LAUSR

The architect Louis Sullivan wrote that “form ever follows function,” highlighting the interrelationship of these two and the importance of form as an indicator of underlying factors and purpose. In the case of the recent work by Luo et al., in which facial dysmorphology features are linked with prenatal exposure to the class of chemicals known as perand polyfluoroalkyl substances (PFAS), formmay follow cause, and perhaps be used as an endophenotypic marker of neurodevelopmental outcomes or susceptibility to exposure. PFAS, often referred to as “forever chemicals,” are a broad group of persistent and widely used chemicals characterized by a carbon–fluorine chain. Many studies have been published over the past 5–10 years investigating exposure to these chemicals in utero, a known susceptibility period for potential teratogenic effects of environmental exposures, in association with a range of child health outcomes. As noted by Luo et al., the literature addressing prenatal PFAS exposure and child neurodevelopment is quite mixed, with no clear consensus on the effects of these chemicals on neurodevelopmental outcomes in childhood, but several suggestions of adverse effects with higher exposure levels. The study by Luo et al., conducted within the Danish National Birth Cohort and including consideration of mixture models, not only builds off this prior work, but also adds a new angle to the question of associations of these chemicals with child neurodevelopmental outcomes. This is done by focusing on an outcome that can be considered as an endophenotype, or a quantitative biological trait. In addition to the need to clarify inconsistencies across studies of neurodevelopmental outcomes, the work by Luo et al. raises the additional question as to whether PFAS exposure may be linked, perhaps in more subtle ways, with other phenotypic outcomes. It is also possible that the associations with neurodevelopmental outcomes may be observable only in those with apparent dysmorphology, as an indicator of their susceptibility to this environmental exposure. These remain open questions for the field to address. Dysmorphology and atypical facial features have long been linked with neurologic deficits and considered indicators of toxicant exposure or genetic syndromes, as demonstrated most classically in fetal alcohol syndrome and Down syndrome, respectively. Yet, less severe differences in facial features, or variation in these characteristics, may be linked with environmental exposures and have also been purported to indicate underlying neurologic deficits. Interestingly, a shortened palprebal fissure length, the primary facial feature demonstratingan association in theworkbyLuo et al., has been suggested as an indicator of impaired development of regions of the forebrain, of relevance for a range of neurodevelopmental conditions. For example, autism spectrum disorder, a neurodevelopmental condition characterized by deficits in social communication and presence of repetitive behaviors, has been linked with alterations in forebrain regions including the amygdala, hippocampus, hypothalamus, and cerebellum (among others). Fragile X syndrome, a genetic condition with a distinctly increased rate of autism diagnosis, is marked by unique facial characteristics, although evidence has suggested neuroanatomical findings may be distinct in fragile X vs. autism. Ongoing work has examined dysmorphology in autism, suggesting facial dysmorphology is relatively common (recently estimated at 17% of cases in a wellcharacterized study) and not entirely due to genetic conditions. The finding that variation in facial features thought to be markers of altered neurodevelopment may be linked with environmental risk factors therefore provides intriguing additional context for our pursuit of better understanding of the etiology of complex neurodevelopmental conditions. Future extensions of this work should use quantitative metrics of dysmorphology, such as geometric morphometrics for 3dimensional measures, in order to more fully assess the range of associations across these traits. In addition, the use of causal mediation models (although not without assumptions and limitations) will be of interest in determining whether and how these features relate to later neurodevelopmental outcomes in the same children. It will also be of interest to consider joint effects with other environmental factors, including following up on the suggestion of Luo et al. of modification by alcohol intake. An additional area of need is determining whether and how these facial features may relate to continuous latent traits of neurodevelopmental outcomes, and if more subtle dysmorphology holds public health relevance. One caveat for consideration is the need for sensitivity surrounding the study of facial features and communication regarding results from such work, so that variation itself is not uniformly construed in a negative light or conflated with variation related to race or ethnicity. Finally, broadening our consideration of endophenotypes, using multiple continuous measures, and including endophenotypes across domains (given the high degree of comorbidity across the neurodevelopmental spectrum) are likely to be fruitful avenues for better understanding the ways in which environmental exposures, such as PFAS, may contribute to the etiology of neurodevelopmental outcomes and phenotypes.