Association Between Advanced Paternal Age and Congenital Heart Defects: A Systematic Review and Meta-analysis

A Systematic Review and Meta-analysis

F. Joinau-Zoulovits; N. Bertille; J.F. Cohen; B. Khoshnood


Hum Reprod. 2020;35(9):2113 

In This Article


In this meta-analysis, we found that advanced paternal age (>35 years) was significantly associated with a 16% increase in the odds of CHD and the dose–response meta-analysis shows a linear association between paternal age and CHD. This association remained stable when meta-analysis was restricted to population-based studies (Kazaura et al., 2004; Yang et al., 2007; Materna-Kiryluk et al., 2009; Su et al., 2015), for the entire population, and to studies deemed at low risk of bias. The meta-analysis summary estimate lies within the bounds of 95% CIs for all included studies. This suggests that all individual study results are consistent with the summary measure. There was substantial heterogeneity in the reported estimates across studies, which is probably explained at least in part by methodological issues: (i) different strategies to account for potential confounders, (ii) various definitions of CHD, (iii) differences in study populations and methods of case ascertainment, (iv) inclusion (or not) of TOPFA and fetal deaths (Pradat, 1992; Cedergren et al., 2002; Kazaura et al., 2004) and (v) different age intervals used to define advanced paternal age (Bassili et al., 2000; Abqari et al., 2016). Most e-values were low (<2), suggesting that their results of the corresponding studies were not robust to unmeasured confounders. In particular, estimates may have been biased due to confounding because of lack of adjustment for maternal age (Miller et al., 2011) and assisted reproductive technologies given the association of the latter with CHD (Tararbit et al., 2011).

There were limitations in the retrieved evidence due to issues with design and analysis. These limitations included paucity of data for investigating the risk of isolated CHD (i.e. not associated with chromosomal or other anomalies) (Baltaxe and Zarante, 2006; Materna-Kiryluk et al., 2009), lack of adjustment for potential confounders such as maternal age, and inconsistent definitions of advanced paternal age. We could not use adjusted estimates in meta-analysis because, in studies with multivariable adjustment, age categories were inconsistent across studies. The data available did not allow us to look at adjusted estimates across studies using a common cut-off value for paternal age. So, we conducted a dose–response meta-analysis by looking at adjusted estimates from studies that had looked at different paternal age groups. We did not find a statistically significant dose-effect (P = 0.15); however, the point estimate suggested that there may be a dose-effect that was not statistically significant due to insufficient power. Moreover, we investigated the association between paternal age and the overall risk of CHD, but data did not allow to assess the association between paternal age and specific CHDs (Archer et al., 2007; Green et al., 2010), for example, membranous ventricular septal defects (Ewing et al., 1997). This is an important limitation because known teratogens (e.g. valproic acid) (Meador et al., 2008), and established risk factors of congenital anomalies in general, and those of CHD in particular, may be associated with one or a few specific congenital anomalies. In other words, paternal age may be associated with some CHDs, but not with others. Although the funnel plot seemed asymmetric, Egger's test was not significant. A reason could be that Egger's test often lacks statistical power when <10 studies are included (rule of thumb) (Page et al., 2019).

The underlying mechanisms of any association between paternal age and offspring CHD are unclear. According to the literature, several potential mechanisms may account for such associations (Sharma et al., 2015). Advanced paternal age has been associated with poor semen parameters, notably reduced semen volume, sperm motility and sperm morphology (Das et al., 2013). These changes appear to occur mostly after 34 years of age (Kidd et al., 2001; Stone et al., 2013). Furthermore, Singh et al. (2003) showed that the percentage of highly damaged sperm DNA was significantly higher after 35 years of age. In a meta-analysis of 26 studies, Johnson et al. (2015) reported a significant negative association between advanced paternal age and DNA fragmentation. Paternal age also appears to affect telomere shortening (Fumagalli et al., 2012), which can lead to cellular senescence and apoptosis (Rossiello et al., 2014). Also, advanced paternal age can result in accumulation of de novo mutations, resulting in an increased risk of genetic abnormalities in the offspring (Callaway, 2012). One study estimated that de novo mutations increased by 4% per year with advancing paternal age (Kong et al., 2012). The same authors also reported that the heritability of mutations in the offspring was affected by advanced paternal age (Kong et al., 2012). Finally, a novel hypothesis, the so-called 'selfish spermatogonial selection' (Goriely and Wilkie, 2012) has recently been put forward. This phenomenon occurs as rare spermatogonial cells bearing mutations are preferentially selected, leading to their progressive clonal expansion. Exposure to environmental risk factors may also cause sperm genetic changes due to accumulated exposure effects (Jafarabadi, 2007). In addition to age, father's diet and exposure to toxic agents such as certain chemicals (Snijder et al., 2012) or heavy smoking (Cresci et al., 2011) may also affect the offspring's epigenetic factors (Curley et al., 2011). Nevertheless, whether any of the factors noted above may be associated with the risk of CHD, particularly isolated CHD, remains unclear. Hence, whether or not paternal age has a causal effect on the risk of CHD through one or more of the mechanisms noted above remain speculative.

In conclusion, the results of our systematic review and meta-analysis suggest that advanced paternal age may lead to a modest increase in the risk of CHD (16% increase in the odds of CHD, positive and negative predictive values are, respectively, 0.8% and 90%), and although it is not significant, the risk seems to increase with age. However, because the association is modest in magnitude its usefulness as a criterion for targeted screening for CHD seems limited. Moreover, we call for caution when interpreting these findings because of several limitations in the available evidence. Future studies should further explore the association between advanced paternal and specific CHDs. In turn, this may provide new insights into the underlying mechanisms of any causal effects associated with advanced paternal age.