In this cohort of 1.4 million women, preterm delivery was associated with a higher risk of hospitalization for mental illness up to 32 years after pregnancy. Risk of hospitalization was greatest within 2 years of delivery, but persisted throughout follow-up. Associations were present with all types of mental disorders, including depression, stress and anxiety disorders, and personality disorders. Women with moderate to late preterm birth were as much at risk of mental disorder hospitalization as women with extreme and very preterm birth. For some mental disorders, risks were even greater around 34 weeks of gestation or moderate preterm delivery. The findings suggest that women who deliver preterm may be at risk of hospitalization for a range of mental disorders in the short and long term, even when delivery occurs at moderate to late gestational ages.

Few studies have addressed moderate preterm birth in maternal mental health. Most of the literature focuses on depression or stress disorders in women with extreme or very preterm delivery [3, 22, 23]. Very preterm delivery is associated with 1.4 to 2.9 times the risk of postpartum depression compared with term delivery [3, 22]. Up to 22% of women who deliver before 30 weeks develop posttraumatic stress the first year postpartum [23]. Nevertheless, effects may not be limited to extreme or very preterm birth. A study of 91 mothers who delivered moderate to late preterm reported that symptoms of depression, anxiety, and post-traumatic stress were elevated up to 6 months later [6]. In population-based studies from Sweden and the United States, women who delivered between 32 and 36 weeks of gestation had 1.2 times the risk of postpartum depression compared with term [7, 22]. A study of 60 participants suggested that women who deliver moderate to late preterm have more postpartum depressive symptoms than women who deliver very preterm [8].

In our study, risk of maternal mental disorders was greatest for deliveries around 34 weeks of gestation, which is consistent with the few available studies [6,7,8, 22]. However, the association with stress, anxiety, and psychotic disorders was also present at lower gestational ages, suggesting that preterm birth may affect anxiety-linked mental health disorders differently. Extremely premature infants have a high prevalence of morbidity and developmental delay [1], which may be stressful for caregivers. While moderate to late preterm birth is associated with less morbidity, most support programs do not cover these births [24]. These factors may explain why any degree of preterm birth is associated with stress, anxiety, and psychotic disorders. In contrast, inadequate social support for mothers with moderate to late preterm birth may be more likely to lead to other types of mental disorders.

Longitudinal studies are conflicting for outcomes after the postpartum period. In a cohort of 214 mothers, very preterm delivery was associated with anxiety and depressive symptoms 7 years after birth [5]. Another study found that women who delivered very preterm were at risk of anxiety and depression 13 years later [10]. Three separate analyses of a total of 726 women found no association between very preterm delivery and depression, stress, or anxiety up to 25 years later [9, 25, 26]. The authors suggested that the impact of preterm delivery faded by the time children reach adolescence [9, 26]. However, the number of women studied was low. Our analysis of 1.4 million mothers suggests that women who deliver preterm remain at risk of mental illness up to 32 years later. The effects weakened over time but did not disappear.

Research has focused more on maternal depression and stress than other mental disorders [3,4,5,6,7,8,9,10, 22, 23, 25, 26]. Two Swedish cohort studies reported that preterm delivery increased the risk of psychosis 3 months postpartum [27, 28]. In one report, preterm delivery was not associated with suicide attempts the first year postpartum [29]. In contrast, we found that preterm birth was associated with up to 1.5 times the risk of most mental disorders later in life. The low number of cases, narrow range of mental disorders, and short follow-up may explain why previous studies differed from ours.

Preterm birth was associated with stress, anxiety, and personality disorders shortly after delivery, while the impact on depression, psychotic, and bipolar disorders appeared only 10 years later. Immediate consequences of preterm birth, including infant morbidity and need for neonatal intensive care, may be associated with maternal stress and anxiety in the short term. Childhood health and behavioral disorders that occur later may impact maternal depression, psychotic, and bipolar disorders in the long term. Data indicate that parental distress due to caring for a preterm infant may change as a child grows [5]. Respiratory and gastrointestinal disorders may be more concerning during infancy, while neurodevelopmental disability and behavioural problems may become prevalent during childhood and adolescence [2].

Maternal complications leading to preterm delivery may also be associated with mental disorders. Obesity, gestational diabetes, and preeclampsia increase the risk of preterm delivery [15] and are associated with postpartum mental illness [14, 17, 30]. Severe maternal morbidity, including severe preeclampsia, cerebrovascular accidents, and other life-threatening complications, is associated with 1.7 times the risk of mental illness the first year postpartum [16]. Pregnancy complications share immuno-inflammatory pathophysiology with mental illness, including dysregulation of the hypothalamic-pituitary-adrenal axis [17, 30, 31]. In our study, preterm birth was associated with maternal mental health despite adjustment for pregnancy complications. Thus, it is unlikely that maternal morbidity fully explains the association between preterm delivery and mental disorders.

This study has limitations. The outcome included severe maternal mental illness that required hospitalization, not mild problems that never led to hospitalization. We excluded women with mild or severe mental illness at the start of follow-up but could not identify mild disorders that were not reported. As we used administrative data, nondifferential misclassification may have occurred due to coding errors. We were not able to investigate associations for planned versus spontaneous preterm delivery. We could not account for childhood morbidity or psychiatric comorbidity. Although we controlled for several confounders, we cannot rule out residual confounding as we had no data on ethnicity, marital status, social support, psychotropic medications, or psychotherapy. A confounder would have to be associated with at least two times the risk of preterm birth and mental illness to fully confound our associations [32]. Quebec is a multicultural province, but studies across diverse countries, cultures, and ethnicities are needed before generalizing at a national level.