BACKGROUND A secondary analysis was conducted to compare treatment outcomes for anxious depression and nonanxious depression in previous published OPERATION trials of a variety of antidepressants and augmentation strategies for patients with treatment-resistant depression (TRD). METHODS A total of 375 patients that met DSM-IV criteria for major depressive disorder (MDD) and the stage 2 TRD criteria (described by Thase & Rush) were enrolled. Anxious depression was defined as MDD with a HRSD-17 anxiety/somatization factor score ≥7. Data were derived from an earlier study, designed to compare efficacy and tolerability of fixed dosage of extended-release venlafaxine, mitazapine, paroxetine, and risperidone, sodium valproate, buspirone, trazodone or thyroid hormone augmenting to paroxetine in those patients. Treatment outcomes were compared between patients with anxious and nonanxious TRD. RESULTS Nearly 70% of participants had anxious depression. Remission rates were significantly lower and ratings of adverse event frequency were significantly greater in patients with anxious TRD than in those with nonanxious TRD. Presence of anxious depression predicted worse outcomes. CONCLUSIONS Lack of a placebo control arm prevents us from ruling out placebo effects. The two groups were non-randomly allocated to medications. Only patients with stage 2 TRD were enrolled, which may limit generalizablity to patients without a history of resistance. Comorbid anxiety disorders that might confound the specific treatment effects were not addressed. CONCLUSIONS The findings support and extend the hypothesis that anxious depression is associated with poorer outcomes. It suggests a dimensional assessment of co-occurring anxious features of MDD patients may be clinically feasible for countries like China where difficulties in making comorbidity diagnosis exist.