Healthy pregnancy and prevention of psychosis

Do healthier pregnancies reduce the risk of offspring psychosis? Variants of this question have appeared in recent papers, but with little discussion of how to answer it.

As a starting point, we note that current research on prenatal factors and psychotic disorders is relevant to this question but only addresses it partially and indirectly; that we are not aware of studies that do address it holistically; and that we do not yet know how such studies could be done. We begin by offering a perspective on research on prenatal factors and psychotic disorders. Then we discuss three points that would require consideration before explicitly directing research toward the question at hand.

About 25 years ago, numerous established pregnancy/birth cohorts had already reached adulthood, creating new opportunities for life course investigations1. In the US, for example, investigators launched studies of prenatal factors and psychotic disorders that made use of archived prenatal maternal sera in two large pregnancy cohorts. The development and linkage of national electronic registries in Scandinavia and elsewhere further transformed capacity to investigate prenatal factors and psychotic disorders2. Studies of prenatal exposures based in “natural experiments” also contributed by mitigating sources of confounding that preclude causal inference in traditional observational designs3.

We are presently at the cusp of another leap forward, as national registries are being linked to archived biological data4; pregnancy cohorts of more than 100,000 births are entering the peak age of risk for psychotic disorders, with prenatal genetic and biological data, and ongoing follow‐up5; and natural experiments are being conducted which include neuroimaging as well as diagnoses6.

Relatively strong evidence suggests a role for prenatal infection and nutrition, but prenatal toxic exposures, prenatal stress, and interpregnancy intervals are also viable candidates, to name just a few. New methodologies from epidemiology are increasingly incorporated to strengthen causal inference in these data, meeting challenges such as disentangling the contributions of factors that tend to cluster together due to lifestyle or social conditions. Genomics and population neuroscience are contributing to the converging evidence that prenatal factors matter for psychotic disorders, and yielding insights into mechanisms. We still do not have definitive evidence that a specific modifiable prenatal exposure is a cause of psychotic disorders. There is much room for optimism, however, as new approaches and data bases come to fruition.

As we move closer to definitive results, it becomes important to consider how these results could be incorporated into public health initiatives to promote healthy pregnancy. Some results might yield further evidence for preventive actions already incorporated into healthy pregnancy initiatives, such as recommended vaccinations and nutritional supplements. It seems likely, however, that emerging results will require us to consider public health actions that go beyond these simplest scenarios. Therefore, it would be appropriate in the long‐term to adopt a more holistic public health framework for research. For this purpose, three central points would require consideration: What do we mean by a healthier pregnancy? Should we broaden the offspring outcomes beyond psychotic disorders? What could we gain by focusing on the population distribution of relevant prenatal factors that lie on a continuum?

A universally applicable definition of “healthier pregnancy” is elusive, and any particular measure needs justification for purpose and context. From a life course perspective, characteristics of a pregnancy may be beneficial for some offspring health outcomes and harmful for others. Among many examples, pregnancy characteristics that increase birthweight may reduce offspring risk of psychotic disorders but increase offspring risk of pre‐menopausal breast cancer2, 7, and advanced paternal age at conception may increase risk of psychotic disorders but lower offspring risk of cardiovascular disease8. Moreover, across different contexts, the characteristics and outcomes that need most emphasis will be different.

Neurodevelopmental delays, low cognitive performance, and persistent subclinical psychotic experiences in children are associated with increased risk of subsequent psychotic disorders. These outcomes are manifest earlier and are more common than psychotic disorders; therefore, they are often more amenable to investigation. They have been related to prenatal experiences; however, like for psychotic disorders, the evidence is not definitive. Furthermore, recent work on the structure of psychopathology supports a dimensional transdiagnostic perspective9. From this perspective, preventing these earlier outcomes could substantially reduce risk of psychotic disorders, and probably other psychiatric disorders too, and could have more public health significance. By contrast, we may also find that certain prenatal factors are related to a subgroup of frank psychotic disorders and not to these earlier antecedents; hence the need to investigate the breadth of related outcomes.

Characteristics of a pregnancy may be related to psychotic disorders on a continuum. A large study found that lower birthweight was associated with increased risk of psychotic disorders, but across a broad continuum, implying that a shift in the entire distribution of birthweight (or the causal factor it represents) in the population might do more for prevention of psychotic disorders than reducing the number of low birthweight babies2. Furthermore, across the continuum of birthweight, lower birthweight was associated with all treated psychiatric disorders, not only with psychotic disorders. Caution is needed, however, because the relationship of prenatal factor and outcome may not be linear, but rather J‐shaped or U‐shaped.

We suggest that, alongside the currently dominant approach to research on prenatal factors and psychotic disorders, it could be useful to set the stage for a more holistic program of research on healthy pregnancy and prevention of psychosis. We have highlighted three central questions that could be amenable to research and might be significant for public health interventions. We should bear in mind, however, that the results of such research may offer guidance, but may not provide unequivocal answers.

Finally, we note that, with few exceptions3, studies of prenatal factors and psychotic disorders have been done in high‐income countries. This makes it difficult to generalize any holistic framework to lower‐resource settings, where maternal exposures and conditions affecting pregnancy are different, and access to prenatal care is more limited. Interventions may need to be integrated into broadly conceived programs, such as the Maternal Health Thematic Fund10; and we may need to consider, for example, whether reducing maternal mortality and obstetric fistulas could result in less childhood trauma and thereby benefit offspring neurodevelopment. A global approach to healthy pregnancy and psychosis will depend upon the expansion of research to diverse low‐ and middle‐income country settings.