A meta-analysis of genetic effects associated with neurodevelopmental disorders and co-occurring conditions

Abstract

A systematic understanding of the aetiology of neurodevelopmental disorders (NDDs) and their co-occurrence with other conditions during childhood and adolescence remains incomplete. In the current meta-analysis, we synthesized the literature on (1) the contribution of genetic and environmental factors to NDDs, (2) the genetic and environmental overlap between different NDDs, and (3) the co-occurrence between NDDs and disruptive, impulse control and conduct disorders (DICCs). Searches were conducted across three platforms: Web of Science, Ovid Medline and Ovid Embase. Studies were included only if 75% or more of the sample consisted of children and/or adolescents and the studies had measured the aetiology of NDDs and DICCs using single-generation family designs or genomic methods. Studies that had selected participants on the basis of unrelated diagnoses or injuries were excluded. We performed multilevel, random-effects meta-analyses on 296 independent studies, including over four million (partly overlapping) individuals. We further explored developmental trajectories and the moderating roles of gender, measurement, geography and ancestry. We found all NDDs to be substantially heritable (family-based heritability, 0.66 (s.e. = 0.03); SNP heritability, 0.19 (s.e. = 0.03)). Meta-analytic genetic correlations between NDDs were moderate (grand family-based genetic correlation, 0.36 (s.e. = 0.12); grand SNP-based genetic correlation, 0.39 (s.e. = 0.19)) but differed substantially between pairs of disorders. The genetic overlap between NDDs and DICCs was strong (grand family-based genetic correlation, 0.62 (s.e. = 0.20)). While our work provides evidence to inform and potentially guide clinical and educational diagnostic procedures and practice, it also highlights the imbalance in the research effort that has characterized developmental genetics research.

Fig. 1. Visual summary of the three core aims of the current meta-analysis.

Aim 1 (orange and light blue): estimate family-based genetic (h²), shared environmental (c²) and non-shared environmental (e²) influences as well as SNP heritability (h²_SNP) for all NDDs identified by the DSM-5. Aim 2 (orange and red): provide grand estimates of family-based genetic (r_A), shared environmental (r_C) and non-shared environmental (r_E) correlations and SNP-based genetic correlations ($r_{{\mathrm{G}}_{\mathrm{SNP}}}$) between different NDDs. Aim 3 (navy blue and red): provide grand estimates of r_A, r_C, r_E and $r_{{\mathrm{G}}_{\mathrm{SNP}}}$ between NDDs and DICCs. The results for c², e², r_C and r_E are presented in Supplementary Note 1.

Fig. 2. Diagram of searches and screening.

Overview of the screening and selection process across primary and secondary searches, along with statistics of inter-rater reliability.

Fig. 3. Genetic and environmental sources of variation in NDDs.

Meta-analytic family and SNP-based heritability (h²) of NDDs and shared environmental influences (c²) and non-shared environmental influences (e²) on variation in NDDs. The number preceding each bar on the y axis denotes the number of studies identified that provided estimates for specific NDDs. The error bars signify standard errors of the grand estimates of heritability and environmental influences. The results for c² and e² are discussed in Supplementary Note 1.

Fig. 4. Genetic and environmental correlations between NDDs and DICCs.

Strength of the meta-analytic genetic (r_A), shared environmental (r_C) and non-shared environmental (r_E) correlations between NDDs and their homotypic (other NDDs) and heterotypic (DICCs) co-occurrences. The outer layer of each circle shows all the different NDDs and DICCs for which meta-analytic correlation estimates could be computed. Each coloured connector path indicates the strength of association between disorders; the thicker the connector path, the stronger the correlation between the two disorders. The results for family r_C and r_E are presented in Supplementary Note 1.

Fig. 5. Sex differences.

Distributions of the sex-specific meta-analytic estimates for the heritability (h²) of NDDs and environmental contributions to NDDs. The top left panel shows the distributions of sex-specific estimates for the transdiagnostic meta-analysis; the remaining panels show the same estimates for specific NDDs for which a sufficient number of studies (>2) reporting sex-specific estimates was identified. The results for sex-specific c², e², r_C and r_E estimates are presented in Supplementary Note 1.

Fig. 6. Developmental trajectories.

a,b, Age-related differences in family-based heritability (h²) and shared (c²) and non-shared (e²) environmental influences on NDDs (a) and SNP heritability (b). Developmental stages include childhood (4–7 years), middle childhood (8–10 years) and adolescence (11–24 years). The error bars represent standard errors of grand estimates of heritability and environmental influences. The number located near each point estimate denotes the number of studies identified that provided estimates for specific developmental stages. For intellectual disabilities and motor disorders, we could not identify a sufficient number of studies (>1) reporting age-dependent estimates, and we were consequently unable to derive meta-analytic estimates. The results for age-stratified c², e², r_C and r_E are reported in Supplementary Note 1.

Fig. 7. Geographical differences.

a, Differences in family-based heritability (h²) and shared environmental (c²) and non-shared environmental (e²) influences across all NDDs. b, Geographical differences in the genetic (r_A), shared environmental (r_C) and non-shared environmental (r_E) overlap between NDDs. The areas shaded in grey are regions for which not enough relevant studies were identified (<2 studies). Geographical differences in r_A, r_C and r_E between NDDs and DICCs are presented in Supplementary Fig. 28. The results for c² and e² as well as r_C and r_E are discussed in Supplementary Note 1.