The objective of this paper was to investigate the incidence of extra-intestinal infections caused by Clostridium difficile. During a 10-year period, the microbiology laboratory of our institution isolated 2034 isolates of C. difficile. Of the 2034 isolates, 21 (1.08%) were obtained from extra-intestinal sources. This represents an incidence of extra-intestinal isolation of four cases per 100 000 admissions. We were able to review the records of 17 patients for our study. The isolates in 12 patients were obtained from structures or fluids anatomically close to the colon and included the following infections: peritonitis in five cases (three primary and two secondary), intra-abdominal abscesses in three patients and abdominal wound infections in four cases. The infections in the other five patients were not in the anatomic vicinity of the colon. They included one case with a brain abscess, two episodes of bacteremia and two cases of foot infections (one chronic osteomyelitis). In all but one case, C. difficile isolation was obtained as part of a polymicrobial flora. The isolates were frequently non-toxigenic and the extra-intestinal infections occurred without concomitant diarrhea or prior anti-microbial therapy. Out of the 17 patients, eight died and nine survived. Death could not be directly attributed to C. difficile in any of the cases. The isolation of C. difficile outside the intestinal tract is very uncommon. Its clinical significance should be interpreted with caution.

The optimum treatment for prosthetic joint infections has not been clearly defined. We report our experience of the management of acute haematogenous prosthetic joint infection (AHPJI) in patients during a 3-year prospective study in nine Spanish hospitals. Fifty patients, of whom 30 (60%) were female, with a median age of 76 years, were diagnosed with AHPJI. The median infection-free period following joint replacement was 4.9 years. Symptoms were acute in all cases. A distant previous infection and/or bacteraemia were identified in 48%. The aetiology was as follows: Staphylococcus aureus, 19; Streptococcus spp., 14; Gram-negative bacilli, 12; anaerobes, two; and mixed infections, three. Thirty-four (68%) patients were treated with a conservative surgical approach (CSA) with implant retention, and 16 had prosthesis removal. At 2-year follow-up, 24 (48%) were cured, seven (14%) had relapsed, seven (14%) had died, five (10%) had persistent infection, five had re-infection, and two had an unknown evolution. Overall, the treatment failure rates were 57.8% in staphylococcal infections and 14.3% in streptococcal infections. There were no failures in patients with Gram-negative bacillary. By multivariate analysis, CSA was the only factor independently associated with treatment failure (OR 11.6; 95% CI 1.29-104.8). We were unable to identify any factors predicting treatment failure in CSA patients, although a Gram-negative bacillary aetiology was a protective factor. These data suggest that although conservative surgery was the only factor independently associated with treatment failure, it could be the first therapeutic choice for the management of Gram-negative bacillary and streptococcal AHPJI, and for some cases with acute S. aureus infections.