Perinatal depression is defined as the occurrence of a depressive episode in pregnancy or within six weeks post partum (DSM-5; American Psychiatric Association, 2013). It is a major global health concern, affecting the health and well-being of more than 10% of pregnant women, new mothers and their families each year (Gavin et al., 2005). Accordingly, a growing body of strong evidence addresses the prevalence, risks and consequences of this disorder. Specifically, a large number of studies have investigated biopsychosocial risk factors for perinatal depression. In our 2015 systematic review of biological and psychosocial predictors of postpartum depressive symptoms (Yim, Tanner Stapleton, Guardino, Hahn-Holbrook, & Dunkel Schetter, 2015), we identified 199 relevant studies. One of our major findings was that the separate literatures investigating biological and psychosocial contributors to postpartum depression risk had evolved independently and remained distinct. That is, of the 199 studies, a mere eleven considered at least one biological and one psychological predictor together in the same study; of those eleven, only six included the biological and psychological variables in the same statistical model. Thus, we concluded our previous review paper with a “Call for Integration.” Since the publication of our systematic review, the number of studies testing biopsychosocial factors in postpartum depressive symptoms has increased somewhat. In an effort to quantify this increase, we again systematically searched the literature in the same manner as in our initial systematic review conducted five years ago (Yim et al., 2015). Publication dates were limited to January of 2014 to December of 2018 for the present search. Briefly, combinations of the key words “depression,” “postpartum” and “postnatal” were combined with relevant broad and specific biological terms (e.g., biological, hormone, cortisol) and psychosocial terms (e.g., psychological, stress, demands). Similar to our previous search, we included only English language publications and peer-reviewed human studies within the first year following birth that reported a statistical association between the relevant variables. However, because we were only interested in studies that included both a biological and psychosocial predictor, we required this when searching – in deviation from our 2015 search. In other words, search findings had to include the depression-related keyword, a biological term, and a psychosocial term to be included here. We identified 69 records in PubMed and 42 in PsycINFO, resulting in 98 records after removal of 13 duplicates (see flow diagram in Fig. 1). Four additional manuscripts were identified through a reference section review. Of these 102 records, 61 were deemed clearly irrelevant after abstract review. The remaining 41 full text articles were assessed for eligibility, and 32 studies were excluded on the basis of the criteria stated. The remaining nine papers met inclusion criteria and are summarized in detail in Table 1. What stands out is that the rate of publications of integrative studies has indeed increased. We now identified nine integrative papers published across roughly three years, whereas in our previous review we identified six papers published over the course of 13 years. We interpret this to reflect an important shift among researchers toward recognition of the importance of integrative approaches. Moreover, we note that the new studies identified are less likely to rely on solely cortisol measures and instead include other important biological markers such as oxytocin and markers of inflammation treated as both cross-sectional correlates and longitudinal predictors of postpartum depressive symptoms. To summarize briefly the publications over the last three years, studies of oxytocin confirm our previous conclusion that evidence is largely in favor of an inverse association with postpartum depressive symptoms (e.g., Eapen et al., 2014), with some important qualifications. For example, one study found this association only among individuals who also scored high on a measure of psychosocial stress (Zelkowitz et al., 2014), and another found no association between endogenous oxytocin and postpartum depressive symptoms overall, but yielded a positive association with synthetic oxytocin administered at birth, which held in a model covarying relationship status (Gu et al., 2016). One last study found that childhood abuse and postpartum depression interacted to modulate oxytocin receptor DNA methylation (Kimmel et al., 2016). In studies assessing inflammatory markers, one suggests that inflammatory markers (IL-6, CRP) were associated with depressive symptoms within two days of hospital admission for delivery, and controlling for stressful life events and low partner support (Liu, Zhang, Gao, & Zhang, 2016). Another investigation found that TNF-α and IL-10 were associated with EPDS scores measured at 1 week post partum, but this association did not hold in a multivariate model controlling for the effects of perceived stress (Dunn, Paul, Ware, & Corwin, 2015). Similarly, in a path model revealing significant direct effects of perceived stress on EPDS symptoms, neither a latent inflammation variable nor diurnal salivary cortisol emerged as significant (Ruyak, Lowe, Corwin, Neu, & Boursaw, 2016). Also concerning HPA axis function, one study showed that the association between an HPA axis-related polymorphism and depressive symptoms at six weeks post partum was mediated by neuroticism (Iliadis et al., 2017). And, a single study reported no evidence of an association of postpartum depressive symptoms and self-reported social support or adiponectin, a hormone involved in glucose and lipid metabolism (Rebelo, Farias, Struchiner, & Kac, 2016). This brief summary of the literature published between our previous review and the publication of this special issue provides further

中文翻译：

---

围产期抑郁症状的生物心理社会预测因素：走向综合方法

围产期抑郁症定义为在怀孕期间或产后六周内发生抑郁发作（DSM-5；美国精神病学协会，2013 年）。它是一个主要的全球健康问题，每年影响超过 10% 的孕妇、新妈妈及其家人的健康和福祉（Gavin 等，2005）。因此，越来越多的有力证据表明了这种疾病的流行、风险和后果。具体而言，大量研究调查了围产期抑郁症的生物心理社会危险因素。在我们 2015 年对产后抑郁症状的生物学和社会心理预测因素的系统回顾中（Yim、Tanner Stapleton、Guardino、Hahn-Holbrook 和 Dunkel Schetter，2015 年），我们确定了 199 项相关研究。我们的主要发现之一是，研究产后抑郁风险的生物学和社会心理因素的单独文献是独立发展的，并且保持不同。也就是说，在 199 项研究中，只有 11 项在同一研究中至少考虑了一种生物学和一种心理预测因素；在这 11 个中，只有 6 个将生物和心理变量包含在同一统计模型中。因此，我们以“呼吁整合”结束了我们之前的评论论文。自我们的系统评价发表以来，测试产后抑郁症状中生物心理社会因素的研究数量有所增加。为了量化这种增长，我们再次以与五年前进行的初始系统审查相同的方式系统地检索了文献（Yim 等，2015）。本次检索的出版日期仅限于 2014 年 1 月至 2018 年 12 月。简而言之，将关键词“抑郁”、“产后”和“产后”与相关广泛和特定的生物学术语（例如生物学、激素、皮质醇）和社会心理术语（例如心理、压力、需求）相结合。与我们之前的搜索类似，我们仅纳入了出生后第一年内报告相关变量之间统计关联的英语出版物和同行评审的人类研究。然而，因为我们只对同时包含生物学和社会心理预测因子的研究感兴趣，所以我们在搜索时需要这样做——这与我们 2015 年的搜索有所不同。换句话说，搜索结果必须包括与抑郁症相关的关键词，一个生物学术语，以及此处包含的社会心理术语。我们在 PubMed 中确定了 69 条记录，在 PsycINFO 中确定了 42 条记录，在删除了 13 个重复项后得到了 98 条记录（见图 1 中的流程图）。通过参考部分审查确定了另外四份手稿。在这 102 条记录中，经过摘要审查，有 61 条被认为明显无关。其余 41 篇全文文章进行了资格评估，根据规定的标准排除了 32 项研究。其余9篇论文符合纳入标准，详细总结在表1中。值得注意的是，综合研究的发表率确实有所提高。我们现在确定了大约三年内发表的九篇综合论文，而在我们之前的审查中，我们确定了在 13 年内发表的六篇论文。我们将其解释为反映了研究人员对整合方法重要性的认识的重要转变。此外，我们注意到，确定的新研究不太可能仅依赖于皮质醇测量，而是包括其他重要的生物标志物，例如催产素和炎症标志物，它们被视为产后抑郁症状的横断面相关性和纵向预测因子。简要总结过去三年的出版物，催产素研究证实了我们之前的结论，即证据在很大程度上支持与产后抑郁症状呈负相关（例如，Eapen 等，2014），具有一些重要的资格。例如，一项研究发现这种关联仅在那些在心理社会压力方面也得分很高的个体中（Zelkowitz 等人，2014），另一项发现内源性催产素与产后抑郁症状之间总体上没有关联，但与出生时使用的合成催产素呈正相关，处于模型共变关系状态（Gu et al., 2016）。最后一项研究发现，童年虐待和产后抑郁症相互作用以调节催产素受体 DNA 甲基化（Kimmel 等人，2016 年）。在评估炎症标志物的研究中，一项研究表明炎症标志物（IL-6、CRP）与住院分娩后两天内的抑郁症状有关，并控制压力性生活事件和伴侣支持度低（Liu、Zhang、Gao 和张，2016）。另一项调查发现 TNF-α 和 IL-10 与产后 1 周测量的 EPDS 评分相关，但这种关联在控制感知压力影响的多变量模型中并不成立（Dunn、Paul、Ware 和 Corwin，2015 年）。同样，在揭示感知压力对 EPDS 症状的显着直接影响的路径模型中，潜在炎症变量和昼夜唾液皮质醇均不显着（Ruyak、Lowe、Corwin、Neu 和 Boursaw，2016 年）。同样关于 HPA 轴功能，一项研究表明，HPA 轴相关多态性与产后六周的抑郁症状之间的关联是由神经质介导的（Iliadis 等，2017）。而且，一项研究报告没有证据表明产后抑郁症状与自我报告的社会支持或脂联素（一种参与葡萄糖和脂质代谢的激素）有关（Rebelo、Farias、Struchiner 和 Kac，2016 年）。