International Consensus Statement for the Screening, Diagnosis, and Treatment of Adolescents with Concurrent Attention-Deficit/Hyperactivity Disorder and Substance Use Disorder

Attention-deficit/hyperactivity disorder (ADHD) and substance use disorder (SUD) often co-occur, in both adults and adolescents seeking treatment [1]. The co-occurrence of ADHD and SUD complicates screening, diagnosis, and treatment and is associated with poor treatment outcomes (e.g., [2]). Childhood ADHD is a prominent risk factor for substance misuse and the development of SUD in adolescence and early adulthood, even when the most common comorbidities during childhood – conduct disorder and oppositional defiant disorder – are accounted for (e.g., [3, 4]). Some studies suggest that stimulant treatment of children with ADHD has a protective effect on the development of SUD in adolescence and early adulthood (e.g., [5, 6]), but other studies did not find such a protective effect (e.g., [7, 8]).

Existing guidelines on ADHD pay little attention to adolescents with comorbid SUD as a distinctive subgroup. Most guidelines only provide some general advice to screen adolescents with ADHD for substance misuse and SUD, recommend to use medications with little or no misuse potential, and/or mention that clinicians should be alert for signs of misuse or diversion of ADHD medication in this group (Guidelines of the Netherlands [9], Australia [10], Scotland [11], Germany [12], Canada [13], the USA [14], and the UK [15]). Moreover, most studies on treatment efficacy of ADHD in youth were conducted in mixed samples of children and adolescents (5–18 years) without separate reporting on the adolescent subgroup (12–18 years). This is unfortunate, because compared with children and adults with ADHD, adolescents with ADHD experience different challenges and treatment demands [16, 17].

Early detection of comorbid ADHD among adolescents receiving SUD treatment and detection of SUD and age-inappropriate use of substances among adolescents receiving ADHD treatment is of vital importance for successful treatment of these comorbid disorders. Although reliable and valid screening instruments and structured diagnostic interviews are available for both disorders [18‒21], systematic screening for ADHD in SUD treatment settings and, conversely, for SUD in psychiatric settings is not routine in clinical practice. Consequently, comorbid ADHD and SUD often remains undetected and hence untreated.

Evidence on the psychosocial and pharmacological treatment of comorbid ADHD and SUD in adolescents is limited. There are currently no meta-analyses or randomized controlled trials (RCTs) that have investigated the efficacy of psychosocial treatments or complementary interventions in adolescents with concurrent ADHD and SUD. However, there are 11 RCTs that examined psychosocial interventions in adolescents with ADHD but without SUD comorbidity, either in a school setting [22‒26] or in a clinical setting [27‒32]. A few of these studies provide some indication that adolescents with ADHD may benefit from psychosocial treatment; however, methodological quality of the studies was generally low and treatment types, treatment components, and treatment settings were heterogeneous. Hence, there is no robust evidence to recommend specific psychosocial treatments for ADHD in adolescents with or without comorbid SUD. Recently, the first trial comparing cognitive behavioral therapy (CBT) for SUD with CBT targeted at both ADHD and SUD in adults with concurrent SUD and ADHD showed a favorable add-on effect of this integrated CBT over CBT for SUD only on ADHD symptoms but not on substance use [33]. Taken together, these findings suggest that CBT may be an effective treatment not only for SUD but also for comorbid ADHD in SUD patients.

Five RCTs have been conducted to study the effect of complementary treatments in adolescents with ADHD but without SUD: cognitive training [26, 34], neurofeedback [35], dietary supplementation with omega-3/6 fatty acids [36], and physical exercise [37]. None yielded robust evidence that these complementary interventions reduce ADHD symptoms in these patients. It is, therefore, unlikely that these interventions will be effective in ADHD adolescents with SUD.

To examine the efficacy of pharmacotherapy in patients with ADHD and concurrent SUD, Cunill et al. [38] conducted a meta-analysis of 13 RCTs and reported mixed results: pharmacotherapy had a (small) beneficial effect on ADHD symptoms, but no effect on substance use or abstinence. Unfortunately, in this meta-analysis, no distinction was made between adult and adolescent patients, and no attention was paid to the heterogeneity within and between samples, to the methodological difference between studies or to variations in dosing. In their systematic review of ADHD pharmacotherapy trials in SUD patients, Carpentier and Levin [39] provide some possible explanations for the mostly negative study findings, including the high prevalence of polydrug use in most study samples, the presence of psychiatric comorbidity in addition to ADHD and SUD, the possibility of suboptimal dosing of stimulant medication, the generally low treatment retention in the trials, and the influence of concurrent treatment, mostly CBT.

To date, only 4 (placebo-)controlled studies have been conducted on the efficacy of pharmacotherapy in adolescents with concurrent ADHD and SUD with a total of less than 500 patients: 2 trials with long-acting methylphenidate [40, 41], 1 trial with pemoline [42], and 1 trial with atomoxetine [43] as the active compound. None of these studies showed a robust treatment effect on either ADHD or SUD. One study found an effect on ADHD, but included only 16 patients [41]; one study reported an effect on ADHD, but with contradictory findings between the primary and secondary ADHD outcome measures [42]; one study reported an effect on ADHD and SUD, but with contradictory findings between the primary and secondary ADHD outcomes and between the substance use self-report and urinalysis data [40]; and one study found no effect on ADHD and SUD on any of the outcome measures [43]. In contrast, virtually, all trials of ADHD pharmacotherapy in adolescents with ADHD but without SUD comorbidity showed a moderate-to-large effect of the medication on ADHD symptoms, including studies with long-acting methylphenidate [44‒47], lisdexamfetamine [48], extended-release mixed amphetamine salts [49], pemoline [50], atomoxetine [51], and long-acting guanfacine [52‒54] as the active compound.

Pharmacological treatments in adolescents with concurrent ADHD and SUD were well tolerated in all 4 studies [40‒43]. Adverse events were generally more prevalent in adolescents in the active medication groups, but these were mostly mild and transient. Reported study-related serious adverse events (SAEs) in the active medication groups were absent or rare (≤1 SAE) in all studies, with no excess of SAEs in any of the active medication groups, compared with placebo. Negative interactions between study medication and the adolescent’s substance use were only investigated in the trial of Riggs et al. [40] and were reported only by 2.8% of the patients receiving methylphenidate. There was no indication that stimulant medication resulted in deterioration of SUD.

While generally well tolerated, stimulant medications and atomoxetine have been associated with a range of short- and long-term cardiovascular adverse effects. Childress [55] reviewed the extended-release stimulants [56] for treating ADHD and found an overall small increase in blood pressure and heart rate for both amphetamines and methylphenidate. In a recent, comprehensive meta-analysis of RCTs on this subject, Liang et al. [56] found that both methylphenidate and atomoxetine in children and adolescents were associated with increased heart rate and systolic blood pressure but were not associated with the number of adverse cardiac events, compared with placebo. Earlier large-scale registry studies have found no evidence that ADHD medications, including stimulants, were associated with increased risk of serious cardiovascular events in children and young adults [57] or in young and middle-aged adults [58].

In conclusion, research on the diagnosis and treatment of adolescents with ADHD and SUD is scarce, and the evidence from the few available studies in this subgroup is not strong enough to justify robust treatment recommendations. Therefore, as part of the initiative of the International Collaboration on ADHD and Substance Abuse (ICASA), we performed a modified Delphi study based on scientific and clinical evidence with a large multidisciplinary international group of experts who were asked to rate a set of statements regarding (a) the effect of ADHD treatment of children on the development of SUD in adolescence and early adulthood, and (b) the possibilities for screening, diagnosis, and treatment of adolescent patients with co-occurring ADHD and SUD. ICASA previously initiated the development of a comparable consensus document on the screening, diagnosis, and treatment of adult patients with concurrent ADHD and SUD [59].

We conducted a modified Delphi process [60] with the following steps:

1. A systematic literature review was performed by 3 members of the expert group (V.H., H.Ö., and R.S.) [61] on (a) the effect of ADHD treatment of children on the development of SUD in adolescence and (b) the possibilities for screening, diagnosis, and treatment of adolescent patients with co-occurring ADHD and SUD [61].

2. Based on this systematic review, initial statements for a consensus document were selected by the authors of the review (V.H., H.Ö., and R.S.) and 4 members of ICASA: W.v.d.B., G.v.d.G., A.S., M.N., and M.H.

3. A multidisciplinary international group of 62 experts was invited for the consensus process: 44 ICASA members (71%) and 18 child and adolescent ADHD-SUD experts nominated by these ICASA members (29%). Of these 62 invited experts, 55 participated in the Delphi process: 37 (67%) ICASA-members and 18 (33%) invited child and adolescent experts. The 55 participating experts (including the 8 experts involved in the selection of the statements) were from 17 countries and 5 continents (Europe, North America, Australia, Asia, and Africa) and included 15 (child and adolescent) psychiatrists, 14 (child and adolescent) psychologists, 16 addiction physicians, 6 scientists, 2 pharmacists, 1 mental health nurse, and 1 epidemiologist.

4. All statements (resulting from step 2) and the systematic review (resulting from step 1) were sent to the group of 55 experts. They were asked to (a) rate all statements on a scale of 1–5 (1 = most disagree, 2 = disagree, 3 = neutral, 4 = agree, and 5 = strongly agree) based on their knowledge, clinical experience, and – if desired – the systematic review, (b) provide comments on the content and/or the phrasing of the statements if they felt that this was needed or useful, and (c) propose any additional statements they felt would be useful. Consensus between experts for each statement was defined a priori as at least 95% of all the ratings being greater than or equal to 3 (strongly agree, agree, and neutral), and thus, no more than 2 experts with a rating on a specific statement equal to or smaller than 2 (disagree and most disagree). Although arbitrary, we considered a “neutral” rating in our dichotomous consensus measure to reflect (some) agreement rather than disagreement, because the rater at least did not disagree with the statement. The same procedure was used in a recent consensus article on baclofen as a viable treatment for alcohol dependence [62].

5. After all ratings and comments were received and consensus was calculated, all statements without consensus, statements with consensus but useful comments that resulted in a different phrasing, and additional statements proposed by the experts were sent out for a second round of rating by all experts.

6. Based on this second round of ratings, consensus was calculated for this subgroup of statements using the same a priori defined rules. Statements that still did not reach expert consensus were sent out to all experts for a final rating and a final request for comments. In addition, all experts received the full text of the international consensus statement (including abstract, introduction, methods, and discussion/conclusion) with a request to comment on the text.

7. Based on the final round of ratings and comments, the text of the paper was finalized by the senior authors of the current article (H.Ö., R.S., W.v.d.B., and V.H.) and submitted for publication.

Of the 55 experts in the consensus group, 52 (95%) participated in the first round, 55 (100%) participated in the second round, and again 55 (100%) participated in the third round of ratings and comments on statements and on the text of the consensus document.

Based on our a priori definition of consensus, 10 of the 36 selected initial statements did not reach consensus in the first round (Table 1). These were adapted and were sent out again for the second round of rating and commenting. In addition, we identified 3 statements with consensus, but with comments that stimulated us to slightly rephrase these statements. These adapted statements were also sent out for the second round of rating and commenting. Finally, some experts proposed additional statements, and of these, we added 1 statement (Table 1, Statement 30) that was also sent out for the second round of rating and commenting. Thus, the second round of rating and commenting involved 14 of the 37 (36 + 1) statements.

In the second round of ratings and comments, 11 of the 14 (adapted) statements reached consensus. The remaining 3 statements were further adapted and sent out for a third round of rating and commenting, together with the draft text of the consensus document and a request to provide comments.

In the third round of rating and commenting, 2 of the 3 (adapted) statements reached consensus. Hence, consensus was reached on 36 of the 37 statements. The statement without consensus (Table 1, Statement 22) concerned the requirement of abstinence before starting stimulant pharmacotherapy in adolescents with ADHD and concurrent SUD. In contrast to the majority of experts, some experts required full abstinence before starting any pharmacological treatment, some were against the use of any stimulant in the treatment of these patients (independent of abstinence), and some experts were against the alternative use of bupropion.

In summary, consensus was reached on 23 of the 36 statements in round 1 (64%) and – after adaptation of the remaining 13 statements plus 1 new statement – on 11 of the 14 statements in round 2 (overall consensus 23 + 11 of 37 statements = 34/37 = 92%). After adaptation of the remaining 3 statements, consensus in round 3 was reached on 2 of the 3 statements, resulting in an overall consensus about 36 of the 37 statements (97%).

The effect of treatment of childhood ADHD on the development of adolescent SUD and the screening, diagnosis, and treatment of concurrent ADHD and SUD in adolescents has received little attention in research and in treatment guidelines. The evidence-base on these issues is limited and not robust enough for strong clinical recommendations. At the same time, adolescent patients with ADHD and concurrent SUD are in serious need of treatment to improve clinical and psychosocial outcomes and to prevent chronicity. In an attempt to fill this need, we performed a modified Delphi study with the aim of obtaining a set of consensus statements pertaining to these issues.

The study shows that a multidisciplinary international group of experts was able to reach a high level of consensus on 36 of 37 statements about the treatment of childhood ADHD to prevent the development of SUD later in life and about the screening, diagnosis, and treatment of adolescents with co-occurring ADHD and SUD. Consensus was easily reached in the first round for 64% of the statements and – after adaptations of the original statements – this percentage mounted to 92% in the second and 97% in the third round. The only statement with no consensus concerned the requirement of abstinence or reduced/stabilized substance use before prescribing psychostimulants to adolescents with concurrent ADHD and SUD. The reasons for this lack of consensus were multiple. Some experts preferred to minimize the potential risks by demanding strict abstinence as a precondition for pharmacotherapy, while others were worried that a strict abstinence policy would exclude a large group of patients from receiving effective treatment. Some other experts were against the use of psychostimulants in these patients altogether (i.e., independent of whether patients were abstinent at the start of such a treatment) mainly because of the risk of abuse and/or diversion. Finally, some experts were strongly opposed to the use of bupropion as an alternative to psychostimulants in non-abstinent patients, mainly due to a lack of evidence of its effect and the risk of bupropion-induced seizures. It is important to note that the existing national treatment guidelines on treating ADHD in children and adolescents show similar disagreements on these issues.

The current study has both strengths and limitations. The main strengths are (1) the presence of a systematic review as a shared knowledge base for all participating experts, (2) a sizable group of experts from different countries with different treatment cultures, and (3) a very high initial and final response rate. The main limitations are (1) possible bias in the selection of the experts, given that most participating scientists and clinicians had a special interest in the topic of ADHD and SUD, child and adolescent experts were recruited through co-optation by the ICASA Network, and pediatricians and patient representatives were absent in the expert group, and (2) the lack of an official approval of our consensus statement by scientific, professional, and patient organizations. In order to (partly) remedy the latter limitation, we will circulate the results of this consensus article directly to the international scientific, professional, and patient organizations in the mental health and addiction field. In addition, we will present the results at national and international conferences.

The current set of consensus statements can be used by clinicians and patients together in a shared decision-making process to select the most appropriate treatments and to reach optimal outcomes for adolescents with complex problems based on the integration of scientific knowledge, clinical experience, and patient preference. Finally, we would like to emphasize that our consensus statement on concurrent ADHD and SUD is not a replacement for an evidence-based guideline and that more high-quality studies, including both RCTs and long-term naturalistic follow-up studies are needed [63].

We gratefully acknowledge the ICASA network for initiating the research of this study.

The authors have no ethical conflicts to disclose.

The authors have no conflicts of interest to declare.

The research was supported by a financial contribution of ICASA to the Parnassia Addiction Research Centre (PARC) (Dr. Özgen, Dr. Spijkerman, Dr. Hendriks).

V.H., H.Ö., and R.S. conducted the systematic review that informed the development of this consensus statement. H.Ö., R.S., M.N., M.H., A.S., G.v.d.G., W.v.d.B., and V.H. were responsible for the selection of the initial statements for this consensus document and drafted the initial document and its revisions. The senior authors H.Ö., R.S., W.v.d.B., and V.H. were responsible for collecting and coordinating the feedback during each consensus round and for finalizing the text of the paper. All authors contributed to the manuscript and approved its final version.

Heval Özgen, Renske Spijkerman, Wim van den Brink and Vincent Hendriks contributed equally to this paper.