The aim of the present study was to compare six media, three selective and three nonselective media, to determine the best combination of media for the primary isolation of Helicobacter pylori. Over a period of 8 months, mucosal antral biopsy specimens were obtained from 97 dyspeptic patients undergoing endoscopy. Biopsy samples were plated in parallel on all six media. Egg yolk emulsion agar (EYE), Skirrow's medium and modified Thayer-Martin medium were used as selective media; modified chocolate agar (MCHOC), Triptycase Soy Agar (TSA) and brain heart infusion agar were used as nonselective media. Overall, by using these six media, H. pylori were recovered from biopsy specimens from 48 of 97 patients, yielding an isolation rate of 49%. Comparison of all possible combinations of the six media showed that the highest rate of isolation of H. pylori was 100% (48 of 48) with EYE-MCHOC, followed by 97% (47 of 48) when EYE-SK was used. Conversely, it was found that none of the media used alone yielded a 100% rate of recovery (the maximum recovery rate was 92%, which was achieved with EYE). These results indicate that the association of EYE and MCHOC yielded the maximum recovery of H. pylori from gastric biopsy specimens. Therefore, the use of selective and nonselective media in parallel offers optimal recovery rates with only a slight increase in costs.