Prenatal complications and neurodevelopmental outcomes in offspring: interactions and confounders

Reproductive psychiatry, a more narrowly concentrated field within women's mental health, focuses on the research and treatment of psychiatric illness related to the female reproductive life cycle, especially during times of reproductive transitions including the menstrual cycle, infertility, and perinatal and perimenopause periods. In the current issue of Acta Psychiatrica Scandinavica, several articles reflect the breadth of research within reproductive psychiatry (1, 2) and the related field of maternal–child health (3-5). Rotberg and colleagues examined the association of the birth of a child with a major congenital anomaly with later risk of new-onset maternal psychiatric disorders (1). Their evidence of an increased risk of new-onset psychiatric illness, which peaked in the first-year postpartum yet remained elevated for an extended time period, highlights a need for adequate screening in this high-risk population. Larsen et al. examined oral contraceptive exposure in healthy women and its association with brain serotonin 4 receptor binding compared with non-users (2). Their findings add valuable insight into a possible molecular link between oral contraceptive use and risk of depression in women. Focusing on prenatal exposures to children, Brannigan and colleagues examined associations and mediators between prenatal maternal smoking, birthweight, and persistent offspring psychiatric symptoms (3). Their findings extended the literature by demonstrating that the association between prenatal maternal smoking and child and adolescent internalizing symptoms was mediated by birthweight and emphasized the need for perinatal smoking cessation programmes. Within this research realm, two additional articles focused on the neurodevelopmental impact of preeclampsia (4, 5). Preeclampsia is classified into hypertensive disorders of pregnancy, for which a negative impact on maternal and offspring health has been previously suggested (6, 7), although there are striking inconsistencies in the literature (7). In this issue of Acta Psychiatrica Scandinavica, the study of Maher and colleagues reported an increased risk of attention-deficit/hyperactivity disorder (ADHD) in infants born small for gestational age (SGA) to mothers with preeclampsia (4); mainly characterized by inattention, hyperactivity, and impulsivity, ADHD is a common neurodevelopmental disorder with major impact on quality of life and functional performance (8). Combining different types of national register data, Maher and colleagues estimated an increase of 15% in ADHD risk for infants exposed to preeclampsia, 43% for infants exposed to preeclampsia and having been born SGA, and 32% for subjects with SGA only (4). Therefore, we may hypothesize synergistic effects between SGA and preeclampsia on ADHD risk. This notion of synergism provides some novel insight into assessing the risk of neurodevelopmental sequelae of preeclampsia. Previous findings from a meta-analysis of case–control and cohort studies had reported the association between ADHD and preeclampsia alone, although effect sizes were slightly higher (30%) in the pooled sample (7). The findings of this meta-analysis also suffered from substantial heterogeneity (7). In fact, because of the observational nature of these studies, confounders introduce a well-known source of bias (9). On the other hand, the use of specifiers, such as SGA in the design of Maher's study, points out that the ADHD–preeclampsia association may appear in a subgroup of dyads with preeclampsia. Maher and colleagues also performed several sensitivity analyses, the most striking finding of which suggested a ADHD risk increase of 74% in preeclampsia-exposed preterm birth offspring (<34 gestational weeks) (4). An additional strength of the study of Maher and colleagues was the inclusion of a sibling-matched analysis to account for familial and genetic confounders. The results of the sibling-matched analysis replicated the findings in the population-based cohort providing similar effect sizes. However, the estimated E-values, as an estimate of unmeasured confounders, implied that residual confounding may essentially underlie the increased ADHD risk in subjects exposed to preeclampsia alone, but less likely in subjects exposed to preeclampsia and having been born SGA. In light of the findings of the study of Maher and colleagues, we consider two principal implications: First, these findings point out that the design of future studies assessing the effects of hypertensive disorders of pregnancy, and specifically preeclampsia, on neurodevelopmental outcomes in offspring, will need to assess interplays of maternal and fetal conditions to provide more accurate estimates of associations. Inserting more granular data instead of proxies in these analyses will more sufficiently tackle residual confounding that would potentially explain away reported associations. This knowledge will advance the understanding of related pathways including moderators/modifiers. For example, in Maher et al's sample residual confounding was likely to account for the association of preeclampsia with an increased ADHD risk; when assessing the ADHD–preeclampsia association in subjects born SGA, it was unlikely that unmeasured confounding was implicated (4). Second, the Maher findings identified a subgroup of mother–fetus dyads at high risk of ADHD in offspring, that is, women with preeclampsia giving birth to SGA infants. It is important to understand the risk profile of this dyad, where the impairment of fetal development in light of preeclampsia takes place. Indeed, there may be pathophysiological underpinnings linking preeclampsia and intrauterine growth restriction (for which SGA is frequently used as a proxy) in these dyads. On a related note, the co-occurrence of preeclampsia and SGA is two-fold higher in women with early-onset compared with late-onset preeclampsia (10). Indeed, evidence from obstetric assessments outlines a unique profile for pregnant women developing preeclampsia and giving birth SGA (11, 12); apart from elevated uric acid levels throughout pregnancy (12), the differences in hemodynamic profiles in pregnant women with preeclampsia and SGA vs. with preeclampsia without SGA have been reported (11). These differences included a low cardiac output and high systemic vascular resistance affecting uteroplacental circulation (11); based on this pattern, the authors proposed a symptom continuum with co-occurrence of preeclampsia and SGA reflecting more severe clinical pictures. In these severe clinical forms, alterations in uteroplacental circulation seem to impact fetal development (11, 13). Accordingly, future data will need to study better characterized subcohorts of women with preeclampsia using specifiers to unravel most-at-risk subgroup profiles. Further research is also required to validate (or reject) the concept of physiological continuum underlying preeclampsia effects on offspring. The notion of 'dose–response' effects with severe forms of preeclampsia having more deleterious effects is also indirectly supported by another analysis performed by Maher and colleagues in the same issue of Acta Psychiatrica Scandinavica (5); the authors analyzed register data in an intergeneration fashion suggesting an ADHD risk increase of 34% in subjects exposed to preeclampsia in both grandmother and mother compared with 8% and 23% for grandmother only and mother only, respectively (5). Regardless of involved genetic correlates (14) and although SGA was not inserted in the analysis, this finding may support some intergenerational patterns for the mechanisms linking preeclampsia-associated and neurodevelopmental outcomes. This analysis also reported increased risks of autism spectrum disorders in subjects, both the mother and grandmother of who had developed preeclampsia (5). The lack of specificity regarding neurodevelopmental effects of preeclampsia has been consistently reported in previous works (7). Nevertheless, the determinants of the type of neurodevelopmental disorders associated with preeclampsia remain poorly understood (15). In summary, the Swedish register-based study of Maher and colleagues highlights the interplay of preeclampsia and restricted intrauterine growth on the risk of ADHD using a sophisticated methodological approach. Future research will need to develop similar approaches, but also integrate data on maternal mental health; in fact, preeclampsia has been previously related to increased risk of postpartum depression (16). Therefore, sophisticated interactive models assessing co-occurrence of mental and somatic burden in perinatal women may facilitate the better understanding of the pathways that may regulate the impact of maternal mental and somatic conditions on fetal short- and long-term development. Ultimately, this knowledge will enhance the role of reproductive psychiatry in the early identification of mother–infant dyads at risk of health complications. Dr. Deligiannidis receives grant support from the U.S. National Institutes of Health (grants R01MH120313 and R01MH118269), the Feinstein Institutes for Medical Research, Vorso Corporation, and Sage Therapeutics. She serves as a consultant to Sage Therapeutics. Dr. Schoretsanitis reports no conflicts of interest.