Clinical Outcomes of Treated and Untreated C. difficile PCR-Positive/Toxin-Negative Adult Hospitalized Patients: a Quasi-Experimental Noninferiority Study

ABSTRACT

Clostridioides difficile infection (CDI) is routinely diagnosed by PCR, with or without toxin enzyme immunoassay testing. The role of therapy for positive PCR and negative toxin remains unclear. The objective of this study was to determine whether clinical outcomes of PCR⁺/cycle threshold-based toxin (CT-toxin)⁻ individuals vary by result reporting and treatment strategy. We performed a quasiexperimental noninferiority study comparing clinical outcomes of PCR⁺/CT-toxin⁻ individuals by reporting PCR result only (most patients treated) with reporting CT-toxin result only (most patients untreated) in a single-center, tertiary academic hospital. The primary outcome was symptomatic PCR⁺/CT-toxin⁺ conversion at 8 weeks. Secondary outcomes included 7-day diarrhea resolution, hospital length of stay, and 30-day all-cause mortality. A total of 663 PCR⁺/CT-toxin⁻ test results were analyzed from 632 individuals with a median age of 61 years (interquartile range [IQR], 44 to 72) and 50.4% immunocompromised. Individuals in the preintervention group were more likely to have received CDI therapy than those in the intervention group (91.5 versus 15.1%; P < 0.001). Symptomatic toxin conversion at 8 weeks and hospital length of stay failed to establish the predefined thresholds for noninferiority. Lack of diarrhea resolution at 7 days and 30-day all-cause mortality was similar and established noninferiority (20.0 versus 13.7%; adjusted odds ratio [aOR], 0.57; 90% confidence interval [CI], 0.32 to 1.01; P = 0.1; and 8.6 versus 6.5%; aOR, 0.46; 90% CI, 0.20 to 1.04; P = 0.12). These data support the safety of withholding antibiotics for selected hospitalized individuals with suspected CDI but negative toxin.

INTRODUCTION

Clostridioides difficile is the most important source of nosocomial diarrhea and is associated with poor clinical outcomes (1). Widespread transition to nucleic acid amplification tests (NAATs) for the diagnosis of C. difficile infection (CDI) has resulted in a significant increase in documented CDI incidence rate due to its higher analytical sensitivity (2, 3). However, in the absence of a reference standard for CDI, the optimal diagnostic testing algorithm for CDI remains debated (4). Two main strategies are recommended: use of a NAAT alone for symptomatic individuals with unexplained and new-onset diarrhea (≥3 unformed stools in 24 h) or use of a two-step approach, such as NAAT followed by a toxin test for positive results (5). Current guidelines recommend both strategies and only differentiate on the basis of institutional guidelines for appropriate samples for testing (5). Consistent with the pathogenesis model in which C. difficile toxins mediate disease, individuals with both positive NAAT and direct toxin assays have been shown to have a longer duration of symptoms, increased length of stay, and increased mortality compared to individuals with only positive NAAT (6–12). Similarly, NAAT⁺/toxin⁻ patients had symptom duration and mortality similar to that of NAAT⁻/toxin⁻ patients (6–12). However, there remains clinical uncertainty and lack of evidence regarding the need for antimicrobial therapy for the subgroup of NAAT⁺/toxin⁻ individuals.

The Stanford Health Care (SHC) Clinical Microbiology Laboratory previously validated a toxin prediction algorithm based on the PCR (PCR) cycle threshold (C_T) value, enabling prediction of a C_T-based toxin (termed CT-toxin) result with a sensitivity of 96.0% and negative predictive value of 99.8% at a C_T of 26.35 when using C. Diff Quik Chek Complete rapid membrane enzyme immunoassay (RMEIA) toxin (TechLab, Blacksburg, VA) as the reference method (13). CT-toxin was 92.0% sensitive when using either EIA or cell culture cytotoxicity neutralization assay (CCNA) as the reference standard (13). Starting in November 2017, given the concern for C. difficile overtreatment in PCR⁺/CT-toxin⁻ patients and after consultation with the institutional antimicrobial stewardship program (ASP) and infection prevention and control (IPAC) group, reporting was changed to include only the CT-toxin result as a surrogate for EIA and CCNA. This study aimed to assess if PCR⁺/CT-toxin⁻ individuals in the intervention group (reporting CT-toxin result only), who mostly did not receive CDI treatment, showed noninferior clinical outcomes compared to PCR⁺/CT-toxin⁻ individuals in the preintervention group (reporting PCR result only), who mostly received CDI treatment.

MATERIALS AND METHODS

Ethics.

This study was approved by the Stanford institutional review board (IRB; protocol number 54837), and individual consent was waived.

Study design.

We performed a quasiexperimental study comparing two C. difficile test result periods at SHC, a 605-bed tertiary academic medical center (Fig. S1). This center serves an adult patient population with comprehensive programs for the care of immunocompromised hosts. C. difficile testing was performed on the Xpert C. difficile/Epi tcdB PCR assay (Cepheid Inc., Sunnyvale, CA) per manufacturer instructions. The predicted toxin result (CT-toxin) was calculated based on a C_T value cutoff of 27.5, using an algorithm previously validated at our institution using the same assay (13). This cutoff resulted in a sensitivity of 99.0% and negative predictive value of 99.8% using rapid membrane EIA (for the detection of fecal free toxin TcdA and/or TcdB) as the reference method. When using either EIA or CCNA as the reference standard, the sensitivity is 92.0% with a negative predictive value of 99.0%. Hospitalized individuals aged 18 years and above with at least one PCR⁺/CT-toxin⁻ result were included. The preintervention period consisted of C. difficile test reporting that included only qualitative PCR results (positive/negative) and ranged from 21 February 2015 to 9 October 2016. Unreported CT-toxin results from this time period were predicted per the algorithm described above, and patients with PCR⁺/CT-toxin⁻ results were retrospectively identified. The intervention period consisted of prospective C. difficile test reporting that included the qualitative CT-toxin result only (positive/negative) and ranged from 1 November 2017 to 24 June 2020. The decision to initiate CDI therapy was left at the discretion of the providers, and there was no direct ASP intervention. Patients receiving empirical CDI treatment prior to test results were not excluded from the analysis. The period between 9 October 2016 and 1 November 2017 consisted of dual reporting of PCR and CT-toxin results. Thus, testing did not change across study periods; only reporting of results changed over time. Results from this dual reporting period, which have been described previously (14), were not included in this current study, so as to limit the impact of provider knowledge of the PCR result in influencing initiation of therapy in PCR⁺/CT-toxin⁻ patients. Aside from the C. difficile test reporting changes, there was no significant change in institutional guidelines for CDI antibiotic therapy and prevention approaches between the preintervention and intervention study periods. Prior to the modification in C. difficile test result reporting, documentation was circulated to clinical teams to inform them of the changes. In August 2017, 1 year after initiation of the dual reporting period, the institutional guidelines were updated to reflect the reporting with PCR and CT-toxin and emphasized the importance of a positive CT-toxin along with a compatible syndrome in making a CDI diagnosis.

The primary outcome of the study was symptomatic C. difficile PCR-positive, toxin-positive conversion (PCR⁺/CT-toxin⁺) within 8 weeks of the index CDI diagnosis. Secondary outcomes included lack of symptom resolution (unresolved diarrhea) at 7 days, hospital length of stay, and 30-day all-cause mortality. The SHC Microbiology Laboratory implemented C. difficile testing criteria in July 2015 restricting inpatient testing to individuals with 3 or more unformed stools per 24 h and absence of laxative use in the preceding 48 h, along with preexisting criteria of no previous C. difficile test in the preceding week (15–17). These criteria were followed without modification during this study.

Data collection.

A customized electronic medical data extraction report was performed to retrieve all C. difficile tests during the preintervention and intervention periods, in addition to demographic data, clinical characteristics, hospital admission data, and mortality. Additional individual chart review was performed by two users (C.A.H. and S.F.) to assess CDI therapy in detail and assess CDI-attributable complications. Data collection was performed to include any antibiotic received; however, only oral or intrarectal vancomycin, oral or intravenous metronidazole, and/or oral fidaxomicin were considered CDI therapy for the purposes of this study. Only individuals with a PCR⁺/CT-toxin⁻ result during the intervention and preintervention periods were included. From the identified PCR⁺/CT-toxin⁻ index test result, individuals were monitored for 8 weeks for symptomatic PCR⁺/CT-toxin⁺ conversion and for 30-day all-cause mortality. Resolution of symptoms at 7 days was defined as the absence of 3 or more unformed stools, as documented in the medical chart by validated nursing protocols at 7 days after the index PCR⁺/CT-toxin⁻ result (15). Data were considered missing if this information was not recorded in the medical chart. CDI therapy was defined as administration of 7 days or more of CDI therapy. Severe CDI definition was based on the IDSA 2018 definition and includes white blood cell (WBC) count of >15,000 cells/mL and/or serum creatinine of ≥1.5 mg/dL (5). Repeat tests with a PCR⁺/CT-toxin⁻ result from a single individual were included in the study if more than 8 weeks had elapsed from the original PCR⁺/CT-toxin⁻ test result. Strain typing data were collected for 027/NAP1/BI and non-027/NAP1/BI from the Xpert C. difficile/Epi tcdB PCR instrument.

Statistical analysis.

Statistical analysis was performed using R version 4.0.0. Differences in descriptive statistics across groups were t tests for continuous variables and chi square tests for categorical variables. PCR⁺/CT-toxin⁺ symptomatic conversion at 8 weeks, symptom persistence at 7 days, and all-cause 30-day mortality were analyzed by logistic regression analysis, both crude and adjusted for the a priori determined potential confounders of age, sex, CDI severity, history of CDI in the preceding 6 months, and active CDI therapy. Thus, given not all individuals with a PCR⁺/CT-toxin⁻ result in the preintervention period were treated and not all individuals in the intervention period were untreated, CDI treatment status was accounted for in the adjusted analysis, which forms the basis of the main results reported in this study. Hospital length of stay was assessed by Cox proportional hazards regression analysis in both unadjusted and adjusted models for the same potential confounders as those described above. For each analysis, a priori noninferiority margins were defined as an adjusted odds ratio (aOR) of 1.1 for 30-day all-cause mortality and an aOR or adjusted hazard ratio (aHR) of 1.15 for the primary and other secondary outcomes. The 90% confidence intervals were computed to account for the noninferiority design (18). A statistical threshold of a P value of ≤0.10 was considered significant, and complete case analysis was performed. Censoring was performed based on mortality and individuals lost to follow-up. Based on estimated symptomatic conversion rates of 8% in the preintervention group and 16% in the intervention group, a two-sided alpha level of 0.05, and power of 80%, the required sample size was calculated to be 268 individuals per group.

RESULTS

A total of 13,468 C. difficile tests were performed during the preintervention and intervention periods. Of these, 663 PCR⁺/CT-toxin⁻ tests from 632 individuals were included in the study (Fig. 1). Individuals with repeat testing had a median of 2 tests (interquartile range [IQR], 2 to 3) included in the preintervention period and 2 tests (IQR 2 to 3) in the intervention period. The median age in the entire cohort was 61 (IQR, 44 to 72), and about half (51.3%) were female (Table 1). Immunocompromised status was present in 334 (50.4%) individuals, most commonly from solid malignancy (23.4%), hematologic malignancy (21.9%), and solid-organ transplant (8.8%). A total of 19 individuals across both study periods underwent repeat CDI testing beyond 8 weeks from the first test, and the median time between tests was 193 days (interquartile range, 128 to 246 days). Proportion of PCR+ results that were 027/NAP1/BI strain was 17% in the preintervention versus 13.8% in the intervention period (P = 0.05). The proportion of individuals treated for CDI was significantly higher in the preintervention group than the intervention group (91.5% versus 15.1%; P <0.001). Of the 63 individuals treated in the intervention group, 44 (69.8%) received metronidazole for ≥7 days as part of antibiotic therapy for another syndrome in the following proportions: 70.5% for suspected gastrointestinal source, 20.5% for sepsis or septic shock, and 9.1% for unknown or other source. The remaining 19 patients received CDI therapy preemptively for suspected CDI despite the negative CT-toxin result, most frequently based on the management recommendations of the infectious diseases team, due to high clinical suspicion, recent history of a positive C. difficile result, known recurrent disease, or a combination thereof. Similarly, repeat CDI testing within 8 weeks was more frequent in the preintervention group than the intervention group (30.5% versus 19.2%; P = 0.001). For the 663 CDI episodes included, a total of 527 (79.0%) were from individuals with antibiotic use documented in the EMR, with intravenous vancomycin (269; 51.0%), piperacillin-tazobactam (206; 39.1%), and ceftriaxone (219; 41.6%) being the most common agents received.

FIG 1.

Flowchart of the study. CT-toxin, cycle threshold-based toxin. Indeterminate refers to an indeterminate PCR result.

TABLE 1.

Demographic and clinical characteristics of individuals from the 663 tests included in this study^a

Parameter	Overall (n = 663)	Preintervention (n = 246)	Intervention (n = 417)	P value (LR)
Age, yr, median (IQR)	61 (44–72)	59 (41–69)	63 (46–73)	0.02
Female sex, no. (%)	340 (51.3)	120 (48.8)	220 (52.8)	0.3
Immunocompromised, no. (%)	334 (50.4)	115 (46.8)	219 (52.5)	0.2
Hematologic malignancy, no. (%)	145 (21.9)	43 (17.5)	102 (24.5)	0.04
BMT, no. (%)	57 (8.6)	23 (9.4)	34 (8.2)	0.7
Solid malignancy, no. (%)	155 (23.4)	52 (21.1)	103 (24.7)	0.3
Solid-organ transplant, no. (%)	58 (8.8)	26 (10.6)	32 (7.7)	0.3
Treatment for CDI in the 6 mo prior to initial testing, no. (%)	61 (9.2)	27 (11.0)	34 (8.2)	0.3
Antibiotics in the 30 days prior to initial testing, no. (%)
Yes	498 (75.1)	190 (77.2)	308 (73.9)	0.3
No	165 (24.9)	56 (22.8)	109 (26.1)
Patient location at time of CDI testing, no. (%)
Inpatient ward	139 (21.0)	52 (21.1)	87 (20.9)	0.2
ED	505 (76.2)	183 (74.4)	322 (77.2)
Outpatient	19 (2.9)	11 (4.5)	8 (1.9)
Need for ICU transfer within 7 days of CDI test, no. (%)	95 (14.3)	44 (17.9)	51 (12.3)	0.07
WBC within 24 h of CDI test, medianb (IQR)	8.2 (4.7–12.8)	8.5 (5.5–12.9)	8.0 (4.4–12.6)	0.1
Creatinine within 24 hours of initial test, medianc (IQR)	0.9 (0.7–1.4)	1.0 (0.8–1.7)	0.9 (0.6–1.4)	0.07
Albumin within 24 hours of initial test, mediand (IQR)	2.8 (2.3–3.4)	2.7 (2.1–3.1)	3.0 (2.4–3.6)	<0.01
Severe CDIe	235 (35.4)	80 (32.5)	155 (37.2)	0.2
CT value, positive initial toxin result (median, IQR)	31.9 (29.9–34.1)	32.3 (29.8–34.7)	31.8 (29.9–33.8)	0.08
CDI therapy, no. (%)	288 (43.4)	225 (91.5)	63 (15.1)	<0.001
Metronidazole treatment, no. (%)	132 (19.9)	85 (34.6)	47 (11.3)	<0.001
Vancomycin treatment, no. (%)	80 (12.1)	73 (29.7)	7 (1.7)	<0.001
Metronidazole and vancomycin treatment, no. (%)	69 (10.4)	61 (24.8)	8 (1.9)	<0.001
Fidaxomicin alone or in combination, no. (%)	7 (1.1)	6 (2.4)	1 (0.2)	0.01
Repeat CDI testing within 8 wk after initial test, no. (%)
Yes	155 (23.4)	75 (30.5)	80 (19.2)	0.001
No	508 (76.6)	171 (69.5)	337 (80.8)
No. of repeat tests within 8 wk after initial test, median (IQR)	1 (1–2)	1 (1–2)	1 (1–1)	0.07
Hospital readmission within 30 days of discharge, no. (%)	205 (30.9)	74 (30.1)	131 (31.4)	0.7

^aBMT, bone marrow transplant; CDI, C. difficile infection; C_T value, cycle threshold value; ED, emergency department; ICU, intensive care unit; IQR, interquartile range; LR, logistic regression; WBC, white blood cell. P values were calculated using t tests (proportion) or Wilcoxon rank sum test (median) for continuous variables and chi square tests for categorical variables.

^bTotal of 22 data points missing.

^cTotal of 16 data points missing.

^dTotal of 150 data points missing.

^eSevere CDI definition based on the IDSA 2018 definition and includes WBC of >15,000 cells/mL or serum creatinine of ≥1.5 mg/dL (5).

Noninferiority testing of the clinical outcomes between the two groups revealed similar measures of effect between the unadjusted and adjusted analyses (Table 2 and Fig. 2). Symptomatic C. difficile PCR⁺/CT-toxin⁺ conversion at 8 weeks appeared similar between the two groups but failed to establish the predefined threshold for noninferiority (5.3 versus 6.7%; aOR, 0.90; 90% confidence interval [CI], 0.37 to 2.16; P = 0.84). In contrast, the proportion of unresolved diarrhea at 7 days met the predefined margin of noninferiority (20.0 versus 13.7%; aOR, 0.57; 90% CI, 0.32 to 1.01; P = 0.1). Hospital length of stay was suggested to be longer in the intervention group than in the preintervention group (10.8 versus 6.9 days; aHR, 1.20; 90% CI, 0.88 to 1.64; P = 0.26), failing to meet noninferiority criteria. Finally, all-cause 30-day mortality was established as noninferior in the intervention group (8.6 versus 6.5%; aHR, 0.46; 90% CI, 0.2 to 1.04; P = 0.12). Of the 21 deaths (8.6%) in the preintervention group and 26 deaths (6.5%) in the intervention group, only 4 deaths in the preintervention group were partially or fully attributable to CDI.

TABLE 2.

Primary and secondary outcomes of the study in the intervention group (CT-toxin only reporting) compared to the preintervention group (PCR only reporting)

Outcome	No. (%) or median (IQR) in:		Unadjusted HR or OR (90% CI)	Unadjusted P value	aHR or OR (90% CI)	Adjusted P valuee	Noninferiority margin	Evidence of noninferiority established
	Preintervention group	Intervention group
Symptomatic C. difficile PCR+/CT-toxin+ conversion within 8 wka	13 (5.3)	28 (6.7)	1.29 (0.73–2.28)	0.46	0.90 (0.37, 2.16)	0.84	1.15	No
Unresolved diarrhea at 7 daysb	40 (20.0)	57 (13.7)	0.63 (0.44–0.92)	0.04	0.57 (0.32, 1.01)	0.10	1.15	Yes
Median hospital length of stayc	10.8 (4.2–22.1)	6.9 (3.0–17.8)	1.30 (1.11–1.53)	0.001	1.20 (0.88, 1.64)	0.26	1.15	No
30-Day all-cause mortalityd	21 (8.6)	26 (6.5)	0.74 (0.45–1.22)	0.32	0.46 (0.20, 1.04)	0.12	1.10	Yes

^aAnalysis performed by logistic regression model.

^bAnalysis performed by logistic regression model; 46 data points missing.

^cAnalysis performed by Cox regression model.

^dAnalysis performed by logistic regression model; 17 data points missing.

^eAdjusted for age, sex, immunocompromised host status, CDI therapy in preceding 6 months, and active CDI therapy; significance level of 0.10. Thus, the effect of CDI therapy in each group is accounted for in the adjusted analysis.

FIG 2.

Noninferiority plot of the primary and secondary outcomes of the study. aHR, adjusted hazard ratio; aOR, adjusted odds ratio; wk, week. The two vertical dotted lines represent the margins of noninferiority: 1.1 for 30-day mortality and 1.15 for the other outcomes.

A total of 13 individuals (5.3%) in the preintervention group and 28 (6.7%) individuals in the intervention group were diagnosed with symptomatic C. difficile PCR⁺/CT-toxin⁺ conversion at 8 weeks. Of these episodes, 22 (53.7%) were classified as severe disease, 13 (31.7%) individuals required new hospitalization for CDI, and the median length of hospitalization was 3 days (IQR, 2 to 13) (Table S1). Furthermore, there was no CDI-attributable mortality among individuals with positive conversion.

DISCUSSION

The clinical importance of toxin positivity in CDI has been challenged in observational studies, with evidence for and against worse clinical outcomes compared to individuals with negative toxin immunoassay results (19–22). However, the clinical impact of CDI treatment versus no treatment in PCR⁺/toxin⁻ CDI suspects has not been investigated using experimental study designs. In this study, of over 600 hospitalized symptomatic adults with a PCR⁺/CT-toxin⁻ result, restricting the C. difficile result report to the CT-toxin result alone was associated with a significant reduction in antimicrobial therapy for suspected CDI compared to reporting the PCR result alone. There was no major signal for an increase in adverse events as a result of this change in diagnostic test result reporting, and noninferiority thresholds were achieved for unresolved diarrhea at 7 days and 30-day all-cause mortality between the two reporting strategies. However, although symptomatic PCR⁺/CT-toxin⁺ conversion within 8 weeks and hospital length of stay were similar in the two groups, we did not establish the predefined thresholds to support noninferiority.

As observed in this study, a major benefit of reporting the CT-toxin result alone is the favorable impact on reducing CDI overtreatment. Indeed, a reduction in treatment rate from 91.5% to 15.1% in PCR⁺/CT-toxin⁻ cases between the two intervention periods was observed. Of the individuals in the intervention group who received therapy, the most common justification was administration of metronidazole as part of empirical combination antimicrobial therapy to cover a gastrointestinal source of infection, with or without sepsis. Combining these data with our previous report of a decrease from 96 to 51.2% of treatment with reporting of both PCR and CT-toxin results (14) and of 71% treatment in another setting with reporting of both PCR and EIA toxin results (23) supports the increasing benefit of only reporting the toxin result to reduce unnecessary therapy. Of note, rates of empirical therapy did not increase during the CT-toxin reporting period, and there were no changes to the infection control guidelines for room cleaning, hand hygiene, or isolation procedures during this study. Furthermore, our institutional hospital-onset CDI (HO-CDI) rates declined by approximately half with introduction of CT-toxin only reporting. Reduction of unnecessary C. difficile therapy is an important antimicrobial stewardship program target given its association with increased prevalence of resistant organisms such as vancomycin-resistant enterococci (VRE) (24). Furthermore, averting antibiotic therapy may limit host microbiota dysbiosis (4), which is associated with reduced colonization resistance to gastrointestinal pathogens and increased subsequent risk of CDI (25, 26) and proposed morbidity and mortality (27).

Although we showed similar clinical outcomes in treated and untreated PCR⁺/CT-toxin⁻ CDI suspects, this study presents several limitations. First, the single-center, retrospective study design did not account for confounders the way a randomized-controlled trial could have. However, given the challenging nature of the study question due to conflicting findings from observational studies, this study design represents a pragmatic approach to study the need for CDI therapy in PCR⁺/toxin⁻ patients. Second, this study did not exclude individuals on the basis of a positive CDI result prior to study entry, which resulted in inclusion of a small fraction of patients who received CDI therapy in the intervention group despite a negative test result. Third, use of the CT-toxin approach presents some accuracy differences compared to EIA, which is currently the main viable toxin testing alternative for timely clinical decision making. Although CT-toxin is 99% sensitive and has very high negative predictive value (99.8%) compared with toxin EIA used as the reference method (13), CT-toxin is more sensitive than EIA and detects the majority of EIA-negative but CCNA-positive CDI individuals. Furthermore, CT-toxin overcalls toxin positivity in PCR⁺/toxin⁻ individuals (13). In addition, only the Xpert C. difficile/Epi tcdB PCR assay was used in this study. Thus, our findings have to be reproduced using toxin EIA to determine whether results generated in this study are considered applicable to other settings using toxin EIA testing and with other PCR assays to evaluate alternative C_T cutoffs (13). Fourth, this study included a large proportion of immunocompromised populations and few individuals with gastrointestinal pathology, which may limit generalizability of results. Fifth, we could not account for the different antibiotics and/or administration of intravenous immune globulin (IVIG) administered for non-CDI purposes during the 8-week follow-up period. Finally, despite the selection of individuals with PCR⁺/toxin⁻ results for the study, we observed a high proportion (>1/3) of individuals with severe disease based on the most recent IDSA guidelines, which may confound the interpretation of CT-toxin-negative results. Similarly, 22/41 (53.7%) individuals with positive conversion within 8 weeks satisfied criteria for severe disease, of which 13 required hospital admission for CDI. However, the definitions for severity are not standardized and may be overly sensitive in this hospitalized population with several potential medical reasons underlying elevated creatinine or WBC level.

In summary, this study extended our understanding of feasibility and safety of reporting C. difficile toxin results alone. It showed a significant reduction in antibiotic treatment for CDI in patients with PCR⁺/CT-toxin⁻ results while demonstrating noninferiority for diarrhea resolution at 7 days and 30-day all-cause mortality compared to PCR⁺/CT-toxin⁻ patients who were treated. However, the noninferiority criteria were not met for 8-week symptomatic toxin conversion and hospitalization length of stay, limiting the generalizability of findings. Further studies are required to more comprehensively describe the clinical impact of such an intervention.