0454 CRANIOFACIAL MEASUREMENTS COMBINED WITH PROPORTIONS OF GENETIC ANCESTRY ARE USEFUL TO INFORM OSA SEVERITY

• Published in: Sleep (New York, N.Y.) Vol. 40; no. suppl_1; p. A169
• Main Authors: Mazzotti, DR, Veatch, OJ, Keenan, B, Sutherland, K, Leinwand, S, Oliveira, L, Moura, T, Hirotsu, C, Schwab, R, Cistulli, P, Pack, AI, Tufik, S, Bittencourt, L
• Format: Journal Article
• Language: English
• Published: US Oxford University Press 28.04.2017
• Subjects: Cluster analysis; Gender; Morphology; Sleep apnea
• ISSN: 0161-8105; 1550-9109
• DOI: 10.1093/sleepj/zsx050.453

Abstract Introduction: Craniofacial measurements have been shown to be important predictors of obstructive sleep apnea (OSA). Evidence indicates that these measures may inform OSA severity. Studies also suggest that genetic factors influence both the risk for OSA and craniofacial morphology. It is possible that underlying genetic architecture modifies the relationship between craniofacial morphology and OSA severity. We hypothesized that the proportion of genetic heterogeneity that is attributable to different ancestries would influence this relationship. Methods: Craniofacial photographic measurements, in-lab sleep studies, and genome-wide genotyping data were collected from 454 admixed individuals that were ascertained via the population-based Sao Paulo Epidemiological Sleep Study (EPISONO) cohort. Ancestry proportions were calculated based on four major populations from the 1,000 Genomes Project. We then determined the correlation structure across these ancestry proportions, 40 craniofacial measurements, gender, age, body mass index (BMI), and neck circumference (NC). Additionally, we conducted unsupervised clustering on these data to determine if these traits could identify clinically-meaningful OSA subgroups. Results: Genetic ancestry proportions were significantly correlated and predicted several craniofacial measurements, adjusted for age, gender, BMI and NC. Cluster analysis identified two main subgroups, distinguished by gender, NC, overall craniofacial dimensions, and proportions of genetic ancestry. Individuals in Cluster 1 (N=211) had increased apnea-hypopnea indices (AHI, β=7.6;95%CI=4.4–10.7), oxygen desaturation indices (β=7.4;95%CI=4.3–10.4) and more severe OSA, defined by AHI≥30 (OR=2.8;95%CI=1.7–4.9) compared to Cluster 2 (N=243). Increased lower anterior face volume and nose width, as well as decreased African and American ancestry proportions, were strong independent predictors of being assigned to the more severe OSA cluster. Conclusion: Using craniofacial photographic measurements, genetic ancestry and anthropometric characteristics, we identified clinically-meaningful OSA subgroups. We expect that measurements of lower anterior face volume and nose width may be useful to inform severity of OSA, in lieu of polysomnography. Furthermore, population genetic structure potentially modifies the relationship of craniofacial morphology with OSA severity. We are currently analyzing genomic data to determine if specific variants are associated with the more severe OSA cluster. Support (If Any): AFIP, CNPq and FAPESP (#2014/1259-2).