Background: Non-invasive testing forH. pylori (serology or radiolabelled¹⁴C urea breath testing) is increasingly used without upper gastrointestinal endoscopy in selected patients with new onset of dyspepsia (the ‘test and treat’ strategy).

Aims: To determine whether non-invasive testing was associated with delay in referral and diagnosis in patients with upper GI malignancy.

Patients and methods: Patients diagnosed with gastric or oesophageal malignancy from 1^stJanuary 1997- 31^st December 1999 were identified from histopathology records. Non-invasive testing for H. pylori within 2 years of diagnosis of malignancy was identified from case notes, general practitioners' records and microbiology records.

Results: 152 cases were identified, with mean age 70 years (range 34–91y, 4 patients younger than 45 years). 86 were male. 59 malignancies were oesophageal, 8 junctional, 77 gastric and one duodenal adencarcinoma; in 7 the precise anatomical origin was undetermined. 22 patients (14.5%) had been tested forH. pylori prior to diagnosis (tested group), all by serology (9 oesophageal, 2 junctional and 11 gastric carcinomas); 20/22 were older than 45 years. The median interval between H. pylori serology testing and referral for endoscopy was 21 days (range 1–615 days, mean 109 days). There was a delay of more than 12 weeks between serology and referral in 6 patients; 17/22 patients had locally advanced or metastatic disease at diagnosis. Of all patients, median delay between referral to the endoscopy service and diagnosis of malignancy was 9.1 days (range 1–181 days, mean 24.0 days; medians 12v 9.1 days in tested and untested groups respectively, p=NS). Of the patients aged 45 or older, 20/148 (13.3%) had H. pyloriserology.

Conclusion: Serological testing forH. pylori was used in a significant minority of patients aged 45 years or older who were subsequently diagnosed with upper GI malignancy, and in some …