Clostridioides difficile infection (CDI) rates in Department of Veterans Affairs (VA) acute-care facilities have decreased nationwide by 37% in the 7 years since the implementation of a VA CDI control guideline in 2012. ^{Reference Roselle, Evans, Simbartl, McCauley, Lipscomb and Clifton1} For much of that time, the diagnosis was made using a compatible clinical illness and a positive nucleic acid amplification test (NAAT). This test has a sensitivity of ˜96%, but because it detects bacterial DNA, it may be positive in stool containing toxin-producing bacteria, bacteria carrying the nonexpressed toxin gene, or DNA left behind by dead bacteria. These factors led some investigators to suggest that the test was too sensitive and resulted in unnecessary isolation and treatment of patients whose true cause of diarrhea was something else. ^{Reference Polage, Gyorke and Kennedy2} For this reason, some have advocated the use of an enzyme immunoassay (EIA), which detects C. difficile toxin in the stool rather than detecting the genes responsible for toxin production. Unfortunately, commercially available toxin EIAs have a sensitivity in the range of 60%–70%, meaning that some patients with CDI may be missed if this test is used alone. One solution is to use an NAAT, and if positive, to perform a follow-up toxin EIA. The National Healthcare Safety Network of the Centers for Disease Control and Prevention adopted this 2-step process and recommended that the results of the final test (usually the toxin EIA) be reported for surveillance purposes. Thus, NAAT-positive/toxin EIA–negative patients would be diagnosed negative for CDI and would not be reported in a national surveillance database.

Others have reported, however, that clinicians ultimately treat between 22% to 77% of NAAT positive/toxin EIA negative patients with oral drugs such as vancomycin, fidaxomicin, or metronidazole, which are indicated for CDI. ^{Reference Hitchcock, Holubar, Hogan, Tompkins and Banaei3–Reference Lenggenhager, Zanella, Poncet, Kaiser and Schrenzel7} This finding suggests that CDI cases might be underreported. These publications come from single centers where patient management and antibiotic prescribing may reflect a “community standard of practice.” We wished to assess how common this practice is within the national VA healthcare system. A discrepancy between laboratory results and patient management may be germane to CDI rates reported to national surveillance programs.

Results

All 134 acute-care VA medical centers were included in the database, and 34 (26.6%) did 2-step testing for CDI during the study period. More than 1.25 million oral prescriptions were written for inpatients and outpatients during the study period including 25,393 prescriptions for oral vancomycin, metronidazole, or fidaxomicin. In total, 88,922 NAATs were performed for unique patients for the diagnosis of CDI and 14,414 (16.2%) of these were positive (Fig. 1). Also, 11,233 toxin EIAs were performed for unique patients, and 10,018 (89.2%) of these were negative. In the 34 facilities that did 2-step testing, a positive NAAT result was followed by a negative toxin EIA result within 36 hours for 2,374 unique patients. Of these, 702 (29.5%) were prescribed one of the oral drugs indicated for the treatment of CDI within 72 hours of the negative toxin EIA result. Vancomycin was prescribed 92.9% of the time, metronidazole was prescribed 6.7% of the time, and fidaxomicin was prescribed 0.4% of the time. This pattern occurred in 25 (74%) of the 34 facilities doing 2-step testing and occurred in facilities widely distributed across the United States.

Fig. 1.

Clostridioides difficile nucleic acid amplification tests (NAAT) and toxin enzyme immunoassays (EIA) for unique patients performed from October 2018 through March 2021 in VA medical facilities.