Epidemiologic data on inclusion body myositis (IBM) are scarce, and possibly biased, because they are derived from larger neuromuscular centers. The present nationwide collaborative cross-sectional study, which culminated on July 1, 1999, resulted in identification of 76 patients with IBM and the establishment of a prevalence of 4.9 patients with IBM per million inhabitants in the Netherlands. Several discrepancies suggest that this may be an underestimation. The most frequently identified pitfall in diagnosing IBM was an erroneous diagnosis of polymyositis or motor neuron disease.