Abstract
IMPORTANCE: In addition to confirmed prenatal alcohol exposure and severe neurodevelopmental deficits, three cardinal facial features are included in the diagnostic criteria for fetal alcohol spectrum disorder. It is not understood whether subtle facial characteristics occur in children without a diagnosis but who were exposed to a range of common pregnancy drinking patterns and, if so, whether these persist throughout childhood. OBJECTIVE: To determine whether subtle changes in facial shape with prenatal alcohol exposure found in 12-month-old children were evident at age 6 to 8 years using extended phenotyping methods and, if so, whether facial characteristics were similar to those seen in fetal alcohol spectrum disorder. DESIGN, SETTING, AND PARTICIPANTS: In a prospective cohort study in Melbourne, Victoria, Australia, commencing in July 2011 with follow-up through April 2021, pregnant women were recruited in the first trimester from low-risk, metropolitan, public maternity clinics over a period of 12 months. Three-dimensional craniofacial images from 549 children of European descent taken at age 12 months (n = 421 images) and 6 to 8 years (n = 363) were included. Data analysis was performed from May 2021 to October 2024. EXPOSURES: Predominantly low to moderate prenatal alcohol exposure in the first trimester or throughout pregnancy compared with controls without prenatal alcohol exposure. MAIN OUTCOMES AND MEASURES: Following hierarchical facial segmentation, phenotype descriptors were computed. Hypothesis testing was performed for 63 facial modules to analyze different facial parts independently using principal component analysis and response-based imputed predictor (RIP) scores. Comparison was made with a clinical discovery sample of facial images of children with a confirmed diagnosis of partial or full fetal alcohol syndrome. RESULTS: A total of 549 children took part in the 3-dimensional craniofacial image analysis, of whom 235 (42.8%) contributed an image at both time points. Time 1 included 421 children, comprising 336 children (159 [47.3%] female) with any prenatal alcohol exposure and 85 control children (45 [52.9%] female); time 2 included 363 children, comprising 260 children with any prenatal alcohol exposure (125 [48.1%] female; mean [SD] age, 6.9 [0.7] years) and 103 control children (53 [51.5%] female; mean [SD] age, 6.8 [0.7] years). At both time points, there was consistent evidence for an association between prenatal alcohol exposure and the shape of the eyes (eg, module 15: RIP partial Spearman ρ, 0.19 [95% CI, 0.10-0.29; P < .001] at 6-8 years) and nose (eg, module 5: RIP partial Spearman ρ, 0.19 [95% CI, 0.09-0.27; P < .001] at 6-8 years), whether exposure occurred only in trimester 1 or throughout pregnancy. Facial variations observed differed from those in the clinical discovery sample. CONCLUSIONS AND RELEVANCE: Low to moderate prenatal alcohol exposure was associated with characteristic changes in the face, which persisted until at least 6 to 8 years of age. A linear association between alcohol exposure levels and facial shape was not supported.