The study by Bublitz et al describes a surprisingly strong crude odds ratio of 3.82 between experiencing adoption or foster care placement before the age of 18 years and the risk of delivering a preterm infant at <37 completed weeks of gestation. This relationship demonstrates positive confounding after adjustment for sociodemographic, psychosocial, health behavioral, and biomedical conditions with a fully adjusted odds ratio of 4.22 (adjusted odds ratio, 4.22; 95% confidence interval, 1.17–15.29). If this relationship can be confirmed in other studies and in different populations, this finding may provide evidence in support of one or more mechanisms that lead to preterm birth.

However, although this retrospective study appears to be conceptually sound with statistically significant results, the number of participants was quite small. Three hundred two low-income, but otherwise apparently low-risk women, oversampled for depression and smoking, whose data was drawn from 2 previous studies of maternal mood and behavior and fetal and infant development, were included in the present analysis. Overall, 39 women (13%) had a history of adoption or foster care placement, and 26 women (9%) had a preterm birth. Six percent of the women (n = 16) without a history of placement (n = 263) delivered preterm infants, compared with 18% of those women (n = 7) with a history of adoption or placement (n = 39). Confirmation of these results in other populations with larger sample sizes is crucial. Additionally, little information is provided in this study to help understand the obstetric outcome. Specifically, no information is provided regarding the gestational age or birthweight distribution of the preterm births in women with and without placement history or whether the preterm births were spontaneous or indicated or which obstetric conditions, if any, were associated with the preterm births. Any further study of this issue should define clearly the preterm outcomes by gestational age, by birthweight, by whether the preterm births were indicated or spontaneous, and by likely cause.

Even though this article suffers from small sample size and poor specification of the dependent variable, the findings are intriguing. Assuming this relationship is confirmed, 3 possible mechanisms may be at play. Given that statistical adjustment for a number of psychosocial and behavioral attributes only strengthened the association between placement and preterm delivery, it does not appear likely that this relationship is simply explained by mediation in cases in which maternal depression/anxiety/stress, smoking, or alcohol use arise from the experience of placement and, in turn, increase the risk of preterm birth. So, one can probably rule out this relatively simple explanation. Next, adults who experience out-of-home placement in childhood may be disproportionally more likely to have experienced a suboptimal uterine environment and thus be born preterm. As noted by the authors, information regarding the participant’s gestational age at birth and whether she was exposed to potentially damaging conditions in utero was not obtained. Therefore, the potential generational influences on the risk of delivering preterm infants could not be evaluated. Having this type of information would help to determine whether some characteristic of the mother, such as drug or alcohol exposure in utero or a history of being preterm, might explain partially the reason for both the placement and the increased risk of preterm birth in the pregnancy under study. Finally, a third potential mechanism may be alterations in maternal stress physiology that is related to both adverse uterine environment and poor parenting/maternal attachment in early life.

Evidence suggests that mothers who experience stress during pregnancy have an increased risk of preterm delivery with mechanistic work that focuses on alterations in the maternal neuroendocrine and immune systems that lead to increased inflammation. Recent work has extended this association between stress and adverse birth outcomes to include exposure to preconception stress and risk of preterm birth. For example, Witt et al explored the association between maternal exposure to stressful life events before conception and risk of preterm birth. These authors found a significant interaction between maternal age and exposure such that young women who reported at least 1 preconception stressful event had a 4-fold increased risk for preterm birth.

Experiencing childhood stress has also been shown to have a long-term impact on the hypothalamic-pituitary axis (HPA)and is associated with alterations in corticotrophin-releasing hormone and cortisol production even during subsequent pregnancy. Substantial evidence for an effect of corticotrophin-releasing hormone and cortisol levels on preterm birth exists, although whether altered cortisol levels are the explanation for the increase in the preterm births in women with a history of adoption or foster care that is noted in this study is unknown. Future studies therefore will need to explore the influence of childhood placement on the long-term functioning of the HPA and then determine whether these changes, if any, influence the rate of preterm birth.

In animal models, there is clear evidence that rat pups that are raised by dams that exhibit “nurturing” maternal behaviors (increased licking and grooming and arched back nursing) have a more regulated biologic response to external stressors and less fearful/anxious behavior as adults compared with pups that are raised by less caring dams. Female offspring, as adults, demonstrate the same behavior exhibited by their mothers and thus pass along either a regulated or a dysregulated HPA axis and patterned fear and anxiety behavior. Fascinatingly, if at birth, female rat pups are cross-fostered (for example from a caring to a less caring dam), there is clear biologic evidence of acquisition of a more dysregulated HPA axis and increased fearful and anxious behavior. This acquired HPA axis reactivity and fearful behavioral pattern is retained into adulthood.

Based on the animal models and some association studies, a leading candidate to explain the relationship between adoption and foster care placement and risk of subsequent preterm birth is stress during gestation and/or childhood. Several studies suggest that negative childhood experiences may have long-term consequences for pregnancy outcomes as an adult; one study specifically notes that children with a history of sexual abuse have a greater risk of preterm birth. Bublitz et al suggest a potential causal explanation for their findings, in that children who are placed out of the home may experience extreme stress before placement including poverty, neglect, and abuse. Even without this history, separation from one or both parents is likely, by itself, to be highly stressful.

Finally, over the past decade, there has been a tremendous growth in research that has explored how social conditions or social exposures produce measurable, predictable, long-lasting, and even heritable changes in physiology that leads to alterations in health and well-being. This article contributes evidence that it is increasingly inappropriate to ignore social context when studying risk factors for and potential causes of adverse reproductive outcomes.

Although these findings are interesting and suggestive, it is not clear that screening (specifically for a history of placement) is warranted in clinical practice. Because this is but one study to describe an epidemiologic association with small numbers and an uncertain causal mechanism, we believe that substantial confirmatory work about this relationship in particular, as opposed to a more global understanding of a woman’s lifetime experiences, is necessary before the initiation of any study of interventions to reduce preterm birth in this specific group of women.