|
nif:isString
|
-
Subject recruitment and data collection for CHD GENES have been described [11]. Briefly, subjects were recruited from five main sites (Children’s Hospital of Philadelphia, Columbia University Medical Center, Harvard Medical School including Boston Children’s Hospital and Brigham and Women’s Hospital, Icahn School of Medicine at Mount Sinai, and Yale School of Medicine) and four satellite sites (Children’s Hospital of Los Angeles, Cohen Children’s Medical Center, University College London, and University of Rochester Medical Center) from December 2010 through November 2014. Recruitment methods were center-specific, but generally included ascertainment of cases at the time of hospital admission or an outpatient visit. The study protocol was approved by an Institutional Review Board for each site. All study participants (or their parent/guardian) provided written informed consent. The Institutional Review Board at the University of Texas Health Science Center at Houston approved the study protocol for the data analyzed and presented in this article. Patients with any diagnosis of CHD (except as noted below), regardless of sex, age, and race/ethnicity were eligible to participate. Patients with a genetic diagnosis were eligible to participate, but preference for enrolling such patients may have varied across study sites. Patients with isolated patent foramen ovale, prematurity-related isolated patent ductus arteriosus, pulmonary stenosis related to a twin-twin transfusion, and cardiomyopathy without a CHD were not eligible. Cardiac diagnoses were confirmed by review of imaging (e.g., echocardiogram) and operative reports. Information on genetic testing, genetic physical exams, and extracardiac malformations was abstracted from medical records. In addition, information on cases and their parents was obtained during subject and family interviews. Cases that did not participate in the interviews were excluded from this report. Data collected by interview included race/ethnicity, sex, birth weight, and maternal and paternal ages at the time of the cases’ birth. Data were also collected on maternal characteristics, including pre-pregnancy height and weight (to calculate pre-pregnancy body mass index), pre-gestational diabetes, gestational diabetes, epilepsy or seizure during pregnancy, and education level. For cases who were ≤1 year of age at recruitment, interview data were also collected on maternal smoking and alcohol use during the first trimester, any folic acid supplementation six months before pregnancy, and parity. For cases who were >1 year at recruitment, the interview included questions related to neurodevelopmental outcomes (e.g., attention deficit hyperactivity disorder, autism spectrum). CHD diagnoses assigned using the International Paediatric and Congenital Cardiac Codes (http://www.ipccc.net/) were manually reviewed by two of the authors (S.E and E.G.) and cases were assigned to one of seven types of CHDs: laterality disorder (LAT), conotruncal heart defect (CTD), atrioventricular septal defect (AVCD), left ventricular outflow tract obstruction (LVOT), right ventricular outflow tract obstruction (RVOT), atrial septal defect (ASD), and other. These groups are based on subsets of lesion that are thought to share genetic and mechanistic underpinnings and are defined in Table 1. Cases were categorized using a hierarchical approach. First, cases with a laterality disorder, regardless of other findings, were placed in LAT. Next, cases with abnormal conotruncal anatomy (including specific subtypes of isolated ventricular septal defects), regardless of associated left or right sided obstruction or atrioventricular canal anomalies, were placed in CTD. Then, cases with atrioventricular canal abnormalities with normally related great arteries were categorized as AVSD and cases with left or right sided obstructive lesions with normally related great arteries and normal atrioventricular canals were assigned to LVOT or RVOT, respectively. Finally, cases with an isolated secundum or sinus venosus type atrial septal defect were assigned to ASD. Cases with any other CHD diagnosis were assigned to the other group.
Table data removed from full text. Table identifier and caption: 10.1371/journal.pone.0191319.t001 Diagnostic types of congenital heart defect in the Pediatric Cardiac Genetic Consortium Cohort. a Does not include variants of hypoplastic left heart syndrome such as malaligned atrioventricular canal defect or double outlet right ventricle with mitral atresia.
Based on data from the interviews and medical records, cases were classified as either having 1) an identified genetic diagnosis (i.e. a syndrome or genetic alteration thought to explain the associated CHD), or 2) no genetic diagnosis. For simplicity, we refer to such cases as “syndromic” and “nonsyndromic”, respectively. Cases classified as nonsyndromic by this scheme may have had additional non-cardiac anomalies or reported neurodevelopmental deficits.
For syndromic cases, we reported counts and frequencies for each specific diagnosis. Given the clinical heterogeneity within this group, we excluded syndromic cases from subsequent analyses. For nonsyndromic cases, parental characteristics, case characteristics, and parent-reported neurodevelopmental outcomes were described using counts and frequencies for discrete variables, and means and standard deviations or median and range for continuous variables. Due to differences in the education systems in the United States and United Kingdom, we excluded women who were educated in the United Kingdom in our description of maternal education. Further, we restricted our analyses of neurodevelopmental outcomes to cases who were ≥5 years of age at recruitment, since neurodevelopmental deficits may be under-diagnosed in younger children. In addition to assessing each of 13 parental-reported (yes/no) neurodevelopmental outcomes, we created a composite neurodevelopmental outcome variable, indicating a positive parental report for at least one of four conditions: developmental delay, learning disability, mental retardation, or autism spectrum disorder [13]. We used the chi-square test (or Fisher’s exact test when >20% of cells had an expected cell count <5) to compare the distribution of categorical variables across types of CHDs. For continuous variables, we used ANOVA or the Kruskal-Wallis test to compare the mean or median, respectively, across types of CHDs. For ANOVA analyses, we used Levene’s test to check the assumption of homogeneity of variance. If Levene’s test was significant (p<0.05), we used Welch’s ANOVA. Analyses of all variables, except neurodevelopmental outcomes, were repeated in the subset of cases who were ≤1 year of age at recruitment for the following reasons: 1) inaccurate recall of characteristics or events before or during pregnancy is of greater concern for cases ascertained at older ages than at younger ages; and 2) the distribution of characteristics across types of CHDs may be influenced by survival. Because of the heterogeneity within type of CHDs, analyses were also repeated to compare specific subtypes in the two largest types of CHDs—CTD and LVOT cases. These analyses were restricted to include subtypes that included at least 200 cases. For LVOT, cases with aortic stenosis were combined with cases with bicuspid aortic valve to create a subtype called ‘aortic valve disease.’ Because differences in the distribution of neurodevelopmental outcomes across types of CHDs may be influenced by factors other than the CHD diagnosis, we used logistic regression to control for potential confounders determined a priori from the literature [15]: maternal education, case race/ethnicity, sex, birth weight (low [<2,500g], normal [2,500–4,000g], high [>4,000g]), and extracardiac malformations (yes/no). Further, as neurodevelopmental deficits may be under-diagnosed in younger cases, we also adjusted for case age at the time of recruitment. Adjusted analyses were not conducted for the CTD and LVOT subtypes because of the relatively small numbers of cases with specific outcomes (e.g., double outlet right ventricle with autism spectrum, N = 4). All analyses were conducted using SAS version 9.4 (SAS Institute Inc., Cary, NC). P-values <0.05 were considered statistically significant.
|
![[This page as RDF]](http://data.gesis.org/somesci/static/rdf-icon.gif){kind=icon}