Risk of postpartum depression in association with serum leptin and interleukin-6 levels at delivery: A nested case–control study within the UPPSAT cohort

Psychoneuroendocrinology (2009) 34, 1329—1337

a v a i l a b l e a t w w w. s c i e n c e d i r e c t . c o m

j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / p s y n e u e n

Risk of postpartum depression in association with serum leptin and interleukin-6 levels at delivery: A nested case—control study within the UPPSAT cohort ´n a, Fotios C. Papadopoulos b, Alkistis Skalkidou a,*, Sara M. Sylve ¨m-Poromaa a Matts Olovsson a, Anders Larsson c, Inger Sundstro a

Department of Women’s and Children’s Health, Uppsala University, 751 85 Uppsala, Sweden Department of Neurosciences, Psychiatry, Uppsala University Hospital, 751 85 Uppsala, Sweden c Department of Department of Medical Sciences, Uppsala University, 751 85 Uppsala, Sweden b

Received 30 October 2008; received in revised form 12 February 2009; accepted 4 April 2009

KEYWORDS Biological marker; Postpartum depression; Postnatal depression; Leptin; Interleukin; EPDS

Summary Although postpartum depression (PPD) is a common condition, it often goes undiagnosed and untreated, with devastating consequences for the woman’s ability to perform daily activities, to bond with her infant and to relate to the infant’s father. Leptin, a protein synthesised in the adipose tissue and involved in regulation of food intake and energy expenditure has been related to depressive disorders, but studies report conflicting results. The aim of this study was to evaluate the association between serum leptin levels at the time of delivery and the subsequent development of postpartum depression in women, using data from a populationbased cohort of delivering women in Uppsala, Sweden. Three hundred and forty seven women from which serum was obtained at the time of delivery filled out at least one of three structured questionnaires containing the Edinburgh Scale for Postnatal Depression (EPDS) at five days, six weeks and six months after delivery. Mean leptin levels at delivery did not significantly differ between the 67 cases of PPD and the 280 controls. Using linear regression analysis and adjusting for maternal age, body-mass index, smoking, interleukin-6 levels, duration of gestation and gender of the newborn, the EPDS scores at six weeks and six months after delivery were found to be negatively associated with leptin levels at delivery ( p < 0.05). Serum leptin levels at delivery were found to be negatively associated with self-reported depression during the first six months after delivery. No such association was found concerning serum IL-6 levels at delivery. If these finding are replicated by other studies, leptin levels at delivery could eventually serve as a biological marker for the prediction of postpartum depression. # 2009 Elsevier Ltd. All rights reserved.

* Corresponding author. Tel.: +46 18610000. E-mail address: [email protected] (A. Skalkidou). 0306-4530/$ — see front matter # 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.psyneuen.2009.04.003

1330

1. Introduction Depressed mood in the postpartum period is common, affecting nearly half of all newly delivered women during the first days after delivery (baby blues or postpartum blues). This transient mood disturbance may lead to more severe and persistent depression during the following weeks, eventually fulfilling the diagnostic criteria for major depression. According to the Diagnostic and Statistical Manual of Mental Disorders, fourth edition, text revision (2000), an episode of depression is considered to have postpartum onset if it begins within four weeks after delivery. However, onset within one year after delivery is the time frame commonly used. Postpartum depression is a common disorder affecting approximately 13—20% of all newly delivered women (O’Hara and Swain, 1996; Josefsson et al., 2001). It nevertheless often goes undiagnosed and can thus lead to devastating consequences in women’s ability to perform daily activities, to bond with her infant, and to relate to the infant’s father (Beck, 1995). Detrimental effects of postnatal depression have also been reported on the cognitive and emotional development of children (Sharp et al., 1995; Murray et al., 2003). Leptin is a protein synthesised in the adipose tissue which is coded by the obese (ob) gene. Leptin is believed to be a messenger from adipose tissue to the brain, which acts by binding to specific receptors in the hypothalamus, thus decreasing food intake and increasing energy expenditure (Jequier, 2002). The results of the studies concerning leptin levels in depression have so far been inconsistent with leptin levels being reported to be unaltered (Deuschle et al., 1996), increased (Antonijevic et al., 1998; Gecici et al., 2005) and decreased (Kraus et al., 2001; Jow et al., 2006; Yang et al., 2007) in depressed patients. One study found that high leptin levels are predictive of a new depressive episode during a 5year period only among non-smokers (Pasco et al., 2008). Some studies report increased leptin levels only in females with major depression (Rubin et al., 2002; Esel et al., 2005). Different selection criteria and recruitment strategies, time of blood sampling, depression subtypes (typical versus atypical) and subsequent hyper- or hypo-activity of the hypothalamic—pituitary—adrenal (HPA) axis, co-morbid anxiety disorders and gender differences may at least in part explain these discrepancies. Interleukin-6 is an pro-inflammatory cytokine with a variety of endocrine and metabolic actions apart from its hematologic and immune effects. It stimulates the HPA axis (Papanicolaou et al., 1998), regulates the adrenal cortex cortisol response to adrenocorticotropic hormone (ACTH) stimulation (Zarkovic´ et al., 2008) and is negatively regulated by glucocorticoids. Furthermore, it is secreted during stress and is positively controlled by catecholamines (Papanicolaou et al., 1998). There is now evidence that activation of the inflammatory response system may be involved in the pathophysiology of anxiety states and major depression. There is also strong evidence that major depression is accompanied by increased serum concentrations of interleukin-6, among other markers of the inflammatory response (Maes et al., 1995; Berk et al., 1997; Sluzewska et al., 1995). The postpartum period per se is accompanied by an increased inflammatory capacity and women with postpartum depressive symptomatology have been reported to display significantly

A. Skalkidou et al. higher serum interleukin-6 receptor levels (Maes et al., 2000), lower salivary cortisol levels, lower interferongamma/interleukin-10 ratio (Groer and Morgan, 2007), as well as elevated urinary interleukin-1 beta levels in the early puerperium (Corwin et al., 2003). A recent study has even shown that higher cerebrospinal fluid interleukin-6 levels at the time of delivery are associated with risk for development of postpartum depression (Boufidou et al., 2009). A model has recently been proposed for the development of postpartum depression, which involves both the hypothalamic—pituitary—adrenal axis as well as the innate immune system (Corwin and Pajer, 2008). The HPA axis has profound effects on immunity, metabolism and reproduction. Both hyper-activation and hypo-activation of the axis have been associated with depressive states (Chrousos and Gold, 1992). On the other hand, prolonged or excessive activation of the pro-inflammatory immune response may be involved in the pathogenesis of depression (Raison et al., 2006). A bidirectional association between these two systems has been thoroughly documented (Kiecolt-Glaser and Glaser, 2002; Chrousos and Kino, 2005). Both systems are profoundly affected both during pregnancy and the postpartum state, rendering fertile ground for the development of depressive symptomatology (Corwin and Pajer, 2008). It is well documented that leptin interacts both with the innate immune system, exhibiting a strong pro-inflammatory effect (Loffreda et al., 1998), as well as with the HPA axis, targeting the central nervous system, the pituitary gland and the adrenals (Walker et al., 2004). Animal models suggest that there is a bidirectional regulation of responsiveness to the HPA axis between mother and newborn which is mediated by leptin (Oates et al., 2000). Both leptin and IL-6 have independently been implicated in the pathogenesis of depression. To our knowledge, no studies on the interaction between leptin and IL-6 levels and postpartum depression have been conducted. The aim of this study was to evaluate the potential association between leptin and IL-6 levels around the time of delivery and the subsequent development of postpartum depression in women.

2. Methods This study was undertaken as part of the UPPSAT project, a population-based cohort study in the county of Uppsala, Sweden, investigating correlates of postpartum depression. The study was conducted at the Department of Obstetrics and Gynecology in Uppsala University Hospital. Uppsala county is a medium-sized Swedish county with a population of 323 270 inhabitants and the University Hospital is responsible for all delivering women within the county, as well as for high risk pregnancies from nearby counties.

2.1. Study population During the period between November, 2006 and May, 2007, all eligible women giving birth at Uppsala University Hospital were contacted by their midwife or midwife assistant after delivery and asked about their willingness to take part in a longitudinal study of maternal, paternal and infant wellbeing. Exclusion criteria for the study were (1) not being

Leptin and interleukin-6 levels in association with postpartum depression able to adequately communicate in Swedish, (2) women whose personal data were kept confidential, (3) women with intrauterine demise or with infants immediately admitted in the neonatal intensive care unit. The mothers received both oral and written information about the study objectives and written consent was obtained. The study subjects completed a self-administered structured questionnaire containing the Edinburgh Postnatal Depression Scale (EPDS) five days after delivery. Two more questionnaires were sent to the study subjects, one at 6 weeks after delivery and one at 6 months after delivery. The women were instructed to complete and send back the questionnaires by post. The study protocol was approved by the Independent Research Ethics Board of Uppsala University.

2.2. Outcome measures The mother’s score on the Swedish version of the EPDS (Cox et al., 1987) was used as the primary outcome measure. The EPDS has been validated in Sweden and has a cut-off of 12 points, after which a mother is considered being at high risk of postpartum depression (Wickberg and Hwang, 1996). The EPDS is usually administered at least 2 weeks following delivery, because of the high prevalence of postpartum blues occurring before this time (Lee et al., 2003). Therefore, for the current study, a mother was considered as a case of postnatal depression if she had an EPDS score of more or equal to 14 at five days after delivery, or a score of more or equal to 12 at six weeks postpartum, or a score of more than 12 at six months postpartum.

2.3. Blood samples During the time-frame for the current study, one blood sample from each delivering woman was collected. The blood samples were collected in conjunction with routine intravenous catheterization before delivery. Because consent for the UPPSAT study was obtained after the delivery, blood samples from women who did not to consent to participate in the study were discarded (40% of all blood samples collected). This procedure was used in order to minimize selection bias and to avoid unnecessary extra blood sampling after the women received written information following delivery. Coded blood samples were stored at 4 8C for a maximum of 24 h and then centrifuged. The sera were stored at 70 8C.

2.4. Analysis of leptin and high sensitivity IL-6 Samples were analyzed using commercially available ELISA kits (Leptin kit DY398 and high sensitivity IL-6 kit HS600B, R&D Systems, Minneapolis, MN, USA). The assays employ the quantitative sandwich enzyme immunoassay technique. A monoclonal antibody specific for IL-6 or leptin has been pre-coated onto a microplate. Standards and samples are pipetted into the wells and any IL-6 or leptin present is bound by the immobilized antibody. After washing away any unbound substances, a specific detection antibody is added to the wells. Following a wash to remove unbound antibodyenzyme reagent, a substrate solution is added to the wells. (For the IL-6 assay, an amplifier solution had to be added to

1331

the substrate to achieve a proper color development.) The color develops in proportion to the amount of antigen bound in the initial step. The concentration of IL-6 and leptin in the samples are calculated using the standard curve. The immunoassays are calibrated against highly purified E. coliexpressed recombinant human IL-6 and leptin. The assays have a total coefficient of variation (CV) of approximately 7%.

2.5. Statistical analyses SPSS version 15.0 was used for the statistical analyses. Statistical significance was set at a p-value of <0.05. For the current sub-study of the UPPSAT project, women consuming alcohol during pregnancy were excluded because of reports indicating that serum leptin levels are sensitive to alcohol consumption, an effect mediated by an altered hormone gene expression in the adipose tissue (Pravdova and Fickova, 2006). Differences in the study variables among cases and controls were assessed with the Mann—Whitney U-test, to account for non-normality, or the Fisher’s exact test, because of the small number of individuals involved in these comparisons. Possible correlations of leptin with maternal bodymass index (BMI), maternal age and interleukin-6 levels, as well as with duration of gestation were assessed separately among cases and controls using the Spearman Correlation Coefficient. Differences in mean serum leptin concentration by smoking status during pregnancy and newborn gender were assessed with the Mann—Whitney U-test. Subsequently, a multivariate linear regression model was used, with the EPDS score as the outcome variable and leptin (in increments of 1 standard deviation of the hormone among controls), interleukin-6 (in increments of 1 standard deviation of the hormone among controls), as well as a series of possible confounders (maternal age, body-mass index, smoking, interleukin-6 levels, duration of gestation and gender of the newborn) as predictor variables. The EPDS score at five days, six weeks and six months after delivery were used in the models after logarithmic transformation, in order to account for non-normality. The possible confounders were included in the model based on a positive association with either case— control status or leptin levels, with a cut-off p-value < 0.25. BMI of the mother at six weeks postpartum was used instead of BMI at delivery, as a better proxy of her regular BMI.

3. Results In total, one blood sample, written consent, as well as at least one completed questionnaire were available for 365 women. Eighteen women who reported alcohol use during pregnancy were excluded from the analysis because of the reported significant alterations in leptin levels among alcohol consumers, leaving 347 to be included in the analyses. The number of completed questionnaires, percentage of mothers with self-reported depression and mean, range and standard deviation of the EPDS score at the three postpartum time-points are shown in Table 1. No significant variations in leptin or IL-6 levels were detected in relation to time of blood sampling, number of hours fasting, or cervical status at blood sampling, data not shown. Likewise, no significant correlations between leptin

1332

A. Skalkidou et al.

Table 1 Number of completed questionnaires (N), percentage of mothers with self-reported depression (%) and Mean, Range and Standard deviation (SD) of the EPDS score at the three points postpartum.

5 days 6 weeks 6 months

Table 2

N

% depressed

Mean EPDS

Range EPDS

SD EPDS

256 274 228

12.9% EPDS  14 12.0% EPDS  12 11.4% EPDS  12

6.30 5.84 5.15

20 19 22

4.821 4.372 4.672

Anthropometric, lifestyle, obstetric and hormonal variables in 67 cases of postpartum depression and 280 controls.

Variable

Cases

Controls

p-value *

Age (years) (median, mean  SD) BMI 6 weeks postpartum (kg/m2) (median, mean  SD) Leptin (ng/mL) (median, mean  SD) IL-6 (pg/mL) (median, mean  SD) Duration of gestation (days) (median, mean  SD)

30.0, 30.2  5.1 25.1, 27.0  4.70 26.8, 34.9  34.0 8.4, 10.6  9.5 280, 277  11.38

31.0, 30.9  4.01 24.5, 25.2  4.105 28.9, 38.2  34.1 5.9, 10.2  10.4 280, 279  12.10

0.291 0.407 0.234 0.365 0.362

Variable

Cases

Controls

p-value **

Maternal smoking during pregnancy (N, %) Use of antidepressants at delivery (N, %)

3 (5.6%) 4 (6.0%)

5 (2.7%) 3(1.1%)

0.267 0.028

* **

Mann—Whitney U-Test. Fisher’s exact Test.

and IL-6 levels and the above mentioned variables were found. Table 2 presents the distribution of anthropometric, lifestyle, pregnancy-related and hormonal characteristics among cases and controls. The data serve mostly descriptive purposes and are not directly interpretable due to the existence of mutual confounding. Although mean serum leptin levels were lower in cases compared to controls, this difference did not reach statistical significance using the Mann— Whitney U-test. Accordingly, mean IL-6 levels were higher among cases compared to controls, but this difference did not reach statistical significance either. Users of antidepressants at the time around delivery were found to be at higher risk for postpartum depression using the Fischer’s exact test. In Tables 3a and 3b, the Spearman Correlation Coefficients between leptin, IL-6 and the continuous variables that might act as possible confounders among cases and controls respectively are displayed. As expected, leptin levels at delivery were positively associated with maternal BMI 6 weeks postpartum, but not significantly so with maternal age, possibly because the age-range in our sample was relatively narrow. IL-6 levels were significantly associated with leptin levels.

Among cases, IL-6 levels were found to be correlated with maternal age. In Table 4, the median and mean leptin and IL-6 levels by smoking status during pregnancy and newborn gender are presented. A statistically significant difference was observed in both mean leptin and as IL-6 levels according to the mothers smoking status during pregnancy, with smokers having lower leptin and IL-6 levels than non-smokers. A difference that approached statistical significance was observed according to the gender of the newborn, with mothers of female offspring having higher leptin levels. Table 5 shows the results of the linear regression models with EPDS score as the outcome variable and leptin, IL-6 and possible confounders as predictor variables. EPDS score was introduced as the outcome variable after logarithmic transformation in order to account for non-normality. In these analyses, adjustments were made for maternal BMI at six weeks postpartum, maternal age, duration of gestation, offspring gender, as well as maternal smoking status during pregnancy. The regression coefficients are presented in the table after exponentiation of the coefficients provided by the model (anti-logarithmic transformation). Lower maternal

Table 3a Correlations between serum leptin, maternal age, body-mass index (BMI) at 6 six weeks postpartum, duration of gestation and serum interleukin-6 (IL-6) among the 280 controls.

Table 3b Correlations between serum leptin, maternal age, body-mass index (BMI) at 6 six weeks postpartum, duration of gestation and serum interleukin-6 (IL-6) among the 67 cases of self-reported postpartum depression.

Variable Leptin IL-6 *

Age 0.004 0.083

BMI 0.230 * 0.019

Duration of gestation

IL-6

Variable

0.085 0.088

0.166 * —

Leptin IL-6

p < 0.01, Spearman Correlation Coefficient.

*

Age 0.155 0.395 *

BMI 0.367 * 0.108

Duration of gestation 0.083 0.122

p < 0.01, Spearman Correlation Coefficient.

IL-6 0.185 —

Leptin and interleukin-6 levels in association with postpartum depression

1333

Table 4 Serum leptin levels and interleukin-6 (IL-6) levels (median, mean  standard deviation, SD) in relation to maternal smoking status during pregnancy and newborn gender among the 280 controls. Leptin levels Median, mean  SD Maternal smoking during pregnancy No (N = 228) Yes (N = 8)

27.1, 36.8  33.3 14.0, 18.4  14.6

Newborn gender Girl (N = 163) Boy (N = 183)

31.1, 40.1  36.7 26.4, 35.4  31.4

*

p-value *

IL-6 levels Median, mean  SD

0.045

p-value * 0.332

6.6, 10.3  10.0 4.9, 7.6  8.2 0.100

0.860 6.2, 10.4  10.5 6.3, 10.2  10.0

Mann—Whitney U-Test.

Table 5 Linear regression models for variables associated with self-reported postpartum depression five days, six weeks and six months after delivery. Variables

Model 1 B

Model 2 B

Adjusted R Leptin IL-6 BMI Age Duration of gestation Newborn gender Smoking

0.004 1.00

0.002

Adjusted R 2 Leptin IL-6 BMI Age Duration of gestation Newborn gender Smoking

0.002 0.93

Adjusted R 2 Leptin IL-6 BMI Age Duration of gestation Newborn gender Smoking

0.009 0.89

2

EPDS 5 days

EPDS 6 weeks

EPDS 6 months

1.04

0.002 1.07

0.004 1.02

Model 3 B

Model 4 B

0.024 0.89 1.05 1.04 *

0.041 0.88 1.07 1.04 * 0.98 0.99 * 0.95 1.05

0.026 0.85 * 1.08 1.01

0.025 0.82 * 1.10 1.03 1.00 0.99 0.95 0.87

0.012 0.85 1.02 1.03

0.008 0.82 * 1.06 1.02 0.99 0.99 0.94 0.80

Leptin in one standard deviation increments, IL-6 in one standard deviation increments, maternal age and BMI at six weeks postpartum, duration of gestation, newborn gender and maternal smoking during pregnancy alternatively introduced into various models. Edinburgh Postnatal Depression scale (EPDS) score used as the outcome variable after logarithmic transformation. B, regression coefficients, presented after anti-logarithmic transformation. * p-value < 0.05.

serum leptin levels at delivery were found to be associated with higher risk for self-reported depression five days ( p = 0.082), six weeks ( p = 0.009) and six months after delivery ( p = 0.026). More specifically, leptin levels higher by one standard deviation among controls, which corresponds to 34.06 ng/mL, were associated with an EPDS score lower by 18% six weeks and at six months after delivery. These results were also replicated using multiple logistic regression analysis, having the mother’s case—control status as the outcome variable (data not shown).

4. Discussion The principal finding of this nested case—control study is the independent negative association between serum leptin levels at delivery and self-reported depression five days, six weeks and six months following delivery. To our knowledge, this is the first time that such an observation is reported in the literature. Earlier studies have assessed leptin levels in patients with major depression with mixed results, which can at least in part be explained by different

1334 depression subtypes, particularly melancholic versus atypical depression and the subsequent differences in HPA-axis function (Gold and Chrousos, 2002). Increases in plasma leptin levels are induced by cortisol in a dose dependent manner in healthy individuals (Newcomer et al., 1998). Moreover, methodological issues such as inclusion criteria, co-morbid disorders, controlling or not for confounding factors such as body-mass index, age, gender, alcohol and smoking may contribute to why results on leptin and major depression are still inconclusive. In the present study, every effort was made in order to statistically control for known confounding factors–—the inverse association found between serum leptin levels at delivery and depressive symptoms even six months after delivery remained nevertheless robust. Even though depression during pregnancy is considered a major risk factor for the development of postpartum depression, biological data suggest that these conditions may in fact represent two different types of depression. Antenatal depression is associated with high cortisol levels, mainly due to placental production of Corticotrophin Releasing Hormone (CRH) during pregnancy (Kammerer et al., 2006), whereas postpartum depression tends to be accompanied by decreased cortisol levels. It is thus hypothesized that antenatal depression may be more of the melancholic type, whereas postpartum depression may be more of the atypical type, triggered by cortisol withdrawal after delivery (Kammerer et al., 2006). Glucocorticoids are known to directly increase leptin synthesis and secretion from adipose tissue (Masuzaki et al., 1997). On the other hand, leptin has been shown in animal models to activate the HPA axis by promoting CRH release (Schwartz et al., 1996). Moreover, delivery is accompanied by substantial fluctuations in concentrations of other hormones such as estrogens and progesterone. Estrogens increase up to 1000-fold during pregnancy and return abruptly to near pre-pregnancy levels within a few hours after delivery (Russell et al., 2001). Thus, a hormonal hypothesis for the etiology of postpartum depression was formulated more than thirty years ago (Nott et al., 1976). However, no consistent differences in estrogen and progesterone levels have been found between women who develop postpartum depression and those who do not (Bloch et al., 2003). It is hypothesized that estrogens and progesterone are involved in the development of postpartum depression in a subgroup of women who are differentially sensitive to the mood-destabilizing effects of such hormones (Bloch et al., 2000). Interestingly, leptin is also involved in the regulation of reproductive function possibly through stimulation of the gonadotropin-releasing hormone (GnRH) and luteinizing hormone (LH) in healthy women (Licinio et al., 1998; Gao and Horvath, 2008). There are few reports on the metabolism of leptin and interleukin-6 during pregnancy and early puerperium. In healthy women, leptin levels are reported to rise during pregnancy (Helland et al., 1998; Lage et al., 1999; Liu et al., 2000), fall after delivery (Lage et al., 1999; Liu et al., 2000) and subsequently increase during the first six months postpartum (Lage et al., 1999). Interleukin-6 is reported to rise during pregnancy (Vassiliadis et al., 1998; Aris et al., 2008), peak during the first day after delivery and subsequently decrease (Maes et al., 2000). The protective effect of higher leptin levels at delivery, which was shown in this study, may thus be explained by

A. Skalkidou et al. either a slower decrease of hormones such as cortisol and estrogens during late pregnancy due to leptin (indirect association) or by direct up-regulation of the HPA axis by leptin, in order to compensate for the withdrawal of glucocorticoids and estrogens (direct action). Unfortunately, our study design did not allow for repeated measurements of leptin and other hormones in the postpartum period which might have helped in the elucidation of the effect’s direction. Another pathway for the protective effect of leptin against postpartum depression could be through the serotonergic system, which is central in the pathophysiology of depression (Schildkraut, 1965). An interaction between leptin and serotonin has been demonstrated in animal models (Leibowitz and Alexander, 1998) and it is suggested that leptin-induced effects in the brain may be mediated by enhancement of serotonergic neurotransmission (Yamada et al., 2003). It could be thus hypothesized higher leptin levels at delivery conferred protection against postpartum depression by induction of serotonergic activity. Lastly, a pathway involving breastfeeding can also be hypothesized. Maternal nutritional state has been known to influence the development of responsiveness to stress in the infant. An animal model of developing rats provides evidence that leptin reduces exposure to glucocorticoids and enhances hippocampal development during a sensitive period of brain development (Oates et al., 2000). Therefore, it would be tempting to speculate that high serum concentration of leptin in infants, associated with maternal leptin levels via breastfeeding, is critical in order to maintain the blunted adrenal glucocorticoid secretion during the neonatal period (Walker et al., 2004). This could in turn affect the infant’s temperament and activity level in the early puerperium, and thereafter the risk for maternal postpartum depression (McGrath et al., 2008). Notably, interleukin-6 was not associated with selfreported depression postpartum despite evidence that major depression is accompanied by increased serum concentrations of interleukin-6. Blood sampling in the present study was carried out during delivery, and interleukin-6 levels were not found to have a predictive value for the development of postpartum depression, but differences in interleukin-6 levels among depressed and non-depressed women may eventually have arisen later, several weeks after delivery. Moreover, leptin and interleukin-6 were simultaneously introduced in our statistical models in order to accommodate for their similar endocrinological actions. Interestingly, a similar study has also failed to demonstrate any statistically significant association between IL-6 serum levels at delivery and development of postpartum depression, despite the fact that the association between postpartum depression and IL-6 levels in the cerebrospinal fluid was statistically significant (Boufidou et al., 2009). Among the strengths of this study is its population-based design, its size and the available information on individual basis on most of the plausible confounding factors for the association between serum leptin and IL-6 levels at delivery and the development self-reported postpartum depression, which were controlled for in the analysis, rendering robust results. A possible limitation of the study could be timing of blood sampling, which, for methodological reasons occurred in conjunction with intravenous catheterization during labour. Leptin levels exhibit a diurnal variation (Sinha

Leptin and interleukin-6 levels in association with postpartum depression et al., 1996) and are affected by time fasting (Ahima, 2000), but no such tendency was observed in our sample, possibly because labour represents an especially stressful situation. Serial measurements postpartum, as well as measurements of other hormones (e.g. estrogens and cortisol) could have promoted a better understanding of the hormonal pathophysiology behind postpartum depression. On the other hand, lack of such measurements cannot compromise the predictive value of serum leptin levels at delivery for the development of postpartum depression which was shown in this study. IL-6 has, in this setting, served as a proxy to cortisol levels. EPDS, a self-administered scale, was used instead of a psychiatric interview for the classification of postpartum depression cases, for methodological reasons. The scale has a high sensitivity but lower specificity (Wickberg and Hwang, 1996), but misclassification of cases and controls could only have led to an attenuation a given association. Moreover, the use of a scale provided information on milder form of depressive symptomatology which would not have been possible with the use of strict psychiatric diagnostic criteria. In conclusion, this nested case—control study demonstrates that lower serum leptin levels at delivery are associated with higher risk for the development of postpartum depressive symptomatology. If these findings are replicated by other studies, leptin levels at delivery could eventually serve as a biological marker for the prediction of postpartum depression. Bearing in mind the several similarities between postpartum depression and atypical depression, these results could contribute in the efforts to unravel the complex pathophysiological mechanisms involved in this condition as well. The findings presented herein may open interesting perspectives for future research.

Role of the funding source The funding sources did not have any involvement in the study design, nor in the collection, the analysis or in the interpretation of the data, nor in the writing of the report or the decision to submit the paper for publication.

Conflict of interest None of the authors have reported any biomedical financial interests or potential conflicts of interest.

Acknowledgments The authors would like to sincerely thank all the mothers who participated in this study, as well as all the employees at the Department of Obstetrics, Uppsala University Hospital, who helped with the distribution of questionnaires and blood sampling. The authors would also like to express their gratitude to Lena Moby, Jenny Juhlin and Tina Sa ¨fstro ¨m for their dedication and hard work for the computerization of all data used in this study, as well as to Nikolaos Dessypris for statistical advice. This study was supported by grants from the Swedish Research Council proj K2008-54X-20642-01-3, the Swedish ˚ ke Wiberg Foundation, the Society of Medicine, the A

1335

So ¨nna BBs Minnes¨derstro ¨m-Ko ¨ningska Foundation, Allma fond, the Alexander S. Onassis Public Benefit Foundation and the Gillbergska Foundation.

References Ahima, R.S., 2000. Leptin and the neuroendocrinology of fasting. Front. Horm. Res. 26, 42—56. American Psychiatric Association, 2000. Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, Text Revision. American Psychiatric Association, Washington, DC. Antonijevic, I.A., Murck, H., Frieboes, R.M., Horn, R., Brabant, G., Steiger, A., 1998. Elevated nocturnal profiles of serum leptin in patients with depression. J. Psychiatr. Res. 32, 403—410. Aris, A., Lambert, F., Bessette, P., Moutquin, J.M., 2008. Maternal circulating interferon-gamma and interleukin-6 as biomarkers of Th1/Th2 immune status throughout pregnancy. J. Obstet. Gynaecol. Res. 34, 7—11. Beck, C.T., 1995. The effects of postpartum depression on maternalinfant interaction: a meta-analysis. Nurs. Res. 44, 298—304. Berk, M., Wadee, A.A., Kuschke, R.H., O’Neill-Kerr, A., 1997. Acute phase proteins in major depression. J. Psychosom. Res. 43, 529— 534. Bloch, M., Schmidt, P.J., Danaceau, M., Murphy, J., Nieman, L., Rubinow, D.R., 2000. Effects of gonadal steroids in women with a history of postpartum depression. Am. J. Psychiatry 157, 924— 930. Bloch, M., Daly, R.C., Rubinow, D.R., 2003. Endocrine factors in the etiology of postpartum depression. Compr. Psychiatry 44, 234— 246. Boufidou, F., Lambrinoudaki, I., Argeitis, J., Zervas, I.M., Pliatsika, P., Leonardou, A.A., Petropoulos, G., Hasiakos, D., Papadias, K., Nikolaou, C., 2009. CSF and plasma cytokines at delivery and postpartum mood disturbances. J. Affect. Disord. 115, 287—292. Chrousos, G.P., Gold, P.W., 1992. The concepts of stress and stress system disorders. Overview of physical and behavioral homeostasis. JAMA 267, 1244—1252. Chrousos, G.P., Kino, T., 2005. Interactive functional specificity of the stress and immune responses: the ying, the yang, and the defense against 2 major classes of bacteria. J. Infect. Dis. 192, 551—555. Corwin, E.J., Pajer, K., 2008. The psychoneuroimmunology of postpartum depression. J. Womens Health 17, 1529—1534. Corwin, E.J., Bozoky, I., Pugh, L.C., Johnston, N., 2003. Interleukin1beta elevation during the postpartum period. Ann. Behav. Med. 25, 41—47. Cox, J.L., Holden, J.M., Sagovsky, R., 1987. Detection of postnatal depression. Development of the 10-item Edinburgh Postnatal Depression Scale. Br. J. Psychiatry 150, 782—786. Deuschle, M., Blum, W.F., Englaro, P., Schweiger, U., Weber, B., Pflaum, C.D., Heuser, I., 1996. Plasma leptin in depressed patients and healthy controls. Horm. Metab. Res. 28, 714—717. Esel, E., Ozsoy, S., Tutus, A., Sofuoglu, S., Kartalci, S., Bayram, F., Kokbudak, Z., Kula, M., 2005. Effects of antidepressant treatment and of gender on serum leptin levels in patients with major depression. Prog. Neuropsychopharmacol. Biol. Psychiatry 29, 565—570. Gao, Q., Horvath, T.L., 2008. Cross-talk between estrogen and leptin signaling in the hypothalamus. Am. J. Physiol. Endocrinol. Metab. 294, E817—E826. Gecici, O., Kuloglu, M., Atmaca, M., Tezcan, A.E., Tunckol, H., Emul, H.M., Ustundag, B., 2005. High serum leptin levels in depressive disorders with atypical features. Psychiatry Clin. Neurosci. 59, 736—738. Gold, P.W., Chrousos, G.P., 2002. Organization of the stress system and its dysregulation in melancholic and atypical depression: high vs low CRH/NE states. Mol. Psychiatry 7, 254—275.

1336 Groer, M.W., Morgan, K., 2007. Immune, health and endocrine characteristics of depressed postpartum mothers. Psychoneuroendocrinology 32, 133—139. Helland, I.B., Reseland, J.E., Saugstad, O.D., Drevon, C.A., 1998. Leptin levels in pregnant women and newborn infants: gender differences and reduction during the neonatal period. Pediatrics 101, E12. Jequier, E., 2002. Leptin signaling, adiposity, and energy balance. Ann. N.Y. Acad. Sci. 967, 379—388. Josefsson, A., Berg, G., Nordin, C., Sydsjo, G., 2001. Prevalence of depressive symptoms in late pregnancy and postpartum. Acta Obstet. Gynecol. Scand. 80, 251—255. Jow, G.M., Yang, T.T., Chen, C.L., 2006. Leptin and cholesterol levels are low in major depressive disorder, but high in schizophrenia. J. Affect. Disord. 90, 21—27. Kammerer, M., Taylor, A., Glover, V., 2006. The HPA axis and perinatal depression: a hypothesis. Arch. Womens Ment. Health 9, 187— 196. Kiecolt-Glaser, J.K., Glaser, R., 2002. Depression and immune function: central pathways to morbidity and mortality. J. Psychosom. Res. 53, 873—876. Kraus, T., Haack, M., Schuld, A., Hinze-Selch, D., Pollmacher, T., 2001. Low leptin levels but normal body mass indices in patients with depression or schizophrenia. Neuroendocrinology 73, 243— 247. Lage, M., Garcia-Mayor, R.V., Tome, M.A., Cordido, F., Valle-Inclan, F., Considine, R.V., Caro, J.F., Dieguez, C., Casanueva, F.F., 1999. Serum leptin levels in women throughout pregnancy and the postpartum period and in women suffering spontaneous abortion. Clin. Endocrinol. (Oxf.) 50, 211—216. Lee, D.T., Yip, A.S., Chan, S.S., Tsui, M.H., Wong, W.S., Chung, T.K., 2003. Postdelivery screening for postpartum depression. Psychosom. Med. 65, 357—361. Leibowitz, S.F., Alexander, J.T., 1998. Hypothalamic serotonin in control of eating behavior, meal size, and body weight. Biol. Psychiatry 44, 851—864. Licinio, J., Negrao, A.B., Mantzoros, C., Kaklamani, V., Wong, M.L., Bongiorno, P.B., Mulla, A., Cearnal, L., Veldhuis, J.D., Flier, J.S., McCann, S.M., Gold, P.W., 1998. Synchronicity of frequently sampled, 24-h concentrations of circulating leptin, luteinizing hormone, and estradiol in healthy women. Proc. Natl. Acad. Sci. USA 95, 2541—2546. Liu, Y., Wang, W., Hitomi, M., 2000. Serum leptin levels in normal pregnant women and their babies. Zhonghua. Fu. Chan. Ke. Za. Zhi. 35, 392—395. Loffreda, S., Yang, S.Q., Lin, H.Z., Karp, C.L., Brengman, M.L., Wang, D.J., Klein, A.S., Bulkley, G.B., Bao, C., Noble, P.W., Lane, M.D., Diehl, A.M., 1998. Leptin regulates proinflammatory immune responses. FASEB J. 12, 57—65. Maes, M., Meltzer, H.Y., Bosmans, E., Bergmans, R., Vandoolaeghe, E., Ranjan, R., Desnyder, R., 1995. Increased plasma concentrations of interleukin-6, soluble interleukin-6, soluble interleukin-2 and transferrin receptor in major depression. J. Affect. Disord. 34, 301—309. Maes, M., Lin, A.H., Ombelet, W., Stevens, K., Kenis, G., De Jongh, R., Cox, J., Bosmans, E., 2000. Immune activation in the early puerperium is related to postpartum anxiety and depressive symptoms. Psychoneuroendocrinology 25, 121—137. Masuzaki, H., Ogawa, Y., Hosoda, K., Miyawaki, T., Hanaoka, I., Hiraoka, J., Yasuno, A., Nishimura, H., Yoshimasa, Y., Nishi, S., Nakao, K., 1997. Glucocorticoid regulation of leptin synthesis and secretion in humans: elevated plasma leptin levels in Cushing’s syndrome. J. Clin. Endocrinol. Metab. 82, 2542—2547. McGrath, J.M., Records, K., Rice, M., 2008. Maternal depression and infant temperament characteristics. Infant. Behav. Dev. 31, 71—80. Murray, L., Cooper, P.J., Wilson, A., Romaniuk, H., 2003. Controlled trial of the short- and long-term effect of psychological treatment

A. Skalkidou et al. of post-partum depression: 2. Impact on the mother-child relationship and child outcome. Br. J. Psychiatry 182, 420—427. Newcomer, J.W., Selke, G., Melson, A.K., Gross, J., Vogler, G.P., Dagogo-Jack, S., 1998. Dose-dependent cortisol-induced increases in plasma leptin concentration in healthy humans. Arch. Gen. Psychiatry 55, 995—1000. Nott, P.N., Franklin, M., Armitage, C., Gelder, M.G., 1976. Hormonal changes and mood in the puerperium. Br. J. Psychiatry 128, 379— 383. Oates, M., Woodside, B., Walker, C.D., 2000. Chronic leptin administration in developing rats reduces stress responsiveness partly through changes in maternal behavior. Horm. Behav. 37, 366— 376. O’Hara, M.W., Swain, A., 1996. Rates and risk of postpartum depression–—a meta-analysis. Int. Rev. Psychiatry 8, 37—54. Papanicolaou, D.A., Wilder, R.L., Manolagas, S.C., Chrousos, G.P., 1998. The pathophysiologic roles of interleukin-6 in human disease. Ann. Intern. Med. 128, 127—137. Pasco, J.A., Jacka, F.N., Williams, L.J., Henry, M.J., Nicholson, G.C., Kotowicz, M.A., Berk, M., 2008. Leptin in depressed women: cross-sectional and longitudinal data from an epidemiologic study. J. Affect. Disord. 107, 221—225. Pravdova, E., Fickova, M., 2006. Alcohol intake modulates hormonal activity of adipose tissue. Endocr. Regul. 40, 91—104. Raison, C.L., Capuron, L., Miller, A.H., 2006. Cytokines sing the blues: inflammation and the pathogenesis of depression. Trends Immunol. 27, 24—31. Rubin, R.T., Rhodes, M.E., Czambel, R.K., 2002. Sexual diergism of baseline plasma leptin and leptin suppression by arginine vasopressin in major depressives and matched controls. Psychiatry Res. 113, 255—268. Russell, J.A., Douglas, A.J., Ingram, C.D., 2001. Brain preparations for maternity—adaptive changes in behavioral and neuroendocrine systems during pregnancy and lactation. An overview. Prog. Brain Res. 133, 1—38. Schildkraut, J.J., 1965. The catecholamine hypothesis of affective disorders: a review of supporting evidence. Am. J. Psychiatry 122, 509—522. Schwartz, M.W., Seeley, R.J., Campfield, L.A., Burn, P., Baskin, D.G., 1996. Identification of targets of leptin action in rat hypothalamus. J. Clin. Invest. 98, 1101—1106. Sharp, D., Hay, D.F., Pawlby, S., Schmucker, G., Allen, H., Kumar, R., 1995. The impact of postnatal depression on boys’ intellectual development. J. Child Psychol. Psychiatry 36, 1315— 1336. Sinha, M.K., Ohannesian, J.P., Heiman, M.L., Kriauciunas, A., Stephens, T.W., Magosin, S., Marco, C., Caro, J.F., 1996. Nocturnal rise of leptin in lean, obese, and non-insulin-dependent diabetes mellitus subjects. J. Clin. Invest. 97, 1344—1347. Sluzewska, A., Rybakowski, J.K., Laciak, M., Mackiewicz, A., Sobieska, M., Wiktorowicz, K., 1995. Interleukin-6 serum levels in depressed patients before and after treatment with fluoxetine. Ann. N.Y. Acad. Sci. 762, 474—476. Vassiliadis, S., Ranella, A., Papadimitriou, L., Makrygiannakis, A., Athanassakis, I., 1998. Serum levels of pro- and anti-inflammatory cytokines in non-pregnant women, during pregnancy, labour and abortion. Mediators Inflamm. 7, 69—72. Walker, C.D., Deschamps, S., Proulx, K., Tu, M., Salzman, C., Woodside, B., Lupien, S., Gallo-Payet, N., Richard, D., 2004. Mother to infant or infant to mother? Reciprocal regulation of responsiveness to stress in rodents and the implications for humans. J. Psychiatry Neurosci. 29, 364—382. Wickberg, B., Hwang, C.P., 1996. The Edinburgh Postnatal Depression Scale: validation on a Swedish community sample. Acta Psychiatr. Scand. 94, 181—184. Yamada, J., Sugimoto, Y., Hirose, H., Kajiwara, Y., 2003. Role of serotonergic mechanisms in leptin-induced suppression of milk intake in mice. Neurosci. Lett. 348, 195—197.

Leptin and interleukin-6 levels in association with postpartum depression Yang, K., Xie, G., Zhang, Z., Wang, C., Li, W., Zhou, W., Tang, Y., 2007. Levels of serum interleukin (IL)-6, IL-1beta, tumour necrosis factor-alpha and leptin and their correlation in depression. Aust. N. Z. J. Psychiatry 41, 266—273.

1337

Zarkovic ´, M., Ignjatovic ´, S., Dajak, M., Ciric ´, J., Beleslin, B., Savic ´, S., Stojkovic ´, M., Bulat, P., Trbojevic ´, B., 2008. Cortisol response to ACTH stimulation correlates with blood interleukin-6 concentration in healthy humans. Eur. J. Endocrinol. 159, 649—652.