www.nature.com/npp ARTICLE OPEN Shared genetic architecture between attention-deficit/ hyperactivity disorder and lifespan Laura Vilar-Ribó 1,2,3,10, Judit Cabana-Domínguez 1,2,3,10, Lourdes Martorell 3,4, Josep Antoni Ramos-Quiroga1,2,3,5, ✉ Sandra Sanchez-Roige 6,7, Abraham A. Palmer6,7, Elisabet Vilella 3,4, Marta Ribasés 1,2,3,8, Gerard Muntané 3,4,9,10 and 1,2,3,8,10 ✉ María Soler Artigas © The Author(s) 2023 There is evidence linking ADHD to a reduced life expectancy. The mortality rate in individuals with ADHD is twice that of the general population and it is associated with several factors, such as unhealthy lifestyle behaviors, social adversity, and mental health problems that may in turn increase mortality rates. Since ADHD and lifespan are heritable, we used data from genome-wide association studies (GWAS) of ADHD and parental lifespan, as proxy of individual lifespan, to estimate their genetic correlation, identify genetic loci jointly associated with both phenotypes and assess causality. We confirmed a negative genetic correlation between ADHD and parental lifespan (rg = −0.36, P = 1.41e−16). Nineteen independent loci were jointly associated with both ADHD and parental lifespan, with most of the alleles that increased the risk for ADHD being associated with shorter lifespan. Fifteen 1234567890();,: loci were novel for ADHD and two were already present in the original GWAS on parental lifespan. Mendelian randomization analyses pointed towards a negative causal effect of ADHD liability on lifespan (P = 1.54e−06; Beta = −0.07), although these results were not confirmed by all sensitivity analyses performed, and further evidence is required. The present study provides the first evidence of a common genetic background between ADHD and lifespan, which may play a role in the reported effect of ADHD on premature mortality risk. These results are consistent with previous epidemiological data describing reduced lifespan in mental disorders and support that ADHD is an important health condition that could negatively affect future life outcomes. Neuropsychopharmacology (2023) 48:981–990; https://doi.org/10.1038/s41386-023-01555-x INTRODUCTION problems, such as oppositional defiant disorder, conduct Attention-deficit/hyperactivity disorder (ADHD) is a neurodeve- disorder, mood and anxiety disorders, and substance use lopmental disorder that emerges in childhood and often persists disorder [8, 12–14]; (ii) comorbid somatic disorders such as into adulthood, affecting approximately 5.3% of children and obesity [3], asthma [15] and diabetes [16, 17]; (iii) harmful adolescents and 2.8% of adults [1, 2]. It is characterized by age- lifestyle behaviors, such as unhealthy eating habits or smoking inappropriate symptoms of inattention, impulsivity, and hyper- [18–20]; and (iv) educational underachievement, low income, activity, which have a severe impact on the individual’s social, social adversity, delinquency and aggression [3, 21]. However, emotional and psychological functioning, often representing an despite the high rate of comorbid conditions also linked to an entry point into a poor life trajectory [3]. excess mortality [12], these do not fully explain the risk of death There is increasing evidence linking ADHD to a shorter life observed in individuals with ADHD [4, 5, 8], indicating that expectancy and mortality rates in individuals with ADHD are two ADHD itself is a health condition that confers an increased risk of to five times higher than in individuals without ADHD [4, 5]. mortality. For instance, inattention and impulsivity may directly Besides natural causes [5], the higher risk of early mortality in increase proneness to risk-taking behaviors and therefore risk individuals with ADHD appears to be largely due to misadventure for accidental injuries, leading to reduced lifespan [5]. including a high propensity for accidents and injuries and an ADHD and lifespan are complex traits influenced both by elevated risk of suicide [4–9]. genetic and environmental factors. Heritability is estimated to be ADHD often co-occurs with other mental and somatic around 70–80% for ADHD and 7–16% for human lifespan [22]. comorbid disorders, traits and behaviors that are likely to Genome-wide association studies (GWAS) identified genetic loci increase mortality rates [10, 11]. These include (i) mental health associated with both ADHD and lifespan [23, 24], although a large 1 Psychiatric Genetics Unit, Group of Psychiatry, Mental Health and Addiction, Vall d’Hebron Research Institute (VHIR), Universitat Autònoma de Barcelona, Barcelona, Spain. 2 Department of Mental Health, Hospital Universitari Vall d’Hebron, Barcelona, Spain. 3Biomedical Network Research Centre on Mental Health (CIBERSAM), Instituto de Salud Carlos III, Madrid, Spain. 4Hospital Universitari Institut Pere Mata, Institut d’Investigació Sanitària Pere Virgili (IISPV)-CERCA, Universitat Rovira i Virgili, Reus, Spain. 5Department of Psychiatry and Forensic Medicine, Universitat Autònoma de Barcelona, Barcelona, Spain. 6Department of Psychiatry, University of California San Diego, La Jolla, CA, USA. 7 Department of Medicine, Vanderbilt University Medical Center, Nashville, TN, USA. 8Department of Genetics, Microbiology, and Statistics, Faculty of Biology, Universitat de Barcelona, Barcelona, Spain. 9Institut de Biologia Evolutiva (UPF-CSIC), Department of Medicine and Life Sciences, Universitat Pompeu Fabra, Parc de Recerca Biomèdica de Barcelona, Barcelona, Spain. 10These authors contributed equally: Laura Vilar-Ribó, Judit Cabana-Domínguez, Gerard Muntané, María Soler Artigas. ✉email: muntaneg@peremata.com; maria.soler@vhir.org Received: 31 May 2022 Revised: 3 February 2023 Accepted: 20 February 2023 Published online: 11 March 2023 L. Vilar-Ribó et al. 982 part of their heritability still remains to be explained. Some genes Functional annotation known to be related with the risk of developing ADHD, such as Functional annotation of all SNPs with a conjFDR value <0.10 and an LD those related to the dopaminergic system, have also been r2 ≥ 0.6 with one of the independent significant SNPs was performed in associated with shorter life expectancy [19], and a negative FUMA (Functional Mapping and Annotation of Genome-Wide Association genetic correlation between ADHD and parental age at death has Studies, https://fuma.ctglab.nl/) [30]. We combined data from the Combined Annotation Dependent Depletion (CADD) scores [31], which been described [24]. These data support observational studies predict how deleterious the SNP effect is on protein structure/function showing an association between ADHD with both elevated based on 63 functional annotations, and RegulomeDB scores [32], a mortality risk and reduced estimated life expectancy in adulthood, categorical score that estimates the regulatory functionality of SNPs based and suggest that the underlying genetic background of ADHD and on existing functional data (annotation to cis-eQTLs, expression quantita- lifespan may overlap. tive trait loci) and evidence for transcription factor binding. CADD ≥ 12.37 In the present study, we aim to examine the shared genetic was considered as the threshold for deleterious variants, RegulomeDB architecture and the nature of the relationship between ADHD scores <3 were likely to have a regulatory function, and minimum and parental lifespan, as a proxy for individual lifespan, using chromatin states between 1–7 were considered open chromatin states. The NHGRI-EBI GWAS catalog [33], the release from the 15th of September available GWAS data on both phenotypes by: (i) estimating their 2021, was used to identify traits previously associated with the SNPs of genetic correlation; (ii) performing a cross-trait analysis and (iii) interest and we queried SNPs for known brain eQTLs using the Genotype- testing the causal role of ADHD on lifespan. Tissue Expression (GTEx) v8 [34] and BRAINEAC [35]. We also used FUMA to map SNPs to genes based on physical proximity (using default parameters) and eQTL in brain (based on GTEx v8 and BRAINEAC) to test for enrichment MATERIALS AND METHODS on gene ontology and biological pathways of the mapped genes. GWAS samples and data processing All analyses were corrected for multiple comparisons using False GWAS summary statistics on ADHD were obtained from Demontis et al., Discovery Rate. comprising a total of 19,099 individuals with ADHD and 34,194 healthy controls, all of European ancestry [24]. Summary statistics on parental Causality analyses lifespan were obtained from Timmers et al. and included data from about GWAS summary statistics. To assess the causal effect of the genetic 1 million individuals of European ancestry [23]. The GWAS summary liability of ADHD on lifespan, the summary statistics from Pilling et al. [36], statistics were referenced to a set of 9,546,816 SNPs generated from the generated from a linear mixed-effects model in 208,118 individuals of 1000 Genomes Project Phase 3 European reference panel (http:// European ancestry, was used instead of Timmers et al. [23]. The latter used www.internationalgenome.org/). SNPs that were non-biallelic, without a survival analysis to study parental lifespan, which may produce a rsIDs, duplicated, or with strand-ambiguous alleles were removed. We also significant bias in MR analyses [37]. LDSC was used to calculate genetic filtered out SNPs with INFO scores <0.9 in the summary statistics files, correlations between both studies and between ADHD and parental those mapping to the extended major histocompatibility complex (MHC, lifespan in the study by Pilling et al. [25, 36]. For mediation analyses, as genomic position in hg 19; chr6:25,119,106–33,854,733) and the 8p23.1 defined below we used summary statistics from Sanchez-Roige et al. for region (chr8:7,200,000–12,500,000), which are prone to rearrangements, total impulsivity score, lack of premeditation and positive urgency [38]. SNPs located on the X, Y and mitochondrial chromosomes, and SNPs with sample sizes 5 standard deviations away from the mean. Finally, a common set of 3,206,697 SNPs were kept in both summary statistics. All Mendelian randomization. Causality between ADHD (as the exposure), P values were adjusted for standard genomic control (GC). and parental lifespan (as the outcome), was assessed by two-sample MR using the TwoSampleMR and MRPRESSO R packages [39, 40]. After clumping (r2 = 0.05, kb = 500) with PLINK 1.09 [29], independent SNPs Genetic correlation and pleiotropy assessment were selected using a P value threshold of 5e−08 in the ADHD GWAS to be Linkage Disequilibrium Score Regression (LDSC) was used to calculate used as instruments. The multiplicative random effects inversed-variance genome-wide genetic correlations across the traits studied [25]. weighted (IVW) was used as the main method to obtain the average effect The shared polygenic architecture between the two traits was assessed across genetic variants. For IVW results to be valid, genetic variants used as by means of stratified cross-phenotype Q-Q plots. P values for the primary instruments must meet three assumptions: (i) robust association with the trait were plotted conditioning on different association strengths (P < 1, exposure, (ii) absence of horizontal pleiotropy, or association with the 0.1, 0.01 and 1e−03) with the secondary trait. Thus, the visualization of a outcome through an exposure-independent pathway, and (iii) indepen- leftward shift in the primary trait of interest, as a function of increasingly dence of confounders influencing exposure and outcome. Additional MR strict P value thresholds in the secondary trait, was an indicator of a shared methods were implemented, as sensitivity analyses, for significant IVW polygenic architecture between the two traits. To test for SNP-based results (IVW P < 0.05) to assess the robustness of the findings under weaker heritability enrichment of a trait conditioned on different association assumptions: (i) the weighted median method, which under equal weights strengths with a secondary trait, we used stratified LDSC [26]. As a requires at least half of the variants to be valid instruments and is robust to reference panel for linkage disequilibrium (LD) we used the 1000 Genomes outliers [41]; (ii) the MR-PRESSO method, which tests for horizontal Project Phase 3 European reference panel [27]. pleiotropy (MR-PRESSO global test), and if detected, eliminates horizontal pleiotropic outliers and then performs the IVW method using the Cross-trait analysis remaining instruments [40]; and (iii) the MR-Egger method, which is To identify genetic loci jointly associated with ADHD and parental affected by outlier data points but allows all genetic variants to have lifespan we estimated the conjunction FDR (conjFDR), defined as “the pleiotropic effects assuming that these effects are independent of posterior probability that a given SNP is null for both phenotypes variant–exposure associations [42]. MR-Egger also implements a pleiotropy simultaneously when the P values for both phenotypes are as small as or test, however, when the NO Measurement Error (NOME) assumption is smaller than the observed P values”, using pleioFDR (https://github.com/ violated (I2gx < 0.9) MR-Egger causal estimates are biased towards the null, precimed/pleiofdr) [28]. We kept all SNPs with conjFDR <0.1 for and the type I error of the pleiotropy test can be inflated. For this reason, functional studies and reported independent SNPs with a conjFDR when I2gx < 0.6, MR-Egger results were disregarded. In addition, hetero- <0.05. Independent genomic loci were identified through clumping geneity tests and leave-one-out analyses were performed and scatter, (r2 = 0.05, kb = 500) using the 1000 Genomes Project Phase 3 European funnel and forest plots were generated. The analysis in the opposite as the reference panel for LD computation and PLINK 1.09 [27, 29]. We direction, testing the effect of parental lifespan on ADHD, was considered evaluated the directional effects of shared loci between ADHD and as a negative control. Finally, as an alternative method, we also used the ρ‐ parental lifespan by comparing z-scores between the original GWAS HESS implementation to prioritize putative causal models between pairs of summary statistics. Overlap between hits from the cross-trait analysis traits, as previously described [43, 44]. and genome-wide significant associations reported in the original GWAS results (P < 5e−08) was assessed according to distance (+/−250 kb) and Multivariate Mendelian Randomization (MVMR). To explore whether a linkage disequilibrium (r2 > 0.1) between of the cross-trait lead SNPs and putative causal effect of the genetic liability of ADHD on lifespan is previous reported hits. mediated by impulsive personality traits we used MVMR and summary Neuropsychopharmacology (2023) 48:981 – 990 L. Vilar-Ribó et al. 983 statistics from the GWAS by Sanchez-Roige et al. undertaken in over 20,000 To explore the landscape of pleiotropic effects further, we individuals [38]. Out of the ten traits analyzed by Sanchez-Roige et al. [38]. examined the direction of the effects of the lead SNPs of all we selected as potential mediators those reported to have a significant independent loci from the cross-trait analysis on both ADHD and genetic correlation with ADHD, prioritizing total scores when available, parental lifespan and found an opposite direction of effect for 95% namely total impulsivity score (measured with the BIS-11 [45]), lack of of them (n = 18), with alleles that increased the risk for ADHD also premeditation and positive urgency (both measured using the UPPS-P Impulsive Behavior Scale [46]). We used the same clumping parameters as shortening lifespan (Fig. 2B). ADHD hits [24] showed a significant for the main analysis (r2 = 0.05, kb = 500) and, given that no genome-wide negative correlation with parental lifespan with p-HESS, whereas significant SNPs were available, a threshold of 5e−06 was chosen to select no significant correlation was found between parental lifespan hits independent SNPs for these traits. [23] and ADHD, which is consistent with a putative causal relationship of ADHD on shortened lifespan (Supplementary Causal Analysis Using Summary Effect estimates (CAUSE). We also explored Figure 3). In line with these results, MR analyses showed evidence causality between ADHD and lifespan using the cause R package [47] of a negative causal effect of ADHD liability on lifespan (IVW considering independent variants (r2 = 0.05, kb = 500). The CAUSE Beta = −0.07 and P = 1.54e−06; weighted median Beta = −0.05 method uses a more permissive threshold for variant selection than and P = 6.52e−03; Supplementary Table 6 and Fig. 4). There was MR (P < 1e−03). In addition, CAUSE allows all variants to show no evidence of horizontal pleiotropy according to MR-PRESSO uncorrelated pleiotropy, also accounted for by MR-Egger or MR-PRESSO, but it also allows a subset of variants to show correlated pleiotropy, (global test P = 0.43) or heterogeneity (I2 = 11.15). The results when they affect exposure and outcome through a shared heritable were not driven by a single SNP (Supplementary Fig. 4). Variants factor. CAUSE compares two nested models by measuring how well the included in the MR analyses are shown in Supplementary Table 7. posterior distributions of a particular model fit the data: (i) a sharing MR testing the causal relationship of parental lifespan on ADHD, as model, which only allows for pleiotropic effects and no causal effects; a negative control, showed no significant results (IVW P = 0.66). and (ii) a causal model, which also allows both for pleiotropic and causal MVMR analyses showed that the effect of ADHD liability on effects. lifespan remained unchanged when accounting for total impul- sivity score or lack of premeditation, however it decreased (from −0.07 to −0.06) when taking into account the effect of positive RESULTS urgency, suggesting that positive urgency may be mediating We found strong evidence of negative SNP-based genetic around 11% of the effect of ADHD liability on lifespan correlation between ADHD and parental lifespan, used as a proxy (Supplementary Table 8). CAUSE did not provide evidence of for individual lifespan (rg = −0.36, P = 1.41e−16). Partitioning correlated pleiotropy (Supplementary Table 9) or of a causal effect ADHD SNP heritability on different parental lifespan P value of ADHD liability on lifespan (Beta = −0.01 and CI = (−0.03, 0)) thresholds showed enrichment of ADHD SNP heritability (Supplementary Tables 6 and 9). Given that survival analyses, (P = 2.73e−07 and P = 1.85e−03 conditioning ADHD on parental including the study conducted by Timmers et al. used in the lifespan P < 0.1 and P < 0.01, respectively; Supplementary Table 1a). present study, may bias MR results, we conducted the causality Furthermore, conditioning parental lifespan SNPs on ADHD P values analyses using summary statistics on parental lifespan from a showed parental lifespan SNP heritability enrichment (P = 1.09e second GWAS by Pilling et al. [36]. The genetic correlation −06 and 7.69e−03 conditioning parental lifespan on ADHD P < 0.1 between both summary statistics on parental lifespan was very and P < 0.01, respectively; Supplementary Table 1b). These enrich- high (rg = −0.93, P = 2.98e−184), and parental lifespan from the ments were consistent with stratified cross-phenotype Q-Q plots latest was also negatively correlated with ADHD (rg = −0.41, showing a stronger leftward deflection from the null expectation P = 2.66e−14). when conditioning ADHD on increasing levels of association for parental lifespan and vice-versa (Fig. 1). The cross-trait analysis showed a total of 19 independent DISCUSSION genomic loci associated with both ADHD and parental lifespan The present study provides the first evidence of a common with a conjFDR < 0.05 (Fig. 2A and Supplementary Fig. 1), 15 of genetic background between ADHD and lifespan. Extensive which were not identified in the original GWAS on ADHD and two literature supports that individuals with ADHD have an increased were already present in the original GWAS on parental lifespan risk of premature death and a shorter life expectancy, which (Table 1). Functional annotation of all SNPs with conjFDR < 0.1 at increases in females and depends on age at first ADHD diagnosis these 19 loci (n = 479 SNPs) revealed that 92.1% of these loci lay [4, 8]. In addition, the presence of other comorbid psychiatric on regions of open chromatin and most of the signals were conditions, such as oppositional defiant disorder, conduct disorder intergenic, intronic or located in intronic non-coding RNA (ncRNA) or substance use disorder, further increases ADHD mortality [4, 8]. genes. In addition, several SNPs in these genomic risk loci were This excess mortality and decreased life expectancy in ADHD likely to affect the binding of transcription factors or had CADD subjects are mainly driven by unnatural causes being accidents scores >12.37, suggesting high deleteriousness (Fig. 3, Table 1 the most common cause of early death [4, 9]. In addition, an and Supplementary Table 2). In addition, we found that 42% of increased risk of suicide, traffic violations or trauma have been the SNPs were eQTL for at least one gene in one brain area, also described in ADHD [8, 48–52]. At the same time, the impact of according to GTEx v8 [34] and BRAINEAC [35] (Supplementary the disorder on other adverse life-course outcomes, such as Table 3). Finally, 23 SNPs at 11 different genomic risk loci were poorer educational attainment, unemployment, delinquency or previously associated with different traits, mainly related to lower socioeconomic status, may also increase the risk of mortality lifetime risky behaviors (e.g., smoking, general risk tolerance and [11]. Strengthening the results of observational studies, we number of sexual partners), psychiatric disorders (e.g., schizo- provide evidence of shared genetic signatures and negative phrenia, ADHD and depression) and metabolic alterations (e.g., genetic correlation between ADHD and lifespan, which further metabolic syndrome, triglyceride or cholesterol levels and blood supports the existing evidence that ADHD represents an entry pressure; Supplementary Table 4). The 19 risk loci identified point into a negative life trajectory [3]. mapped 40 genes (Supplementary Table 5), which were enriched To provide potential pleiotropic molecular mechanisms under- in genes previously associated with cognitive performance, lying this association, we performed a cross-trait analysis on ADHD smoking and metabolite levels according to the GWAS catalog and parental lifespan and identified 19 independent loci jointly [33] (Supplementary Fig. 2), but no association with either associated with both traits, including 15 novel hits for ADHD biological pathways or differential tissue expression from GTEx [23, 24]. All of them but one (95%) showed consistent direction of were found. the effect, with risk alleles for ADHD shortening lifespan, which is Neuropsychopharmacology (2023) 48:981 – 990 L. Vilar-Ribó et al. 984 Fig. 1 Stratified cross-phenotype Q-Q plots. Nominal versus empirical (−log10) P values (corrected for inflation) are shown in A ADHD as a function of significance with parental lifespan and B parental lifespan as a function of significance with ADHD, at the level of P < 0.1 (red line), P < 0.01 (yellow line), and P < 1e−03 (purple line). The blue line indicates the standard enrichment of A ADHD or B parental lifespan including all SNPs, irrespective of their association with the secondary trait (i.e., parental lifespan or ADHD, respectively). The gray dashed line indicates the null distribution of P values. LD score regression intercepts for ADHD and parental lifespan full summary statistics were 1.04 and 1.05 respectively. PLS parental lifespan. Neuropsychopharmacology (2023) 48:981 – 990 L. Vilar-Ribó et al. 985 rs9927158 3 ADHD&PLS rs219218 rs6980093 rs1902330 rs11210887 rs57349798 rs2503773 rs10147849 rs7613360 rs10101132 2 rs770087 log10(conjFDR) rs7563362 rs11633054 rs10063426 rs2272272 rs12614708 rs281297 rs12441947 rs11780592 1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 1819 202122 Chromosome Fig. 2 Cross-trait analysis results. A Manhattan plot for independent (r2 < 0.05) loci associated with both ADHD and parental lifespan after excluding SNPs in the MHC and the 8p23.1 regions. The dashed black line represents the PFDR threshold of 0.05. B Pleiotropy plot. For lead SNPs (n = 19), P values and the direction of the effects (z-scores) of the derived alleles are plotted for parental lifespan (x-axis) against ADHD (y-axis). PLS parental lifespan. consistent with the negative genetic correlation between ADHD associated with ADHD and/or reduced life expectancy-related and lifespan, adding further evidence for the contribution of a phenotypes, including other psychiatric disorders (e.g., schizo- shared biological architecture. Interestingly, functional annotation phrenia and major depression), lifetime risk behaviors (e.g., of top hits from the cross-trait analysis highlighted loci previously number of sexual partners and risk tolerance) or metabolic Neuropsychopharmacology (2023) 48:981 – 990 986 Table 1. List of loci jointly associated with ADHD and parental lifespan (PLS) at conjunction FDR < 0.05. Locus SNP CHR BP nearestGene ANNOVAR CADD RDB minChr Effect Z-score P value Z-score P value conjFDR New locus in State allele ADHD ADHD PLS PLS ADHD_PLS ADHD PLS 1 rs11210887 1 44076019 PTPRF Intronic 12.97 6 4 A −6.41406 1.48E−10 3.84263 1.22E−04 1.80E−02 n y 2 rs2272272 1 110009802 SYPL2 Intronic 22.2 NA 1 A 3.65022 2.63E−04 3.45524 5.50E−04 4.77E−02 y y 3 rs7563362 2 620297 AC068490.2 ncRNA_intronic 0.133 6 5 A −3.67479 2.35E−04 3.51328 4.43E−04 4.18E−02 y n 4 rs12614708 2 22568238 AC009965.2 Intergenic 0.259 6 5 A 3.68099 2.39E−04 −3.4884 4.86E−04 4.43E−02 y y 5 rs219218 2 205007037 AC093326.3 Intergenic 0.041 5 2 T 4.15596 3.13E−05 −4.0265 5.66E−05 1.05E−02 y y 6 rs7613360 3 49916710 ACTBP13 Intergenic 2.031 5 5 T 4.11314 3.71E−05 −3.6695 2.43E−04 2.85E−02 y y 7 rs10063426 5 154775154 CTC-447K7.1 Intergenic 1.814 6 9 A −3.69334 2.17E−04 3.48961 4.84E−04 4.42E−02 y y 8 rs57349798 6 37486052 RP11- ncRNA_intronic 1.314 NA 9 A −4.5238 6.27E−06 3.852 1.17E−04 1.75E−02 y y 436D23.1 9 rs2503773 6 98537145 RP1- ncRNA_intronic 2.452 5 2 A −3.9189 9.52E−05 3.7788 1.58E−04 2.24E−02 y y 153P14.8 10 rs6980093 7 114162740 FOXP2 Intronic 0.41 6 5 A 4.9354 8.39E−07 −3.9201 8.85E−05 1.44E−02 n y 11 rs10101132 8 9616553 TNKS Intronic 9.165 7 4 A 3.80143 1.40E−04 −3.8661 1.11E−04 2.89E−02 y y 12 rs11780592 8 27418747 GULOP ncRNA_intronic 1.308 1f 4 A 3.62034 2.97E−04 −5.2548 1.48E−07 4.65E−02 y y 13 rs1902330 12 49939953 KCNH3 Intronic 8.355 5 5 A −4.26642 2.00E−05 3.90708 9.34E−05 1.50E−02 y y 14 rs770087 12 89744773 DUSP6 Exonic 23 NA 1 A −5.29553 1.23E−07 3.59893 3.20E−04 3.40E−02 n y 15 rs10147849 14 33304431 AKAP6 Intergenic 6.449 6 5 T −3.89787 1.00E−04 3.71885 2.00E−04 2.52E−02 y y 16 rs281297 15 47685504 SEMA6D Intronic 4.125 5 5 T −5.22589 1.89E−07 3.556 3.77E−04 3.78E−02 n y 17 rs11633054 15 77747276 HMG20A Intronic 7.219 7 4 A −3.59311 3.20E−04 4.16936 3.05E−05 4.87E−02 y y 18 rs12441947 15 81031886 ABHD17C ncRNA_intronic 0.885 7 5 T 3.75846 1.64E−04 −3.4859 4.90E−04 4.46E−02 y y L. Vilar-Ribó et al. 19 rs9927158 16 72256156 RP11- Intergenic 0.365 7 1 T 4.56195 4.86E−06 −4.961 7.01E−07 2.52E−03 y n 328J14.1 ANNOVAR functional variant classification based on position in or outside of a gene, CADD Combined Annotation-Dependent depletion score, which predict how deleterious the SNP effect is on protein structure/function, in bold scores >12.37 suggesting high deleteriousness, RBD RegulomeDB scores predict likelihood of regulatory functionality, in bold scores lower than 3 that indicate higher likelihood. Further information about RDB scores can be found in Table 2 from the original paper [32]; minChrState minimum chromatin state across 127 tissue types, in bold scores lower than 6 that indicate more open chromatin, Also shown the effect allele (used to calculate beta), p values and effect sizes from the original summary statistics on Attention-Deficit/Hyperactivity disoder (ADHD) and parental lifespan (PLS), NA Not available. Neuropsychopharmacology (2023) 48:981 – 990 L. Vilar-Ribó et al. 987 Fig. 3 Functional categories, Regulome DB scores, and Minimum Chromatin States for SNPs within loci jointly associated with ADHD and parental lifespan. Regulome DB score predicts likelihood of regulatory functionality, where lower scores indicate higher likelihood. Further information can be found in Boyle et al. 2012 [32]. Minimum Chromatin State across 127 tissue and cell types, lower scores indicate higher accessibility, with states 1–7 referring to open chromatin states. Fig. 4 Mendelian randomization results with ADHD as exposure and parental lifespan as outcome. Scatter plot of SNP effect estimates on ADHD vs. effect estimates on parental lifespan. Lines are drawn for inverse-variance weighted, weighted median and CAUSE, with the slope of each line corresponding to the estimated causal effect. Given that Igx2 = 0.36, MR-Egger results are not reported. alterations (e.g., metabolic syndrome, and triglyceride and encoding tankyrase1, a protein involved in telomere maintenance cholesterol levels). This further supports that shared risk factors [53], which is an essential cellular function closely related to may not be specific of ADHD or lifespan but also contribute to ageing and longevity [54]. Interestingly, several genetic variants in other related disorders and behaviors that could mediate, in part, this locus are cis-eQTL for TNKS in putamen, a key area in the basal this relationship. For instance, ADHD may directly increase ganglia previously related with ADHD [55–57]. In addition, TNKS propensity for risky behaviors and thus lead to a shorter life was previously associated with multiple psychiatric conditions expectancy [5]. (e.g., risk tolerance and neuroticism), neurological (e.g., Alzhei- The pleiotropic risk loci identified were enriched in introns and mer’s disease and epilepsy) and metabolic disorders (e.g., blood open chromatin regions, and almost half of them (42%) were cis- pressure, obesity, diabetes and stroke) [58–64], which are highly eQTLs tags for at least one gene in the brain, supporting their comorbid with ADHD and associated with increased morbidity putative causal effect. Among the identified risk loci, we highlight and mortality [65–67]. We also found genetic variants in other new relevant candidate genes. For instance, the TNKS gene two loci encompassing AKAP6 and SEMA6D genes, previously Neuropsychopharmacology (2023) 48:981 – 990 L. Vilar-Ribó et al. 988 associated with increased risk of schizophrenia, lower cognitive the disorder and that the role of ADHD on the overall life ability and educational attainment [68], all of them related with expectancy across age groups deserves further investigation. In shorter life expectancy [69–72]. In addition, we found other loci addition, we did not control our analysis for other potential including genes with relevant brain functions: SYPL2, which confounders that may bias our results. Increased mortality in encodes a vesicular transmembrane protein that belongs to the individuals with ADHD depends on sex or the presence of co- synaptophysin family, key elements for the regulation of neuronal occurring disorders [4, 8]. Also severity of ADHD symptoms, synaptic vesicles [73, 74]; and HMG20A, a non-histone chromoso- impairment and/or pharmacological treatment may influence life mal factor that regulates gene expression through changes in outcomes [83–85] and their role in the relationship between chromatin conformation, also involved in the regulation of ADHD, premature death and reduced overall life expectancy neuronal differentiation [75, 76]. deserve further investigation. Our results seem to reflect an effect of ADHD on premature In summary, our results are in agreement with observational mortality risk and are consistent with previous epidemiological studies supporting ADHD as the entrance into a detrimental life data describing a shortened lifespan in many mental disorders trajectory, support negative genetic correlation between ADHD and [4, 12]. A recent study suggests that there is no causal relationship lifespan and show common genetic loci shared between them, most between schizophrenia and parental lifespan [69], which can of which showed risk-increasing effects on ADHD and reduced indicate that our results may not be extensive to other psychiatric overall life expectancy. These results confirm the general pattern of disorders. Further studies are needed to clarify how specific the increased mortality rates and reduced overall life expectancy effect on parental lifespan is to ADHD. We also found suggestive associated with ADHD and highlight the need for further studies evidence of a mediating role for positive urgency (i.e. the on larger datasets to better understand the common genetic proclivity for rash action when feeling positive emotion) in the architecture underlying these complex relationships. relationship between the genetic liability of ADHD and lifespan. This is consistent with a reported predictive effect of positive urgency on ADHD-related traits such as illegal drug use and risky REFERENCES sexual behavior [77]. Nevertheless these findings should be 1. Polanczyk G, De Lima MS, Horta BL, Biederman J, Rohde LA. The worldwide interpreted with caution, since the evidence found for a causal prevalence of ADHD: a systematic review and metaregression analysis. Am J relationship of ADHD liability on shorter life expectancy was not Psychiatry. 2007;164:942–8. 2. Fayyad J, Sampson NA, Hwang I, Adamowski T, Aguilar-Gaxiola S, Al-Hamzawi A, conclusive. While IVW MR, together with all the sensitivity analyses et al. The descriptive epidemiology of DSM-IV Adult ADHD in the World Health performed and putative causality inferred with p‐HESS supported Organization World Mental Health Surveys. ADHD Atten Deficit Hyperact Disord. a negative causal effect of genetic liability of ADHD on lifespan, 2017;9:47–65. CAUSE did not confirm this causal relationship. This highlights that 3. Franke B, Michelini G, Asherson P, Banaschewski T, Bilbow A, Buitelaar JK, et al. inferring causal relationships between related traits remains a Live fast, die young? A review on the developmental trajectories of ADHD across challenge to date and that new methods and larger sample sizes the lifespan. Eur Neuropsychopharmacol. 2018;28:1059–88. are needed to understand the common genetic architecture 4. Dalsgaard S, Ostergaard SD, Leckman JF, Mortensen PB, Pedersen MG. Mortality underlying such complex relationships. in children, adolescents, and adults with attention deficit hyperactivity disorder: a Our results, however, should be interpreted in the context of nationwide cohort study. Lancet. 2015;385:2190–6. 5. Sun S, Kuja-Halkola R, Faraone SV, D’Onofrio BM, Dalsgaard S, Chang Z, et al. some considerations: Association of psychiatric comorbidity with the risk of premature death among First, parental lifespan of genotyped subjects was used as a children and adults with attention-deficit/hyperactivity disorder. JAMA Psy- surrogate trait for individual lifespan in the present generation. This chiatry. 2019;76:1141–9. kinship cohort design is supported by previous results reporting that 6. Prasad V, West J, Sayal K, Kendrick D. Injury among children and young people parental lifespan predicts the long-term mortality of their offspring. with and without attention-deficit hyperactivity disorder in the community: the However, the use of indirect genotypes reduces the effective sample risk of fractures, thermal injuries, and poisonings. Child Care Health Dev. size of the study [23, 78]. This, together with the modest heritability 2018;44:871–8. estimates for human lifespan (ranging from 7% to 16%) [23, 79] 7. Stickley A, Koyanagi A, Ruchkin V, Kamio Y. Attention-deficit/hyperactivity dis- limits the power of our study and precludes the identification of order symptoms and suicide ideation and attempts: findings from the Adult Psychiatric Morbidity Survey 2007. J Affect Disord. 2016;189:321–8. shared genetic variants with smaller effect sizes. 8. Barbaresi WJ, Colligan RC, Weaver AL, Voigt RG, Killian JM, Katusic SK. Mortality, Second, to avoid the potential bias introduced by using ADHD, and psychosocial adversity in adults with childhood ADHD: a prospective summary statistics generated through a survival analysis [37], study. Pediatrics. 2013;131:637–44. such as that conducted by Timmers et al. [23], we selected the 9. Catalá-López F, Hutton B, Page MJ, Driver JA, Ridao M, Alonso-Arroyo A, et al. results of Pilling et al. [36] to assess causality between ADHD and Mortality in persons with autism spectrum disorder or attention-deficit/hyper- parental lifespan. The smaller sample size of this study activity disorder. JAMA Pediatr. 2022;176:e216401. (n = 208,118) further reduced the statistical power and may have 10. Faraone SV. Attention deficit hyperactivity disorder and premature death. Lancet. led to the inconsistent results observed [36]. CAUSE was 2015;385:2132–3. developed with the aim of avoiding false positive results due to 11. Erskine HE, Norman RE, Ferrari AJ, Chan GCK, Copeland WE, Whiteford HA, et al. Long-term outcomes of attention-deficit/hyperactivity disorder and conduct correlated horizontal pleiotropy, but in the present study it did not disorder: a systematic review and meta-analysis. J Am Acad Child Adolesc Psy- detect a causal effect but neither did it detect this kind of chiatry. 2016;55:841–50. pleiotropy. Recent studies suggest that in some scenarios CAUSE 12. Nordentoft M, Wahlbeck K, Hällgren J, Westman J, Ösby U, Alinaghizadeh H, et al. tends to underestimate causal effect sizes in comparison to MR Excess mortality, causes of death and life expectancy in 270,770 patients with and produce overly conservative P values [80, 81], which may recent onset of mental disorders in Denmark, Finland and Sweden. PLoS One. suggest, in our case, limited power rather than lack of a causal 2013;8:e55176. relationship, supporting further studies with larger sample sizes to 13. Wilens TE, Spencer TJ. Understanding attention-deficit/hyperactivity disorder clarify this issue. from childhood to adulthood. Postgrad Med. 2010;122:97. Third, despite the shared genetic architecture underlying ADHD 14. Kendler KS, Ohlsson H, Sundquist J, Sundquist K. Alcohol use disorder and mortality across the lifespan: a longitudinal cohort and co-relative analysis. JAMA in children and adults [82], the risk of premature death also Psychiatry. 2016;73:575–81. depends on age at diagnosis, where individuals diagnosed with 15. Cortese S, Sun S, Zhang J, Sharma E, Chang Z, Kuja-Halkola R, et al. Association ADHD in adulthood appear to have a higher risk of death than do between attention deficit hyperactivity disorder and asthma: a systematic review those diagnosed in childhood or adolescence [4]. These data may and meta-analysis and a Swedish population-based study. The Lancet. Psychiatry. suggest that persistent ADHD represents a more severe form of 2018;5:717–26. Neuropsychopharmacology (2023) 48:981 – 990 L. Vilar-Ribó et al. 989 16. Chen MH, Pan TL, Hsu JW, Huang KL, Su TP, Li CT, et al. Risk of type 2 diabetes in 41. Bowden J, Davey, Smith G, Haycock PC, Burgess S. Consistent estimation in adolescents and young adults with attention-deficit/hyperactivity disorder: A Mendelian randomization with some invalid instruments using a weighted nationwide longitudinal study. J Clin Psychiatry. 2018;79:17m11607. median estimator. Genet Epidemiol. 2016;40:304–14. 17. Kapellen TM, Reimann R, Kiess W, Kostev K. Prevalence of medically treated 42. Bowden J, Davey Smith G, Burgess S. Mendelian randomization with invalid children with ADHD and type 1 diabetes in Germany - analysis of two repre- instruments: effect estimation and bias detection through Egger regression. Int J sentative databases. J Pediatr Endocrinol Metab. 2016;29:1293–7. Epidemiol. 2015;44:512–25. 18. Ptacek R, Stefano GB, Weissenberger S, Akotia D, Raboch J, Papezova H, et al. 43. Pickrell JK, Berisa T, Liu JZ, Ségurel L, Tung JY, Hinds DA. Detection and inter- Attention deficit hyperactivity disorder and disordered eating behaviors: links, pretation of shared genetic influences on 42 human traits. Nat Genet. risks, and challenges faced. Neuropsychiatr Dis Treat. 2016;12:571–9. 2016;48:709–17. 19. Barkley RA, Smith KM, Fischer M. ADHD risk genes involved in dopamine sig- 44. Major M, Freund MK, Burch KS, Mancuso N, Ng M, Furniss D, et al. Integrative naling and metabolism are associated with reduced estimated life expectancy at analysis of Dupuytren’s disease identifies novel risk locus and reveals a shared young adult follow-up in hyperactive and control children. Am J Med Genet Part genetic etiology with BMI. Genet Epidemiol. 2019;43:629–45. B Neuropsychiatr Genet. 2019;180:175–85. 45. Patton JH, Stanford MS, Barratt ES. Factor structure of the Barratt impulsiveness 20. Groenman AP, Oosterlaan J, Rommelse N, Franke B, Roeyers H, Oades RD, et al. scale. J Clin Psychol. 1995;51:768–74. Substance use disorders in adolescents with attention deficit hyperactivity dis- 46. Cyders MA, Littlefield AK, Coffey S, Karyadi KA. Examination of a short English order: a 4-year follow-up study. Addiction. 2013;108:1503–11. version of the UPPS-P Impulsive Behavior Scale. Addict Behav. 2014;39:1372–6. 21. Mohr-Jensen C, Steinhausen HC. A meta-analysis and systematic review of the 47. Morrison J, Knoblauch N, Marcus JH, Stephens M, He X. Mendelian randomization risks associated with childhood attention-deficit hyperactivity disorder on long- accounting for correlated and uncorrelated pleiotropic effects using genome- term outcome of arrests, convictions, and incarcerations. Clin Psychol Rev. wide summary statistics. Nat Genet. 2020;52:740–7. 2016;48:32–42. 48. Ertan C, Özcan ÖÖ, Pepele MS. Paediatric trauma patients and attention deficit 22. van den Berg N, Beekman M, Smith KR, Janssens A, Slagboom PE. Historical hyperactivity disorder: correlation and significance. Emerg Med J. 2012;29:911–4. demography and longevity genetics: back to the future. Ageing Res Rev. 49. Balazs J, Kereszteny A. Attention-deficit/hyperactivity disorder and suicide: a 2017;38:28–39. systematic review. World J Psychiatry. 2017;7:44. 23. Timmers PRHJ, Mounier N, Lall K, Fischer K, Ning Z, Feng X, et al. Genomics of 1 50. Curry AE, Metzger KB, Pfeiffer MR, Elliott MR, Winston FK, Power TJ. Motor vehicle million parent lifespans implicates novel pathways and common diseases and crash risk among adolescents and young adults with attention-deficit/hyper- distinguishes survival chances. Elife. 2019;8:1–40. activity disorder. JAMA Pediatr. 2017;171:756–63. 24. Demontis D, Walters RK, Martin J, Mattheisen M, Als TD, Agerbo E, et al. Discovery 51. London AS, Landes SD. Attention Deficit Hyperactivity Disorder and adult mor- of the first genome-wide significant risk loci for attention deficit/hyperactivity tality. Prev Med. 2016;90:8–10. disorder. Nat Genet. 2019;51:63–75. 52. Chen VCH, Chan HL, Wu SI, Lee M, Lu ML, Liang HY, et al. Attention-deficit/ 25. Bulik-Sullivan B, Finucane HK, Anttila V, Gusev A, Day FR, Loh PR, et al. An atlas of hyperactivity disorder and mortality risk in Taiwan. JAMA Netw Open. genetic correlations across human diseases and traits. Nat Genet. 2019;2:e198714. 2015;47:1236–41. 53. Damale MG, Pathan SK, Shinde DB, Patil RH, Arote RB, Sangshetti JN. Insights of 26. Finucane HK, Bulik-Sullivan B, Gusev A, Trynka G, Reshef Y, Loh P-R, et al. Parti- tankyrases: a novel target for drug discovery. Eur J Med Chem. 2020;207:112712. tioning heritability by functional annotation using genome-wide association 54. Riera CE, Merkwirth C, De Magalhaes Filho CD, Dillin A. Signaling networks summary statistics. Nat Genet. 2015;47:1228–35. determining life span. Annu Rev Biochem. 2016;85:35–64. 27. Auton A, Abecasis GR, Altshuler DM, Durbin RM, Abecasis GR, Bentley DR, et al. A 55. Wang Y, Zuo C, Xu Q, Hao L, Zhang Y. Attention-deficit/hyperactivity disorder is global reference for human genetic variation. Nature 2015;526:68–74. characterized by a delay in subcortical maturation. Prog Neuro- 28. Andreassen OA, Djurovic S, Thompson WK, Schork AJ, Kendler KS, O’Donovan Psychopharmacology. Biol Psychiatry. 2021;104:110044. MC, et al. Improved detection of common variants associated with schizophrenia 56. Hoogman M, Bralten J, Hibar DP, Mennes M, Zwiers MP, Schweren LSJ, et al. by leveraging pleiotropy with cardiovascular-disease risk factors. Am J Hum Subcortical brain volume differences in participants with attention deficit Genet. 2013;92:197–209. hyperactivity disorder in children and adults: a cross-sectional mega-analysis. The 29. Purcell S, Neale B, Todd-Brown K, Thomas L, Ferreira MAR, Bender D, et al. PLINK: Lancet. Psychiatry 2017;4:310–9. A tool set for whole-genome association and population-based linkage analyses. 57. Ceceli AO, Natsheh JY, Cruz D, Tricomi E. The neurobehavioral mechanisms of Am J Hum Genet. 2007;81:559–75. motivational control in attention-deficit/hyperactivity disorder. Cortex 30. Watanabe K, Taskesen E, Van Bochoven A, Posthuma D. Functional mapping and 2020;127:191–207. annotation of genetic associations with FUMA. Nat Commun. 2017;8:1–11. 58. Wachinger C, Nho K, Saykin AJ, Reuter M, Rieckmann A. A longitudinal imaging 31. Kircher M, Witten DM, Jain P, O’roak BJ, Cooper GM, Shendure J. A general genetics study of neuroanatomical asymmetry in Alzheimer’s Disease. Biol Psy- framework for estimating the relative pathogenicity of human genetic variants. chiatry. 2018;84:522–30. Nat Genet. 2014;46:310–5. 59. Song M, Liu J, Yang Y, Lv L, Li W, Luo XJ. Genome-wide meta-analysis identifies 32. Boyle AP, Hong EL, Hariharan M, Cheng Y, Schaub MA, Kasowski M, et al. two novel risk loci for epilepsy. Front Neurosci. 2021;15:722592. Annotation of functional variation in personal genomes using RegulomeDB. 60. Sung YJ, Winkler TW, de las Fuentes L, Bentley AR, Brown MR, Kraja AT, et al. A Genome Res. 2012;22:1790–7. large-scale multi-ancestry genome-wide study accounting for smoking behavior 33. MacArthur J, Bowler E, Cerezo M, Gil L, Hall P, Hastings E, et al. The new NHGRI-EBI identifies multiple significant loci for blood pressure. Am J Hum Genet. Catalog of published genome-wide association studies (GWAS Catalog). Nucleic 2018;102:375–400. Acids Res. 2017;45:D896–901. 61. Karlsson Linnér R, Biroli P, Kong E, Meddens SFW, Wedow R, Fontana MA, et al. 34. Aguet F, Brown AA, Castel SE, Davis JR, He Y, Jo B, et al. Genetic effects on gene Genome-wide association analyses of risk tolerance and risky behaviors in over 1 expression across human tissues. Nature. 2017;550:204–13. million individuals identify hundreds of loci and shared genetic influences. Nat 35. Ramasamy A, Trabzuni D, Guelfi S, Varghese V, Smith C, Walker R, et al. Genetic Genet. 2019;51:245–57. variability in the regulation of gene expression in ten regions of the human brain. 62. Luciano M, Hagenaars SP, Davies G, Hill WD, Clarke TK, Shirali M, et al. Association Nat Neurosci. 2014;17:1418–28. analysis in over 329,000 individuals identifies 116 independent variants influen- 36. Pilling LC, Kuo CL, Sicinski K, Tamosauskaite J, Kuchel GA, Harries LW, et al. cing neuroticism. Nat Genet. 2018;50:6. Human longevity: 25 genetic loci associated in 389,166 UK biobank participants. 63. Hill WD, Weiss A, Liewald DC, Davies G, Porteous DJ, Hayward C, et al. Genetic Aging. 2017;9:2504–20. contributions to two special factors of neuroticism are associated with affluence, 37. Cho Y, Rau A, Reiner A, Auer PL. Mendelian randomization analysis with survival higher intelligence, better health, and longer life. Mol Psychiatry. 2020;25:3034. outcomes. Genet Epidemiol. 2021;45:16–23. 64. Baselmans BML, Jansen R, Ip HF, van Dongen J, Abdellaoui A, van de Weijer MP, 38. Sanchez-Roige S, Fontanillas P, Elson SL, Gray JC, de Wit H, MacKillop J, et al. et al. Multivariate genome-wide analyses of the well-being spectrum. Nat Genet Genome-wide association studies of impulsive personality traits (BIS-11 and 2019 513. 2019;51:445–51. UPPSP) and drug experimentation in up to 22,861 adult research participants 65. Spann SJ, Ottinger MA. Longevity, Metabolic Disease, and Community Health. identify loci in the CACNA1I and CADM2 genes. J Neurosci. 2019;39:2562–72. Prog Mol Biol Transl Sci. 2018;155:1–9. 39. Hemani G, Zheng J, Elsworth B, Wade KH, Haberland V, Baird D, et al. The MR- 66. Mackenbach JP, Karanikolos M, Looman CWN. The rise of mortality from mental Base platform supports systematic causal inference across the human phenome. and neurological diseases in Europe, 1979-2009: Observational study. BMC Public Elife. 2018;7:e34408. Health. 2014;14:840. 40. Verbanck M, Chen C-Y, Neale B, Do R. Detection of widespread horizontal 67. Walker ER, McGee RE, Druss BG. Mortality in mental disorders and global disease pleiotropy in causal relationships inferred from Mendelian randomization burden implications: a systematic review and meta-analysis. JAMA Psychiatry. between complex traits and diseases. Nat Genet. 2018;50:693–8. 2015;72:334. Neuropsychopharmacology (2023) 48:981 – 990 L. Vilar-Ribó et al. 990 68. Lam M, Hill WD, Trampush JW, Yu J, Knowles E, Davies G, et al. Pleiotropic Meta- FUNDING Analysis of Cognition, Education, and Schizophrenia Differentiates Roles of Early This work was supported by the Agència de Gestió d’Ajuts Universitaris i de Recerca Neurodevelopmental and Adult Synaptic Pathways. Am J Hum Genet. (AGAUR, 2017SGR-00444 and 2017SGR-1461), the Instituto de Salud Carlos III (PI18/ 2019;105:334–50. 01788, P19/01224, PI20/00041, PI21/00612, PI22/00464, FI18/00285 to LVR and CP22/ 69. Muntané G, Farré X, Bosch E, Martorell L, Navarro A, Vilella E. The shared genetic 00128 to MSA), the Ministry of Science, Innovation and Universities (IJC2018-035346-I architecture of schizophrenia, bipolar disorder and lifespan. Hum Genet. to MSA and RYC2021-031324-I to JCD); the European Regional Development Fund 2021;140:441–55. (ERDF), the European Union H2020 Programme (H2020/2014–2020) under grant 70. Olfson M, Gerhard T, Huang C, Crystal S, Stroup TS. Premature mortality among agreements no. 728018 (Eat2beNICE), no. 848228 (DISCOvERIE) and no. 2020604 adults with schizophrenia in the United States. JAMA Psychiatry. 2015;72:1172–81. (TIMESPAN) and the ECNP Network ‘ADHD across the Lifespan’ and “la Marató de 71. Danler C, Pfaff K. The impact of an unequal distribution of education on TV3” (202228-30 and 202228-31). SSR was supported by funds from the California inequalities in life expectancy. SSM Popul Heal. 2021;16:100954. Tobacco-Related Disease Research Program (TRDRP; Grant Number T29KT0526 and 72. Halpern-Manners A, Raymo JM, Warren JR, Johnson KL. School performance and T32IR5226), and NIH/NIDA DP1DA054394. AAP was supported by NIH grant mortality: the mediating role of educational attainment and work and family P50DA037844 and R01AA029688 and Tobacco-Related Disease Research Program trajectories across the life course. Adv Life Course Res. 2020;46:100362. Grant number: 28IR-0070. 73. White DN, Stowell MHB. Room for two: the synaptophysin/synaptobrevin com- plex. Front Synaptic Neurosci. 2021;13:47. 74. Kwon SE, Chapman ER. Synaptophysin regulates the kinetics of synaptic vesicle COMPETING INTERESTS endocytosis in central neurons. Neuron. 2011;70:847–54. JARQ was on the speakers bureau and/or acted as consultant for Janssen-Cilag, 75. Ceballos-Chávez M, Rivero S, García-Gutiérrez P, Rodríguez-Paredes M, García- Novartis, Shire, Takeda, Bial, Shionogi, Sincrolab, Novartis, BMS, Medice, Domínguez M, Bhattacharya S, et al. Control of neuronal differentiation by Technofarma, Rubió and Raffo in the last 3 years. He also received travel awards sumoylation of BRAF35, a subunit of the LSD1-CoREST histone demethylase (air tickets + hotel) for taking part in psychiatric meetings from Janssen-Cilag, complex. Proc Natl Acad Sci USA. 2012;109:8085–90. Rubió, Shire, Takeda, Shionogi, Bial and Medice. The Department of Psychiatry 76. Artegiani B, Labbaye C, Sferra A, Quaranta MT, Torreri P, Macchia G, et al. The Mental Health chaired by him received unrestricted educational and research interaction with HMG20a/b proteins suggests a potential role for β-dystrobrevin support from the following companies in the last 3 years: Janssen- Cilag, Shire, in neuronal differentiation. J Biol Chem. 2010;285:24740–50. Oryzon, Roche, Psious, and Rubió. AAP is on the scientific advisory board, owns 77. Zapolski TCB, Cyders MA, Smith GT. Positive urgency predicts illegal drug use and stock options and a patent (https://patents.google.com/patent/US20160038559A1) risky sexual behavior. Psychol Addict Behav. 2009;23:348–54. for Vivid Genomics (https://www.vividgenomics.com/). All other authors declare no 78. Vågerö D, Aronsson V, Modin B. Why is parental lifespan linked to children’s competing interests. chances of reaching a high age? A transgenerational hypothesis. SSM - Popul Heal. 2018;4:45–54. 79. Kaplanis J, Gordon A, Shor T, Weissbrod O, Geiger D, Wahl M, et al. Quantitative ADDITIONAL INFORMATION analysis of population-scale family trees with millions of relatives. Science. 2018;360:171–5. Supplementary information The online version contains supplementary material 80. Darrous L, Mounier N, Kutalik Z. Simultaneous estimation of bi-directional causal available at https://doi.org/10.1038/s41386-023-01555-x. effects and heritable confounding from GWAS summary statistics. Nat Commun. 2021;12:7274. Correspondence and requests for materials should be addressed to Gerard Muntané 81. Yuan Z, Liu L, Guo P, Yan R, Xue F, Zhou X. Likelihood-based Mendelian rando- or María Soler Artigas. mization analysis with automated instrument selection and horizontal pleiotropic modeling. Sci Adv. 2022;8:eabl5744. Reprints and permission information is available at http://www.nature.com/ 82. Rovira P, Demontis D, Sánchez-Mora C, Zayats T, Klein M, Mota NR, et al. Shared reprints genetic background between children and adults with attention deficit/hyper- activity disorder. Neuropsychopharmacology. 2020;45:1617–26. Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims 83. McCarthy S, Cranswick N, Potts L, Taylor E, Wong ICK. Mortality associated with in published maps and institutional affiliations. attention-deficit hyperactivity disorder (ADHD) drug treatment: A retrospective cohort study of children, adolescents and young adults using the general prac- tice research database. Drug Saf. 2009;32:1089–96. 84. Chang Z, Lichtenstein P, D’Onofrio BM, Sjölander A, Larsson H. Serious transport accidents in adults with attention-deficit/hyperactivity disorder and the effect of Open Access This article is licensed under a Creative Commons medication a population-based study. JAMA Psychiatry. 2014;71:319–25. Attribution 4.0 International License, which permits use, sharing, 85. Lichtenstein P, Halldner L, Zetterqvist J, Sjölander A, Serlachius E, Fazel S, et al. adaptation, distribution and reproduction in any medium or format, as long as you give Medication for Attention Deficit–Hyperactivity Disorder and Criminality. N Engl J appropriate credit to the original author(s) and the source, provide a link to the Creative Med. 2012;367:2006–14. Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory AUTHOR CONTRIBUTIONS regulation or exceeds the permitted use, you will need to obtain permission directly GM, MR and MSA conceived the project, GM, MR and MSA participated in the study from the copyright holder. To view a copy of this license, visit http:// design, JCD, GM and LVR and participated in data acquisition, JCD, GM, MSA and LVR creativecommons.org/licenses/by/4.0/. undertook the statistical analyses, JARQ, JCD, LM, GM, MR, MSA, LVR, EV, SSR and AAP. participated in the manuscript preparation. All authors contributed to the interpretation of the findings and revised and approved the final version of the manuscript. © The Author(s) 2023 Neuropsychopharmacology (2023) 48:981 – 990