The pharmacotherapeutic treatment of panic disorder can be hampered by several factors. These factors, which can be measured at baseline, can reduce the likelihood of successful treatment. A better knowledge of these factors can ensure that the optimum treatment is chosen. From the literature it is known that patients with very frequent panic attacks, with severe avoidance, with high scores on anxiety or depression scales and/or a comorbid Axis II-diagnosis have a worse prognosis. In part I of this review several studies are discussed which investigated possible predictors of short term treatment response. In part II predictors of long term treatment response are reviewed and critically discussed.