Outcomes 50 Years After Preterm Birth: A Golden Opportunity to Reflect on Pathways Toward Thriving

Since the outset of the field decades ago, neonatology has been criticized for advancing survival rates at the cost of comorbid illness and quality of life. The work of Walters et al published in this issue of Pediatrics shows that those fears may have been misplaced.¹ This study describes that compared with term-born controls, surviving preterm-born participants were less likely to have significant cardiovascular events despite having a higher risk for hypertension and similar risk of diabetes, prediabetes, and dyslipidemia. The 2 groups also had similar self-reported general health ratings and functional abilities in middle age.¹ 

The authors are to be commended for bringing data to bear in understanding the long-term health implications of preterm birth on infants in the post-steroid era of neonatal medicine. These positive findings are surprising given the heightened risk for later-life morbidity described in other large, population-based cohort studies of adults born preterm. The high and stable rates of early morbidities in, for example, bronchopulmonary dysplasia,² retinopathy of prematurity,³ and neurosensory impairment⁴ over the past decades reinforce the notion of developmental vulnerability imparted by preterm birth. Perhaps these unexpectedly positive results stem from the relatively mild degree of prematurity in the study cohort. Regardless, we must interpret them in the context of accumulating evidence from developmental biology, epigenetics, and epidemiology that supports the significant role of preterm birth and associated morbidity on life course health development.⁵ 

This paper illustrates several of the methodological challenges of conducting longitudinal follow-up that threaten study validity. Participant attrition and censoring due to mortality pose significant challenges to interpreting these results. It is widely understood that people electing to participate in longitudinal studies differ in many ways from those lost to follow-up,⁶ often on socioeconomic factors⁷ inextricably linked with health outcomes. Many such characteristics vary with time, often because of prematurity-related morbidity. Preterm birth itself is associated with lower annual income, lower upward mobility, and higher downward mobility in adulthood compared with term-born peers.⁸ That participants and nonparticipants in this study share a similar profile on a limited list of baseline factors tells us little about how the groups compare on key health-related social drivers over the span of 50 years.

The role of mortality is of particular concern in longitudinal studies following cohorts into adulthood. Individuals with poorer health have higher mortality risk at any given time, a risk that increases with advancing age. In this cohort, the survival advantage of the term-born control group abated after infancy, with a higher all-cause mortality rate compared with that of the group born preterm. Survivors in either group are definitionally healthier than those lost due to nonsurvival. This effect may be amplified among those born preterm because their health may be monitored more closely by caregivers in childhood and by themselves as adults. Censoring due to mortality introduces bias that makes drawing meaningful conclusions on health differences between the 2 groups of survivors difficult.

The use of a composite outcome of cardiovascular events, risk factors, and death poses additional challenges. Investigators frequently choose composite outcomes of death/morbidity due to the competing risks of mortality and nonterminal morbidities such as neurodevelopmental impairment in preterm populations. The outcomes for this study are not mutually exclusive but rather occur in a semicompeting paradigm in which an individual can experience either, both, or neither outcome. When aged 50 years, a participant has 4 possible outcomes: alive and morbidity-free, alive and experiencing 1 or more cardiovascular morbidities, dead from a cardiovascular disorder, or dead from an unrelated disease (eg, cancer). Failure to account for these multiple outcomes in a jointly modeled composite variable can lead to inaccurate risk estimates.⁹ Applying advanced methods to address the effects of semicompeting risks in longitudinal cohort studies will help answer important questions about health outcomes for infants born preterm.

The fatalistic premise that gestational age alone determines future health risk is challenged by the breadth of outcomes observed among surviving preterm infants. In a recent review, we outline the role of neonatal multimorbidity experienced by preterm infants as both reflective of and contributing to underlying vulnerability to age-related health and developmental problems over time.¹⁰ Although connections between neonatal and adult multimorbidity and their impact on functional capacity are only partly defined, they may play an important role in discriminating those at greatest risk. The tenets of the life course health development framework further underscore the importance of early life experiences and the role of complex biological and environmental systems in determining health.⁵ Understanding the multilevel contextual influences on health and development and the interconnections between preterm birth and multimorbid conditions across the life span requires significant investments in population-level data collection; theory-based, hypothesis-driven research questions; and methodological strategies designed for analyzing complex systems.