A 53-year-old man, who had been cured of colon carcinoma following a colonoscopic procedure in 1990, visited the Department of Gastroenterology, Osaka Medical Center for Cancer and Cardiovascular Diseases, in 1993 to remove a recurrent polyp. He had a pedunculated polyp in his ascending colon which was resected by endoscopic mucosal resection technique. The size of the head of the polyp was 14×8×7 mm. Histologically, it was a tubular adenomatous polyp with moderate atypia. After informed consent was obtained, the patient was enrolled into the Interventional Trial for Colorectal Cancer Prevention in Osaka.⁴ One of the baseline tests of the trial was an examination for faecal flora 1 month after resection of colorectal polyps. The patient collected his faeces using an airtight plastic bag with free-oxygen absorbing agent and brought it to our hospital, while it was kept cool with ice. After measuring the pH immediately by using a pH meter (Horiba pH meter F-8), the faeces was sent to Yakult Central Institute for Microbiological Research for culture within 36 h after collection. The flora analysis was done by the anaerobic roll tube method or aerobic agar plate method. In our case, the proportion of the number of the genus Streptococcus against whole faecal flora was so high that we identified this bacterium as S bovis using the API Strep identification system (API System SA, France). For reference, the composition and percentage (average and standard deviation) of faecal flora of 125 patients with two or more colorectal neoplasms 1 month after endoscopic resection is shown in table 1. The proportion of the genus Streptococcus in whole faecal flora was around 43% in this case, while the total percentage of other genera (excluding Bifidobacterium and Enterobacteriaceae) was <1%.

After obtaining the patient’s informed consent, administration of BLP (three times a day after each meal) was started with the aim of improving of his intestinal flora. Six months after administration of BLP, the percentage of Streptococcus was reduced from 43% to 9%; moreover, the lowered percentage of Streptococcus compared with baseline was maintained for 2 years after the administration of BLP was stopped (table 1, fig 1). Other genera, except Enterobacteriaceae and Bifidobacterium, did not show any specific changes (table 1). The proportion of Enterobacteriaceae was reduced by the administration; however, these bacteria rebounded up again after BLP was stopped. On the other hand, Bifidobacterium increased during the administration period, but dropped sharply after the treatment was stopped. Faecal pH was unrelated to the change in the faecal flora. Any side effects by BLP or any change of bowel movements were not recognised. Follow-up colonoscopy after 4 years revealed a pedunculated polyp in the transverse colon, 7 mm in diameter, which was shown to be a tubular adenoma with moderate atypia upon histological examination.

The first description of colorectal cancer occurring in association with S bovis bacteraemia was attributed to Mccoy and Mason in 1951. S bovis is known to be a causal bacterium of endocarditis and accounts for approximately 15% of bacterial endocarditis.¹ Many reports regarding an association between endocarditis or bacteraemia due to S bovis and colorectal cancer have been published.^1,2 The patients with endocarditis caused by S bovis more frequently suffered from colorectal cancer than those patients infected with other causal bacteria.¹ Burns et al⁵ reported that patients with colorectal cancer have S bovis in their faeces more frequently than healthy people. Klein et al² reported that the faecal carrier rate of S bovis was significantly higher in patients with colorectal cancer or inflammatory bowel disease than in controls, while an association between S bovis and the colorectal cancer could not be recognised by Dubrow et al.⁶