Melia et al. Biology of Sex Differences (2019) 10:55 https://doi.org/10.1186/s13293-019-0273-3 RESEARCH Open Access Sex-specific association between the cortisol awakening response and obsessive- compulsive symptoms in healthy individuals Cristian Sebastian Melia1,3, Virginia Soria1,2,3* , Neus Salvat-Pujol1, Ángel Cabezas2,4, Roser Nadal2,5, Mikel Urretavizcaya1,2,3, Alfonso Gutiérrez-Zotes2,4, José Antonio Monreal2,6, José Manuel Crespo1,2,3, Pino Alonso1,2,3, Elisabet Vilella2,4, Diego Palao2,6, José Manuel Menchón1,2,3 and Javier Labad2,6 Abstract Background: Previous studies have shown associations between obsessive-compulsive disorder (OCD) and hypothalamic- pituitary-adrenal axis activity (HPA). We aimed to investigate the association between obsessive-compulsive (OC) symptoms and HPA axis functionality in a non-clinical sample and to explore whether there are sex differences in this relationship. Methods: One hundred eighty-three healthy individuals without any psychiatric diagnosis (80 men, 103 women; mean age 41.3 ± 17.9 years) were recruited from the general population. The Obsessive-Compulsive Inventory Revised (OCI-R) was used to assess OC symptoms. State-trait anxiety, perceived stress, and stressful life events were also assessed. Saliva cortisol levels were determined at 6 time points (awakening, 30 and 60 min post-awakening, 10:00 a.m., 23:00 p.m. and 10:00 a.m. the following day of 0.25 mg dexamethasone intake [that occurred at 23:00 p.m.]). Three HPA axis measures were calculated: cortisol awakening response (CAR), cortisol diurnal slope, and cortisol suppression ratio after dexamethasone (DSTR). Multiple linear regression analyses were used to explore the association between OC symptoms and HPA axis measures while adjusting for covariates. Our main analyses were focused on OCI-R total score, but we also explored associations with specific OC symptom dimensions. Results: No significant differences were observed between males and females in OC symptoms, anxiety measures, stress, or cortisol measures. In the multiple linear regression analyses between overall OC symptoms and HPA axis measures, a female sex by OC symptoms significant interaction (standardized beta = − 0.322; p = 0.023) for the CAR (but not cortisol diurnal slope nor DSTR) was found. Regarding specific symptom dimensions, two other sex interactions were found: a blunted CAR was associated with obsessing symptoms in women, whereas a more flattened diurnal cortisol slope was associated with ordering symptoms in men. Conclusions: There are sex differences in the association between OC symptoms and HPA axis measures in healthy individuals. Keywords: Cortisol, Stress, Obsessive, OCD, Sex differences * Correspondence: vsoria@bellvitgehospital.cat 1 Department of Psychiatry, Bellvitge University Hospital, Bellvitge Biomedical Research Institute (IDIBELL), Neurosciences Group—Psychiatry and Mental Health, Barcelona, Spain 2 Centro de Investigación Biomédica en Red de Salud Mental (CIBERSAM), Carlos III Health Institute, Madrid, Spain Full list of author information is available at the end of the article © The Author(s). 2019 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Melia et al. Biology of Sex Differences (2019) 10:55 Page 2 of 11 Introduction an increased CAR and reduced cortisol suppression after Obsessive-compulsive disorder (OCD) is a common dexamethasone administration. Saliva sampling also al- chronic long-lasting psychiatric disorder, with a preva- lows the study of the cortisol response to psychosocial lence of 2.3% that interferes with all aspects of life of the stress. In this line, previous studies exploring the cortisol individual [1]. OCD is characterized by the presence of response to the Trier Social Stress Test, a social stress obsessions, repetitive disturbing and uncontrollable procedure, in patients with OC personality disorder have thoughts that persist despite the patient’s efforts to sup- revealed an attenuated response in male patients com- press or ignore them, and compulsions and repetitive and pared to male controls, whereas no significant differ- ritualized behaviors or mental acts aimed at neutralizing ences were found in women [14]. obsession-induced anxiety [2]. Obsessive-compulsive There is a wealth of data describing sex differences in (OC) symptoms are not exclusive of OCD. They are HPA axis, documented as early as the neonatal period prevalent in non-clinical samples, with clinically relevant and at all individual levels of the HPA axis [15]. Evi- symptoms in about 21.7% of the general population [3]. dence on sex differences in the HPA axis response to The vulnerability to compulsive activity can be predicted stress is conflicting, with some studies showing no dif- by a spectrum of neuropsychological mechanisms such as ferences. But the overall picture seems to indicate that impaired motor inhibition, cognitive inflexibility and an the response to psychological stress in adult men is with imbalance in goal-directed vs habit learning [4]. A series greater increases in cortisol in comparison to women of cortico-striato-thalamo-cortical circuits that are associ- [5]. However, the study of sex differences in the HPA ated with these cognitive changes are believed to underpin axis response is more complex if we consider potential the expression of compulsive behaviors [4]. moderating variables. For instance, the type of the stres- The hypothalamic-pituitary-adrenal (HPA) axis, the sor seems to influence the HPA axis response: men show major stress response system of the body, is known to greater cortisol responses to a mathematical and verbal be involved in the susceptibility to develop psychiatric challenge whereas women show greater cortisol re- disorders and physical conditions such as infectious dis- sponses to a social rejection challenge [16]. Age and gen- eases, cardiovascular problems, autoimmune processes, der are also moderators of the HPA response to chronic fatigue syndrome, and rheumatoid arthritis [5]. psychosocial stress. In a study exploring ACTH and cor- Stressful life events and dysregulations of the HPA axis tisol responses to the TSST in healthy elderly adults, are thought to play a role in the pathogenesis of the younger adults, and children, there was no age effect in OCD [6, 7] and other anxiety disorders [8]. The first the subgroup of women, whereas younger men had studies to explore the relationship between HPA axis higher ACTH responses compared to older men [17]. In functionality and OCD measured HPA hormones in relation to free salivary cortisol, an enhanced cortisol re- cerebrospinal fluid (CSF) and blood. For instance, ele- sponse in elderly men compared to elderly women was vated values of corticotropin releasing hormone (CRH) observed, while no gender differences emerged in nei- in CSF [9], increased nocturnal secretion of adrenocorti- ther young adults nor children [17]. cotropic hormone (ACTH) [10], and higher basal serum The study of potential sex differences regarding HPA cortisol values [11, 12] have been described in OCD axis activity in the field of OCD has been an underre- patients. searched area. This approach is important because sex In the last two decades, the study of HPA axis func- plays a role in the clinical expression of the illness: tionality using saliva samples has attracted great interest women show, when compared to men, more contamin- [13]. Saliva collection allows cortisol determination at ation/cleaning symptoms [18–21], more aggressive ob- different moments of the day with a non-invasive pro- sessions and hoarding symptoms [20], more somatic cedure. Therefore, saliva sampling can be used for study- obsessions [19], less sexual/religious obsessions [18, 22], ing several dynamic tests of the HPA axis, such as the and less checking and repeating compulsions [22]. Con- cortisol awakening response (CAR), the diurnal cortisol versely, other studies have failed to find sex differences rhythm, or the negative feedback of the HPA axis with in OCD symptom dimensions [23]. However, in that the study of the cortisol suppression by dexamethasone. study correlations between the distinct types of OCD We have used this approach to study the role of these symptom dimensions were stronger in men compared to HPA axis measures in patients with OCD, major depres- women. It is plausible that sex differences in HPA axis sive disorder (MDD) and healthy individuals [6]. A more activity could contribute to the different clinical expres- flattened diurnal cortisol slope calculated between 10 sion of obsessive-compulsive symptoms between men and 23 h was observed for OCD patients with comorbid and women. However, previous studies studying HPA MDD. We also found that trait anxiety was a moderator axis measures in OCD [6, 9–12] did not specifically ex- of the relationship between OCD and HPA axis mea- plore whether there were sex differences on these sures, as OCD patients with greater trait anxiety showed relationships. Melia et al. Biology of Sex Differences (2019) 10:55 Page 3 of 11 The main aim of our study was to explore the role of Several psychometric scales were administered to as- the HPA axis in the clinical expression of obsessive- sess anxiety and stress measures of all participants. compulsive (OC) symptoms in a non-clinical sample. To Current and trait anxiety were measured with the State- our knowledge, this issue has not been studied up to Trait Anxiety Inventory [29]. Stressful life events that date. As there are sex differences in the presentation of occurred in the 6 previous months were assessed with OC symptoms and HPA axis activity, we also aimed to the Holmes-Rahe Social Readjustament Scale [30], that explore whether there are sex differences in this has been also used in Spanish populations [31]. Per- relationship. ceived stress during the previous month was assessed with the 14-item Perceived Stress Scale (PSS) [32]. Methods Study sample Collection of saliva samples One hundred eighty-three healthy individuals (80 men, Saliva samples were obtained using Salivette (Sarstedt 103 women; mean age 41.3 ± 17.9 years) were recruited AG & Co., Nümbrecht, Germany) containers. Partici- from the general population by advertisements in the pants were instructed to collect repeated saliva samples community. Recruitment was conducted in two prov- at home during a regular day and to avoid intense phys- inces (Barcelona, Tarragona) from Catalonia (Spain) by ical activity and stressful situations. Fifteen minutes be- clinical researchers from the Department of Psychiatry fore sample collection, participants were told not to eat, at Bellvitge University Hospital (Hospitalet de Llobregat, drink, smoke, or brush their teeth. Barcelona) and from the Hospital Universitari Institut Saliva samples were obtained at awakening (T1), 30 Pere Mata (Reus, Tarragona). All participants had no (T2) and 60 (T3) min after awakening, at 10 a.m. (T4) past or current history of psychiatric disorders (includ- and at 11 p.m. (T5) on the same day. Participants were ing OCD) and a score below 7 on the 28-item Spanish told to take a very low dose of dexamethasone (0.25 mg) adaptation of the Goldberg General Health Question- at 11 p.m. just after T5 sampling. Another saliva sample naire (GHQ-28) [24]. Exclusion criteria were age less was obtained the day after at 10 a.m. (T6). than 18 years, a diagnosis of a psychiatric disorder in- The samples were kept refrigerated and returned per- cluding substance abuse or dependence (except nico- sonally by each participant. After receipt, the Salivettes tine), mental retardation, neurological disorders, severe were stored at − 20 °C and sent to the BioBank from the medical conditions, pregnancy, or puerperium and cor- Institut de Investigació Sanitaria Pere Virgili (IISPV) for ticosteroid treatment in the previous 3 months. Two centrifugation (3000 rpm for 5 min) and aliquotation, women were receiving contraceptive pills. The research then frozen at − 20 °C until analysis. protocol was approved by the Ethics Committees of Bell- vitge University Hospital and Hospital Universitari Sant Cortisol measurements Joan, and all participants provided written informed con- An enzyme-linked immunosorbent assay (ELISA) was sent after having received a full explanation of the study. performed to determine the cortisol levels in the saliva samples (IBL International, Hamburg, Germany). The Clinical assessment intra-assay and inter-assay coefficients of variation were From a semi-structured interview, a sort of sociodemo- under 8%. The sensitivity of the assay was 0.08 nmol/L. graphic and clinical variables and substance use were To assess the HPA axis function, three dynamic tests assessed. In order to calculate the body mass index, were used: the Cortisol Awakening Response (CAR), the weight and height were measured in all participants, slope between morning and evening cortisol, and the using the formula (kg)/height(m2). Dexamethasone Suppression Test Ratio (DSTR) with The assessment of the OC symptoms was performed very low doses of dexamethasone (0.25 mg). using the Obsessive-Compulsive Inventory Revised The CAR is a physiological process, consisting in the (OCI-R). The OCI-R consists of a self-reported test of increase of cortisol levels as a response to awakening in 18 items divided on six subscales: washing, checking, or- the morning [33]. It combines features of a reactivity dering, obsessing, hoarding, and neutralizing. Each item index (response to awakening) with aspects tied to circa- is scored on a 5-point scale (0–4 points), and the total dian regulation and it has been related to a wide range score is the sum of the scores on all items. It also pro- of psychosocial, physical, and mental health parameters vides scores for the six subscales [25]. The OCI-R has [13]. The CAR was calculated, as suggested by Pruessner proved its validity in both clinical [25, 26] and non- and colleagues, using the area under the curve with re- clinical samples [27] and has been validated to Spanish spect to the increase [34]. The calculation included the [28]. A cut-off score of 21 has been recommended, with sampling points of T1-T3 samples. scores at or above this level indicating the likely pres- The diurnal cortisol slope is defined as the rate of de- ence of OCD [25]. cline in cortisol levels during the day, from morning to Melia et al. Biology of Sex Differences (2019) 10:55 Page 4 of 11 evening. It was calculated using the T4 and the T5 sam- we conducted multiple regression analyses, considering ples. Previous studies have calculated the diurnal cortisol the HPA axis measures as the dependent variable in each slope using both awakening to bedtime or fixed time case. The following independent variables were entered in points (morning to evening), and the implications of this each model with the enter procedure: female gender, age, choice has received little attention [35]. In a previous BMI, smoking, OCI-R scores, Holmes-Rahe Social study by our group, we detected significant differences Readjustament Scale stress score, STAI-trait score, and between OCD patients with or without major depression STAI-state anxiety score. The decision for including these in a fixed time points diurnal cortisol slope (calculated variables was based on the fact that they are known moder- between 10 a.m. and 11 p.m.) but not in the awakening ators of HPA axis activity [5, 6, 13, 39]. As PSS showed a to evening diurnal cortisol slope [6]. Some authors have high correlation with STAI anxiety subscales, we opted for suggested that the CAR is influenced by different bio- not including this variable in the multiple linear regression logical mechanisms than the rest of the diurnal cortisol analyses. Potential sex by OCI-R interactions were tested rhythm [36]. For this reason, in this study, we have de- with a forward selection procedure. Therefore, only signifi- cided to use the diurnal cortisol slope using fixed time cant interactions were included in the final equations. points that do not include the awakening response (be- First, we performed three separate multiple regression tween 10 a.m. and 11 p.m.). analysis using, in each one, a different HPA axis measure The DSTR reflects the cortisol suppression ratio to (CAR, diurnal cortisol slope, and DSTR) as the dependent dexamethasone, a glucocorticoid receptor agonist and variable. In these models, the OC symptoms, as the total provides information about negative feedback of the score of the OCI-R, together with stressful life events and HPA axis. Thus, a lack of suppression after dexametha- anxiety measures, were considered the independent sone administration is considered a measure of gluco- variables. corticoid resistance. The DSTR was defined as the ratio We also performed additional multiple regression ana- of cortisol between T4 and T6 samples (equivalent to lyses as an exploratory approach considering different the ratio between cortisol at 10 am of two consecutive subscales of the OCI-R (checking, hoarding, neutralizing, days: before and after dexamethasone administration). obsessing, ordering, and washing) in relation to each HPA Higher ratios were indicative of greater suppression after axis measure. In these multiple linear regression analyses, dexamethasone administration. one equation was performed for each OCI-R subscore that was included in the model as an independent variable, Statistical analysis along with the same covariates as described before. Poten- Data processing was performed using SPSS 23.0 (SPSS, tial sex by OCI-R subscores interactions were also tested. IBM, USA). Normal distribution for all variables was ex- As only two women were receiving oral contraceptive plored using histograms and normality tests (Kolmogorov- pills, we did not control for this treatment in multivari- Smirnov). In order to test for deviations from normality, ate analyses. However, we repeated the main analyses the distribution of all continuous variables was explored. after excluding these two participants to be sure that the Measures with a skewed distribution were log transformed results did not change. (ln) before its use in parametric tests (e.g., T test, Pearson Sample size calculation was performed with G power correlations) or in linear regression analyses when used as 3.1.9.2. (Franz Faul, Universität Kiel, Germany). With an dependent variables. This was the case for OCI-R and alpha error of 0.05 and a beta error of 0.20 (statistical Holmes-Rahe Social Readjustment scales. However, for power of 80%), considering an effect size (f2) of 0.1 (small cortisol values, we opted for a power transformation to medium), the required sample size was determined to (X’ = (X0.26 − 1)/0.26) to normalize the data, as pro- be 172 for testing the main hypothesis with multiple linear posed by Miller and Plessow [37]. The diurnal cortisol slope regression analyses that included 10 predictors. was calculated with and without transformed cortisol con- centrations. However, when parametric tests and multivari- Results ate analyses were performed, slopes were calculated using Characteristics of the sample transformed cortisol values. To calculate the DSTR, Demographic and clinical characteristics of the sample untransformed cortisol values were used, performing the are shown in Table 1. Significant sex differences were power transformation after calculating the ratio [38]. found in BMI, as men had higher BMI than women. No T tests were used to compare continuous data between significant sex differences were found regarding age, men and women. Pearson correlations were used to explore OCI-R total score and the OCI-R subscales and the the relationship between continuous variables. Significance different psychometric tests evaluated. There were no was set at p < 0.05 (bilateral). sex differences in cortisol concentrations at the different To explore the relationship between the status of the sampling points or the three derived HPA axis measures HPA axis, the OC symptoms, and the psychometric scales, (Table 2). Melia et al. Biology of Sex Differences (2019) 10:55 Page 5 of 11 Table 1 Clinical characteristics from the sample Men Women p (N = 80) (N = 103) Age (years) 41.4 (18.4) 41.3 (17.6) 0.969 Education (years) 12.8 (3.7) 13.5 (3.8) 0.207 BMI (kg/m2) 26.0 (4.6) 24.0 (4.7) 0.006 Smokers, n (%) 19 (23.8) 16 (15.5) 0.161 Tobacco consumption (cig/d), all participants 3.3 (8.1) 1.4 (4.2) 0.058 Tobacco consumption (cig/d), only smokers 14.2 (11.4) 9.3 (6.6) 0.144 OCI-R OCI-R total score 9.8 (8.8) 9.7 (8.9) 0.895 OCI-R hoarding score 2.4 (2.2) 2.7 (2.6) 0.823 OCI-R checking score 1.6 (1.8) 1.4 (2.1) 0.103 OCI-R ordering score 3.0 (2.5) 3.0 (2.1) 0.547 OCI-R neutralizing score 0.7 (1.5) 0.6 (1.2) 0.616 OCI-R washing score 0.6 (1.4) 0.7 (1.3) 0.254 OCI-R obsessing score 1.5 (2.2) 1.4 (1.9) 0.972 OCI-R ≥ 21, n (%) 10 (12.5) 12 (11.7) 0.861 STAI STAI-Trait Anxiety subscore 13.0 (8.0) 14.9 (8.7) 0.131 STAI-State Anxiety subscore 11.7 (7.9) 10.9 (6.7) 0.462 Perceived stress scale 17.0 (7.2) 17.7 (7.5) 0.523 Holmes-Rahe Social Readjustment Scale Number of stressful life events 2.8 (2.6) 3.0 (2.6) 0.572 Stress Score 82.3 (8.8) 90.1 (80.3) 0.356 Abbreviations: BMI body mass index, cig/d cigarettes per day, OCI-R Obsessive-Compulsive Inventory—Revised, STAI State-Trait Anxiety Inventory. Quantitative variables are presented as mean and standard deviation (SD). Table 2 HPA axis measures by sex Men Women p N = 80 N = 103 Awakening cortisol (T1), nmol/L 16.0 (11.7) 17.2 (9.0) 0.256 30-min post-awakening cortisol (T2), nmol/L 25.0 (12.6) 24.6 (14.1) 0.499 60-min post-awakening cortisol (T3), nmol/L 19.9 (12.5) 21.3 (12.6) 0.559 10 a.m. cortisol (T4), nmol/L 13.4 (12.2) 12.1 (8.3) 0.780 11 p.m. cortisol (T5), nmol/L 3.7 (5.6) 3.1 (2.3) 0.965 10 a.m. post-dexamethasone cortisol (T6), nmol/L 7.7 (5.6) 3.1 (2.3) 0.229 CAR† 35.8 (59.8) 26.0 (61.6) 0.295 Diurnal cortisol slope‡ − 0.95 (0.69) − 1.08 (0.71) 0.393 DSTR§ 4.7 (7.1) 10.3 (20.4) 0.086 Abbreviations: CAR cortisol awakening response, DSTR dexamethasone suppression test ratio All variables presented in mean (SD). For cortisol concentrations (T1-T6), diurnal cortisol slope and DSTR, raw data are shown. P values for these HPA axis measures are generated with transformed cortisol values (in the transformation of the DSTR, we calculated the ratio with untransformed cortisol values and performed the power transformation after calculating the ratio) † CAR was calculated with transformed cortisol values ‡ Diurnal cortisol slope was calculated using 10 a.m. and 11 p.m. samples § DSTR = 10 a.m. cortisol/10 a.m. post-dexamethasone cortisol Melia et al. Biology of Sex Differences (2019) 10:55 Page 6 of 11 Correlation analyses events (r = 0.30; p = 0.010) and Holmes-Rahe stress We explored the relationship of cortisol measurements score (r = 0.29, p = 0.021) only in males. with the OCI-R and the psychometric tests. Taking all data No correlation was found between OC symptoms and together, the CAR correlated significantly with stressful life HPA axis measures in the whole sample, but, stratifying events (r = 0.17, p = 0.021), but not with perceived stress or by sex, a blunted CAR was associated with ordering symp- trait/state anxiety. When stratifying by sex (Table 3), toms (r = − 0.22, p = 0.032) in women, whereas a more CAR correlated with the number of stressful life flattened diurnal cortisol slope was associated with Table 3 Sex-stratified correlation analyses exploring the association between psychometric scales and hypothalamic-pituitary- adrenal axis measures Men (N = 80) Number HRSRS stress PSS total STAI-state STAI-trait CAR Diurnal DSTR of SLEs score score anxiety anxiety cortisol slope Number of SLEs 1 0.835** 0.163 0.109 0.080 0.300** 0.210 − 0.133 HRSRS stress score 0.835** 1 0.073 − 0.001 − 0.059 0.290* 0.225 − 0.147 PSS total score 0.163 0.073 1 0.595 ** 0.603** 0.133 0.202 − 0.145 STAI-state anxiety 0.109 − 0.001 0.595** 1 0.624** − 0.091 0.012 − 0.066 STAI-trait anxiety 0.080 − 0.059 0.603 ** 0.624 ** 1 0.129 0.219 0.164 CAR 0.300** 0.290* 0.133 − 0.091 0.129 1 0.332** − 0.064 ** Diurnal cortisol slope 0.210 0.225 0.202 0.012 0.219 0.332 1 0.043 DSTR − 0.133 − 0.147 − 0.145 − 0.066 0.164 − 0.064 0.043 1 OCI-R total score 0.180 −0.031 0.238* 0.235* 0.445** 0.110 0.185 0.019 OCI-R washing 0.242* − 0.012 0.298 ** 0.310 ** 0.406** 0.207 0.241* − 0.063 OCI-R checking 0.073 − 0.140 0.352** 0.329** 0.464** 0.118 0.004 0.037 OCI-R ordering 0.162 − 0.022 0.153 0.194 0.352** 0.062 0.135 0.056 OCI-R obsessing 0.070 − 0.023 0.235 * 0.246 * 0.402** 0.171 0.163 − 0.046 OCI-R hoarding 0.146 0.035 0.060 0.015 0.164 0.151 0.255* − 0.186 OCI-R neutralizing 0.292** 0.045 0.207 0.232 * 0.381** 0.105 0.149 − 0.039 Women (N = 103) Number HRSRS stress PSS total STAI-state STAI-trait CAR Diurnal DSTR of SLEs score score anxiety anxiety cortisol slope Number of SLEs 1 0.832** 0.120 − 0.184 − 0.003 0.091 0.045 − 0.078 HRSRS stress score 0.832** 1 0.248* 0.016 0.254* 0.092 0.080 − 0.013 PSS total score 0.120 0.248 * 1 0.495 ** 0.630** − 0.067 0.051 0.074 STAI-state anxiety − 0.184 0.016 0.495** 1 0.611** 0.013 − 0.058 0.088 STAI-trait anxiety − 0.003 0.254 * 0.630 ** 0.611 ** 1 − 0.147 − 0.017 0.130 CAR 0.091 0.092 − 0.067 0.013 − 0.147 1 0.381** 0.148 Diurnal cortisol slope 0.045 0.080 0.051 − 0.058 − 0.017 0.381** 1 0.151 DSTR − 0.078 − 0.013 0.074 0.088 0.130 0.148 0.151 1 OCI-R total score 0.128 0.270 * 0.354 ** 0.131 0.470** − 0.176 − 0.013 − 0.036 OCI-R washing 0.066 0.085 0.100 0.019 0.224* 0.019 0.043 0.057 OCI-R checking 0.042 0.163 0.239 * − 0.069 0.323** − 0.120 − 0.031 0.046 OCI-R ordering 0.087 0.277** 0.311** 0.185 0.352** − 0.220* − 0.046 − 0.127 OCI-R obsessing − 0.085 0.005 0.455 ** 0.224 * 0.561** − 0.196 0.110 0.095 OCI-R hoarding 0.232* 0.298** 0.168 − 0.014 0.232* 0.001 0.021 − 0.032 OCI-R neutralizing 0.065 0.238* 0.406** 0.145 0.459** − 0.138 0.007 − 0.019 Pearson’s correlation coefficients are shown. Significant results are marked with an asterisk ( p < 0.05; **p < 0.01) * Abbreviations: SLEs stressful life events, HRSRS Holmes-Rahe Social Readjustment Scale, STAI State-Trait Anxiety Inventory, CAR cortisol awakening response, DSTR dexamethasone suppression test ratio, OCI-R Obsessive-Compulsive Inventory—Revised. Melia et al. Biology of Sex Differences (2019) 10:55 Page 7 of 11 hoarding (r = 0.26; p = 0.024) and washing (r = 0.24; When repeating the analyses excluding the two p = 0.033) symptoms in men. women taking oral contraceptive pills, the results did not change. Multiple linear regression analyses Results from the three different multiple regression ana- Discussion lysis considering OC symptoms (OCI-R total score) as Our study suggests that obsessive-compulsive symptoms the main independent variable and HPA axis measure- in healthy individuals are associated with subtle abnor- ments as dependent variables, are shown in Table 4. We malities in HPA axis measures and this association is found a female sex by OC symptoms significant inter- moderated by sex. A blunted CAR was associated with action (standardized beta = - 0.322; p=0.023), indicating the severity of both overall OC symptoms and obsessing that the relationship between OC symptoms and the symptoms in women. A more flattened diurnal cortisol CAR differs between men and women. This interaction slope was associated with ordering symptoms in men. has been depicted in Fig. 1, as there is a positive associ- All these findings were adjusted for stressful life events ation between OC symptoms and the CAR in men, and trait and state anxiety, suggesting an independent whereas a negative association is found in women. No effect of OC symptoms on HPA axis activity. associations were found between OCI-R scores and There is little information in the scientific literature other HPA axis measures (diurnal cortisol slope, DSTR). regarding the relationship among OC symptoms and Trait anxiety was associated with an increased DSTR HPA axis status. In fact, to our knowledge there is only (Table 3). one study that has explored the differences in CAR or When repeating the multiple linear regression analyses diurnal cortisol slope in patients with OCD [6]. In that for each OCI-R subscore, a female sex by OC symptoms study from our group, we did not find significant differ- from the obsessing dimension was found in the analyses ences between OCD patients and healthy individuals in exploring the effect on the CAR (standardized beta = − CAR although OCD patients with comorbid major de- 0.252, p = 0.032), which means that women with more pression had a more flattened cortisol diurnal slope. Few obsessions show a more blunted CAR response. In the studies have explored the HPA axis sensitivity of the analyses regarding the diurnal cortisol slope, OC symp- negative feedback with the classical DST test (adminis- toms from the ordering dimension were associated with tration of 1 mg of dexamethasone and assessing cortisol a more flattened diurnal cortisol slope (standardized beta suppression in plasma), suggesting that OCD patients = 0.365, p = 0.006), and a significant female sex by OC show less suppression to dexamethasone than healthy ordering symptoms significant interaction was also controls [40], particularly if they have comorbid major found (standardized beta = − 0.370, p = 0.019). These re- depression [41, 42]. However, in recent studies from our sults indicate that men with ordering symptoms show a group using a very low dose of dexamethasone and more blunted diurnal cortisol slope whereas women assessing the DSTR in saliva, no differences were found show a more negative slope. There were no significant between OCD and healthy controls nor between OCD associations between OCI-R dimensions nor female sex patients with or without comorbid depression [6]. In by OCI-R dimensions interactions in those equations for that study a positive association between trait anxiety the DSTR. and DSTR was found, which is in accordance with our Table 4 Results of the multiple linear regression analyses exploring the relationship between obsessive-compulsive symptoms and HPA axis measures CAR Diurnal cortisol slope DSTR Beta p Beta p Beta p Female gender 0.115 0.306 − 0.085 0.302 0.070 0.388 Age − 0.316 0.001 − 0.040 0.698 0.146 0.149 BMI 0.029 0.756 − 0.074 0.445 − 0.090 0.346 Smoking (cig/day) 0.054 0.489 0.028 0.726 − 0.020 0.801 OCI-R total score 0.225 0.062 0.060 0.507 − 0.123 0.167 HRSRS stress score 0.023 0.777 0.082 0.348 − 0.027 0.752 STAI-Trait Anxiety 0.037 0.742 0.134 0.251 0.264 0.022 STAI-State Anxiety − 0.071 0.469 − 0.147 0.151 − 0.120 0.234 Female gender × OCI-R interaction − 0.322 0.023 Abbreviations: CAR cortisol awakening response, DSTR dexamethasone suppression test ratio, BMI body mass index, OCI-R Obsessive-Compulsive Inventory—Revised, HRSRS Holmes-Rahe Social Readjustment Scale, STAI State-Trait Anxiety Inventory Melia et al. Biology of Sex Differences (2019) 10:55 Page 8 of 11 Fig. 1 Scatterplot of the relationship between obsessive-compulsive symptoms and the cortisol awakening response by sex results in the current study, suggesting that trait anxiety orbitofrontal cortex (OFC), putamen, and right cerebel- is more relevant for the DST than OC symptoms. lum volume. In contrast, findings related to the order- Although no previous studies have explored whether ing/symmetry dimension are less clear, and have there are sex differences in the relationship between OC included both bigger and smaller volume of OFC, as well symptoms and HPA axis measures in healthy individ- as bigger volume of other frontal regions such as the uals, some studies have explored whether there are sex dorsal anterior cingular cortex and medial frontal cortex differences in the association between neuroticism and [45]. Other studies have reported reduced hippocampal HPA axis indices. In a previous study that included volumes in OCD patients with more severe ordering and undergraduated students and that assessed neuroticism checking symptoms [46]. The hippocampus, which ex- with the NEO five-factor inventory, salivary cortisol at erts negative feedback on the HPA axis via glucocortic- midday (between 10:30 a.m. and 2:30 p.m.) were posi- oid and mineralocorticoid receptors, shows sex tively associated with neuroticism in men and negatively differences in response to stressors, with less remodeling in women [43]. In another study that included university of hippocampal CA3 dendrites in females after chronic students that were also assessed with the NEO 5-factor stress [47]. The hippocampus is also thought to play a inventory, men had more flattened diurnal cortisol role in the cortisol awakening response, as subjects with slopes but not differences in the CAR [44]. These two hippocampal damage show a blunted CAR [48]. Thus, it studies are in line with our results that suggest more is plausible that some findings regarding sex differences flattened cortisol slopes in healthy individuals with more on the association between OC symptoms and HPA axis ordering OC symptoms. However, it is difficult to com- measures (e.g., blunted CAR and more prominent obses- pare our results with the scientific literature because sing symptoms in women) could be explained by sex dif- particular studies focused on OC symptoms are lacking. ferences in neurobiological substrates or brain regions Clearly, further studies are needed to replicate our find- that are involved both in HPA axis regulation and the ings. The distinct associations between particular OC clinical expression of OC symptoms. In line with this hy- symptom dimensions and HPA axis measures might be pothesis, in the MRI study by Ress et al. [46] which explained by differences in the brain correlates of these assessed the relationship between OC symptom dimen- symptom dimensions. As discussed in a recent review sions and hippocampal volumes, greater volume reduc- [45], several OCD studies have found that increased se- tions in the hippocampus were observed in a subset of verity of aggressive/checking obsessions to be related to OCD patients that reported greater obsessing and check- smaller gray matter volume in the temporal lobes, ex- ing symptoms. Interestingly, in this study, there was a tending into the amygdala and insula, as well as the left significant main effect of sex on global hippocampal Melia et al. Biology of Sex Differences (2019) 10:55 Page 9 of 11 volume, which suggests that sex is a moderating variable HPA axis activity and OC symptoms in women and in the relationship between OC symptoms and hippo- men. In summary, our study suggests that sex is a mod- campal volumes. If this is the case, it is plausible that erator of the relationship between OC symptoms and there might also exist sex differences in the relation- HPA axis functionality, as women with more obsessions ship between the CAR and obsessing symptoms, be- showed a more blunted CAR response. Future studies cause of the important role of the hippocampus on need to replicate our findings in clinical populations in- both the regulation of the negative feedback of the cluding OCD patients in order to know whether these HPA axis and the CAR. sex differences are also present in a more severe pheno- Some methodological decisions and limitations of our type. Our study suggests that distinct OC symptom di- study need to be discussed. We aimed to study the rela- mensions might show different associations with HPA tionship between OC symptoms and HPA axis measures axis measures, and points out to the possibility that in a non-clinical sample. This decision limits the these differences could be secondary to potential alter- generalizability of the findings to individuals with OCD. ations in neural substrates that are involved in the regu- However, it allows to test the association between less lation of the HPA axis and the clinical expression of OC severe OC symptoms and subtle HPA axis abnormalities symptoms. Although this is a speculative issue, future re- that are not influenced by the severity of a clinical diag- search is needed to shed light on this issue. In addition, nosis, comorbid conditions (e.g., major depression), or psychoneuroendocrinological studies exploring the asso- treatments. This “less biased”’ approach might help to ciation between OC or anxiety symptoms with HPA axis better explore whether a dysregulation of the HPA axis activity need to control for potential sex differences. is contributing to OC symptoms. We did only assess the Acknowledgements CAR over one day. As dexamethasone was administered The authors are very grateful to all of the study participants, as well as to the at 11:00 p.m., we dismissed the possibility of collecting staff from Bellvitge University Hospital and Institut Pere Mata, who helped to further CAR sampling the next day. Although partici- collect the sample used in this study. We would also like to thank the staff of Biopsychology at the Department of Psychology of the Technische pants were instructed to collect the saliva samples at Universität Dresden for analyzing salivary cortisol samples and the home at the specific times, the timing of the sampling technicians from the Biobanc IISPV of Reus and the Biobank HUB-ICO- was not verified with objective methods (e.g., electronic IDIBELL. monitoring systems). Inaccurate sample timing can bias Authors’ contributions CAR estimates [13]. Therefore, it is possible that differ- JL, VS, and JMM designed the study. NSP, AC, AGZ, MU, JMC, and PA ences in the accuracy of sampling times might be driven participated in the recruitment of participants. EV was responsible for the processing of biological samples. JL, VS, and RN analyzed the data. CSM by obsessive-compulsive symptoms (e.g., more accurate reviewed the scientific literature and wrote the first draft of the manuscript. sampling timing in people with more obsessive- All authors participated in the discussion of the results and approved the compulsive symptoms). As already explained in previous final manuscript. studies from our group [6, 49], we used a very low dose Funding of dexamethasone (0.25 mg) because salivary cortisol This study was supported in part by grants from the Carlos III Health Institute presents more profound suppression than does plasma through the Ministry of Science, Innovation and Universities (PI10/01753, cortisol and because we aimed to explore the DSTR as a PI10/01607, PI15/01386 and PI15/00662), the European Regional Development Fund (ERDF) “A way to build Europe”, the Fundació La Marató continuous measure (ratio), and the use of higher dexa- de TV3 (092230/092231), CIBERSAM, and the Catalan Agency for the methasone doses would not have allowed us to detect Management of University and Research Grants (AGAUR 2017 SGR 1247 and subtle alterations in HPA axis regulation. The cross- 2017 SGR 632). Javier Labad received an Intensification of the Research Activity Grant (SLT006/17/00012) by the Health Department of the sectional design of our study prevents us from inferring Generalitat de Catalunya during 2018 and 2019. Samples from patients causality in the association between OC symptoms and included in this study were processed and preserved by the Biobanc IISPV HPA axis measures. Some exploratory analyses were not and the Biobank HUB-ICO-IDIBELL, integrated in the Spanish National Bio- banks Network (PT17/0015/0024) and Xarxa Banc de Tumors. corrected for multiple testing (e.g., associations between The funders had no role in the study design, data collection and analysis, distinct OC symptom dimensions and different HPA decision to publish, or preparation of the manuscript. axis measures) [50]. However, it is important to under- Availability of data and materials score that these results on OC symptom dimensions are Additional data regarding the current study are available from the exploratory and that they need to be confirmed in fur- corresponding author on reasonable request. ther confirmatory studies. Ethics approval and consent to participate The research protocol was approved by the Ethics Committees of Bellvitge University Hospital and Hospital Universitari Sant Joan, and all participants Perspectives and significance provided written informed consent after have received a full explanation of the study. Although we did not find sex differences in OC symp- toms, state/trait anxiety or HPA axis measures in healthy Consent for publication individuals, we found a different association between Not applicable Melia et al. Biology of Sex Differences (2019) 10:55 Page 10 of 11 Competing interests 16. Stroud LR, Salovey P, Epel ES. Sex differences in stress responses: The authors declare that they have no competing interests. social rejection versus achievement stress. Biol Psychiatry. 2002;52: 318–27. Author details 17. Kudielka BM, Buske-Kirschbaum A, Hellhammer DH, Kirschbaum C. HPA axis 1 Department of Psychiatry, Bellvitge University Hospital, Bellvitge Biomedical responses to laboratory psychosocial stress in healthy elderly adults, Research Institute (IDIBELL), Neurosciences Group—Psychiatry and Mental younger adults, and children: impact of age and gender. Health, Barcelona, Spain. 2Centro de Investigación Biomédica en Red de Psychoneuroendocrinology. 2004;69:113–32. Salud Mental (CIBERSAM), Carlos III Health Institute, Madrid, Spain. 18. Labad J, Menchon JM, Alonso P, Segalas C, Jimenez S, Jaurrieta N, et al. 3 Department of Clinical Sciences, School of Medicine, Universitat de Gender differences in obsessive-compulsive symtpom dimensions. Depress Barcelona, Barcelona, Spain. 4Hospital Universitari Institut Pere Mata, IISPV, Anxiety. 2008;838:832–8. Universitat Rovira i Virgili, Reus, Spain. 5Institut de Neurociències, Universitat 19. Fullana MA, Vilagut G, Rojas-Farreras S, Mataix-Cols D, de Graaf R, Autònoma de Barcelona, Cerdanyola del Vallès, Spain. 6Department of Demyttenaere K, et al. Obsessive-compulsive symptom dimensions in Mental Health, Parc Taulí Hospital Universitari, I3PT, Universitat Autònoma de the general population: results from an epidemiological study in six Barcelona, Sabadell, Spain. European countries. J Affect Disord. 2010;124:291–9. 20. Torresan RC, Ramos-Cerqueira ATA, Shavitt RG. do Rosário MC, de Mathis Received: 25 July 2019 Accepted: 21 November 2019 MA, Miguel EC, et al. Symptom dimensions, clinical course and comorbidity in men and women with obsessive-compulsive disorder. Psychiatry Res. 2013;209:186–95. 21. Zhang X, Liu J, Cui J, Liu C. Study of symptom dimensions and clinical References characteristics in Chinese patients with OCD. J Affect Disord. 2013;151:868–74. 1. Ruscio AM, Stein DJ, Chiu WT, Kessler RC. The epidemiology of obsessive- 22. Cherian AV, Narayanaswamy JC, Viswanath B, Guru N, George CM, Bada compulsive disorder in the National Comorbidity Survey Replication. Mol Math S, et al. Gender differences in obsessive-compulsive disorder: Psychiatry. Nature Publishing Group. 2010;15:53–63. findings from a large Indian sample. Asian J Psychiatr. 2014;9:17–21. 2. Leckman JF, Denys D, Simpson HB, Mataix-Cols D, Hollander E, Saxena S, 23. Raines AM, Oglesby ME, Allan NP, Mathes BM, Sutton CA, Schmidt NB. et al. Obsessive-compulsive disorder: a review of the diagnostic criteria and Examining the role of sex differences in obsessive-compulsive symptom possible subtypes and dimensional specifiers for DSM-V. Depress. Anxiety. dimensions. Psychiatry Res. Elsevier Ireland Ltd. 2018;259:265–9. 2010;27:507–27. 24. Lobo A, Perez-Echeverria MJ, Artal J. Validity of the scaled version of the 3. Fineberg NA, Hengartner MP, Bergbaum CE, Gale TM, Gamma A, Ajdacic- General Health Questionnaire (GHQ-28) in a Spanish population. Psychol Gross V, et al. A prospective population-based cohort study of the Med. 1986;16:135–40. prevalence, incidence and impact of obsessive-compulsive 25. Foa EB, Huppert JD, Leiberg S, Langner R, Kichic R, Hajcak G, et al. The symptomatology. Int J Psychiatry Clin Pract. 2013;17:170–8. Obsessive-Compulsive Inventory: development and validation of a short 4. Fineberg NA, Apergis-Schoute AM, Vaghi MM, Banca P, Gillan CM, Voon V, version. Psychol Assess. 2002;14:485–96. et al. Mapping compulsivity in the DSM-5 obsessive compulsive and related 26. Huppert JD, Walther MR, Hajcak G, Yadin E, Foa EB, Simpson HB, et al. The disorders: cognitive domains, neural circuitry, and treatment. Int. J. OCI-R: validation of the subscales in a clinical sample. J Anxiety Disord. Neuropsychopharmacol. 2018;21:42–58. 2007;21:394–406. 5. Kudielka BM, Kirschbaum C. Sex differences in HPA axis responses to stress: 27. Hajcak G, Huppert JD, Simons RF, Foa EB. Psychometric properties of the A review. Biol Psychol. 2005;69:113–32. OCI-R in a college sample. 2004;42:115–123. 6. Labad J, Soria V, Salvat-Pujol N, Segalàs C, Real E, Urretavizcaya M, et al. 28. Fullana MA, Tortella-feliu M, Mataix-cols D. Psychometric properties of the Hypothalamic-pituitary-adrenal axis activity in the comorbidity between Spanish version of the Obsessive – Compulsive Inventory — Revised in a obsessive-compulsive disorder and major depression. non-clinical sample. 2005;19:893–903. Psychoneuroendocrinology. 2018;93:20–8. 29. Spielberger C. Manual for the State-Trait Anxiety Inventory (STAI). Consult 7. Real E, Labad J, Alonso P, Segalàs C, Jiménez-Murcia S, Bueno B, et al. Psychol Press. 1983:4–26. Stressful life events at onset of obsessive-compulsive disorder are associated 30. Holmes TH, Rahe RH. The Social Readjustment Rating Scale. J Psychosom with a distinct clinical pattern. Depress Anxiety. 28:367–76. Res. 1967;11:213–8. 8. Faravelli C, Lo SC, Lelli L, Pietrini F, Lazzeretti L, Godini L, et al. The role of 31. De La Revilla L. De Los Ríos Álvarez AM, Luna Del Castillo JDD. Factores que life events and HPA axis in anxiety disorders: a review. Curr Pharm Des. intervienen en la producción de los problemas psicosociales. Aten Primaria. 2011;18:5663–74. 2007;39:305–11. 9. Fossey MD, Lydiard RB, Ballenger JC, Laraia MT, Bissette G, Nemeroff CB. 32. Cohen S, Kamarck T, Mermelstein R. A global measure of perceived stress. J Cerebrospinal fluid corticotropin-releasing factor concentrations in patients Health Soc Behav. 1983;24:385–96. with anxiety disorders and normal comparison subjects. Biol Psychiatry. 33. Pruessner JC, Wolf OT, Hellhammer DH, Buske-Kirschbaum A, von Auer K, Jobst 1996;39:703–37. S, et al. Free cortisol levels after awakening: a reliable biological marker for the 10. Kluge M, Schüssler P, Künzel HE, Dresler M, Yassouridis A, Steiger A. assessment of adrenocortical activity. Life Sci. 1997;61:2539–49. Increased nocturnal secretion of ACTH and cortisol in obsessive compulsive 34. Pruessner JC, Kirschbaum C, Meinlschmid G, Hellhammer DH. Two formulas disorder. J Psychiatr Res. 2007;41:928–33. for computation of the area under the curve represent measures of total 11. Gustafsson PEAE, Ivarsson T, Nelson N. Diurnal cortisol levels and cortisol hormone concentration versus time-dependent change. response in youths with obsessive-compulsive disorder. Psychoneuroendocrinology. 2003;28:916–31. Neuropsychobiology. 2008;57:14–21. 35. Adam EK, Quinn ME, Tavernier R, McQuillan MT, Dahlke KA, Gilbert KE. Diurnal 12. Morgado P, Freitas D, Bessa JM, Sousa N, Cerqueira JJ. Perceived stress in cortisol slopes and mental and physical health outcomes: a systematic review obsessive-compulsive disorder is related with obsessive but not compulsive and meta-analysis. Psychoneuroendocrinology. 2017;83:25–41. symptoms. Front Psychiatry. 2013;4:21. 36. Clow A, Hucklebridge F, Stalder T, Evans P, Thorn L. The cortisol awakening 13. Stalder T, Kirschbaum C, Kudielka BM, Adam EK, Pruessner JC, Wüst S, response: more than a measure of HPA axis function. Neurosci Biobehav et al. Assessment of the cortisol awakening response: expert Rev. 2010;35:97–103. consensus guidelines. Psychoneuroendocrinology [Internet]. 2016;63: 37. Miller R, Plessow F. Transformation techniques for cross-sectional and 414–32. longitudinal endocrine data: application to salivary cortisol concentrations. 14. Kanehisa M, Kawashima C, Nakanishi M, Okamoto K, Oshita H, Masuda K, Psychoneuroendocrinology. 2013;38:941–6. et al. Gender differences in automatic thoughts and cortisol and alpha- 38. Salvat-Pujol N, Labad J, Urretavizcaya M, de Arriba-Arnau A. Segalàs C, Real amylase responses to acute psychosocial stress in patients with obsessive- E, et al. Hypothalamic-pituitary-adrenal axis activity and cognition in major compulsive personality disorder. J Affect Disord. 2017;217:1–7. depression: the role of remission status. Psychoneuroendocrinology. 2017; 15. Panagiotakopoulos L, Neigh GN. Development of the HPA axis: where 76:38–48. and when do sex differences manifest? Front Neuroendocrinol. 2014;35: 39. Champaneri S, Xu X, Carnethon MR, Alain G, Seeman T, Desantis AS, 285–302. et al. Diurnal salivary cortisol is associated with body mass index and Melia et al. Biology of Sex Differences (2019) 10:55 Page 11 of 11 waist circumference: the Multi-Ethnic Study of Atherosclerosis. Obesity. 2013;21:1–17. 40. Catapano F, Monteleone P, Maj M, Kemali D. Dexamethasone suppression test in patients withprimary obsessive-compulsive disorder and in healthy controls. Neuropsychobiology. 1990. 41. Vallejo J, Olivares J, Marcos T, Martínez-Osaba MJ, Ribera F, Bulbena A. Dexamethasone suppression test and primary obsessional compulsive disorder. Compr Psychiatry. 1988. 42. Kuloǧlu M, Atmaca M, Önal S, Geçici Ö, Bulut V, Tezcan E. Neopterin levels and dexamethasone suppression test in obsessive-compulsive disorder. Psychiatry Res. 2007. 43. DeSoto MC, Salinas M. Neuroticism and cortisol: the importance of checking for sex differences. Psychoneuroendocrinology. 2015;23:53–6. 44. Garcia-Banda G, Chellew K, Fornes J, Perez G, Servera M, Evans P. Neuroticism and cortisol: pinning down an expected effect. Int J Psychophysiol. 2014;91:132–8. 45. Thorsen AL, Kvale G, Hansen B, van den Heuvel OA. Symptom dimensions in Obsessive-Compulsive Disorder as predictors of neurobiology and treatment response. Curr. Treat. Options Psychiatry. 2018;5:182–94. 46. Reess TJ, Rus OG, Gürsel DA, Schmitz-Koep B, Wagner G, Berberich G, et al. Association between hippocampus volume and symptom profiles in obsessive–compulsive disorder. NeuroImage Clin. 2018;17:474–80. 47. McEwen BS, Nasca C, Gray JD. Stress effects on neuronal structure: hippocampus, amygdala, and prefrontal cortex. Neuropsychopharmacology. 2016;41:3–23. 48. Buchanan TW, Kern S, Allen JS, Tranel D, Kirschbaum C. Circadian regulation of cortisol after hippocampal damage in humans. Biol Psychiatry. 2004;56: 651–6. 49. Salvat-Pujol N, Labad J, Urretavizcaya M, de Arriba-Arnau A, Segalàs C, Real E, et al. Hypothalamic-pituitary-adrenal axis activity and cognition in major depression: the role of remission status. Psychoneuroendocrinology. 2017; 76:38–48. 50. Bender R, Lange S. Adjusting for multiple testing - when and how? J Clin Epidemiol. 2001;54:343–9. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.