Findings focus on depressed teens.

PHOENIX -- The only prospective, randomized, controlled comparison of treatment for resistant depression in adolescents is yielding new insights about efficacy and suicidality through several post hoc analyses.

The selective serotonin reuptake inhibitor fluoxetine is the only medication approved to treat severe depression in adolescents. Until publication of the Treatment of SSRI-Resistant Depression in Adolescents (TORDIA) trial (JAMA 2008;299:901-13), there were no data to guide clinicians on treating the roughly 40% of adolescents with major depressive disorders who do not respond to initial SSRI therapy.

The TORDIA trial randomized 334 patients--aged 12-18 years and not responsive to 2 months of initial treatment with an SSRI--to 12 weeks of treatment with a different SSRI, a different SSRI plus cognitive-behavioral therapy (CBT), venlafaxine (Effexor), or venlafaxine plus CBT. Switching medication and adding CBT produced a significantly higher rate of response (55%), compared with a medication switch alone (41%). Efficacy was similar whether switching to an SSRI (paroxetine [Paxil], citalopram [Celexa], or fluoxetine [Prozac]) or to venlafaxine, but venlafaxine caused more skin problems and increases in pulse and diastolic blood pressure.

In a 12-week extension of the TORDIA trial, providers could switch medications, augment with drug therapy or CBT, or use all of those approaches. Response rates increased by week 24, mainly because of the effects of medication, and some patients relapsed, leaving an overall 13%-15% increase in response rates, Dr. Graham J. Emslie reported at a meeting of the New Clinical Drug Evaluation Unit sponsored by the National Institute of Mental Health.

Other investigators in the trial reported during the same session on separate post hoc analyses suggesting that venlafaxine and benzodiazepines might increase the risk of suicidal events and that more severely depressed adolescents benefit from CBT (contrary to findings in studies of CBT as initial therapy).

They cautioned that post hoc analyses should not be the basis for changes in clinical practice, but the findings do raise important questions for further study.

"The goal of treating depression is remission of symptoms, not just response" to therapy, noted Dr. Emslie, professor and chief of psychiatry at the University of Texas Southwestern Medical Center at Dallas. Remission rates did not differ significantly between groups, but increased with medication use by week 24.

Remission rates increased from 28% with an SSRI or 29% with venlafaxine at week 12 to 39% and 37%, respectively, by week 24. Remission rates in patients who got CBT plus medication were 31% at 12 weeks and 34% at 24 weeks, and patients with medication alone (no CBT) were 26% at 12 weeks and 41% at 24 weeks, Dr. Emslie and his associates reported.

Patients whose symptoms ultimately remitted showed greater improvement than did nonremitters as early as week 6, a separation that was maintained over time.

In 175 patients who did not respond to treatment by week 12, further treatment produced remission in 18% by week 24 and a therapeutic response but not remission in 11%. "These people had failed two adequate trials of two antidepressants, and if they got additional therapy after week 12, they showed further improvement," said Dr. Emslie, who received research funds from or is a consultant or speaker for Eli Lilly & Co., GlaxoSmithKline Inc., Wyeth-Ayerst Laboratories, Shire PLC, Biobehavioral Diagnostics Co., McNeil Consumer & Specialty Pharmaceuticals, Organon USA Inc., Somerset Pharmaceuticals Inc., and Forest Laboratories Inc., some of which make medications used in the study.

The TORDIA trial was interrupted for 6-12 months at its six sites when the Food and Drug Administration issued a black box warning about increased suicide risk in children and adolescents taking antidepressants, based on spontaneously reported suicides or attempted suicides. When the trial resumed, researchers monitored patients for suicidal ideation and events, and compared data on the 153 patients who were monitored weekly with the previous 181 patients whose suicidal events were reported spontaneously.

"What we found is a little bit surprising," said Dr. David A. Brent, the lead investigator in the TORDIA study and chief of child and adolescent psychiatry at the University of Pittsburgh Medical Center. He has no potential conflicts of interest related to the TORDIA trial.

There were 48 suicide-related events in the 12-week trial, with no difference between the spontaneously reported or monitored groups in the frequency of serious events but an increase in milder events (ideation rather than action) in the monitored group.

"That subgroup is something we weren't picking up in the first half, and I'm not sure how clinically significant they are," Dr. Brent said. The spontaneously reported events seemed to be representative of the more serious events, lending credence to the FDA analyses.

Adolescents with higher self-reported stress or high suicidal ideation were more likely to have suicidal events if treated with venlafaxine. "Things are not looking good for venlafaxine," he added.

CBT was somewhat protective in patients without non-suicidal self-injurious behavior such as cutting, but not in patients with those behaviors, perhaps because the CBT in this study did not specifically target hopelessness, he said.

Ten children who were treated with benzodiazepines were at more than five times greater risk for a suicide event, even after controlling for disease severity, compared with those not on benzodiazepines.

"I'd certainly avoid using venlafaxine in kids who were highly suicidal unless they're allergic to everything else," Dr. Brent said. "I think we need to more carefully evaluate our use of benzodiazepines."

An analysis of medication exposure in the TORDIA trial suggested that patients in the middle ranges of plasma levels had the greatest clinical response, Dr. Karen D. Wagner said. This suggests that physicians might use blood levels to help manage dosing.

Patients started on doses of 20 mg of an SSRI or 150 mg of venlafaxine, and after 6 weeks, could increase doses as high as 40 mg of an SSRI or 225 mg of venlafaxine. Plasma levels were not normally distributed, and so were described as percentiles. Patients with levels in the 25th-75th percentiles were more likely to respond to treatment and more likely to develop side effects than were patients with higher or lower levels.

Dr. Wagner received research funds from or is a consultant or adviser to Abbott Laboratories, AstraZeneca Pharmaceuticals LP, Bristol-Myers Squibb Co., Cyberonics Inc., Eli Lilly, Forest Laboratories, GlaxoSmithKline, Janssen LP, Jazz Pharmaceuticals Inc., Novartis Pharmaceuticals Corp., Ortho-McNeil Inc., Pfizer Inc., Organon, Otsuka Pharmaceutical Co., Solvay Pharmaceuticals Inc., UCB Pharma, Wyeth-Ayerst, and Johnson & Johnson, some of which make medications used in the study.

BY SHERRY BOSCHERT

San Francisco Bureau