Résumé : Background: Late-onset glycogen storage disease type II is associated with variable muscle phenotypes. Epidemiological data suggest that its prevalence is lower in Belgium than in bordering countries like The Netherlands. Objective: We investigated whether such low estimated prevalence is due to missed diagnoses. Methods: We screened 100 patients with muscle phenotypes of undetermined origin using a dried blood spot test for alpha-acid glucosidase (GAA) activity. Patients with low activity at screening were re-tested by the same method and, if low activity was confirmed, GAA gene analysis was performed. Results: The screening test revealed lower than normal GAA activity in 15 patients, but in only two of them it was low enough to be considered in the disease range. Retesting confirmed lower than normal GAA activity in five patients, but in all of them it was above the disease range. A single patient carried a heterozygous known pathogenic GAA mutation, whose significance in this case remains undetermined. Conclusions: We conclude that reported low prevalence estimates in Belgium are not likely to be due to an underdiagnosis bias. Lower prevalence compared to neighbouring The Netherlands may be due to different ethnic stratification of our patients. Diagnostic strategies should keep into account the expected prevalence of a disease in specific populations.