We studied 2253 consecutive dyspeptic patients, without clinical evidence of organic disease, who were referred to our open access endoscopy service. The aim was to assess whether the various clinical patterns of dyspepsia can be considered a valid guideline for the appropriate use of endoscopy. According to the symptomatological patterns, our patients were defined as sufferers from 1) ulcer-like (973 patients), 2) reflux-like (857), and 3) dysmotility-like dyspepsia (423). In our patient population, which reflects the general population of our city, the dysmotility-like type of dyspepsia was the least frequent (19%), whereas the ulcer-like (43%) and the reflux-like (38%) dyspepsia were almost equivalent. A negative endoscopy (35.7%) occurred significantly (p < 5 x 10(-4)) more often in dysmotility-like than in ulcer-like (26.3%) and reflux-like dyspepsia (25.7%). Furthermore, in dysmotility-like dyspepsia, we observed no malignancies in patients less than 60 yr old, and no gastric ulcers in patients less than 50 yr old. In the latter subgroup of patients (under 50 yr), duodenal ulcers and esophagitis were rare (occurring in only one and five, respectively, out of 145 patients). In ulcer-like and reflux-like dyspepsia, abnormal endoscopic findings occurred frequently (in 73.5% and 74.1%, respectively), and no relationship with patients' age was observed. Our data indicate that patients under 50 yr old with dysmotility-like dyspepsia can be considered a kind of population for which endoscopy is inappropriate. However, because the prevalence of dysmotility-like dyspepsia was 19% (423/2253) in our patient sample, and only 7.15% of them were under 50 yr old (161/2253), we can obtain only a small percentage of reduction in endoscopic service load if the guideline of age < 50 yr is adopted.