Journal of Affective Disorders 298 (2022) 577–589 Contents lists available at ScienceDirect Journal of Affective Disorders journal homepage: www.elsevier.com/locate/jad Personality traits as a risk factor for postpartum depression: A systematic review and meta-analysis Meritxell Puyané a, b, *, Susana Subirà b, Anna Torres c, d, e, Alba Roca c, Lluïsa Garcia-Esteve c, e, Estel Gelabert b a Tecnocampus, Universitat Pompeu Fabra, Research group in Attention to Chronicity and Innovation in Health (GRACIS), Mataró, Spain b Department of Clinical and Health Psychology, Universitat Autònoma de Barcelona, Bellaterra, Spain c Unitat de Salut Mental Perinatal CLINIC-BCN, Hospital Clínic, Barcelona, Spain d Department of Psychiatry and Clinical Psychology, Universitat de Barcelona, Barcelona, Spain e Institut D’Investigacions Biomèdiques August Pi i Sunyer (IDIBAPS), Barcelona, Spain A R T I C L E I N F O A B S T R A C T Keywords: Background: Certain personality traits increase vulnerability to depression, but the evidence linking personality Postpartum depression and postpartum depression (PPD) is less robust. This systematic review aimed to identify personality traits that Personality trait increase the risk of PPD. Risk factor Methods: We systematically reviewed studies retrieved from PubMed/Medline, PsycINFO, Scopus, CINAHL, and Neuroticism Cochrane, following the PRISMA guidelines for reporting. We carried out a meta-analysis on the association between neuroticism and PPD. Results: A total of 34 studies were analyzed. Of these, 31 considered at least one trait associated with PPD; 10 studies considered at least one trait not associated with PPD. The meta-analysis included 13 studies, concluding that neuroticism was associated with PPD (OR: 1.37; 95%CI: 1.22–1.53; p<0.001). Limitations: Study design and approach to personality assessment influence results. Prospective longitudinal studies of persons with no prior history of mood disorder would provide stronger evidence about whether particular personality traits predict PPD. Most studies reviewed used self-report measures to assess personality. Study design and approach to personality assessment influence results, and indications of publication bias were found. Conclusions: Neuroticism is the personality trait most widely studied in relation to PPD. Our meta-analysis found this trait is strongly related with PPD. Moreover, vulnerable personality style and trait anxiety are also associated with PPD. Screening for these traits might help identify women at risk, improving prevention, early detection, and possibly treatment. 1. Introduction women develop depression within a year of giving birth, and depressive symptoms persist beyond the first year in 25% of these (Falana and Women are especially vulnerable to psychiatric disorders during Carrington, 2019). pregnancy and in the postpartum period, and mood disorders are the The vulnerability-stress model explains how different factors can most common maternal psychiatric disorders after childbirth (Stocky affect susceptibility to psychopathologic disorders; predisposing factors and Lynch, 2000). Postpartum mood disorders are often classified into are conceptualized as risk factors or protective factors, depending on three categories, in order of increasing severity: blues, depression, and whether they increase or decrease the likelihood that a stressor will lead psychosis (O’Hara, 1986; Brockington, 2004). Postpartum depression to psychopathology (Grant and McMahon, 2008; Ingram and Luxton, (PPD) is usually defined as an episode of major depressive disorder 2005). (although minor depression is sometimes included) that occurs in the The etiology of PPD is not totally clear. Until recently, PPD was postpartum period (O’Hara and McCabe, 2013). Between 10% to 20% of thought to be mainly due to biological changes after childbirth; * Corresponding author at: School of Health Sciences, c/ Av. d’Ernest Lluch, 32, 08302 Barcelona, Spain. E-mail address: mpuyane@tecnocampus.cat (M. Puyané). https://doi.org/10.1016/j.jad.2021.11.010 Received 19 April 2021; Received in revised form 30 October 2021; Accepted 3 November 2021 Available online 8 November 2021 0165-0327/© 2021 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). M. Puyané et al. Journal of Affective Disorders 298 (2022) 577–589 however, the predominant role of psychosocial factors in increasing the If insufficient information was available to determine whether the risk of developing PPD is now widely recognized. Risk factors for PPD study reported in the article met the inclusion criteria, we attempted to can be divided into three main categories: psychosocial, clinical, and contact the corresponding author and/or publisher; in the absence of personality or temperamental features (Ambrosini et al., 2011). The confirmation, the article was excluded. We also excluded studies whose main risk factors identified in systematic reviews are lower socioeco­ reported effect size cannot be converted to odds ratios (OR). Studies nomic class, life stress during pregnancy, difficult pregnancy, marital without multivariable analysis were excluded. dysfunction, poor social support, and a personal history of psychopa­ thology (O’Hara and Swain, 1996; Robertson et al., 2004). 2.4. Study selection To identify personality traits that increase vulnerability to PPD, re­ searchers must decide which personality trait or traits to examine and We used dedicated systematic review software (Covidence; Mel­ which instrument to use to assess vulnerability (Boyce et al., 2001). bourne, Australia) (Babineau, 2014) to retrieve the references selected Despite this, there is conceptual overlap between clinical depression and in the searches and automatically remove duplicates. Two independent personality traits as measured by the studies. Certain personality traits researchers (MP and EG) examined the consistency of the search and increase vulnerability to depression (Akiskal, 1983; Hirschfeld, 1999), suitable of studies in light of the inclusion and exclusion criteria. All and various of these have been proposed as risk factors for PPD, Title, abstract and full text screening were conducted by both re­ including dependency (Hirschfeld, 1983; Birtchnell, 1984), neuroticism searchers. Disagreement was resolved through discussion, until (Coppen and Metcalfe, 1965, ; Duberstein et al., 2008, ; Lamers et al., consensus was reached. 2012), obsessionality (Allsopp and Williams, 1991), perfectionism The full texts of all articles deemed relevant were downloaded and (Kawamura, 2001; Wei et al., 2014) and interpersonal sensitivity examined to determine whether they met the inclusion criteria and to (Boyce, 1996). The link between personality traits and vulnerability to remove duplicates that might have been missed in the automated depression is well established (Mulder, 2002), but the findings linking screening. personality and PPD are inconsistent and less robust. Determining factors that increase the likelihood of developing PPD 2.5. Data collection process and data items would help identify women at risk, improving efforts at prevention and early detection. This systematic review aimed to synthesize the findings Studies which met all inclusion criteria were reviewed by both re­ in the growing body of literature exploring which personality traits, if searchers (MP and EG), extracting the year of publication, sample size, any, are associated with PPD. population characteristics, procedure, statistical analyses, and main findings from each paper and recording it in an Excel spreadsheet 2. Methods designed by researchers. Discrepancies in data extraction were resolved through discussion. 2.1. Transparency 2.6. Assessment of study quality and risk of bias The inclusion criteria and methods of analysis were specified in advance, and the protocol was registered on PROSPERO (Protocol The Newcastle-Ottawa Quality Assessment Scale (NOS) was used by number: CRD42020157523). We followed the PRISMA guidelines both researchers (MP and EG) to evaluate the methodological quality of (Liberati et al., 2009) for reporting this systematic review. studies based on three aspects: selection of study groups, comparability of study groups, and ascertainment of either the exposure of interest in 2.2. Sources and search strategy case-control studies or outcome of interest in cohort studies (Wells, 2015). Any discrepancies between the authors were resolved through To identify relevant articles, we searched the PubMed/Medline, discussion. We used the NOS to code the risk of bias in the results each PsycINFO, Scopus, CINAHL, and Cochrane databases for articles pub­ study on a 10-point scale (0–9, where scores 0–5 indicate high risk of lished between 1990 and December 2020 for the terms “postpartum bias and scores 6–9 indicate low risk of bias). We excluded studies with depression”, “postnatal depression”, “personality”, “personality traits”, scores <5 (Adejuwon et al., 2018; Denis and Luminet, 2018; Dudley and “women” with Boolean connectors (AND, NOT, OR). We used et al., 2001; Gourounti, 2015 and Vliegen et al., 2006). Table 1 reports different strategies to search within each database; we will be happy to the quality score for each study included in this systematic review. provide the full search strategy for each database upon request. As an example, the Medline search algorithm was ("depression, post­ 2.7. Data analysis partum"[MeSH Terms] OR "postpartum depression" OR "postnatal depression") AND ("personality"[MeSH Terms] OR "personality"). First, we analyzed the studies included in the review qualitatively to The search was completed in December 2020; it was re-run just determine whether sufficient data were available to allow a meta- before the final analyses, and studies not identified in previous searches analysis of the relationship between any personality trait(s) and PPD. were retrieved for inclusion. On finding that sufficient data were available for the trait neuroticism, we used a random effects model to calculate the magnitude of the mean 2.3. Criteria for study selection effect and its 95% confidence interval in the 13 studies evaluating this trait as a risk factor for developing PPD. Importantly, some of these 13 To be included, articles were required to: (a) quantitatively examine studies report various OR because they assessed PPD at more than one the relationship between PPD (up to 12 months after delivery) and at timepoint. We excluded OR corresponding to measurements obtained in least one personality trait (characteristic patterns of thoughts, feelings, the immediate postpartum period (to avoid confusion with postpartum and behaviors that are consistent and stable) using standardized ques­ blues) and included only those corresponding to measurements obtained tionnaires, inventories, or scales with published psychometric proper­ later in the first postpartum trimester. If multiple models were reported, ties. Personality traits have been included according to their conceptual only the final model was included. definition, regardless of the questionnaires used to assess them (b) To combine effect estimates, we performed a random-effects meta- present the results of cross-sectional, cohort, or case-control studies; (c) analysis using the DerSimonian and Laird approach. To estimate the be published in the period comprising January 1990 through December heterogeneity among studies, we calculated the I-squared statistic and 2020; (d) exclude participants with personality disorders and (e) be its p-value. Results of the meta-analysis were plotted in a forest plot. available in English. To explore which features of the studies were statistically associated 578 M. Puyané et al. Journal of Affective Disorders 298 (2022) 577–589 Table 1 Main characteristics of the studies included in the systematic review. Author, Year Country Design N Depression measure/s Assessment timepoint/ Personality Assessment Risk of sDepression measure/s timepoint/s bias Personality Axfors et al., 2017 Sweden Cohort 1618 EPDS ≥12 Pregnancy, 6 weeks and 6 SSP. STAI-trait Pregnancy 7 months PP Besser et al. 2007 Israel Cohort 209 CES-D ≥ 16 Pregnancy and 8 weeks PP DEQ Pregnancy 7 Besser and Priel Israel Cohort 146 CES-D Pregnancy and 8 weeks PP DEQ Pregnancy 7 2003 Boyce et al., 2001 Australia Cohort 717 EPDS >12, SCID-I 2 days PP and 6, 12, 18 and VPSQ 2 days PP 7 24 weeks PP Boyce and Hichey Australia Cohort 425 EPDS >12, SCID-I 2 days PP and 6,12, 18 and VPSQ 2 days PP 7 2005 24 weeks PP Boyce et al., 1991 Australia Cohort 140 EPDS ≥13, BDI ≥11 Pregnancy, 1, 3, 6 month PP EPI, IPSM Pregnancy 7 Chang et al., 2014 Taiwan Cross- 213 CES-D PP MPI PP 8 sectional De Venter et al. Belgium Cohort 187 EPDS ≥13 12 and 24 weeks PP DS-14 Pregnancy 7 2015 Dennis et al., 2004 Canada Cohort 594 EPDS > 9 1 week PP VPSQ 1 week PP 7 Fisher et al., 2002 Australia Cross- 109 EPDS, POMS PP VPSQ PP 7 sectional Gelabert et al., Spain Cohort 309 EPDS ≥10, DIGS 8 and 32 weeks PP VPSQ, EPQ-RS, 2–3 day PP 7 2011 FMPS, TCI Gelabert et al., Spain Case- 237 EPDS ≥10, SCID-I PP FMPS, EPQ-RS PP 8 2012 control Gonidakis et al., Greece Cohort 402 EPDS ≥12, BQ, MADRS First week PP and 1, 3, 6 STAI-Trait. SSPS, 1 day PP 7 2008 months PP MOCI, WI Grant et al., 2008 Australia Cohort 100 EPDS: Antenatal ≥13, Pregnancy and 32 weeks PP STAI-Trait Pregnancy and 32 7 PP ≥10, MINI weeks PP Hipwell et al., 2004 United Cohort 94 EPDS ≥12.5, BDI >12, 7–10 days PP and T3: 2 EPQ Pregnancy 7 Kingdom DSC months PP Iliadis et al., 2015 Sweden Cohort 137 EPDS: Antenatal ≥13, Pregnancy and 6 weeks and 6 SSP Pregnancy and 6 7 PP≥12, DSRS months PP months PP Imsiragic et al. Croatia Cohort 262 EPDS ≥9 3–5 days and 6–9 weeks PP NEO-FFI 3–5 days and 6–9 7 2014 weeks PP Johnstone et al., Australia Cohort 490 EPDS >12 8 weeks PP VPSQS, SCID-II 1 week PP 7 2001 Maia et al., 2012 Portugal Cohort 386 BDI-II, DIGS Pregnancy and 3 months PP MPS Pregnancy 7 Maliszewska et al., Poland Cross- 387 EPDS >12 or item 10+, 4–8 weeks PP NEO-FFI PP 8 2017a sectional PHQ-9 Maliszewska et al., Poland Cross- 101 EPDS ≥13 First week PP NEO-FFI First week PP 8 2016a sectional Maliszewska et al., Poland Cross- 548 EPDS ≥13, PHQ-9 First week and 3 months PP NEO-FFI First week PP 8 2017b sectional Maliszewska et al., Poland Cross- 546 EPDS ≥13, PHQ-9 First week PP NEO-FFI First week PP 8 2016b sectional Marín-Morales Spain Cohort 116 EPDS, SCL-90-R pregnancy and four months NEO-FFI Pregnancy 7 et al., 2014 PP Martín-Santos Spain Cohort 1804 EPDS>9, DIGS 2–3 days, 8 and 32 weeks PP EPQ-RS 2–3 days PP 7 et al., 2012 Masih et al., 2007 Australia Cohort 76 EPDS ≥13, BDI Pregnancy and 8 weeks PP. PSI-II Pregnancy-and 8 7 weeks PP Meltzer-Brody Netherland Cohort 682 EPDS retrospectively 4 years follow up NEO-FFI Baseline 7 et al., 2013 ≥12, CIDI Phillips et al., 2010 Australia Cross- 157 EPDS, SCID-I PP <12 months VPSQS PP <12 months 8 sectional Priel and Besser Israel Cohort 73 CES-D Pregnancy and 8 weeks PP DEQ Pregnancy 7 1999 Saisto et al., 2001 Finland Cohort 211 BDI-II Early pregnancy and late NEO-PI Pregnancy 7 pregnancy and 2–3 months PP Tian et al., 2012 China Case- 4567 EPDS, CIDI PP EPQ. NEO-FFI PP 8 control Van Bussel et al. Belgium Cohort 403 EPDS, HADS-R Pregnancy and 8–12 weeks NEO-FFI Pregnancy 7 2009 and 20–25 weeks PP Verkerk et al., 2005 Netherland Cohort 277 EPDS >11, RDC Pregnancy and 3, 6, and 12 DPQ Pregnancy 7 months PP Vliegen and Luyten Belgium Case- 92 BDI PP DEQ PP 8 2009 control EPDS, Edinburgh Postnatal Depression Scale; BDI, Beck Depression Inventory; DIGS, Diagnostic Interview for Genetic Studies; PHQ-9, Patient Health Questionnaire-9; POMS, Profile of Mood States Questionnaire; BQ, Kennerley’s Blues Questionnaire; MADRS, Montgomery Asberg Depression rating scale; MINI, Mini International Neuropsychiatric Interview; DSRS, Depression Self-. Rating Scale; SCL-90-R, Symptom Checklist-90 Revised CIDI, Composite International Diagnostic Interview; SCID-I, Structured Clinical Interview for DSM Axis I; CES- D, Radloff’s Center for Epidemiological Studies – Depression scale; HADS-R, Hospital Anxiety and Depression Scale; RDC, Research Diagnostic Criteria; PP, postpartum; MPS, Multidimensional Personality Questionnaire; NEO-FFI, NEO Five-Factor Inventory; EPQ, Eysenck Personality Questionnaire; PSI-II, The Personality Style Inventory-Revised; VPSQS, Vulnerable Personality Style. 579 M. Puyané et al. Journal of Affective Disorders 298 (2022) 577–589 Questionnaire; DEQ, Depressive Experiences Questionnaire; NEO-PI, Neo Personality Inventory; DPQ, Dutch Personality Questionnaire; DEQ, The Depressive Expe­ riences Questionnaire; SSP, Karolinska Scales of Personality; STAI-trait, State-Trait Anxiety Inventory; EPI, Eysenck Personality Inventory; IPSM, Interpersonal Sensitivity Measure; MPI, Maudsley Personality Inventory; DS-14, Standard Assessment of Negative Affectivity, Social Inhibition, and Type D Personality; TCI, Temperament and Character Inventory; FMPS, Frost Multidimensional Perfectionism Scale; MOCI, Maudsley Obsessive-Compulsive Inventory; SCID-II; Structured Clinical Interview for DSM Axis II; PP, postpartum. Fig. 1. Flowchart for study selection. 580 M. Puyané et al. Journal of Affective Disorders 298 (2022) 577–589 with the magnitudes of the effect, first we fit a bivariate meta-regression this systematic review. They were conducted in 15 different countries in model assuming mixed effects for each of the characteristics, where the Europe, North America, Asia, and Oceania; most were done in Europe (n outcome variable was the effect estimate (log odds ratio) and the = 10). explanatory variables were characteristics of studies that might influ­ The most common type of study was cohort studies (n = 24), fol­ ence the size of effect. Finally, we fit a multiple meta-regression model lowed by cross-sectional studies (n = 7) and case-control studies (n = 3). assuming mixed effects including the most relevant features that Most studies recruited women at hospitals and maternity clinics. Sample explained some of the variability in the magnitude of the effect observed size ranged from 73 to 4567 subjects. in the bivariate study. These features are: study design (Cohort vs Case- All studies used screening tools to detect PPD, mainly the Edinburgh control), country (Europe Nordic countries vs Europe except Nordic Postnatal Depression Scale (EPDS) (Cox, 1987), used in 27 studies; 8 countries), Timepoint(s) of personality assessment (Pregnancy vs Post­ studies also used clinical interviews to diagnose PPD according to partum), Edinburgh Postnatal Depression Scale cutoff (<12 vs ≥ 12), ICD-10 or DSM-IV criteria. The timepoints at which PPD was assessed diagnostic approach of PPD (Diagnoses vs Symptomatology), timepoint varied widely among studies. Most studies (n = 22) assessed PPD more (s) of depression assessment (1 week after birth vs 6–8 weeks after birth, than once. 1 week after birth vs 24 weeks of 1 year after birth) and type of variable Diverse tools were used to assess personality, the most commonly to assess neuroticism (Categorical variable vs Continuous variable). employed being the Neo Five Factor Inventory (NEO-FFI) (Costa and Mc To analyze small-study effects, we used a funnel plot and the Egger Crae, 1992) (n = 9). Most studies assessed personality only once (n = test, representing the 13 magnitudes of the effect (OR) against its stan­ 30), usually between the second and third trimesters of pregnancy; dard errors. Lastly, we performed a sensitivity analysis, repeating the however, the timepoints at which personality was assessed varied meta-analysis 13 times, as many times as individual studies, excluding widely. The studies analyzed different personality traits, most one study each time. We used Stata statistical software version 16.1 for commonly neuroticism (n = 18) and extraversion (n = 10). all analyses. 3. Results 3.2. Association between personality and postpartum depression The PRISMA flow diagram in Fig. 1 summarizes the study selection Most studies (n = 31) found associations between some personality process. The initial search yielded 3664 articles. After duplicates were traits and PPD, thus concluding that the associated traits should be excluded, the titles and abstracts of 2071 studies were screened, leaving considered risk factors. Some studies (n = 10) found no association 133 to be screened through full text. Of these, 99 were excluded; thus, between certain personality traits and PPD, thus concluding that these the final sample consisted of 34 studies. traits should not be considered risk factors. Table 2 summarizes the 10 studies that found no association between personality traits and PPD. As seen in Table 1, 7 of these were longitu­ 3.1. Characteristics of the studies included in the review dinal (cohort) studies and 3 were cross-sectional studies. All used screening tools; the EPDS was used in 9 studies, with cutoffs for PPD Table 1 presents the main characteristics of the studies included in ranging from 9 to 13. One study (Masih et al., 2007) also used the BDI to Table 2 Results of studies affirming that certain personality trait is not associated with postpartum depression. Author, year Depression assessment timepoint (after Personality trait β p R2 p Adjusted OR (95% p delivery) CI) 1. De Venter et al. (2015) 12 weeks Type D personality – – – – 2.22 (0.66–7.48) ns 24 weeks Type D personality – – – – 2.49 (0.73–8.50) ns 2. Imsiragic et al. (2014) 3–5 days Extraversion – – – – 0.95 (0.88–1.03) ns 3–5 days Agreeableness – – – – 1.04 (0.97–1.12) ns 3–5 days Conscientiousness – – – – 0.97(0.90–1.04) ns 3–5 days Openness to experience – – – – 0.98(0.93–1–04) ns 6–9 weeks Extraversion – – – – 1.05(0.95–1.15) ns 6–9 weeks Agreeableness – – – – 0.94(0.86–1.03) ns 6–9 weeks Conscientiousness – – – – 0.97(0.89–1.06) ns 6–9 weeks Neuroticism – – – – 1.07(0.96–1.20) ns 3. Maia et al. (2012) 12 weeks Perfectionism.- Self-oriented 0.005 ns – – 1.005 (0.980–1.032) ns Perfectionism.-Other’s high 0.016 ns – – 1.016 (0.979–1.054) ns standards Perfectionism.- Conditional 0.003 ns – – 1.003 (0.947–1.063) ns acceptance 4. Maliszewska et al. 4–8 weeks Extroversion – – – – 0.96 (0.75–1.22) ns (2017a) Openness to experience – – – – 0.86 (0.69–1.07) ns Agreeableness – – – – 1.03 (0.84–1.26) ns Conscientiousness – – – – 0.89 (0.72–1.10) ns 5. Maliszewska et al. First week Neuroticism – – – – 1.89 (0.94–3.81) ns (2016a) 6. Marín-Morales et al. 16 weeks Extroversion − 0.104 ns – – – – (2014) Conscientiousness − 0.086 ns – – – – 7. Masih et al. (2007) 8 weeks Autonomy personality 0.20 – 0.321 ns – – 8. Meltzer-Brody et al. 4 years Extraversion – – – ns – – (2013) 9. Phillips et al. (2010) <12 months Vulnerable personality 0.10 ns – – – – 10. Van Bussel et al. (2009) 20–25 weeks Agreeableness − 0.02 ns – – – – 20–25 weeks Conscientiousness 0.17 ns – – – – 20–25 weeks Extroversion − 0.09 ns – – – – 581 M. Puyané et al. Journal of Affective Disorders 298 (2022) 577–589 evaluate symptoms of PPD, and another (Maia et al., 2012) used the BDI personality (Phillips et al., 2010). but not the EPDS. The timepoints for PPD assessment ranged from the Table 3 summarizes the 31 studies that found an association between immediate postpartum to four years after delivery. one or more personality traits and PPD; 22 of these were cohort studies The traits that were not considered risk factors for PPD in these (n = 22). As seen in Table 1, all used screening tools; the EPDS was used studies were extraversion (Imsiragic et al., 2014; Maliszewska et al., in 25. In addition to screening tools, 8 studies (Boyce et al., 2001; Boyce 2017a; Marín-Morales et al., 2014; Meltzer-Brody et al., 2013, and Van and Hichey., 2005; Gelabert et al., 2011, 2012; Grant et al., 2008; Bussel et al., 2009), neuroticism (Imsiragic et al., 2014 and Maliszewska Martín-Santos et al., 2012; Meltzer-Brody et al., 2013; Tian et al., 2012) et al., 2016a) perfectionism (Maia et al., 2012), and vulnerable used clinical interviews to confirm the diagnosis of PPD according to Table 3 Results from studies affirmed that certain personality trait is associated with PPD. Author, year Depression assessment timepoint (after Personality trait β p R2 p Adjusted OR (95% p delivery) CI) 1. Axfords et al. (2017) 6 months Neuroticism – – – – 5.44 (2.77–10.69) <0.001 Anxiety trait – – – – 3.42 (1.34–8.71) <0.001 2. Besser et al. (2007) 8 weeks Dependency − 0.19 <0.05 0.37 <0.001 – – Self-criticism 0.43 <0.001 0.37 <0.001 – – Self-criticism 0.36 <0.001 0.39 <0.01 – – 3. Besser and Priel (2003) 8 weeks Dependency − 0.22 <0.01 0.22 <0.001 – – Self-criticism 0.20 <0.01 0.22 <0.001 – – 4. Boyce et al. (2001) 24 weeks Vulnerable personality – – – – 1.18 (1.10–1.20) <0.001 Organised/responsive – – – – 0.792 (0.66–0.94) <0.001 5. Boyce and Hichey (2005) 24 weeks Vulnerable personality – – – – 2.82 (1.06–7.45) <0.001 Organised/responsive – – – – 3.69 (1.26–10.84) <0.001 6. Boyce et al. (1991) 24 weeks Interpersonal sensitivity 0.29 <0.01 0.24 <0.001 2.90 <0.05 Extraversion − 0.11 <0.01 0.19 <0.001 2.24 <0.05 Neuroticism 0.10 <0.01 0.19 <0.001 0.69 <0.05 7. Chang et al. (2014) 1 year Neuroticism – – – – 1.25 (1.17–1.34) <0.001 8. Dennis et al. (2004) 1 week Vulnerable personality 0.20 – – – 1.21 (1.13–1.31) <0.001 9. Fisher et al. (2002) postpartum Vulnerable personality – – – – 1.2 (1.02–1.39) <0.02 10. Gelabert et al. (2011) 32 weeks Vulnerable personality – – – – 1.16 (1.07–1.26) <0.001 11. Gelabert et al. (2012) postpartum Perfectionism 1.092 <0.05 – – 2.98 (1.23–7.21) <0.05 Neuroticism 1.421 <0.01 – – 4.14 (1.72–9.94) <0.05 12. Goniadakis et al. (2008) First week, 1, 3, 6 months Obsessivity (cleaning) 0.3 <0.05 – – – 13. Grant et al. (2008) 32 weeks Anxiety trait 1.81 <0.05 – – 6.12 (1.37–27.41) <0.05 14. Hipwell et al. (2004) 8 weeks Neuroticism 0.25 <0.05 0.17 <0.001 – Author, year Assessment timepoint (after delivery) Personality trait β p R2 p Adjusted OR (95% p CI) 15. Iliadis et al. (2015) 6 weeks Neuroticism – – – – 5.0 (2.2–11.5) <0.05 24 weeks Neuroticism – – – – 7.9 (3.1–20.0) <0.05 16. Imsiragic et al. (2014) 3–5 days Neuroticism – – – – 1.16 (1.07–1.25) <0.05 6–9 weeks Openess – – – – 0.92 (0.86–0.99) <0.05 17. Johnstone et al. (2001) 8 weeks Vulnerable personality – – – – 7.54 <0.05 18. Maliszewska et al. 4–8 weeks Neuroticism – – – – 1.50 (1.17–1.92) 0.001 (2017a) 19. Maliszewska et al. First week Extraversion – – – – 0.51(0.27–0.95) 0.04 (2016a) 20. Maliszewska et al. First week 3 months Neuroticism – – – – 1.37 (1.05–1.77) 0.02 (2017b) 21. Maliszewska et al. First week Neuroticism – – – – 1.65 (1.32–2.06) <0.001 (2016b) Extraversion – – – – 0.77 (0.62–0.97) <0.05 22. Marín-Morales et al. 16 week s Neuroticism 0.42 – 0.247 <0.001 – – (2014) 23. Martín-Santos et al. 8 weeks Neuroticism 0.047 <0.001 – – 1.05(1.02–1.07) <0.05 (2012) 32 weeks Neuroticism 0.040 0.021 – – 1.04 (1.02–1.08) <0.05 32 weeks Neuroticism 0.048 <0.001 – – 1.05(1.02–1.08) <0.05 24. Masih et al. (2007) 8 weeks Sociotropic personality 0.25 0.42 <0.05 – 25. Meltzer-Brody et al. 4 years follow up Neuroticism – – – - <0.05 (2013) 26. Priel and Besser (1999) 8 weeks Dependency − 0.24 <0.01 0.14 <0.01 – – Self-criticism 0.41 <0.001 0.14 <0.01 – – 27. Saisto et al. (2001) 8–12 weeks Neuroticism – – – – 3.37 (1.04–10.86) <0.05 28. Tian et al. (2012) postpartum Neuroticism – – – – 1.12 (1.09–1.21) <0.05 29. Van Bussel et al. (2009) 12 weeks Neuroticism 0.27 <0.001 0.12 <0.001 – – 20–25 weeks Neuroticism 0.39 <0.001 0.23 <0.001 – – 20–25 weeks Openess 0.28 <0.01 0.23 <0.001 – – 30. Verkerk et al. (2005) 12 weeks Neuroticism, Introversion – – – – 3.08 (1.10–8.63) <0.05 24 weeks Neuroticism, Introversion – – – – 4.64 (1.65–13.16) <0.01 1 year Neuroticism, Introversion – – – – 6.83 (1.97–23.74) <0.01 31. Vliegen and Luyten postpartum Dependency − 0.11 – 0.09 <0.001 – – (2009) Self-Criticism 0.03 – 0.09 <0.001 – – Dependency and Self- 0.22 <0.01 0.05 <0.01 – – Criticism 582 M. Puyané et al. Journal of Affective Disorders 298 (2022) 577–589 ICD-10 or DSM-IV criteria. The sample size was less than 400 women in timing of the assessment of neuroticism 100% (Appendix A). Fitting a 18 studies; Tian et al.’s (2012) case-control study included 4567 women. mixed-effects multiple meta-regression model revealed that the timing The trait most commonly studied as a possible risk factor for PPD was of the assessment of neuroticism explained 100% of the observed vari­ neuroticism, analyzed in 18 (53%) of the studies; 17 (94%) of these ance among the magnitudes of the effect (Appendix B). These results found that neuroticism was a risk factor for PPD. However, the magni­ should be treated with caution since so that values are likely to be tude of the risk (OR) varied widely, ranging from 1.05 (95%CI: inflated R-squared values should be interpreted in light of the small 1.02–1.07) in a convenience cohort (n = 1804) (Martin-Santos et al., sample size. 2012) to 7.90 (95%CI: 3.1–20.0) in a cohort (n = 137) of women Fig. 3 plots the 13 magnitudes of the effect (OR) against their stan­ screened for PPD (EPDS ≥ 12) 24 weeks postpartum (Iliadis et al., 2015). dard errors. The asymmetry in the magnitudes of the effects shown in The vulnerable personality was analyzed in six studies, where ORs these funnel plots was confirmed in the Egger test (p<0.001). Therefore, ranged from 1.16 (95%CI: 1.07–1.26) (Gelabert et al., 2011) to 2.82 publication bias cannot be ruled out, because the tests for small-study (95%CI: 1.06–7.45) (Boyce and Hichey, 2005). effects indicate that smaller studies were more likely to present large effect sizes. In the sensitivity analysis, the average effect size ranged from 1.30 to 3.3. Meta-analysis of the studies that analyzed neuroticism associated 1.51 and the heterogeneity ranged from 84.3% to 89.2%. with PPD 4. Discussion We performed a meta-analysis of the association between neuroti­ cism and PPD (13 studies). We systematically analyzed studies published in the three last de­ Our meta-analysis found this trait was associated with PPD (OR: cades where the relationship between personality traits and PPD was 1.37; 95%CI: 1.22–1.53; p<0.001). However, given the great hetero­ investigated. Our analysis shows that discrepant findings are likely due geneity among the studies (I2 = 88.3%; p <0.001), further analyses were mainly to differences in the timepoint(s) when personality traits and necessary to clarify these results. This heterogeneity was confirmed with PPD were assessed, the approach used to diagnose PPD, and the per­ the forest plot (Fig. 2). sonality trait under study. However, the design of these studies makes it Therefore, given the great variability in the magnitude of the effect impossible for us to determine causality in the relationship between among the studies, we analyzed the characteristics of the studies to try to personality and PPD. Moreover, studies researching the association be­ determine the source of this variability. tween personality and depression often evaluate both constructs Among the characteristics of the studies analyzed, the country where simultaneously. Although many of the cohort studies included here the study was conducted (p = 0.084), the timing of the assessment of considered depressive symptoms as confounding factors, only one of neuroticism (p <0.001), and the type of variable used to assess neurot­ them excluded women with concurrent depression at the time of per­ icism (continuous or categorical) (p = 0.002) in bivariate analyses were sonality assessment. the most important. These variables accounted for much of the vari­ ability in explained variance, with country explaining 37.64% of the variance, the type of variable used to assess neuroticism 84.69%, and the Fig. 2. Forest plot of the meta-analysis of random effects of neuroticism as a personality trait associated with postpartum depression. 583 M. Puyané et al. Journal of Affective Disorders 298 (2022) 577–589 Fig. 3. Funnel plot showing the odds ratios of the 13 studies included in the meta-analysis of the effect of neuroticism as a personality trait associated with postpartum depression. 4.1. Time of assessment are not diagnostic tests, and they might overestimate the prevalence of PPD. The timepoint(s) when PPD and personality were assessed varied widely among studies. Taken together, these results reflect the pro­ gression of depressive symptoms during the postpartum period; how­ 4.3. Personality traits studied ever, the fact that PPD was assessed at different timepoints makes it difficult to compare studies and limits the generalizability of our The current study aimed to systematically review the evidence for findings. personality traits as a risk factor for PPD. We found that a wide variety of In some of these, this assessment took place within days after traits had been explored; however, among the multitude of personality childbirth (immediate postpartum). The results of these studies should traits that could have been analyzed, most studies focused on very few be taken with caution, as depressive symptoms in this period may be traits. All the studies analyzed are vulnerable to selective analysis indistinguishable from postpartum blues (O’Hara and Wisner, 2014). reporting and chance findings. Moreover, most report only significant It is important consider this aspect when interpreting the results. associations, possibly disregarding null associations, so we cannot rule Depression is more common in the third trimester of pregnancy, when it out a potential reporting bias at the study level. Thus, caution is war­ is sometimes associated with delusional and suicidal thoughts. The onset ranted in interpreting the findings of any of the studies included in this of depression during pregnancy triples the risk of postpartum depression review as well as those of the meta-analysis, which, taken together, may (Robertson et al., 2004) represent a selected sample of analyses. In multiple studies, personality traits were assessed only in the Among the traits that have not aroused much interest in relation to postpartum period. Postpartum personality assessments can be distorted PPD are sociotropic-autonomous personality, studied only by Masih by depressive symptoms, whether related to PPD or to depression et al. (2007), and Type D personality, studied only by De Venter et al. beginning prior to childbirth. Episodes of depression and residual effects (2015). Likewise, only one study examined the relationship between can affect personality traits, and depressive states can modify scores for obsessivity and PPD. Gonidakis et al. (2008) concluded that women some personality traits such as neuroticism and extraversion (Griens obsessed with cleaning who feel that their baby cries excessively and et al., 2002). The effects of depression on personality assessments is had experienced stressful life events during pregnancy and prolonged problematic because there can be considerable overlap between per­ maternity blues after delivery ran a higher risk of developing PPD during sonality construct and depression. Many of the studies included in this the first 6 months after delivery. These findings are in line with those review did not take into consideration the fact that personality features reported in Russell’s (2013) systematic review, which showed that can be state-dependent, so their results should be taken with caution. obsessive-compulsive disorder often debuts or exacerbates during pregnancy. Self-criticism and dependency were explored in three studies. Self- 4.2. Diagnostic approach: screening tests with or without clinical criticism refers to an excessive preoccupation with self-definition, con­ interviews trol, and perfection; dependency refers mainly to fears about abandon­ ment and loss. Two early studies carried out in Israel (Priel and Besser, Whereas all studies included in the review assessed PPD through 1999; Besser and Priel, 2003) found that self-criticism assessed during instruments that screen for depressive symptoms, only few of these the third trimester of pregnancy predicted depressive symptoms 8 weeks (Boyce et al., 2001; Boyce and Hichey, 2005; Gelabert et al., 2011, 2012; postpartum in first-time mothers. By contrast, dependency was nega­ Grant et al., 2008; Maia et al., 2012; Meltzer-Brody et al., 2013; and tively associated with symptoms of PPD, suggesting that this personality Phillips et al., 2010) also used clinical interviews (DSM, ICD) to assess dimension might protect against PPD. Moreover, clinically depressed the same subjects, thus increasing the validity of results. Screening tests mothers had higher levels of self-criticism than non-depressed mothers 584 M. Puyané et al. Journal of Affective Disorders 298 (2022) 577–589 (Priel and Besser, 1999; Besser and Priel, 2003). Another study (Vliegen delivery and Imsiragic et al. (2014) found that it decreased the risk of and Luyten, 2009) found that depressed mothers tended to show developing depression 6–9 weeks after delivery. increased levels of dependency, but this trend failed to reach Nine studies in this review (Maliszewska et al., 2016a, 2016b; Ver­ significance. kerk et al., 2005; Boyce et al., 1991) evaluated the relationship between Boyce and Parker (1989) developed the Interpersonal Sensitivity extroversion/introversion and PPD. Extroversion is related to positive Measure (IPSM) as a self-report measure of the construct of interper­ emotions such as happiness, optimism, and enthusiasm as well as to sonal sensitivity, which they defined as “undue and excessive awareness security, activation, and interest for social interaction (Watson and of and sensitivity to, the behavior and feelings of others”. This construct Clark, 1997; Yang and Ha, 2019). This trait is associated with reduced has also been described as a general sensitivity to social feedback, vig­ vulnerability to affective disorders (Kotov, 2010), increased social sup­ ilance with regard to others’ reactions, increased concern about the port (Rhonda et al., 2012), and successful adjustment to childbirth behavior and statements of others, and fear of perceived or actual crit­ stressors (Johnston et al., 2013). Introversion could be associated with icism by others (Boyce et al., 1993). One prospective study included in the onset of emotional disorders and poor adjustment during the peri­ the review (Boyce et al., 1991) found significantly increased risks for natal period (Peñacoba-Puente, 2016). Four studies in this review PPD in women with high interpersonal sensitivity, demonstrating the (Maliszewska et al., 2016a, 2016b; Verkerk et al., 2005; Boyce et al., ability of the IPSM to predict the development of initial depressive ep­ 1991) found that women with low extroversion/high introversion had a isodes, the recurrence of depression six months following childbirth, and higher risk of PPD. By contrast, five studies in this review (Maliszewska non-remission of depressive symptoms among depressed inpatients. et al., 2017a; Marín-Morales et al., 2014; Meltzer-Brody et al., 2013; Much more common in studies about PPD is the vulnerable person­ Imsiragic et al., 2014; Van Bussel et al.,2009) found no relationship ality style. Boyce et al. (2001) developed the Vulnerable Personality between this personality dimension and PPD, corroborating the findings Style Questionnaire (VPSQ), used in numerous prospective studies of of prospective studies that found that extroversion/introversion does not PPD to measure a composite trait including dimensions generally predict vulnerability to depression (Hirschfeld, 1989). considered to be associated with vulnerability, such as anxiety, worry, By far the trait explored in the greatest number of studies included in sensitivity, unassertiveness, volatility, and to a lesser extent obsession­ this review was neuroticism. Originally defined by Eysenck (1967), ality. Several studies (Johnstone et al., 2001; Fisher et al., 2002; Dennis neuroticism is often considered as an endophenotype of the genetic et al., 2004; Boyce and Hichey, 2005; Gelabert et al., 2011) concluded predisposition for affective disorders (Lahey, 2009), as genetic factors that vulnerability was a risk factor for PPD; Boyce and Hichey (2005) that influence individual variation in neuroticism overlap substantially reported that high VPSQ scores were associated with an increased risk of with those that influence the manifestations of other internalizing dis­ major depression in the six months after birth even after controlling for orders (Hettema, 2006). In this review, 17 studies support the conclu­ other variables. sion that neuroticism is associated with PPD (Axfors et al., 2017; Boyce Various studies (Vliegen et al., 2006; Grant et al., 2008; Gourounti, et al., 1991; Chang et al., 2014; Gelabert et al., 2012; Hipwell et al., 2015; Axfors et al., 2017) concluded that the trait anxiety was a risk 2004; Iliadis et al., 2015; Imsiragic et al., 2014; Maliszewska et al., factor for PPD, and this trait contributes significantly to explaining the 2017a, 2017b; Maliszewska et al., 2016b; Marín-Morales et al., 2014; severity of PPD in all the studies in this review. Measures of adult Meltzer-Brody et al., 2013; Saisto et al., 2001; Tian et al., 2012; Van attachment and trait anxiety during late pregnancy might help identify Bussel et al., 2009. Verkerk et al., 2005; Martín-Santos et al., 2012). women at high risk of developing PPD (Axfors et al., 2017). Baseline Neurotic individuals tend to experience intense negative emotions in trait anxiety might also predict postnatal anxiety and mood disorders: response to stressful situations, and pregnancy and childbirth can be Grant et al. (2008) found that women with elevated levels of stressful. Neurotic individuals often perceive themselves as ineffective at self-reported anxiety were significantly more likely to meet diagnostic coping and engage in worrying, rumination, or emotional avoidance. criteria for an anxiety or mood disorder during the seven months Interest in neuroticism has increased in recent years (Barlow, 2014), and following birth. various authors have concluded that neuroticism is among the most Another trait that was examined in various studies is perfectionism. important predictive factors for depression both in pregnancy and in the Perfectionism is a risk factor for many psychopathological disorders postpartum (Dennis and Boyce, 2004; Podolska et al., 2010). Addition­ (Limburg, 2017), and different studies have found associations between ally, neuroticism has been associated with low perceived social support perfectionism and depression (Enns et al., 2005; Huprich, 2008). (Swickert et al., 2010) and poor adjustment to childbirth stressors Although one of the studies included in this review (Maia et al., 2012) (Johnston et al., 2013). concluded that perfectionism was not a risk factor for PPD, other Nevertheless, not all studies fully support the hypothesis that confirmed that this trait is a risk factor for PPD (Gelabert et al., 2012). neuroticism might predict PPD. Two studies in this review (Imsiragic Using one-dimensional assessment measures, Egan et al. (2011) et al., 2014; Maliszewska et al., 2016a) concluded that neuroticism is not demonstrated that perfectionism was a predictor of postpartum a risk factor for PPD. However, Maliszewska et al. (2016a) studied depressive symptoms. Using Hewitt and Flett’s construct of multidi­ depressive symptoms in the first week after childbirth, possibly con­ mensional perfectionism (HMPS, 1991), Hewitt and Flett (1991) found founding concepts of postpartum blues and PPD. In stark contrast to that self-oriented perfectionism and socially-prescribed perfectionism these findings, Imsiragic et al.’s (2014) found that neuroticism predicted were associated with depressive symptoms during pregnancy, but only EPDS score at 3–5 days postpartum (corresponding to the postpartum the socially prescribed perfectionism dimension predicted depressive blues that was not associated with neuroticism in Maliszewska et al. postpartum symptoms. This result highlight the importance of inter­ (2016a)), but did not predict scores at 6–9 weeks postpartum. personal dimension and extrinsic motivation in maladaptative In recent years, the growing interest in PPD has led to initiatives that perfectionism. aim to go beyond merely identifying PPD after it develops to identifying Four of the studies in this review (Maliszewska et al., 2017a; Mar­ women at risk of developing depression during pregnancy or after giving ín-Morales et al., 2014; Imsiragic et al., 2014 and Van Bussel et al., birth (Gelabert et al., 2021). To this end, the International Marcé Society 2009) evaluated McCrae and Costa’s (1999) “Big Five” personality traits for Perinatal Mental Health’s 2013 position statement on the psycho­ (openness to experience, conscientiousness, agreeableness, extraver­ logical evaluation and detection of depression during the perinatal sion, and emotional stability) as risk factors for PPD. All concluded that period (Austin, 2014) recommends assessing personality traits, among agreeableness and conscientiousness were not risk factors for PPD, but other risk factors. their findings on openness to experience differed. Whereas Maliszewska et al. (2017a) found that this trait was not related PPD, Van Bussel et al. (2009) concluded that it increased the risk of depression 25 weeks after 585 M. Puyané et al. Journal of Affective Disorders 298 (2022) 577–589 4.4. Meta-analysis Our meta-analysis focused only on neuroticism, but it would be interesting to conduct meta-analyses on other traits when data become We performed a meta-analysis of the association between neuroti­ available. cism and PPD. We found wide heterogeneity among the 13 studies included in the meta-analysis. Our findings from the meta-analysis seem 5. Conclusion to confirm those of our qualitative analysis included in the review. The relationship between personality traits and PPD is complex. Our 4.5. Strengths and limitations findings support the hypothesis that certain personality traits are asso­ ciated with increased vulnerability to PPD; among these, the evidence is To our knowledge, this is the first systematic review to analyze strongest for neuroticism. The evidence suggests that vulnerable per­ personality traits as possible risk factors for PPD and the first meta- sonality style and trait anxiety are also associated with PPD. Further analysis to measure the effect of neuroticism as risk factor for PPD. studies with designs that can achieve more robust results (e.g., pro­ Search terms were carefully selected to capture all possible studies in the spective longitudinal studies of women without a prior history of mood five indexed databases. The participation of two reviewers helped disorders diagnosed by clinical interviews as well as screening tools) ensure the validity of our results. should aim to determine the best timepoint to assess personality traits However, several limitations should be considered. Not uncommon (especially neuroticism, vulnerable personality style, and trait anxiety) for evidence syntheses, the body of studies included here likely suffers and PPD as well as to understand the mechanisms through which these from publication bias. traits increase the likelihood of PPD. To mitigate potential selective To guard against this bias, we tried to ensure that our literature analysis reporting, authors are encouraged to prospectively register search was as comprehensive as possible. Nevertheless, only studies their study protocol and statistical analysis plan before collecting the published in English were eligible for inclusion, although the articles data. included reported studies done in a wide range of countries, including To ensure early detection and optimal treatment, we need to un­ those where other languages are dominant. Moreover, we used funnel derstand PPD better. plots and the Egger test and found that small studies were more likely to report larger effects, which may be related to publication bias. Contributors There is also potential reporting bias at the study level. Most (71%) of the included studies report only significant associations, and we Author MP: conceptualized the study and took primary responsibility cannot know whether the authors disregarded negative results. for preparing the manuscrit and carry out all the aspects of the sys­ Furthermore, the design of the studies included in the review precludes tematic review. Author EG: assisted with the design of the study, the possibility of discussing a causal relationship between personality reviewing potential studies to include, assessing the quality of studies and PPD. included, and manuscript preparation. Author SS: provided conceptual Our analyses were limited by differences in the methods used in input for the study design and contributed to the preparation of the different studies. Whereas some studies used clinical interviews to di­ manuscript. Authors AT, AR and LG: Supervision and review. agnose depression, most used only screening tools relying on self- reporting to detect depressive symptoms; furthermore, different Role of the funding source studies also used different cutoff scores of the screening tools to diagnose PPD, thus making it difficult to compare their results. This work was supported partially by Fundació TecnoCampus Self-reported personality tests can be complicated by subjects’ moods, limited insight, and response styles; additionally, it can be Author declaration statement difficult for these tests to distinguish traits from the effects of stable environmental contexts (Chmielewski and Watson, 2009). Additionally, We confirm that the manuscript has been read and approved by all temperament and personality are not a fixed, static set of characteristics, named authors and that there are no other persons who satisfied the but rather are dynamic constructs that develop over the lifespan and criteria for authorship but are not listed. We further confirm that the change in response to maturation and life circumstances. For example, order of authors listed in the manuscript has been approved by all of us. life stressors and major shifts in social roles and relationships can We confirm that we have given due consideration to the protection of contribute to personality change (Fraley and Roberts 2005, Kandler intellectual property associated with this work and that there are no et al. 2010). The differences in the timepoints when personality traits impediments to publication. and PPD were assessed also make it difficult to compare the results of the different studies. Additionally, because the studies that used multivar­ Declaration of Competing Interest iate analyses adjusted for different variables, their results might not be comparable. We declare no conflicts of interest. We introduced an analysis as a mean to investigate statistical het­ erogeneity in a meta-analysis: meta-regression analysis. Meta-regression Acknowledgments has low power to detect statistically significant associations when there is a small number of studies, so it should be viewed as exploratory. We acknowledge the support of the Secretaria d’Universitats i Finally, although a wide variety of traits were explored in the studies Recerca del Departament d’Economia i Coneixement, Generalitat de included, most studies focused on only one or two traits. Catalunya (SGR2014/1114; SGR2017/1798). Appendix A Results of mixed effects bivariate meta-regression models of characteristics of studies. 586 M. Puyané et al. Journal of Affective Disorders 298 (2022) 577–589 Characteristics of studies eβ (95% confidence interval)* se p R2 Design Cohort (Reference) - 0.84 0.211 7.61% Case-control 1.84 (0.67, 5.04) Country Nordic countries (Europe) (Reference) - 0.94 0.084 37.64% Europe except Nordic countries 2.23 (0.88, 5.66) Timepoint(s) of personality assessment Pregnancy (Reference) - 0.06 <0.001 100% Postpartum 0.25 (0.15, 0.41) EPDS cutoff <12 (Reference) - 0.78 0.686 4.24% ≥ 12 1.29 (0.32, 5.18) Diagnostic approach Diagnoses (Reference) - 0.74 0.417 3.8% Symptomatology 1.51 (0.51, 4.47) Timepoint(s) of depression assessment 1 week after birth (Reference) - 0.74 0.701 27.22% 6–8 weeks after birth 1.26 (0.34, 4.65) 0.93 0.573 24 weeks of 1 year after birth 1.45 (0.35, 6.03) Type of variable to assess neuroticism Categorical variable (Reference) - 0.09 0.002 84.69% Continuous variable 0.35 (0.19, 0.63) eβ: exp (β coefficient); se: standard error; p: probability; R2: proportion of variance explained. * eβ below 1= smaller effect size; eβ above 1= higher effect size. EPDS: Edinburgh Postnatal Depression Scale Cohort studies (n = 6); Case control studies (n = 2); Nordic countries (n = 4); Europe except Nordic countries (n = 7); Pregnancy personality assessment (n = 4); Postpartum personality assessment (n = 9); Diagnoses PPD approach (n = 3); Symptomatology PPD approach (n = 10); 1 week after birth depression assessment (n = 3); 6–8 weeks after birth depression assessment (n = 4); 24 weeks after birth depression assessment (n = 5); Categorical variable (n = 5); Continuous variable (n = 8). Appendix B Results of mixed effects multiple meta-regression model of characteristics of studies. Characteristics of studies eβ (95% confidence interval)* se p R2 Country Nordic countries (Europe) (Reference) - 0.36 0.285 Europe except Nordic countries 1.36 (0.74, 2.49) Timepoint(s) of personality assessment Pregnancy (Reference) - 0.15 0.036 100% Postpartum 0.39 (0.16, 0.92) Type of variable to assess neuroticism Categorical variable (Reference) - 0.32 0.573 Continuous variable 0.79 (0.32, 1.97) eβ: exp (β coefficient); se: standard error; p: probability; R2: proportion of variance explained. * eβ below 1= smaller effect size; eβ above 1= higher effect size. EPDS: Edinburgh Postnatal Depression Scale. References Boyce, P., Hickey, A., Gilchrist, J., Talley, N.J., 2001. The development of a brief personality scale to measure vulnerability to postnatal depression. Arch. Womens. Ment. Health 3, 147–153. https://doi.org/10.1007/s007370170012. Adejuwon, G.A., Adekunle, I.F., Ojeniran, M., 2018. Social support and personality traits Boyce, P., Parker, G., Barnett, B., Cooney, M., Smith, F., 1991. Personality as a as predictors of psychological wellbeing of postpartum nursing mothers in Oyo state, vulnerability factor to depression. Br. J. Psychiatry 159, 106–114. https://doi.org/ Nigeria. Int. J. Caring Sci. 11, 704. 10.1192/bjp.159.1.106. Akiskal, H.S., 1983. The Relationship of Personality to Affective Disorders: a Critical Boyce, P., Parker, G., 1989. Development of a scale to measure interpersonal sensitivity. Review. Arch. Gen. Psychiatry 40, 801. Aust. N. Z. J. Psychiatry 23, 341. Allsopp, M., Williams, T., 1991. Self-report measures of obsessionality, depression and Boyce, P., Hickey, A., 2005. Psychosocial risk factors to major depression after childbirth. social anxiety in a school population of adolescents. J. Adolesc. 14, 149–156. Soc. Psychiatry Psychiatr. Epidemiol. 40, 605–612. https://doi.org/10.1007/ https://doi.org/10.1016/0140-1971(91)90027-O. s00127-005-0931-0. Axfors, C., Sylvén, S., Ramklint, M., Skalkidou, A., 2017. Adult attachment’s unique Boyce, P., Hickie, I., Parker, G., Mitchell, P., Wilhelm, K., Brodaty, H., 1993. Specificity contribution in the prediction of postpartum depressive symptoms, beyond of interpersonal sensitivity to non-melancholic depression. J. Affect. Disord. 27, personality traits. J. Affect. Disord. 222, 177–184. https://doi.org/10.1016/j. 101–105. https://doi.org/10.1016/0165-0327(93)90082-U. jad.2017.07.005. Brockington, I., 2004. Postpartum psychiatric disorders. Lancet 363 (9405), 303–310. Barlow, D.H., 2014. The Origins of Neuroticism. Perspect. Psychol. Sci. 9, 481. https://doi.org/10.1016/S0140-6736(03)15390-1, 2004 Jan 24PMID: 14751705. Besser, A., Priel, B., 2003. Trait vulnerability and coping strategies in the transition to Chang, H.P., Chen, J.Y., Huang, Y.H., Tyan, J.Y., Yeh, C.J., Su, P.H., Chin-Hung Chen, V., motherhood. Curr. Psychol. 22, 57–72. https://doi.org/10.1007/s12144-003-1013- 2014. Prevalence and factors associated with depressive symptoms in mothers with 7. infants or toddlers. Pediatr. Neonatol. 55, 470–479. https://doi.org/10.1016/j. Birtchnell, J., 1984. Dependence and its relationship to depression. Br. J. Med. Psychol. pedneo.2013.12.009. 57, 215. Chmielewski, M., Watson, D., 2009. What is being assessed and why it matters: the Boyce, P., 1996. An overview of depression-prone personality traits and the role of impact of transient error on trait research. J. Pers. Soc. Psychol. 97 (1), 186–202. interpersonal sensitivity. Aust. N. Z. J. Psychiatry 30, 90. https://doi.org/10.1037/a0015618. 587 M. Puyané et al. Journal of Affective Disorders 298 (2022) 577–589 Coppen, A., Metcalfe, M., 1965. Effect of a Depressive Illness on M.P.I. Scores. Br. J. Limburg, K., 2017. The Relationship Between Perfectionism and Psychopathology: a Psychiatry 111, 236–239. https://doi.org/10.1192/bjp.111.472.236. Meta-Analysis. J. Clin. Psychol. 73, 1301. Cox, J.L., 1987. Detection of Postnatal Depression: development of the 10-item Maia, B.R., Pereira, A.T., Marques, M., Bos, S., Soares, M.J., Valente, J., Gomes, A.A., Edinburgh Postnatal Depression scale. Br. J. Psychiatry 150, 782. Azevedo, M.H., Macedo, A., 2012. The role of perfectionism in postpartum Denis, A., Luminet, O., 2018. Cognitive factors and post-partum depression: what is the depression and symptomatology. Arch. Womens. Ment. Health 15, 459–468. https:// influence of general personality traits, rumination, maternal self-esteem, and doi.org/10.1007/s00737-012-0310-2. alexithymia? Clin. Psychol. Psychother. 25, 359–367. https://doi.org/10.1002/ Maliszewska, K., Bidzan, M., Świątkowska-Freund, M., Krzysztof, P., 2017b. Screening cpp.2168. for maternal postpartum depression, and associations with personality traits and Dennis, C.L.E., Janssen, P.A., Singer, J., 2004. Identifying women at-risk for postpartum social support. A Polish follow-up study 4 weeks and 3 months after delivery. depression in the immediate postpartum period. Acta Psychiatr. Scand. 110, Psychiatr. Pol. 51, 889–898. https://doi.org/10.12740/pp/68628. 338–346. https://doi.org/10.1111/j.1600-0447.2004.00337.x. Maliszewska, K., Bidzan, M., Światkowska-Freund, M., Preis, K., 2016b. Personality type, Duberstein, P.R., Pálsson, S.P., Waern, M., Skoog, I., 2008. Personality and risk for social support and other correlates of risk for affective disorders in early puerperium. depression in a birth cohort of 70-year-olds followed for 15 years. Psychol. Med. 38, Ginekol. Pol. 87, 814–819. https://doi.org/10.5603/GP.2016.0094. 663–671. https://doi.org/10.1017/S0033291707002620. Maliszewska, K., Bidzan, M., Światkowska-Freund, M., Preis, K., 2017a. Medical and Dudley, M., Roy, K., Kelk, N., Bernard, D., 2001. Psychological correlates of depression in psychosocial determinants of risk of postpartum depression: a cross-sectional study. fathers and mothers in the first postnatal year. J. Reprod. Infant Psychol. 19, Acta Neuropsychiatr. 29, 347–355. https://doi.org/10.1017/neu.2017.4. 187–202. https://doi.org/10.1080/02646830124397. Maliszewska, K., Światkowska-Freund, M., Bidzan, M., Preis, K., 2016a. Relationship, Egan, S.J., Wade, T.D., Shafran, R., 2011. Perfectionism as a transdiagnostic process: a social support, and personality as psychosocial determinants of the risk for clinical review. Clin. Psychol. Rev. https://doi.org/10.1016/j.cpr.2010.04.009. postpartum blues. Ginekol. Pol. 87, 442–447. https://doi.org/10.5603/ Enns, M.W., Cox, B., Clara, I., 2005. Perfectionism and Neuroticism: a Longitudinal Study GP.2016.0023. of Specific Vulnerability and Diathesis-Stress Models. Cognit. Ther. Res. 29, 463. Marín-Morales, D., Carmona-Monge, F.J., Peñacoba-Puente, C., 2014. Personality, Falana, S.D., Carrington, J.M., 2019. Postpartum Depression: are You Listening? Nurs. depressive symptoms during pregnancy and their influence on postnatal depression Clin. North Am. 54, 561–567. https://doi.org/10.1016/j.cnur.2019.07.006. in Spanish pregnant Spanish women. [Personalidad, síntomas depresivos en el Fisher, J.R.W., Feekery, C.J., Rowe-Murray, H.J., 2002. Nature, severity and correlates of embarazo y su influencia en la depresión postparto en gestantes españolas]. An. psychological distress in women admitted to a private mother-baby unit. J. Paediatr. Psicol. 30, 908–915. https://doi.org/10.6018/analesps.30.3.153101. Child Health 38, 140–145. https://doi.org/10.1046/j.1440-1754.2002.00723.x. Martín-Santos, R., Gelabert, E., Subirà, S., Gutierrez-Zotes, A., Langorh, K., Jover, M., Gelabert, E., Subirà, S., García-Esteve, L., Navarro, P., Plaza, A., Cuyàs, E., Navinés, R., Torrens, M., Guillamat, R., Mayoral, F., Canellas, F., Iborra, J.L., Gratacos, M., Gratacòs, M., Valdés, M., Martín-Santos, R., 2012. Perfectionism dimensions in Costas, J., Gornemann, I., Navinés, R., Guitart, M., Roca, M., De Frutos, R., major postpartum depression. J. Affect. Disord. 136, 17–25. https://doi.org/ Vilella, E., Valdé, M., García Esteve, L.G., Sanjuan, J., 2012. Research Letter: is 10.1016/j.jad.2011.08.030. neuroticism a risk factor for postpartum depression? Psychol. Med. 42, 1559–1565. Gelabert, E., Subirà, S., Plaza, A., Torres, A., Navarro, P., Ímaz, M.L., Valdés, M., García- https://doi.org/10.1017/S0033291712000712. Esteve, L., Martín-Santos, R., 2011. The Vulnerable Personality Style Questionnaire: Masih, S., Spence, S.H., Oei, T.P.S., 2007. Sociotropic and autonomous personality and psychometric properties in Spanish postpartum women. Arch. Womens. Ment. stressful life events as predictors of depressive symptoms in the postpartum period. Health 14, 115–124. https://doi.org/10.1007/s00737-010-0186-y. Cognit. Ther. Res. 31, 483–502. https://doi.org/10.1007/s10608-006-9071-4. Gelabert, E., Subirà. S., Navinés R., 2021. Personality and perinatal depression. In M. Meltzer-Brody, S., Boschloo, L., Jones, I., Sullivan, P.F., Penninx, B.W., 2013. The EPDS- Serati (Ed). Perinatal depression. Milan, Italy: Nova Science Publishers. Lifetime: assessment of lifetime prevalence and risk factors for perinatal depression Gonidakis, F., Rabavilas, A.D., Varsou, E., Kreatsas, G., Christodoulou, G.N., 2008. A 6- in a large cohort of depressed women. Arch. Womens. Ment. Health 16, 465–473. month study of postpartum depression and related factors in Athens Greece. Compr. https://doi.org/10.1007/s00737-013-0372-9. Psychiatry 49, 275–282. https://doi.org/10.1016/j.comppsych.2007.05.018. Mulder, R.T., 2002. Reviews and Overviews Personality Pathology and Treatment Gourounti, K., 2015. Psychosocial risk factors of depression in pregnancy: a survey study. Outcome in Major Depression: a Review. Psychiatry Interpers. Biol. Process. Heal. Sci. J. 9, 1–6. 359–371. Grant, K.A., McMahon, C., Austin, M.P., 2008. Maternal anxiety during the transition to O’Hara, M.W., 1986. Social support, life events, and depression during pregnancy and parenthood: a prospective study. J. Affect. Disord. 108, 101–111. https://doi.org/ the puerperium. Arch. Gen. Psychiatry 43, 569. 10.1016/j.jad.2007.10.002. O’Hara, M.W., Swain, A., 1996. Rates and risk of postpartum depression–a meta-analysis. Griens, A.M.G.F., Jonker, K., Spinhoven, P., Blom, M.B.J., 2002. The influence of Int. Rev. Psychiatry 8, 37. depressive state features on trait measurement. J. Affect. Disord. 70, 95–99. https:// O’Hara, M.W., Mc Cabe, J.E., 2013. Postpartum depression: current status and future doi.org/10.1016/S0165-0327(00)00371-2. directions. Annu. Rev. Clin. Psychol. 9, 379–407. https://doi.org/10.1146/annurev- Hettema, J.M., 2006. Association between glutamic acid decarboxylase genes and clinpsy-050212-185612. anxiety disorders, major depression, and neuroticism. Mol. Psychiatry 11, 752. O’Hara, M.W., Wisner, K.L., 2014. Perinatal mental illness: definition, description and Hipwell, A.E., Reynolds, S., Crick, E.P., 2004. Cognitive vulnerability to postnatal aetiology. Best Pract. Res. Clin. Obstet. Gynaecol. 28, 3–12. https://doi.org/ depressive symptomatology. J. Reprod. Infant Psychol. 22, 211–227. https://doi. 10.1016/j.bpobgyn.2013.09.002. org/10.1080/02646830410001723797. Peñacoba-Puente, C., 2016. Post-Partum Depression, Personality, and Cognitive- Hirschfeld, R.M., 1989. Premorbid personality assessments of first onset of major Emotional Factors: a Longitudinal Study on Spanish Pregnant Women. Health Care depression. Arch. Gen. Psychiatry 46, 345. Women Int 37, 97. Hirschfeld, R.M., 1983. Personality and depression. Empirical findings. Arch. Gen. Phillips, J., Sharpe, L., Matthey, S., Charles, M., 2010. Subtypes of postnatal depression? Psychiatry 40, 993. A comparison of women with recurrent and de novo postnatal depression. J. Affect. Hirschfeld, R.M., 1999. Personality disorders and depression: comorbidity. Depress. Disord. 120, 67–75. https://doi.org/10.1016/j.jad.2009.04.011. anxiety. 10, 142–146. Podolska, M.Z., Majkowicz, M., Bidzan, M., Pankiewicz, P., Sipak-Szmigiel, O., Huprich, S.K., 2008. Depressive personality disorder, dysthymia, and their relationship Podolski, J., 2010. Increased neuroticism aggravates the risk of depressiye symptoms to perfectionism. Depress 25, 207. pregnant women. Stud. Psychol. (Bratisl) 10, 53–66. Iliadis, S.I., Koulouris, P., Gingnell, M., Sylvén, S.M., Sundström-Poromaa, I., Ekselius, L., Priel, B., Besser, A., 1999. Vulnerability to postpartum depressive symptomatology: Papadopoulos, F.C., Skalkidou, A., 2015. Personality and risk for postpartum dependency, self-criticism and the moderating role of antenatal attachment. J. Soc. depressive symptoms. Arch. Womens. Ment. Health 18, 539–546. https://doi.org/ Clin. Psychol. 18, 240–253. https://doi.org/10.1521/jscp.1999.18.2.240. 10.1007/s00737-014-0478-8. Rhonda, M., Swickert, R.J., Hittner, J.B., Foster, A., 2012. A Proposed Mediated Path Imširagić, A.S., Begić, D., Vuković, I.S., Šimićević, L., Javorina, T., 2014. Multivariate between Gender and Posttraumatic Growth: the Roles of Empathy and Social analysis of predictors of depression symptomatology after childbirth. Psychiatr. Support in a Mixed-Age Sample. PSYCH_2012123116055437.pdf 3, 1142–1147. Danub. 26, 416–421. Russell, E.J., 2013. Risk of Obsessive-Compulsive Disorder in Pregnant and Postpartum Johnston, R.G., OstMed, D.O., Brown, A.E., 2013. Maternal trait personality and Women: a Meta-Analysis. J. Clin. Psychiatry 74, 377. childbirth: the role of extraversion and neuroticism. Midwifery 29, 1244–1250. Saisto, T., Salmela-Aro, K., Nurmi, J.E., Halmesmäki, E., 2001. Psychosocial predictors of https://doi.org/10.1016/j.midw.2012.08.005. disappointment with delivery and puerperal depression: a longitudinal study. Acta Johnstone, S.J., Boyce, P.M., Hickey, A.R., Morris-Yates, A.D., Harris, M.G., 2001. Obstet. Gynecol. Scand. 80, 39–45. https://doi.org/10.1034/j.1600- Obstetric risk factors for postnatal depression in urban and rural community 0412.2001.800108.x. samples. Aust. N. Z. J. Psychiatry 35, 69–74. https://doi.org/10.1046/j.1440- Stocky, A., Lynch, J., 2000. Acute psychiatric disturbance in pregnancy and the 1614.2001.00862.x. puerperium. Bailliere’s. Best Pract. Res. Clin. Obstet. Gynaecol. 14, 73–87. https:// Kawamura, K.Y., 2001. Perfectionism, Anxiety, and Depression: are the Relationships doi.org/10.1053/beog.1999.0064. Independent? Cogn. Ther. 25, 291. Tian, T., Li, Yihan, Xie, D., Shen, Y., Ren, J., Wu, W., Guan, C., Zhang, Z., Zhang, D., Kotov, R., 2010. Linking. Psychol. Bull. 136, 768. Gao, C., Zhang, X., Wu, J., Deng, H., Wang, G., Zhang, Y., Shao, Y., Rong, H., Gan, Z., Lahey, B.B., 2009. Public Health Significance of Neuroticism. Am. Psychol. 64, 241. Sun, Y., Hu, B., Pan, J., Li, Yi, Sun, S., Song, L., Fan, X., Zhao, X., Yang, B., Lv, L., Lamers, F., Rhebergen, D., Merikangas, K.R., De Jonge, P., Beekman, A.T.F., Penninx, B. Chen, Yunchun, Wang, X., Ning, Y., Shi, S., Chen, Yiping, Kendler, K.S., Flint, J., W.J.H., 2012. Stability and transitions of depressive subtypes over a 2-year follow- Tian, H., 2012. Clinical features and risk factors for post-partum depression in a large up. Psychol. Med. 42, 2083–2093. https://doi.org/10.1017/S0033291712000141. cohort of Chinese women with recurrent major depressive disorder. J. Affect. Disord. Liberati, A., Altman, D.G., Tetzlaff, J., Mulrow, C., Gøtzsche, P.C., Ioannidis, J.P.A., 136, 983–987. https://doi.org/10.1016/j.jad.2011.06.047. Clarke, M., Devereaux, P.J., Kleijnen, J., Moher, D., 2009. The PRISMA statement for Van Bussel, J.C.H., Spitz, B., Demyttenaere, K., 2009. Depressive symptomatology in reporting systematic reviews and meta-analyses of studies that evaluate health care pregnant and postpartum women. An exploratory study of the role of maternal interventions: explanation and elaboration. J. Clin. Epidemiol. Pergamon e1–e34. antenatal orientations. Arch. Womens. Ment. Health 12, 155–166. https://doi.org/ https://doi.org/10.1016/j.jclinepi.2009.06.006. 10.1007/s00737-009-0061-x. 588 M. Puyané et al. Journal of Affective Disorders 298 (2022) 577–589 Verkerk, G.J.M., Denollet, J., Van Heck, G.L., Van Son, M.J.M., Pop, V.J.M., 2005. Watson, D., Clark, L.A., 1997. Extraversion and Its Positive Emotional Core, in: handbook Personality factors as determinants of depression in postpartum women: a of Personality Psychology. Elsevier 767–793. https://doi.org/10.1016/b978- prospective 1-year follow-up study. Psychosom. Med. 67, 632–637. https://doi.org/ 012134645-4/50030-5. 10.1097/01.psy.0000170832.14718.98. Wei, Y., Ullah, G., Schiff, S.J., 2014. Unification of neuronal spikes, seizures, and Vliegen, N., Luyten, P., 2009. Dependency and self-criticism in post-partum depression spreading depression. J. Neurosci. 34, 11733–11743. https://doi.org/10.1523/ and anxiety: a case control study. Clin. Psychol. Psychother. 16, 22–32. https://doi. JNEUROSCI.0516-14.2014. org/10.1002/cpp.597. Yang, S.K., Ha, Y., 2019. Predicting posttraumatic growth among firefighters: the role of Vliegen, N., Luyten, P., Meurs, P., Cluckers, G., 2006. Adaptive and maladaptive deliberate rumination and problem-focused coping. Int. J. Environ. Res. Public dimensions of relatedness and self-definition: relationship with postpartum Health 16. https://doi.org/10.3390/ijerph16203879. depression and anxiety. Pers. Individ. Dif. 41, 395–406. https://doi.org/10.1016/j. paid.2005.11.029. 589