- Email: [email protected]
Socioeconomic status in children is associated with hair cortisol levels as a biological measure of chronic stress
Psychoneuroendocrinology 65 (2016) 9–14
Contents lists available at ScienceDirect
Psychoneuroendocrinology journal homepage: www.elsevier.com/locate/psyneuen
Socioeconomic status in children is associated with hair cortisol levels as a biological measure of chronic stress J. Vliegenthart a , G. Noppe a,b , E.F.C. van Rossum b , J.W. Koper b , H. Raat c , E.L.T. van den Akker a,∗ a Department of Pediatrics, Division of Endocrinology, Sophia Children’s Hospital, Erasmus MC, University Medical Center Rotterdam, Rotterdam, The Netherlands b Department of Internal Medicine, Division of Endocrinology, Erasmus MC, University Medical Center Rotterdam, Rotterdam, The Netherlands c Department of Public Health, Erasmus MC, University Medical Center Rotterdam, Rotterdam, The Netherlands
a r t i c l e
i n f o
Article history: Received 30 June 2015 Received in revised form 25 November 2015 Accepted 30 November 2015 Keywords: SES Scalp hair Cortisol Children Stress
a b s t r a c t Introduction: Low socioeconomic status (SES) may be associated with a high risk of lifestyle-related diseases such as cardiovascular diseases. There is a strong association between parental SES, stress and indicators of child health and adult health outcome. The exact mechanisms underlying this association have not yet been fully clarified. Low SES may be associated with chronic stress, which may lead to activation of the hypothalamic-pituitary-adrenal (HPA)-axis, resulting in a higher circulating level of the stress hormone cortisol. Therefore, chronic stress may mediate the association between low SES and elevated cortisol levels and its adverse outcomes. Aim: We investigated whether SES was associated with a chronic measure of cortisol exposure in a child population. Methods: Cortisol and cortisone were measured in scalp hair in 270 children and adolescents, aged 4–18 years, enrolled through school visits. Neighborhood level SES was based on a score developed by the Netherlands Institute for Social Research using postal codes, and this includes neighborhood measures of income education and unemployment. Maternal and paternal education level were used as indicators of family SES. Results: Neighborhood level socioeconomic status score was significantly associated with hair cortisol (ˇ = −0.103, p = 0.007, 95%CI [−0.179, −0.028]) and hair cortisone (ˇ = −0.091, p = 0.023, 95%CI [−0.167, −0.015]), adjusted for age and sex. Additionally, hair cortisol was significantly correlated with maternal education level and hair cortisone was significantly correlated with paternal education level. Conclusion: The results of our study suggest that the widely shown association between low family SES and adverse child health outcomes may be mediated by chronic stress, given the chronically higher levels of cortisol in children and adolescents in families with low SES. It is especially notable that the association between SES and cortisol was already found in children of young age as this can have major consequences, such as increased risk of cardio metabolic diseases in later life. © 2015 Elsevier Ltd. All rights reserved.
1. Introduction There is a strong association between parental socioeconomic status (SES) and indicators of child health (Evans and Kantrowitz,
∗ Corresponding author: E.L.T. van den Akker, Erasmus MC-SP1536, P.O. box 2060, 3000 CB Rotterdam, The Netherlands. E-mail addresses: [email protected] (J. Vliegenthart), [email protected] (G. Noppe), [email protected] (E.F.C. van Rossum), [email protected] (J.W. Koper), [email protected] (H. Raat), [email protected] (E.L.T. van den Akker). http://dx.doi.org/10.1016/j.psyneuen.2015.11.022 0306-4530/© 2015 Elsevier Ltd. All rights reserved.
2002; Evans and Kim, 2010; Clearfield et al., 2014). The association of parental SES is partly driven by environmental exposures, such as nutrition, lifestyle and healthcare. However, these variables do not explain the complete picture as the association exists even in the context of adequate healthcare and nutrition (Adler et al., 1994). Low SES is also associated with increased stress by violence in the neighborhood and at home, disorganization in school environments, environmental toxins, crowding, and noise (Evans and Kantrowitz, 2002; Evans, 2006; Evans and Kim, 2010). Psychological stressors can elicit cortisol activation through the biological stress response, depending on the characteristics of the stressor
10
J. Vliegenthart et al. / Psychoneuroendocrinology 65 (2016) 9–14
(Dickerson and Kemeny, 2004; Felner, 2014). This involves a physiological response to situations in which the individual experiences stress. The stress response leads to activation of the hypothalamicpituitary-adrenal (HPA)-axis. Elevated HPA-axis activity results in increased levels of circulating cortisol. In animal studies, where randomization is feasible, intervention trials show that chronic stress exposure leads to activation of the HPA-axis. For example, cortisol concentrations measured in hair of rhesus macaques were found to be elevated in response to a prolonged environmental stressor (being relocated to a newly constructed building) (Davenport et al., 2006). In humans, where randomization trials are not feasible, association studies show a link between low SES, chronic stress and dysregulation of the HPA-axis (Hajat et al., 2010; Sheridan et al., 2013; Vaghri et al., 2013; Clearfield et al., 2014). Van Uum et al., for example, found that hair cortisol contents were significantly greater in chronic pain patients than in controls and that HCC was also increased in patients with major chronic stress (Van Uum et al., 2008). Several studies found an association between lower SES and higher levels of cortisol, many found no association, with some reporting an opposite relationship (Lupien et al., 2000; Dowd et al., 2009; Sheridan et al., 2013; Clearfield et al., 2014). Lupien et al. measured basal morning salivary cortisol levels in 217 children with different SES levels ranging from 6 to 10 years of age. They found that children with low SES had significantly higher salivary cortisol levels than children with high SES (Lupien et al., 2000). Clearfield et al. compared diurnal salivary cortisol and maternal-infant synchrony in 32 infants, aged 6–12 months and their mothers (with low or high SES). Low-SES infants and mothers exhibited higher average salivary cortisol, compared to high-SES infants (Clearfield et al., 2014). Sheridan et al. (2013) compared maternal self-rated social status and salivary cortisol in 40 children, aged 8–12 years and also found a negative association (Sheridan et al., 2013). However some other studies even found opposite associations e.g. low SES linked to lower cortisol levels (Fuller-Rowell et al., 2012, Sturge-Apple et al., 2012). Thus, both activation and suppression of the HPA-axis have been found in these studies. SES is associated with dysregulation of the HPA-axis, but literature is inconsistent concerning the nature of this association, varying from a positive association to a negative association, a blunted cortisol response or a flattened daily rhythm (Lupien et al., 2000; Dowd et al., 2009; Hajat et al., 2010; Sheridan et al., 2013; Vaghri et al., 2013; Clearfield et al., 2014). These inconsistencies may arise from the way the cortisol is measured. Cortisol concentrations in these studies are mainly measured in saliva, urine, or serum. However, due to the circadian rhythm, pulsatile secretion, and the daily variation in the secretion of cortisol, none of these sampling matrices is an optimal measure of chronic cortisol concentrations. A more suitable matrix appears to be scalp hair, because hair samples can provide a historical timeline of cortisol exposure (Manenschijn et al., 2011a; Manenschijn et al., 2012; Stalder and Kirschbaum, 2012). Scalp hair at the posterior vertex displays regular growth at an average rate of 1 cm/month (Hayashi et al., 1991; Sauve et al., 2007) providing a retrospective calendar of mean cortisol exposure and HPA-axis activity of recent months to years. Hair cortisol analysis has previously been applied to investigate different causes of stress, such as shift work, unemployment and chronic pain (Davenport et al., 2006; Dettenborn et al., 2010; Manenschijn et al., 2011b; Staufenbiel et al., 2013). In addition to cortisol, cortisone might be a useful biomarker for stress research. The biologically inactive hormone cortisone mainly originates from cortisol metabolism by enzymatic activity of 11hydroxysteroid dehydrogenase type 2. Cortisone levels in scalp hair of young children experiencing psychosocial stress were found to be elevated (Vanaelst et al., 2013).
Chronic stress can negatively impact health, resulting in an increased incidence of obesity, type 2 diabetes mellitus, and cardiovascular disease (Everson-Rose and Lewis, 2005; Manenschijn et al., 2013). The risk of obesity in young children from lower SES groups is 4 times higher than that in the higher SES groups (Veldhuis et al., 2013). Low SES, in turn, is associated with a high risk of lifestylerelated diseases such as cardiovascular diseases (Shrewsbury and Wardle, 2008). These cardiovascular diseases are partly lifestylerelated, but could also be caused by other factors such as stress or its biological mediator cortisol (Manenschijn et al., 2013). We hypothesized that children with a lower SES experience more stress, leading to increased activation of the HPA-axis resulting in higher levels of cortisol. We investigated whether SES was associated with chronic measurements of cortisol and cortisone, measured in scalp hair in children and adolescents from the general population of the Netherlands, aged 4–18 years. 2. Methods We included children from the general population of the Netherlands in a cross-sectional observational study. Children aged 4–18 years were included. Data collection was performed between January 2011 and July 2014. Two elementary schools (one in the city of Amersfoort and one in Rotterdam) and two secondary schools in Rotterdam participated in the study. The schools are located in different districts of the city (difference in SES, income, etc.) and are of different educational levels (low, mid and high) to create the best possible reflection of the general population. Parents and children of school classes (1 class per grade) were invited to participate in the study. All parents and children above the age of 12 years provided informed consent, children under the age of 12 years provided informed assent. The participation rate approximated 50%. Exclusion criteria for the children were: suffering from chronic diseases, chronic use of medication, in particular glucocorticoids (both based on a completed questionnaire), or scalp hair length shorter than 2 centimeters. Data were encoded before they were entered in the database. This study was approved by the Medical Ethics Committee of the Erasmus MC (MEC-2012-25). Anthropometric data were collected in all children, including height (cm) and weight (kg). Standing height was measured in centimeters with the precision of 1 mm, with a wall mounted stadiometer. Body weight was measured in kilograms with the precision of 2 decimal places with a calibrated electric-weight model. The standard deviation scores (SDS) of weight, height and body mass index (BMI: body mass divided by the square of the body height) were calculated based on data from the fourth Dutch nationwide growth study 1997, using Growth Analyser (Fredriks et al., 2000). 2.1. Hair cortisol Hair samples of about 100 hairs were taken from the posterior vertex of the scalp, as close to the scalp as possible, using small surgical scissors. The hair was taped to a piece of paper and the scalp end was marked. The samples were stored in an envelope at room temperature until analysis (Manenschijn et al., 2011a). The proximal 3 cm of hair, reflecting the most recent 3 months, was cut in 1 cm segments, weighed on an electronic scale and transferred to disposable glass tubes. Subsequently, the hair samples were washed with isopropanol for 2 min, after which the samples were dried for at least 2 days. Deuterated cortisol was added as an internal standard, and steroids were extracted in 1.5 ml of methanol (MeOH) for 18 h at 25 ◦ C. After extraction, the samples
J. Vliegenthart et al. / Psychoneuroendocrinology 65 (2016) 9–14
are centrifuged for 5 min at 4500 rpm at 4 ◦ C. The supernatant is transferred to glass tubes, and the extract was dried at 37 ◦ C under a stream of nitrogen. The extract is then reconstituted in 2% MeOH in water and further cleaned using “solid phase extraction” on a 96-well plate SPE (Oasis HLB 96-well plate, 30 mg, Waters Corporation, Milford, USA), evaporated under a stream of nitrogen, and reconstituted in 30% MeOH in water for analysis. The cortisol was then analyzed on a Xevo TQS Liquid Chromatography-Tandem Mass Spectrometer (Waters Corporation, Milford, USA) (Noppe et al., 2015). The measured concentration was then divided by the hair weight in order to calculate the concentration per mg hair. The outcome is called hair cortisol concentration (HCC) and is in picograms per milligram. The analysis of hair cortisone is identical to hair cortisol (Noppe et al., 2015). Intra-individual coefficients of variation were below 15% for cortisol and cortisone (Noppe et al., 2015). 2.2. Socioeconomic status As an indicator of socioeconomic status we used Dutch neighborhood level status scores, which are scores calculated by the Netherlands Institute for Social Research (Sociaal en Cultureel Planbureau; SCP) that specify the relative social status of a neighborhood, compared to other neighborhoods in the Netherlands. The higher the score, the higher the status of a neighborhood. A low score indicates a low neighborhood status. The status score in 2010 reaches from −7.25 to 3.19. Within our study population, there are 91 different zip codes for all participants and the mean number of residents per zip code district is 8922 and varies between 835 and 20,285 people. The social status of a neighborhood is derived from the following characteristics of the people who live there: education, income and labor market status. Neighborhoods in this case are 4-digit postal code areas (only the numbers, not letters). The status scores are composed of: the average income in a neighborhood, the percentage of people with a low income, the percentage of low-educated people and the percentage of people not working. In a period of four years, data is collected per six-digit (the four numbers plus two letters) postal code area. Data are aggregated to the level of the four-digit postcode area. This means that the status scores of 2010 contain data for the period 2006–2010. Through factor analysis, these characteristics are summarized in a composite feature: the social status (The Netherlands Institute for Social Research, 2012). In addition, at the level of the family, we used both maternal and paternal reported highest education level as indicator of family SES. At the elementary schools the parents completed a questionnaire in which they filled in their educational level. At the secondary schools the children completed the questionnaire, but when there was indistinctness about parental educational level or other unanswered questions, the parents were contacted to clarify this. Education level was classified according to the Dutch education system: no education, primary education (elementary school), secondary education (secondary school), vocational studies, community college and university. Of these categories we made three education levels: low (no and primary education), mid (secondary education and vocational studies) and high (community college and university) (Statistics Netherlands 2004a). 2.3. Ethnicity Ethnicity was determined by country of birth of the children and their parents, derived from the completed questionnaires. Ethnicity is divided into three groups: native Dutch, western background and non-western background. Ethnicities are determined by a report of Statistics Netherlands. When a child, or one or both parents were not born in the Netherlands, he is seen as an immigrant. A western immigrant comes from one of the countries in Europe (excluding
11
Turkey), North America and Oceania, or Indonesia or Japan. A child whose ethnic background is one of the countries in the continents of Africa, Latin America and Asia (excluding Indonesia and Japan) or Turkey is a non-western immigrant (Statistics Netherlands, 2004b). 2.4. Statistics Statistical analysis was performed using IBM SPSS Statistics 20 (IBM Corp. Armonk, New York, NY, USA). The Gaussian distribution of variables was assessed by the Kolmogorov–Smirnov test. We regarded a variable to deviate from the normal distribution if the p-value was smaller than 0.05. Non-normally distributed variables were log-transformed to achieve a normal distribution. Because the SES-score was not normally distributed despite log-transformation, we calculated the Spearman’s correlation coefficient between the SES-score and HCC. Using a hierarchical linear regression approach, we assessed the association between SES-score, adjusted for sex and age, with HCC as the primary model, adding ethnicity in the second block. The random effects in the hierarchical model were included as random intercept effects. The  and 95% CI of SES-score on HCC before and after adjustment for ethnicity was reported. 3. Results In total, 296 children participated in the study, 17 children were excluded because of missing analysis of hair cortisol and 9 children were excluded because of a missing SES score. Therefore, the total number of children included in our study is 270. Of these, 131 are in elementary school and 139 children are in secondary school. Baseline characteristics of the total group, and of two subgroups based on school type, are shown in Table 1. Of all participants, 7% was overweight (BMI–SDS > 1.1) and 2% was obese (BMI–SDS > 2.3) (Fredriks et al., 2000). There is a significant negative correlation between status score and hair cortisol concentration (r = −0.193; p = 0.001), as shown in Fig. 1. With the Spearman’s correlation test we observed significant associations between maternal education and hair cortisol, and between paternal education and hair cortisol and cortisone (Table 2). No associations between cortisol and maternal or paternal education were found with the linear regression model. Combination of maternal and paternal education into one variable gave the same result. With linear regression analysis, adjusted for sex and age, we found a significant association in the total group of both hair cortisol (ˇ = −0.103, p = 0.007, 95%CI [−0.179, −0.028]) and hair cortisone (ˇ = −0.091, p = 0.023, 95%CI [−0.167, −0.015]) with the SES score (Table 2). When ethnicity was added to the linear regression model of hair cortisol, the ˇ changed from −0.103 to −0.064 (95%CI [−0.155, 0.027]). In the linear regression model of hair cortisone, the ˇ changed from −0.091 to −0.049 (95%CI [−0.142,−0.044]) (Table 2). Adjustments of the association for BMI–SDS did not change these results. 4. Discussion This study shows a negative association between indicators of family socioeconomic status and hair cortisol concentration as well as a negative association between socioeconomic status and hair cortisone concentration in children and adolescents. This association was independent of BMI, but the strength of the association was influenced by ethnicity. To our knowledge, this is the first study reporting the association between SES and chronic cortisol exposure measured in scalp hair in children aged 4–18 years. Only one previous study, by Vaghri et al., studied hair cortisol and SES in
12
J. Vliegenthart et al. / Psychoneuroendocrinology 65 (2016) 9–14
Table 1 Baseline characteristics of the whole group of children from the general population that participated in this study and divided by participant educational level.
Total Sex Age BMI–SDS Ethnicity
SES score Maternal education
Paternal education
Cortisol (pg/mg) Cortisone (pg/mg)
N= Male (%) Mean (range) Mean (SD) Native Dutch (N%) Western background (N%) Non-western background (N%) Missing (N%) Mean (SD) N (%) Low Mid N (%) N (%) High N (%) Low N (%) Mid High N (%) Mean (range) Mean (range)
Elementary school
Secondary school
Total
131 70 (53%) 7.97 (3.66–12.58) −0.30 (1.01) 99 (75.6%) 15 (11.5%) 3 (2.3%) 14 (10.6%) 0.85 (0.92) 0 17 (18%) 75 (82%) 0 14 (15%) 79 (85%) 3.05 (0.28–38.26) 13.55 (1.99–159.89)
139 70 (50%) 15.17 (11.89–18.78) 0.15 (.98) 87 (62.6%) 12 (8.6%) 39 (28.1%) 1 (0.7%) −0.14 (1.86) 5 (4%) 35 (29%) 83 (67%) 3 (3%) 31 (26%) 86 (71%) 3.67 (0.40–34.32) 16.23 (3.93–50.30)
270 140 (52%) 11.67 (3.66–18.78) −0.07 (1.01) 186 (68.9%) 27 (10.0%) 42 (15.6%) 15 (5.5%) 0.34 (1.86) 5 (2%) 52 (24%) 158 (74%) 3 (1%) 45 (21%) 165 (78%) 3.37 (0.28–38.26) 14.94 (1.99–159.89)
SDS: standard deviation score, SES: socioeconomic status, pg/mg: picograms per milligram. Cortisol and cortisone are measured in hair.
Fig. 1. Correlation between SES and HCC.
HCC is in picograms per milligram (pg/mg) and is shown on a logarithmic scale.
Table 2 Associations between measures of SES and hair cortisol or hair cortisone.
Status score Corrected for age & sex Corrected for age & sex & BMI Corrected for age & sex & ethnicity Status score Paternal education Maternal education
Cortisol
Cortisone
ˇ = −0.103** [−0.179,−0.028] ˇ = −0.094* [−0.170,−0.017] ˇ = −0.064[−0.155,0.027] = −0.193** = −0.135* = −0.164*
ˇ = −0.091* [−0.167,−0.015] ˇ = −0.081* [−0.158,−0.004]  = -0.049[−0.142,0.044] = −0.159** = −0.199* = −0.120
Linear regression analyses coefficients, ˇ. Spearman’s correlation coefficient .
J. Vliegenthart et al. / Psychoneuroendocrinology 65 (2016) 9–14
children, with mean age 4.6 years. A negative association between parental education and hair cortisol was found (Vaghri et al., 2013). Previous studies have analyzed the association between SES and cortisol, but did not use a chronic measure of cortisol exposure, such as scalp hair analysis. Authors generally reported salivary cortisol concentrations (Lupien et al., 2000; Sheridan et al., 2013; Clearfield et al., 2014). Studies vary widely in sample size, selection criteria and number of measures collected. Previous study outcomes are not consistent with regard to the nature of the relation between SES and cortisol. Although several studies found an association between lower SES and higher levels of cortisol, many found no association, with some reporting an opposite relationship (Lupien et al., 2000; Dowd et al., 2009; Sheridan et al., 2013; Clearfield et al., 2014). We found ethnicity to be an important confounder and we cannot exclude a role of ethnicity as a driving force. Relatively few studies have investigated SES in combination with differences by ethnicity and cortisol. In adults, Stalder and Kirschbaum (2012), found an association between salivary cortisol and ethnicity and SES in a multi-ethnic population study (Hajat et al., 2010). In African American and Hispanic adolescents, significantly flatter diurnal cortisol slopes (due to higher evening cortisol) were found, which could have negative health consequences (DeSantis et al., 2007). In this study, examined socio environmental factors (including parental SES) failed to explain the observed racial/ethnic differences in diurnal cortisol rhythms (DeSantis et al., 2007). HCC in volunteers from settlement communities in Kenya were significantly higher in females, divorced volunteers and those whose income was below minimum wage. However, no evidence was found for increased chronic stress (measured by hair cortisol content) between members of slum settlements that were adjacent and those that were distant to large floriculture farms (Henley et al., 2014). However, we cannot directly compare these results with our study because of the different SES measures that were used. In our study there was a change of  with a loss of significance in the linear regression analysis between HCC and the SES-score after adjustment for ethnicity. Ethnicity is so highly linked to SES that to address this unequal distribution, studies with much higher numbers are needed to unravel the effects of both SES and ethnicity independently. Ethnicity as defined in the current study may be a misnomer, as it is not a comparison of multiple ethnicities, but rather the comparison of offspring of native Dutch parents versus immigrants, which usually represent a subgroup of an ethnical population. In the Netherlands, these ethnical minorities tend to be lower educated, have less income and have more health problems. The underlying causes for the association between SES and hair cortisol is unclear, it seems likely to hypothesize that low SES leads to psychological stress and thereby to elevated hair cortisol. The generalizability of these results remains to be investigated in future studies of larger populations with larger numbers of specific ethnicities, which offer the possibility of elucidating the most important confounding factors. The negative association between SES and cortisol, observed in the present study, is in line with most previous studies measuring salivary cortisol associations with SES in children. Markedly higher cortisol levels have been found in low SES adolescents and children compared to middle and high SES adolescents and children from diverse age categories: e.g. aged 9–18 years (Chen et al., 2010), aged 6–9 years (Lupien et al., 2000; Evans and English, 2002) and in infants as young as 6–12 months of age (Clearfield et al., 2014). However some other studies did not find an association between SES and salivary cortisol levels in 8–11 year old children (Sheridan et al., 2013), or even found opposite associations e.g. low SES linked to lower cortisol levels (Fuller-Rowell et al., 2012; Sturge-Apple., 2012).
13
SES can be scored in different ways. In most studies, SES was scored using parental education level or income. In our study we used three indicators of family SES, i.e. maternal and paternal educational level and a neighborhood level SES score. When we analyzed educational level separately, we found significant negative correlations between maternal education level and hair cortisol concentration, and also between paternal education level and both hair cortisol and cortisone concentrations. The negative association that was found implies that children with low educated parents are likely to have higher stress hormone levels. A limitation of cross-sectional observational studies is to determine cause and consequence of the associations found. Although the underlying cause for the association between SES and hair cortisol in our study is unclear, it seems likely that chronic stress may mediate the association between low SES and elevated cortisol levels. Also, the generalizability of these results remains to be investigated in future studies of larger populations, which offer the possibility of pinning down the most important confounding factors. On the other hand, the magnitude of the effect in our study is relatively large, with an effect size ˇ of −0,103 for the association between status score and (log-transformed) hair cortisol, because both parameters can be influenced by multiple other factors such as non-registered major life events, individual, social or family factors. SES scores can be measured in many different ways, each with its own advantage and disadvantages. We derived SES from postal codes, representing mean SES values from a street area and not from an individual. As a consequence, SES distribution is wider in secondary school children compared to elementary school children. The advantage of neighborhood SES analysis is that multiple SESrelated factors were included in the calculation of this score, which is a valid method to compare SES (Knol, 2012). A strength of our study is that we used a new, highly sensitive and reliable method to measure chronic cortisol exposure in scalp hair. Because of this method we have a better understanding of cortisol concentrations over a prolonged period of time and not just one moment in time (e.g. serum or saliva). We measured the hair cortisol and hair cortisone concentrations with liquid chromatography tandem mass spectrometry (LC–MS/MS), a very sensitive method to accurately measure steroids, which is advocated to be the golden standard to measure steroids (Handelsman and Wartofsky 2013; Noppe et al., 2015). The above-mentioned LC–MS/MS has the advantage that besides cortisol, we could also measure cortisone, which could be of added value for reliable measurement of stress (Vanaelst et al., 2013). Both cortisol and cortisone showed the same relationship with SES. Low SES is thought to lead to negative health consequences. Multiple studies show that chronic elevated stress has negative health outcomes (Everson-Rose and Lewis 2005; Shrewsbury and Wardle 2008). Chronic elevated cortisol levels were shown to be associated with increased risk of the metabolic syndrome, increased visceral fat mass, hypertension, insulin resistance, and dyslipidemia, which all result in an increased cardiovascular risk (Manenschijn et al., 2013; Stalder et al., 2013). The results of our study suggest that the widely shown association between low family socioeconomic status and adverse child health outcomes may be mediated by chronic stress, given the chronically higher levels of cortisol in children and adolescents in families with low socioeconomic status. The measurement of scalp hair cortisol seems to provide a unique tool to study chronic stress. It is especially notable that the association between SES and cortisol was already found in children of young age as this can have important consequences later in life, and poses an increased risk of cardio metabolic diseases in adulthood. The next step for future studies is to elucidate the role of cortisol in more specific pathways
14
J. Vliegenthart et al. / Psychoneuroendocrinology 65 (2016) 9–14
of adverse socioeconomic circumstances and health consequences in childhood. Conflict of interest The authors have no conflict of interests or financial statements to disclose. Acknowledgements The authors are grateful to Ms. Margriet E. Bisschoff, research nurse, Erasmus MC, Rotterdam, The Netherlands, for her help in participant recruitment and measurements. We also thank the primary and secondary schools, the children and their parents for their participation. We are also grateful to the Thrasher research Fund for financial support (grant TRF-11643). References Adler, N.E., Boyce, T., Chesney, M.A., Cohen, S., Folkman, S., Kahn, R.L., Syme, S.L., 1994. Socioeconomic status and health. The challenge of the gradient. Am. Psychol. 49 (1), 15–24. Chen, E., Cohen, S., Miller, G.E., 2010. How low socioeconomic status affects 2-year hormonal trajectories in children. Psychol. Sci. 21 (1), 31–37. Clearfield, M.W., Carter-Rodriguez, A., Merali, A.R., Shober, R., 2014. The effects of SES on infant and maternal diurnal salivary cortisol output. Infant Behav. Dev. 37 (3), 298–304. Davenport, M.D., Tiefenbacher, S., Lutz, C.K., Novak, M.A., Meyer, J.S., 2006. Analysis of endogenous cortisol concentrations in the hair of rhesus macaques. Gen. Comp. Endocrinol. 147 (3), 255–261. DeSantis, A.S., Adam, E.K., Doane, L.D., Mineka, S., Zinbarg, R.E., Craske, M.G., 2007. Racial/ethnic differences in cortisol diurnal rhythms in a community sample of adolescents. J. Adolesc. Health 41 (1), 3–13. Dettenborn, L., Tietze, A., Bruckner, F., Kirschbaum, C., 2010. Higher cortisol content in hair among long-term unemployed individuals compared to controls. Psychoneuroendocrinology 35 (9), 1404–1409. Dickerson, S.S., Kemeny, M.E., 2004. Acute stressors and cortisol responses: a theoretical integration and synthesis of laboratory research. Psychol. Bull. 130 (3), 355–391. Dowd, J.B., Simanek, A.M., Aiello, A.E., 2009. Socio-economic status, cortisol and allostatic load: a review of the literature. Int. J. Epidemiol. 38 (5), 1297–1309. Evans, G.W., English, K., 2002. The environment of poverty: multiple stressor exposure, psychophysiological stress, and socioemotional adjustment. Child Dev. 73 (4), 1238–1248. Evans, G.W., Kantrowitz, E., 2002. Socioeconomic status and health: the potential role of environmental risk exposure. Annu. Rev. Publ. Health 23, 303–331. Evans, G.W., Kim, P., 2010. Multiple risk exposure as a potential explanatory mechanism for the socioeconomic status-health gradient. Ann. N. Y. Acad. Sci. 1186, 174–189. Evans, G.W., 2006. Child development and the physical environment. Annu. Rev. Psychol. 57, 423–451. Everson-Rose, S.A., Lewis, T.T., 2005. Psychosocial factors and cardiovascular diseases. Annu. Rev. Publ. Health 26, 469–500. Felner, U., 2014. Adrenal gland disorders. In: Endocrine Pathophysiology. Lippincott Williams & Wilkins, pp. 207. Fredriks, A.M., van Buuren, S., Burgmeijer, R.J., Meulmeester, J.F., Beuker, R.J., Brugman, E., Roede, M.J., Verloove-Vanhorick, S.P., Wit, J.M., 2000. Continuing positive secular growth change in The Netherlands 1995–1997. Pediatr. Res. 47 (3), 316–323. Fuller-Rowell, T.E., Doan, S.N., Eccles, J.S., 2012. Differential effects of perceived discrimination on the diurnal cortisol rhythm of African Americans and Whites. Psychoneuroendocrinology 37 (1), 107–118. Hajat, A., Diez-Roux, A., Franklin, T.G., Seeman, T., Shrager, S., Ranjit, N., Castro, C., Watson, K., Sanchez, B., Kirschbaum, C., 2010. Socioeconomic and race/ethnic differences in daily salivary cortisol profiles: the multi-ethnic study of atherosclerosis. Psychoneuroendocrinology 35 (6), 932–943.
Handelsman, D.J., Wartofsky, L., 2013. Requirement for mass spectrometry sex steroid assays in the Journal of Clinical Endocrinology and Metabolism. J. Clin. Endocrinol. Metab. 98 (10), 3971–3973. Hayashi, S., Miyamoto, I., Takeda, K., 1991. Measurement of human hair growth by optical microscopy and image analysis. Br. J. Dermatol. 125 (2), 123–129. Henley, P., Lowthers, M., Koren, G., Fedha, P.T., Russell, E., VanUum, S., Arya, S., Darnell, R., Creed, I.F., Trick, C.G., Bend, J.R., 2014. Cultural and socio-economic conditions as factors contributing to chronic stress in sub-Saharan African communities. Can. J. Physiol. Pharmacol. 92 (9), 725–732. Knol, F., 2012. Statusontwikkeling van wijken in Nederland 1998–2010 (Dutch). Institute for Social Research, The Netherlands. Lupien, S.J., King, S., Meaney, M.J., McEwen, B.S., 2000. Child’s stress hormone levels correlate with mother’s socioeconomic status and depressive state. Biol. Psychiatry 48 (10), 976–980. Manenschijn, L., Koper, J.W., Lamberts, S.W., van Rossum, E.F., 2011a. Evaluation of a method to measure long term cortisol levels. Steroids 76 (10–11), 1032–1036. Manenschijn, L., van Kruysbergen, R.G., de Jong, F.H., Koper, J.W., van Rossum, E.F., 2011b. Shift work at young age is associated with elevated long-term cortisol levels and body mass index. J. Clin. Endocrinol. Metab. 96 (11), E1862–E1865. Manenschijn, L., Koper, J.W., van den Akker, E.L., de Heide, L.J., Geerdink, E.A., de Jong, F.H., Feelders, R.A., van Rossum, E.F., 2012. A novel tool in the diagnosis and follow-up of (cyclic) Cushing’s syndrome: measurement of long-term cortisol in scalp hair. J. Clin. Endocrinol. Metab. 97 (10), E1836–E1843. Manenschijn, L., Schaap, L., van Schoor, N.M., van der Pas, S., Peeters, G.M., Lips, P., Koper, J.W., van Rossum, E.F., 2013. High long-term cortisol levels, measured in scalp hair, are associated with a history of cardiovascular disease. J. Clin. Endocrinol. Metab. 98 (5), 2078–2083. Noppe, G., de Rijke, Y.B., Dorst, K., van den Akker, E.L., van Rossum, E.F., 2015. LC–MS/MS based method for long-term steroid profiling in human scalp hair. Clin. Endocrinol. 83 (Oxf). Sauve, B., Koren, G., Walsh, G., Tokmakejian, S., Van Uum, S.H., 2007. Measurement of cortisol in human hair as a biomarker of systemic exposure. Clin. Invest. Med. 30 (5), E183–191. Sheridan, M.A., How, J., Araujo, M., Schamberg, M.A., Nelson, C.A., 2013. What are the links between maternal social status, hippocampal function, and HPA axis function in children? Dev. Sci. 16 (5), 665–675. Shrewsbury, V., Wardle, J., 2008. Socioeconomic status and adiposity in childhood: a systematic review of cross-sectional studies 1990–2005. Obesity (Silver Spring) 16 (2), 275–284. Stalder, T., Kirschbaum, C., 2012. Analysis of cortisol in hair-state of the art and future directions. Brain Behav. Immun. 26 (7), 1019–1029. Stalder, T., Kirschbaum, C., Alexander, N., Bornstein, S.R., Gao, W., Miller, R., Stark, S., Bosch, J.A., Fischer, J.E., 2013. Cortisol in hair and the metabolic syndrome. J. Clin. Endocrinol. Metab. 98 (6), 2573–2580. Statistics Netherlands 2004a, Standaard Onderwijsindeling 2003. Statistics Netherlands 2004b, Allochtonen in Nederland (in Dutch). Staufenbiel, S.M., Penninx, B.W., Spijker, A.T., Elzinga, B.M., van Rossum, E.F., 2013. Hair cortisol, stress exposure, and mental health in humans: a systematic review. Psychoneuroendocrinology 38 (8), 1220–1235. Sturge-Apple, M.L., Davies, P.T., Cicchetti, D., Manning, L.G., 2012. Interparental violence, maternal emotional unavailability and children’s cortisol functioning in family contexts. Dev. Psychol. 48 (1), 237–249. The Netherlands Institute for Social Research, (2012) Status scores. Vaghri, Z., Guhn, M., Weinberg, J., Grunau, R.E., Yu, W., Hertzman, C., 2013. Hair cortisol reflects socio-economic factors and hair zinc in preschoolers. Psychoneuroendocrinology 38 (3), 331–340. Van Uum, S.H., Sauve, B., Fraser, L.A., Morley-Forster, P., Paul, T.L., Koren, G., 2008. Elevated content of cortisol in hair of patients with severe chronic pain: a novel biomarker for stress. Stress 11 (6), 483–488. Vanaelst, B., Michels, N., De Vriendt, T., Huybrechts, I., Vyncke, K., Sioen, I., Bammann, K., Rivet, N., Raul, J.S., Molnar, D., De Henauw, S., 2013. Cortisone in hair of elementary school girls and its relationship with childhood stress. Eur. J. Pediatr. 172 (6), 843–846. Veldhuis, L., Vogel, I., van Rossem, L., Renders, C.M., Hirasing, R.A., Mackenbach, J.P., Raat, H., 2013. Influence of maternal and child lifestyle-related characteristics on the socioeconomic inequality in overweight and obesity among 5-year-old children; the be active, eat right study. Int. J. Environ. Res. Publ. Health 10 (6), 2336–2347.
Contents lists available at ScienceDirect
Psychoneuroendocrinology journal homepage: www.elsevier.com/locate/psyneuen
Socioeconomic status in children is associated with hair cortisol levels as a biological measure of chronic stress J. Vliegenthart a , G. Noppe a,b , E.F.C. van Rossum b , J.W. Koper b , H. Raat c , E.L.T. van den Akker a,∗ a Department of Pediatrics, Division of Endocrinology, Sophia Children’s Hospital, Erasmus MC, University Medical Center Rotterdam, Rotterdam, The Netherlands b Department of Internal Medicine, Division of Endocrinology, Erasmus MC, University Medical Center Rotterdam, Rotterdam, The Netherlands c Department of Public Health, Erasmus MC, University Medical Center Rotterdam, Rotterdam, The Netherlands
a r t i c l e
i n f o
Article history: Received 30 June 2015 Received in revised form 25 November 2015 Accepted 30 November 2015 Keywords: SES Scalp hair Cortisol Children Stress
a b s t r a c t Introduction: Low socioeconomic status (SES) may be associated with a high risk of lifestyle-related diseases such as cardiovascular diseases. There is a strong association between parental SES, stress and indicators of child health and adult health outcome. The exact mechanisms underlying this association have not yet been fully clarified. Low SES may be associated with chronic stress, which may lead to activation of the hypothalamic-pituitary-adrenal (HPA)-axis, resulting in a higher circulating level of the stress hormone cortisol. Therefore, chronic stress may mediate the association between low SES and elevated cortisol levels and its adverse outcomes. Aim: We investigated whether SES was associated with a chronic measure of cortisol exposure in a child population. Methods: Cortisol and cortisone were measured in scalp hair in 270 children and adolescents, aged 4–18 years, enrolled through school visits. Neighborhood level SES was based on a score developed by the Netherlands Institute for Social Research using postal codes, and this includes neighborhood measures of income education and unemployment. Maternal and paternal education level were used as indicators of family SES. Results: Neighborhood level socioeconomic status score was significantly associated with hair cortisol (ˇ = −0.103, p = 0.007, 95%CI [−0.179, −0.028]) and hair cortisone (ˇ = −0.091, p = 0.023, 95%CI [−0.167, −0.015]), adjusted for age and sex. Additionally, hair cortisol was significantly correlated with maternal education level and hair cortisone was significantly correlated with paternal education level. Conclusion: The results of our study suggest that the widely shown association between low family SES and adverse child health outcomes may be mediated by chronic stress, given the chronically higher levels of cortisol in children and adolescents in families with low SES. It is especially notable that the association between SES and cortisol was already found in children of young age as this can have major consequences, such as increased risk of cardio metabolic diseases in later life. © 2015 Elsevier Ltd. All rights reserved.
1. Introduction There is a strong association between parental socioeconomic status (SES) and indicators of child health (Evans and Kantrowitz,
∗ Corresponding author: E.L.T. van den Akker, Erasmus MC-SP1536, P.O. box 2060, 3000 CB Rotterdam, The Netherlands. E-mail addresses: [email protected] (J. Vliegenthart), [email protected] (G. Noppe), [email protected] (E.F.C. van Rossum), [email protected] (J.W. Koper), [email protected] (H. Raat), [email protected] (E.L.T. van den Akker). http://dx.doi.org/10.1016/j.psyneuen.2015.11.022 0306-4530/© 2015 Elsevier Ltd. All rights reserved.
2002; Evans and Kim, 2010; Clearfield et al., 2014). The association of parental SES is partly driven by environmental exposures, such as nutrition, lifestyle and healthcare. However, these variables do not explain the complete picture as the association exists even in the context of adequate healthcare and nutrition (Adler et al., 1994). Low SES is also associated with increased stress by violence in the neighborhood and at home, disorganization in school environments, environmental toxins, crowding, and noise (Evans and Kantrowitz, 2002; Evans, 2006; Evans and Kim, 2010). Psychological stressors can elicit cortisol activation through the biological stress response, depending on the characteristics of the stressor
10
J. Vliegenthart et al. / Psychoneuroendocrinology 65 (2016) 9–14
(Dickerson and Kemeny, 2004; Felner, 2014). This involves a physiological response to situations in which the individual experiences stress. The stress response leads to activation of the hypothalamicpituitary-adrenal (HPA)-axis. Elevated HPA-axis activity results in increased levels of circulating cortisol. In animal studies, where randomization is feasible, intervention trials show that chronic stress exposure leads to activation of the HPA-axis. For example, cortisol concentrations measured in hair of rhesus macaques were found to be elevated in response to a prolonged environmental stressor (being relocated to a newly constructed building) (Davenport et al., 2006). In humans, where randomization trials are not feasible, association studies show a link between low SES, chronic stress and dysregulation of the HPA-axis (Hajat et al., 2010; Sheridan et al., 2013; Vaghri et al., 2013; Clearfield et al., 2014). Van Uum et al., for example, found that hair cortisol contents were significantly greater in chronic pain patients than in controls and that HCC was also increased in patients with major chronic stress (Van Uum et al., 2008). Several studies found an association between lower SES and higher levels of cortisol, many found no association, with some reporting an opposite relationship (Lupien et al., 2000; Dowd et al., 2009; Sheridan et al., 2013; Clearfield et al., 2014). Lupien et al. measured basal morning salivary cortisol levels in 217 children with different SES levels ranging from 6 to 10 years of age. They found that children with low SES had significantly higher salivary cortisol levels than children with high SES (Lupien et al., 2000). Clearfield et al. compared diurnal salivary cortisol and maternal-infant synchrony in 32 infants, aged 6–12 months and their mothers (with low or high SES). Low-SES infants and mothers exhibited higher average salivary cortisol, compared to high-SES infants (Clearfield et al., 2014). Sheridan et al. (2013) compared maternal self-rated social status and salivary cortisol in 40 children, aged 8–12 years and also found a negative association (Sheridan et al., 2013). However some other studies even found opposite associations e.g. low SES linked to lower cortisol levels (Fuller-Rowell et al., 2012, Sturge-Apple et al., 2012). Thus, both activation and suppression of the HPA-axis have been found in these studies. SES is associated with dysregulation of the HPA-axis, but literature is inconsistent concerning the nature of this association, varying from a positive association to a negative association, a blunted cortisol response or a flattened daily rhythm (Lupien et al., 2000; Dowd et al., 2009; Hajat et al., 2010; Sheridan et al., 2013; Vaghri et al., 2013; Clearfield et al., 2014). These inconsistencies may arise from the way the cortisol is measured. Cortisol concentrations in these studies are mainly measured in saliva, urine, or serum. However, due to the circadian rhythm, pulsatile secretion, and the daily variation in the secretion of cortisol, none of these sampling matrices is an optimal measure of chronic cortisol concentrations. A more suitable matrix appears to be scalp hair, because hair samples can provide a historical timeline of cortisol exposure (Manenschijn et al., 2011a; Manenschijn et al., 2012; Stalder and Kirschbaum, 2012). Scalp hair at the posterior vertex displays regular growth at an average rate of 1 cm/month (Hayashi et al., 1991; Sauve et al., 2007) providing a retrospective calendar of mean cortisol exposure and HPA-axis activity of recent months to years. Hair cortisol analysis has previously been applied to investigate different causes of stress, such as shift work, unemployment and chronic pain (Davenport et al., 2006; Dettenborn et al., 2010; Manenschijn et al., 2011b; Staufenbiel et al., 2013). In addition to cortisol, cortisone might be a useful biomarker for stress research. The biologically inactive hormone cortisone mainly originates from cortisol metabolism by enzymatic activity of 11hydroxysteroid dehydrogenase type 2. Cortisone levels in scalp hair of young children experiencing psychosocial stress were found to be elevated (Vanaelst et al., 2013).
Chronic stress can negatively impact health, resulting in an increased incidence of obesity, type 2 diabetes mellitus, and cardiovascular disease (Everson-Rose and Lewis, 2005; Manenschijn et al., 2013). The risk of obesity in young children from lower SES groups is 4 times higher than that in the higher SES groups (Veldhuis et al., 2013). Low SES, in turn, is associated with a high risk of lifestylerelated diseases such as cardiovascular diseases (Shrewsbury and Wardle, 2008). These cardiovascular diseases are partly lifestylerelated, but could also be caused by other factors such as stress or its biological mediator cortisol (Manenschijn et al., 2013). We hypothesized that children with a lower SES experience more stress, leading to increased activation of the HPA-axis resulting in higher levels of cortisol. We investigated whether SES was associated with chronic measurements of cortisol and cortisone, measured in scalp hair in children and adolescents from the general population of the Netherlands, aged 4–18 years. 2. Methods We included children from the general population of the Netherlands in a cross-sectional observational study. Children aged 4–18 years were included. Data collection was performed between January 2011 and July 2014. Two elementary schools (one in the city of Amersfoort and one in Rotterdam) and two secondary schools in Rotterdam participated in the study. The schools are located in different districts of the city (difference in SES, income, etc.) and are of different educational levels (low, mid and high) to create the best possible reflection of the general population. Parents and children of school classes (1 class per grade) were invited to participate in the study. All parents and children above the age of 12 years provided informed consent, children under the age of 12 years provided informed assent. The participation rate approximated 50%. Exclusion criteria for the children were: suffering from chronic diseases, chronic use of medication, in particular glucocorticoids (both based on a completed questionnaire), or scalp hair length shorter than 2 centimeters. Data were encoded before they were entered in the database. This study was approved by the Medical Ethics Committee of the Erasmus MC (MEC-2012-25). Anthropometric data were collected in all children, including height (cm) and weight (kg). Standing height was measured in centimeters with the precision of 1 mm, with a wall mounted stadiometer. Body weight was measured in kilograms with the precision of 2 decimal places with a calibrated electric-weight model. The standard deviation scores (SDS) of weight, height and body mass index (BMI: body mass divided by the square of the body height) were calculated based on data from the fourth Dutch nationwide growth study 1997, using Growth Analyser (Fredriks et al., 2000). 2.1. Hair cortisol Hair samples of about 100 hairs were taken from the posterior vertex of the scalp, as close to the scalp as possible, using small surgical scissors. The hair was taped to a piece of paper and the scalp end was marked. The samples were stored in an envelope at room temperature until analysis (Manenschijn et al., 2011a). The proximal 3 cm of hair, reflecting the most recent 3 months, was cut in 1 cm segments, weighed on an electronic scale and transferred to disposable glass tubes. Subsequently, the hair samples were washed with isopropanol for 2 min, after which the samples were dried for at least 2 days. Deuterated cortisol was added as an internal standard, and steroids were extracted in 1.5 ml of methanol (MeOH) for 18 h at 25 ◦ C. After extraction, the samples
J. Vliegenthart et al. / Psychoneuroendocrinology 65 (2016) 9–14
are centrifuged for 5 min at 4500 rpm at 4 ◦ C. The supernatant is transferred to glass tubes, and the extract was dried at 37 ◦ C under a stream of nitrogen. The extract is then reconstituted in 2% MeOH in water and further cleaned using “solid phase extraction” on a 96-well plate SPE (Oasis HLB 96-well plate, 30 mg, Waters Corporation, Milford, USA), evaporated under a stream of nitrogen, and reconstituted in 30% MeOH in water for analysis. The cortisol was then analyzed on a Xevo TQS Liquid Chromatography-Tandem Mass Spectrometer (Waters Corporation, Milford, USA) (Noppe et al., 2015). The measured concentration was then divided by the hair weight in order to calculate the concentration per mg hair. The outcome is called hair cortisol concentration (HCC) and is in picograms per milligram. The analysis of hair cortisone is identical to hair cortisol (Noppe et al., 2015). Intra-individual coefficients of variation were below 15% for cortisol and cortisone (Noppe et al., 2015). 2.2. Socioeconomic status As an indicator of socioeconomic status we used Dutch neighborhood level status scores, which are scores calculated by the Netherlands Institute for Social Research (Sociaal en Cultureel Planbureau; SCP) that specify the relative social status of a neighborhood, compared to other neighborhoods in the Netherlands. The higher the score, the higher the status of a neighborhood. A low score indicates a low neighborhood status. The status score in 2010 reaches from −7.25 to 3.19. Within our study population, there are 91 different zip codes for all participants and the mean number of residents per zip code district is 8922 and varies between 835 and 20,285 people. The social status of a neighborhood is derived from the following characteristics of the people who live there: education, income and labor market status. Neighborhoods in this case are 4-digit postal code areas (only the numbers, not letters). The status scores are composed of: the average income in a neighborhood, the percentage of people with a low income, the percentage of low-educated people and the percentage of people not working. In a period of four years, data is collected per six-digit (the four numbers plus two letters) postal code area. Data are aggregated to the level of the four-digit postcode area. This means that the status scores of 2010 contain data for the period 2006–2010. Through factor analysis, these characteristics are summarized in a composite feature: the social status (The Netherlands Institute for Social Research, 2012). In addition, at the level of the family, we used both maternal and paternal reported highest education level as indicator of family SES. At the elementary schools the parents completed a questionnaire in which they filled in their educational level. At the secondary schools the children completed the questionnaire, but when there was indistinctness about parental educational level or other unanswered questions, the parents were contacted to clarify this. Education level was classified according to the Dutch education system: no education, primary education (elementary school), secondary education (secondary school), vocational studies, community college and university. Of these categories we made three education levels: low (no and primary education), mid (secondary education and vocational studies) and high (community college and university) (Statistics Netherlands 2004a). 2.3. Ethnicity Ethnicity was determined by country of birth of the children and their parents, derived from the completed questionnaires. Ethnicity is divided into three groups: native Dutch, western background and non-western background. Ethnicities are determined by a report of Statistics Netherlands. When a child, or one or both parents were not born in the Netherlands, he is seen as an immigrant. A western immigrant comes from one of the countries in Europe (excluding
11
Turkey), North America and Oceania, or Indonesia or Japan. A child whose ethnic background is one of the countries in the continents of Africa, Latin America and Asia (excluding Indonesia and Japan) or Turkey is a non-western immigrant (Statistics Netherlands, 2004b). 2.4. Statistics Statistical analysis was performed using IBM SPSS Statistics 20 (IBM Corp. Armonk, New York, NY, USA). The Gaussian distribution of variables was assessed by the Kolmogorov–Smirnov test. We regarded a variable to deviate from the normal distribution if the p-value was smaller than 0.05. Non-normally distributed variables were log-transformed to achieve a normal distribution. Because the SES-score was not normally distributed despite log-transformation, we calculated the Spearman’s correlation coefficient between the SES-score and HCC. Using a hierarchical linear regression approach, we assessed the association between SES-score, adjusted for sex and age, with HCC as the primary model, adding ethnicity in the second block. The random effects in the hierarchical model were included as random intercept effects. The  and 95% CI of SES-score on HCC before and after adjustment for ethnicity was reported. 3. Results In total, 296 children participated in the study, 17 children were excluded because of missing analysis of hair cortisol and 9 children were excluded because of a missing SES score. Therefore, the total number of children included in our study is 270. Of these, 131 are in elementary school and 139 children are in secondary school. Baseline characteristics of the total group, and of two subgroups based on school type, are shown in Table 1. Of all participants, 7% was overweight (BMI–SDS > 1.1) and 2% was obese (BMI–SDS > 2.3) (Fredriks et al., 2000). There is a significant negative correlation between status score and hair cortisol concentration (r = −0.193; p = 0.001), as shown in Fig. 1. With the Spearman’s correlation test we observed significant associations between maternal education and hair cortisol, and between paternal education and hair cortisol and cortisone (Table 2). No associations between cortisol and maternal or paternal education were found with the linear regression model. Combination of maternal and paternal education into one variable gave the same result. With linear regression analysis, adjusted for sex and age, we found a significant association in the total group of both hair cortisol (ˇ = −0.103, p = 0.007, 95%CI [−0.179, −0.028]) and hair cortisone (ˇ = −0.091, p = 0.023, 95%CI [−0.167, −0.015]) with the SES score (Table 2). When ethnicity was added to the linear regression model of hair cortisol, the ˇ changed from −0.103 to −0.064 (95%CI [−0.155, 0.027]). In the linear regression model of hair cortisone, the ˇ changed from −0.091 to −0.049 (95%CI [−0.142,−0.044]) (Table 2). Adjustments of the association for BMI–SDS did not change these results. 4. Discussion This study shows a negative association between indicators of family socioeconomic status and hair cortisol concentration as well as a negative association between socioeconomic status and hair cortisone concentration in children and adolescents. This association was independent of BMI, but the strength of the association was influenced by ethnicity. To our knowledge, this is the first study reporting the association between SES and chronic cortisol exposure measured in scalp hair in children aged 4–18 years. Only one previous study, by Vaghri et al., studied hair cortisol and SES in
12
J. Vliegenthart et al. / Psychoneuroendocrinology 65 (2016) 9–14
Table 1 Baseline characteristics of the whole group of children from the general population that participated in this study and divided by participant educational level.
Total Sex Age BMI–SDS Ethnicity
SES score Maternal education
Paternal education
Cortisol (pg/mg) Cortisone (pg/mg)
N= Male (%) Mean (range) Mean (SD) Native Dutch (N%) Western background (N%) Non-western background (N%) Missing (N%) Mean (SD) N (%) Low Mid N (%) N (%) High N (%) Low N (%) Mid High N (%) Mean (range) Mean (range)
Elementary school
Secondary school
Total
131 70 (53%) 7.97 (3.66–12.58) −0.30 (1.01) 99 (75.6%) 15 (11.5%) 3 (2.3%) 14 (10.6%) 0.85 (0.92) 0 17 (18%) 75 (82%) 0 14 (15%) 79 (85%) 3.05 (0.28–38.26) 13.55 (1.99–159.89)
139 70 (50%) 15.17 (11.89–18.78) 0.15 (.98) 87 (62.6%) 12 (8.6%) 39 (28.1%) 1 (0.7%) −0.14 (1.86) 5 (4%) 35 (29%) 83 (67%) 3 (3%) 31 (26%) 86 (71%) 3.67 (0.40–34.32) 16.23 (3.93–50.30)
270 140 (52%) 11.67 (3.66–18.78) −0.07 (1.01) 186 (68.9%) 27 (10.0%) 42 (15.6%) 15 (5.5%) 0.34 (1.86) 5 (2%) 52 (24%) 158 (74%) 3 (1%) 45 (21%) 165 (78%) 3.37 (0.28–38.26) 14.94 (1.99–159.89)
SDS: standard deviation score, SES: socioeconomic status, pg/mg: picograms per milligram. Cortisol and cortisone are measured in hair.
Fig. 1. Correlation between SES and HCC.
HCC is in picograms per milligram (pg/mg) and is shown on a logarithmic scale.
Table 2 Associations between measures of SES and hair cortisol or hair cortisone.
Status score Corrected for age & sex Corrected for age & sex & BMI Corrected for age & sex & ethnicity Status score Paternal education Maternal education
Cortisol
Cortisone
ˇ = −0.103** [−0.179,−0.028] ˇ = −0.094* [−0.170,−0.017] ˇ = −0.064[−0.155,0.027] = −0.193** = −0.135* = −0.164*
ˇ = −0.091* [−0.167,−0.015] ˇ = −0.081* [−0.158,−0.004]  = -0.049[−0.142,0.044] = −0.159** = −0.199* = −0.120
Linear regression analyses coefficients, ˇ. Spearman’s correlation coefficient .
J. Vliegenthart et al. / Psychoneuroendocrinology 65 (2016) 9–14
children, with mean age 4.6 years. A negative association between parental education and hair cortisol was found (Vaghri et al., 2013). Previous studies have analyzed the association between SES and cortisol, but did not use a chronic measure of cortisol exposure, such as scalp hair analysis. Authors generally reported salivary cortisol concentrations (Lupien et al., 2000; Sheridan et al., 2013; Clearfield et al., 2014). Studies vary widely in sample size, selection criteria and number of measures collected. Previous study outcomes are not consistent with regard to the nature of the relation between SES and cortisol. Although several studies found an association between lower SES and higher levels of cortisol, many found no association, with some reporting an opposite relationship (Lupien et al., 2000; Dowd et al., 2009; Sheridan et al., 2013; Clearfield et al., 2014). We found ethnicity to be an important confounder and we cannot exclude a role of ethnicity as a driving force. Relatively few studies have investigated SES in combination with differences by ethnicity and cortisol. In adults, Stalder and Kirschbaum (2012), found an association between salivary cortisol and ethnicity and SES in a multi-ethnic population study (Hajat et al., 2010). In African American and Hispanic adolescents, significantly flatter diurnal cortisol slopes (due to higher evening cortisol) were found, which could have negative health consequences (DeSantis et al., 2007). In this study, examined socio environmental factors (including parental SES) failed to explain the observed racial/ethnic differences in diurnal cortisol rhythms (DeSantis et al., 2007). HCC in volunteers from settlement communities in Kenya were significantly higher in females, divorced volunteers and those whose income was below minimum wage. However, no evidence was found for increased chronic stress (measured by hair cortisol content) between members of slum settlements that were adjacent and those that were distant to large floriculture farms (Henley et al., 2014). However, we cannot directly compare these results with our study because of the different SES measures that were used. In our study there was a change of  with a loss of significance in the linear regression analysis between HCC and the SES-score after adjustment for ethnicity. Ethnicity is so highly linked to SES that to address this unequal distribution, studies with much higher numbers are needed to unravel the effects of both SES and ethnicity independently. Ethnicity as defined in the current study may be a misnomer, as it is not a comparison of multiple ethnicities, but rather the comparison of offspring of native Dutch parents versus immigrants, which usually represent a subgroup of an ethnical population. In the Netherlands, these ethnical minorities tend to be lower educated, have less income and have more health problems. The underlying causes for the association between SES and hair cortisol is unclear, it seems likely to hypothesize that low SES leads to psychological stress and thereby to elevated hair cortisol. The generalizability of these results remains to be investigated in future studies of larger populations with larger numbers of specific ethnicities, which offer the possibility of elucidating the most important confounding factors. The negative association between SES and cortisol, observed in the present study, is in line with most previous studies measuring salivary cortisol associations with SES in children. Markedly higher cortisol levels have been found in low SES adolescents and children compared to middle and high SES adolescents and children from diverse age categories: e.g. aged 9–18 years (Chen et al., 2010), aged 6–9 years (Lupien et al., 2000; Evans and English, 2002) and in infants as young as 6–12 months of age (Clearfield et al., 2014). However some other studies did not find an association between SES and salivary cortisol levels in 8–11 year old children (Sheridan et al., 2013), or even found opposite associations e.g. low SES linked to lower cortisol levels (Fuller-Rowell et al., 2012; Sturge-Apple., 2012).
13
SES can be scored in different ways. In most studies, SES was scored using parental education level or income. In our study we used three indicators of family SES, i.e. maternal and paternal educational level and a neighborhood level SES score. When we analyzed educational level separately, we found significant negative correlations between maternal education level and hair cortisol concentration, and also between paternal education level and both hair cortisol and cortisone concentrations. The negative association that was found implies that children with low educated parents are likely to have higher stress hormone levels. A limitation of cross-sectional observational studies is to determine cause and consequence of the associations found. Although the underlying cause for the association between SES and hair cortisol in our study is unclear, it seems likely that chronic stress may mediate the association between low SES and elevated cortisol levels. Also, the generalizability of these results remains to be investigated in future studies of larger populations, which offer the possibility of pinning down the most important confounding factors. On the other hand, the magnitude of the effect in our study is relatively large, with an effect size ˇ of −0,103 for the association between status score and (log-transformed) hair cortisol, because both parameters can be influenced by multiple other factors such as non-registered major life events, individual, social or family factors. SES scores can be measured in many different ways, each with its own advantage and disadvantages. We derived SES from postal codes, representing mean SES values from a street area and not from an individual. As a consequence, SES distribution is wider in secondary school children compared to elementary school children. The advantage of neighborhood SES analysis is that multiple SESrelated factors were included in the calculation of this score, which is a valid method to compare SES (Knol, 2012). A strength of our study is that we used a new, highly sensitive and reliable method to measure chronic cortisol exposure in scalp hair. Because of this method we have a better understanding of cortisol concentrations over a prolonged period of time and not just one moment in time (e.g. serum or saliva). We measured the hair cortisol and hair cortisone concentrations with liquid chromatography tandem mass spectrometry (LC–MS/MS), a very sensitive method to accurately measure steroids, which is advocated to be the golden standard to measure steroids (Handelsman and Wartofsky 2013; Noppe et al., 2015). The above-mentioned LC–MS/MS has the advantage that besides cortisol, we could also measure cortisone, which could be of added value for reliable measurement of stress (Vanaelst et al., 2013). Both cortisol and cortisone showed the same relationship with SES. Low SES is thought to lead to negative health consequences. Multiple studies show that chronic elevated stress has negative health outcomes (Everson-Rose and Lewis 2005; Shrewsbury and Wardle 2008). Chronic elevated cortisol levels were shown to be associated with increased risk of the metabolic syndrome, increased visceral fat mass, hypertension, insulin resistance, and dyslipidemia, which all result in an increased cardiovascular risk (Manenschijn et al., 2013; Stalder et al., 2013). The results of our study suggest that the widely shown association between low family socioeconomic status and adverse child health outcomes may be mediated by chronic stress, given the chronically higher levels of cortisol in children and adolescents in families with low socioeconomic status. The measurement of scalp hair cortisol seems to provide a unique tool to study chronic stress. It is especially notable that the association between SES and cortisol was already found in children of young age as this can have important consequences later in life, and poses an increased risk of cardio metabolic diseases in adulthood. The next step for future studies is to elucidate the role of cortisol in more specific pathways
14
J. Vliegenthart et al. / Psychoneuroendocrinology 65 (2016) 9–14
of adverse socioeconomic circumstances and health consequences in childhood. Conflict of interest The authors have no conflict of interests or financial statements to disclose. Acknowledgements The authors are grateful to Ms. Margriet E. Bisschoff, research nurse, Erasmus MC, Rotterdam, The Netherlands, for her help in participant recruitment and measurements. We also thank the primary and secondary schools, the children and their parents for their participation. We are also grateful to the Thrasher research Fund for financial support (grant TRF-11643). References Adler, N.E., Boyce, T., Chesney, M.A., Cohen, S., Folkman, S., Kahn, R.L., Syme, S.L., 1994. Socioeconomic status and health. The challenge of the gradient. Am. Psychol. 49 (1), 15–24. Chen, E., Cohen, S., Miller, G.E., 2010. How low socioeconomic status affects 2-year hormonal trajectories in children. Psychol. Sci. 21 (1), 31–37. Clearfield, M.W., Carter-Rodriguez, A., Merali, A.R., Shober, R., 2014. The effects of SES on infant and maternal diurnal salivary cortisol output. Infant Behav. Dev. 37 (3), 298–304. Davenport, M.D., Tiefenbacher, S., Lutz, C.K., Novak, M.A., Meyer, J.S., 2006. Analysis of endogenous cortisol concentrations in the hair of rhesus macaques. Gen. Comp. Endocrinol. 147 (3), 255–261. DeSantis, A.S., Adam, E.K., Doane, L.D., Mineka, S., Zinbarg, R.E., Craske, M.G., 2007. Racial/ethnic differences in cortisol diurnal rhythms in a community sample of adolescents. J. Adolesc. Health 41 (1), 3–13. Dettenborn, L., Tietze, A., Bruckner, F., Kirschbaum, C., 2010. Higher cortisol content in hair among long-term unemployed individuals compared to controls. Psychoneuroendocrinology 35 (9), 1404–1409. Dickerson, S.S., Kemeny, M.E., 2004. Acute stressors and cortisol responses: a theoretical integration and synthesis of laboratory research. Psychol. Bull. 130 (3), 355–391. Dowd, J.B., Simanek, A.M., Aiello, A.E., 2009. Socio-economic status, cortisol and allostatic load: a review of the literature. Int. J. Epidemiol. 38 (5), 1297–1309. Evans, G.W., English, K., 2002. The environment of poverty: multiple stressor exposure, psychophysiological stress, and socioemotional adjustment. Child Dev. 73 (4), 1238–1248. Evans, G.W., Kantrowitz, E., 2002. Socioeconomic status and health: the potential role of environmental risk exposure. Annu. Rev. Publ. Health 23, 303–331. Evans, G.W., Kim, P., 2010. Multiple risk exposure as a potential explanatory mechanism for the socioeconomic status-health gradient. Ann. N. Y. Acad. Sci. 1186, 174–189. Evans, G.W., 2006. Child development and the physical environment. Annu. Rev. Psychol. 57, 423–451. Everson-Rose, S.A., Lewis, T.T., 2005. Psychosocial factors and cardiovascular diseases. Annu. Rev. Publ. Health 26, 469–500. Felner, U., 2014. Adrenal gland disorders. In: Endocrine Pathophysiology. Lippincott Williams & Wilkins, pp. 207. Fredriks, A.M., van Buuren, S., Burgmeijer, R.J., Meulmeester, J.F., Beuker, R.J., Brugman, E., Roede, M.J., Verloove-Vanhorick, S.P., Wit, J.M., 2000. Continuing positive secular growth change in The Netherlands 1995–1997. Pediatr. Res. 47 (3), 316–323. Fuller-Rowell, T.E., Doan, S.N., Eccles, J.S., 2012. Differential effects of perceived discrimination on the diurnal cortisol rhythm of African Americans and Whites. Psychoneuroendocrinology 37 (1), 107–118. Hajat, A., Diez-Roux, A., Franklin, T.G., Seeman, T., Shrager, S., Ranjit, N., Castro, C., Watson, K., Sanchez, B., Kirschbaum, C., 2010. Socioeconomic and race/ethnic differences in daily salivary cortisol profiles: the multi-ethnic study of atherosclerosis. Psychoneuroendocrinology 35 (6), 932–943.
Handelsman, D.J., Wartofsky, L., 2013. Requirement for mass spectrometry sex steroid assays in the Journal of Clinical Endocrinology and Metabolism. J. Clin. Endocrinol. Metab. 98 (10), 3971–3973. Hayashi, S., Miyamoto, I., Takeda, K., 1991. Measurement of human hair growth by optical microscopy and image analysis. Br. J. Dermatol. 125 (2), 123–129. Henley, P., Lowthers, M., Koren, G., Fedha, P.T., Russell, E., VanUum, S., Arya, S., Darnell, R., Creed, I.F., Trick, C.G., Bend, J.R., 2014. Cultural and socio-economic conditions as factors contributing to chronic stress in sub-Saharan African communities. Can. J. Physiol. Pharmacol. 92 (9), 725–732. Knol, F., 2012. Statusontwikkeling van wijken in Nederland 1998–2010 (Dutch). Institute for Social Research, The Netherlands. Lupien, S.J., King, S., Meaney, M.J., McEwen, B.S., 2000. Child’s stress hormone levels correlate with mother’s socioeconomic status and depressive state. Biol. Psychiatry 48 (10), 976–980. Manenschijn, L., Koper, J.W., Lamberts, S.W., van Rossum, E.F., 2011a. Evaluation of a method to measure long term cortisol levels. Steroids 76 (10–11), 1032–1036. Manenschijn, L., van Kruysbergen, R.G., de Jong, F.H., Koper, J.W., van Rossum, E.F., 2011b. Shift work at young age is associated with elevated long-term cortisol levels and body mass index. J. Clin. Endocrinol. Metab. 96 (11), E1862–E1865. Manenschijn, L., Koper, J.W., van den Akker, E.L., de Heide, L.J., Geerdink, E.A., de Jong, F.H., Feelders, R.A., van Rossum, E.F., 2012. A novel tool in the diagnosis and follow-up of (cyclic) Cushing’s syndrome: measurement of long-term cortisol in scalp hair. J. Clin. Endocrinol. Metab. 97 (10), E1836–E1843. Manenschijn, L., Schaap, L., van Schoor, N.M., van der Pas, S., Peeters, G.M., Lips, P., Koper, J.W., van Rossum, E.F., 2013. High long-term cortisol levels, measured in scalp hair, are associated with a history of cardiovascular disease. J. Clin. Endocrinol. Metab. 98 (5), 2078–2083. Noppe, G., de Rijke, Y.B., Dorst, K., van den Akker, E.L., van Rossum, E.F., 2015. LC–MS/MS based method for long-term steroid profiling in human scalp hair. Clin. Endocrinol. 83 (Oxf). Sauve, B., Koren, G., Walsh, G., Tokmakejian, S., Van Uum, S.H., 2007. Measurement of cortisol in human hair as a biomarker of systemic exposure. Clin. Invest. Med. 30 (5), E183–191. Sheridan, M.A., How, J., Araujo, M., Schamberg, M.A., Nelson, C.A., 2013. What are the links between maternal social status, hippocampal function, and HPA axis function in children? Dev. Sci. 16 (5), 665–675. Shrewsbury, V., Wardle, J., 2008. Socioeconomic status and adiposity in childhood: a systematic review of cross-sectional studies 1990–2005. Obesity (Silver Spring) 16 (2), 275–284. Stalder, T., Kirschbaum, C., 2012. Analysis of cortisol in hair-state of the art and future directions. Brain Behav. Immun. 26 (7), 1019–1029. Stalder, T., Kirschbaum, C., Alexander, N., Bornstein, S.R., Gao, W., Miller, R., Stark, S., Bosch, J.A., Fischer, J.E., 2013. Cortisol in hair and the metabolic syndrome. J. Clin. Endocrinol. Metab. 98 (6), 2573–2580. Statistics Netherlands 2004a, Standaard Onderwijsindeling 2003. Statistics Netherlands 2004b, Allochtonen in Nederland (in Dutch). Staufenbiel, S.M., Penninx, B.W., Spijker, A.T., Elzinga, B.M., van Rossum, E.F., 2013. Hair cortisol, stress exposure, and mental health in humans: a systematic review. Psychoneuroendocrinology 38 (8), 1220–1235. Sturge-Apple, M.L., Davies, P.T., Cicchetti, D., Manning, L.G., 2012. Interparental violence, maternal emotional unavailability and children’s cortisol functioning in family contexts. Dev. Psychol. 48 (1), 237–249. The Netherlands Institute for Social Research, (2012) Status scores. Vaghri, Z., Guhn, M., Weinberg, J., Grunau, R.E., Yu, W., Hertzman, C., 2013. Hair cortisol reflects socio-economic factors and hair zinc in preschoolers. Psychoneuroendocrinology 38 (3), 331–340. Van Uum, S.H., Sauve, B., Fraser, L.A., Morley-Forster, P., Paul, T.L., Koren, G., 2008. Elevated content of cortisol in hair of patients with severe chronic pain: a novel biomarker for stress. Stress 11 (6), 483–488. Vanaelst, B., Michels, N., De Vriendt, T., Huybrechts, I., Vyncke, K., Sioen, I., Bammann, K., Rivet, N., Raul, J.S., Molnar, D., De Henauw, S., 2013. Cortisone in hair of elementary school girls and its relationship with childhood stress. Eur. J. Pediatr. 172 (6), 843–846. Veldhuis, L., Vogel, I., van Rossem, L., Renders, C.M., Hirasing, R.A., Mackenbach, J.P., Raat, H., 2013. Influence of maternal and child lifestyle-related characteristics on the socioeconomic inequality in overweight and obesity among 5-year-old children; the be active, eat right study. Int. J. Environ. Res. Publ. Health 10 (6), 2336–2347.