Line of Service-Specific Performance and Antibiotic Prescribing Habits Following Introduction of a Two-Step Diagnostic Approach Using NAAT Followed by Enzyme Immunoassay in Cancer Patients with Suspected Clostridium difficile Infection

Abstract Background Patients with cancer are at an increased risk for C. difficile infection (CDI). A two-step approach with a Nucleic Acid Amplification Test (NAAT) followed by enzyme immunoassay (EIA) increases diagnostic sensitivity and specificity and can be used to guide antibiotic therapy. We retrospectively investigated the relative performance of the two-step approach in cancer patients with solid tumors (ST), hematologic malignancies (HS), and hematopoietic stem cell transplant recipients (HSCT). Methods We identified 204 patients with a positive NAAT test for CDI as determined by GI multiplex (Biofire) or by Illumigene (Meridian, Bioscience) in whom a reflex EIA was performed for C. difficile A/B toxins between November 2015 and February 2017. Patients were stratified into ST, HM, HSCT groups. We compared the proportion of discordant NAAT+, EIA- results among the three groups. We then compared the clinical presentation and antibiotic use for patients in the NAAT+/EIA- to those with NAAT+/EIA+ results. Results Overall an EIA+ result was found in 53 (26%) patients. The proportion of patients with NAAT+/EIA+ results was significantly different between the three lines of service; ST 31/86 (36%), HM 16/62 (26%), and HSCT 6/56 (11%) P < 0.01. A trend towards a higher proportion of positive results was observed for ST compared with the HM group (P = 0.06). Results were similar between the HM and HSCT group. However, patients in the ST were more likely to have a positive EIA when compared with HSCT patients (36% vs. 11% P < 0.01). Clinical presentation and healthcare-association were similar in all three groups regardless of the EIA result. Despite the low proportion of EIA+ confirmatory results, the majority of patients (196/204 96%) received antibiotic therapy targeting CDI. Discontinuation of CDI antibiotics prior to 10 days of therapy was similar in the EIA+ (12%) vs. EIA- (10%). Conclusion The relevance of discordant results needs to be interpreted in the context of the line of service/patient care unit. The presence of CDI as determined by NAAT/EIA is low in patients with other potential causes of diarrhea such as in HSCT recipients. A substantial proportion of cancer patients are treated unnecessarily for CDI. Disclosures All authors: No reported disclosures.