Postnatal versus perinatal depression

The term “postnatal depression”, as distinct from postpartum psychosis or “maternity blues”, is used to refer to a depressive disorder affecting approximately 10–15% of postnatal women, and often satisfying DSM IV diagnostic criteria for Major Depressive Disorder (Boyce and Condon 2004). The DSM IV contains an onset specifier “postnatal” for Major Depressive Disorder (American Psychiatric Association 1994). This terminology has contributed greatly to increasing awareness in both health professionals and the lay public of this important disorder, as well as its potential severe impact upon the woman, her infant and her partner.

However, it can be argued that this term might well have outlived its usefulness, and that the term “perinatal depression” may be more appropriate in the light of current research findings.

There are at least four reasons why replacing “postnatal depression”, with “perinatal depression” might be desirable. First, depression during pregnancy is just as common as depression in the postnatal period (Evans et al. 2001; Lusskin et al. 2007). Second, the significance of depression during pregnancy has been greatly underestimated. There is increasing evidence that this (and other psychiatric disorders in pregnancy) may have direct adverse effects on foetal, obstetric, and neonatal outcomes. These include miscarriage, premature delivery, small-for-dates infants, stillbirth and possibly reduced head circumference (Hobel et al. 2008; Marcus 2009; Wisborg et al. 2008)). Such outcomes would appear to be independent of potential confounding factors such as alcohol intake, other substance abuse, smoking, caffeine intake, poor antenatal care, etc. Thus, psychological stress in pregnancy appears to have a direct adverse biological impact on the intrauterine environment in both humans and animals (Alder et al. 2007; Koubovec et al. 2005; Bonari et al. 2004). In addition, this direct adverse impact can have long-term implications for the child in terms of neuro-developmental disadvantage, which appears to be independent of the postnatal environment (Talge et al. 2007).

Third, pregnancy (especially first pregnancy) can be seen as a time of psychological preparation for parenting the future child (Condon 1990). An abundance of research supports the growth of antenatal attachment throughout pregnancy, and the potential for transfer of this attachment to the infant following delivery (Condon and Dunn 1988). First pregnancy is also a time for the couple to prepare for the transition from a dyadic to a triadic relationship, as well as the role transition to parenthood. Depression can inhibit antenatal attachment, and also interfere with achieving these other important developmental tasks (Condon 1990).

Fourth, depression during pregnancy is both under-diagnosed and under-treated. Currently, only about 20% of women with major depression in the antenatal or postnatal periods received treatment (Marcus 2009). Many health professionals are much more alert to depression in the postnatal period, and many still believe the myth that pregnancy is protective against depression (Lusskin et al. 2007). In addition, reluctance to make the diagnosis of depression during pregnancy may result from an unwillingness to prescribe antidepressant medication during pregnancy, and a lack of access to non-pharmacological treatments. In reality, well-defined recent guidelines are available for performing the risk-benefit analysis of prescribing psychotropic medications in pregnancy, and often the risks of not prescribing are substantially greater than those of prescribing (Psychotropic Expert Group Therapeutic Guidelines 2008). Moreover, untreated antenatal depression may be complicated by substance abuse, continued smoking, self-harm and poor antenatal care, all of which pose significant risks to the foetus.

Official adoption of the term “perinatal depression” in the DSM V would serve to refocus the attention of health professionals on the antenatal, as well as the postnatal period. Most women are in regular contact with health professionals during pregnancy, which would further facilitate the diagnosis, recognition of, screening for and early intervention in antenatal depression. The latter is a potentially disabling disorder affecting a very large number of women, and is a major public health problem which could be much better addressed.

Childbirth-related post-traumatic stress disorder

Post-traumatic stress disorder (PTSD) can be the result of a number of different pregnancy-related events. These include: termination of a wanted pregnancy; termination of an unwanted pregnancy because of foetal abnormality; early miscarriage; the woman’s subjective experience of either a “normal” or “traumatic” delivery; stillbirth; early neonatal death; relinquishment of the neonate for adoption and others (Condon 1986b; Boyce et al. 2002; Boyce and Condon 2000). The most well-researched of these is Post-traumatic Stress Disorder (as defined by DSM IV criteria) as a consequence of childbirth, which has been shown to complicate 1.5–6% of deliveries (average approximately 2.5–3%) (Ayers 2007; Soderquist et al. 2006; Alder et al. 2006). Considerably more women develop post-traumatic stress disorder symptoms, but do not fully satisfy DSM IV diagnostic criteria (Cigoli et al. 2006). Given the number of women undergoing childbirth, this represents a very considerable female psychiatric morbidity.

In addition to the suffering caused to women as a result of PTSD symptoms (i.e: recurrent nightmares, flashbacks, intrusive recollections, hyperarousal and emotional numbing), both the maternal-infant and partner relationships can be adversely impacted (Beck 2006). Prospective studies suggest that, once established, spontaneous remission of childbirth-related PTSD is uncommon, at least in the first 6–12 postnatal months (Zaers et al. 2008).

Different studies have identified different factors contributing to the development of childbirth-related post-truamatic stress disorder, aside from a delivery perceived by the woman (but not necessarily by attending health professionals) as traumatic. Such factors include first delivery, pre-existing fears of childbirth, anxiety in late pregnancy, dissociation experiences at delivery, high neuroticism scores, low educational level, and impoverished social support in the postnatal period (Cigoli et al. 2006; Engelhard et al. 2006 Soderquist et al. 2006; Olde et al. 2005).

Simple and reliable screening questionnaires for perinatal PTSD have been published (Callahan et al. 2006). However, the diagnosis is not difficult to make at clinical interview, provided the relevant questions are asked. Unfortunately however, the diagnosis is often missed by obstetricians and other health professionals, or mistaken for postnatal depression.

The inclusion of childbirth-related Post-tramatic Stress Disorder in the DSM V, either as a disorder in its own right or as an onset specifier for PTSD, would be an important means of emphasising the need for better recognition as well as the possibility of routine screening for this disorder. It would also pave the way for better education of health professionals in obstetric settings. As a result, the current unfortunate reality, characterized by the vast majority of sufferers remaining untreated, could begin to be addressed.

Disorders of parent-to-infant attachment

Over the last 15 years, since the publication of the DSM IV, the field of infant psychiatry has developed from a little-known area, which was often seen as “on the fringe”, to a respected scientific sub-specialty with sound theoretical underpinnings, a respectable empirical evidence base, reliable assessment techniques and effective treatments (Velderman et al. 2006).

Maternal-infant attachment has been, and remains, a central construct of this discipline, although more recently, some long overdue attention has been devoted to paternal-infant attachment (Condon et al. 2004, 2007; Condon 2006). The attachment construct and its assessment remains somewhat controversial. Some argue that attachment resides primarily in the infant, whilst others maintain that it is the emotional “bond” of the parent to the infant. Many believe it to be more a consequence of the mutuality and reciprocity inherent in the parent-child relationship. These different notions of attachment determine the approach one takes to its assessment (Condon and Corkindale 1998). The traditional approach utilises observation of parent-infant interaction (almost always mother), often involving a separation- reunion paradigm such as the Strange Situation, with the observed behaviours being rated along predefined dimensions (Ainsworth 1964). Such techniques are time-consuming, labour-intensive and expensive, which limits their applicability in clinical situations, and usually dictates small sample sizes in research settings.

We have adopted an alternative approach using parental self-report questionnaires, in which parents objectively report on the frequency and intensity of a number of predefined thoughts, feelings and behaviours involving their infant. We conceptulise the latter as “indicators” of parent-to-infant attachment. Such an approach is time-efficient and enables the use of larger sample sizes in research. This approach was a natural extension of our previous work on assessment of antenatal attachment. The psychometric properties of these questionnaires have been published (Condon 1993; Condon and Corkindale 1997, 1998; Condon et al. 2007).

Much remains to be learnt about the pathology of parent-to-infant attachment and its assessment. Dysfunctional parent-child relationships, even those involving various kinds of abuse, do not necessarily equate with lack of attachment. Also, a parent-child relationship which seems “good” in the absence of more detailed scrutiny, may camouflage a dysfunctional attachment relationship in which affection is conditional upon the child’s (or even the infant’s) achievement or “performing” as a trophy child.

A related area, the pathology of the maternal-foetal attachment relationship, remains almost entirely unexplored. The first accounts of physical foetal abuse were published in the mid 1980s (Condon 1986c, 1987). In fact, this represents one extreme end of a wide spectrum of foetal abuse and neglect, which few health professionals dealing with pregnant women ever enquire about. A classification system for these would highlight their importance in terms of early intervention during pregnancy, and potential prevention of dysfunctional attachment in the future parent child relationship.

It is not surprising that, given its publication date of 1994, the DSM IV fails to do justice to Infant Psychiatry. Separation Anxiety Disorder (309.21), focuses exclusively on the child rather than the caregiver, and the criteria are largely applicable only to older children. The criteria for Reactive Attachment Disorder of Infancy and Early Childhood (313.89) acknowledge some potential shortcomings in terms of the carer failing to provide basic physical and emotional care, but criterion A is again likely to be applicable to children rather than infants. The same criticism applies to Parent-Child Relational Problem (V 61.20).

Hopefully, the DSM IV’s omission of disorders involving the pathology of early parent-to-infant attachment will be corrected in DSM V. The construct of maternal and paternal sensitivity to their infant is now well understood, and its underlying dimensions have been defined (De Wolff and Ijzendoorn 1997; Smith and Pederson 1988; Pederson et al. 1998). Operationalisation of these into a set of criteria for deficits in sensitivity would now be feasible, and such an assessment could be carried out in a clinical situation involving mother-infant observation. The above-mentioned brief self-report questionnaires may also have a role, e.g. for screening for disorders of parent-to-infant attachment.

The DSM is a powerful force for guiding the education of mental health professionals. The incorporation of the disorders commonly encountered in infant psychiatry into the DSM V would highlight the possibility of early recognition and intervention in maternal and paternal infant dyads. A specific focus on the parent is required, as well as the infant. Health professionals in contact with families of infants would be more attuned to problems of parent-to-infant attachment and parental sensitivity, whether these are primary or secondary to co-morbid parental psychiatric disorders.

Perinatal bereavement

The DSM IV’s “Bereavement” (62.82), provides some very brief guidelines for differentiating normal bereavement from Major Depressive Disorder, but makes no meaningful attempt to address the pathology of grieving per se. This is surprising, given the large amount of literature on this topic which makes clear that disorders of grieving can be highly distressing and disabling their own right, without necessarily satisfying diagnostic criteria for Major Depressive Disorder . This omission potentially obscures a number of important and treatable conditions.

Perinatal bereavement is unfortunately a more common occurrence than many believe despite advances in obstetric and neonatal technologies. In modern Western environments, approximately 5 per thousand babies are stillborn, and a similar number will die in the first few days after birth (Stanton et al. 2006). There are a large number of factors which potentially complicate the grieving process following perinatal loss. First, despite a wealth of evidence that the antenatal attachment of most pregnant women and expectant fathers to their foetus progressively increases over the course of pregnancy (Condon 1993; Condon and Corkindale 1997; Condon et al 2004, 2007), many health professionals and lay people, greatly underestimate the psychological significance of the breaking of this attachment bond through antenatal pregnancy loss or stillbirth (Condon 1986a, 1986b, 2000; Frost and Condon 1996). As a result, pathological mourning in this setting, may go unrecognized, untreated or be mistreated as “postnatal depression” (with antidepressant medication).

Second, wide variation still exists in hospital practice in terms of viewing and holding the deceased infant, provision of memorabilia, performance of autopsy, funeral/memorial services etc (Gold et al. 2007; Badenhorst and Hughes 2007). Sometimes guidelines are interpreted as “protocols” without consideration of individual parents’ needs. What might be helpful for some parents may be harmful and traumatising for others. A firm evidence base for any of these practices is yet to be developed, but some progress is being made (Cacciatore et al. 2008; Surkan et al. 2008; Saflund and Wredling 2006).

Third, it is self evident that the loss of a child, whose potential is yet to be realised, is more difficult to grieve than that of an elderly adult. However, in the case of stillbirth, the attachment to the lost object may be very strong, yet the object itself has existed only in the imagination, and has never been encountered as a living being. Normal grieving involves remembering the characteristics of the lost object and reliving the relationship with that object. Grief therapy involving stillbirth may involve helping the parent “create” an object, which is sufficiently realistic for grieving to occur (Condon 1986a). As one parent expressed it “how can I say goodbye to someone I have never said hello to”. In the same vein, Vaillant (1985) states “The psychological work of mourning is to remember more than it is to say goodbye….grief work is remembering, not forgetting; it is a process of internalizing, not extruding.”

Fourth, pregnancy, soon after stillbirth, may sometimes be an ‘antidote’ to grief through replacement. However, the “appropriate” time to delay subsequent pregnancy is controversial and depends greatly on the individual parents and the circumstances of the stillbirth (Surkan et al. 2008; Barr 2006; Turton et al. 2006). Any recommendation based on a fixed time period is inevitably an over-generalisation (Condon 2000).

Fifth, the affective component of sadness following stillbirth is frequently complicated by a significant component of anger which is often directed at the health professionals involved. In the author’s experience, it is rare for guilt to not also be a component of parental stillbirth distress. The latter frequently leads the mother to focus on trivial events during pregnancy for which she blames herself. The father is also not immune to self-blame (for example believing that sexual intercourse may have contributed). In the author’s experience, these desperate pursuits of attribution or causality are often primitive defenses against facing the frightening realisation that catastrophic events can occur randomly, and can befall “good” as well as “bad” people. Thus, rage at the staff or crippling guilt are “preferable” to accepting randomness. If the patient can believe someone is to blame, the problem can potentially be remedied on the next pregnancy. However, the anxiety accompanying the acceptance that such events are beyond our control, and occur randomly, can be overwhelming for some parents (Condon 2000).

Thus, perinatal bereavement is often quite psychologically complex, can cause very substantial distress and frequently requires specific psychotherapy. There are many potential obstacles to progression through the four stages of normal bereavement, viz: denial, angry protest, despair/ grief and resolution. “Replacement pregnancy” may indicate failure to progress beyond the first stage of denial. Abnormally prolonged grieving, may indicate that the parent is dealing with sadness (stage 3), but has not yet dealt with coexisting anger and/or guilt (stage 2) to allow progress to resolution (stage 4). Clearly, in any grieving patient, it is a very important to exclude Major Depressive Disorder and Post-Traumatic Stress Disorder.

It is to be hoped that the architects of the DSM V will go beyond 62.82 (“Bereavement”), and at least present a simple evidence-based typology of the pathology of grief. Instruments to assist distinguishing grief and depression have been published (Toedter et al. 1988). Such a DSM V classification would provide an impetus for better education of health professionals in this area, leading to improved recognition and treatment of pathological grieving through grief therapy. The beneficiaries would not only be parents suffering perinatal bereavement, but also those experiencing other bereavements whose resolution is potentially fraught. The latter would include combat-related deaths, deaths following terrorist attacks, suicides, homicides and death following HIV infection.