In response to Dr Kumar's comments there are some important issues that need addressing. First, Dr Kumar introduces his letter in saying that we conclude that sertraline is not effective in preventing recurrent episodes of depression. This is not the case. We only stated that there is no evidence that sertraline has a prophylactic efficacy when used at the dose that achieved remission.

Second, we agree that if we had conducted an end-point analysis on the 27 subjects who completed the 100-week follow-up, then the study would have been relatively meaningless. We conducted a survival analysis on 113 subjects. This is a well-founded and recognised method of analysis of this type of study.

Third, we concur with Dr Kumar in that the prophylactic management of recurrent depression is critical. We do not advocate treating each episode as a new episode when prophylactic management is indicated. However, we do make the point that preventive techniques should be based on evidence of efficacy and effectiveness.

Dr Kumar suggests that we should have adopted a protocol that enabled increase in dose ‘when the clinical situation demanded it’; presumably when we thought a patient was experiencing the early stages of a recurrence. This misses the point of the paper. Our study (which is of a power similar to or greater than equivalent studies in this field) shows that the dose of sertraline required to achieve remission does not have prophylactic efficacy. This is important, as what evidence there is suggests that therapeutic doses of dothiepin (Old Age Depression Interest Group, 1993), nortriptyline (Reference Reynolds, Frank and PerelReynolds et al, 1999) and citalopram (Reference Klysner, Bent-Hansen and HansenKlysner et al, 2002) do have prophylactic efficacy. The implications for guidelines concerning the long-term management of older people with depression are self-evident.

Dr Kumar has failed to present arguments that undermine our conclusions. There is no evidence that the dose of sertraline required to achieve remission has prophylactic efficacy. The 8.4% reduction in risk of recurrence (over 100 weeks) that it offers is unlikely to instil clinical confidence in prophylactic efficacy when evidence indicates that other drugs for which the dose does not need to be changed are available.