About two-thirds of depressed patients respond to a standard course of serotonin specific reuptake inhibitor (SSRI's) within 3-4 weeks. While some clinicians advise continued watchful waiting after this time or switch to a different reuptake-blocker based antidepressant, result of such conservative strategies are usually disappointing. For severe depression electroconvulsive therapy (ECT) is an option and for atypical depressions monoamine oxide inhibitors (MAO) inhibitors often give relief at this point. A unique strategy with both theoretical and practical implications is lithium augmentation (Fava et al, 1994). Addition of lithium to SSRI failures at 3-4 weeks is consistently and sometimes dramatically found to be helpful. This is considered true even by those authors who advocate use of lithium under usual circumstances only in bipolar patients.

Lithium in recent years has been joined as a mood stabilizer by carbamazepine and valproate. Phenytoin, ignored for many years as a possible anticonvulsant mood stabilizer, has been recently reported in double-blind controlled trials to be anti-manic (Mishory et al, 2000) and also prophylactic in BP disorder (Mishory et al, 2003).

Data on mood stabilizers other than lithium as augmentors in SSRI failures are sparse. Carbamazepine (Steinacher et al, 2002) and valproate (Barbee et al, 2002) have been used. Given our recent preliminary results of phenytoin's efficacy in unipolar depression (Nemets et al, 2005) and its analogy to lithium as a mood stabilizer, it seems important to study phenytoin as a possible augmentation of SSRI failures.

We have published a negative study previously of inositol as an augmentation of SSRI failures, enrolling forty-two patients over two years (Nemets et al, 1999). Antidepressant failures are easier to recruit from referring physicians in our center than are untreated patients, whom clinicians are reluctant to refer for new drug studies given the adequacy of standard treatment in 2/3 of them. Thus we estimate that we could enroll 20 patients per year in such a study. Survey of the literature of Li augmentation suggests that 40 phenytoin vs. 40 placebo should give adequate power to detect a significant phenytoin effect if the phenytoin effect is similar to that of lithium.