Introduction
Diagnosis of Clostridioides difficile infection (CDI) is challenging, with multiple testing algorithms proposed Reference McDonald, Gerding and Johnson1,Reference Gateau, Couturier, Coia and Barbut2 . With increased availability of molecular syndromic panels in clinical laboratories, many patients without typical risk factors are routinely tested for C. difficile using multiplex assays, adding further complexity to interpretation of results. Some groups have suggested suppressing C. difficile results from panels Reference Park, Hitchcock, Gomez and Banaei3 ; performing additional confirmatory testing with enzyme immunoassays (EIAs) or secondary molecular tests Reference Murphy, Fowler, Iwen and Fey4,Reference O’Neal, Murray, Dash, Al-Hasan, Justo and Bookstaver5 ; or, sending panel results for clinical review. Furthermore, the impact of panels on C. difficile positivity rates and clinical outcomes has not been fully characterized.
Our laboratory implemented an infectious diarrhea panel (IDP) in late 2023. This multiplex panel is performed on all stool samples submitted from outpatients and inpatients (<72 hours of admission) presenting with acute gastroenteritis. To assess the impact of IDP on CDI rates, we investigated C. difficile laboratory results and clinical interpretation for a 6-month period pre- and post-IDP implementation.
Methods
Prior to IDP implementation, patients presenting to two acute tertiary care hospitals (St. Paul’s Hospital and Mount Saint Joseph Hospital, Vancouver, BC) and nearby long-term care sites and outpatient clinics with clinical suspicion for CDI had stool samples submitted for dedicated two-step C. difficile testing. Stool samples first underwent molecular detection of the tcdB gene (Xpert® C. difficile/Epi, Cepheid). Positive samples underwent further analysis by enzyme immunoassay (EIA) for direct detection of C. difficile toxin A/B and glutamate dehydrogenase (GDH) antigen (C. Diff QuikChek Complete, Techlab). If toxin was not detected by EIA, final results were reported as “Indeterminate” rather than “Positive.” Clinical review of each inpatient case (including patients admitted from the Emergency Department [ED]) was conducted by the hospital’s antimicrobial stewardship program (ASP), determining whether patients were infected (new-onset ≥3 loose stools in 24 hours without an alternate diagnosis) or colonized (alternate diagnosis identified based on clinical review of stool charts, medications [e.g., laxatives or other medications associated with diarrhea], laboratory results, underlying conditions, and final discussion with the patient’s attending physician) Reference Lowe, Shakeraneh, Lee, Sharma and Leung6,Reference Zou, Leung and Champagne7 .
After implementation of IDP, dedicated two-step C. difficile testing remained available for inpatients when clinically suspected; however, all patients presenting with acute gastroenteritis for any reason had stool samples tested by IDP (BioFire® FilmArray® Gastrointestinal [GI] Panel, bioMérieux), replacing traditional stool bacterial culture and ova & parasite examination. Samples with C. difficile incidentally detected by IDP underwent further testing by EIA. Clinical review by infection prevention and control (IPAC) and ASP was conducted for all inpatient cases, as previously described Reference Lowe, Shakeraneh, Lee, Sharma and Leung6,Reference Zou, Leung and Champagne7 .
Results were retrospectively reviewed for a 6-month period after implementation of IDP (February-July 2024). Rates of positivity and colonization versus CDI were compared to the same 6-month period before IDP implementation (February–July 2023). Fisher’s exact test (GraphPad QuickCalcs) was used for statistical analysis where applicable, with p < 0.05 considered significant.
Results
In the post-IDP period, C. difficile was the most frequently detected pathogen on IDP in our patient population (10.8% detection rate). A higher number of stool samples underwent C. difficile testing compared to the pre-IDP period (1,661 vs 1,049), with a greater proportion submitted from ED and outpatients with a younger median age (Table 1). Of note, the number of orders for dedicated C. difficile testing decreased by nearly half compared to the pre-IDP period. Of samples testing positive for C. difficile by IDP, a significantly higher proportion (27%) tested negative for both GDH and toxin EIA compared to dedicated C. difficile testing in the same post-IDP period (6%, p < 0.001) and the pre-IDP period (11%, p < 0.001).
Table 1. Results of stool samples submitted to the microbiology laboratory during the pre-IDP (February to July 2023) and post-IDP (February to July 2024) periods
* p < 0.001 compared to Infectious Diarrhea Panel (IDP).
NAT, nucleic acid amplification test, either by BioFire GI Panel or Xpert C. difficile/Epi test as indicated.
EIA, enzyme immunoassay.
GDH, glutamate dehydrogenase antigen.
Reference McDonald, Gerding and Johnson1 Stool samples rejected if <6 on Bristol Stool Chart or repeated within 7 days. IDP orders were additionally rejected if sample collected >72 hours after hospital admission.
Clinical review of inpatient cases revealed a significantly greater proportion of patients tested by IDP was interpreted as “Colonized” compared to patients in the pre-IDP period (46.9% vs 27.5%, p = 0.01). Even when comparing to patients who had dedicated C. difficile tests performed within the same post-IDP period, colonization rates were higher among patients tested by IDP only, although this difference did not quite reach statistical significance (46.9% vs 37.3%, p = 0.30). Clinical outcomes including critical care admission, surgical intervention, and 30-day all-cause mortality did not significantly differ between the pre- and post-IDP periods (Table 2).
Table 2. Classification of inpatient C. difficile cases by clinical review, including clinical outcomes in the pre- and post-IDP periods
* p = 0.01 compared to the pre-IDP (Infectious Diarrhea Panel) period.
**p < 0.01 compared to the pre-IDP period and to IDP results in the post-IDP period.
ICU, intensive care unit.
Discussion
Molecular syndromic panels have several advantages including improved efficiency and turnaround time; however, routine testing for C. difficile regardless of patients’ pre-test probability (or prevalence of the condition in the population being tested) may not be optimal. Patients with CDI typically have risk factors and clinical presentations that differ from those with foodborne or community-acquired infectious diarrhea, and molecular assays for C. difficile toxin genes may be detecting asymptomatic carriers of C. difficile rather than those with CDI Reference Polage, Gyorke and Kennedy8 . Our study demonstrates two different patient populations being tested for C. difficile in the pre- and post-IDP periods, with an impact on clinical interpretation of the results.
The detection rate of C. difficile by IDP in our study was similar to what has been previously described in other centres using the BioFire GI panel (9.7–16.3%) Reference Park, Hitchcock, Gomez and Banaei3–Reference O’Neal, Murray, Dash, Al-Hasan, Justo and Bookstaver5 . The majority of these samples tested negative for toxin EIA in our study, consistent with previous reports (57–78%) Reference Murphy, Fowler, Iwen and Fey4,Reference Wadskier, Cowman and Szymczak9,Reference Pender, Throneberry, Grisel, Leung and Lopansri10 . Our laboratory previously observed negative EIA toxin in 60% of samples when using a laboratory-developed test Reference Zou, Leung and Champagne7 and 67% when using Xpert Reference Lowe, Shakeraneh, Lee, Sharma and Leung6 as the initial molecular assay in a two-step algorithm, which increased to 73% using IDP in this study. A key finding was the significant increase in the proportion of samples testing negative for both GDH and toxin EIA compared to our pre-IDP period. No samples with both negative GDH and toxin EIA were observed in our centre’s previous study using a laboratory-developed assay as the initial molecular test Reference Zou, Leung and Champagne7 .
Clinical review was conducted on only a subset of the cases (inpatients), which demonstrated a higher proportion being interpreted as “Colonization” when tested by IDP. This increase may be driven by the higher rate of GDH and toxin EIA negative samples in this cohort, and also potentially reflects the lower CDI pre-test probability in patients undergoing IDP. The significant decrease in dedicated C. difficile orders during the post-IDP period suggests many clinicians ordered IDP instead of dedicated C. difficile testing. This may be concerning from an ASP perspective, as previous studies have demonstrated patients with positive C. difficile results by a molecular assay are likely to receive treatment regardless of their pre-test probability or EIA results Reference Pender, Throneberry, Grisel, Leung and Lopansri10 . Unfortunately, IPAC and ASP surveillance teams in our centre are not able to review C. difficile results from outpatients and those discharged from ED; it is unclear how clinicians in the community may be interpreting and managing indeterminate C. difficile IDP results.
This study has additional limitations, including potential confounding factors during the pre- and post-IDP periods that may have affected C. difficile positivity rates and clinical interpretation; however, this study design was necessary to evaluate real-world data after implementation of a new testing method. Additional outcomes of interest including antibiotic usage and symptom resolution are not routinely collected by IPAC and ASP teams and were not available for analysis. Sample size was limited due to C. difficile positivity rates in our patient population.
With a two-step algorithm in use, the IDP did not significantly alter C. difficile “Positive” and “Indeterminate” rates, although a greater proportion of inpatient cases was interpreted as colonization after clinical review. To prevent potential over-treatment of C. difficile IDP results, it would be important to continue the two-step algorithm and clinical review, and consider suppression of C. difficile results from molecular syndromic panels in populations where clinical consultation is not available or for which colonization rates are high.
Acknowledgments
We are grateful to the St. Paul’s Hospital microbiology lab staff and Infection Prevention and Control staff for their dedication to quality testing and patient care.
Financial support
This study did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Competing interests
V.L and M.G.R report speaker fees from bioMérieux unrelated to this work. The remaining authors have no conflicts of interest to declare.
Open access
