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

Disclosures

Hum Reprod. 2020;35(9):2113 

In This Article

Results

Literature Search

A total of 185 references were identified through MEDLINE and EMBASE (Figure 1). Six additional studies were retrieved from the list of references in the set of articles initially identified. Of these 191 references, we excluded 25 duplicate reports. Based on title and abstract screening, 127 references were further excluded, leaving 39 studies for full-text assessment. Twenty-one full-text studies were excluded for the following reasons: specific CHDs were included (N = 9), irrelevant outcome (N = 4, circulatory system, children mortality, congenital malformations), wrong exposure (N = 4, maternal age, child gender, socio-occupational status, no paternal data), could not have access to full text (N = 1), lack of comparison group (N = 1), systematic review (N = 2) (Supplementary Table SII).

Figure 1.

Flow of studies through the review.

In total, we included 18 studies in our systematic review (Tay et al., 1982; Stoll et al., 1989; Zhan et al., 1991; Pradat, 1992; Olshan et al., 1994; Bassili et al., 2000; Cedergren et al., 2002; Kazaura et al., 2004; Baltaxe and Zarante, 2006; Yang et al., 2007; Materna-Kiryluk et al., 2009; Cresci et al., 2011; Fung et al., 2013; Liu et al., 2015; Su et al., 2015; Wang et al., 2015; Yuan et al., 2015; Abqari et al., 2016). Among these studies, five reported paternal age only as means (N = 5), four studies investigated the association between paternal age and CHD using a different threshold for advanced paternal age, which impeded inclusion in the meta-analysis (cut-off at 30 years and risk of CHD in two (Baltaxe and Zarante, 2006; Yuan et al., 2015), cut-off at 25 years in one (Abqari et al., 2016) and <20, 20–39, ≥40 years in one (Bassili et al., 2000)). Nine provided sufficient data on the association between paternal age ≥35 years and the risk of CHD for inclusion in the quantitative meta-analysis.

Characteristics of Included Studies

The characteristics of included studies are presented in Table I. The included studies were conducted in 10 different countries and published between 1991 and 2016. Six studies assessed the risk of isolated CHD, and seven studies included all CHDs combined (isolated, chromosomal and syndromic). Eight studies included live births only whereas five comprised live births, fetal deaths and TOPFA. Four studies were population-based case-control studies, and five were clinical case-control studies. The number of CHD cases per study ranged from 39 to 15 216 subjects (median N = 706). The number of controls per study ranged from 52 to 5 228 452 subjects (median N = 6222). Overall, the systematic review encompassed ~34 000 cases of CHD and 10 000 000 controls. Twelve studies controlled for maternal age.

NOS quality scores ranged from 3 to 11, with a median score of eight. Five studies had low risk of bias scores (Cedergren et al., 2002; Kazaura et al., 2004; Yang et al., 2007; Materna-Kiryluk et al., 2009; Su et al., 2015), three had medium risk scores and one had a high risk of bias score (Supplementary Table SIII). The most common concern for bias resulted from selection bias and especially in the selection of controls (hospital controls/drawn from a different source (exposition biases) or no description).

Qualitative Results of the Nine Studies Excluded From the Meta-analysis

Means of father's age on cases was between 27.8 and 33 years and on controls between 28.2 and 34 years. None of the means were significantly different between the cases and the controls (Tay et al., 1982; Stoll et al., 1989; Cresci et al., 2011; Fung et al., 2013; Liu et al., 2015).

Two studies (Baltaxe and Zarante, 2006; Yuan et al., 2015) found that paternal age ≥30 years was significantly associated with an increase in the odds of CHD. The same was found from Abqari et al. (2016) for paternal age ≥25 years and from Bassili et al. (2000) for paternal age ≥40 years.

Association Between Advanced Paternal Age and CHD: Meta-analysis

The meta-analysis included data from nine studies, encompassing 34 447 CHD cases and 9 882 564 controls without CHD. Results of the meta-analysis showed that advanced paternal age was significantly associated with an increase in the odds of CHD (summary OR 1.16, 95% CI 1.07–1.25, I 2 = 81%) (Figure 2), with positive and negative predictive values of 0.8% and 90%, respectively. The e-value of included articles was between 1.17 and 7.42 (Supplementary Table SIV). The funnel plot seemed asymmetric (Figure 3). However, Egger's test did not show any statistical evidence for publication bias (P = 0.84).

Figure 2.

Association between advanced paternal age and CHD (congenital heart defects): main meta-analysis (nine studies, random-effects model).

Figure 3.

Publication bias visualized by a Funnel Plot. OR, odds ratio.

Results of stratified meta-analysis were stable, i.e. OR between 1.02 and 1.24 (Table II) including those carried out on population-based studies, studies with low risk of bias and studies that focused on the entire population (live births, fetal deaths and TOPFA). Statistical heterogeneity (I 2) ranged from 31% to 91% depending on the subset of studies in the meta-analysis, indicating a fairly substantial degree of heterogeneity across studies.

Dose–Response Meta-analysis

Twelve studies (Zhan et al., 1991; Pradat, 1992; Olshan et al., 1994; Bassili et al., 2000; Cedergren et al., 2002; Kazaura et al., 2004; Yang et al., 2007; Materna-Kiryluk et al., 2009; Su et al., 2015; Wang et al., 2015; Yuan et al., 2015; Abqari et al., 2016) using different paternal age categories could be included in the dose–response meta-analysis. With the cubic spline model, we found no evidence of a nonlinear relation between paternal age and risk of CHD (P = 0.52) (Figure 4). Test of a linear dose–response curve between paternal age and risk of CHD showed a positive relation, which was not statistically significant (P = 0.15).

Figure 4.

Dose–response meta-analysis. The measures of associations in the selected studies were ORs or RRs (relative risk) depending on the study design.

Comments

3090D553-9492-4563-8681-AD288FA52ACE

processing....