Bacteroides fragilis

Associated infections

• Intra-abdominal abscess and secondary peritonitis
• Bacteraemia
• Female genital tract infection and pelvic abscess
• Deep surgical site and wound infection
• Brain and liver abscess
• Diabetic foot infection and other necrotising soft tissue infection
• Perirectal and perianal abscess

Transmission route

B. fragilis infection is endogenous. It arises almost exclusively from the patient's own colonic flora entering a normally sterile site through perforation, surgery, penetrating trauma, bowel ischaemia, malignancy or another breach of the bowel wall. The determining factor is not exposure to an external source but loss of the anatomical barrier that normally keeps a dense anaerobic population confined to the lumen. This is why the risk profile of B. fragilis infection maps onto surgical and gastrointestinal events rather than onto environmental exposures.

Person-to-person transmission and waterborne transmission are not recognised routes. The organism cannot persist in an aerated environment for long enough to establish a transmission chain, and there is no vector or fomite pathway of practical significance in healthcare. Isolation of B. fragilis from a patient is interpreted as a marker of gut flora translocation, not as evidence of acquisition from another patient or from the facility.

The organism has no established role as a coloniser of water systems used in reprocessing. It is oxygen-sensitive, requires a reducing and nutrient-rich environment, and does not form the aerobic biofilm relevant to plumbing and reprocessor water paths. It is not among the indicator organisms used for routine surveillance of final rinse water in any current guidance, and would not be recovered by the aerobic culture methods those surveillance programmes normally employ. Where B. fragilis appears in a reprocessing context at all, the transmission question to ask is not how it travelled through a water system but how faecal material reached the sample.

The practical route by which B. fragilis can enter a reprocessing setting is therefore transfer of patient-derived intestinal soil. A colonoscope or gastroscope carries substantial faecal or luminal material at the point of use, and inadequate bedside pre-cleaning followed by inadequate manual cleaning can leave that soil in a channel, at a valve seat, in an elevator mechanism or in a damaged or worn surface. Alternatively the organism may be introduced at the point of sampling, through contaminated sampling equipment, contact with a soiled surface, or handling of the sample container by someone who has been working on the dirty side.

Relevance in endoscopy and reprocessing

B. fragilis has no documented role in endoscope-associated water transmission, no recognised pseudo-outbreak history in reprocessing surveillance, and no biofilm behaviour relevant to rinse water plumbing or CSD and dental water lines. Stating that plainly is more useful than manufacturing a relevance the organism does not have. It is not on any surveillance indicator list for final rinse water, it will not grow under the aerobic conditions used for routine water testing, and it is not part of the microbial ecology that reprocessing water controls are designed to manage.

What a detection does indicate is gross contamination with intestinal content, and that is a serious finding in its own right. B. fragilis is a marker of colonic soil in a location where colonic soil should not be present. In an endoscopy service, the most plausible explanation for a confirmed isolate from a reprocessed device is that patient-derived material survived the cleaning stage. This is significant because cleaning is the step on which everything downstream depends: high-level disinfection is validated against a cleaned surface, and organic soil both physically shields organisms and chemically consumes the disinfectant. An organism that cannot live in water but that has appeared on a device that has completed a full reprocessing cycle points at a cleaning failure with considerable specificity.

The conditions that permit this are known and checkable. They include omitted or delayed bedside pre-cleaning, allowing bioburden to dry inside channels; inadequate or omitted channel brushing, or use of brushes that are worn, wrong-sized or not passed the full length; incomplete disassembly of valves, caps and elevator mechanisms; retained soil in a damaged, scored or delaminated channel that cannot be cleaned to a validated standard; and overloaded or interrupted workflow that compresses cleaning time. Because B. fragilis is oxygen-sensitive and fastidious, its recovery also implies a substantial inoculum was present, since small numbers would be unlikely to survive handling and transport to the laboratory. A confirmed isolate should therefore be treated as evidence of a significant residual bioburden event, with quarantine of the implicated device and inspection of its channels, rather than as a low-grade quality deviation.

Interpreting a detection

A detection of B. fragilis in a water or final rinse sample is not a plausible water-system finding. There is no mechanism by which an obligate anaerobe of the human colon establishes itself in an aerated, treated, low-nutrient water distribution system, and no remediation of the water treatment train would address the finding. The interpretive question is therefore not whether the water is colonised but how intestinal material entered the sample or the device.

The first checks are of sampling and laboratory process, followed by cleaning practice. Confirm the identification, since anaerobic Gram-negative bacilli can be misassigned and an unexpected organism warrants confirmation before an investigation is launched on it. Establish what culture conditions were used, since recovery of a strict anaerobe from a routine aerobic water surveillance plate would itself be anomalous and suggests either that anaerobic culture was specifically requested or that the report is in error. Review the chain of custody for the sample: who collected it, whether gloves were changed, whether the sample was taken by someone who had been handling contaminated devices, whether the sampling port and connectors were disinfected, whether the container was sterile and opened only at the point of collection, and whether transport and processing were prompt. Determine whether the laboratory processed any faecal or clinical anaerobic specimens in the same batch or on the same bench, since cross-contamination in the laboratory is a real alternative explanation.

If the identification is confirmed and sampling artefact can be reasonably excluded, the finding escalates immediately. This is not an organism where a single isolate is watched and a trend is awaited. A confirmed B. fragilis isolate from a reprocessed endoscope implies residual faecal soil on a device that was presented as ready for patient use, which is a patient safety issue rather than a water quality trend. The appropriate response is quarantine of the affected device pending resolution, review of the manual cleaning process including brushing technique, brush condition, disassembly and cleaning chemistry contact time, borescope or equivalent internal inspection of the device channels for damage or retained debris, and review of the workflow pressures that may have compressed the cleaning stage. Infection prevention should be notified, and the unit should consider whether other devices processed in the same session or by the same method require assessment. Where the isolate cannot be confirmed or the sampling chain was demonstrably compromised, the finding should be documented as a probable artefact, the sampling technique corrected, and the sample repeated under controlled conditions, but the cleaning review is still worth completing given the consequence if the first interpretation is wrong.

Antimicrobial resistance

The Bacteroides fragilis group is among the most antimicrobial-resistant of the anaerobic genera, and resistance rates have risen consistently over successive surveillance periods. More than ninety per cent of isolates are resistant to penicillin through production of beta-lactamases, principally cephalosporinases, which also compromises many cephalosporins. Beta-lactamase inhibitor combinations retain activity against most isolates but not all. Resistance to clindamycin has increased to the point that it is no longer regarded as reliable empirical therapy for anaerobic infection, and resistance to moxifloxacin and to the tetracyclines is common.

Carbapenem resistance is the development of greatest concern. It is mediated principally by the cfiA gene, which encodes a metallo-beta-lactamase, and is found in a small but significant minority of isolates. Expression of cfiA depends on the presence of an upstream insertion sequence, so a strain can carry the gene while testing susceptible and then become resistant on therapy through insertion sequence activation, which makes phenotypic testing alone an incomplete guide. Metronidazole resistance, mediated by nim genes, remains uncommon but is documented and has been reported in association with treatment failure.

Because susceptibility is not reliably predictable from species identification alone, routine susceptibility testing of clinically significant B. fragilis group isolates is recommended, and local surveillance data should inform empirical regimens. None of this alters reprocessing requirements. Anaerobic resistance mechanisms confer no tolerance of validated high-level chemical disinfection or of thermal processes, and B. fragilis is in fact among the less robust organisms encountered in a reprocessing setting given its oxygen sensitivity and desiccation intolerance. Its significance is entirely as a marker of residual faecal soil, and the corrective action is better cleaning rather than stronger disinfection.