Pharmacological and clinical dilemmas of prescribing in co-morbid adult attention-deficit/hyperactivity disorder and addiction

Abstract

The present article reviews whether available efficacy and safety data support the pharmacological treatment of adult attention-deficit/hyperactivity disorder (ADHD) in patients with concurrent substance use disorders (SUD). Arguments for and against treating adult ADHD with active SUD are discussed. Findings from 19 large open studies and controlled clinical trials show that the use of atomoxetine or extended-release methylphenidate formulations, together with psychological therapy, yield promising though inconclusive results about short term efficacy of these drugs in the treatment of adult ADHD in patients with SUD and no other severe mental disorders. However, the efficacy of these drugs is scant or lacking for treating concurrent SUD. No serious safety issues have been associated with these drugs in patients with co-morbid SUD-ADHD, given their low risk of abuse and favourable side effect and drug–drug interaction profile. The decision to treat adult ADHD in the context of active SUD depends on various factors, some directly related to SUD-ADHD co-morbidity (e.g. degree of diagnostic uncertainty for ADHD) and other factors related to the clinical expertise of the medical staff and availability of adequate resources (e.g. the means to monitor compliance with pharmacological treatment). Our recommendation is that clinical decisions be individualized and based on a careful analysis of the advantages and disadvantages of pharmacological treatment for ADHD on a case-by-case basis in the context of active SUD.

Introduction

The prevalence of adult attention-deficit/hyperactivity disorder (ADHD) in patients with substance use disorder (SUD) is high. According to a recent systematic review [1], the prevalence rate was 23.3% (95% CI 17.7%, 30.1%), although prevalence varies depending on the psychoactive substances involved and the diagnostic instrument. Given the close bidirectional relationship between both disorders, it seems highly improbable that the elevated co-morbidity between adult ADHD and SUD is due to chance alone. Some authors have even suggested that SUD and ADHD could be different aspects of the same overarching condition [2] or that the same endophenotype (i.e. impulsivity) may underlie both disorders [3]. According to the aforementioned systematic review, approximately one-fifth of patients at SUD treatment centres may also require drug therapy for ADHD, the primary treatment for adult ADHD [4].

Pharmacological treatment of adult ADHD in patients with SUD presents numerous clinical problems for clinicians. Although various drugs have proven effective in the treatment of adult ADHD in patients who do not suffer from SUD [5,6], it is unclear whether these drugs are also safe and effective in co-morbid SUD-ADHD. As a result, clinicians are faced with an important clinical dilemma related to the uncertainties surrounding the use of these medications in patients with co-morbid SUD-ADHD. It is not clear whether pharmacological treatment for adult ADHD should be administered in patients with co-morbid SUD. In the present review, we evaluate the available evidence to determine whether there is sufficient evidence to support pharmacological treatment in these cases. We have reviewed the relevant clinical and laboratory studies of adults with SUD to evaluate the efficacy and/or safety of the drugs used to treat ADHD.

A second dilemma confronting clinicians is the advisability of administering pharmacological treatment for adult ADHD in active substance abusers. Given that the symptoms of both disorders tend to overlap each other, correctly diagnosing ADHD in such cases is extremely challenging [7,8]. In this review, we present the arguments for and against treating adult ADHD in patients with active SUD and we review the relevant literature related to these arguments. We have structured our discussion to follow the same steps carried out in our clinical decision-making process. Finally, we give our recommendations based on the information presented here as well as our own clinical experience.

This article, authored by a group of clinician-scientists with more than 10 years of clinical experience in the treatment of severe addictions and/or personality disorders, was written with a practical purpose: to provide a decision making framework for clinical practice at centres that specialize in treating adult SUD. Consequently, the medical literature on childhood or adolescent ADHD is only reviewed when necessary (i.e. when the available information about that particular aspect of adult ADHD is scant). It is important to keep in mind that the present article is not a review of the risk of developing SUD. Rather, the focus is on the risks of exacerbating SUD and the potential for misuse/abuse of the drugs prescribed to treat adult ADHD. In order to identify relevant studies and literature reviews, we used the PubMed interface to search MEDLINE (from inception up to March 2012). A comprehensive search strategy, favouring sensitivity over specificity, was performed using different combinations of SUD-specific (e.g. substance abuse, drug dependence, alcoholism, cannabis, cocaine, heroin, nicotine), ADHD-specific (e.g. attention deficit hyperactivity disorder, attention deficit disorders, attention deficit disorder with hyperactivity) and treatment-specific (e.g. amphetamine, atomoxetine, bupropion, methylphenidate, modafinil, pemoline) free-text and MeSH terms. Details on the full search strategy are available from the corresponding author upon request. Reference lists of relevant high quality literature reviews were also reviewed to identify additional pertinent studies.

In short, in the present review we evaluate two important dilemmas related to drug safety, efficacy and interactions that clinicians face when treating patients with co-morbid SUD-ADHD: (i) Should adult ADHD be treated pharmacologically in the context of an addiction? and (ii) Must we wait until SUD is under control before initiating pharmacological treatment of adult ADHD?

Should adult ADHD be treated pharmacologically in the context of an addiction?

Numerous clinical trials have found that pharmacological treatment of ADHD is safe and effective for adults without co-morbid SUD. However, because patients with SUD were excluded from these trials, the results may not be generalizable to the patient group addressed in this current review. Consequently, our first aim is to determine if pharmacological treatment of adult ADHD is safe and effective in patients with concurrent SUD-ADHD. Secondly, in patients with concurrent disorders, we must consider the possibility that the drugs used to treat one disorder could also affect the second disorder. It is important to determine if there is any impact, and if so, whether that impact is positive, negative or neutral. In other words, what are the possible repercussions of pharmacological treatment for ADHD on SUD?

Efficacy of pharmacological treatment for ADHD on adults with SUD-ADHD: Overview with an emphasis on the repercussions on the addictive disorder

In the mid-1980s, a not inconsiderable number of case reports were published on the treatment of adult ADHD in patients with SUD (primarily cocaine dependence, although a few cases of alcohol dependence were also reported) treated with pemoline [9,10] or bromocriptine [11,12[. Later, various open studies [13–20] reported positive findings (see Table 1) regarding the clinical utility of several different drugs (stimulants and some non-stimulants) in patients with this dual disorder. These results then led to the launch of several controlled clinical trials [21–31] (see Table 2).

Table 1.

Uncontrolled clinical studies (with n ≥ 10 subjects) of stimulants and non-stimulants in adults with attention-deficit hyperactivity disorder and substance use disorder

First author (year) [Reference]	n	Age (years) (SD) and male proportion	Co-morbid substance use disorder	Study design, setting and duration	Medication and mean maintenance dose and/or range		Concurrent treatment	Retention	Main outcomes	Adverse events	Comments
Levin (1998) [13]	12	34 (1.4)	Cocaine dependence	Open; out-patient	MPH SR		Relapse prevention	66.7%	ADHD: Significant reduction in ADHD symptoms.	The most common AEs were dry mouth, increased heart rate, jitteriness and agitation. No patient was discontinued due to AEs	Psychiatric co-morbidity was common. Few medication doses were missed due to lack of attendance (91% compliance with clinical appointments)
		60%		12 weeks	68 mg day−1				SUD: Drug use severity, cocaine craving and cocaine use (both self-reported and confirmed by urinalyses) decreased significantly
					40–80 mg day−1
Castaneda (1999) [14]	19	36.6 (10.4)	Cocaine dependence	Open; out-patient	FLX	20 mg day−1	Medication-free patients (with the exception of two bipolar patients maintained on valproic acid)	88.9%	ADHD: 12 patients (63.2%) were kept on a fully effective single treatment regimen (suppression of almost 80% of the initial ADHD symptoms) for 36 to 52 weeks (31.6% on MPH ER).	AEs were rare and mild. No evidence of any abusive use of any stimulant medication	Stable patients in full or partial remission from cocaine dependence. Duration of remission prior to ADHD treatment ranged from 6 to 26 months (mean = 13.9). Most patients had at least one associated psychiatric diagnosis. Private practice study
		89.5%		1 year	BPR	200 mg day−1			SUD: Two patients (10.5%) had two 1 day cocaine slips each. Two other patients (10.5%) had a full-blown relapse
					PML	37.5–75 mg day−1
					MPH ER	40–120 mg day−1
					d-AMP ER	60 mg day−1
					m-AMP ER	15 mg day−1
					Progressive introduction (in the order listed) and discontinuation of ineffective treatment
Upadhyaya (2001) [15]	10	34.5 (5.7)	Alcohol abuse or dependence (100%); cocaine abuse or dependence (20%)	Open; out-patient	VLFX		No other psychotropic medication. Supportive psychotherapy. Possible additional participation in Alcoholics Anonymous and routine outpatient group and/or individual therapy for alcohol abuse	33.3%	ADHD: Significant improvement in ADHD symptoms.	One patient (10%) dropped out because of AEs (headaches) after week 3	Exclusion criteria included current Axis I diagnosis or lifetime schizophrenia diagnosis. Four patients (40%) had a history of major depression
		40%		12 weeks	Max = 300 mg day−1				SUD: Significant improvement in alcohol craving intensity, and a trend toward improvement in frequency
Somoza (2004) [16]	41	36	Cocaine dependence	Open; out-patient	MPH IR		Cognitive-behavioural therapy	70.7%	ADHD: The sample as a whole, as well as the compliant and non-compliant subgroups, experienced significant improvement in their ADHD symptoms.	MPH IR was well-tolerated by participants. No serious AEs and no persistent unresolved AEs rated as moderate or severe. Higher doses were not associated with greater number of AEs. One participant was discontinued due to AEs (chest tightness, restlessness, and tingling in arm). One patient discharged for abusing MPH IR	Sample primarily composed of crack-cocaine dependent patients. Participants were required to provide at least one urine positive for benzoylecgonine (cocaine metabolite) during the 1 week screening period and to provide a negative urine sample on the day of study enrolment. Measures of treatment compliance included MPH plasma levels and staff ratings of participant's compliance (to be categorized as compliant, participants were required to have scored above the median for both indicators)
		76%		10 weeks	Max = 60 mg day−1				SUD: Subjective efficacy measures (both patient-and clinician-rated) suggested that patients (the whole sample as well as the compliant and non-compliant subgroups) evidenced statistically significant improvement. Urine results indicated that only the compliant subgroup succeeded in maintaining low levels of use
Levin (2009) [17]	20	39.3 (6.8)	Cocaine dependence	Open; out-patient	ATX		Cognitive-behavioural therapy	25%	ADHD: 50% response rate.	Most patients tolerated the medication well. Two participants (10%) were discontinued from the medication component of treatment due to AEs (a) chest pain and b) nervousness, anxiety, insomnia and fatigue, respectively)	Patients with no additional severe psychopathology, but actively using cocaine. Good medication compliance (monitored by self-report and riboflavin fluorescence testing)
		95%		12 weeks	80–100 mg day−1				SUD: No significant decrease in cocaine use (determined by urine toxicology and self-report)
Adler (2010) [18]	18	36.8 (10.0)	Drug and/or alcohol dependence (mainly poly-substance users; 33.3% of whom reporting cocaine as primary drug of use)	Open; residential treatment facility	ATX		NR	27.8%	ADHD: Improvement of investigator-and participant-rated ADHD symptoms.	No serious AEs. Headache and abdominal pain/cramps were the most commonly reported TEAEs. None of the AEs resulted in discontinuation of treatment	Abstinence from all substances for at least 2 weeks prior to study entry (mean length of abstinence at baseline was 14.1 [SD = 18.5] weeks). Exclusion criteria included lifetime diagnosisof bipolar disorder, schizophrenia, or schizoaffective disorder
		83.3%		10 weeks	Max = 120 mg day−1				SUD: Reduction in the intensity, frequency, and length of cravings
Wilens (2010) [19]	32	32.0 (8.5)	Substance use disorders; mainly alcohol and drug abuse or dependence (68.8%)	Open; out-patient	BPR SR		Patients treated with naltrexone, methadone or nicotine replacement therapy were admitted	59.4%	ADHD: 65.6% response rate.	No serious AEs. No AEs related to interactions with substances of abuse. No clinically meaningful changes in cardiovascular signs. Four patients (12.5%) withdrew due to AEs, including elevated blood pressure, exacerbation of panic attacks and one case of significant rash	Exclusion criteria included clinically unstable psychiatric conditions (e.g. unstable SUD) and lifetime bipolar or psychotic disorders
		81.3%		6 weeks	Max = 400 mg day−1				SUD: No clinically significant reductions in either self-reported substance use or investigator-rated scores.
									ADHD x SUD: ADHD responders were more likely to have improved SUD at end point. SUD responders were more likely to have improved ADHD symptoms at endpoint
McRae-Clark (2011) [20]	14	33.9 (13.2)	Past stimulant misuse, abuse, or dependence	Open; out-patient	MPH TS	1.82 (0.36) mg h−1	Exclusion criteria included current treatment with a psychoactive medication	85.7%	ADHD: Significant improvements from baseline were found on ADHD symptoms.	Most frequent AEs were localized skin reactions at the site of patch application. Eight patients (57.1%) reported mild-to-moderate localized irritation, pruritis, or rash. Two participants (14.3%) were removed due to AEs (severe localized skin irritation, and re-emergence of depressive symptoms, respectively)	Current abuse or dependence on substances other than caffeine or nicotine was not permitted
		42.9%		8 weeks		1.1–2.2 mg h−1				SUD: All submitted urine samples (Mean [SD] = 7.6 [1.8]) were negative for stimulants. Only one participant (7.1%) reported use of oral stimulants at week 6
						Daily wear time: 9.05 (1.43) h

Response rate refers to the percent of participants meeting the standard response criterion for the primary outcome measure (e.g. CGI-I score < 3 [i.e. much to very much improved] or a ≥ 30% reduction in symptoms in ADHD rating scales when comparing the last observation to baseline); ADHD, Attention-deficit hyperactivity disorder; AE, Adverse event; ATX, Atomoxetine; BPR, Bupropion; d-AMP, Dexamphetamine; ER, Extended release; FLX, Fluoxetine; IR, Immediate release; m-AMP, Methamphetamine; MPH, Methylphenidate; NR, Not reported; PML, Pemoline; SR, Sustained release; SUD, Substance use disorder; TEAE, Treatment-emergent adverse effect; TS, Transdermal system; VLFX, Venlafaxine.

Table 2.

Controlled clinical studies of stimulants and non-stimulants in adults with attention-deficit hyperactivity disorder and substance use disorder

First author (year) [Reference]	n	Age (years) (SD) and male proportion	Co-morbid substance use disorder	Study design, setting and duration	Medication and mean maintenance dose and/or range	Concurrent treatment	Retention	Main outcomes	Adverse events	Comments
Levin (2002) [21]	11	30.7 (5.2)	Cocaine dependence	Single-blind; out-patient	BPR	Relapse prevention	91.0%	ADHD: Significant reduction of ADHD symptoms no significant differences between BPR and MPH SR.	BPR was well-tolerated. The most frequently reported AEs were agitation, dry mouth, insomnia, constipation, and dizziness	The MPH SR group served as a historical control group (patients were not randomized to either BPR or MPH SR). Most patients had at least one current Axis I diagnosis (mainly alcohol or marijuana abuse or dependence)
		100%		12 weeks	250–400 mg day−1			SUD: Significant reduction of cocaine craving between baseline and the last 2 study weeks. Significant reduction of cocaine use between the 4 weeks before study entry and the last 4 weeks of study. No significant differences between BPR and MPH SR for cocaine craving or use
Schubiner (2002) [22]	48	37.1 (6.6)	Cocaine dependence	Double-blind, placebo-controlled, parallel-group; out-patient	MPH	Cognitive-behavioural therapy	MPH: 45%	ADHD: Significant differences between groups on physician-and self-rated efficacy indexes.	Occurrence of insomnia or trouble sleeping was higher in the MPH group. No participants required discontinuation of MPH because of AEs	High rate of Axis I co-morbidity (mainly affective and anxiety disorders). Compliance (measured by self-report) was high. The sample was stratified by gender. No group differences in the proportion of participants guessing their study medication correctly
		90%		13 weeks	78.8 (19.4)		Placebo: 58%	SUD: No group differences in proportion of days of reported cocaine use, percentage of cocaine-negative urine samples, duration of continuous abstinence, and craving
					Max = 90 mg day−1
Carpentier (2005) [23]	25	31.9 (5.8)	Drug and/or alcohol dependence (primary substances of use: cocaine [32%] and alcohol [28%])	Double-blind, placebo-controlled, cross-over; in-patient	MPH	No concurrent use of any other medication in most cases	76%	ADHD: 36% response rate (Placebo: 20%), NS.	MPH group showed significantly more AEs than patients randomized to placebo	Abstinence from all substances for at least 3 weeks prior to study entry. High psychiatry co-morbidity rate
		88.0%		8 weeks	Max = 45 mg day−1			SUD: NR (participants had to remain abstinent for the total duration of the study)
Levin (2006) [24]	98	39.0 (7.3)	Opioid dependence and cocaine dependence or abuse	Double-blind, placebo-controlled, parallel-group; out-patient	MPH SR	Medication and treatment as usual in a methadone programme; cognitive-behavioural therapy	MPH SR: 65.6%	ADHD: 34% (MPH SR) and 49% (BPR SR) response rates (Placebo: 46%), NS.	The medications were well-tolerated. No group differences were observed with respect to AEs. The most frequently reported AEs were fatigue (Placebo, 9%) and increased sweating (MPH SR, 6% and BPR SR, 9%). No evidence of abuse	Exclusion criteria included current psychiatric disorders (other than ADHD and SUD) and physiological dependence on either sedatives or alcohol. Compliance (measured by self-report and by riboflavin fluorescence testing) was good and did not differ across groups
		57.1%		12 weeks	40–80 mg day−1		BPR SR: 69.7%	SUD: No significant differences across the groups in any substance use outcome
					BPR SR		placebo: 75.8%
					200–400 mg day−1
Levin (2007) [25]	106	37.0 (6.5)	Cocaine dependence	Double-blind, placebo-controlled, parallel-group; out-patient	MPH SR	Cognitive-behavioural therapy	MPH SR: 43.4%	ADHD: 47.2% response rate (placebo: 54.7%), NS.	No group differences were observed with respect to AEs. Only two patients were discontinued because of AEs (one in the MPH SR group due to persistent insomnia). Five participants had their doses lowered (four in the MPH SR group due to rash, insomnia, depressed mood, and clenched jaw, respectively)	Exclusion criteria included current psychiatric disorders (other than ADHD and SUD) and physiological dependence on opioids, sedatives or alcohol. Compliance (measured by self-report and by riboflavin fluorescence testing) was good and did not differ between groups
		83.0%		14 weeks	40–60 mg day−1		placebo: 45.3%	SUD: Approximately 70% of the weeks were cocaine-positive. 17 patients (16%) achieved 2 consecutive weeks of abstinence; NS. Both groups significantly reduced their craving over time; NS. The MPH SR group had a lower likelihood of cocaine use over time.
								ADHD x SUD: MPH SR-responders (based on a semi-structured clinical interview) were more likely to have a reduction in cocaine use compared to non-responders
Wilens (2008) [26]	147	34.6 (10.0)	Alcohol abuse or dependence	Double-blind, placebo-controlled, parallel-group; out-patient	ATX	None (outside of 12-step programmes)	ATX: 44.4%	ADHD: Primary and secondary measures of ADHD symptoms were significantly improved in the ATX group compared with placebo.	No serious AEs or specific drug-drug reactions related to current alcohol use. Discontinuation rates due to AEs were low in both groups and not significantly different. AEs significantly more prevalent in ATX-treated patients were nausea, dry mouth, decreased appetite, dizziness, fatigue, constipation, urinary hesitation, hot flush, and paraesthesia. Pulse, blood pressure, and QTcF change from baseline measurements were similar between groups	Patients had to be abstinent from alcohol at least 4 days (maximum 30 days) before randomization. Exclusion criteria included diagnosis of current bipolar disorder, major depressive disorder, or psychosis. History of non-alcohol SUD did not preclude participation provided that patients were not actively abusing other substances. Psychotherapy, pharmacological, or other interventions for SUD (other than 12-step programmes) were not permitted
		85.0%		12 weeks	89.9 (17.6) mg day−1		Placebo: 64%	SUD: Time to initial relapse to heavy drinking showed no difference between groups. ATX-treated patients had a significantly lower rate of cumulative heavy drinking days.
					Max = 100 mg day−1			ADHD x SUD: No statistically significant differences between ATX responders and non-responders on any alcohol use variable
Konstenius (2010) [27]	24	37.4 (9.9)	Amphetamine dependence	Double-blind, placebo-controlled, parallel-group; out-patient	OROS MPH	Skills training programme	OROS MPH: 59%	ADHD: Significant reduction of ADHD symptoms in both groups; NS.	Most AEs were mild in severity and reversible. Most commonly reported AEs were headache and nausea. Only one severe AE (blurred vision, which disappeared with dose reduction) occurred in one participant	Exclusion criteria included current or past diagnosis of any other substance dependence except nicotine, and history of any major psychiatric disorder or any psychiatric condition requiring medication. All patients were required to stay abstinent from all psychoactive substances (minimum of 4 weeks) before entering the trial
		75%		13 weeks	72 mg day−1		Placebo: 84%	SUD: No significant differences between groups in AMP use (self-report and urinalysis), craving for AMP, time to relapse or cumulative abstinence duration
McRae-Clark (2010) [28]	38	29.9 (11.5)	Marijuana dependence	Double-blind, placebo-controlled, parallel-group; out-patient	ATX	Motivational interviewing	ATX: 47.4%	ADHD: Participants treated with ATX showed greater improvement on the CGI-I scale than participants randomized to placebo. No treatment group differences on other ADHD rating scales.	All ATX-treated participants experienced at least one AE compared with 84% of participants randomized to placebo; NS. The risk of gastrointestinal AEs was 2.25 times higher for ATX-treated patients	Exclusion criteria included dependence on any other substance (with the exception of caffeine or nicotine), history of psychotic disorder, current major depression, and current treatment with psychoactive medication
		76.3%		12 weeks	Max = 100 mg day−1		Placebo: 36.8%	SUD: No statistically significant differences between treatment groups in marijuana use outcomes
Winhusen (2010) [29]	255	37.8 (10.0)	Nicotine dependence	Double-blind, placebo-controlled, parallel-group; out-patient	OROS MPH	Nicotine replacement therapy and smoking cessation counselling	OROS MPH: 84.3%	ADHD: 71% response rate (placebo: 44%), statistically significant difference.	No participant discontinued due to an AE. TEAEs reported at significantly higher rates in the OROS MPH group included dyspepsia, decreased appetite, heart rate increase, and palpitations. Tolerability for treatment and the nicotine patch did not differ between groups	Exclusion criteria included current non-nicotine SUD, current major depression, current anxiety disorder (except specific phobias), antisocial personality disorder, lifetime diagnoses of bipolar disorder or psychosis, and positive urine screen for an illicit drug. Medication adherence was high and did not differ significantly between treatment groups
		56.5%		11 weeks	Max = 72 mg day−1		Placebo: 84.4%	SUD: No significant difference between treatment groups in prolonged abstinence or point-prevalence abstinence rates.
								ADHD x SUD: No significant responder x treatment interaction effects for prolonged abstinence and point prevalence abstinence
Kollins (2011) [30]	17	35.1 (10.3)	History of SUD (abuse of alcohol, marijuana, m-AMP, or opioids; dependence on alcohol, AMP, or m-AMP)	Double-blind, placebo-controlled, parallel-group; out-patient	LDX	NR	History of SUD: 83.3%	ADHD: 65% response rate (without a history of SUD: 59%), NS	LDX was generally well tolerated. Two participants (11.8%) with a history of SUD discontinued due to psychiatric TEAEs. TEAE profile of participants with a history of SUD was similar to that of participants with no history of SUD	Exploratory, post hoc analysis on data from a forced dose titration study (n = 420). Exclusion criteria included a current co-morbid Axis I or II diagnosis, a positive urinalysis, and SUD (excluding nicotine) within the past 6 months. No patient with a history of SUD was randomly assigned to the placebo group (n = 62)
		54.3%		4 weeks	30, 50, or 70 mg day−1		No history of SUD: 82.9%
Kollins (in press) [31]	32	31.4 (8.7)	Nicotine dependence	Double-blind, placebo-controlled, parallel-group; out-patient	LDX	Nicotine replacement therapy	LDX: 82.4%	ADHD: Significant reductions in self-reported and clinician-rated ADHD symptoms compared with placebo.	LDX was generally well tolerated with respect to AEs and cardiovascular functioning. One participant (5.9%) in the LDX group was discontinued because of a severe AE (marked anxiety) possibly related to study drug. No significant differences between groups in cardiovascular function.	Exclusion criteria included the presence of any other psychiatric condition and use of psychoactive medication or illegal drugs
		62.5%		6 weeks	70 mg day−1		Placebo: 93.3%	SUD: No effects compared with placebo on rates of continuous abstinence (verified by self-report and CO levels)

Response rate refers to the percent of participants meeting the standard response criterion for the primary outcome measure (e.g. CGI-I score < 3 [i.e. much to very much improved] or a ≥ 30% reduction in symptoms in ADHD rating scales when comparing the last observation to baseline). ADHD, Attention-deficit hyperactivity disorder; AE, Adverse event; AMP, Amphetamine; ATX, Atomoxetine; BPR, Bupropion; LDX, Lisdexamfetamine dimesylate; m-AMP, Methamphetamine; MPH, Methylphenidate; NR, Not reported; NS, No statistically significant difference between groups; OROS, Osmotic release oral system; QT_cF, Corrected QT interval by Fridericia's formula; SR, Sustained release; SUD, Substance use disorder; TEAE, Treatment-emergent adverse effect.

Alcohol dependence and ADHD

Only one controlled study has been carried out in patients with alcohol dependence and ADHD [26]. The 12 week study included 147 patients (all with a diagnosis of alcohol abuse or dependence) assigned to treatment with atomoxetine or placebo. Subjects were required to abstain from alcohol in the 4 days prior to the start of the study. Patients in the atomoxetine treatment group showed significantly better improvement in ADHD than the placebo group. Although no significant differences between groups were observed regarding time to initial relapse to heavy drinking, the atomoxetine treatment group presented a significantly lower rate of cumulative heavy drinking days. When responders to atomoxetine were compared with non-responders, no significant differences were observed in the variables related to alcohol use. Although various adverse events (e.g. nausea, dry mouth) were observed more frequently in the treatment group, this finding did not yield any significant differences between groups in the discontinuation rate due to adverse events. No specific drug–drug reactions related to current alcohol use were observed.

Amphetamine dependence and ADHD

The only controlled study performed to assess the efficacy of pharmacological treatment of adult ADHD in patients with co-morbid amphetamine dependence [27] found no differences between the treatment and placebo groups. Subjects were assigned to treatment with osmotic-release oral system (OROS) methylphenidate or placebo. Although ADHD symptoms decreased in both groups, no significant differences were observed in symptom improvement or for any of variables related to the addictive disorder (e.g. use, craving, time to relapse, cumulative abstinence duration). It should be noted that patients were required to remain drug free (minimum of 4 weeks) to participate in the trial. Although most patients reported adverse events (e.g. headache, nausea), these were mild in severity and reversible.

Cannabis dependence and ADHD

In adults with ADHD, cannabis is the most commonly abused illegal substance [32]. However, only one controlled study has been carried out to evaluate the efficacy of a pharmacological agent to treat adult ADHD in cannabis-dependent patients [28]. In that study, participants were assigned to treatment with atomoxetine or placebo. Patients in the atomoxetine treatment group had a greater improvement on the Clinical Global Impression of Improvement (CGI-I) scale although not on other ADHD rating scales. No significant between group differences were observed in any of the marijuana use outcomes. One noteworthy (though not statistically significant) finding was the increased risk of gastrointestinal adverse events, which were 2.25 times greater in the atomoxetine treatment group vs. placebo.

Cocaine dependence and ADHD

Three double-blind, placebo-controlled trials have been carried out in patients with this dual disorder, with contradictory results regarding the efficacy of the various methylphenidate formulations that were evaluated [22,24,25]. We discuss one of these studies, in which opioids were the primary drugs of abuse, in Opioid dependence and ADHD below.

The first clinical trial performed in adults with cocaine dependence and ADHD evaluated the therapeutic efficacy of methylphenidate [22]. The authors found that methylphenidate was significantly better than placebo in improving ADHD symptoms as measured by both physician-rated and self-rated indexes. Nevertheless, no significant differences were found on any of the variables relative to cocaine-dependence (e.g. self-reported cocaine use days, percentage of cocaine-negative urinalyses, craving levels). Although insomnia or trouble sleeping were significantly more common in the methylphenidate group, none of the participants abandoned treatment due to adverse events.

A second controlled trial in patients with this dual disorder evaluated the efficacy of sustained-release methylphenidate vs. placebo [25]. The use of methylphenidate did not significantly decrease ADHD symptoms vs. placebo. Moreover, the placebo response (54.7%) was remarkably strong. Although no significant between group differences were found for most variables related to the cocaine addiction (e.g. percentage of patients able to remain abstinent for 2 consecutive weeks, craving levels), the methylphenidate group did show a reduced likelihood of cocaine use over time. Moreover, differences were observed within the active treatment group between responders and non-responders. Patients whose ADHD symptoms improved after methylphenidate therapy were more likely to reduce cocaine use than patients who showed no improvement.

Nicotine dependence and ADHD

According to one study in young adults, the number of ADHD symptoms has a linear relationship with the probability of being a regular smoker [33]. Nevertheless, to date, only two controlled studies have assessed whether ADHD medications enhance response to smoking cessation treatment in smokers with co-morbid ADHD [29,31]. Concomitant nicotine replacement therapy was used in both of these studies, which we describe below.

In the first of these studies [29], patients treated with OROS methylphenidate were compared with a placebo treated group. The active treatment group had a higher response rate, defined as a ≥ 30% reduction in the DSM-IV ADHD Rating Scale [34] and a 1-point or greater reduction in Clinical Global Impression-Severity scale from baseline measures to the final week of the study. Nevertheless, for other variables, including prolonged abstinence, point prevalence abstinence rates, and treatment by responder interaction, no significant differences were observed between the groups. It is important to note that patients in the active treatment arm were more likely to experience adverse events (e.g. dyspepsia, decreased appetite) although this did not result in significant differences between the groups in the discontinuation rate due to adverse events. Likewise, no significant between group differences were observed in nicotine patch tolerability.

The second controlled study in patients with nicotine-dependence and co-morbid ADHD [31] reported that lisdexamfetamine dimesylate (LDX) was significantly more effective than placebo in reducing both self-and clinician-reported ADHD symptoms. However, between-group differences in continuous abstinence rates (verified by self report and CO levels) and cardiovascular functioning were not significant.

Opioid dependence and ADHD

Only one randomized clinical trial has been carried out to evaluate ADHD treatment in patients with concomitant opioid dependence. That trial involved 98 patients on methadone maintenance treatment randomized to sustained release methylphenidate, sustained release bupropion or placebo [24]. None of the active medications was found to be more effective than placebo on any of the outcome variables related to ADHD, opioid dependence, or concomitant cocaine dependence. The strong clinical response to placebo was noteworthy: on the self-rated Adult ADHD Rating Scale (AARS) [35], 46% of participants in the placebo group reported a large decrease in ADHD symptoms. No cases of misuse of the methylphenidate or bupropion sustained release preparations were observed, nor were any cases of intensification of cocaine use reported in these two treatment groups.

Safety and abuse liability of stimulant and non-stimulant medications for ADHD in patients with co-morbid SUD: Experimental and laboratory findings

In addition to the data provided by controlled clinical trials, additional data on the safety and abuse liability of some of the medications used to treat adult ADHD have been obtained through a few experimental studies involving drug users.

Safety

One study assessed the effects of intravenous cocaine administration during treatment with sustained-release methylphenidate (up to 60 mg day⁻¹) in cocaine dependent adult patients with ADHD [36]. Importantly, this study found that the cardiovascular effects of cocaine administration were not clinically significant in the study group, thus suggesting that maintenance on sustained release methylphenidate is safe in cocaine dependent patients with ADHD.

Another study, in which patients were recruited without regard to ADHD, evaluated the safety of immediate release methylphenidate (up to 90 mg day⁻¹) in cocaine dependent individuals receiving concurrent intravenous cocaine [37]. The principal findings of that study can be summarized as follows: (i) methylphenidate in combination with cocaine is well-tolerated, (ii) methylphenidate does not significantly alter the pharmacokinetics of cocaine, (iii) no clinically significant ECG findings were observed with respect to interaction between the two substances and (iv) methylphenidate reduces some of the positive subjective effects of cocaine.

Finally, another study examined the safety, tolerability and subject-related effects of acute intranasal cocaine administration during atomoxetine maintenance (up to 80 mg day⁻¹) in cocaine dependent individuals [38]. Results showed that atomoxetine attenuates the systolic pressure increasing effects while enhancing the heart rate increasing effects of cocaine, but otherwise has no behavioural effects. The authors conclude that cocaine is well-tolerated during atomoxetine maintenance.

Abuse liability

Two separate studies reported that the abuse liability of atomoxetine in drug users is scant and notably lower than non-modified release preparations of methylphenidate. The subjective effects of atomoxetine (up to 90 mg day⁻¹) and methylphenidate (up to 40 mg day⁻¹) were compared in recreational drug users with no history (aside from nicotine) of drug or alcohol dependence [39]. The authors concluded that atomoxetine does not induce subjective effects that are similar to those elicited by methylphenidate. Another study compared the abuse liability of atomoxetine (45, 90, and 180 mg day⁻¹), methylphenidate (90 mg day⁻¹), phentermine (60 mg day⁻¹), desipramine (100 and 200 mg day⁻¹) and placebo in subjects with a diagnosis of current stimulant abuse [40]. Findings revealed that participants liked methylphenidate and phentermine significantly more than placebo, atomoxetine and desipramine. In contrast, none of the atomoxetine doses were liked significantly more than placebo. Mean liking scores for atomoxetine were similar to, or significantly lower than, mean scores for desipramine.

Other studies have found that the abuse liability of OROS methylphenidate in recreational drug users is much lower than that of non-modified release methylphenidate. In one randomized crossover study [41], participants received single oral doses of placebo, methylphenidate (60 mg) and OROS methylphenidate (108 mg). Findings showed that the subjective effects of methylphenidate and OROS methylphenidate differed significantly from placebo. Even though the dose of OROS methylphenidate was higher than the methylphenidate dose, abuse-related subjective effects were consistently lower for OROS methylphenidate. A second study [42] evaluated the abuse-related subjective effects of comparable doses of OROS methylphenidate (54 mg and 108 mg) and methylphenidate (50 mg and 90 mg). In that study, only the lower dose of OROS methylphenidate (54 mg) failed to produce statistically significant differences when compared with placebo. At comparable dose levels, OROS-methylphenidate produced lower positive and stimulant subjective effects than methylphenidate. Moreover, low dose methylphenidate (50 mg) produced greater subjective effects than high dose OROS methylphenidate (108 mg).

A recent study investigated the abuse liability of single oral doses of 50, 100 and 150 mg of LDX in adults with a current diagnosis of stimulant abuse [43]. Although no significant difference in liking scores between non-modified release 40 mg dexamphetamine and 150 mg LDX were observed, liking scores for 100 mg LDX were significantly lower than for 40 mg dexamphetamine.

The results of the studies reviewed in this subsection strongly suggest that drugs used to treat adult ADHD can be classified into three categories according to their abuse liability, listed here from higher to lower abuse liability: (i) non-modified release stimulants (e.g. immediate release methylphenidate and dexamphetamine), (ii) extended release or long acting stimulants (e.g. OROS-methylphenidate, LDX) and (iii) non-stimulant agents (e.g. atomoxetine).

Final considerations

The methodology and the results of the controlled trials carried out in SUD patients with co-morbid adult ADHD raise a number of practical issues, addressed below, that should be taken into account by clinicians analyzing the possible answers to the first dilemma.

First, it appears that although the various drugs (both stimulants and non-stimulants) assessed in these studies can improve ADHD in the short term, they are less effective in ADHD patients with addictive disorders than in patients without these co-morbid disorders. Two meta-analyses, both of which concluded that methylphenidate appears to be less efficacious in adult patients with co-morbid SUD, support this conclusion [44,45]. This lack of efficacy appears to remain unchanged even when patients refrain from using psychoactive substances (including the primary dependence-related substance) [23]. It seems unlikely that this lack of efficacy is due to poor adherence to pharmacological treatment because treatment compliance was high in the double-blind placebo clinical trials that have investigated this issue [22,24,25,29]. Moreover, one open study did not find any association between differences in treatment compliance and ADHD improvement [16]. Consistent with this counterintuitive result, a recent review concluded that the available evidence is ambiguous in terms of whether differences in adherence affect clinical outcomes in patients treated for adult ADHD [46].

Second, these ADHD drugs appear to have little to no efficacy in the treatment of co-morbid addictive disorders. Although this lack of efficacy may be somewhat disappointing, one positive aspect of note is that these ADHD drugs do not appear to exacerbate addictive disorders. None of the controlled clinical trials performed to date have found any evidence of exacerbation, nor have any cases of medication abuse or misuse been reported. Nevertheless, it should be noted that a few rare cases of abuse/misuse have been reported in uncontrolled clinical studies.

Third, ADHD medications seem to exert a strong placebo effect on ADHD symptoms in patients with addictive disorders. Some studies report placebo response rates as high as 45%–55% [24,25,29]. The high placebo response may be related to a non-specific effect of addiction treatment on impulsivity and aggressiveness. Such an effect probably lessens the severity of some clinical manifestations of ADHD. Another possible explanation is that patients with addictions, and/or the clinicians who treat them, have high expectations that medical treatment will improve ADHD symptoms. In any case, this information needs to be considered when interpreting the results discussed above, and to ensure adequate sample sizes in any new controlled clinical trials.

Fourth, in the large majority of controlled studies carried out to date, participants received some type of psychosocial intervention (mainly cognitive behavioural therapy). Consequently, clinicians must consider the reported efficacy of these drugs, and even the placebo response rates, in the context of concomitant psychosocial treatment. However, it is important to note that the psychological interventions in most of the studies were primarily designed to treat the addictive disorder. This approach is very different from the one taken in most clinical studies designed to assess the efficacy of a specific drug in the treatment of ADHD patients without co-morbid SUD. In those studies, psychological interventions (when offered) are performed to reduce ADHD symptoms.

Fifth, and last, none of the controlled clinical trials discussed here stratified participants by dependence severity or by ADHD subtype. These variables may be important according to a recent secondary analysis [47] of the data collected in the original Winhusen et al. study [29]. In the original study [29], OROS methylphenidate was not superior to placebo in improving response to smoking cessation treatment. However, the secondary analysis revealed that ADHD subtypes present divergent smoking cessation responses to OROS methylphenidate or placebo as a function of severity of nicotine dependence [47]. More specifically, the secondary analysis found: (i) a significant benefit of OROS methylphenidate vs. placebo for the combined (inattentive and hyperactive/impulsive) ADHD subtype when nicotine dependence is high, (ii) a worse response to active drug among smokers with the predominantly inattentive subtype when nicotine dependence is high and (iii) no added benefit of OROS methylphenidate regardless of ADHD subtype when nicotine dependence is low.

Conclusion

Evidence from large open studies and controlled trials in patients with SUD and co-morbid ADHD suggests that the drugs typically used in the treatment of adult ADHD, particularly long acting or extended release preparations of stimulants and the non-stimulant atomoxetine (all of which are generally administered in conjunction with diverse psychological interventions), offer promising though not conclusive efficacy results in the short term treatment of ADHD in patients with SUD but without other severe mental disorders. Despite the lack of long term studies, these drugs have an adequate tolerability profile in adult patients with SUD and ADHD co-morbidity. However, at the same time, the efficacy of these drugs is limited or non-existent in SUD treatment. Nevertheless, the adequate tolerability profile of long acting stimulant preparations and atomoxetine in this group of patients, together with their favourable drug–drug interaction profile, suggest that these drugs are unlikely to raise any important safety issues when used to treat patients with SUD-ADHD co-morbidity.

Must we wait until SUD is under control before initiating pharmacological treatment of adult ADHD?

The two possible responses to this question (i.e. yes or no) could lead to clinically divergent scenarios. On one hand, delaying ADHD treatment until SUD is controlled would create ideal conditions for diagnosing and treating ADHD because it would eliminate the interference from the SUD. However, achieving satisfactory control of SUD is very difficult in severe cases, such as those often associated with ADHD. Moreover, waiting until SUD is under control might require an extended or even indefinite postponement in starting ADHD treatment. On the other hand, we could opt to treat the ADHD without waiting to control SUD. This would improve the ADHD and might make it easier to control the SUD. However, this plan requires that ADHD be diagnosed and treated in a clinically complex context, thereby increasing the probability of therapeutic failure. Moreover, the reinforcing effect of the stimulants used to treat ADHD increase the risk of SUD exacerbation.

In clinical practice, it is not possible to avoid making a treatment decision. We must, inevitably, select one of these two approaches to manage co-morbid SUD and ADHD even though, as we discussed previously, conclusive results regarding the efficacy of pharmacological treatment of ADHD in patients with SUD are not available. However, despite the lack of a definitive treatment approach, a thorough understanding of the advantages and disadvantages of pharmacological treatment of ADHD in patients with active SUD might simplify the clinical decision making process. The purpose of the present section of this review is to assess the relevant aspects of this dilemma. However, this analysis is limited almost exclusively to ADHD medications with the least abuse liability (i.e. extended release stimulants and atomoxetine), which are precisely those medications that are typically used to treat adult ADHD in patients with SUD.

ADHD treatment can help to control SUD better

This approach relies on the possibility that adult ADHD is a causal factor for SUD. If this hypothesis is true, then pharmacological management of ADHD would be an aetiological treatment for SUD. Moreover, the drugs used to treat ADHD might have a direct, positive effect, unmediated by ADHD, on SUD.

Adult ADHD as a causal factor for SUD

Before a decision can be made to treat or not adult ADHD in order to control SUD, it is essential to determine whether the persistence of ADHD in adults is a causal factor of persistent SUD. To assess fully whether adult ADHD contributes to SUD persistence, we must first ascertain whether adult ADHD promotes the lifetime incidence of SUD. To our knowledge, this possible association has yet to be investigated. Most studies that have been performed in this area have evaluated the association between childhood or adolescent ADHD and substance use or addiction during adolescence or in young adults [48–50].

ADHD as a causal factor in the development of SUD

At present, the available data on the role of childhood or adolescent ADHD as a causal factor for SUD is contradictory. To evaluate this association, it would be necessary to investigate whether childhood ADHD is an independent predictor of SUD or if the pharmacological treatment of childhood ADHD prevents the emergence of SUD. However, in both of these scenarios, the presence or not of conduct disorder must also be considered due to the high co-morbidity between ADHD and conduct disorder [51] and because conduct disorder is one of the most robust risk factors for SUD [48,50,52].

According to a recent series of meta-analyses, childhood ADHD is a predictor for SUD in young adults [49]. However, this finding must be considered in context. Most of the studies included did not adjust for the presence of conduct disorder. Moreover, findings from the two well-designed studies that did adjust for conduct disorder were contradictory. In one study, the association between childhood ADHD and SUD was found to persist into young adulthood [53] while the other found that this association was not maintained [54].

One meta-analytic review found that treatment of ADHD with stimulants during childhood and adolescence prevented the development of SUD [55]. However, once again, findings were not adjusted for the presence of conduct disorder. Several subsequent studies, in which the results were adjusted for the presence of conduct disorder (or antisocial personality disorder, the adult equivalent to conduct disorder), have shown that the pharmacological treatment of ADHD neither increases nor reduces the risk of SUD in adults [56–58].

If conduct disorder alone were a causal factor for SUD, treating ADHD to control SUD would not be effective. However, this line of reasoning may not apply to ADHD patients who also suffer from conduct disorder. Substance use in young adults with a history of childhood ADHD alone (i.e. without conduct disorder) is less intense and diverse than substance use in patients who had both ADHD and conduct disorder in childhood [59]. This finding supports the hypothesis that childhood ADHD and conduct disorder interact synergistically, leading to a particularly severe form of SUD [48]. Two 5 week, double-blind clinical trials found that methylphenidate reduces conduct disorder in children with [60] and without ADHD [61]. However, it is clearly not possible to extrapolate results obtained in patients with childhood ADHD and conduct disorder to adults with ADHD and antisocial personality disorder. That said, antisocial personality disorder is probably a predictor for non-medical stimulant use during treatment for adult ADHD.

Adult ADHD as a causal factor in persistent SUD

It is unclear whether ADHD contributes or not to the emergence of SUD in young adulthood, although published data support the hypothesis that adult ADHD contributes to persistent SUD. Several transversal studies have reported that patients with adult ADHD experience more severe SUD than their counterparts without adult ADHD. This difference (unadjusted for the presence of conduct disorder) has been observed in studies in which patients with co-morbid SUD-ADHD presented more substance use disorders [62], more intense drug craving [63], greater severity of substance use [64], and, importantly for SUD persistence, a worse response to various addiction treatments [65,66]. In studies that did adjust for conduct disorder and other mental disorders, adult ADHD was associated with earlier onset [67] and more persistent SUD [68].

The clinical alterations that characterize adult ADHD could contribute to SUD persistence through two non-exclusive mechanisms. Awareness of these mechanisms has important therapeutic implications for clinicians. The first mechanism is related to alterations in attention [69] and impulsivity [70], both of which favour persistent SUD. In addition, ADHD is often associated with deficits in executive functions [71,72], and such dysfunctions have also been considered to have an aetiological relation to SUD [73]. The second mechanism by which adult ADHD might contribute to SUD persistence is through the use of substances of abuse to self-medicate ADHD [74]. This self-medication hypothesis assumes that ADHD symptoms improve when a specific substance is consumed. However, this may not be true. A laboratory study of cocaine abusers with ADHD found that acute administration of cocaine worsened self-ratings for ‘Able to concentrate’, and ‘Calm’ [36]. Several studies suggest that nicotine is used to self-medicate ADHD. Indeed, for this reason, pharmacological treatment of ADHD is considered crucial in smoking cessation treatments [75]. During abstinence from nicotine, smokers with ADHD show worse performance on tasks requiring attentional control and response inhibition when compared with counterparts without ADHD [76]. In addition, two double-blind crossover studies suggest that a nicotine patch is more effective than a placebo patch in improving the ability of smokers with ADHD to concentrate 8–12 h after interruption of nicotine consumption [77,78].

Utility of pharmacological treatment for ADHD to treat SUD

Adult ADHD has been linked to a dysregulation of brain dopamine and norepinephrine systems [79], which explains the therapeutic efficacy of drugs that act as catecholaminergic agonists. Given that SUD has also consistently been linked to dysregulation of the brain dopamine system [80,81], stimulants and other drugs that affect that system might be effective in treating SUD regardless of their impact on ADHD. The results discussed below show, however, that despite this possible mechanism, drugs used to treat ADHD have not, in general, proven effective in treating SUD.

One meta-analysis reviewed a total of nine controlled trials and concluded that stimulants are not effective in treating cocaine dependence [82]. In that meta-analysis, the authors evaluated two clinical trials in which patients with adult ADHD were excluded and only one clinical trial that specifically included ADHD. The remaining six trials either did not evaluate ADHD or did not report whether ADHD was evaluated or not. Two double-blind placebo-controlled pilot trials found that modafinil improved mnestic function in patients with methamphetamine dependence without ADHD [83] or without other Axis I psychiatric disorders (exclusion due to ADHD was not specified) [84].

An 8 week, randomized, double-blind, clinical trial of smokers without ADHD concluded that extended release methylphenidate was not effective in treating nicotine dependence [85]. In contrast, a double-blind, crossover laboratory study [86] found that atomoxetine reduced symptoms of nicotine abstinence in patients without ADHD. Atomoxetine was also effective in treating nicotine addiction in subjects who completed a 21 day, parallel, double-blind, clinical trial in which patients with any other Axis I psychiatric disorder were excluded (exclusion due to ADHD was not specified) [87]. However, five of the nine participants treated with atomoxetine in that study abandoned treatment due to the adverse events associated with the drug. An 11 week, open label study of cannabis dependent patients without co-morbid ADHD concluded that atomoxetine should not be used to treat this addictive disorder because efficacy was limited or non-existent and more than 75% of participants experienced clinically significant adverse gastrointestinal events [88].

Patient belief that pharmacological treatment of ADHD is a priority

In our clinical experience, the acceptability of a diagnosis of ADHD is very high in young adults with SUD and some even request treatment because they are convinced that they suffer from ADHD. This attitude towards an ADHD diagnosis contrasts with the low acceptability often observed when a diagnosis of SUD is made. We suspect that the high acceptability for ADHD diagnosis in SUD patients may be related to the patient's desire to receive pharmacological interventions to treat his/her problems, or it may reflect an intention to use stimulants for non-medical purposes.

Patients who show a high acceptability for an ADHD diagnosis sometimes assume that the disorder is the cause of their history of disruptive behaviour and SUD. Such patients may harbour the hope that pharmacological treatment of their ADHD will also solve their other problems. Obviously, this expectation implies a high risk of abandonment of SUD treatment if the request for pharmacological treatment of ADHD is not met quickly.

In patients who believe that pharmacological treatment for ADHD is a priority and who show a strong interest in receiving medical treatment for this disorder, we recommend the following approach. First and foremost, determine if the patient fulfils diagnostic criteria for adult ADHD. If the diagnosis is confirmed, or at least considered highly probable, the next step is to redirect the patients' expectations towards realistic therapeutic objectives. To do this, the patient should be informed of the following facts: (i) remission of ADHD is not necessarily associated with SUD remission because SUD does not have a single cause, (ii) both ADHD and SUD require specific therapeutic interventions, (iii) adult ADHD does not always respond well to pharmacological treatment and (iv) treatment of SUD will likely improve ADHD symptoms. Once the patient's expectations have been lowered, pharmacological treatment for ADHD can be initiated under the close supervision of nursing staff, as described in Risk of non-medical drug use.

Diagnostic uncertainty in adult ADHD

Diagnosis of adult ADHD requires that clinicians evaluate not only current symptoms, but that they also retrospectively evaluate symptoms that were present in the past (i.e. during childhood and adolescence). However, these clinical evaluations are particularly difficult to perform in patients with active SUD [7,89].

Clinical evaluation of ADHD during adulthood in a patient with SUD is difficult due to changes that occur in attention, motor activity and/or impulsiveness [90] in the context of substance intoxication or withdrawal. Given the overlap between ADHD and SUD symptoms, it is essential that clinicians be prepared to evaluate for ADHD when patients interrupt their substance use, as occurs when they are admitted to a detoxification unit. No definite criteria are available to determine how long a patient should abstain from substance use before an assessment of ADHD can be performed. However, some experts assert that 1–4 weeks of abstinence is sufficient to allow current ADHD to be assessed with a reasonable degree of diagnostic certainty [91]. The clinical evaluation should be performed by an experienced clinician who is capable of deciding whether the attention or impulsivity problems observed are to be expected in a drug dependent patient who has not used substances for 1–4 weeks, or whether such symptoms would be better explained by ADHD.

Substance abuse can hinder the retrospective evaluation of childhood or adolescent ADHD. Firstly, ADHD can emerge after age 7 years [92] while substance use can appear in early adolescence, thus making it difficult to determine retrospectively if the patient suffered ADHD symptoms before starting to use psychoactive substances. Moreover, mnestic difficulties related to substance use probably reduce the validity of any assessment of past ADHD [89]. Furthermore, the impact of SUD on relationships with the family of origin reduces the likelihood of having access to external informants capable of describing patient functioning during the early stages of life.

Patients with SUD and ADHD during adulthood can be divided into two groups: those with adult ADHD and those that only present ADHD in the context of substance abuse [11,62]. In patients with adult ADHD, the onset of the disorder may have occurred before age 7 years (as required by DSM-IV criteria for a diagnosis of ADHD), or after age 7 years, in which case the DSM-IV diagnosis is ‘ADHD, not otherwise specified’ [90]. However, such a distinction is not clinically relevant in SUD patients because it does not discriminate with regard to SUD prevalence and associated problems [92,93].

Risk of non-medical drug use

Non-medical drug use can involve abuse related to SUD, misuse to enhance performance of cognitive, athletic, sexual or other functions, and diversion, usually to obtain the substance of abuse preferred by the patient. Obviously, patients with SUD are at a particularly high risk of engaging in non-medical use of ADHD medications. In some cases, several of the various non-medical uses of stimulants described above are observed in a single patient. To give just one example, a male treated with extended release methylphenidate at our clinic took the drug after dinner on evenings when he planned to go out to bars/discotheques. He told us that he used the drugs to avoid becoming sleepy and to increase his sexual function and alcohol tolerance.

To minimize the risk of non-medical stimulant use, pharmacological treatment of adult ADHD in SUD patients usually involves drugs that have an abuse liability that is considered either low (extended release stimulants) or very low (atomoxetine) [39,40]. Extended release stimulants have been found to have a reinforcing effect in some patients with a history of recreational use or abuse of substances [94], even in cases where the proper route of administration is used [41], and especially at elevated dose levels [42,43]. In our experience, the risk of non-medical use of extended release stimulants can be reduced considerably by administering the medication under the close supervision of nursing care, as occurs with other medications (e.g. methadone) in use at SUD treatment centres.

Nursing supervision involves evaluating the patient 1–3 times per week to assess his/her state, administering the dose scheduled for that day, providing the doses planned for home use until the next nursing appointment, and collecting a urine sample for laboratory analysis (i.e. to check for the presence of substances of abuse and, if relevant, methylphenidate). Urinary analysis permits detection of any changes in the pattern of substance use and allows us to determine if the patient is not taking methylphenidate due to diversion. The frequency of nursing supervision can be progressively reduced if treatment compliance is adequate. Nursing supervision should be closer in patients that present a greater risk of non-medical stimulant use. In our clinical experience, high risk patients are those who request immediate treatment with stimulants (see Patient belief that pharmacological treatment of ADHD is a priority), those who report a prior history of non-medical stimulant use, and patients who are addicted to various substances or suffer from other mental disorders in addition to co-morbid SUD-ADHD. To our knowledge, no studies have been performed to identify predictors for the non-medical use of stimulants in adults with SUD.

The availability of drugs with a low or very low risk of abuse, the high level of safety in terms of interactions between these drugs and substances of abuse [37,38], and the clinical skills of professionals who work in SUD treatment centres all suggest that the risk of non-medical drug use is not a strong argument against treating ADHD with uncontrolled SUD.

Risk of exacerbating other mental disorders

A large percentage of patients with adult ADHD present other mental disorders in addition to SUD [95], some of which have a bidirectional relationship with SUD (e.g. bipolar disorder and eating disorders). As discussed in the previous subsection, the presence of such disorders can increase the risk of non-medical stimulant use. In addition, the clinical manifestations of other mental disorders, added to the co-morbidity of SUD-ADHD, can be exacerbated through the use of the drugs used to treat ADHD, or can be confused with the side effects of these drugs. For example, a patient with bipolar disorder could suffer mania in response to stimulant administration [96] and the manic state could worsen the SUD. Irritability, suicidal thoughts, self-harm behaviours and behavioural changes are characteristic symptoms of borderline personality disorder [90], yet they have also been reported as a side effect of atomoxetine administration [97]. As a result, it can be difficult to assess whether a decompensation of borderline personality disorder has occurred in adults with co-morbid SUD-ADHD who receive atomoxetine. In addition to the complexities of these varied clinical situations, we must also consider the lack of data regarding their management. Most clinical trials performed to evaluate drugs used to treat ADHD in adults with SUD exclude patients with other severe mental disorders (see Efficacy of pharmacological treatment in ADHD on adults with SUD-ADHD: Overview with an emphasis on the repercussions on the addictive disorder). For all these reasons, the risk of exacerbating other mental disorders that have the potential for significant interactions with the co-morbid conditions of SUD-ADHD is a strong argument against treating adult ADHD before SUD has been controlled.

Conclusion and final recommendation

The decision to treat or not treat ADHD in the context of active substance abuse depends on numerous factors. Among these, the most notable are the bidirectional relationship between SUD and ADHD (which can vary depending on the type of substance involved), the patient's opinion about the importance of ADHD in his/her particular case, the degree of diagnostic uncertainty for ADHD, the risk of exacerbating SUD or other mental disorders by treating the ADHD, the experience of the therapeutic team in the management of SUD-ADHD co-morbidity, and the ability to closely monitor the patient's response to treatment and adherence to pharmacological treatment.

Our recommendation is to individualize the treatment decision required by this dilemma. This should be done on a case-by-case basis in which the arguments for and against immediate pharmacological treatment of ADHD are carefully weighed. The results of such an assessment will indicate when to initiate ADHD treatment and the most appropriate choice of drug.

Competing Interests

All authors have completed the Unified Competing Interest form at http://www.icmje.org/coi_disclosure.pdf (available on request from the corresponding author) and declare no support from any organization for the submitted work; no financial relationships with any organizations that might have an interest in the submitted work in the previous 3 years (except in the case of JP, who declares having acted as a speaker in activities sponsored by Eli Lilly) and no other relationships or activities that could appear to have influenced the submitted work.