Aspiration is a common phenomenon in patients with oropharyngeal dysphagia. It can be studied using fiberoptic endoscopic evaluation of swallowing (FEES). FEES is well known and widely used in the diagnosis and treatment of swallowing disorders. However, various protocols exist, and there is no consensus on the examination protocol. The objective of this prospective study was to determine the FEES protocol derived estimates of sensitivity (Se') to detection of aspiration in dysphagic patients. The study estimated the probability of aspiration as a function of the number of swallow trials in dysphagic patients using FEES. The derived sensitivity was calculated based on presence or absence of aspiration in a ten-swallow trial protocol as arbitrary 'gold standard'. Eighty-four persons were included, comprising two patient populations with oropharyngeal dysphagia. Dysphagia in one group was due to head and neck cancer and possible oncological treatment effects on swallowing; in the other it was a result of neurological disease. All patients underwent a standardized FEES examination using ten swallows of thin liquid followed by ten swallows of thick liquid, all in boluses of 10 cc each. FEES recordings were rated for aspiration by an expert panel blinded to patients' identity and clinical history. Descriptive statistics, Kaplan-Meier survival analysis techniques, and Log Rank/Mantel-Cox tests were used. In both patient populations the aspiration risk was underestimated when using a limited number (three or four) of swallow trials. The oncology and neurology patients differed significantly in the number of swallow trials required to determine aspiration for thin liquids (median values 2 and 7 respectively, P = 0.006). FEES protocols using a limited number of swallow trials can underestimate the aspiration risk in both oncological and neurological patients suffering from oropharyngeal dysphagia, especially when using boluses with a thin liquid consistency.