Seung Hun Lee1, Moo Il Kang, Seong Hee Ahn, Kyeong-Hye Lim, Gun Eui Lee, Eun-Soon Shin, Jong-Eun Lee, Beom-Jun Kim, Eun-Hee Cho, Sang-Wook Kim, Tae-Ho Kim, Hyun-Ju Kim, Kun-Ho Yoon, Won Chul Lee, Ghi Su Kim, Jung-Min Koh, Shin-Yoon Kim. 1. Division of Endocrinology and Metabolism (S.H.L., S.H.A., K.-H.L., B.-J.K., G.S.K., J.-M.K.), Asan Medical Center, University of Ulsan College of Medicine, Seoul 138-736, Korea; Department of Endocrinology and Metabolism (M.I.K., K.-H.Y.), The Catholic University of Korea, College of Medicine, Seoul 137-701, Korea; DNA Link (G.E.L., E.-S.S., J.-E.L.), Seoul 138-736, Korea; Department of Internal Medicine (E.-H.C., S.-W.K.), Kangwon National University College of Medicine, Chuncheon 200-722, Korea; Skeletal Diseases Genome Research Center and Department of Orthopedic Surgery (T.-H.K., H.-J.K., S.-Y.K.), Kyungpook National University School of Medicine, Daegu 702-701, Korea; and Department of Preventive Medicine (W.C.L.), The Catholic University of Korea, College of Medicine, Seoul 137-701, Korea.
Abstract
CONTEXT: Osteoporotic fracture risk is highly heritable, but genome-wide association studies have explained only a small proportion of the heritability to date. Genetic data may improve prediction of fracture risk in osteopenic subjects and assist early intervention and management. OBJECTIVE: To detect common and rare variants in coding and regulatory regions related to osteoporosis-related traits, and to investigate whether genetic profiling improves the prediction of fracture risk. DESIGN AND SETTING: This cross-sectional study was conducted in three clinical units in Korea. PARTICIPANTS: Postmenopausal women with extreme phenotypes (n = 982) were used for the discovery set, and 3895 participants were used for the replication set. MAIN OUTCOME MEASURE: We performed targeted resequencing of 198 genes. Genetic risk scores from common variants (GRS-C) and from common and rare variants (GRS-T) were calculated. RESULTS: Nineteen common variants in 17 genes (of the discovered 34 functional variants in 26 genes) and 31 rare variants in five genes (of the discovered 87 functional variants in 15 genes) were associated with one or more osteoporosis-related traits. Accuracy of fracture risk classification was improved in the osteopenic patients by adding GRS-C to fracture risk assessment models (6.8%; P < .001) and was further improved by adding GRS-T (9.6%; P < .001). GRS-C improved classification accuracy for vertebral and nonvertebral fractures by 7.3% (P = .005) and 3.0% (P = .091), and GRS-T further improved accuracy by 10.2% (P < .001) and 4.9% (P = .008), respectively. CONCLUSIONS: Our results suggest that both common and rare functional variants may contribute to osteoporotic fracture and that adding genetic profiling data to current models could improve the prediction of fracture risk in an osteopenic individual.
CONTEXT: Osteoporotic fracture risk is highly heritable, but genome-wide association studies have explained only a small proportion of the heritability to date. Genetic data may improve prediction of fracture risk in osteopenic subjects and assist early intervention and management. OBJECTIVE: To detect common and rare variants in coding and regulatory regions related to osteoporosis-related traits, and to investigate whether genetic profiling improves the prediction of fracture risk. DESIGN AND SETTING: This cross-sectional study was conducted in three clinical units in Korea. PARTICIPANTS: Postmenopausal women with extreme phenotypes (n = 982) were used for the discovery set, and 3895 participants were used for the replication set. MAIN OUTCOME MEASURE: We performed targeted resequencing of 198 genes. Genetic risk scores from common variants (GRS-C) and from common and rare variants (GRS-T) were calculated. RESULTS: Nineteen common variants in 17 genes (of the discovered 34 functional variants in 26 genes) and 31 rare variants in five genes (of the discovered 87 functional variants in 15 genes) were associated with one or more osteoporosis-related traits. Accuracy of fracture risk classification was improved in the osteopenicpatients by adding GRS-C to fracture risk assessment models (6.8%; P < .001) and was further improved by adding GRS-T (9.6%; P < .001). GRS-C improved classification accuracy for vertebral and nonvertebral fractures by 7.3% (P = .005) and 3.0% (P = .091), and GRS-T further improved accuracy by 10.2% (P < .001) and 4.9% (P = .008), respectively. CONCLUSIONS: Our results suggest that both common and rare functional variants may contribute to osteoporotic fracture and that adding genetic profiling data to current models could improve the prediction of fracture risk in an osteopenic individual.