Familial co-occurrence of congenital heart defects follows distinct patterns

Abstract

Aims: Congenital heart defects (CHD) affect almost 1% of all live born children and the number of adults with CHD is increasing. In families where CHD has occurred previously, estimates of recurrence risk, and the type of recurring malformation are important for counselling and clinical decision-making, but the recurrence patterns in families are poorly understood. We aimed to determine recurrence patterns, by investigating the co-occurrences of CHD in 1163 families with known malformations, comprising 3080 individuals with clinically confirmed diagnosis.

Methods and results: We calculated rates of concordance and discordance for 41 specific types of malformations, observing a high variability in the rates of concordance and discordance. By calculating odds ratios for each of 1640 pairs of discordant lesions observed between affected family members, we were able to identify 178 pairs of malformations that co-occurred significantly more or less often than expected in families. The data show that distinct groups of cardiac malformations co-occur in families, suggesting influence from underlying developmental mechanisms. Analysis of human and mouse susceptibility genes showed that they were shared in 19% and 20% of pairs of co-occurring discordant malformations, respectively, but none of malformations that rarely co-occur, suggesting that a significant proportion of co-occurring lesions in families is caused by overlapping susceptibility genes.

Conclusion: Familial CHD follow specific patterns of recurrence, suggesting a strong influence from genetically regulated developmental mechanisms. Co-occurrence of malformations in families is caused by shared susceptibility genes.

Figure 1

The number of unique pairs of diagnoses per family was tailed. The total number of times a pair was observed between families was used in calculating the odds ratio for different pairs of diagnoses. (A) Example of a family with three affected individuals. (B) Possible pairs of diagnoses observed between individuals. (C) A maximum of one unique pair was included per family. See Supplementary material online, Table S2 for abbreviations.

Figure 2

Gender ratio, concordance and discordance in 1163 congenital heart defects families. A family was considered to be concordant (green bar) if all affected individuals shared the same diagnosis. A family was considered to be discordant (blue bar) if all affected individuals had different diagnoses. If a family had family members that shared some of the diagnoses it was considered as both concordant and discordant (orange bar). *Statistical significant differences.

Figure 3

Comparison of familial relatedness with concordance and discordance. Relatedness is shown as a rate of first-degree relatedness, of which first-degree relatives have a rate of 1. The rate decreases for more distant relatives.

Figure 4

Patterns of co-occurrence of phenotypes in congenital heart defects families. The log-odds ratio (LOR) for 1640 pairs of malformations is displayed as a heatmap. Only pairs occurring ≥10 times in the dataset were included. To enhance readability only LOR values ≤ −1 and ≥1 are shown. The data were grouped according to similarities in LOR using hierarchical cluster analysis. An anatomical term for each group is suggested on the right. LVOTO, left ventricular outflow tract obstruction; RVOTO, right ventricular outflow tract obstruction; n.d., not determined. International Pediatric Congenital Cardiac Code diagnoses are indicated in full and abbreviated.

Figure 5

Overlap of deduced susceptibility genes between discordant phenotypes. Mouse (A) and human (B) susceptibility genes were identified (numbers shown in bar chart) and malformations were analysed in pairs (M1 and M2, indicated on the y-axis).