Psychiatric causes of maternal deaths: Lessons from the Confidential Enquiries into Maternal Deaths

The UK Confidential Enquiries into Maternal Deaths, published triennially, are over 50 years old. Its forebears are even older; enquiries into maternal deaths began early in the 19th century in Scotland. In the 20th century the numbers of women dying from childbirth has steadily declined, influenced by many factors, including improved public health and maternity care, smaller family size, blood transfusions, and antibiotics, to name but a few. The introduction of the Abortion Act in 1967 was followed by a marked reduction of deaths in pregnancy from the consequences of illegal abortion. The rate and causes of maternal death have always been influenced by changes in reproductive epidemiology and technology, and continue to be so. Maternal deaths in pregnancy and in the 6 weeks following delivery are required to be reported to the Coroner, if directly related to childbirth. However, there are other causes of maternal death due to conditions exacerbated by pregnancy: for example, diabetes, cardiac disease, epilepsy. These are referred to as indirect deaths. Women who die from conditions unrelated to pregnancy or childbirth are counted and described as coincidental deaths. Over the years as the direct causes of maternal death have fallen, the indirect causes of maternal death have achieved more prominence and case ascertainment has improved. Improvements in medical care and in particular intensive care have resulted in some women developing their fatal condition within 6 weeks of childbirth, only to die beyond it. For this reason, the UK Enquiry extended their period of surveillance beyond 6 weeks to include late maternal deaths, both a small number of late direct deaths and a larger number of late indirect deaths. Suicide in pregnancy and following delivery has always been included in the Enquiries. However, prior to 1994 the cases were not separately analysed and were included in the group of late Coincidental Deaths (i.e. not thought to be related to pregnancy or childbirth). The 1994–1996 Enquiry, under the Directorship and Editorship of Dr Gwyneth Lewis and Professor James O’Drife, heralded a change in presentation of the Enquiry.

Postpartum psychosis—important clues to the aetiology of mood disorders

It is of great regret that although corresponding with him at the start of my research career, I never met Channi Kumar face to face. His work, however, as evidenced by this book, remains an important influence on our field. I share his belief in the ‘maternal brain as a model for investigating mental illness’ (Kumar 2001), and this conviction has underlined much of my research. In this chapter I will discuss the concept of postpartum psychosis (PP), explore what we know about the relationship of these episodes to other mood and psychotic disorders, and consider research strategies aimed at understanding the nature of the postpartum trigger. I will argue that the nosological confusion surrounding this condition has been unhelpful and that it is time, perhaps, to consider whether we should revive postpartum psychosis as a diagnostic concept. Episodes of mood disorder in relation to pregnancy and childbirth are very common. In our group we have recently examined the history of perinatal episodes in over 1,500 women with mood disorder who have participated in our genetic studies and find that approximately two thirds of parous women, with both bipolar and unipolar disorder, have experienced a significant mood episode in the perinatal period (Di Florio et al. 2013). PP refers to some of the most severe forms of postpartum psychiatric disorder. Although the boundaries of this condition are not easy to define, the core concept is the acute onset of a manic or affective psychosis in the immediate postpartum period. Depending on the definition employed, the incidence is approximately 1 in 1,000 deliveries (Jones et al. 2010). Women may go from being very well to severely ill within hours. Affective (mood) symptoms, both elation and depression, are prominent, as is a disturbance of consciousness marked by an apparent confusion, bewilderment, or perplexity. As the name suggests, psychotic phenomena occur, with delusions and hallucinations prominent. Some women with severe manic episodes, but who do not show psychotic symptoms, may receive the diagnosis, although it is also possible to reserve the label for those women with frank psychotic presentations.

Not a native son

In the summer of 1953, aged 7, I arrived with my father at the port of Southampton from the colony of Nigeria. We were making for Ledsham Court School, a boarding school in St Leonards-on-Sea, Sussex. It was a stately building sitting among many green acres. After about an hour with the headmistress, Mrs Redfarn, my father said goodbye, turned and returned to Nigeria. I did not know then that I would not see or hear from him for 10 years, by which time I had forgotten what he looked like. Ledsham’s only black pupil began his academic life speaking no English. I was duly placed in the kindergarten with daily lessons in the native tongue. After catching up with my age group, in addition to the core subjects I was thereafter given instruction in Latin, ancient Greek, poetry, and nature study. To eradicate ‘that funny African accent’ I was solely accorded a daily class of elocution for a year—one hour a day with a speech therapist, held in a long, oak-panelled gallery with a book on my head to improve my deportment. Although in receipt of the beginnings of a good classical education, I was also given what I came to understand was a prototypical quantity of punishment for a ‘darkie’—for most of that first year I was caned daily and frequently ‘sent to Coventry’ for the slightest indiscretion, usually for not understanding the customs and traditions of an alien white culture. Thus, for refusing to eat salad on my first day, I received ‘three of the best’. The staff were undoubtedly ignorant of the eggs that parasites can lay on raw vegetables in a tropical climate like Nigeria, where all vegetables were cooked and salad was unheard of. Perhaps, I thought with a child’s naivety, that with all the mosquitoes and eating of salad, no wonder West Africa was called the white man’s grave in my books and comics. I woke up—for I had clearly landed in the mother country in the wrong skin colour. It hurt. I had arrived knowing myself to be Yoruba. Suddenly, I was called ‘coloured’ and ‘darkie’.

Mother- and father-to-infant emotional involvement

A mother’s specific emotional and hormonal state after childbirth ensures her emotional involvement and adequate parental behaviour. Soon after delivery, or even in late pregnancy, the mother’s emotional state—in particular, an increased sensitivity—becomes fully adapted to the identification and satisfaction of the infant’s physical and psychological needs. Winnicott (1956, 1960) was perhaps one of the first authors to point out the presence of a particular emotional state in recently delivered mothers—‘primary maternal preoccupation’, referring to the mother’s correct identification and immediate satisfaction of the infant’s physical and psychological needs. Winnicott (1990) later defined and described four main tasks to be fulfilled in the maternal role, including the emotional involvement with the child, which he termed ‘holding’. Holding tasks are: (1) to provide protection and care to the child, (2) to take into account the child’s limitations and dependency status, (3) to provide the necessary care for the child’s growth and development, and (4) to love the child. In the meantime, Yalom et al. (1968) and Pitt (1973) both described the ‘postpartum/maternity blues—a transient state of emotional dysphoria, emerging within a few hours to 2 weeks after childbirth, in about 50 to 70% of puerperal women, and characterized by intermittent mild fatigue, tearfulness, worry, difficulty in thinking, and sleep disturbances. Progesterone and oestrogen levels, which gradually increase during pregnancy, fall suddenly after delivery, returning to prepregnancy levels in just 3 days. This rapid decline, the most severe threat to a women’s hormonal and emotional balance, has been proposed as the main cause of postpartum/maternity blues (e.g. Pitt 1973; Yalomand et al. 1968). The mother’s behavioural sensitivity to such a drop in reproductive hormones was later associated with higher reactivity to the infant’s stimuli and greater proximity with the neonate (e.g. Barrett and Fleming 2011; Carter 2005; Fleming et al. 1997; Miller and Rukstalis 1999), and was proposed as serving the function of eliciting mother-to-infant involvement, to ensure that the infant receives the required care to survive (e.g. Carter 2005; Figueiredo 2003; Pedersen 1997). The evolutionary point of view had its clearest proponent in John Bowlby (1969/1982, 1980) who proposed the presence of a behavioural system (that is, an organized set of behaviours) in parents—the ‘caregiving system’, to guarantee the proximity and protection of the child.

Personal reflections on the early development of the EPDS

Shortly after returning to the London Hospital from Uganda in 1974, and still jet-lagged and culture-shocked, I had an unexpected call from a Dr Kumar at the Maudsley Hospital whose name was then unfamiliar to me. Stephen Wolkind (Child Psychiatrist at the London Hospital working with Professor Pond) had informed him that I had completed a study of postnatal depression in East Africa and had used Goldberg’s Standardised Psychiatric Interview (SPI) translated into Luganda. Could we meet, and could I advise him on the use of the SPI? I was surprised, flattered and motivated by this request. We met in Turner St, London E1. This was the beginning of a friendly and mutually respectful collaboration, which facilitated the later development of the Edinburgh Postnatal Depression Scale (EPDS) (Cox et al. 1987), helped launch the 1980 meeting in Manchester, when the Marcé Society was founded, and motivated Phase One of the International Transcultural Postnatal Depression Study. Channi Kumar was a fine team player, and as a leader had that knack of making you feel respected and at ease. His greeting ‘Come in dear boy and have a seat’ when he ushered you to a chair piled high with research papers, was characteristic of his style and productivity. We would then talk, not only about screening scales, but about College matters and the Perinatal Special Interest Group (which later became a Specialty Section), as well as our ‘Blue Skies’ research programmes. My interest in perinatal psychiatry began when, as an impressionable medical student, I first met Brice Pitt at Claybury Hospital. He was carrying out a study of ‘atypical’ postnatal depression and was devising a self-report questionnaire to detect increases in depression scores after birth. This early experience, together with a postgraduate seminar some years later, must surely have been on my mind when I was asked by Allen German on my arrival in Uganda, what research I was planning to do. I replied that I wished to replicate Assael’s finding (1972) that a quarter of pregnant women at Kasangati had mental health problems, and I was curious to know whether African women experienced depression as described by Pitt (1968)—and if not, what were the differences.

The psychoses of childbearing

It is 50 years since the late Ralph Paffenbarger (1961) wrote a famous article on ‘the picture puzzle of postpartum psychosis. In order to solve this puzzle, it is necessary to clarify the term ‘postpartum psychosis’. One must first exclude a wide variety of disorders, occurring after childbirth, which are not ‘psychoses’. This may seem obvious, but, at one time, some psychoanalysts included disorders of the mother-infant relationship under ‘postpartum schizophrenia’ (Zilboorg 1929). One must then draw a clear boundary between organic and non-organic psychoses. The birth process is so complex, and has so many complications, that there are (depending on definition) 15–18 distinct organic psychoses occurring in pregnancy, parturition or the puerperium (Brockington 2006). Nineteenth century alienists found it difficult to distinguish these from puerperal mania, and this was not finally achieved until the work of Chaslin (1895) & Bonhöffer (1910) at the turn of the twentieth century. Even the most common of these organic psychoses—eclamptic psychosis and infective delirium—are now rare in Europe, North America, and Japan; but these nations, where most of the research is done, contribute less than 10% of the world’s births. In the rest of the world they may be important, and they may still interfere with epidemiological, genetic, and neuroscientific studies of non-organic psychoses. As for the non-organic psychoses, a few are psychogenic, but most have manic depressive features. The term ‘puerperal affective psychosis’, however, does not suffice, because there is an extensive literature on ‘atypical psychoses’, under names like hallucinatorische Irresein der Wochnerinnen (Furstner 1875), amentia, cycloid psychosis, and acute polymorphic psychosis. That is why some psychiatrists still claim that ‘puerperal psychosis’ is a specific disorder, with its own clinical features—those ‘specific features’ are the polymorphic symptoms found in ‘atypical psychoses’, and occur in women at other times, and in men. Ralph Paffenbarger’s ‘picture puzzle’, therefore, applies to the combined group of puerperal bipolar and acute polymorphic psychoses.

Child abuse in the United States

Child abuse is a major cause of morbidity and mortality in the United States and other countries. It is the second leading cause of death among children in the US. All 50 States, the District of Columbia, and the US Territories have mandatory child abuse and neglect reporting laws that require certain professionals and institutions to report suspected maltreatment to a child protective services (CPS) agency. Four major types of maltreatment are considered: neglect, physical abuse, psychological maltreatment, and sexual abuse (Centers for Disease Control and Prevention 2010). Once an allegation or referral of child abuse is received by a CPS agency, the majority of reports receive investigations to establish whether or not an intervention is needed. Some reports receive an alternative response in which safety and risk assessments are conducted, but the focus is on working with the family to address issues. Investigations involve gathering evidence to substantiate the alleged maltreatment. Data from reports on child abuse is derived from the National Child Abuse and Neglect Data System (NCANDS), which aggregates and publishes statistics from state child protection agencies. The first report from NCANDS was based on data for 1990. Case-level data include information about the characteristics of reports of abuse and neglect that are made to CPS agencies, the children involved, the types of maltreatment that are alleged, the dispositions of the CPS responses, the risk factors of the child and the caregivers, the services that are provided, and the perpetrators (Centers for Disease Control and Prevention 2010). During 2010, the NCANSDS reported that an estimated 3.3 million referrals estimated to include 5.9 million children were received by CPS agencies. Of the nearly 2 million reports that were screened and received a CPS response, 90.3% received an investigation response and 9.7% received an alternative response (Centers for Disease Control and Prevention 2010). Of the 1,793,724 reports that received an investigation in 2010, 436,321 were substantiated; 24,976 were found to be indicated (likely but unsubstantiated); and 1,262,118 were found to be unsubstantiated. Three-fifths of reports of alleged child abuse and neglect were made by professionals.

The Children Act 1989: Success or failure

Heralded as the most progressive legislation of the world, the Children Act of 1989 revolutionized children’s law in England and Wales. It is underpinned by six principles: the supremacy of the child’s interest in all decisions concerning their upbringing and education; the recognition that it is best for any chid to be brought up by their blood family, that his religious and ethnic background must be respected, and that siblings should not be separated; the abolition of the stigma of illegitimacy and its replacement with the attribution at birth of paternal responsibility to the child’s father; the unification of public and private law, and the creation of the ‘menu’ of Residence, Contact, Prohibition, and Specific Issue orders available to the court; the establisment of the new principle that time is of the essence in all cases relating to children; and the creation of the presumption that ‘no order is better than an order’ thus the ingerence of the court must be minimal. I believed in those principles and in the benefits that the Children Act would bring to my clients—children and parents alike. I had some reservations: the system was expensive to implement on two counts: first, it gave the child a ‘guardian’ (a qualified social worker appointed by the court through CAFCASS, a governmental agency), as well as their own solicitor paid for by Legal Aid, as was the representative of the parents, who had the right to instruct independent experts; second, because its requirements of social services and other agencies involved further training and increased resources, as well as further involvement of the judiciary, and increased court time. Hornby and Levy were at the forefront of its implementation: our entire staff received in-house training that was open to other disciplines, within the spirit of cooperation between agencies that permeated the Act and its implementation. I also lectured in Britain and abroad and was proud to tell others that social services were under a duty to keep families united, rather than removing children from parents, and make efforts to return to the family the child removed from it, or if this failed, to place the child within the extended family, or with adoptive parents, within a year.

The intergenerational transmission of stress: psychosocial and biological mechanisms

I met Channi for the first time when I was a senior house officer (trainee) in psychiatry at the Maudsley Hospital, and I worked under his supervision for 6 months, in 1998. At that time, Channi was the only Consultant Perinatal Psychiatrist at the Maudsley, covering the Liaison Services at King’s College Hospital, the outreach work, and the Mother and Baby Unit. And, of course, he was leading the academic section. It is perhaps the best tribute to his memory that it takes now three consultants and two academics to do the work that he was then doing all by himself! I was already interested in neuroendocrinology, and Channi was fascinated by the possibility that hormones might have a role in the mental health problems of the perinatal period. At that time, the notion that hormonal changes in pregnancy could have long-lasting effects on the offspring was still at its infancy, and I remember fondly the many discussions on this topic with Channi, sitting at his famous old desk. Channi was a pioneer in this field: he was the first to emphasize the dramatic impact of depression in pregnancy on the wellbeing of mothers and children. I am honoured to be able to continue this line of research today. The intergenerational transmission of stress has powerful clinical and social consequences, consolidating social adversity and psychopathology in future generations. The 2007 Policy Briefing by the World Health Organization Regional Office for Europe, ‘Preventing child maltreatment in Europe: a public health approach’ (WHO 2007), recognizes that ‘there is an association between maltreatment in childhood and the risk of later . . . becoming a perpetrator of violence or other antisocial behaviour as a teenager or adult’. The report also highlights that the costs are both overt (for example, medical care for victims, treatment of offenders, and legal costs for social care) and less obvious (for example, criminal justice and prosecution costs, specialist education, and mental health provision). In Europe, only the United Kingdom has calculated the total economic burden, estimated to be ￡735 million in 1996 (WHO 2007).

Young motherhood, perinatal depression, and children’s development

The aim of this chapter is to examine young women’s experience of mental health problems during the perinatal period. We shall argue that women who were young at the time of their transition to parenthood are at elevated risk for perinatal depression, in their first and subsequent pregnancies. Evidence for the impact of perinatal depression on children’s development will be outlined, and we propose that the elevated rates of mental health problems among young mothers may partly account for the increased prevalence of adverse outcomes often seen among their children. However, for these young women and their offspring, the impact of perinatal depression may be compounded by many other social, psychological, and biological risk factors, and young women’s circumstances may exacerbate their own and their children’s difficulties. Therefore any clinical strategies regarding the identification and treatment of depression during the antenatal and postnatal months may need to take into account the age of women, with women bearing children earlier and later than the average presenting different challenges for health professionals. Across the industrialized nations the demographics of parenthood are changing, with both men and women first becoming parents at increasingly older ages (Bosch 1998; Martin et al. 2005; Ventura et al. 2001). In the UK for example, the average maternal age at first birth in 1971 was 23.7 years, compared to the present figure of 29.5 years (ONS 2012). Correspondingly, over the last four decades, birth rates for women aged 30 and over have increased extensively, whilst those for women in their teenage years and early twenties have declined (ONS 2012, 2007). Since the 1970s, the proportion of children born to women aged 20–24 in the UK has been decreasing, with women aged 30–34 years now displaying the highest birth rates (ONS 2010). These changes in the demography of parenthood are not confined to the UK with similar trends toward delayed first births observed across Western Europe (Ventura et al. 2001), the United States (Mirowsky 2002), New Zealand (Woodward et al. 2006) and Australia (Barnes 2003). Thus, a transition to parenthood during adolescence and the early 20s is non-normative for Western women, and the implications of this ‘off-time’ transition (Elder 1997, 1998) for the mother’s and the child’s mental health warrants attention.