Introduction: Coronary Artery Risk Detection in Appalachian Communities (CARDIAC) is a project that has assessed childhood risk factors for cardiovascular disease (CVD) for nearly twenty years. We have screened approximately 95,000 children from WV that now include 60,403 with lipid panels. Eight-thousand-five-hundred-fifty-seven persons have one or more siblings who have participated in the project. To begin to consider genetic primordial antecedents of CVD in WV schoolchildren, we linked sibships for analysis.
Methods: Linkage software designed to deduplicate EHRs proved poorly adaptable to the problem of family matching. Successive excel spreadsheets were used to directly match either mother or father’s first and last name with care taken to avoid similar names by requiring additional matching parameters. A group of children were identified with younger siblings subsequently enrolled in CARDIAC and the two older children were matched for analysis when 3 siblings were found. Body Mass Index z-score was calculated for each child. Fasting lipid levels were obtained by venipuncture and measured by a commercial laboratory (LabCorp).
Results: Children (5259)with at least one sibling with LDL levels available for both were found along with 1384 third siblings. The mean LDL level for the first child was 90.5 mg/dl and 92.1 mg/dl for the second child. The mean BMI z-score was +0.83 for the group of 95,000 children, +0.73 for the first sibling and +0.72 for the second sibling. Sibling BMI z-scores were strongly correlated. Multivariate analysis demonstrated that the LDL level for the second child correlated strongly with both the older sibling’s LDL level and with the second child’s BMI z-score. This model showed R2 = 0.15 and the following prediction equation:
LDLsibling2 = 56.3 + 0.377*LDL-proband + 2.24*BMI-z-score-sibling2
p = 4.33 x 10 -175 for LDL and p = 4.06 x 10 -15 for BMI z-score
Six-hundred-sixty-eight children had an LDL level greater than 130 mg/dl including 88 sets of siblings. 12 pairs of siblings each had an LDL level > 160 mg/dl.
Conclusion: the analysis demonstrates a strong correlation between sibling LDL-c levels that is not as well explained by the correlation of BMI z-score as by LDL-c. This analysis along with our recent study of LDL and height will help to characterize CVD risk in this population to focus future investigations in children and families.