In-utero and perinatal influences on suicide risk: a systematic review and meta-analysis

Summary

Background

Adverse in-utero and perinatal conditions might contribute to an increased suicide risk throughout the lifespan; however, existing evidence is sparse and contradictory. We aimed to investigate in-utero and perinatal exposures associated with suicide, suicide attempt, and suicidal ideation.

Methods

We did a systematic review and meta-analysis and searched MEDLINE, Embase, and PsycINFO from inception to Jan 24, 2019, for population-based prospective studies that investigated the association between in-utero and perinatal factors and suicide, suicide attempt, and suicidal ideation. Only papers published in English in peer-reviewed journals were considered. Two researchers independently extracted formal information (eg, country, year, duration of follow-up) and number of cases and non-cases exposed and non-exposed to each risk factor. We calculated pooled odds ratios (ORs) with 95% CIs using random-effects models and used meta-regression to investigate heterogeneity. This study was registered with PROSPERO, number CRD42018091205.

Findings

We identified 42 eligible studies; they had a low risk of bias (median quality score 9/9 [IQR 8–9]). Family or parental characteristics, such as high birth order (eg, for fourth-born or later-born vs first-born, pooled OR 1·51 [95% CIs 1·21–1·88]), teenage mothers (1·80 [1·52–2·14]), single mothers (1·57 [1·31–1·89]); indices of socioeconomic position, such as low maternal (1·36 [1·28–1·46]) and paternal (1·38 [1·27–1·51]) education; and fetal growth (eg, low birthweight 1·30 [1·09–1·55] and small for gestational age 1·18 [1·00–1·40]) were associated with higher suicide risk. Father's age, low gestational age, obstetric characteristics (eg, caesarean section), and condition or exposure during pregnancy (eg, maternal smoking or hypertensive disease) were not associated with higher suicide risk. Similar patterns of associations were observed for suicide attempt and suicidal ideation; however, these results were based on a lower number of studies. In meta-regression, differences in length of follow-up explained most between-study heterogeneity (inital I² ranged from 0 to 79·5).

Interpretation

These findings suggest that prenatal and perinatal characteristics are associated with increased suicide risk during the life course, supporting the developmental origin of health and diseases hypothesis for suicide. The low number of studies for some risk factors, especially for suicide attempt and ideation, leaves gaps in knowledge that need to be addressed. The mechanisms underlying the reported associations and their causal nature still remain unclear.

Funding

Horizon 2020 (EU).

Introduction

Suicide is an important cause of mortality worldwide, with an estimated global burden of 800 000 deaths each year,¹ and its prevention is of paramount importance. The need for prevention extends to non-fatal suicide attempt and severe suicidal ideation because both carry serious long-term consequences,² including increased risk of death by suicide.³

Suicide is a complex and multifactorial phenomenon with both proximal (eg, mental disorders, economic breakdowns) and distal (eg, childhood adversities) risk factors.⁴ In keeping with the developmental origins of health and disease (DOHaD) model,4, 5, 6, 7, 8 an increasing number of studies have reported associations of in-utero and perinatal factors with suicide and mental disorders.9, 10 Although the contribution of in-utero and perinatal conditions is well established for a range of non-communicable diseases, such as cardiovascular and metabolic comorbidities,11, 12, 13 evidence is less consistent for psychiatric disorders and suicide.14, 15 For example, previous studies have reported an increased risk of suicide for individuals born with a low birthweight¹⁶ or individuals exposed to maternal smoking during pregnancy,¹⁷ whereas others have been unable to replicate these associations.18, 19 Such inconsistencies might be accounted for by heterogeneity of studies in terms of country of origin, study design, and length of follow-up. Further, because both suicide and most in-utero or perinatal factors (such as low birthweight) are rare at the population level, failure to find an association in some studies might be attributable to the insufficient statistical power, despite large sample sizes. We are unaware of any systematic review of the in-utero and perinatal influences on suicide that use statistical approaches to increase the statistical power (such as meta-analysis).

Research in context

Evidence before this study

We systematically searched MEDLINE without data and language limitations up to March 6, 2018, using the keywords (“suicid*” OR “suicide attempt” OR “attempted suicide” OR “suicidal ideation” OR “suicide thoughts” OR “suicidal behavior” OR “suicidal behaviours” OR “completed suicide” OR “suicide risk” OR “suicidality”) AND (“birth order” OR “parity” OR “multiparity” OR “family size” OR “sibship size” OR “number of siblings” OR “birth weight” OR “underweight” OR “maternal age” OR “mother age” OR “paternal age” OR “father age” OR “parent* age” OR “foetal fetal growth” OR “fetal growth” OR “early life” OR “early-life” OR “perinatal” OR “prenatal”) AND (“longitudinal” OR “population-based” OR “cohort” OR “observational” OR “prospective” OR “registry” OR “register-based”). Most of the retrieved studies investigated risk factors measured during childhood and a few investigated risk factors during the in-utero and perinatal periods, resulting in 20 relevant studies. No systematic review or meta-analysis was identified. Most studies came from high-quality cohorts and population-based registers. Some studies suggested increased suicide risk for some risk factors (eg, increased birth order and low socioeconomic position at birth). However, although similar in quality and design, some studies reported different estimates for the same risk factor; some studies reported similar estimates (eg, similar hazard ratios or odds ratios) but different confidence intervals, so that only some of the studies could conclude for a significant difference between individuals exposed and non-exposed to the same risk factor; some studies had different follow-up periods—ie, some extending until adolescence or young adulthood, others extending until adulthood; almost all studies presented both crude and adjusted estimates, but the choice of the adjustment variables was highly variable across studies. In particular, some studies adjusted for factors occurring between the exposure (eg, low birthweight) and the outcome (eg, suicide), which can bias the estimate of the contribution of the exposure. These differences preclude firm conclusions about the role of these factors on suicide risk.

Added value of this study

To our knowledge, this is the first systematic review and meta-analysis summarising the available evidence about the in-utero and perinatal influences on suicide. We found that, across the high-quality included studies, family or parental characteristics at childbirth (eg, higher birth order, low mother's age, and single motherhood), low socioeconomic position (eg, low parental education and socioeconomic status), and restricted fetal growth (eg, low birthweight and small for gestational age) were associated with increased suicide risk. No evidence for an association was found for risk factors such as advanced father's age, low gestational age, obstetric characteristics, and mother's smoking in pregnancy. Moreover, accounting for the participant's age at follow-up explained the heterogeneity across studies for most comparisons, without changing the study conclusions.

Implications of all the available evidence

Evidence suggests that prenatal and perinatal characteristics are associated with increased suicide risk during the life course, with main influences represented by unfavourable family or parental characteristics and restricted fetal growth. However, more studies are necessary to establish the role of those risk factors on suicide attempt and ideation, extend the evidence to understudied influences, and probe their causal role.

We did a systematic review and meta-analysis of the available evidence on the association between in-utero and perinatal risk factors (including in-utero conditions, maternal-related factors, and socioeconomic conditions at birth) and suicide, suicide attempt, and suicidal ideation. Findings from such a study might provide insight that supports the DOHaD hypothesis for suicide risk, thus informing public health preventive interventions focusing on early maternal and child care.