Diagnostic and Therapeutic Considerations for Oncology Patients With Clostridium difficile Infection

Diagnosis and Classification Despite the impressive burden ofC difficile infection in the United States, it has become clear that there are many subtle complexities in diagnosis and classification of C difficile infection, which is also the case in oncology patients. Although toxigenic culture for C difficile is the gold-standard diagnostic technique, this method is infrequently performed because of its complex laboratory protocol requiring experienced technologists, delayed turnaround time for results, and limited scalability.Nucleic acid amplification tests are highly sensitive but have correspondingly high rates of falsepositive results. Even a test with 99% sensitivity and 99% specificity (or combination test algorithms with similar performance characteristics) will produce falsepositive results. As Neeman and Freifeld point out, overdiagnosis of C difficile infection in these patients puts them at risk for overtreatment. Overtreatment, in turn, worsens the risk of intestinal dysbiosis, defined as abnormal intestinal microbiota composition or lack of diversity. Oncology patients are already at high risk of developing intestinal dysbiosis through frequent and prolonged health care exposures, and providers should be mindful of other factors that may exacerbate this risk. As examples, in amousemodel, metronidazole was shown to have prolonged detrimental changes in gut microbiota composition, and in an industry-supported study, use of enteric vancomycin was shown to increase risk of colonization with Candida and vancomycin-resistant Enterococcus. Additional work is needed to better understand ideal tests and frequency of C difficile testing in oncology patients. There are multiple scoring systems for severity of C difficile infection, but the importance of classification of infection severity should depend on the implications for clinical care. It is not clear that such classification improves outcomes if therapies such as oral vancomycin or fecal microbiota transplantation are increasingly considered as first-line treatments. The potential to introduce collider bias is another challenge of such scoring systems when they include multiple factors related to C difficile infection that could be either causes or effects (eg, intensive care unit admission or hospitalization) rather than causal factors alone. Collider bias is the creation of an artificial statistical relationship between two factors when conditioning on common effects of a disease rather