Are you a Health Professional? Jump over to the doctors only platform. Click Here

ADHD: The truth about the link between pharmacotherapy and substance use disorder

Print Friendly, PDF & Email

There have been many publicised theories about the link between the use of medications in the treatment of attention deficit hyperactivity disorder (ADHD) and the possible increased risk these children may be of developing a substance use disorder in later life. Despite the paucity of current literature the findings in the existing studies have been relatively consistent. Two important studies were performed by Wilens T, Biederman J and colleagues that analysed the truthfulness of this theory to attempt to clear the waters on this subject.1,2

The aims of both of the studies carried out by Wilens T, Biederman J and colleagues were essentially the same, the aims were to refute or support the current assumption that use of medication in ADHD results in increased incidence of substance use disorder as defined by the DSM III-R criteria.1,2 The null hypothesis they used in their study was that there is no link between the development of substance use disorder and the use of pharmacotherapy in boys treated for ADHD. There were two possible alternate hypotheses, the first of which was where a positive link between medication use in ADHD and subsequent development of substance use disorder exists, and the second of which being that the use of medication would reduce the risk of substance use disorder by eliminating the features of ADHD that result in maladaptive behaviours.1,2

The study performed by Wilens T, Biederman J and colleagues in 1999 was performed on boys who were 15 years or older, this consisted of a group of 56 medicated patients, 19 non-medicated patients and 137 non-ADHD subjects.1 Those boys who were medicated at baseline were medicated for an average of 4.4 years and none of those boys who were non-medicated proceeded to start taking pharmacotherapy during the course of the study.1 The confounders of low socioeconomic status, lifetime risk of conduct disorder and lifetime history of substance use disorder were corrected for using multiple logistic regression. 1

It was found that the substances most commonly abused were alcohol and marijuana with other drugs such as hallucinogens, stimulants and cocaine being much less commonly abused.1 The study showed that statistically significant differences in the incidence of substance use disorder did exist between the medicated and non-medicated groups when followed up 4 years after baseline.1

The multiple logistic regression analysis which controlled for confounders computed odds ratios for substance use disorder as a whole and each of substance use disorders including alcohol, marijuana, hallucinogens, cocaine and other stimulants.1 The regression analysis showed that those subjects with ADHD, who were medicated at baseline, were at lower risk of substance use disorder than the control subjects.1 This was witnessed by an odds ratio of 0.15 at follow up indicating a protective effect.1 The unmedicated group of patients showed a significantly increased risk of substance abuse disorder at follow up with an odds ratio of 6.3.1

Wilens T, Biederman J and colleagues in 1999 concluded in their study, that in boys aged 15 years and over, the risk of developing a substance use disorder, subsequent to the use of medication in the treatment of ADHD, was associated with a lower risk than in those who were unmedicated.1

The study also concluded that by not using pharmacotherapy in ADHD the risk of developing a substance abuse disorder was much greater than those who were on therapy.1

The findings of this study were consistent with a review of the literature performed at the time of publication.1 It is important however, to note that this study had methodological limitations which prevented a risk comparison of substance abuse disorder between medicated ADHD subjects and non-ADHD subjects.1 Other shortcomings of the study were that despite correcting for some measurable confounders using a multiple logistic regression analysis, many other immeasurable confounders may have contributed to the results.1

Additionally there weren’t enough numbers in the analysis to give the study sufficient statistical power, this resulted in a lack of precision with the odds ratio estimates with large 95% confidence intervals.1 The study was only performed on boys aged 15 years and over, therefore the results of this study can only apply to this age and gender group.1

Since findings of this study suggest that medications have a protective effect on the relationship between ADHD and substance use disorder, it is safe to assume that it is possible that the early recognition of ADHD with appropriate medical treatment could prevent substance abuse in later life.

Although the study had several limitations its findings were important as it potentially showed how pharmacological treatment of ADHD can result in better outcomes for these children in the long term by helping to prevent the dreaded outcomes of substance use disorders thus helping these individuals better integrate into society.

The same group of researchers, in light of the limitations of their previous study and the growing pool of data on the subject, performed a meta-analysis in 2003 of all the available long term studies on the topic. 2 These were identified using a literature search.2 They tested the same three hypotheses listed in their previous study.2 The literature search revealed six studies that tested the hypothesis in question which were subsequently used in the meta-analysis.2 Five of these six studies were prospective longitudinal studies, the remaining study was a retrospective study performed on adults with ADHD.2 Two of the five prospective longitudinal studies followed up the participants for 4 years and the 3 remaining studies followed the children into adulthood.2 97% of the medicated participants in the studies were treated with stimulants (methylphenidate or amphetamine), the rest were treated with other undisclosed therapies. 2

The results of these studies were expressed as odds ratios; an odds ratio less than one indicated that there was an increased risk of developing substance use disorder with the use of stimulants, and an odds ratio greater than one being protective against this association.2 The odds ratios were deemed to be statistically significant if their 95% confidence interval did not include one and their P-value was less than 0.05.2 Combined, the six studies had a total of 674 medicated subjects and 360 unmedicated subjects. 2 There were seven statistically significant odds ratios taken from four studies which were greater than 1, indicating a protective effect of pharmacotherapy.2 Two studies showed an adverse effect of stimulants with four odds ratios from these being less than 1.2

The meta-analysis of the pooled data yielded a statistically significant odds ratio of 1.9 indicating a protective effect of the combined data, illustrating that there is up to a 2 fold reduction in the likelihood of substance use disorder in those treated with pharmacotherapy in ADHD.2 The analysis was run multiple times, each time excluding the results of one of the studies to exclude the possibility that the results of the meta-analysis were due to the results from a single study.2

It was found that no one study explained the results of the analysis and the results remained statistically significant despite one of the six studies being omitted each time.2 The meta-analysis also showed that the protective effect of pharmacotherapy in ADHD was greater in adolescence than in adulthood (OR of 5.8 vs 1.4).2

The results of this meta-analysis showed that there was a statistically significant two fold reduction in the risk of developing substance use disorder in children with ADHD that were treated with pharmacotherapy compared to those with ADHD that received no treatment.2 Four of the six individual studies showed this relationship with statistically significant odds ratios, however two of the six showed a contradictory association.2

An explanation of the contradictory individual results of some of the studies has been accounted for by the overrepresentation of conduct disorder in the medicated group, a condition which has strong links to substance use disorders.2

The shortcomings of this meta-analysis were that there were a poor number of studies (N = 6) included in the analysis, however there were sufficient numbers in the studies included in the analysis to provide adequate statistical power. 2 There were immeasurable confounding factors that could have contributed to the results of the analysis. For example socioeconomic status of the patients and their families may be related to higher rates of substance abuse in that group and may also be responsible for lack of medication use in those children. 2 The majority of data from the six studies was predominantly on males with ADHD therefore the findings cannot be extrapolated to make inferences about females with ADHD.2

Despite the shortcomings of this meta-analysis there were many interesting findings that have relevance to clinical practice. This study found a statistically significant 2 fold reduction in substance use disorders with the use of pharmacotherapy in ADHD.

This finding encourages doctors to diagnose and medically treat ADHD early so as to reduce the risk of these children developing substance abuse disorders. Substance use disorders can have severe mental, physical and economical impacts on patients with ADHD. It is therefore in the patient’s best interest to avoid such a catastrophic end-point, medications will hopefully provide a vehicle to ensure they steer away from the path of substance use.


  1. Biederman J, Wilens T, Mick E, Spencer T and Faraone SV. Pharmacotherapy of attention-deficit/hyperactivity disorder reduces risk for substance use disorder. Pediatrics 1999; 104(2): 1-5.
  2. Wilens TE, Faraone SV, Biederman J, Gunawardene S. Does stimulant therapy of attention-deficit/hyperactivity disorder beget later substance abuse? A meta-analytic review of the literature. Pediatrics 2003; 111(1): 179-85.

Print Friendly, PDF & Email


Posted On: 16 July, 2007
Modified On: 16 January, 2014


Created by: myVMC