Both Taiwan and Sweden have universal single-payer health insurance systems that in effect track their entire national populations. With detailed health and other records on millions of individuals, with no significant exclusions, one can essentially eliminate sampling error, and also explore how associations vary by degree of familial/genetic relationship.

A Taiwanese research team used the Taiwan National Health Insurance Research Database to follow 708,517 family triads (father-mother-child) from 2001 through 2011. That's a total of over 2.1 million persons. The database covers over 99% of Taiwan's population.

Noting that previous studies had found links between maternal autoimmune diseases and ADHD in their offspring and that research on associations with paternal autoimmune diseases had been inconclusive, they were particularly interested in exploring the latter.

Children born from 2001 through 2008 were enrolled in the study. The investigators then noted the presence or absence of any autoimmune disease in their parents from 1996 through childbirth.

In Taiwan, expert panels review diagnostic information of severe systemic autoimmune diseases to confirm the diagnosis. Once confirmed, patient co-payments are waived. ADHD diagnoses are by board-certified psychiatrists.

To reduce the effect of confounding variables, adjustments were made for family demographic data (income level and residence), parental ages, parental mental disorders, and sex of children.

The presence of any maternal autoimmune diseases was associated with a 60% greater risk of ADHD in offspring. The risk was especially elevated for inflammatory bowel diseases (2.4 times the risk) and ankylosing spondylitis (twice the risk).

The presence of any paternal autoimmune diseases was also associated with an elevated risk of ADHD in offspring, although only about half as much as for maternal autoimmune diseases, with a 33% greater risk overall. The association was especially pronounced for psoriasis and ankylosing spondylitis, both doubling the risk of ADHD in offspring.

Meanwhile, half a world away, a joint Swedish, Norwegian, and U.S. team used the Swedish national registries to dig further into these associations. They did this by examining data not only from mothers and fathers, but from full siblings, aunts, uncles, and cousins as well, to probe genetic links.

The team used the Swedish registers to identify 5,178,225 individuals born in Sweden between 1960 and 2010 for whom the identity of the biological mother was known, excluding all who died or emigrated before age 10. They then used the registers to identify the aforementioned relatives.

The researchers only included autoimmune diseases with at least two thousand diagnosed individuals in the cohort, to avoid small sample effects.

They adjusted for sex and year of birth, but not "for another covariate that is often adjusted for (e.g. maternal education, family income, parental psychiatric disorder, parental AD [autoimmune disease] as these are likely not true confounders of the association between ADHD and ADD, but may rather represent either mediator between ADHD and AD's, or proxies of ADHD and/or AD risk or alternatively proxies for the associations we aim to measure."

The team found statistically significant associations between ADHD and autoimmune diseases in all categories of relatives. Mothers of children with ADHD were 29% more likely to have an autoimmune disease than those of typically developing children; fathers were 14% more likely to have an autoimmune disease; full siblings 19% more likely; aunts 12% more likely; uncles 7% more likely; and cousins 4% more likely.

Quantitative genetic modeling produced a significant genetic correlation, but no significant environmental correlation. Genetic correlation explained most, if not all, the covariance between ADHD and any autoimmune disease.

The authors concluded, "ADHD was to some degree more strongly associated with maternal than paternal AD's, but by using aunts and uncles in a genetically informative study design, we demonstrate that this difference cannot be readily explained by AD-mediated maternal effects. Quantitative genetic modeling further indicates that the familial co-aggregation of ADHD and ADs is partly due to shared genetic factors. In addition, biological aunts, uncles, and cousins must be assumed to share the little environment with the index individuals, in further support of shared genetic factors underlying the familial co-aggregation. Moreover, both epidemiological and molecular genetics studies have demonstrated positive genetic correlations between ADHD and ADs, in agreement with our findings."

The authors emphasize that these results do not warrant screening for autoimmune diseases among asymptomatic individuals with ADHD.

A Swedish-Danish-Dutch team used the Swedish Medical Birth Register to identify the almost 1.7 million individuals born in the country between 1980 and 1995. Then, using the Multi-Generation Register, they identified 341,066 pairs of full siblings and 46,142 pairs of maternal half-siblings, totaling 774,416 individuals.

The team used the National Patient Register to identify diagnoses of ADHD, as well as neurodevelopmental disorders (autism spectrum disorder, developmental disorders, intellectual disability, motor disorders), externalizing psychiatric disorders (oppositional defiant and related disorders, alcohol misuse, drug misuse), and internalizing psychiatric disorders (depression, anxiety disorder, phobias, stress disorders, obsessive-compulsive disorder).

The team found that ADHD was strongly correlated with general psychopathology overall (r =0.67), as well as with the neurodevelopmental (r = 0.75), externalizing (r =0.67), and internalizing (r = 0.67) sub factors.

To tease out the effects of heredity, shared environment, and non-shared environment, a multivariate correlation model was used. Genetic variables were estimated by fixing them to correlate between siblings at their expected average gene sharing (0.5for full siblings, 0.25 for half-siblings). Non-genetic environmental components shared by siblings (such as growing up in the same family) were estimated by fixing them to correlate at 1 across full and half-siblings. Finally, non-shared environmental variables were estimated by fixing them to correlate at zero across all siblings.

This model estimated the heritability of the general psychopathology factor at 49%, with the contribution of the shared environment at 7 percent and the non-shared environment at 44%. After adjusting for the general psychopathology factor, ADHD showed a significant and moderately strong phenotypic correlation with the neurodevelopmental-specific factor (r = 0.43), and a significantly smaller correlation with the externalizing-specific factor (r = 0.25).

For phenotypic correlation between ADHD and the general psychopathology factor, genetics explained 52% of the total correlation, the non-shared environment 39%, and the shared familial environment only 9%. For the phenotypic correlation between ADHD and the neurodevelopmental-specific factor, genetics explained the entire correlation because the other two factors had competing effects that canceled each other out. For the phenotypic correlation between ADHD and the externalizing-specific factor, genetics explained 23% of the correlation, shared environment 22%, and non-shared environment 55%.

The authors concluded that "ADHD is more phenotypically and genetically linked to neurodevelopmental disorders than to externalizing and internalizing disorders, after accounting for a general psychopathology factor. ... After accounting for the general psychopathology factor, the correlation between ADHD and the neurodevelopmental-specific factor remained moderately strong, and was largely genetic in origin, suggesting substantial unique sharing of biological mechanisms among disorders. In contrast, the correlation between ADHD and the externalizing-specific factor was much smaller and was largely explained by-shared environmental effects. Lastly, the correlation between ADHD and the internalizing subfactor was almost entirely explained by the general psychopathology factor. This finding suggests that the comorbidity of ADHD and internalizing disorders are largely due to shared genetic effects and non-shared environmental influences that have effects on general psychopathology."

Behavioral disinhibition is a trait associated with both ADHD and several genes that affect dopamine signaling. A new study by three American medical researchers set out to examine how these ADHD risk genes - DRD4 (dopamine 4 receptor density), DAT1 (dopamine 1 transporter), and DBH(dopamine beta-hydroxylase) - affect estimated life expectancy in young adulthood.

The method used was a longitudinal study of 131 hyperactive children and 71 matched controls through early adulthood. The original evaluations were done in 1979-1980, when both groups were children in the 4 to 12 age range. They were reevaluated in1987-1988 as adolescents aged 12 to 20. The next follow-up was in 1992-1996 in early adulthood, aged 19 to 25. The final follow-up was in 1998-2004, for adults aged 24 to 32. All agreed to physical examinations that formed the basis for calculating estimated life expectancy using actuarial tables that factor in the effects of smoking, body mass index, alcohol, and other risk factors of on expected longevity. Participants also provided blood samples that enabled gene typing.

For the DAT1 gene, participants who had the homozygous-repeat allele (9/9) had a five-year reduction in estimated life expectancy relative to those with the ten-repeat allele (10/10). Those with the intermediate (9/10) configuration had a three-year reduction in estimated life expectancy.

For the DBH Taq1 gene, those with a heterozygous (A1/A2) combination had almost a three-year reduction in estimated life expectancy relative to those with homozygous (A1/A1 or A2/A2)configurations.

For DRD4, on the other hand, no significant differences were found in estimated life expectancy.

In a related study, several background traits were found to be significantly predictive of variance estimated life expectancy. The largest of these was behavioral disinhibition, followed by verbal IQ, self-rated hostility, and a nonverbal fluency test. But no significant differences were found between any of the gene polymorphisms on any of these four measures, indicating that the present gene associations were independent of the background traits.

The researchers next sought to determine which variables used in the estimated life expectancy calculations were associated with the two significant genes. For DBH, one variable stood out. Those with the A1/A2 heterozygous pairings had almost twice the alcohol consumption of those with homozygous pairings (p = 0.023).

For DAT1, two variables stood out. Overall, the 9/9 pairings smoked two and a half times as much as the 10/10pairings, with the 9/10 pairings midway between the extremes (p = 0.036). They were also 73 percent more likely to be smokers relative to the 10/10 pairings, and 61 percent more likely relative to the 9/10 pairings. They also had significantly less education than the 10/10 pairings, with the 9/10 pairings again being intermediate (p = 0.027).

An obvious limitation of the study was its small sample size. The authors cautioned, our findings should be considered quite preliminary and in need of much greater research before being given much weight in the literature or public policy.

"With these limitations in mind, they concluded, the present study demonstrated that two ADHD risk genes (DB Hand DAT1) independently contributed to a reduction in ELE [estimated life expectancy] beyond the second-order variables of behavioral disinhibition, IQ, hostility, and nonverbal fluency that contributed in the related study to variation in ELE. The gene polymorphisms seemed to be influencing ELE through their affiliation with first-order or more proximal factors related to ELE such as education, smoking, alcohol use, and possibly exercise."

Our genes are very important for the development of mental disorders-including ADHD, where genetic factors capture up to 75% of the risk. Until now, the search for these genes had yet to deliver clear results.   In the 1990s, many of us were searching for genes that increased the risk for ADHD because we know from twin studies that ADHD had a robust genetic component.  Because I realized that solving this problem required many DNA samples from people with and without ADHD, I created the ADHD Molecular Genetics Network, funded by the US NIMH.  We later joined forces with the Psychiatric Genomics Consortium (PTC) and the Danish psych group, which had access to many samples.  
The result is a study of over 20,000 people with ADHD and 35,000 who do not suffer from it - finding twelve locations (loci) where people with a particular genetic variant have an increased risk of ADHD compared to those who do not have the variant.  The results of the study have just been published in the scientific journal Nature Genetics, https://www.nature.com/articles/s41588-018-0269-7.
These genetic discoveries provide new insights into the biology behind developing ADHD. For example, some genes have significance for how brain cells communicate with each other, while others are important for cognitive functions such as language and learning.
Our study used the genome-wide association study (GWAS)methodology because it allowed us to discover genetic loci anywhere on the genome.  The method assays DNA variants throughout the genome and determines which variants are more common among ADHDvs. control participants.  It also allowed for the discovery of loci having very small effects.  That feature was essential because prior work suggested that, except for very rare cases, ADHD risk loci would individually have small effects.
The main findings are:

A) we found 12 loci on the genome that we can be certain harbor DNA risk variants for ADHD.  None of these loci were traditional candidate genes' for ADHD, i.e., genes involved in regulating neurotransmission systems that are affected by ADHD medications.  Instead, these genes seem to be involved in the development of brain circuits.  
B) we found a significant polygenic etiology in our data, which means that there must be many loci(perhaps thousands) having variants that increase the risk for ADHD.  We will need to collect a much larger sample to find out which specific loci are involved;

We also compared the new results with those from a genetic study of continuous measures of ADHD symptoms in the general population. We found that the same genetic variants that give rise to an ADHD diagnosis also affect inattention and impulsivity in the general population.  This supports prior clinical research suggesting that, like hypertension and hypercholesteremia, ADHD is a continuous trait in the population.  These genetic data now show that the genetic susceptibility to ADHD is also a quantitative trait comprised of many, perhaps thousands, of DNA variants
The study also examined the genetic overlap with other disorders and traits in analyses that ask the questions: Do genetic risk variants for ADHD increase or decrease the likelihood a person will express other traits and disorders.   These analyses found a strong negative genetic correlation between ADHD and education. This tells us that many of the genetic variants which increase the risk for ADHD also make it more likely that a person will perform poorly in educational settings. The study also found a positive correlation between ADHD and obesity, increased BMI, and type-2 diabetes, which is to say that variants that increase the risk of ADHD also increase the risk of overweight and type-2 diabetes in the population. This work has laid the foundation for future work that will clarify how genetic risks combine with environmental risks to cause ADHD.  When the pieces of that puzzle come together, researchers will be able to improve the diagnosis and treatment of ADHD.

Behavioral disinhibition is a trait associated with both ADHD and several genes that affect dopamine signaling. Anew study by three American medical researchers set out to examine how threaded risk genes – DRD4 (dopamine 4 receptor density), DAT1 (dopamine 1transporter), and DBH (dopamine beta-hydroxylase) – affect estimated life expectancy in young adulthood.

The method used was a longitudinal study of 131 hyperactive children and 71 matched controls through early adulthood. The original evaluations were done in 1979-1980, when both groups were children in the 4 to 12 age range. They were reevaluated in1987-1988 as adolescents aged 12 to 20. The next follow-up was in 1992-1996 in early adulthood, aged 19 to 25. The final follow-up was in 1998-2004, as adults aged 24 to 32. All agreed to physical examinations that formed the basis for calculating estimated life expectancy using actuarial tables that factor in the effects of smoking, body mass index, alcohol, and other risk factors on expected longevity. Participants also provided blood samples that enabled gene typing.

For the DAT1gene, participants who had the homozygous nine-repeat allele (9/9) had an a five-year reduction in estimated life expectancy relative to those with the ten-repeat allele (10/10). Those with the intermediate (9/10) configuration had a three-year reduction in estimated life expectancy.

For the DBHTaq1 gene, those with a heterozygous (A1/A2) combination had almost a three-year reduction in estimated life expectancy relative to those with homozygous (A1/A1 or A2/A2) configurations.

For DRD4, on the other hand, no significant differences were found for estimated life expectancy.

In a related study, several background traits were found to be significantly predictive of variance in estimated life expectancy. The largest of these was behavioral disinhibition, followed by verbal IQ, self-rated hostility, and a nonverbal fluency test. But no significant differences were found between any of the gene polymorphisms on any of these four measures, indicating that the present gene associations were independent of the background traits.

The researchers next sought to determine which variables used in the estimated life expectancy calculations were associated with the two significant genes. For DBH, one variable stood out. Those with the A1/A2 heterozygous pairings had almost twice the alcohol consumption of those with homozygous pairings (p = 0.023).

For DAT1, two variables stood out. Overall, the 9/9 pairings smoked two and a half times as much as the 10/10 pairings, with the 9/10 pairings midway between the extremes(p = 0.036). They were also 73 percent more likely to be smokers relative to the 10/10 pairings, and 61 percent more likely relative to the 9/10 pairings. They also had significantly less education than the 10/10 pairings, with the 9/10pairings again being intermediate (p = 0.027).

An obvious limitation of the study was its small sample size. The authors cautioned, “our findings should be considered quite preliminary and in need of much greater research before being given much weight in the literature or in public policy.

“With these limitations in mind,” they concluded, “the present study demonstrated that two ADHD risk genes (DBH and DAT1) independently contributed to a reduction in ELE [estimated life expectancy] beyond the second order variables of behavioral disinhibition, IQ, hostility, and nonverbal fluency that contributed in the related study to variation in ELE. The gene polymorphisms seemed to be influencing ELE through their affiliation with first-order or more proximal factors related to ELE such as education, smoking, alcohol use, and possibly exercise.”