Background: Clostridium difficile (CD) is a major cause of hospital-associated diarrhea, and both disease severity and prevalence have increased over the past decade. Our laboratory has used a 2-step algorithm for CD laboratory testing since 2008. In 2011, we replaced the toxin A/B enzyme immunoassay (EIA) confirmatory step-2 part of the procedure with the illumigene molecular assay. This retrospective study evaluated the clinical impact this change had on physicians’ decisions to not start or to stop empiric CD treatment based on negative test results.
Methods: Medical records on inpatients from June to December 2010 with glutamate dehydrogenase (GDH) negative test results for Clostridium difficile were first reviewed for physicians’ actions. Additional records from inpatients with GDH-positive/toxin A/B EIA–negative specimens from this same time period and the same number of inpatients with GDH-positive/illumigene-negative specimens from June to December 2011 were also reviewed and the treatment decisions compared.
Results: Of 145 patients with GDH-negative test results, 74.5% were never treated for C difficile infection, 20.7% had treatment stopped within 3 days of the test result, and only 4.8% were continued on treatment. Of 60 patients with GDH-positive/toxin A/B EIA–negative test results, 26.7% were never treated, 36.7% had treatment stopped within 3 days, but 36.7% were continued on treatment for more than 3 days. The toxin A/B portion of the 2-step algorithm was changed in early 2011 to the illumigene C difficile molecular assay. Of 60 patients with GDH-positive/illumigene-negative test results, 36.7% were never treated, 48.3% had treatment stopped within 3 days, and only 15% of patients were continued on treatment.
Conclusions: Results show that GDH-negative results influence physicians not to treat or to stop empiric treatment for C difficile infection. In addition, more patients were continued on treatment (>3 days) with a toxin A/B EIA–negative result and then with an illumigene-negative result (36.7% vs. 15% respectively, P = 0.007).