar]
- 57.Hancox RJ, Landhuis CE. Correlation between measures of insulin resistance in fasting and non-fasting blood. Diabetol Metab Syndr. 2011;3(1):23. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 58.Freiberg JJ, Tybjaerg-Hansen A, Jensen JS, Nordestgaard BG. Nonfasting triglycerides and risk of ischemic stroke in the general population. JAMA. 2008;300(18):2142-2152. [DOI] [PubMed] [Google Scholar]
- 59.Langsted A, Nordestgaard BG. Nonfasting versus fasting lipid profile for cardiovascular risk prediction. Pathology. 2019;51(2):131-141. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The datasets generated during and/or analyzed during the current study are not publicly available but are available from the corresponding author on reasonable request.