Real-world evaluation of a two-step testing algorithm for Clostridioides difficile infection

Abstract

During the 4 years after implementation of the 2-step Clostridioides difficile infection (CDI) testing algorithm, 70% to 78% of patients with suspected CDI and a positive nucleic acid amplification test but a negative toxin test (NAAT+/TOX−) received CDI treatment. Overall, 73% of NAAT+/TOX− patients were classified as having probable or possible CDI.

Nucleic acid amplification tests (NAATs) have excellent sensitivity for the diagnosis of Clostridioides difficile infection (CDI). ¹ However, asymptomatic carriers of toxigenic C. difficile with unformed stool due to other causes (eg, laxatives) may test positive by NAAT, resulting in unnecessary treatment and inflation of CDI rates. ^1–3 One strategy to address this issue has been to use 2-step test algorithms in which results of stool toxin testing and clinical assessments are used to guide management for patients with positive initial screening assays. ^1–3 With this approach, a negative toxin assay suggests asymptomatic carriage of a toxin-producing strain of C. difficile or mild CDI that may resolve without treatment. ^1,3,4 In comparison to patients with NAAT-positive/toxin-positive (NAAT+/TOX+) results, patients with NAAT-positive/toxin-negative (NAAT+/TOX−) results have reduced risk for mortality, severe illness, and recurrence. ^1,3,4 However, clinical assessment remains essential because some NAAT+/TOX− patients may have severe, complicated infections. ^5,6 Although 2-step CDI test algorithms are increasingly used, limited information is available on their performance in real-world settings. Whether clinicians will defer treatment of NAAT+/TOX− patients who are asymptomatic or have mild symptoms remains unknown.

Methods

Setting

In June 2018, a 650-bed academic medical center switched from a standalone NAAT test for CDI to a 2-step algorithm with an initial NAAT with toxin testing of NAAT-positive samples. Toxin testing was performed using the TOX A/B QUIK CHEK enzyme immunoassay for toxins A and B (Alere). NAAT and toxin results were reported simultaneously. NAAT+/TOX− test results included the comment “Possible C. difficile infection or carriage of (toxigenic) C. difficile” (Supplementary Fig. 1 online). The laboratory rejected formed stool specimens.

Fig. 1. Percentage of patients with suspected Clostridioides difficile infection (CDI) with nucleic acid–positive and enzyme immunoassay for toxin-negative test results receiving treatment for CDI during the 4-year period after implementation of a 2-step CDI test algorithm.

Since 2014 the antimicrobial stewardship program reviewed all NAAT+ test results and provided feedback regarding appropriate CDI treatment, dose, and duration. After implementation of the 2-step testing protocol, feedback was provided on interpretation of NAAT+/TOX− results if the diagnosis of CDI was considered unlikely. In November 2018, an electronic medical best-practice advisory (Supplementary Fig. 1 online) was implemented in the inpatient setting that alerted clinicians that testing might not be indicated for patients who had received laxatives within 48 hours, had <3 stools documented in a 24-hour period, or had a positive CDI test within 14 days or a negative test within 7 days.

Investigator categorization and outcomes of NAAT+ patients during the initial 6 months of 2-step testing

From July through December 2018, we conducted a cohort study of a random sample of 85 patients with NAAT-positive stool results. Demographic and clinical data were abstracted from the medical record at the time of the positive NAAT result and for 3 months thereafter. The person conducting the review was blinded to toxin results. Based on predetermined criteria, NAAT+ patients were categorized as having probable, possible, unlikely, or indeterminate cases of CDI. ³

Probable CDI cases had ≥3 unformed stools per day with no alternative explanation and had received antibiotics or chemotherapy or been diagnosed with CDI within 3 months or had ≥5 unformed stools per day with leukocytosis or radiographic findings consistent with CDI and no clear alternative explanation. Possible CDI cases had 3–5 unformed stools per day with no alternative explanation but with no prior antibiotic or chemotherapy exposure. Unlikely CDI cases had <3 unformed stools per day and no leukocytosis or radiographic findings consistent with CDI or ileus OR had ≥3 unformed stools per day but with a definite alternative explanation. Cases were indeterminate if there was insufficient information to provide a classification.

For NAAT+/TOX+ and NAAT+/TOX− patients, bivariate analyses were performed to compare baseline characteristics, investigator-determined category of CDI, and the frequency of CDI treatment and of retesting within 3 months. Data were analyzed using R version 3.5.0 software (R Foundation for Statistical Computing, Vienna, Austria).

Frequency of treatment of NAAT+/TOX− patients during a 4-year period

We evaluated the frequency of CDI treatment prescribed for 240 NAAT+/TOX− patients during the 4-year period after the protocol was implemented. From July to December of each year, 10 NAAT+/TOX− patients per month were randomly selected for medical record review. Patients were considered to have received CDI treatment if ≥4 days of treatment were prescribed.

Results

Investigator categorization and outcomes of NAAT+ patients during the initial 6 months of 2-step testing

Table 1 shows a comparison of the baseline characteristics and outcomes of the 58 (68%) NAAT+/TOX− and 27 (32%) NAAT+/TOX+ patients. In comparison to NAAT+/TOX+ patients, NAAT+/TOX− patients were significantly more likely to have prior CDI and significantly less likely to be categorized as probable CDI, treated for CDI, or retested for CDI within 3 months. Of the 58 NAAT+/TOX− patients, 49 (84%) had received non-CDI antibiotics within 3 months and 7 (12%) were receiving laxatives.

Table 1. Baseline Characteristics and Outcomes of Patients Tested for Clostridioides difficile Infection (CDI) with Positive Versus Negative Enzyme Immunoassay (EIA) for Toxin

^a Units unless otherwise indicated.

^b White blood cell count data were available for 60 of 85 patients.

^c Severe CDI defined as white blood cell count of ≥15,000 cells/mL or a serum creatinine level >1.5 mg/dL (white blood cell count or serum creatinine available for 61 of 85 patients).

^d ≥4 days of CDI treatment received.

^e For initially treated patients, treatment within 3 months represents a recurrence; for untreated patients, treatment represents inadequately treated initial infection or recurrence from an episode that initially resolved without treatment.

^f The toxin negative patient was categorized at the time of the positive polymerase chain reaction test as indeterminate CDI but progressed to fulminant CDI with ileus; the toxin positive patient was initially categorized as probable CDI.

Frequency of treatment of NAAT+/TOX− patients during a 4-year period

Of the 240 NAAT+/TOX− patients assessed during the 4-year period after implementation of the 2-step algorithm, 178 (74%) were treated for CDI. The percentage of NAAT+/TOX− patients receiving treatment remained between 70% and 78% throughout the follow-up period (Fig. 1).

Discussion

During the 6 months after implementation of the 2-step testing algorithm, 79% of patients with NAAT+/TOX− results were treated for CDI. Most of these patients were classified as having probable or possible CDI based on factors such as antibiotic exposure, recent prior CDI, and absence of obvious alternative explanations for diarrhea. Only 21% of patients with NAAT+/TOX− results were classified as unlikely CDI, and most were not treated. Guerrero et al ⁶ also reported that the clinical presentation of many patients with NAAT+/TOX− results was similar to patients with NAAT+/TOX+ results. The percentage of patients with NAAT+/TOX− results receiving CDI treatment remained between 70% and 78% during a 4-year period after the change to the 2-step test procedure.

Our findings highlight the fact that many patients with NAAT+/TOX− test results have risk factors for and clinical presentations consistent with CDI. In practice, it may be difficult to limit treatment of such patients. Others have also reported that clinicians may frequently treat patients with NAAT+/TOX− results despite electronic medical record guidance suggesting that a negative toxin assay may indicate colonization. ⁷

Further study is needed to determine how often patients with NAAT+/TOX− results categorized as having probable or possible CDI improve without CDI treatment. Although a recent report suggested that it may be safe to withhold treatment for select hospitalized NAAT+/TOX− individuals, ⁸ some of these patients may have severe, complicated infections. ^5,6 Bignardi et al ⁹ reported that 9% of glutamate dehydrogenase–positive but toxin-negative patients had persistent or recurrent diarrhea and subsequently were diagnosed with CDI based with a positive toxin test. In our initial cohort 1 (8%) of 12 patients with NAAT+/TOX− results not initially treated for CDI had a subsequent NAAT+/TOX+ result and was treated for CDI.

Our study had several limitations. The evaluation was conducted in a single institution using one type of toxin assay, and feedback on interpretation of NAAT+/TOX− results was only provided if the diagnosis of CDI was considered unlikely. Additional studies are needed in other settings.

Because ultrasensitive and quantitative toxin measurement has been shown to correlate with severe disease, severe CDI-attributable outcomes, and recurrence, ¹⁰ studies are needed using these ultrasensitive toxin assays. Most NAAT+/TOX− patients received CDI therapy. Studies are needed to determine whether diarrhea in patients with NAAT+/TOX− results will resolve without treatment.