Severe major depressive disorder (MDD) is associated with poorer response to treatment than milder forms of depression, but there is a lack of consensus about the optimal treatment approach. We conducted a qualitative review of the literature for clinical trials of severe MDD in adults (1975 - December 2008), to assess the effects of five treatment approaches: placebo, medication, psychotherapy, combined treatment, and medication augmentation. We found significant variability in treatment outcomes across studies of severe MDD, deriving largely from differences in study design and variability in study samples identified as “severely” depressed. Severely ill patients respond more poorly to placebo and psychotherapy than mildly or moderately depressed patients. Medication monotherapy is superior to psychotherapy alone in the treatment of severe MDD. Combination treatment with medication and psychotherapy shows greater utility in severe, rather than mild or moderate, MDD. Lithium is the best established augmenting agent for severe MDD, particularly in combination with a tricyclic antidepressant. This review indicates the best initial clinical approach to treating severe major depressive disorder is through a combination of medication monotherapy and structured psychotherapy. Future trials in severely ill patients should more fully characterize the clinical and biological aspects of the enrolled patients.