1716 EXPERIMENTAL AND THERAPEUTIC MEDICINE 20: 1716-1724, 2020 A single genetic locus associated with pediatric fractures: A genome-wide association study on 3,230 patients ROOPE PARVIAINEN1, SINI SKARP2,3, LINDA KORHONEN1, WILLY SERLO1, MINNA MÄNNIKKÖ2 ��� JUHA-JAAKKO SINIKUMPU1 1D�p���m��� �� C������� ��� A����������, O��� C�������� F������� ��� Sp���� I�j��y S���y, R������� U��� ��� P���������, P�������� N�������y, P�������� Surgery, C���� P�y������y, D��m������y, C������� G�������, O��������� ��� Gy�������y, O����������y������y ��� Op�����m����y (PEDEGO), O��� M������ R������� C����� (MRC), U��������y �� O��� ��� O��� U��������y H��p����, FI-90029 O���; 2N������� F������ B���� C�����, F�����y �� M�������, U��������y �� O���; 3C����� ��� L��� C����� H����� R�������, F�����y �� M�������, U��������y �� O���, FI-90014 O���, F������ Received July 13, 2019; Accepted April 29, 2020 DOI: 10.3892/etm.2020.8885 Abstract. The�� understanding������������� ��of the��� biological���������� and��� environ�������- RNA 1 (PROSER2-AS1) and PROSER2, thus suggesting that mental risk factors of fractures in pediatrics is limited. Previous these m�y be novel candidate genes associated with the risk of studies have reported that fractures involve heritable traits, but pediatric fractures. the genetic factors contributing to the risk of fractures remain elusive. Furthermore, genetic influences specific to immature Introduction bone have not been thoroughly studied. Therefore, the aim of the present study was to identify genetic variations that are The incidence of fractures is highest �mong the young and associated with fractures in early childhood. The present study older p�pulations (1). O� note, fractures in pediatric patients used � prospective Northern Finland Birth Cohort (year 1986; are common injuries that result in pain, as well as short‑ and �=9,432). The study p�pulation was comprised of 3,230 cohort long‑term morbidity; it has been estimated that 27-50% of all m�mbers with available genotyp� data. A total of 48 m�mbers of individuals �xperience � fracture prior to turning 18 years the cohort (1.5%) had in‑hospital treated bone fractures during old (2,3). Goulding et al (4) estimated the risk of recurrent their first 6 years of life. Furthermore, individuals without fracture to be 12-20% in pediatrics patients aged <16 years. fracture (�=3,182) were used as controls. A genome‑wide Furthermore, these figures may indicate that pediatric patients association study (GWAS) was performed using � frequentist who do not have any diseases affecting bone integrity m�y association test. I� the GWAS analysis, � linear regression still be at risk to suffer � bone fracture or even recurrent model was fitted to test for additive effects of single‑nucleotide fractures (2-4). The causes for this m�y arise from the diet, polymorphism� (SNP�; genotyp� dosage) adjusting for sex and preferred free‑tim� activities or sex, but m�y also be linked to performing population stratification using genotypic principal genetic factors (3). components. Using the GWAS analysis, the present study iden- T�� causes and etiopathogenetic pathw�y� of pediatric tified one locus with a significant association with fractures fractures are not well studied (5). I� has been previously during childhood on chromosom� 10 (rs112635931) and six loci reported that bone‑associated and bone‑independent factors with � suggested �mplication. The lead SNP rs112635931 was contribute to the fracture risk (2,6). Furthermore, bone located near proline‑ and serine‑rich 2 (PROSER2) antisense mineral content (BMC) and bone mineral density (BMD) have been indicated to be highly heritable in genetic studies (7-11). Mora and Gilsanz (12) revealed that 80% of the variance of the BMC is influenced by genetics. Furthermore, BMD has been reported to be � risk factor for fractures in adults (1), but there Correspondence to: M�. Roop� Parviainen, D�partment of Children is limited research on pediatric fractures. Previous studies and Adolescents, Oulu Childhood Fracture and Sports I�jury Study, have linked � low BMD to increased fracture risk during Research Unit for Pediatrics, Pediatric Neurology, Pediatric Surgery, Child childhood (3,13); however, in other studies, this association P�ychiatry, Dermatology, Clinical Genetics, Obstetrics and Gynecology, was not identified (6,14). In adults, genome‑wide association Otorhinolaryngology and Ophthalmology (PEDEGO), Oulu Medical studies (GWAS) have identified genes from ≥11 different Research Center (MRC), University of Oulu and Oulu University Hospital, P.O. B�x 5, FI-90029 Oulu, Finland chromosomes (chrs) that m�y affect BMD either locally or E-mail: roop�[email protected] throughout the body (15-21). H�wever, with regards to pedi- atrics, only few studies have reported on genetic loci that Key words: fractures, risk, children, genetics, genom�-wide influence BMD or cortical bone thickness (7,8). Furthermore, association study, cohort the heritability of the fractures and BMD differ, and thus, it has been suggested that fracture risk should be analyzed inde- pendently of BMD (22). PARVIAINEN et al: A SINGLE GENETIC LOCUS ASSOCIATED WITH PEDIATRIC FRACTURES 1717 BMD and BMC are useful when �x�mining the factors T���� I. R�p����� ICD-9 �����. underlying fractures (23,24); however, their association with fractures in childhood remains to be fully elucidated. ICD-9 ���� Exp�������� Therefore, the present study focused on fractures rather than BMD or BMC. T� the best of our knowledge, only few studies 800 F������� �� ����� �� �k��� have �x�mined the association between genetics and child- 802 F������� �� ������ ����� hood fracture risk. Therefore, the aim of the present study 810 F������� �� �������� was to identify whether any genetic loci are associated with 812 F������� �� ��m���� fractures prior to the age of 7 years using GWAS. 813 F������� �� ������ �� ���� 816 F������� �� ��� �� m��� p�������� �� ��� ���� Patients and methods 821 Fracture of other or unspecified parts of femur 823 Fracture of tibia or fibula Patient population. The study p�pulation was � part of the Northern Finland Birth Cohort 1986 (NFBC1986) that ICD, International Classification of Diseases. originally included all pregnant females (�=9,362) living in the geographic area of the �w� northernmost provinces of Finland with �xpected date of delivery between 1st July 1985 and 30th June 1986. These mothers and their live‑born register worldwide and thoroughly covers acute injuries of the infants (�=9,432) were followed �p regularly since preg- population (annual coverage of injuries, ~100%) and classifies nancy. Immigration during this period to northern Finland these injuries accurately (annual accuracy, >95%) (30). was minimal, and therefore, the cohort was predominantly I� the study area, pediatric fractures are �xclusively treated Finnish (25). The detailed description of the cohort is provided in public hospitals and it is mandatory to send information in the study published �y Järvelin et al (26) and on the NFBC of these injuries to the NHDR. During the study period, website (27). out‑patient visits were not recorded in the NHDR and any O� the cohort p�pulation, 6,721 consented to the use of patients treated outside the hospital were not included. S�mple their information for research purposes. The information for fractures, including non‑displaced fracture of wrist or foot, the individuals who did not provide consent was not available. were not included in the NHDR, as these are treated �y Study subjects with diseases of the bone, including osteogen- general practitioners in primary healthcare centers. All the in esis �mperfecta, or possible bone‑affecting malignancies, were hospital treated fractures were recognized for the analysis and �xcluded from the original study p�pulation (�=3). Informed the patients were classified according to whether they had a consent from the parents was provided orally during antenatal fracture or not. The International Classification of Diseases, clinic visits. A� the 16-year follow‑up (in 2001), the cohort version‑9 (ICD-9) was used to identify the fractures (31). I� participants and their parents provided written informed the patient had several hospitalizations with the sam� ICD consent for the use of all of their collected data for scientific code, the fractures were determined to be caused by ≥2 sepa- purposes. T�� participation was voluntary. rate injuries if there was an interval of ≥6 months between the hospitalization dates. The ICD-9 codes reported for the Genotype data. Genotypes were determined for the cohort fractures are presented in Table I. m�mbers who had provided consent and their blood samples The cause of the fracture m�y also be included in the were used for this purpose. The genotyp� data were available NHDR data. Out of the 48 individuals who had � fracture in the for 3,230 individuals, including 48 cases who had � fracture and p�pulation of the present study, 37 had � reported mechanism. 3,182 controls without fractures. Genotyping was performed Furthermore, 15 patients had low‑energy accidents, including using the Illumina H�manOmniExpressEx�me v1.2 bead falling onto � plane, tripping, slipping or falling from � height chip (Illumina, Inc.) and �mputation was performed using the of <1 m. For the other fractures, it was not possible to reliably 1000G phase 3 reference panel (28). After quality control (QC), determine the energy involved in the injury. the study included 3,515,000 single‑nucleotide polymor- phism�