Prevalence of self-reported mental disorders in pregnancy and associations with adverse neonatal outcomes: a population-based cross-sectional study

Women with a history of severe mental illness may experience relapse or deterioration of their condition in the perinatal period. For example, a prospective investigation found 43% of a cohort of women with a history of major depression relapsed during their pregnancy [3]. In bipolar disorder, risk of pregnancy-associated relapse may be as high as 50% [4]. A specific association exists between bipolar disorder and postpartum psychosis, a severe psychiatric condition with associated risks to the life of the mother and, potentially, her baby if not recognised and managed appropriately. In women with bipolar disorder and a personal or family history of postnatal psychosis the risk increases to approximately 60% [5], although, as for other perinatal mental illnesses, the disorder may also arise de novo. For women with schizophrenia, the postpartum period in particular can be associated with a risk of psychotic relapse [6] for which careful monitoring is required. Other conditions associated with a risk of perinatal relapse include anxiety disorders such as generalised anxiety disorder and obsessive-compulsive disorder [7] as well as eating disorders [8]. Mental disorders frequently co-occur with other problems and stressors in pregnancy such as substance misuse [9] and intimate partner violence (IPV) [10] which can further increase the risk of relapse.

In addition to the impact of pregnancy upon a mother’s mental health (whether in the context of a pre-existing mental disorder or otherwise), mental disorders in pregnancy have been associated with adverse neonatal outcomes in previous studies. Results of two meta-analyses indicate an association between maternal depression and preterm birth, although conflicting results were found for low birth weight [11, 12]. Maternal bipolar disorder has been associated with small-for-gestational-age births [13]. Schizophrenia has been associated with intra-uterine growth restriction, reduced APGAR scores at birth and congenital abnormalities [6] as well as low birth weight [14]. Women with eating disorders are also at risk of delivering low birth weight infants [15]. In a population-based study, self-reported anxiety disorders were associated with low APGAR score, although not with low birth weight or preterm birth [16].

Several factors may influence observed associations between maternal mental disorder and adverse neonatal outcomes. Smoking is more common in people with mental health problems [17] and is associated with adverse neonatal outcomes such as intrauterine growth retardation and placental complications [18]. Alcohol use disorders are frequently co-morbid with other mental illnesses, particularly mood and anxiety disorders [19]. A population-based prospective cohort study found associations between drinking one or more alcoholic drink per day during pregnancy and outcomes of low birth weight and preterm birth [20]. Maternal age is also relevant as women aged > 35 years appear to be at increased risk of preterm birth as well as low birth weight [21]. Pregnancy in adolescence has been associated with risk of premature birth [22] and young mothers may be particularly vulnerable for poor mental health outcomes [23]. Risk of obesity and cardiovascular diseases such as hypertension is increased in people with severe mental illness [24]. This likely relates to an interplay of multiple psychosocial factors as well as the effects of psychotropic medications [25]. In pregnancy, obesity [26] and hypertensive disorders such as chronic hypertension [27] and pre-eclampsia [28] have been associated with adverse neonatal outcomes. Preterm birth and low birth weight have also been associated with low maternal BMI [29, 30] which may occur in eating disorders such as anorexia nervosa. Past or active eating disorders in pregnancy have been associated with low birth weight [15]. Fetal abnormalities in pregnancy have a range of possible contributory causes. In the context of maternal mental disorders, fetal abnormalities may be related in part to use of prescribed [31] and illicit drugs [32] with potential teratogenic effects. Social factors are also pertinent; for example, low socioeconomic status [33] and lone parenting [34] have been linked epidemiologically with low birth weight, and are also associated with mental disorders [35, 36]. Experience of IPV is more common in mothers with mental illness compared to those without mental illness [37] and is in itself associated with adverse outcomes for the neonate [38, 39].

Despite the recognised importance of mental illness in pregnancy, limited data exist regarding the proportion of women presenting to maternity services who have a history of mental disorder. The aims of the present study were two-fold. Firstly, we aimed to determine the prevalence of self-reported severe mental disorders in women presenting to maternity services in Northern Ireland. Secondly, we aimed to assess the relationship between self-reported history of maternal mental disorder and preterm birth, low infant birth weight and low APGAR scores, after adjusting for a range of potential confounding variables. We hypothesised that maternal history of mental disorder would be associated with adverse neonatal outcomes.

This study took the form of a cross-sectional database analysis using data derived from the Northern Ireland Maternity System (NIMATS), an electronic database used by midwives throughout Northern Ireland to record routine clinical data for all women presenting to maternity services. As well as a variety of antenatal, perinatal and postnatal information, all mothers are screened for a history of mental disorders by interview with a midwife at the booking appointment in accordance with UK guidelines [40]. Midwives ask women if they have a history of the following disorders: schizophrenia; bipolar disorder; severe obsessive-compulsive disorder; severe eating disorder; severe depression; and postpartum psychosis. Midwives rely upon women self-reporting and/or information from their general practitioner via the referral to maternity services.

Data were obtained from NIMATS for all women presenting to maternity services in Northern Ireland during the period January 2010 – December 2015. All antenatal and neonatal data were recorded in the same database. Data extraction was performed in collaboration with the Northern Ireland Health and Social Care Honest Broker Service (HBS). All HBS processes are in line with data protection principles, confidentiality requirements and the Information Commissioner’s Office codes of practice. HBS provides a safe and secure environment in which data are processed. The dataset was anonymised before provision to the research team in a ‘safe haven’ setting and all outputs were subject to statistical disclosure control. Researchers who directly accessed the dataset completed Safe User of Research data Environments (SURE) accredited training.

Frequency data were obtained for the number of pregnancies where the mother reported a history of one or more of the following mental disorders on routine screening by midwives at their booking appointment: schizophrenia; bipolar disorder; severe obsessive-compulsive disorder; severe eating disorder; severe depression; and postpartum psychosis. Mental disorders aside from these are recorded in the database as ‘other mental disorder’. Data were also collected regarding antenatal factors including the mothers’ age and body mass index (BMI) at the booking appointment; use of alcohol; smoking status; employment status of the mother; parenting status of the mother (lone parent or not); disclosure of IPV on routine enquiry; presence of fetal abnormality; and presence of hypertensive disorder in pregnancy. Data for antenatal smoking status was based on maternal self-report as either current smoker or current non-smoker. Alcohol status was also based on maternal self-report as either current drinker or non-drinker. Neonatal outcome data included gestational age at birth; infant birth weight; and APGAR score at 1 and 5 min postpartum.

Statistical analysis was performed in SPSS Statistics software version 24 (Armonk, NY: IBM Corp). Logistic regression was used to derive odds ratios for associations between the exposure (self-reported history of at least one mental disorder at the booking appointment) and the following outcomes: low birth weight (defined according to WHO criteria as < 2500 g); preterm birth (defined according to WHO criteria as gestational age < 37 weeks at delivery); and APGAR score < 7 at 1 and 5 min (APGAR score < 7 indicates that the neonate requires medical attention). Analyses were performed on the basis of the mother reporting a history of at least one mental disorder (rather than for each disorder individually), producing a dichotomous exposure variable (history of mental disorder/no history of mental disorder). The assumption of non-multicollinearity was satisfied by inspection of variance inflation factors [41]. We excluded pregnancies with multiple births due to the known association between multiple births and adverse perinatal outcomes [42]. Pregnancies where at least one variable value was missing were excluded from the regression analyses.

For adjusted analyses, covariates were chosen based on prior associations with mental disorder and neonatal outcomes. These included: self-reported antenatal smoking status at booking appointment (smoker/non-smoker); self-reported antenatal alcohol status at booking appointment (drinker/non-drinker); maternal age at booking < 18 or > 35 years (with reference to maternal age 18–35 years); maternal BMI at booking < 18.5 kg/m² or > 25 kg/m² (with reference to maternal BMI 18.5–25 kg/m²); whether or not the mother reported a previously-undisclosed experience of IPV on routine enquiry at booking (new disclosure/no new disclosure); employment status of mother (employed/unemployed); parenting status (lone parent/not lone parent); fetal abnormality registered during pregnancy (present/not present); and hypertensive disorder (pre-eclampsia, chronic hypertension or gestational hypertension) registered during pregnancy (present/not present). For the outcome of APGAR score < 7, preterm birth and low birth weight were included as additional covariates.