Background
Attention-deficit/hyperactivity disorder (ADHD) is a neurodevelopmental disorder with onset in childhood; the disorder is characterized by a heterogeneous etiology and developmental course. Although severity of symptoms, especially of hyperactivity-impulsivity, often diminishes over time, prospective longitudinal studies have shown that the impairing symptoms of the disorder persist into adulthood in approximately two thirds of children with ADHD, with associated impairment across multiple domains [
1]. Recent studies have challenged the established notion of ADHD as an exclusive childhood-onset disorder (onset prior to 12 years of age) and reported onset of ADHD in adolescence [
2] and adulthood [
3‐
5].
Various risk factors have been associated with a more persistent course, such as higher ADHD symptom severity, presence of comorbidity (in particular conduct and mood disorders), family history of ADHD, and psychosocial and environmental adversities [
6]. Longitudinal studies of children with ADHD have consistently identified early substance use in adolescence as an important negative outcome of the disorder [
7‐
9]. Although many studies have examined the link between ADHD and the onset of substance use disorders (SUD; alcohol and/or drug use disorder) and nicotine dependence (ND) [
10,
11], few have investigated the association between SUD/ND and the course of ADHD over time. While comorbid conduct disorder and oppositional defiant disorder among children with ADHD increase the risk for SUD [
12], several studies have shown that ADHD is an independent risk factor for the development of SUD in adolescence and adulthood [
10,
11]. Moreover, ND may increase the risk of other SUD [
13]. For example, in a longitudinal study, Biederman et al. [
14] found that ADHD youth who smoked cigarettes were more likely to subsequently use other substances and to develop SUD, compared to other ADHD youth. This risk is further increased when ADHD symptoms persist into adulthood, with higher prevalence rates of SUD found among individuals with persistent than among those with remitted ADHD [
10,
15].
It is unclear whether there is also a link between late-onset ADHD, referring to the onset of ADHD after 12 years of age, and SUD/ND. Several birth cohort studies have reported a 2.5–10.7% prevalence of a form of ADHD that first emerged in mid/late adolescence or (young) adulthood, so-called late-onset ADHD [
3‐
5]. ADHD with onset after age 12 had patterns of psychiatric comorbidity, functional impairment, familial transmission, and intelligence similar to childhood-onset ADHD [
2]. Careful assessment of ADHD symptoms that emerged after childhood is crucial, since false positive cases of adult-onset ADHD are common [
16]. Moreover, the reliability of an age at onset assessment could be increased by using multiple informants or by selecting (young) adult participants from prospective follow-up studies.
The present study examined whether the course of ADHD is associated with increased risks for developing SUD (alcohol and/or drug use disorder) and ND from childhood through young adulthood. We report findings from a 10-year prospective, longitudinal study of a subsample of the Dutch International Multicenter ADHD Genetics (IMAGE) study cohort, including individuals with persistent ADHD (n = 62), remittent ADHD (n = 12), late-onset ADHD (n = 18), unaffected siblings (n = 50), and healthy controls (n = 47). Given the extant literature that ADHD is a risk factor for SUD, we hypothesized that the persistent ADHD group would show higher rates of SUD and ND compared to the remittent ADHD group, while we expected higher prevalence rates of SUD and ND in the late-onset ADHD group compared to the stable unaffected groups (unaffected siblings and healthy controls). By including a high-risk group of full biological siblings of ADHD probands, this study allowed us to investigate the ADHD course and development of SUD and ND over time in this group, characterized by increased genetic and/or environmental risk for ADHD.
Discussion
To our knowledge, this study provides the first data on the association between SUD/ND and the course of ADHD among probands with ADHD and their biological siblings in a 10-year prospective longitudinal study. Results showed that ADHD persisters were at significantly higher risk for the subsequent development of SUD relative to healthy controls, in contrast to a similar level of SUD in ADHD remitters vs healthy controls. ADHD persisters had also higher (albeit non-significant) prevalence rates of ND than ADHD remitters and healthy controls. A similar pattern of results was found for those with late-onset ADHD who had a higher prevalence of SUD compared to healthy controls, while no higher prevalence rates of SUD were found in stable unaffected siblings compared to healthy controls. Further, those with late-onset ADHD were at significantly higher risk for the subsequent development of ND compared to healthy controls, in contrast to a similar level of SUD in stable unaffected siblings versus healthy controls. Although not all results reached statistical significance, the pattern of findings suggests SUD and ND are associated with a negative ADHD outcome (i.e., persistent ADHD and/or late-onset ADHD) and further emphasize the need for clinicians to make a comprehensive assessment of substance use when diagnosing ADHD.
Previous research has suggested that the persistence of ADHD is a risk factor for the development of SUD [
15]. Our study replicated that finding and extended the analysis by showing the same pattern of results in biological siblings with late-onset ADHD. The findings suggest that late-onset ADHD is also associated with ND. ND is often described as a gateway drug to illicit drug use [
28], and studies consistently identify the increased risk of ND in individuals with ADHD [
13]. In our samples, both persistent and remitted ADHD had higher rates of ND than the controls, although the effects were not significant. This is likely due to small sample sizes and hence limited statistical power. Given that SUD and ND might be associated with adverse outcome of ADHD, identification of early risk factors and preventive interventions in children at risk for a persistent course of ADHD as well as their biological siblings might be crucial.
Disentangling the nature of the association between ADHD and SUD/ND remains challenging. Although the literature is consistent in showing that the onset of ADHD precedes the development of SUD, high levels of substance use in adolescence may adversely affect still-maturing prefrontal brain regions leading to behavior regulation deficits associated with ADHD [
29]. In this perspective, substance use could negatively influence the course of ADHD in addition to the widely accepted ADHD-to-substance use pathway. Alternatively, because SUD often manifests in adolescence and young adulthood, the SUDs itself might elicit symptoms of ADHD and be mistakenly identified as “late-onset of ADHD” [
30]. A third possible explanation for the association between ADHD course and SUD/ND is that the relation between ADHD and SUD/ND is largely explained by a common third factor that increases both the risk for a detrimental ADHD course as well as increase the risk for SUD/ND, for instance, environmental factors and/or shared genetic lability to both disorders [
31,
32].
Our findings offer some support for the possibility of late-onset ADHD, a proportion of biological siblings unaffected at baseline met DSM-5 criteria of ADHD in young adulthood, with age of onset after 12 but prior to 18 years of age. Since all participants were comprehensively assessed on ADHD and comorbid disorders by multi-informant questionnaires and interviews in childhood, as well as in (young) adulthood, it is unlikely that these cases represented individuals with undetected childhood symptoms (i.e., late-identified rather than late-onset) [
33‐
35] or false positive cases of late-onset ADHD. Although emotional lability scores at follow-up were higher in the late-onset ADHD group compared to the stable unaffected siblings group, these levels were not in the clinical range. An explanation for late-onset ADHD in biological siblings might be that unaffected siblings carry vulnerable heritable traits for ADHD and often already have increased levels of ADHD symptoms in childhood. However, their ADHD symptoms might not be severe enough to merit a clinical diagnosis. Clinical levels of ADHD might manifest later in life, when the demands of life increase or until they can no longer rely on or compensate by protective factors, such as high cognitive ability [
4]. Future studies should investigate mechanisms associated with later onset ADHD in adolescence or adulthood, e.g., neuropsychological profile that possibly could differentiate individuals developing late-onset ADHD versus stable unaffected individuals.
Previous family studies of ADHD have shown that ADHD and SUD co-aggregate in families [
30,
36]. In accordance with these findings, we found an increased risk of SUD among unaffected siblings of ADHD probands. Importantly, these findings indicate that there is no direct relationship between ADHD and SUD and that shared genetic lability and/or family environment risk factors to both disorders might contribute to the development of both. Previous findings from our study showed no increased risk of SUD in unaffected biological siblings of ADHD probands [
11]; however, our participants in this study were still relatively young (mean 17 years) and might not have completely traversed the developmental pathway to substance use and ND. The present finding supports the notion that SUD increases during the late adolescent and young adult years. Taken together, our findings support the hypothesis that ADHD is a familial risk factor for SUD and that biological siblings represent a group who are at high risk for the subsequent development of both SUD and late-onset ADHD.
Some methodological limitations should be taken into account in the interpretation of the results. Given that several of the subgroup samples were relatively small, detecting significant group differences was more difficult. However, our results were in the expected direction, indicating that significant results are expected with a larger sample. Furthermore, SUD and ND were assessed by self-reports. We used adult cutoff scores for the self-report questionnaires, while not all subjects reached adulthood at the time of assessment. Although this approach would not have biased results to finding spurious case-control differences, it may have influenced our estimates of prevalence.
Conclusions
This study contributes to the understanding of the association between substance use and the course of ADHD over time. In particular, children with persistent ADHD have higher risks of developing alcohol and/or drug dependence over time compared to healthy controls, while this was not found for individuals with remittent ADHD. The same pattern of findings was found for siblings developing late-onset ADHD versus those who remained unaffected. This suggests that SUD and ND seem to be associated with a negative ADHD outcome. Although the mechanisms governing these associations are not yet fully understood, the findings from this study underscore the clinical and public health significance of SUD in the course of ADHD. This study emphasizes the importance of preventive interventions in biological siblings, considering their increased risks of developing both SUD and ADHD.
Acknowledgements
We thank all the families and teachers that took part in this study and all students for their assistance in data collection.