|Home | About | Journals | Submit | Contact Us | Français|
Currently, ADHD treatment is often determined empirically through trial and error until an adequate response is obtained or side effects occur. ADHD is highly heritable and there is wide individual variability in response to ADHD medications, suggesting that the mechanism of action of stimulant medications may provide clues for genetic predictors of response. The promise of ADHD pharmacogenetics is far reaching, and includes the potential to develop individualized medication regimens that improve symptom response, decrease risk for side effects, improve long-term tolerability, and thus contribute to long-term treatment compliance and improved general effectiveness. Early ADHD pharmacogenetic studies have focused predominantly on catecholamine pathway genes and response to methylphenidate. Future efforts will also examine a wider range of stimulant and non-stimulant medications on a range of outcome measures and time periods. Based upon these studies, the potential for personalizing ADHD treatment in clinical practice will be determined.
Attention-Deficit/Hyperactivity Disorder (ADHD) is a common and treatable behavioral syndrome that typically emerges during childhood or adolescence and often persists into adulthood. There are an increasing number of stimulant and non-stimulant medication options for ADHD, with numerous new compounds in development . In acute treatment studies, stimulant medications generally demonstrate large effect sizes on ADHD symptom reduction relative to placebo, with slightly smaller effect sizes in adult studies . Surprisingly, only a minority of children and adolescents with ADHD remain on medication consistently , despite persistence of symptoms and impairments. Even among responders, there is marked variability in optimal dosage, duration of effect, and tolerability. Moreover, in spite of acute symptom reduction from ADHD medications, as yet there is little evidence of long-term response and improvement in functioning among those children who received treatment [4, 5]. Furthermore, there are few reliable predictors of medication response. In the absence of these data, treatment is often determined “empirically” in clinical practice through a gradual titration of different dosages and trial and error approach to different medications.
ADHD is a highly heritable disorder . The search for candidate genes associated with ADHD has been largely driven by the understanding that medications for the disorder have drug targets in the catecholamine system . Conversely, it is likely that variability in individual drug response might be related to genetic factors. Spurred by completion of the human genome project  there is considerable interest in the translation of molecular genetics research into meaningful clinical applications. Emerging findings in ADHD pharmacogenetics and pharmacogenomics underlie attempts to apply molecular genetics findings to optimize individual patient therapies. Two prior reviews summarized findings in ADHD pharmacogenetics [9, 10]. This chapter will review definitions, summarize and update research advances, make recommendations for future investigations, and discuss the potential clinical implications of pharmacogenetics and pharmacogenomics in the management of ADHD.
Pharmacogenetics is the study of genetic variability in medication response . Pharmacogenetics encompasses how genetic variation influences both pharmacokinetics, particularly drug metabolism, and pharmacodynamics, in terms of symptom response and side effect profiles. Pharmacogenetics had its l beginnings in the 1950’s when clinicians noted increased patterns of adverse reactions within certain families and ethnic groups. In contrast to pharmacogenetics, pharmacogenomics is a more recent term, which broadly encompasses efforts to utilize the human genome to better understand and develop pharmacological treatments . Pharmacogenomics refers more specifically to the study of variations in genes and gene products and their relationship to medication response . Based on advances in molecular biology, such as the advent of economical, high throughput gene sequencing and genotyping, and the mapping of the human genome, pharmacogenetic and pharmacogenomic studies have the potential to inform individualized treatment decisions in medicine, and subsequently improve long-term patient outcomes . The promise of ADHD pharmacogenetics is far reaching, and includes the potential to develop individualized medication regimens that improve symptom response, decrease risk for side effects, improve long-term tolerability, and thus contribute to long-term treatment compliance and improved general effectiveness.
Obvious candidates for pharmacogenomic investigations include polymorphic drug-metabolizing enzymes, drug transporters, and targets that affect disease and drug-related pathways . The dramatic and rapid effects of stimulant medications in the management of ADHD symptoms suggests several strong candidates for pharmacogenetic analysis. Numerous candidate gene studies have examined the relationships between polymorphisms in the dopaminergic system and ADHD susceptibility, although only a small percentage of the variance in ADHD susceptibility can be attributed to any one gene . Genome-wide linkage and association studies, including haplotype mapping, also have potential relevance to ADHD pharmacogenomics . With increased identification of single nucleotide polymorphisms (SNP’s), SNP tagging of common haplotypes can be economically used to identify regions of interest which better represent common genetic variation in association studies with drug response .
Candidate genes associated with increased risk for ADHD based on pooled odds ratios across three or more studies are the dopamine receptors (DRD4 and DRD5), dopamine transporter (DAT1), dopa-β-hydroxylase (DBH), serotonin receptor (HTR1B), serotonin transporter (5-HTT), and synaptosomal-associated protein (SNAP-25) . Other genes of potential interest in pharmacogenetic studies of ADHD include catehol-O-methyltransferase, (COMT), and the adrenergic α2-receptor (ADRA2A). In each of these genes, deep resequencing and association analysis has not been reported to date.
While knowledge about the presumed mechanisms of action of ADHD medications initially informed searches for polymorphisms related to increased risk for the disorder, these same polymorphisms are logical candidates to predict medication outcomes . Several preliminary studies suggest that candidate genes involved in catecholamine pathways, i.e. genes related to dopamine or norepinephrine, influence individual patient responses to ADHD treatments (Table 1). However, results from several of these reports are contradictory and the nature, magnitude, and direction of purported genetic effects remain unclear. Moreover, the majority of studies have examined predictors of response to methylphenidate, and little attention has been given to other agents. Nonetheless, this seminal research suggests that genetic variability does contribute to treatment response variability in individual patients and provides a foundation for definitive investigations in larger trials.
Several reasons support the choice of DAT1 as a candidate gene for ADHD treatment response. An association between ADHD and the 10-repeat (480-base-pair) allele of a variable number tandem repeat (VNTR) in the 3′-untranslated region (3′-UTR) of DAT1 was initially described in 1995 . and replicated in multiple, but not all, follow-up reports . Methylphenidate, and to some extent amphetamine, specifically targets and blocks the dopamine transporter [19, 20]. Numerous neuroimaging studies reveal that ADHD patients express increased dopamine transporter densities in striatal regions , and that individuals with the 10-repeat allele exhibit approximately 50% greater densities than other genotypes . This suggests that ADHD medications which block the dopamine transporter densities might serve to attenuate the effects of underlying brain pathophysiology. Similarly, functional polymorphisms in DAT1 might influence response to ADHD medications.
The first pharmacogenetic study of ADHD examining the ability of catecholamine related genes to predict response to methylphenidate . This report found a significant difference in response rate among 30 stimulant naïve, African-American youth based on DAT1 genotype. Specifically, 86% of nonresponders were homozygous for the 10-repeat allele compared with 31% of responders (N = 16, 31%) (χ2 (1) = 6.9, p=.008). Two subsequent studies reported a similar relationship between DAT1 genotypes genotype and symptom response. The first examined symptom reductions in 50 European-Brazilian males with ADHD after open titration with methylphenidate up to .7mg/kg/day . Response was defined as ≥ 50% reduction in baseline ADHD ratings, and individuals who failed to meet this threshold were more likely to be homozygous for the 10-repeat allele (Fisher’s exact test; one tailed p=.04). A second study that assessed dopamine transporter binding in addition to ADHD symptom reduction in 11 Korean subjects found that only 27% of subjects homozygous for the 10-repeat allele met response criteria compared with 100% of subjects without this genotype (χ2 (1)=5.2, p=.06) .
In contrast to the above, one retrospective and three prospective, placebo controlled studies reported improved response to methylphenidate in ADHD youth who were homozygous or heterozygous for the 10-repeat polymorphism. One analysis based on parental retrospective report in 119 Irish children found that individuals with one or two copies of the 10-repeat allele were more likely to have improved, and a linear relationship existed between the number of 10-repeats and degree of improvement . Similarly, the presence of one or two 10-repeat alleles was associated with higher rates of symptom reduction and reduced impairment in 47 children and adolescents treated with 18 mg., 36mg. and 54mg of OROS methylphenidate . Individuals homozygous for the less common 9-repeat allele demonstrated a non-linear dose-response curve, had more stimulant related side effects, and remained more impaired during treatment
Similar findings were reported for the 9/9 genotype group in a double-blind, placebo controlled trial of 159 children with ADHD conducted in Montreal . Although children with either the 9/10 and 10/10 genotypes displayed a significant positive response to 10 mg. of MPH, the 9/9 genotype group displayed a negative response on parent symptom ratings only ( effect size = −. 43. Interestingly, the effect of DAT1 genotype on response was not found on teacher ratings.
In one of the few studies not conducted with methylphenidate. Lott et al  reported that college student volunteers with the 9/9 were less able to “feel” amphetamine effects relative to other genotype groups. It should be noted that this was not an ADHD sample, and consisted of college student volunteers. These recent studies suggest a differential effect of stimulants on individuals homozygous for the 9-repeat relative to genotypes containing the 10-repeat.
However, several uncontrolled studies found no association between DAT1 and methylphenidate response, including a sample of 82 Dutch children treated prospectively with < .6 mg/kg/day , and a retrospective analysis of 186 youth with ADHD (dose not specified) in the United Kingdom . In addition, an attempt to replicate previous findings in Brazilian youth  examined response rates during open label treatment in 111 subjects detected no association between DAT1 genotype and outcomes .
In a study of 81 preschoolers with ADHD treated with methylphenidate, there were no genotype effects of DAT1 on a composite measure based on parent and teacher symptom ratings [33}, However, on parent ratings of ADHD symptoms, there was negative effect size ( −. 58) for the homozygous 9 repeat genotype. Of note, the difference between parent and teacher ratings iand pharmacogenetic effects of DAT1 is similar to that reported by Joober et al .
Recently, the first pharmacogenetic study of adults with ADHD reported no relationship between DAT1 genotype and response in 66 subjects titrated to a maximum methylphenidate dose of 1.3mg/kg/day . However, the sample included only 3 individuals with the 9/9 genotype, and llmited statistical power to detect an effect for this genotype group.
The dopamine receptor DRD4 is a presumed target of post-synaptic catecholaminergic activity and a likely candidate for predicting variability in medication response. The association of the 7-repeat (48-base-pair) VNTR polymorphism in the coding region of DRD4 with ADHD is one of the most replicated findings in psychiatric genetics, yielding odds ratios ranging from 1.4 to 1.9 . In vitro studies further indicate that the 7-repeat allelic variant is functionally less responsive to dopamine [35, 36]. Consistent with this idea, one study demonstrated in 45 subjects that patients with at least one copy of the 7-repeat allele required higher doses of methylphenidate for optimal symptom reduction . Conversely, in a separate report on 83 Korean children, subjects who were homozygous for the 4-repeat polymorphism were much more likely to exhibit positive responses on parent and teacher behavioral ratings than those with other genotypes . Other studies have found no relationship between DRD4 genotype and ADHD symptom change [24, 30, 32, 33].
The Preschool ADHD Treatment Study (PATS) was notable in that it focused not only on symptom reduction, but also examined the potential role of genotype in predicting side effects . PATS participants homozygous for the 4-repeat allele were three times more likely to develop abnormal picking behaviors with methylphenidate treatment, while those with one or two copies of the 7-repeat allele were more four times more likely to exhibit social withdrawal with increasing dose. A second study found no associate between genotype and side effects, particularly appetite loss and sleep difficulties .
One of the few studies to examine pharmacogenetic predictors of response to ADHD treatments other than methylphenidate found that children with at least one copy of the DRD4 4-repeat allele showed a trend towards improved response on atomoxetine . Response to methylphenidate, in contrast, was unrelated to DRD4 genotype. Furthermore, improvement on the hyperactivity subscale of the ADHD Rating Scale was maximized with the absence of any 7-repeat variant.
A relatively unstudied gene with potential effects on ADHD medication response is Synaptosomal -Associated Protein (SNAP25), a neuron-specific protein implicated in exocytotic catecholamine release. Several studies of SNAP25 have revealed small, but significant, increased ADHD risk associated with two single nucleotide polymorphisms (SNPs) (T1069C and T1065G) separated by four base pairs at the 3′ end of the gene . In the coloboma mouse, chromosome deletion of SNAP25 gives rise to hyperactive symptoms that are relieved by amphetamine, but not methylphenidate . This is consistent with presumed differences in the mechanisms of action for these compounds, in which amphetamine, but not methylphenidate compensates for reduced exocytotic catecholamine release by reversing the catecholamine diffusion gradient across the transporter.
In methylphenidate treated preschool children with ADHD, individuals homozygous for the T allele at 1065 demonstrated moderately increased improvements, while those who were homozygous for T at 1069 exhibited poorer medication response . More interestingly, children who were homozygous for the less common G allele at 1065 were two to three times more likely to develop sleep difficulties and irritability than those with at least one copy of the T allele. Additionally, those who were homozygous for the less common C allele at 1069 were two to four times more likely to develop tics and other abnormal movements.
In addition to their effects on the dopamine system, stimulants also block reuptake at norephinephrine transporters and indirectly act to increase norepinephrine concentrations at adrenergic alpha receptors. Genes for the norepinephrine transporter (NET) and adrenergic alpha 2A receptor (ADRA2A) are therefore also likely candidates to assess genetic contributions to variability in ADHD treatment response. Although not confirmed in meta-analysis, polymorphisms at two SNPs in NET have been associated with ADHD . An association between ADRA2A and ADHD has also been previously demonstrated . Norepinephrine transporter blockade is also the presumed mechanism of activity for the non-stimulant ADHD medication, atomoxetine .
One study evaluated the relationship between the G1278A polymorphism at NET and methylphenidate in 45 Han Chinese youth with ADHD and found that individuals homozygous for the less common A/A genotype had decreased symptom reductions compared with the G/A or AA groups . The authors noted, since the G1278A allele has no known functional activity, that the allele might be in linkage disequilibrium with another allele responsible for outcome differences.
Potential effects for polymorphisms at the noradrenergic receptor ADRA2A were assessed in 104 children and adolescents after one and three months of methylphenidate treatment . Following post hoc secondary analyses, the investigators reported an interaction between the C1291G polymorphism and improvement over time that explained 30% of the variance in the inattentive score of the Swanson, Nolan, and Pelham (SNAP-IV) Rating Scale. Subjects with at least one copy of the less common G allele showed improved responses in the inattentive domain (F1,104=8.5; p<.004). There was no effect on overall SNAP-IV ratings or Hyperactive-Impulsive scores, nor any direct effect of genotype on response.
ADHD pharmacogenetic studies have principally examined the potential effects of genetic variability on drug targets, i.e. transporters and receptors [9, 10]. Little attention has been devoted to the potential effects of genetic variability on drug metabolism and pharmacokinetics, although these lines of inquiry frequently provide the basis for pharmacogenetic investigations . This might be due to the fact that while methylphenidate is specific in terms of its site of action, its metabolic pathways are poorly understood. It is believed that d,l-methylphenidate undergoes esterification in the blood stream via enzymatic activity of carboxylesterase to d/l ritalinic acid and l-ethylphenidate. One recent report demonstrated inhibition of this metabolic pathway with alcohol ingestion, and also identified one subject as a methylphenidate poor-metabolizer associated with a polymorphism at the carboxylesterase-1 gene . Further study and replication are required to see if this finding underlies the pharmacokinetic variability of methylphenidate generally seen in clinical practice.
Unlike the metabolic products of methylphenidate which are renally excreted, amphetamine undergoes metabolism via hepatic cytochrome P450 (CYP) isozymes. In mammals, amphetamine is metabolized along two major pathways which are differentially employed by various species . In the first pathway, hydroxylation of amphetamine via CYP2D6 yields p-hydroxy-amphetamine. Many psychotropic medications are metabolized by CYP 2D6, although it is believed to play a minor role for amphetamine . However, up to 20% of Caucasians and varying percentages of other racial groups are poor metabolizers due to polymorphisms at CYP 2D6, which can have implications for individual patients . In the second pathway, amphetamine undergoes deamination via CYP 3A4 to l-phenylpropane2-one which is subsequently excreted as inactive benzoic acid. The CYP 3A4 amphetamine pathway is dominant in humans.
In a study of an extended release preparation of mixed amphetamine salts, mean drug plasma concentrations following acute dosing were 25% higher in African American children . Previous studies have noted ethnic differences in CYP 3A4-mediated drug metabolism, with Caucasian subjects demonstrating the highest levels of activity . One allelic variant is heterozygous in 64% of African Americans and has been shown to have decreased metabolic activity . Although racial differences in amphetamine pharmacokinetics have not been otherwise examined, and while a definitive association between amphetamine metabolism and polymorphisms at CYP 3A4 has not been demonstrated, these finding represent a unique lead for further examination of the potential role of genetic variability in explaining the pharmacokinetic variability of amphetamines used to treat ADHD.
Atomoxetine, a currently approved nonstimulant medication for ADHD, is metabolized by the CYP2D6 isozyme system. Drug development trials for atomoxetine were notable in that consideration of subjects’ CYP2D6 status influenced dosing titration algorithms and subsequently derived approved dosing limits. A recent meta-analysis pooled outcome data from several atomoxetine clinical trials . Poor metabolizers displayed greater symptom improvement than extensive metabolizers, most likely due to higher plasma drug concentration levels, and were more likely to remain in therapy. Higher rates of appetite decrease and insomnia were reported more frequently in poor metabolizers, who also exhibited greater increases in medication-related pulse and blood pressure changes.
Candidate gene studies presume some knowledge of the biological system under investigation, and require specific hypotheses regarding putative effects of particular polymorphisms. In contrast, genome-wide investigations require no a priori hypotheses related to specific genes, but scan the entire genome in an attempt to identify areas harboring genes that contribute some effect on outcome. Several genome-wide scans have identified regions linked with ADHD risk [7, 54], including fine mapping by one group to DAT1(55). Genome-wide approaches also have potential utility in pharmacogenetic investigations. One study employed quantitative trait analysis in a genome-wide scan to assess for linkage with methylphenidate response . A linkage peak of moderate significance was identified on chromosome 7, with additional peaks on chromosomes 3, 5, and 9. Further study, including genome-wide association with high density SNP chips, will be necessary to identify the specific genes that influence medication response.
Preliminary evidence suggests that genetic variability plays some role in predicting treatment response, but that results differ depending on whether parents or teachers are informants. Current studies are also constrained by the type of outcome measures used. Many studies rely on simple dichotomous outcomes, such as responder versus non-responder, which have limited power to detect effects compared with analyses of quantitative measures. Correlations between multiple outcome measures in the same subjects are also known to be fairly weak, raising the question as to which outcome measure best defines positive response . Future pharmacogenetic studies would benefit from consensus on the optimal measures to assess outcome. It remains uncertain, particularly in regards to symptom reduction, the relative contributions of direct genetic effects and effects resulting from medication dose and formulation, as well as individual patient variables including diet and gut motility.
Another contributor to differences in study findings is that pharmacogenetic effects may be different in different ethnic groups. This implies the variants being studies may only be in linkage disequilibrium with other variants.
The majority of ADHD pharmacogenetic studies published to date have examined response to methylphenidate. While this is an obvious choice given the known pathophysiology of DAT1, its association with ADHD, and its serving as a specific target for stimulant action, results from preliminary reports have been inconsistent and contradictory. Some of these discrepant findings might be due to methodological issues, as more consistent findings appear to be emerging from placebo controlled, prospective studies. In addition to small sample sizes, other limitations of most existing trials include open-label or retrospective assessment and medication doses that are not specified or considerably lower than used in community practice for optimal benefit . Since the effects of methylphenidate on ADHD symptoms often follow a linear dose response curve , these lower doses might bias against finding significant treatment effects when a less robust dose is utilized.
A critical methodological issue that remains unaddressed is the proper approach to defining genotypes for analysis. In order to minimize the potential for spurious findings and increased Type I errors, investigators must limit their analyses to minimal genotype combinations. For some genes, the risk polymorphisms for ADHD are the less common variants (eg. the 7-repeat allele of DRD4), while for other genes, such as DAT1, it is the more common variant that is associated with the disorder. For DAT1, the 10/10 and 10/9 genotypes are most common. Notably, earlier studies of DAT1 combined these two common genotypes, which assumed a dominant effect of either the 9 or 10 allele, but failed to test for a recessive effect of the 9/9 genotype. Alternative grouping of genotypes based on the presence of one or more 9 allele has led to different results. Future candidate gene studies would benefit from consensus on optimal strategies to define genotype groupings. Most importantly, without previous evidence of dominance of one allele, genotypes should not be lumped together. Statistical power is a combination of sample size and effect size. Dominance (or lack thereof), may differ between association with etiology of disorder and response to treatment.
At a recent meeting of the ADHD Molecular Genetics Network in Brussels, Belgium, the Pharmacogenetics Working Group proposed several principals to promote research in future ADHD pharmacogenetic studies. These include:
Stimulant medications are acknowledged as first-line treatments for ADHD , although the evidence supporting sustained benefit from pharmacotherapy remains less clear . One 5-year prospective study demonstrated that over half of the participants failed to continue medication into the second treatment year, and many of those who did continue therapy reported clinically significant side effects . The authors concluded that side effects and perceived tolerability were major factors in the decision by patients to discontinue treatment. Open-label follow-up studies of clinical trials subjects taking either methylphenidate or amphetamine showed that fewer than 60% of previously stabilized patients remained on medication after 12 months of treatment, although those who remained in therapy showed sustained improvements from baseline [61, 62]. It is therefore clear, that in spite of the overall effectiveness of stimulant medications in the short-term relief of ADHD symptoms, there are considerable impediments to remaining on long-term treatment.
Although most published ADHD pharmacogenetic studies have emphasized genetic predictors of symptom improvement, a more practical clinical application might lie in prediction of side effect risk and medication tolerability. Side effects are major impediments to long term treatment adherence. In the Preschool ADHD Treatment Study, development of irritability and increased emotionality were two major reasons subjects discontinued medication therapy . Interestingly, pharmacogenetic analyses from that trial revealed genetic predictors of irritability, social withdrawal, and abnormal movements . Conceivably, an awareness of increased side effect risk could guide clinicians towards treatment choices that are more likely to be tolerated over time with less patient exposure to medications that are unlikely to provide benefit.
The most promising clinical application of ADHD pharmacogenetics and pharmacogenomics might lie in the development of novel ADHD treatments. The identification of genes that increase susceptibility for ADHD or predict ADHD treatment response might promote the discovery of new drug targets for future pharmaceutical development. It should also be possible, as the cost of genotyping individual patients decreases, to enroll subsets of patients into clinical trials who are more likely to exhibit positive responses and less likely to suffer untoward effects. Ultimately, it is hoped that this type of research will allow clinicians to tailor individual treatment choices based on genotype.
Not surprisingly, pharmacogenetics and pharmacogenomics research efforts in children, adolescents, and adults are expanding world-wide. To date, several promising findings related to diminished response and specific side effects have been reported. Future efforts will also examine a wider range of stimulant and non-stimulant medications on a range of outcome measures and time periods. It is hoped that eventually ADHD treatment outcomes will be improved by a more personalized approach to determining treatments. Potentially, this could result in better understanding of treatment failures and the development of more targeted ADHD treatments. Although research on positive and negative predictive value is necessary prior to attempts to apply preliminary findings to contemporary practice, such studies are currently ongoing. If successful, personalized ADHD therapy will move quickly from the “promise” to the “practical” phase.
This work was supported by Grant No. MH70564 (to Dr. Stein) from the National Institute of Mental Health. The authors would like to thank Edwin H. Cook Jr. M.D. for his thoughtful review and helpful comments.
Dr. Stein has received research support from McNeil, Novartis, Eli Lilly Corporation, Shire Pharamceuticals, Cortex Pharmaceuticals, and Phizer Phamaceuticals and serves as a consultant for Novartis Pharamceuticals. Dr. McGough has received research support from and serves as a consultant to Eli Lilly Corporation, Janssen Pharmaceuticals, Novartis Pharmaceuticals, and Shire Pharmaceuticals.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.