background A recent study shows that the initial choice of an antidepressant matters. Although the prognosis of a depressive episode is often considered as good, 40-60% of the patients do not respond  to the first antidepressant, and remission is only reached in less than 35%. Therefore, clinicians often need to choose among one of the six possible next-step strategies: continuation for another 4-6  weeks, dose-escalation, switching, augmentation, combination or adding/switching to psychotherapy.
aim To provide an updated overview of the literature on the first choice and on next-step pharmacological strategies.
method The Cipriani meta-analysis about the choice of a first antidepressant will be discussed.  Results of literature searches on the next-step strategies will be presented with special attention to continuation for another 4-6 weeks, dose-escalation, switching, augmentation and combination.  Available meta-analyses will be used to summarize the current evidence, otherwise recent studies will be addressed.
results Some antidepressants appear preferable as first choice over others. As next-step 
strategy for non-responders after 4-6 weeks of treatment: 1. Continuation of treatment has been investigated poorly, but was as effective as a switch strategy. 2. Dose-escalation is a common  practice, the evidence, however, is poor. Two new dose-escalation studies show that for SSRIs and  duloxetine no dose-response-relationship exists in the normal clinical ranges. 3. For switching no  straightforward advise can be given for the choice of the next antidepressant. 4. Augmentation with  lithium, especially to tricyclic antidepressants (tcas)is very effective, and although the augmentation
with atypical antipsychotics has been investigated extensively, questions remain about their   appropriateness as a second-step strategy. Combination strategies are used more often after  sequenced treatment alternatives to relieve depression (star*d), but should not be a routine  second-step strategy.
conclusion s soon as one year after the release of the new Dutch multidisciplinary guideline for  depression, new evidence has become available that might influence the algorithm for pharmacotherapy provided by the guideline.