Does the Pharmacotherapy of ADHD Beget Later Substance Abuse? A Meta-Analytic Review of the Literature Clinical Research Program in Pediatric Psychopharmacology.

Does the Pharmacotherapy of ADHD Does the Pharmacotherapy of ADHD Beget Later Substance Abuse?Beget Later Substance Abuse?

A Meta-Analytic Review of the LiteratureA Meta-Analytic Review of the Literature

Clinical Research Program in Pediatric Clinical Research Program in Pediatric PsychopharmacologyPsychopharmacology

Massachusetts General HospitalMassachusetts General Hospital

Harvard Medical SchoolHarvard Medical School

Timothy E. Wilens, M.D.,Timothy E. Wilens, M.D.,

Stephen V. Faraone, Ph.D.Stephen V. Faraone, Ph.D.

Joseph Biederman, M.D.Joseph Biederman, M.D.

Samantha Gunawardene, B.S.Samantha Gunawardene, B.S.

• It is estimated that ADHD affects 4-9 % of youth

• Stimulant drugs are a mainstay of treatment for ADHD and are highly effective in the treatment of ADHD youth and adults

• However concerns remain as to whether the use of stimulants in ADHD youth could increase the risk for substance use disorders (Vitiello, JAACAP, 2001)

Background

• Preclinical animal studies suggest that stimulants (methylphenidate [MPH] & amphetamine [AMPH]) may lead to later kindling and other behaviors indicative of substance abuse (DEA Review MS, 1995; Kollins et al., Pharm, Biochem, Behavior 68: 2001)

• However, other preclinical studies and concerns of species generalization and routes of administration of tested compounds (IV, IM, IP, PO) has confounded animal work (Volkow et al., Arch Gen Psych, 1995; Volkow et al., J Neurosci, 2001; Garasimov et al., J Clin Pharm Ther, 2001)

Background

• Research findings in human subjects/ patients have produced discrepant findings:

A recent study by Biederman et al. 1999 (Pediatrics 104, 1999) and Wilens et al (CPDD, 1998) showed that anti-ADHD pharmacotherapy protected ADHD youth from later SUD

A study by Lambert et al. 1999 (J Learn Dis 31, 1999) showed the opposite, that cocaine and nicotine abuse were associated with previous stimulant treatment

Background

• In contrast, untreated ADHD is a risk factor for later SUD (Shekim et al., Compr Psych, 1990; Biederman et al., Am J Psych, 1993, 1995; Katusic et al, presentation, 2003)

– Noncomorbid ADHD (Wilens et al, JNMD, 1997)

–Conduct or Bipolar Disorder plus ADHD (Hechtman and Weiss, 1985 Mannuzza et al., AGP, 1993; Wilens et al. JAACAP 1999)

Background

Onset of Substance Abuse in Untreated ADHD Adults

(Retrospectively Derived)

0 10 20 30 40 50 600.0

0.2

0.4

0.6

0.8

1.0ADHDControl

Age of Onset

Pro

bab

ilit

y **p<.05 vs control

6

ADHD ADHD Risk of Substance AbuseRisk of Substance Abuse

0%

20%

40%

60%

Youth Adults

ControlADHD

Sharp rise in SA between mid-adolescence and adulthood

??

% w

ith

SU

D

Does the Pharmacotherapy of ADHD Does the Pharmacotherapy of ADHD Beget Later Substance Abuse?Beget Later Substance Abuse?

A Meta-Analytic Review of the LiteratureA Meta-Analytic Review of the Literature

Specific Aim

To examine the putative association between prior exposure to stimulants and later SUD.

Hypotheses• The first is the null hypothesis that pharmacotherapy would have no demonstrable effect on the development of SUD in ADHD children.

• The first alternative hypothesis is that exposure to pharmacotherapy would predict a higher risk for SUD in general, and stimulant abuse in particular.

• The second alternative hypothesis posited that pharmacological management of ADHD would diminish the later risk for SUD.

• One approach to reconciling conflicting findings is the application of meta-analysis

• This method evaluates if the aggregate evidence across all available studies provides evidence for statistical significance

Methods

• We conducted a systematic literature search of prospective and retrospective studies of children, adolescents, and adults with ADHD that had information relating childhood exposure to stimulant therapy and SUD in adolescence or adulthood. We searched journal articles through PubMed at the National Library of Medicine and used additional data from scientific presentations at scientific meetings. Studies had to have at least four years of followup.

Methods

• We identified the following studies:-Loney et al. 1998, 2001 also cited in Paternite et al. 1999 (tx=182, untx=37)

-Lambert et al. 1999 (tx=93, untx=81)

-Biederman et al. 1999& Wilens et al. 1999 (tx=145, untx=45, ctrls)

-Molina et al. 1999 (tx=53, untx=73)

-Huss et al. 1999 (tx=103, untx=103)

-Barkley et al. 2002 (tx=98, untx=21)

-Huss et al., 2003 (tx =92, untx = 69)

Total sample = 766 Tx with stimulants and 429 unTx with stimulants (N = 1195)

Methods

• We applied meta-analysis to all longer-term studies (>4 years) in which pharmacologically treated and untreated groups of ADHD individuals were examined for SUD outcomes

•We used a random effects meta-analysis to analyze the odds-ratios using the method of

Carlin (Statistics in Medicine, 1992)

Methods

• For the analysis, each study provided the two by two table classifying subjects by treatment status([stimulant pharmacotherapy] or not) and the subsequent development of SUD (present or not) using the odds ratio.

•For these studies, the odds-ratio estimates the increase in the odds of NOT developing SUD (i.e. protective effect) among those individuals previously treated pharmacologically compared to individuals with ADHD not treated pharmacologically.

Methods

• Meta-analyses was used to evaluate the direction and strength of the overall association, differential effects on drug or alcohol use disorders, and the potential effects of covariates

• Since for many psychiatric disorders, people with more severe disorders are usually given more intense treatments and for many disorders increasing severity predicts worse outcome, we also assessed studies for evidence of baseline severity differences between the treated and untreated ADHD groups (Faraone et al. 1992)

Methods

0 1 2 3 4 5

Lambert 1998

Biederman 1999

Huss 1999

Loney 1999

Molina 1999

Barkley 2002

Huss 2003

Odds Ratio

*

*

* indicates p<0.05

Drug Studies

protectivedeleterious

*

0 1 2 3 4 5

Lambert 1998

Biederman 1999

Loney 1999

Molina 1999

Barkley 2002

Odds Ratio

*

*

*

* indicates p<0.05

Alcohol Studies

protectivedeleterious

0 1 2 3 4 5

Lambert 1998-Alcohol

Lambert 1998-DrugBiederman 1999-Alcohol

Biederman 1999-Drug

Huss 1999-DrugLoney 1999-Alcohol

Loney 1999-Drug

Molina 1999-Alcohol

Molina 1999-Drug

Barkley 2002-Alcohol

Barkley 2002-Drug

Odds Ratio

Sensitivity Analyses

The results of this meta-analysis examining the impact of early medication treatment for ADHD in childhood on subsequent SUD outcome in adolescent and young adult years shows that stimulant pharmacotherapy for ADHD significantly decreases the risk for subsequent SUD (z= 2.4, p=0.02)

Results

Pooled Odds Ratio = 2.0Pooled Odds Ratio = 2.0

There was a significant effect of study design (z=2.9, p=0.004) indicating that studies in which groups of treated and untreated youth with ADHD had similar baseline severity found larger (protective) odds ratios than studies that had dissimilar baseline severity

Results

• As a group, the data from studies which had similar baseline severity showed a statistically significant protective effect (OR=4.3). •Data points from the two studies (Lambert, J Learn

Dis, 1999; Barkley, Pediatrics, 2002) that did not have similar baseline severity between treatment groups suggest that stimulants increased the risk for SUD outcomes (OR=0.7).

Results

Baseline Severity Issue

More Severe Illness

Poorer Outcome (confounded association with more Tx)

(Wilens et al., Pediatrics:2003)

Baseline Severity Issue

More Severe Illness

Poorer Outcome (confounded association with more Tx)

More Treatment

(Wilens et al., Pediatrics:2003)

Baseline Severity Issue

More Severe Illness

Poorer Outcome (confounded association with more Tx)

More Treatment

(Wilens et al., Pediatrics:2003)

• A meta-analysis regression found no effect of type of substance (drug vs alcohol; 2.4 vs. 4.0; z=1.1, p=0.3).

• The risk for SUD did not differ between medicated ADHD youth and non-ADHD controls

Results

0 1 2 3 4 5

Adulthood

Adolescence

Odds Ratio

*

** indicates p<0.05

Age Effect

protectivedeleterious

•We assessed for publication bias using the method of Egger (British Medical Journal, 315, 1997). This method is based on the fact that the precision of the odds-ratio increases with larger sample sizes. •Egger’s method regresses the standard normal deviate of the odds ratio (the odds-ratio divided by its standard error) against the precision of the odds ratio (the inverse of its standard error).•The publication bias statistic was NOT SIGNIFICANT (t=0.05, p=0.6) indicating that the group of studies that controlled for baseline severity did not overestimate the protective effect of stimulants

Results

• Small number of overall studies (N=7) & subjects (N = 1195)

• Some studies published, some presented (all peer reviewed)

• Majority of youth were males

• The naturalistic, nonrandomized nature of these studies may have created confounds (severity of illness, comorbidity, family history of SUD) that may have independently affected outcome.

• Not all youth stimulant treated (4% other medications)

• Duration and adequacy of TX not delineated

• SUD outcome relied on self or parental report

• Criteria used to denote abuse or dependence of substances varied between studies.

Limitations

• Despite the limitations this meta-analysis indicates that the pharmacotherapy of ADHD does NOT increase the risk for subsequent SUD.

Conclusion

• In contrast, the available data suggest that stimulant medication has a protective effect on later SUD.

Stimulant Treatment of ADHD reduces the risk for later SUD by 50%

• Stimulant treatment of ADHD appears to reduce the risk for substance abuse in half

– Does not “immunize”

against SUD

Conclusion

0

10

20

30

40

50

60

Controls ADHD

SUD in UnTX adults with ADHD & Ctrls

– Appears to reduce the risk for SUD in ADHD youth to “population risk”

Our finding of a less robust protective effect of stimulant treatment in reducing SUD in adulthood (1.7 fold) relative to adolescence (5.8 fold) is noteworthy.

– Adolescents not fully through age of SUD risk (e.g. 19 to 22 years) (Wilens et al., J Nerv Ment Dis, 1997)

– Exposure more distal to SUD– Monitoring of patients receiving medications– Other undetermined issues

Conclusion

From a public health perspective, the findings of a protective effect of stimulants for SUD in youth with ADHD are among the strongest within Child Psychiatry indicating the preventive influence of treatment on the development of substance abuse.

Conclusions

Further studies investigating the long-term SUD outcome and putative mechanisms of reduced SUD risk in youth with ADHD of both sexes treated pharmacologically are necessary.

Conclusion