Nucleic acid amplification tests are commonly used to diagnose Clostridioides difficile infection (CDI). Two-step testing with a toxin enzyme immunoassay is recommended to discriminate between infection and colonization but requires additional resources. Prior studies showed that PCR cycle threshold (CT) can predict toxin positivity with high negative predictive value. Starting in October 2016, the predicted toxin result (CT-toxin) based on a validated cutoff was routinely reported at our facility.
KEYWORDS: PCR, toxin, cycle threshold, Clostridioides difficile
ABSTRACT
Nucleic acid amplification tests are commonly used to diagnose Clostridioides difficile infection (CDI). Two-step testing with a toxin enzyme immunoassay is recommended to discriminate between infection and colonization but requires additional resources. Prior studies showed that PCR cycle threshold (CT) can predict toxin positivity with high negative predictive value. Starting in October 2016, the predicted toxin result (CT-toxin) based on a validated cutoff was routinely reported at our facility. To evaluate the clinical efficacy of this reporting, all adult patients with positive GeneXpert PCR results from October 2016 through October 2017 underwent a chart review to measure the recurrence of or conversion to a CT-toxin+ result and 30-day all-cause mortality. There were 482 positive PCR tests in 430 unique patients, 282 CT-toxin+ and 200 CT-toxin−. Patient characteristics were similar at testing, though CT-toxin+ patients had higher white blood cell (WBC) counts (12.5 × 103 versus 9.3 × 103 cells/μl; P = 0.001). All cases (n = 21) of fulminant CDI had a CT-toxin+ result. Index CT-toxin+ patients were significantly more likely to have a CT-toxin+ result within 90 days than CT-toxin− patients (17.4% [n = 49] versus 8.0% [n = 16], respectively; P = 0.003). Thirty-day all-cause mortality was higher in CT-toxin− patients (11.1% versus 6.8%; P = 0.1), though no deaths in CT-toxin− patients were directly attributable to CDI. Of the 200 CT-toxin− patients, 51.5% (n = 103) were treated for CDI. The rates of conversion to a CT-toxin+ result (8.8% versus 7.2%; P = 0.8) and all-cause mortality (8.8% versus 13.4%; P = 0.3) were similar between treated and untreated CT-toxin− patients, respectively. CT-based toxin prediction may identify patients at higher risk for CDI-related complications and reduce treatment among CT-toxin− patients.
INTRODUCTION
Clostridioides difficile infection (CDI) remains the most common infectious cause of nosocomial diarrhea (1), with an estimated 450,000 cases annually in the United States (2), and carries additional significance for hospitals due to the public reporting requirements and associated financial penalties (3). The incidence of CDI rose dramatically in the first decade of the 21st century, driven by the emergence of the epidemic BI/NAP1/027 strain and the move toward more sensitive testing for the presence of toxigenic C. difficile with nucleic acid amplification tests (NAATs) (4). The transition from toxin assays to NAATs was followed by a 43% to 97% increase in the incidence of CDI without evidence of increased transmission or antibiotic use (5–8), suggesting that the use of NAATs was leading to the identification of patients who had subclinical disease or were merely colonized with C. difficile and had diarrhea from an unrelated cause (9). Additional studies showed that the differential sensitivities of the various tests are, at least in part, determined by organismal burden, with samples positive by toxin assays having significantly higher bacterial loads than samples positive only by NAAT (10–12). Studies on clinical findings and patient outcomes have shown that toxin positivity by enzyme immunoassay (EIA) or cell culture cytotoxicity neutralization assay (CCNA) is associated with increased duration of diarrhea, longer length of stay, and increased mortality rates compared to those in patients who have C. difficile detected only by NAAT, and outcomes were similar between these latter patients and those negative for C. difficile by all testing methods (13–16). These results have led to revised testing guidelines in the United States and Europe, which currently recommend the use of a multistep algorithm to identify toxin positivity rather than use of NAATs alone and to avoid testing patients who do not have clinically significant diarrhea or have a readily identifiable alternative cause of diarrhea such as laxative use (17, 18), as these have been shown to reduce inappropriate testing rates (19). While two-step testing with NAAT and EIA can help differentiate between patients with true CDI and those who may simply be colonized with C. difficile, this algorithm requires additional labor, resources, and time.
There is an alternative testing methodology that leverages additional information from the quantitative PCR result to obtain data similar to that from a two-step methodology within a single test. Using the Xpert C.diff/Epi tcdB PCR (Cepheid, Sunnyvale, CA), Senchyna et al. compared the PCR cycle threshold (CT), which is known to be correlated with organismal load (10), to results from a multistep algorithm using EIA and CCNA and identified a PCR cycle threshold cutoff value at 27.5 that predicted EIA toxin positivity with a sensitivity of 99% and negative predictive value of 99% (20). Subsequent studies by other groups have yielded similar findings (21–24). Dual reporting of the standard qualitative PCR result based on the manufacturer threshold and a predicted toxin result (CT-toxin) was implemented at our institution starting in October 2016 (20). Here, for the first time, we describe the clinical outcome of CT-toxin prediction in adult patients with CT-toxin-positive (CT-toxin+) and PCR-positive/CT-toxin-negative (CT-toxin−) results. The aims of this study were 3-fold: to evaluate whether there was a difference in clinical severity and outcomes based on the CT-toxin result, to determine if reporting CT-toxin along with the qualitative PCR result affected the treatment rate in patients with PCR+/CT-toxin− results, and to determine whether there was a difference in outcomes among CT-toxin− patients by treatment status.
MATERIALS AND METHODS
Ethics.
Per the Stanford University Institutional Review Board (IRB), this was considered a quality improvement project, and IRB approval was waived.
Study design and patient population.
This study was a retrospective, observational cohort study performed at the 613-bed hospital and associated clinics of Stanford Health Care (SHC). This system includes extensive programs for cancer care and hematopoietic stem cell and solid organ transplant programs.
Testing algorithm.
Stool samples meeting institutional criteria for all patients (presence of unformed stool and no prior C. difficile test within 7 days) as well as more specific criteria for inpatients (documented evidence of diarrhea, defined as ≥3 unformed stools in 24 h, and absence of laxative therapy within 48 h) (19) were tested using the Xpert C. diff/Epi tcdB PCR assay (Cepheid, Sunnyvale, CA) according to the manufacturer’s instructions. Prior to intervention implementation, the qualitative PCR result based on the manufacturer’s threshold for detection of the tcdB gene was reported as a positive or negative. Beginning 5 October 2016, the report included the qualitative PCR result and, when positive, was accompanied by the predicted toxin result (CT-toxin). Based on previously reported validation studies performed at our institution, a PCR CT cutoff of ≤27.5 is used to determine predicted toxin positivity; this cutoff resulted in a sensitivity of 99.0% and negative predictive value of 99.2% using EIA (for the detection of fecal free toxins A and/or B) as the reference method (20). When using either EIA or CCNA as the reference standard, the sensitivity is 92.0% with a negative predictive value of 99.0%, and specificity is 78% (20). When the PCR was positive and CT-toxin was negative (PCR+/CT-toxin−), a comment was included that stated “negative toxin indicates colonization. Treatment not indicated. A toxin−/PCR+ result suggests C. difficile colonization rather than true disease. Recent data indicate that only those with toxin-positive results (toxin+/PCR+) need to be treated (13, 14). Toxin result was predicted from the PCR cycle threshold with a sensitivity of 99%.” No direct antimicrobial stewardship intervention was provided, and the ultimate decision on whether CDI treatment was given was left to the discretion of the attending physician.
Data collection.
All inpatient and outpatient adult patients (>18 years of age) with a positive C. difficile PCR from October 2016 to October 2017 were included. Chart review was performed for collection of clinical data at the time of testing, including patient demographics, laboratory data, and whether CDI treatment was administered. Patients were followed for 90 days after the index positive test to determine 30-day all-cause mortality and whether patients had a CT-toxin+ C. difficile result on subsequent testing in the follow-up period. As patients with a PCR+/CT-toxin+ on the index test were considered more likely to have true CDI, a recurrence was defined as resolution of diarrhea followed by a second episode of CT-toxin+ CDI within 90 days. Patients with a negative CT-toxin result on the index test (PCR+/CT-toxin−) were considered more likely to be to be colonized; thus, conversion was defined as resolution of diarrhea followed by a recurrence of diarrhea with a CT-toxin+ result on follow-up testing within 90 days. For patients tested somewhere other than SHC during the follow-up period whose results were available through the linked electronic medical record, a recurrence or conversion was considered present if there was a compatible diarrheal syndrome, they had a documented positive C. difficile test of any type, and CDI treatment was administered. Fulminant CDI was defined as toxic megacolon on imaging or septic shock requiring vasopressors or >4 liters of initial volume resuscitation in the absence of an alternative cause of septic shock, such as documented bacteremia (17). The number of stools per day at testing and time to resolution were abstracted from validated nursing and physician documentation (19); rectal tube output was excluded. Diarrhea was considered resolved when <3 stools/day were recorded. Duration of diarrheal illness was considered missing if there was no clear documentation of the date of resolution. As a conservative estimate, treatment for CDI was considered to have been given if at least 7 days of therapy with intravenous (i.v.) or oral metronidazole, oral vancomycin, fidaxomicin, or a combination of these, was administered. To measure changes in treatment rates in CT-toxin− patients over time, this 13-month study was divided into quarters. Deaths were considered to be CDI associated if the death was a direct or indirect result of the presence of C. difficile based on initial unblinded chart review by infectious diseases specialists with consensus assessment (M. M. Hitchcock and L. S. Tompkins). A separate review of the cause of death by another infectious diseases specialist uninvolved in data collection (C. A. Hogan) and blinded to the C. difficile CT-toxin results was then performed. The assessment of the physician blinded to the results was accepted in cases of disagreement. Directly associated deaths occurred in the setting of fulminant CDI, and indirect deaths resulted from an exacerbation of underlying comorbidities related to the presence of C. difficile.
Statistical analysis.
Statistical testing was performed with the Student’s t test for parametric continuous variables, the Mann-Whitney U test for nonparametric continuous variables, and the chi-square and Fisher exact tests for categorical variables using SPSS v. 24 (IBM Analytics, Armonk, NY). Because of multiple comparisons were performed, Bonferroni’s correction was applied and an α of 0.01 when these methods were used. Univariate logistic regression was used to evaluate the change in treatment rate of CT-toxin− patients over the 4 periods of the study at an α of 0.05. Time-to-event analyses for recurrence of or conversion to a CT-toxin+ result were performed using the Kaplan-Meier method and the log rank test; for these analyses, α levels were set at 0.05. Patients who were lost to follow-up prior to 90 days after testing were right censored in the time-to-event analysis.
RESULTS
Predicted toxin results.
From October 2016 to October 2017, 3,140 Xpert PCR tests were performed, and 483 (16.1%) were positive. One patient who had no clinical data available for review was excluded, leaving 482 total PCR-positive results in 430 unique adult patients (range, 1 to 4 tests per patient) in this analysis. Of these, 282 (58.5%) were CT-toxin+ and 200 (41.5%) were CT-toxin−. When only the first test per patient was considered, 244 (56.7%) had a CT-toxin+ result and 186 (43.3%) had a CT-toxin− result. There were 254 (52.7%) inpatients and 228 (47.3%) outpatients, including 86 (17.8%) patients from the emergency department. Data on the number of stools per day at the time of testing were missing in 19.9% (n = 56) of CT-toxin+ and 24.5% (n = 49) of CT-toxin− patients. These rates were much lower among inpatients (7.6% [n = 11] of 144 CT-toxin+ patients and 5.5% [n = 6] of 110 CT-toxin− patients).
Patient characteristics by CT-toxin.
Patient ages ranged from 18 to 101 years. Clinical characteristics in unique CT-toxin+ and CT-toxin− patients were similar by age, sex, Charlson comorbidity index, testing location, and treatment for CDI in the prior 6 months (Table 1). CT-toxin+ patients were more likely to have received antibiotics in the month prior to testing (80% versus 64%; P < 0.001) (Table 1).
TABLE 1.
Comparison of characteristics of unique patients stratified by CT-toxin result
| Variable | Value |
P value | |
|---|---|---|---|
| CT-toxin negative (n = 186) | CT-toxin positive (n = 244) | ||
| Age (yrs) (mean [SD]) | 57.7 (18.22) | 59.9 (19.35) | 0.25 |
| Male sex (n [%]) | 87 (46.8) | 119 (48.8) | 0.68 |
| Hospitalized in prior month (n [%]) | 116 (62.4) | 165 (67.9) | 0.23 |
| Antibiotics in prior month (n [%]) | 119 (64.0) | 193 (79.8) | <0.001 |
| Treated for C. difficile in prior 6 months (n [%]) | 25 (13.4) | 43 (17.6) | 0.24 |
| Testing location (n [%]) | |||
| Inpatient | 102 (54.8) | 134 (54.9) | 0.13 |
| Outpatient | 47 (25.3) | 77 (31.6) | |
| Emergency department | 37 (19.9) | 33 (13.5) | |
| ICUa testing location (n [%]) | 21 (20.6) | 28 (20.9) | 0.95 |
| Charlson comorbidity index (median [IQR]b) | 4 (2–6) | 4 (2–6) | 0.45 |
| Immunocompromised (n [%]) | |||
| Any | 105 (56.5) | 129 (52.9) | 0.46 |
| BMT or SOTc | 39 (21.0) | 37 (15.2) | 0.12 |
ICU, intensive care unit.
IQR, interquartile range.
BMT, bone marrow transplantation; SOT, solid organ transplantation.
Laboratory and clinical severity at testing.
The average white blood cell (WBC) count was higher in CT-toxin+ patients (12.3 × 103 versus 9.5 × 103 cells/μl; P = 0.001), but serum creatinine and albumin were statistically similar (Table 2). A trend toward a higher number of stools per day at testing in CT-toxin+ patients was observed (6.1 versus 5.6 stools/day; P = 0.06) (Table 2). When the comparison of the initial number of stools per day was limited only to inpatients, the results were similar to those of the entire set (6.0 versus 5.6 stools/day, respectively; P = 0.27). CT-toxin+ patients were more likely to have a NAP1/027 strain (19.5% [n = 55] versus 10% [n = 20]; P = 0.005) (Table 2). Twenty-one patients presented with fulminant CDI, all of whom had CT-toxin+ results; no fulminant cases occurred in CT-toxin− patients (P < 0.001) (Table 2).
TABLE 2.
Comparison of laboratory results and clinical symptoms for all tests at the time of testing and stratified by CT-toxin result
| Variablea | Value |
P value | |
|---|---|---|---|
| CT-toxin negative (n = 200) | CT-toxin positive (n = 282) | ||
| WBC count (103/μl) (mean [SD]) | 9.5 (6.3) | 12.3 (9.7) | 0.001 |
| Creatinine (mg/dl) (mean [SD]) | 1.05 (0.70) | 1.08 (0.78) | 0.68 |
| Albumin (g/liter) (mean [SD]) | 2.69 (0.80) | 2.73 (0.70) | 0.54 |
| Initial no. of stools/day (mean [SD]) | |||
| All locations | 5.6 (2.9) | 6.1 (3.1) | 0.06 |
| Inpatient only | 5.6 (2.8) | 6.0 (3.0) | 0.27 |
| Fulminant CDI (n [%]) | 0 (0) | 21 (7.4) | <0.001 |
| NAP1/027 strain (n [%]) | 20 (10) | 55 (19.5) | 0.005 |
WBC, white blood cell; CDI, Clostridioides difficile infection.
Outcomes by CT-toxin result.
Outcomes stratified by CT-toxin result are shown in Table 3. Treatment rates were significantly different based on CT-toxin result, as 98.2% (n = 274) of CT-toxin+ patients were treated compared to 51.5% (n = 103) of CT-toxin− patients (P < 0.001) (Table 3). For CT-toxin− patients, there was no difference in the overall trend in treatment rate by study quarter when evaluated by logistic regression across the entire study period, though the treatment rate was significantly lower in quarter 4 than in quarter 1 (37.7% versus 58.3%, respectively; P = 0.04) (Fig. 1). Diarrhea resolved earlier in CT-toxin− patients (2.2 versus 3.2 days; P = 0.001) (Table 3) despite the lower rate of CDI treatment. When the comparison was limited only to inpatients, where missing data rates were lower (18.8% [n = 27] of 144 CT-toxin+ and 14.5% [n = 16] of 110 CT-toxin− patients), the time to resolution was still shorter in CT-toxin− patients, but the difference was no longer significant at the α level of 0.01 (2.1 versus 2.8 days; P = 0.02). The rate of a CT-toxin+ result on follow-up testing within 90 days was significantly higher in patients with an index CT-toxin+ result than in patients with a CT-toxin− result (17.4% versus 8.0%, respectively; P = 0.003) (Table 3). Time-to-event analysis showed a significant difference as well (P = 0.006) (Fig. 2). None of these patients developed fulminant CDI in follow-up. Thirty-day all-cause mortality was higher in CT-toxin− patients, though this did not reach statistical significance (11.1% versus 6.8%; P = 0.1) (Table 3). Only a single death (4.5%; n = 22) (Table 3) in the CT-toxin− group was potentially indirectly related to CDI, though there was no evidence of fulminant CDI, and the patient received treatment. All others were due to underlying comorbidities or acute organ failure not related to CDI (see Table S1 in the supplemental material). In contrast, 47% of the deaths in the CT-toxin+ patients were CDI associated (n = 19) (Table 3).
TABLE 3.
Comparison of outcomes for all tests stratified by CT-toxin result
| Variable | Value |
P value | |
|---|---|---|---|
| CT-toxin negative (n = 200) | CT-toxin positive (n = 282) | ||
| CDIa treatment administered (n [%]) | 103 (51.5) | 274 (98.2) | <0.001 |
| Days to diarrheal resolution (mean [SD]) | |||
| All locations | 2.2 (1.7) | 3.2 (2.7) | 0.001 |
| Inpatient only | 2.1 (1.7) | 2.8 (2.6) | 0.02 |
| Recurrence of or conversion to CT-toxin+ result within 90 days (n [%]) | 16 (8.0) | 49 (17.4) | 0.003 |
| 30-day all-cause mortality (n [%]) | 22 (11.1) | 19 (6.8) | 0.10 |
| CDI-associated mortality (n [%]) | 1 (4.5) | 9 (47.4) | |
CDI, Clostridioides difficile infection.
FIG 1.
Treatment rate in CT-toxin− patients by study period. Logistic regression was used to assess treatment rate in CT-toxin− patients during each quarter of the study over the entire period and showed no statistically significant difference in treatment rate in CT-toxin− patients (P = 0.13), though the difference between quarter 1 and quarter 4 was significantly different (P = 0.04).
FIG 2.
Time-to-event analysis of the rate of recurrence in CT-toxin+ patients and conversion to a CT-toxin+ result in CT-toxin− patients within 90 days of the index test. CT-toxin+ patients were significantly more likely than CT-toxin− patients to have a CT-toxin+ result on follow-up testing by the log rank test (P = 0.006). Rates are reported as decimal proportions.
CT-toxin-negative patients only.
(i) Characteristics of CT-toxin-negative patients by treatment status. Patient characteristics among 99 treated and 87 untreated unique patients with CT-toxin− results are shown in Table 4. Patients were similar by age, sex, prior hospitalization, antibiotic exposure, prior CDI treatment, location, and Charlson comorbidity index (Table 4). A higher proportion of the treated patients were immunocompromised, though the difference was nonsignificant (62.6% versus 49.1%; P = 0.07); however, significantly more transplant patients were in the treated group (29.3% [n = 29] versus 11.5% [n = 10]; P = 0.003) (Table 4).
TABLE 4.
Characteristics of unique patients with a CT-toxin− result stratified by CDI treatment status
| Variable | Value |
P value | |
|---|---|---|---|
| Treated (n = 99) | Untreated (n = 87) | ||
| Age (yrs) (mean [SD]) | 58.3 (18.8) | 57.1 (17.6) | 0.68 |
| Male sex (n [%]) | 44 (44.4) | 43 (49.4) | 0.50 |
| Hospitalized in prior month (n [%]) | 59 (60.0) | 57 (65.5) | 0.41 |
| Antibiotics in prior month (n [%]) | 68 (68.7) | 51 (58.6) | 0.15 |
| Treated for CDI in prior 6 months (n [%]) | 17 (17.2) | 8 (9.2) | 0.11 |
| Patient location (n [%]) | |||
| Inpatient | 48 (48.5) | 54 (62.1) | 0.06 |
| Outpatient | 32 (32.3) | 15 (17.2) | |
| Emergency department | 19 (19.2) | 18 (18.2) | |
| ICUa at testing (n [%]) | 7 (14.6) | 14 (25.9) | 0.22 |
| Charlson comorbidity index (median [IQR]) | 4 (2–6) | 4 (1–6) | 0.34 |
| Immunocompromised (n [%]) | |||
| Any | 62 (62.6) | 43 (49.4) | 0.07 |
| BMT or SOTb | 29 (29.3) | 10 (11.5) | 0.003 |
ICU, intensive care unit.
BMT, bone marrow transplantation; SOT, solid organ transplantation.
Among 103 treated and 97 untreated events, there was no difference in WBC count, creatinine, albumin, or number of stools per day at the time of testing between treated and untreated patients with CT-toxin− results (Table 5). Treated patients had a higher mean number of stools per day at testing, but this was not significant at the α level of 0.01 (6.2 versus 5.1 stools/day; P = 0.04) (Table 5). As noted above, no patients with a CT-toxin− result had fulminant CDI at presentation.
TABLE 5.
Comparison of laboratory test values and clinical symptoms at the time of testing in all CT-toxin− patients stratified by treatment status
| Variablea | Value |
P value | |
|---|---|---|---|
| Treated (n = 103) | Untreated (n = 97) | ||
| WBC count (103/μl) (mean [SD]) | 10.4 (7.4) | 8.8 (5.0) | 0.11 |
| Creatinine (mg/dl) (mean [SD]) | 1.11 (0.76) | 0.99 (0.64) | 0.29 |
| Albumin (g/liter) (mean [SD]) | 2.63 (0.81) | 2.74 (0.78) | 0.42 |
| Initial no. stools/day (mean [SD]) | |||
| All locations | 6.0 (2.5) | 5.2 (3.2) | 0.09 |
| Inpatient only | 6.2 (2.6) | 5.1 (2.8) | 0.04 |
| NAP1/027 strain (n [%]) | 13 (12.6) | 7 (7.2) | 0.20 |
WBC, white blood cell.
(ii) Outcomes among CT-toxin-negative patients. Of the 103 treated and 97 untreated patients, there was no difference in the rate of conversion to a CT-toxin+ result on follow-up testing within 90 days based on treatment status (8.8% [n = 9] in treated versus 7.2% [n = 7] in untreated patients; P = 0.8) (Table 6); none of these cases represented fulminant CDI. Time-to-event analysis also showed a nonsignificant difference (Fig. 3). This was not due to differential testing rates, as treated and untreated patients underwent repeat C. difficile testing during the follow-up period at similar rates (28.4% [n = 29] versus 29.9% [n = 29], respectively; P = 0.8). Time to diarrheal resolution was shorter in untreated than in treated patients in the full data set (1.5 versus 2.8 days, respectively; P < 0.001) (Table 6); when limited only to inpatients, the result was similar (1.5 [n = 61] versus 2.7 [n = 49] days, respectively; P = 0.001) (Table 6). Thirty-day all-cause mortality was higher in untreated than in treated patients, but the difference was not significant (13.4% [n = 13] versus 8.8% [n = 9], respectively; P = 0.3) (Table 6). As stated above, only one death in a CT-toxin− patient was potentially indirectly attributable to CDI, and this occurred in a patient who received treatment (Table 3).
TABLE 6.
Comparison of outcomes among all CT-toxin− patients stratified by treatment status
| Variable | Value |
P value | |
|---|---|---|---|
| Treated (n = 103) | Untreated (n = 97) | ||
| No. of days to resolution (mean [SD]) | |||
| All locations | 2.8 (1.8) | 1.5 (1.3) | <0.001 |
| Inpatient only | 2.7 (1.8) | 1.5 (1.4) | 0.001 |
| Conversion to CT-toxin+ result within 90 days (n [%]) | 9 (8.8) | 7 (7.2) | 0.80 |
| All-cause 30-day mortality (n [%]) | 9 (8.8) | 13 (13.4) | 0.30 |
FIG 3.
Time-to-event analysis of the rate of conversion to a CT-toxin+ result in treated and untreated CT-toxin− patients within 90 days of the index test. The rate of conversion to a CT-toxin+ result on repeat testing in treated and untreated CT-toxin− patients was not different by the log rank test (P = 0.86). Rates are reported as decimal proportions.
DISCUSSION
This study supports the clinical use of a previously reported CT cutoff to predict the presence or absence of free toxin (20) and demonstrates the impact of dually reporting the qualitative PCR and predicted toxin results. The CT-toxin assay results correlated with degrees of disease severity on presentation and follow-up, as all fulminant CDI cases occurred among CT-toxin+ patients; 47% (n = 19) of deaths among CT-toxin+ patients were CDI associated, and these patients had a higher average WBC count, longer time to diarrhea resolution, and were more likely to have a CT-toxin+ result during the follow-up period than CT-toxin− patients (Tables 2 and 3). The single death potentially indirectly attributable to CDI in a CT-toxin− patient occurred in a patient who received empirical CDI treatment prior to test collection, which may have affected the accuracy of the CT-toxin result (25). The patient did not have fulminant disease and had advanced metastatic cancer as the underlying cause of death. Overall, these results are consistent with prior literature associating free toxin detection to outcomes (13–16).
Prior to the intervention, overtreatment was a significant problem at our institution, as 96% of patients with a positive PCR were treated when stand-alone PCR was the testing modality (19), and there was little consideration given to the possibility of colonization. During the intervention, the overall treatment rate among patients with a positive C. difficile PCR fell to 78%, an absolute reduction of 18% compared to historical institutional data (19). This was driven entirely by the reduction of treatment given to PCR+/CT-toxin− patients, as only 51.5% of these patients were treated during this study period, demonstrating a significant response to a passive intervention. No interim presentations were made to, or data shared with, front-line clinicians during the study period, and so we hypothesize that the reduction in treatment rates of CT-toxin− patients over time reflects clinicians becoming increasingly comfortable with expectant management of mildly symptomatic patients with CT-toxin− results who could be safely observed. The notable clinical difference between CT-toxin+ and CT-toxin− patients despite a fairly stringent testing criteria based on presence of diarrhea and absence of laxative use at our institution also suggests that diagnostic stewardship is necessary but insufficient to avoid overdiagnosis of CDI with PCR (19).
In the comparison of treated and untreated CT-toxin− patients, these patients were similar at testing by most clinical measures, with the most significant difference being a higher proportion of transplant patients among those treated, suggesting that patient history rather than degree of symptoms at the time of testing may have driven the decision to administer CDI treatment. However, it should be noted that the treated patients did have a higher number of stools per day at testing, and while this did not reach statistical significance at the α level of 0.01, this certainly may have played a role in the decision to treat. The prolonged diarrhea observed among treated CT-toxin− patients compared to that among untreated patients raises the possibility that that the antibiotics themselves may have extended the duration of diarrhea. More likely, though, this reflects the slightly higher rate of diarrhea at testing, leading to a longer time to resolution or correct identification of these patients as having true, though mild, CDI. There were no significant differences in the conversion rate to a CT-toxin+ result or 30-day all-cause mortality between treated and untreated CT-toxin− patients, further suggesting colonization or mild CDI that does not definitively require treatment. The nonsignificantly higher 30-day all-cause mortality among CT-toxin− patients who were not treated is worrisome; however, the numerical difference between the groups was small. None of the deaths in untreated CT-toxin− patients were attributable to CDI, and nearly all patients had a resolution of diarrhea without intervention prior to their passing. In addition, several patients with end-stage cancer were tested immediately prior to their transfer to a facility for hospice care, suggesting that these tests were obtained as required by the accepting facility, which may have artificially increased the apparent mortality rate.
As the majority of laboratories in North America use NAATs to diagnose CDI (26), our findings, along with other studies showing an association between lower CT and disease severity with poor outcomes (22, 23, 27–30), indicate dual reporting of qualitative PCR result along with CT-toxin could be implemented to identify patients who are at higher risk for CDI-related complications and most likely to benefit from CDI treatment while maintaining the high sensitivity (relative to EIA and CCNA) and rapid turnaround time of the qualitative PCR test. This testing algorithm has already shown similar efficacy in a pediatric context (31). However, some have argued that the CT-toxin lacks 100% analytical sensitivity and lacks specificity to differentiate between true CDI and colonization (32, 33), and in a commentary published earlier this year, Sandlund and Wilcox specifically argue for use of alternative testing modalities such as ultrasensitive toxin immunoassays using single-molecule counting technology, which purports to offer a highly sensitive and more specific result than PCR while maintaining rapid turnaround time (34, 35). While a potential option, the ultrasensitive toxin immunoassay currently is not available for use in routine clinical practice. When it does become available, a switch to qualitative ultrasensitive toxin detection is unlikely to obviate clinical decision making and may reduce the ability to identify patients with mild disease or potential C. difficile carriers who may not require therapy but would be targets for infection control or antimicrobial stewardship interventions. Clinical trials would therefore be needed to show the benefit ultrasensitive toxin testing offers to patients with NAAT-positive/EIA-toxin-negative results that do not appear to have CDI-related complications when they are not treated for CDI (13).
There are several limitations to this preliminary study of use of CT-toxin in clinical care. First, as this was a retrospective, single-center observational study, data were only available from medical records, leaving a risk of bias due to missing data or differential loss to follow-up. Attempts to control for this were made by performing a subanalysis with inpatients where data are more complete and right censoring patients in the time-to-event analyses if they were lost to follow-up prior to 90 days. Because this study was not randomized, it is possible that all CT-toxin− patients who had active CDI were properly identified and treated, leading to the appearance of similar outcomes regardless of treatment status. Additional studies are under way to further evaluate the findings presented here. Multicenter, randomized controlled trials would be the ideal study design to evaluate whether CT-toxin− patients benefit from anti-C. difficile therapy. Although highly sensitive, the specificity of CT-toxin is lower than two-step testing. CT-toxin was previously shown to have a false positivity rate of 22% using EIA+ or CCNA+ samples as a reference (20). Whether patients with false-positive CT-toxin results, who harbor a high burden of C. difficile organisms, are at higher risk for CDI-associated complications and thus would benefit from treatment cannot be answered here. Further studies are needed to address this question.
In summary, single-step PCR testing with dual reporting of qualitative PCR and CT-toxin results was shown to discriminate between degrees of CDI disease severity at diagnosis, consistent with prior literature, and had a significant impact on clinical care by reducing CDI therapy among CT-toxin− patients without evidence of clear adverse effects seen among untreated CT-toxin− patients in short-term follow-up.
Supplementary Material
ACKNOWLEDGMENTS
We thank Tanya Walston for assistance with electronic reports from the Xpert instrument.
M.M.H. was supported by the National Institutes of Health (NIH; T32 AI 052073-11 A1 and T32 AI 07502-22). Significant funding support was also provided by L.S.T. from her personal account.
Footnotes
Supplemental material for this article may be found at https://doi.org/10.1128/JCM.01288-19.
REFERENCES
- 1.Bagdasarian N, Rao K, Malani PN. 2015. Diagnosis and treatment of Clostridium difficile in adults: a systematic review. JAMA 313:398–408. doi: 10.1001/jama.2014.17103. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Lessa FC, Mu Y, Bamberg WM, Beldavs ZG, Dumyati GK, Dunn JR, Farley MM, Holzbauer SM, Meek JI, Phipps EC, Wilson LE, Winston LG, Cohen JA, Limbago BM, Fridkin SK, Gerding DN, McDonald LC. 2015. Burden of Clostridium difficile infection in the United States. N Engl J Med 372:825–834. doi: 10.1056/NEJMoa1408913. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Pimentel MP, Klompas M, Kachalia A. 2018. Quality measurement for Clostridium difficile infection: turning lemons into lemonade. BMJ Qual Saf 27:414–416. doi: 10.1136/bmjqs-2018-007806. [DOI] [PubMed] [Google Scholar]
- 4.Martin JS, Monaghan TM, Wilcox MH. 2016. Clostridium difficile infection: epidemiology, diagnosis and understanding transmission. Nat Rev Gastroenterol Hepatol 13:206–216. doi: 10.1038/nrgastro.2016.25. [DOI] [PubMed] [Google Scholar]
- 5.Goldenberg SD, Price NM, Tucker D, Wade P, French GL. 2011. Mandatory reporting and improvements in diagnosing Clostridium difficile infection: an incompatible dichotomy? J Infect 62:363–370. doi: 10.1016/j.jinf.2011.03.007. [DOI] [PubMed] [Google Scholar]
- 6.Gould CV, Edwards JR, Cohen J, Bamberg WM, Clark LA, Farley MM, Johnston H, Nadle J, Winston L, Gerding DN, McDonald LC, Lessa FC, Clostridium difficile Infection Surveillance Investigators, Centers for Disease Control and Prevention . 2013. Effect of nucleic acid amplification testing on population-based incidence rates of Clostridium difficile infection. Clin Infect Dis 57:1304–1307. doi: 10.1093/cid/cit492. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Moehring RW, Lofgren ET, Anderson DJ. 2013. Impact of change to molecular testing for Clostridium difficile infection on healthcare facility-associated incidence rates. Infect Control Hosp Epidemiol 34:1055–1061. doi: 10.1086/673144. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Musher DM, Stager C. 2012. Diagnosis of Clostridium difficile infection. Clin Infect Dis 54:1675–1676. doi: 10.1093/cid/cis259. [DOI] [PubMed] [Google Scholar]
- 9.Koo HL, Van JN, Zhao M, Ye X, Revell PA, Jiang ZD, Grimes CZ, Koo DC, Lasco T, Kozinetz CA, Garey KW, DuPont HL. 2014. Real-time polymerase chain reaction detection of asymptomatic Clostridium difficile colonization and rising C. difficile-associated disease rates. Infect Control Hosp Epidemiol 35:667–673. doi: 10.1086/676433. [DOI] [PubMed] [Google Scholar]
- 10.Dionne LL, Raymond F, Corbeil J, Longtin J, Gervais P, Longtin Y. 2013. Correlation between Clostridium difficile bacterial load, commercial real-time PCR cycle thresholds, and results of diagnostic tests based on enzyme immunoassay and cell culture cytotoxicity assay. J Clin Microbiol 51:3624–3630. doi: 10.1128/JCM.01444-13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Leslie JL, Cohen SH, Solnick JV, Polage CR. 2012. Role of fecal Clostridium difficile load in discrepancies between toxin tests and PCR: is quantitation the next step in C. difficile testing? Eur J Clin Microbiol Infect Dis 31:3295–3299. doi: 10.1007/s10096-012-1695-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Naaber P, Stsepetova J, Smidt I, Rätsep M, Kõljalg S, Lõivukene K, Jaanimäe L, Löhr IH, Natås OB, Truusalu K, Sepp E. 2011. Quantification of Clostridium difficile in antibiotic-associated-diarrhea patients. J Clin Microbiol 49:3656–3658. doi: 10.1128/JCM.05115-11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Polage CR, Gyorke CE, Kennedy MA, Leslie JL, Chin DL, Wang S, Nguyen HH, Huang B, Tang YW, Lee LW, Kim K, Taylor S, Romano PS, Panacek EA, Goodell PB, Solnick JV, Cohen SH. 2015. Overdiagnosis of Clostridium difficile infection in the molecular test era. JAMA Intern Med 175:1792–1801. doi: 10.1001/jamainternmed.2015.4114. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Planche TD, Davies KA, Coen PG, Finney JM, Monahan IM, Morris KA, O'Connor L, Oakley SJ, Pope CF, Wren MW, Shetty NP, Crook DW, Wilcox MH. 2013. Differences in outcome according to Clostridium difficile testing method: a prospective multicentre diagnostic validation study of C. difficile infection. Lancet Infect Dis 13:936–945. doi: 10.1016/S1473-3099(13)70200-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Baker I, Leeming JP, Reynolds R, Ibrahim I, Darley E. 2013. Clinical relevance of a positive molecular test in the diagnosis of Clostridium difficile infection. J Hosp Infect 84:311–315. doi: 10.1016/j.jhin.2013.05.006. [DOI] [PubMed] [Google Scholar]
- 16.Beaulieu C, Dionne LL, Julien AS, Longtin Y. 2014. Clinical characteristics and outcomes of patients with Clostridium difficile infection diagnosed by PCR versus a three-step algorithm. Clin Microbiol Infect 20:1067–1073. doi: 10.1111/1469-0691.12676. [DOI] [PubMed] [Google Scholar]
- 17.McDonald LC, Gerding DN, Johnson S, Bakken JS, Carroll KC, Coffin SE, Dubberke ER, Garey KW, Gould CV, Kelly C, Loo V, Shaklee Sammons J, Sandora TJ, Wilcox MH. 2018. Clinical practice guidelines for Clostridium difficile infection in adults and children: 2017 update by the Infectious Diseases Society of America (IDSA) and Society of Healthcare Epidemiology (SHEA). Clin Infect Dis 66:e1–e48. doi: 10.1093/cid/cix1085. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Debast SB, Bauer MP, Kuijper EJ, European Society of Clinical Microbiology and Infectious Diseases . 2014. European Society of Clinical Microbiology and Infectious Diseases: update of the treatment guidance document for Clostridium difficile infection. Clin Microbiol Infect 20:1–26. doi: 10.1111/1469-0691.12418. [DOI] [PubMed] [Google Scholar]
- 19.Truong CY, Gombar S, Wilson R, Sundararajan G, Tekic N, Holubar M, Shepard J, Madison A, Tompkins L, Shah N, Deresinski S, Schroeder LF, Banaei N. 2017. Real-time electronic tracking of diarrheal episodes and laxative therapy enables verification of Clostridium difficile clinical testing criteria and reduction of Clostridium difficile infection rates. J Clin Microbiol 55:1276–1284. doi: 10.1128/JCM.02319-16. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Senchyna F, Gaur RL, Gombar S, Truong CY, Schroeder LF, Banaei N. 2017. Clostridium difficile PCR cycle threshold predicts free toxin. J Clin Microbiol 55:2651–2660. doi: 10.1128/JCM.00563-17. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Crobach MJT, Duszenko N, Terveer EM, Verduin CM, Kuijper EJ. 2018. Nucleic acid amplification test quantitation as predictor of toxin presence in Clostridium difficile infection. J Clin Microbiol 56:e01316-17. doi: 10.1128/JCM.01316-17. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Garvey MI, Bradley CW, Wilkinson MAC, Holden E. 2017. Can a toxin gene NAAT be used to predict toxin EIA and the severity of Clostridium difficile infection? Antimicrob Resist Infect Control 6:127. doi: 10.1186/s13756-017-0283-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Kamboj M, Brite J, McMillen T, Robilotti E, Herrera A, Sepkowitz K, Babady NE. 2018. Potential of real-time PCR cycle threshold (CT) to predict presence of free toxin and clinically relevant C. difficile infection (CDI) in patients with cancer. J Infect 76:369–375. doi: 10.1016/j.jinf.2017.12.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Kim HN, Kim H, Moon HW, Hur M, Yun YM. 2018. Toxin positivity and tcdB gene load in broad-spectrum Clostridium difficile infection. Infection 46:113–117. doi: 10.1007/s15010-017-1108-y. [DOI] [PubMed] [Google Scholar]
- 25.Sunkesula VCK, Kundrapu S, Muganda C, Sethi AK, Donskey CJ. 2013. Does empirical Clostridium difficile infection (CDI) therapy result in false-negative CDI diagnostic test results? Clin Infect Dis 57:494–500. doi: 10.1093/cid/cit286. [DOI] [PubMed] [Google Scholar]
- 26.Kamboj M, Brite J, Aslam A, Kennington J, Babady NE, Calfee D, Furuya Y, Chen D, Augenbraun M, Ostrowsky B, Patel G, Mircescu M, Kak V, Tuma R, Karre TA, Fry DA, Duhaney YP, Moyer A, Mitchell D, Cantu S, Hsieh C, Warren N, Martin S, Willson J, Dickman J, Knight J, Delahanty K, Flood A, Harrington J, Korenstein D, Eagan J, Sepkowitz K. 2018. Artificial differences in Clostridium difficile infection rates associated with disparity in testing. Emerg Infect Dis 24:584–587. doi: 10.3201/eid2403.170961. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Davies KA, Planche T, Wilcox MH. 2018. The predictive value of quantitative nucleic acid amplification detection of Clostridium difficile toxin gene for faecal sample toxin status and patient outcome. PLoS One 13:e0205941. doi: 10.1371/journal.pone.0205941. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Jazmati N, Hellmich M, Ličanin B, Plum G, Kaasch AJ. 2016. PCR cycle threshold predicts the course of Clostridium difficile infection. Clin Microbiol Infect 22:e7–e8. doi: 10.1016/j.cmi.2015.09.012. [DOI] [PubMed] [Google Scholar]
- 29.Origüen J, Orellana MÁ, Fernández-Ruiz M, Corbella L, San Juan R, Ruiz-Ruigómez M, López-Medrano F, Lizasoain M, Ruiz-Merlo T, Maestro-de la Calle G, Parra P, Villa J, Delgado R, Aguado JM. 2019. Toxin B PCR amplification cycle threshold adds little to clinical variables for predicting outcomes in Clostridium difficile infection: a retrospective cohort study. J Clin Microbiol 57:e01125-18. doi: 10.1128/JCM.01125-18. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Reigadas E, Alcalá L, Valerio M, Marín M, Martin A, Bouza E. 2016. Toxin B PCR cycle threshold as a predictor of poor outcome of Clostridium difficile infection: a derivation and validation cohort study. J Antimicrob Chemother 71:1380–1385. doi: 10.1093/jac/dkv497. [DOI] [PubMed] [Google Scholar]
- 31.Schwenk HT, Bio LL, Kruger JF, Banaei N. 22 November 2018. Clinical impact of Clostridium difficile PCR cycle-threshold-predicted toxin reporting in pediatric patients. J Pediatric Infect Dis Soc doi: 10.1093/jpids/piy117. [DOI] [PubMed] [Google Scholar]
- 32.Wilmore S, Goldenberg SD. 2018. Potential of real-time PCR threshold cycle (CT) to predict presence of free toxin and clinically relevant C. difficile infection in patients with cancer: a reply. J Infect 76:424–426. doi: 10.1016/j.jinf.2018.01.001. [DOI] [PubMed] [Google Scholar]
- 33.Pollock NR, Banz A, Chen X, Williams D, Xu H, Cuddemi CA, Cui AX, Perrotta M, Alhassan E, Riou B, Lantz A, Miller MA, Kelly CP. 2019. Comparison of Clostridioides difficile stool toxin concentrations in adults with symptomatic infection and asymptomatic carriage using an ultrasensitive quantitative immunoassay. Clin Infect Dis 68:78–86. doi: 10.1093/cid/ciy415. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Sandlund J, Wilcox MH. 2019. Ultrasensitive detection of Clostridium difficile toxins reveals suboptimal accuracy of toxin gene cycle thresholds for toxin predictions. J Clin Microbiol 57:e01885-18. doi: 10.1128/JCM.01885-18. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Sandlund J, Bartolome A, Almazan A, Tam S, Biscocho S, Abusali S, Bishop JJ, Nolan N, Estis J, Todd J, Young S, Senchyna F, Banaei N. 2018. Ultrasensitive detection of Clostridioides difficile toxins A and B by use of automated single-molecule counting technology. J Clin Microbiol 56:e00908-18. doi: 10.1128/JCM.00908-18. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.