Introduction
Despite advances in the detection, treatment, and prevention of Clostridiodes difficile (C. diff) infections, they remain prevalent in the community and are the leading cause of hospital-acquired infection in the United States. Reference Mada and Alam1 Polymerase chain reaction (PCR) has been the most commonly used method for detecting C. diff in stool samples, but it has become increasingly apparent that this single test is unable to differentiate between colonization and active infection with C. diff. Older methods using enzyme-linked immunoassays were less sensitive than PCR but had the advantage of detecting toxin A and/or B production which has been shown to correlate better with C. diff colitis. Reference Cymbal, Chatterjee, Baggott and Auron2 As a result, clinicians using only PCR for diagnosis often fail to distinguish between colonization and infection and their institutions report higher rates of C. diff and higher rates of C. diff treatment.
Most facilities now use a combination of tests to diagnose clinically significant C. diff infection. Reference Carroll and Mizusawa3 Studies using a 2-step method (PCR followed by toxin testing), show lower rates of C. diff and improved antibiotic stewardship by reducing antibiotic treatment that is often considered unnecessary in colonized patients. Reference Hitchcock, Holubar, Hogan, Tompkins and Banaei4–Reference Bettger, Giancola, Cybulski, Okulicz and Barsoumian6 Recently, Stamford Hospital added C. diff toxin assay as a reflex test following a positive PCR to help identify true cases of infection, while discouraging treatment of patients who are only colonized. We prospectively investigated whether the inclusion of this additional layer of testing impacted patient outcomes and led to better antibiotic stewardship.
Study design and methods
Prior to study initiation, approval was received by the Institutional Review Board (WCG IRB Work Order #1–1 736 259–1). We performed a retrospective review of inpatients admitted to Stamford Hospital who tested positive for C. diff by PCR (Cepheid GeneXpert Systems, Sunnyvale California) between November 1, 2023, and May 1, 2024. Stamford Hospital is a 305-bed acute care urban teaching hospital in Connecticut with a primary and secondary service area totaling approximately 250,000 people. C. diff testing was performed by provider order on clinical suspicion and required provider’s written certification that the patient have at least three watery bowel movements within 24 hours, and the diarrhea was not due to cathartics, oral contrast, or tube feedings. Additionally, clinical suspicion of colitis was required (ie, leukocytosis, fever, lower abdominal pain, and cramping). Exclusion criteria were patients with chronic diarrhea (at least 4 wk duration of symptoms), history of IBD (Crohn’s disease, ulcerative colitis, microscopic colitis), and recent gastric bypass surgery.
Prior to April 1, 2023, the results of C. diff PCR testing were reported as positive or negative and no toxin resting was performed resulting in essentially all positive cases treated with antibiotics. After April 1, 2023, toxin testing (C. diff Quik Chek, Techlab, Blacksburg, VA) was reflexively added as an additional diagnostic layer and formal recommendations were made to clinicians that patients with a negative toxin test result represent colonization rather than infection and may not need to be treated.
Patient demographic information, including race, sex, age was collected for all patients meeting inclusion criteria from the hospital’s electronic medical records. Additional clinical data, such as leukocyte counts, erythrocyte sedimentation rate (ESR), and C-reactive protein (CRP) were also captured. The presence of clinical features such as abdominal pain, abdominal pain score, bloating, symptoms of colitis, antibiotic treatment for C. diff, and fatality were determined by medical record review.
Statistical analysis
Statistical analyses were conducted using SAS version 9. χ2 tests compared discrete variables, t-tests assessed continuous variables, and Mann–Whitney U tests were used for small sample sizes. A P -value <.05 was deemed statistically significant. No imputation was used for missing data.
Results
A total of 147 patients met inclusion criteria for analysis of which, 57 were toxin-positive (32%) and 90 were toxin-negative (51%). Table 1 shows the overall demographics including race, sex, and age among toxin results group. There were no significant differences in demographic characteristics among the cohorts with respect to race, gender, or age.
Table 1. Overall demographics of PCR positive based on C. diff toxin results
PCR, polymerase chain reaction; c.diff, Clostridiodes difficile; SD, standard deviation.
Table 2 shows the distribution of clinical variables between patients with a positive compared with a negative toxin result. 82% of patients with a positive result toxin had symptoms of colitis compared to 38% of those with a negative toxin result, a statistically significant finding (P < .01). Similarly, a significant difference was found in the white blood cell count (WBC) count: toxin-positive patients had a higher mean WBC count of 15,000 cells/mcl while toxin-negative patients had a WBC count of 11,000 cells/mcl (P < .01). There were no differences between the cohorts in CRP or ESR values. Symptoms of bloating were more common in toxin-positive (48%) compared with toxin-negative (31%) patients (P = .04) but there was no difference in initial or final abdominal pain scores. 98% of patients who had a positive toxin were treated with antibiotics while those with a negative toxin result were significantly less likely to be treated, 56% (P < .01). There was also a significant difference in a fatal outcome by C. diff toxin result: 12 patients in the toxin-positive group expired compared with 4 patients in the toxin negative group (P < .01).
Table 2. Clinical comparison among positive vs negative toxin result
c.diff, Clostridiodes difficile; SD, standard deviation; WBC, white blood cell; CRP, C-reactive protein; ESR, erythrocyte sedimentation rate. * Mann–Whitney U tests were used for small sample sizes.
A sub-analysis was conducted for clinical variables among positive and negative toxin results among those who had symptoms of colitis (n = 81). 34 of these patients (38%) were toxin-negative and 47 (46%) were toxin-positive. Most of the toxin-positive patients (98%) were treated with antibiotic compared to 65% of those who were toxin-negative (P < .01) (Table 3). The average initial pain score among toxin-negative patients was 3.63 and the average final pain score was lowered significantly to 1.09 (P < .01), indicating symptomatic improvement. Symptomatic improvement was also found among toxin-positive patients with average initial pain score of 3.38 and was lowered to .89 (P < .01) (Table 3).
Table 3. Sub-analysis of patients who had symptoms of colitis among positive vs negative toxin result (n = 81)
C.diff, Clostridiodes difficile; SD, standard deviation.
Discussion
This study demonstrates the efficacy of a two-tier testing strategy for C. diff, utilizing a sensitive screening test followed by a confirmatory toxin assay, in accurately discriminating between true infection and asymptomatic colonization. The implementation of this strategy resulted in a significant improvement in antibiotic stewardship, specifically a reduction in inappropriate treatment for individuals without clinically significant C. diff infection.
Our study confirms the results of others, Reference Turner, Krishnan and Nelson7 that implementation of a two-tier testing strategy can help discriminate between “true” C. diff infections (PCR+/Toxin+) and colonization (PCR+/Toxin-). Symptoms of colitis were present in 82% of patients who were toxin-positive and 98% needed antibiotic treatment. Furthermore, these true cases were associated with a statistically higher number of C. diff fatalities compared to patients who were colonized (toxin-negative). However, some toxin-negative patients had symptoms of colitis (38%) accompanied by abdominal pain and elevated ESR and they responded to antibiotics. Therefore the two-tier testing strategy is not infallible and should not be used to deny treatment. A recent survey of physicians indicated that regardless of clinical background and experience, on-going challenges continue to be presented in the treatment of C.diff Reference Nath, Lee, Bararia and Nijhawan8 and therefore clinical judgment cannot be underestimated when treating this patient population.
The observed improvement in antibiotic stewardship is a crucial finding of this study. The reduction in unnecessary antibiotic use aligns with established guidelines emphasizing the importance of targeted therapy for symptomatic C.diff infections Reference Johnson, Lavergne and Skinner9 while avoiding treatment for asymptomatic carriers. This practice not only minimizes the risk of adverse drug reactions and the development of antibiotic resistance but also reduces healthcare costs associated with prolonged and inappropriate antibiotic therapy. In our study, toxin-negative patients did not need, or receive, antibiotic treatment in over 40% of cases. This validated our educational efforts during the transition to the two-tier process and the effectiveness of our order entry “pop-up” reminders. Additionally, following initiation of toxin testing, hospitalwide memos were shared, and staff members were educated through presentations at grand rounds, house staff and hospitalist meetings, nursing and laboratory staff were also included as well.
Several limitations should be considered when interpreting our results. Firstly, the study was conducted at a single institution, potentially limiting the generalizability of our findings to other healthcare settings, although we did include all inpatient C. diff cases over the six-month study period. Secondly, the retrospective nature of the analysis may have introduced biases, such as selection bias or incomplete data capture. Thirdly, inflammatory markers, such as ESR and CRP, were not ordered consistently enough to provide more support in diagnosing colitis. Finally, many patients in this study had a history of recurrent C. diff infections but this was not included as a variable in the study so the impact of this history on our results is unknown.
The implementation of a two-tier testing strategy for C. diff significantly improved the accuracy of C.diff infection diagnosis and promoted antibiotic stewardship. By effectively differentiating between true C.diff infections and colonization, this approach reduced the unnecessary use of targeted antibiotics, contributing to better patient outcomes and mitigating the risks associated with antimicrobial resistance at our community teaching hospital. Future research should focus on validating these findings in diverse clinical settings and optimizing the selection of screening and confirmatory assays for optimal diagnostic accuracy and clinical impact.
Author contributions
DG: Conceptualization, Data curation, Investigation, Methodology, Writing—original draft, review and editing. MP: Conceptualization, Data curation, Investigation, Methodology, Writing—review and editing. MM: Conceptualization, Data curation, Investigation, Methodology, Writing—original draft, review and editing. AS: Conceptualization, Data curation, Investigation, Methodology, Writing—original draft, review, and editing. SK: Formal analysis, Writing—review and editing. FH: Project administration, Resources, Supervision. SJR: Data curation, Investigation, Methodology, Project administration, Resources, Supervision, Formal Analysis, Writing—review and editing.
Financial support
No funding was received to conduct this research.
Competing interests
All authors declare no conflicts of interest.
Declaration of generative AI in scientific writing
During the preparation of this work the authors did not use any generative AI in scientific writing.
Open access
