OBJECTIVE To study the effect of non-steroidal anti-inflammatory drugs (NSAIDs) on gastric cell turnover using an in vitro immunohistochemical method of bromodeoxyuridine (BrDU) uptake. METHODS Thirty patients undergoing routine upper gastrointestinal endoscopy were studied. Sixteen had taken NSAIDs daily for more than 3 months and there were 14 age-matched controls. Endoscopic gastric antral biopsies were obtained and stained immediately using the BrDU technique. Cell proliferation was expressed as a labelling index percentage (LI%) defined as the number of BrDU-labelled nuclei in 10 gastric glands, expressed as a percentage of the total cells in the gastric gland. RESULTS Gastric infection with Helicobacter pylori was excluded in all patients. Of the 16 patients on NSAIDs, four had gastritis, four had erosions or ulceration and eight had a normal examination. Endoscopy was normal in all patients in the control group. The LI% (mean +/- S.E.M.) in the entire NSAID group was 4.09 +/- 0.29 and in the control group 3.57 +/- 0.29. No significant difference was observed. In the NSAID patients with gastritis and erosions or ulceration, the LI% was 4.99 +/- 0.61 and 3.07 +/- 0.32, respectively. There was no significant difference in LI% between the endoscopic subgroups of patients on NSAIDs or between patients on NSAIDs who had normal endoscopy and the control patients. CONCLUSIONS These results provide evidence that refutes the hypothesis that the prevalence of NSAID gastropathy is due to an effect on gastric cell turnover.