BACKGROUND: Loss or reduction of duodenal folds, scalloping of Kerkring folds and a micronodular or mosaic duodenal mucosal pattern have been described in celiac disease (CD), endoscopic findings that are considered reliable in the diagnosis of this disorder. However, most data have been obtained in patients with suspected or certain disease. We assessed the accuracy of the above markers in diagnosing CD in patients with nonulcer dyspepsia. METHODS: In this prospective study, in 705 consecutive dyspeptic patients (284 men, 421 women, mean age 51 +/- SD 15.8 years) duodenal biopsies were obtained only in the presence of typical endoscopic markers, whereas in another 517 (207 men, 310 women, mean age 49.9 +/- SD 16 years) duodenal biopsies were done irrespective of macroscopic findings. CD was diagnosed histologically and on the basis of positive antiendomysium antibody. RESULTS: Endoscopic markers were found in 4 patients of the first group but CD was ruled out. In the second group 5 patients had an endoscopic pattern that was consistent and CD was diagnosed in 3, whereas 3 others with normal endoscopic findings were eventually diagnosed as having CD. Endoscopic markers had a sensitivity of 50% and a specificity of 99.6% (95% CI [11.8, 88.2 and 98.6, 99.9], respectively) with positive and negative predictive values of 60% and 99.4%, respectively. CONCLUSION: The accuracy of endoscopic markers in the diagnosis of CD must be reevaluated in relation to the characteristics of the population studied.
BACKGROUND: Loss or reduction of duodenal folds, scalloping of Kerkring folds and a micronodular or mosaic duodenal mucosal pattern have been described in celiac disease (CD), endoscopic findings that are considered reliable in the diagnosis of this disorder. However, most data have been obtained in patients with suspected or certain disease. We assessed the accuracy of the above markers in diagnosing CD in patients with nonulcer dyspepsia. METHODS: In this prospective study, in 705 consecutive dyspeptic patients (284 men, 421 women, mean age 51 +/- SD 15.8 years) duodenal biopsies were obtained only in the presence of typical endoscopic markers, whereas in another 517 (207 men, 310 women, mean age 49.9 +/- SD 16 years) duodenal biopsies were done irrespective of macroscopic findings. CD was diagnosed histologically and on the basis of positive antiendomysium antibody. RESULTS: Endoscopic markers were found in 4 patients of the first group but CD was ruled out. In the second group 5 patients had an endoscopic pattern that was consistent and CD was diagnosed in 3, whereas 3 others with normal endoscopic findings were eventually diagnosed as having CD. Endoscopic markers had a sensitivity of 50% and a specificity of 99.6% (95% CI [11.8, 88.2 and 98.6, 99.9], respectively) with positive and negative predictive values of 60% and 99.4%, respectively. CONCLUSION: The accuracy of endoscopic markers in the diagnosis of CD must be reevaluated in relation to the characteristics of the population studied.