Priorities in the assessment and management of perinatal mood disorders

Achieving and maintaining maternal well-being is a primary goal in perinatal psychiatry because, naturally, this optimizes the environment for healthy foetal and infant development. Attention has been recently paid to what happens in pregnancy with the knowledge that perturbations to normal foetal development can have long-reaching effects on subsequent health and well-being. The Developmental Origins of Health and Disease (O’Donnell and Meaney, 2017) is a framework that builds on the observations that poor maternal diet and low infant birth weight predict the development of metabolic syndrome in adulthood. But, in addition to affecting physical health, this also applies to mental well-being and indeed the origins of many psychiatric disorders can be traced back to in utero inception, with exogenous factors during pregnancy, such as famine or poor diet quality (with deficiencies in micronutrients), maternal smoking or maternal influenza during pregnancy increasing the risk of schizophrenia, depression and bipolar disorder among offspring (Boyce et al., in press). We also now know, as one would have anticipated, that exposing the foetus to teratogens such as alcohol or sodium valproate during pregnancy has a deleterious effect on neurodevelopment.

The effects of maternal health on the baby are complicated, but there have been an increasing number of studies examining the effects of perinatal depression on the subsequent mental health of their offspring (Stein et al., 2014). Depressed mothers may have sub-optimal interactions with their infants as a result of their illness—because of inattentiveness or internal preoccupation, which is considered to impact infant attachment and contribute to the poor cognitive, interpersonal and emotional outcomes found in their children. These findings highlighted the necessity of identifying and treating postnatal depression in order to prevent the intergenerational transmission of mental health problems. Similar outcomes have now been found for the offspring of women who were depressed during pregnancy (Stein et al., 2014), indicating a need to identify and treat depression during pregnancy.

Undoubtedly, protecting the foetus during pregnancy is paramount for optimal development. This has led for calls to introduce screening in the perinatal period to identify women with depression and treat them appropriately to benefit both the developing foetus and the mother. Treating depression during pregnancy, however, raises a conundrum: could the treatment directed at the mother be harmful to the foetus? This problem is the key reason for the recommendation in clinical practice guidelines regarding the management of perinatal depression—to undertake a careful risk–benefit analysis. The essence of this exercise is to balance any potential harms to the developing foetus from exposure to psychotropic medication against any potential harms to the mother or the developing foetus from the mother being depressed. Such an analysis focuses on the pharmacological treatment for perinatal depression (despite a lack of clinical trial evidence for its efficacy—pregnancy is an exclusion criterion) and does not necessarily apply to the evidence-based psychological interventions that may not be harmful to the developing foetus.

There is a plethora of meta-analyses of observational studies evaluating the safety of antidepressants for the developing foetus (Chisolm and Payne, 2016). In general, they conclude that there is an increased risk of poor neonatal adaptation syndrome and some minor impact on obstetric outcomes. A potential small increased risk of congenital heart defects (with a small increase in the absolute risk, but a significant increase in relative risk) noted in some of the meta-analyses may, however, be due to the underlying behaviors associated with depression. What remains uncertain is whether these results are due to the agents themselves or are a proxy of other risk factors driving prescription. Such findings should be reassuring for clinicians considering the use of antidepressants to treat depression arising during pregnancy. However, the long-term effects of antidepressants on neurological and psychosocial development of the infant are much less certain.

As well as being of benefit to the woman a reason for treating depression during pregnancy is to avoid harm coming to the foetus as a consequence of maternal depression; but the key question to consider is whether maternal depression really is harmful? Depression is an all-encompassing term ranging from having some depressive symptoms to experiencing mild and transient depressive episodes to having a severe and persistent depressive episode. Before concluding that a pregnant woman’s ‘depression’ can cause harm to her developing foetus, we have to be precise about what is meant by depression. Depression is a heterogeneous disorder and possibly more so during pregnancy, when understandable processes of adjustment to major life circumstances and the symptoms of pregnancy itself can mimic depressive symptoms, leading to spurious diagnoses.

Women will often report transient depressive symptoms during pregnancy and may be considered depressed if scoring above the threshold (especially when ‘screened’ when first attending an antenatal clinic) that will not be present when reassessed a few weeks later (Matthey and Ross-Hamid, 2012). Additionally, when women have a clinically significant episode of major depression, there is still heterogeneity between depression presentations that have more biological underpinning (melancholia or bipolar depression) and those that are predominantly psychosocial in origin. The differences between these ‘types’ of depression may be of critical importance when considering the impact of the depressive episode on the developing foetus.

The neurobiological underpinnings of depression, such as disruption to the hypothalamic–pituitary–adrenal axis (O’Donnell and Meaney, 2017), with the associated heightened maternal glucocorticoid levels that can cross the placenta as a result of reduced placental 11-beta-hydroxysteroid dehydrogenase (11β-HSD-2; as a consequence of stress or neuroinflammation) have been proposed as mechanisms having a deleterious impact on foetal programming. However, it is most unlikely that these biological changes are found in all forms of maternal depression.

By contrast, the more psychosocially driven episodes of depression that arise from a background of early life stressors or psychosocial adversity (such as poverty, inequality, poor diet, impaired social relationships and dysfunctional personality style) may be more important drivers of difficulties in the mother–infant relationship and its development than the depression per se. Attributing the problems to the depression will also mean that attention is not focused on the risk factors leading to the depression in the first instance and that the social problems may not be dealt with appropriately.

Undoubtedly, some episodes of depression can have an adverse impact on the foetus, and there is a pressing need to understand the underpinning mechanisms. This requires that we identify the characteristics of the depression that can lead to this, in particular which type of depression (e.g. melancholic vs non-melancholic, bipolar vs unipolar), its severity and the gestational age at which the depression emerged. We also need to provide appropriate treatment to pregnant women to permit healthy development.

An unfortunate side effect of the assertion that depression during pregnancy causes harm to the developing foetus is the guilt and anxiety that it can have on the mother. Thus, the risk–benefit discussion has to include a careful explanation that most depression is not necessarily going to cause harm to her developing foetus. It should be framed positively: that supporting the mother supports the child. We do not argue that it is only some forms of depression that require treatment; depressive symptoms (and depressive episodes) during pregnancy are strong predictors for postnatal depression and require thorough clinical assessment (to avoid the risk of over-pathologizing mild and transient symptoms) and appropriate treatment. Attention also needs to be paid to the potentially remedial psychosocial and environmental factors that have contributed to the depressive episode (or symptoms).

There are priorities in research and clinical practice that need to be addressed. In order to be able to determine whether depression has an adverse impact on foetal programming, we recommend that, in future research, structured diagnostic instruments (that allow subtypes of depression to be identified, such as melancholia or bipolar depression) should be used to assess depression rather than self-report questionnaires. There should be an objective rating of the severity of the episode of depression. The time and duration of the episode should be specified in order to determine if there are critical periods impacting on the foetus. Ideally, credible biomarkers, such as cortisol levels or inflammatory markers, should also be measured. Assessment of relevant psychosocial factors should also be included.

From a clinical perspective, any mother identified as having depression should undergo a rigorous clinical assessment of her depression and its severity alongside an assessment of other factors that could impact on the development of her foetus. Treatment of mild to moderate depression should involve commencing evidence-based psychological treatments (while addressing other psychosocial and lifestyle factors) with antidepressant medication generally reserved for women with melancholia or a moderate to severe depression. This emphasizes a need for specialist perinatal psychiatric services to be established alongside obstetric units so that women with depression (and other psychiatric disorders) during pregnancy can have more comprehensive assessment and appropriate treatment to be delivered. The Maternal Mental Health Alliance in the United Kingdom has been advocating for establishing specialist perinatal service, including mother and baby units, arguing that an investment in perinatal mental health will reap great health benefits for the next generation as well as reducing costs in the long term.

We agree, and feel that there needs to be a change, both in research focus and clinical practice, to advance our knowledge of this critical period of life and the impact that maternal health has on the subsequent development of psychiatric disorders.