with a tip of the hat to Tim, our official research educator to Bipolar at OBT

Substance abuse increases antidepressant-induced mania risk

Researchers have discovered that a lifetime substance use disorder can increase the chance of antidepressant-induced mania in people with bipolar disorder by five fold.

"This finding suggests that randomized controlled trials of antidepressant use in bipolar disorder could systematically underreport antidepressant-induced mania due to exclusion of patients with substance use disorders," note Sumita Manwani (Cambridge Health Alliance, Massachusetts, USA) and colleagues.

To determine whether substance abuse is a risk factor for antidepressant-induced mania, the researchers carried out an observational study in which they identified 98 patients treated for bipolar disorder between 2000 and 2004 from 335 antidepressant trials.

They reviewed the patients' medical charts and histories of substance use disorder according to DSM-IV criteria and antidepressant-induced mania.

In all, 55 patients from 184 trials had substance use disorders, with current substance abuse reported for 42%. Alcohol was most frequently abused, followed by marijuana, cocaine, amphetamines, and opiates or sedatives.

In unadjusted univariate analysis, no relationship was found between substance use disorders and antidepressant-induced mania, with such mania occurring in 20.7% of substance use disorder trials and in 21.4% of non-substance use disorder trials.

After multivariable analysis that took into account other confounding variables, however, substance use disorder was found to increase antidepressant-induced mania, with an odds ratio (OR) of 5.06.

Other risk factors included having bipolar II disorder compared with bipolar I disorder (OR=3.75), being female compared with being male (OR=4.25), and taking tricyclic or serotonin reuptake inhibitors compared with bupropion (OR= 5.12 and 2.85, respectively).

Thus, the association between substance use disorders and antidepressant-induced mania only became apparent after adjustment for potential confounders.

"This observation of 'negative confounding' when confounding factors obscure a real relationship so that univariate analyses are negative, may help explain why previous studies, which only used univariate analyses, reported conflicting results," the investigators write in the Journal of Clinical Psychiatry.

While noting that their study did not take into account the role of rapid cycling, they conclude that "a history of substance abuse was a likely predictor for antidepressant-induced mania."


J Clin Psychiatry 2006; 67: 1341-1345