Table 2.

The Adjusted Effect of ADHD and Pharmacotherapy on SUD Incidence

Unmedicated ADHD Versus ControlsMedicated ADHD Versus Unmedicated ADHDBaseline CD(+) Versus Baseline CD (−)
OR (95% CI)OR (95% CI)OR (95% CI)
Any SUD at follow-up 6.3 (1.8–21.4) 0.15 * (0.04–0.6) 5.5 (2.0–15.3)
 Alcohol A/D 5.8 (1.7–19.3) 0.16 (0.05–0.57) 4.9 (1.8–13.5)
 Marijuana A/D3.1 (0.8–12.5)0.42 (0.11–1.7) 5.4 (1.7–16.9)
 Hallucinogen A/D1.0 (0.1–9.3)0.76 (0.12–5.0) 9.0 (1.7–46.9)
 Cocaine/stimulant A/D7.5 (0.3–163.4)0.2 (0.02–2.1)3.0 (0.2–48.3)
Tobacco A/D0.85 (0.15–4.8)2.4 (0.5–12.3)4.4 (1.5–12.7)
  • * ORs <1 indicate a protective effect, i.e., that the odds of SUD were smaller in the medicated than in the unmedicated ADHD groups.

  • A/D, abuse or dependence.

    From separate logistic regression models predicting each SUD subtype as the dependent variable and the following baseline characteristics as independent variables: ADHD, Medication for ADHD, Conduct disorder, age, SES, any SUD at baseline, and parental history of SUD. For simplicity only the results from the terms of most interest (ADHD, Medication for ADHD, & CD) are presented.

    Underlining indicates P value <.01 according to Wald's χ2 using robust estimates of variance to account for familial associations between siblings.