Abstract
This letter is written in response to a review recently published in the journal. The aim is to highlight a potential methodological limitation common to many studies comparing bipolar patients with few previous episodes versus those with multiple episodes, and in which the results are interpreted as indicating the longitudinal course of the illness.