Touch cytology in diagnosing Helicobacter pylori: comparison of four staining methods

Abstract

Helicobacter pylori (Hp), a major cause of peptic ulcer disease and an important risk factor for gastric malignancy, can be diagnosed by several methods. Touch cytology (TC) of the gastric mucosa has been noted to give good results and has been found to be very simple, inexpensive and rapid. However, evidence regarding the accuracy of different staining methods of TC is lacking. The present study aims at defining the diagnostic accuracy of four different staining methods of TC. Biopsy specimens were taken from the antral mucosa of one hundred consecutive patients referred for upper gastrointestinal endoscopy (UGIE) for various indications. TC slides were processed by four staining methods: Wright, Giemsa, Papanicolaou and Gram. Rapid urease test (RUT) and histological examination of specimens were also performed. The same experienced pathologist evaluated the coded samples. A patient's Hp status was established by minimum concordance of the three tests, including histology, RUT, and 'Touch mean'. The latter was defined positive when at least three of the four TC staining methods were positive. Forty-six patients (46%) were positive for Hp according to Hp status. TC stained by Wright had excellent agreement with both histology (kappa = 0.80, P < 0.001) and RUT (kappa = 0.84, P < 0.001). Regarding Hp status, histology was 100% sensitive and RUT was 100% specific. Wright-stained TC (88.89%) was significantly more specific than both Giemsa- (74.07%; P < 0.05) and Papanicolaou-stained (70.37%; P < 0.05) TC. RUT should still be acknowledged as the primary test in diagnosing Hp following UGIE. If RUT is negative and Hp detection is intended only, Wright-stained TC can safely substitute for histology. However, when assessment for severity of mucosal damage or cell atypias is meant, histology cannot be neglected.