Comparison of Simplexa Universal Direct PCR with Cytotoxicity Assay for Diagnosis of Clostridium difficile Infection: Performance, Cost, and Correlation with Disease

Abstract

Simplexa Clostridium difficile universal direct PCR, a real-time PCR assay for the detection of the C. difficile toxin B (tcdB) gene using the 3M integrated cycler, was compared with a two-step algorithm which includes the C. Diff Chek-60 glutamate dehydrogenase (GDH) antigen assay followed by cytotoxin neutralization. Three hundred forty-two liquid or semisolid stools submitted for diagnostic C. difficile testing, 171 GDH antigen positive and 171 GDH antigen negative, were selected for the study. All samples were tested by the C. Diff Chek-60 GDH antigen assay, cytotoxin neutralization, and Simplexa direct PCR. Of 171 GDH-positive samples, 4 were excluded (from patients on therapy or from whom duplicate samples were obtained) and 88 were determined to be true positives for toxigenic C. difficile. Of the 88, 67 (76.1%) were positive by the two-step method and 86 (97.7%) were positive by PCR. Seventy-nine were positive by the GDH antigen assay only. Of 171 GDH antigen-negative samples, none were positive by PCR. One antigen-negative sample positive by the cytotoxin assay only was deemed a false positive based on chart review. Simplexa C. difficile universal direct PCR was significantly more sensitive for detecting toxigenic C. difficile bacteria than cytotoxin neutralization (P = 0.0002). However, most PCR-positive/cytotoxin-negative patients did not have clear C. difficile disease. The estimated cost avoidance provided by a more rapid molecular diagnosis was outweighed by the cost of isolating and treating PCR-positive/cytotoxin-negative patients. The costs, clinical consequences, and impact on nosocomial transmission of treating and/or isolating patients positive for toxigenic C. difficile by PCR but negative for in vivo toxin production merit further study.

INTRODUCTION

Clostridium difficile infection (CDI) is a major cause of nosocomial diarrhea, leading to morbidity and mortality in hospitalized patients (1). Anaerobic culture and bacterial glutamate dehydrogenase (GDH) antigen tests detect the presence of both nontoxigenic and toxigenic strains of C. difficile. Toxigenic culture and PCR for toxin genes detect only the presence of bacteria capable of making toxin; these tests cannot separate active disease from colonization. Tests for toxin production in vivo, e.g., enzyme-linked immunosorbent assay (ELISA) to detect toxins in stool specimens or cell culture to determine cytotoxicity, have been considered insensitive compared to molecular methods (2). In a previous report, we compared our two-step C. difficile algorithm testing for GDH antigen and cytotoxin to one of the first commercially available PCR tests (3). While PCR was significantly more sensitive than toxin ELISA, its results were not statistically different from those of the two-step algorithm, it failed to detect nine cytotoxin-positive stools, and it was more expensive. Furthermore, patients positive by PCR but not for cytotoxin did not appear to have C. difficile disease.

The objective of this study was to compare the performance of the newly available Simplexa C. difficile universal direct PCR (Focus Diagnostics) for detection of the C. difficile toxin B gene with that of our standard two-step method, namely, the C. Diff Chek-60 GDH antigen assay followed by cytotoxin neutralization. In addition, the costs and savings of the projected change from cytotoxin assay to PCR with regard to reagents, labor, contact isolation (CI), and antibiotics were estimated.

MATERIALS AND METHODS

Clinical samples.

Stool samples submitted for C. difficile testing from patients at Yale-New Haven Hospital were entered into the study from August 2012 to October 2012. All samples were stored at 4°C and tested within 24 h of receipt using the C. Diff Chek-60 GDH antigen ELISA as part of the hospital's standard two-step diagnostic routine. All C. difficile GDH antigen-positive samples with sufficient stool available and an equivalent number of GDH antigen-negative stools were selected on each study day. All study samples were tested within 24 h of receipt by cytotoxin neutralization. Aliquots were labeled with study numbers and then tested by Simplexa C. difficile universal direct PCR. PCR was performed by study personnel blind to the results of the two-step method. Only semisolid or liquid stools, one stool per patient per day and no more than two stools per patient in a 7-day period, were included. Formed stools, repeat samples sent on the same day, and samples obtained from patients on C. difficile therapy were excluded.

Two-step method using C. Diff Chek-60 and a cytotoxicity assay.

The C. Diff Chek-60 (TechLab, Blacksburg, VA) was performed according to the manufacturer's instructions. Briefly, 0.1 ml of specimen was added to 0.4 ml of specimen diluent, vortexed, and then centrifuged at 5,000 × g for 10 min. Next 0.05 ml of conjugate solution was added to the test microwells, followed by 0.1 ml of centrifuged specimen. The wells were covered and incubated for 50 min at 37°C and washed with 0.35 ml wash solution (7 cycles). Then 0.1 ml of substrate was added, and wells were incubated at room temperature for 10 min. After addition of 0.05 ml of stop solution, optical density was measured on a microplate reader. A positive result had an optical density of ≥0.080, and a negative result had an optical density of <0.080 using the spectrophotometric dual wavelength 450/620 nm. All antigen results were reported clinically.

All study samples were stored at 4°C and tested by cytotoxicity assay within 24 h of receipt. Stool samples (0.5 ml) were added to 0.5 ml of phosphate-buffered saline with antibiotics (vancomycin, gentamicin, and amphotericin B) and then vortexed, and the toxin was allowed to elute for 5 min. After centrifugation for 10 min in a microcentrifuge, the supernatant was removed and passed through a 0.45-μm-pore-size filter. Then, 20 μl of filtrate was inoculated in duplicate onto foreskin fibroblast monolayers (MRHF cells; BioWhittaker, Walkersville, MD) in 96-well plates prepared weekly in the laboratory. To assess cytotoxicity, serial 10-fold dilutions (1:10 to 1:10,000) were made without antitoxin. For neutralization, C. difficile antitoxin (20 μl; TechLab, Inc., Blacksburg, VA) was added to two wells inoculated with 1:10 and 1:100 dilutions of sample. Thus, after addition of antitoxin, the final dilution in the first culture well was 1:20. Monolayers were read at 4, 24, and 48 h after inoculation using an inverted microscope. A known positive control, run with each assay, was required to show cytotoxicity in the expected range. A positive result consisted of cytotoxicity that was neutralized by C. difficile antitoxin. Results were recorded as the highest dilution showing specific cytotoxicity. Cytotoxin results were reported for GDH antigen-positive samples only, as per the routine clinical protocol.

Simplexa C. difficile universal direct PCR.

The Simplexa C. difficile universal direct PCR (Focus Diagnostics, San Diego, CA) was performed directly on stool specimens according to the manufacturer's instructions. It utilizes real-time PCR to amplify the C. difficile toxin B (tcdB) gene, an internal control, and bifunctional fluorescent probes/primers for the identification of amplified target DNA. The amplification, detection, and interpretation of the assay are done using the 3M integrated cycler instrument. Results were positive, negative, or invalid. Invalid results were due to technical errors or inhibitors in the sample lysate. According to the manufacturer's instructions, lysates of unresolved samples were frozen at −70°C, thawed, and then retested. The entire procedure requires about 2 h, depending on the number of samples in the run. All samples not tested within 24 h were stored at −20°C and tested within 5 days. PCR testing was done without knowledge of two-step test results, and PCR results were not reported clinically.

Discrepant analysis.

Samples positive only for GDH antigen were considered to represent colonization with nontoxigenic strains of C. difficile. Samples positive by two or more tests were accepted as true positives. Discrepant samples were defined as those positive by either the cytotoxicity assay or PCR but not both. Discrepant samples from patients who were on treatment for CDI at the time of sample collection were excluded from analysis. Chart reviews for inpatients with discrepant results were conducted independently by two of the authors to determine antibiotic therapy, use of contact isolation (CI), frequency and persistence of diarrhea, and subsequent C. difficile testing results and disease.

CDI.

CDI was defined as three or more liquid or semisolid stools in less than 24 h, without other explanation; it resolves on therapy, persists if not treated, or shows a characteristic appearance on colonoscopy or by pathological exam (1).

Statistical analysis.

Statistical analysis was performed using McNemar's test and the unpaired t test.

IRB review.

The work was considered routine clinical practice and was deemed exempt from Institutional Review Board (IRB) review.

RESULTS

Performance of the Simplexa PCR.

A total of 342 samples (171 GDH antigen-positive and 171 GDH-negative samples) were initially tested by the Simplexa PCR. Of 342 samples tested by PCR, only 1 (0.3%) had an invalid internal-control result, and its DNA was successfully amplified on repeat testing. Four GDH-positive samples were excluded from the final analysis: two samples because the patients were on treatment for CDI and two because the number of samples exceeded one per day or two in a 7-day period. In the final analysis, 338 stool samples from 300 patients were included. Results are shown in Table 1. Overall, 79 samples were positive by the GDH antigen assay only, 23 were positive by the GDH antigen assay and positive either by the cytotoxin assay or by PCR, and 65 were positive by all three tests. When threshold cycle (C_T) values for samples that were PCR positive/cytotoxin positive (mean C_T, 30.784 ± 3.199) were compared to C_T values for samples that were PCR positive/cytotoxin negative (mean C_T, 34.936 ± 3.100), the difference between the two groups was significant (P < 0.0001, unpaired t test). However, substantial overlap in C_T values was evident (range, 23.5 to 39.8 versus 28.9 to 39.4, respectively).

TABLE 1.

Results for all three methods

No. of specimens (n = 338)	Test result
	GDH antigen assay	Cytotoxin assay	PCR
79	+	−	−
65	+	+	+
2a	+	+	−
21a	+	−	+
1a	−	+b	−
0	−	−	+
170	−	−	−

^aChart reviews were performed for 19 inpatients.

^bFalsely positive for cytotoxin.

Of 171 GDH-negative samples, 1 was positive by the cytotoxin assay only and 170 were negative for all three tests. Particularly notable was the fact that no samples were positive by PCR only. Of the 88 true-positive samples, 67 (76.1%) were positive by the two-step method and 86 (97.7%) were positive by the Simplexa PCR assay. There was a 92.9% concordance of PCR results with the two-step method. Simplexa PCR was significantly more sensitive than cytotoxin neutralization for detection of toxigenic C. difficile bacteria (P = 0.0002, McNemar's test) (Table 2). However, it was not clear that PCR-positive/cytotoxin-negative patients had CDI.

TABLE 2.

Comparison of Simplexa C. difficile universal direct PCR assay with the two-step GDH antigen/cytotoxin algorithm for detection of toxigenic C. difficile

Simplexa direct PCR result	No. of specimens with indicated two-step GDH antigen/cytotoxin algorithm resulta		Total
	Positiveb	Negative
Positive	65	21	86
Negative	2	250	252
Total	67	271	338

^aThe difference in sensitivity between PCR and the two-step algorithm was significant (P = 0.0002, McNemar's test).

^bOne GDH antigen-negative/cytotoxin-positive specimen was considered a false positive and was not reported clinically.

Clinical correlation by chart review.

Twenty-four of the samples were considered discrepant, and 19 of these were from inpatients (16 GDH-positive/PCR-positive/cytotoxin-negative specimens, 2 GDH-positive/PCR-negative/cytotoxin-positive specimens, and 1 GDH-negative/PCR-negative/cytotoxin-positive specimen). Charts for the 19 inpatients were reviewed to assess diarrhea, antibiotic therapy, subsequent C. difficile testing, and CDI (Table 3). Since the Simplexa assay was not validated for clinical use and PCR results were not reported, clinical decisions were based solely on the standard two-step algorithm. All GDH antigen results were reported, followed by cytotoxin results if GDH was positive, as per standard practice.

TABLE 3.

Chart review of inpatients with discrepant cytotoxin and PCR results^c

Test results (no. of patients)	Patient	On Metro or Vanco when stool was obtained	Treatment after positive GDH antigen result was reported	Stool as recorded in chart	C. difficile test result(s) obtained within 30 days after initial test(s)	GDH antigen result (OD)	Cytotoxin result	Simplexa PCR result (CT)
GDH antigen and PCR positive, cytotoxin negative (16)	1	No	None	Loose stools	None	Positive (>3.0)	Negative	Positive (37.4)
	2	No	Metro p.o. for 3 days	1 liquid stool	None	Positive (>3.0)	Negative	Positive (36.1)
	3	No	None	Loose stool	None	Positive (>3.0)	Negative	Positive (37.4)
	4	No	Metro p.o. for 2 days	2 loose stools	None	Positive (>3.0)	Negative	Positive (32.0)
	5	No	None	Liquid stool	None	Positive (>3.0)	Negative	Positive (33.8)
	6	No	Metro p.o., 2 doses	2 loose stools	GDH antigen negative; cytotoxin assay not done (6 days later)	Positive (2.965)	Negative	Positive (36.8)
	7	No	None	3 loose stools	None	Positive (2.996)	Negative	Positive (39.4)
	8	No	None	1 loose stool	None	Positive (0.413)	Negative	Positive (39.4)
	9	No	Metro p.o. for 7 days	Loose stools	GDH antigen positive/cytotoxin negative (3 days later)	Positive (>3.0)	Negative	Positive (30.8)
	10	No	Metro p.o., 4 doses	Diarrhea prior to admission	GDH antigen positive/cytotoxin negative (6 and 25 days later)	Positive (2.178)	Negative	Positive (33.9)
	11	No	None	Not recorded	GDH antigen positive/cytotoxin negative (3 and 6 days later)	Positive (>3.0)	Negative	Positive (31.8)
	12	No	Metro p.o. for 2 days	1 loose stool	None	Positive (>3.0)	Negative	Positive (35.0)
	13a	Ceftrx-Metro	Metro p.o. for 2 days, Vanco p.o. for 1 day	1 loose stool	None	Positive (>3.0)	Negative	Positive (31.6)
	14	No	Metro p.o., 4 doses	1 loose stool	None	Positive (0.564)	Negative	Positive (39.2)
	15b	Cipro-Metro	Metro p.o., 4 doses	3 loose stools	None	Positive (>3.0)	Negative	Positive (38.0)
	16	No	Metro p.o. for 3 days	4 loose stools	None	Positive (>3.0)	Negative	Positive (28.9)
GDH antigen and cytotoxin positive, PCR negative (2)	17	No	Metro p.o. for 7 days	6 loose stools	None	Positive (0.196)	Positive	Negative (>40)
	18	No	Metro p.o. for 5 days	Multiple loose stools	None	Positive (>3.0)	Positive	Negative (>40)
GDH antigen and PCR negative, cytotoxin positive (1)	19	No	None (antigen negative)	3 loose stools	None	Negative (0.014)	Positive	Negative (>40)

^aSmall bowel obstruction.

^bSurgical-site infection.

^cMetro, metronidazole; Ceftrx, ceftriaxone; Cipro, ciprofloxacin; Vanco, vancomycin; C_T, cycle threshold; OD, optical density reading (≥0.080 is positive); p.o., per os.

Of the 16 GDH-positive/PCR-positive but ultimately cytotoxin-negative inpatients, 6 were not treated, 9 were treated for up to 3 days until the cytotoxin results returned negative, and 1 was treated for 7 days. Of the 16, only 3 met the requirement of 3 diarrheal stools within a 24-h period. Only four had repeat testing within 30 days, and none were cytotoxin positive. One (6.3%) had symptomatic C. difficile infection 6 weeks later, with positive GDH antigen and cytotoxin results and diarrhea.

Two patients positive by the two-step algorithm but negative by PCR had 3 or 4 loose stools per 24 h and were treated according to clinical guidelines. No further samples were sent in 60 days of follow-up. One patient was both GDH antigen and PCR negative but had a low positive cytotoxin result. Only the negative GDH result was reported (not the positive cytotoxin result) as per standard clinical protocol; in a nonstudy situation, the cytotoxin assay would not have been done. The patient was lactose intolerant and had been given milk in error. This was considered a false-positive cytotoxin result.

Since most patients with stools positive by PCR but not for cytotoxin did not meet the criteria of 3 loose stools in 24 h and symptoms resolved with minimal or no therapy, it is highly likely that most if not all were carriers of toxigenic C. difficile.

Cost per reportable result.

The average turnaround times, reagents, materials, and labor costs per test in our laboratory are given in Table 4. If used as the sole test, PCR is significantly more expensive than the two-step algorithm. Costs can be reduced significantly if the GDH antigen ELISA is used as the screening test and PCR is performed only on GDH antigen-positive samples.

TABLE 4.

Time to result and estimated costs for different methods

Test	Turnaround time (h)a	Cost of materials ($)	Cost of labor ($)	Total ($)
GDH antigen screen	2–24	4.73	2.50	7.23
Cytotoxin neutralizationb,c	6–48	11.76	6.46	18.22
Simplexa direct PCRb	2–24	36.05	6.25	42.30

^aBased on one PCR or antigen assay run per day.

^bFour samples per run; kit pricing is volume dependent.

^cWith the 2-step method, GDH antigen-negative results are reported within 2 to 24 h. GDH-positive results are then tested for cytotoxin, and cytotoxin results are reported after 4, 24, and 48 h of incubation. Negative results require 48 h.

Impact of PCR on infection control and antibiotic costs.

Though the cost of laboratory testing is higher for PCR, potential savings by avoidance of 2 days of contact isolation (CI) and antibiotic therapy were investigated for antigen-positive patients who were ultimately found to be cytotoxin negative.

Analysis of GDH antigen-positive specimens revealed 137 unique inpatients, and of these, 76 were GDH antigen positive but ultimately cytotoxin negative. Of these 76, 38 (50%) were already on CI prior to C. difficile testing due to multidrug-resistant organisms; the remaining 38 (50%) were put on contact precautions as a consequence of the positive GDH antigen result. Empirical antibiotic therapy was given to only 41 of 76 patients (54%) for a median of 2 days and a mean of 3 days (range, 1 to 18 days). Metronidazole (Flagyl) was given to 33 (80%) and oral vancomycin to 8 (20%) patients. If PCR was reported at the same time as the GDH antigen-positive result, these costs could be avoided for patients who were ultimately found to be negative. The cost of CI was estimated at $61.44 per person per day, and the average cost of antibiotic therapy was $20.96 per patient per day. However, only 60 of 76 cytotoxin-negative patients were PCR negative, and 16 were PCR positive. Thus, if same-day PCR results replaced the 2-day cytotoxin assay results, the cost savings for 2 days of CI and antibiotic therapy for PCR-negative patients would be outweighed by the costs incurred for 10 days of CI and antibiotic therapy for the additional PCR-positive patients (Table 5). Length of stay would also likely increase.

TABLE 5.

Estimated impact of same-day PCR result on CI and antibiotic usage for inpatients

PCR result for GDH antigen-positive/cytotoxin-negative inpatients	Impact	Total no. of patients	No. (%) of patients requiring CI	Days of CI required	Total CI days	CI cost per day ($)a	No. (%) of patients treated for CDI	No. of antibiotic treatment days per patient	Total antibiotic treatment days	Antibiotic cost ($)b
Negative (same day)	Cost avoidance	60	30 (50)	2	60	3,686.40	32 (54)	2	64	1,341.44
Positive	Cost incurred	16	8 (50)	10	80	4,825.20	16	10	160	3,353.60

^aCI cost per day, $61.44.

^bAntibiotic cost per day, $20.96 (80% were treated with metronidazole for $1.20; 20% were treated with oral vancomycin for $100).

DISCUSSION

Ours is the first study comparing the performance characteristics of the commercial Focus Diagnostics Simplexa C. difficile universal direct PCR and the two-step GDH antigen ELISA/cytotoxin neutralization protocol. Simplexa PCR was more sensitive than the two-step algorithm in detecting toxigenic bacteria (97.7% versus 76.1%, respectively; P = 0.0002). The concordance of the Simplexa PCR assay with cytotoxicity results (92.9%) is similar to that of other commercial PCR assays (3–5). The Simplexa direct assay produced results in approximately 2 h, compared to 6 to 48 h for cytotoxin neutralization, was simple to perform, and was done outside a molecular laboratory.

The C. Diff Chek-60 assay for detection of GDH antigen did not miss any PCR-positive samples, thus confirming its value as an initial screening test in a two-step protocol, using a toxin assay or PCR as the second step (2, 6–8). In a microplate ELISA format, it is a relatively inexpensive test that takes 1 h to complete. Furthermore, since it is read in a spectrophotometer, it may be more sensitive than GDH antigen lateral-flow tests read by eye, but this has not been studied.

Twenty-one PCR-positive/cytotoxin-negative samples were detected, and of these, 16 were from inpatients whose charts were available for review. Since PCR results were not reported to clinicians, these samples were reported as GDH positive and ultimately cytotoxin negative. Investigation revealed that 10 patients received brief treatment while their cytotoxin results were pending. While unlikely, it is possible that clinical outcomes were affected. Six received no treatment for CDI, with no apparent adverse consequences. According to the nurses' notes, only 3 of the 16 had diarrhea that met the criteria for testing at the time that the samples were sent. In a follow-up period of 30 days, only 4 of 16 patients had repeat C. difficile testing and all remained cytotoxin negative. Our previous study of the BD GeneOhm PCR kit also found that the 18 patients positive by PCR only did not have CDI and did well (3).

Two patients were PCR negative/cytotoxin positive. These two patients met the criteria of 3 or more diarrheal stools per day and were treated. In our institution, cytotoxin testing is performed on site by experienced virology laboratory personnel, using cell culture plates freshly prepared on site, and samples are tested starting at a low dilution (i.e., 1:20). Thus, our cytotoxicity results are likely more sensitive than those obtained using higher starting dilutions of 1:50 (5) or 1:100 (4), commercially prepared cell culture, or samples shipped a distance, during which toxin can degrade. Of note, Simplexa PCR missed fewer cytotoxin-positive samples (n = 2) than the BD GeneOhm test (n = 9) in our previous study (3). Falsely negative PCR results can be due to low levels of bacteria, inhibitors, or genetic variance leading to primer or probe mismatch (3, 9, 10).

Our study contributes to the growing concerns regarding tests that target bacteria rather than toxin. Both toxigenic bacterial culture and toxin gene PCR detect toxigenic bacteria, while toxin assays detect in vivo toxin production. More positive results are detected by toxigenic culture and PCR than by toxin assays, and PCR can be associated with a >50% increase in the rate of incidence (11). However, the reported greater sensitivity of PCR may be misleading, as colonized patients are also detected. Furthermore, toxigenic bacteria can remain detectable in stool for weeks after treatment (12).

The positive predictive value of a PCR result for CDI depends on having significant diarrhea (13). Too often, samples are sent after only one or two loose stools. In addition, carriers can have diarrhea due to other causes, such as norovirus gastroenteritis, tube feedings, laxatives, antibiotics, or other causes, and toxigenic C. difficile is merely a bystander (14, 15). Carriers have been increasingly recognized in asymptomatic individuals of all ages, including 10% or more of hospitalized patients and up to 51% of patients in a long-term-care facility (16–21). Attempting to eradicate carriage by treating with antibiotics may have adverse effects on the host-pathogen balance, increasing the risk of subsequent diarrhea, and is not currently recommended (1, 16, 22, 23). Thus, separating colonized patients from those with true CDI is essential to avoid treating patients unnecessarily. To do this, a toxin assay is essential.

Importantly, several recent studies have found that only toxin-positive stools, and not PCR-positive/toxin-negative stools, are associated with high morbidity and mortality as well as longer hospital stays (11, 24, 25). A large multicenter study found that cytotoxin positivity, and not GeneXpert PCR or toxigenic culture positivity, correlated with clinical outcome and thus best defined true cases of CDI (24). Consequently, they recommended against using PCR alone to diagnose CDI due to its low positive predictive value. Instead, they recommend a new diagnostic category of C. difficile excretor for carriers whose diarrhea is probably not due to CDI but who can potentially transmit infection.

It has been postulated that even asymptomatic carriers are a cause of nosocomial transmission and should be placed in isolation (21, 26, 27). A recent study used genotyping to investigate the role of asymptomatic carriers in transmission in a ward and concluded that carriers played an important role (27). However, it is not yet known whether carrier identification and isolation will reduce transmission and improve patient outcomes.

Commercial PCR test kits are more expensive than toxin assays and may require purchase of expensive equipment (3). Thus, compensatory savings from adoption of PCR were sought for the wards. It was anticipated that cost avoidance from a same-day diagnosis would outweigh the costs of PCR reagents. However, by our estimation, isolation and treatment of the additional patients positive by PCR lead to a net increase in costs. Unfortunately, we were not able to assess whether implementing PCR for toxigenic C. difficile will reduce nosocomial transmission and disease and thus generate long-term savings. Such studies are urgently needed.

As in our previous report (3), inappropriate stool sample submission for C. difficile testing, including multiple samples from the same patient on the same day, was common. Thirteen percent of samples were solid stools, which were rejected. Chart review revealed submission of samples from patients with minimal diarrhea and from patients who were already on treatment. Though same-day samples and samples from patients on therapy were excluded from the study, these problems illustrate the difficulty in both enforcing testing guidelines and interpreting a positive PCR result. Nurses observe the stool and collect the specimens, but physicians receive the test results and act on them. To avoid unnecessary treatment and reduce costs of PCR testing and CI, clinicians need to limit C. difficile testing and treatment to patients with a reasonable probability of having disease, such as those patients having 3 or more loose stools per day for 1 to 2 days (1, 28). Accomplishing this is a challenge.

There were a number of limitations of this study. All GDH antigen-positive samples submitted on study days that met study criteria were included, but only a subset of GDH antigen-negative samples were included. Cytotoxin and PCR results were not confirmed by toxigenic culture but were accepted as true if GDH was also positive. C. difficile ribotypes were not determined, which may influence clinical disease determination as well as performance of the GDH antigen assay. The number of stools as recorded in the chart may have been underestimated. Colonoscopy was not performed to establish the presence or absence of C. difficile disease. Rather we relied on cytotoxin to separate CDI from colonization.

In conclusion, Simplexa PCR detected significantly more positive samples than a two-step method for diagnosis of toxigenic C. difficile. The integrated cycler has a small footprint and can be obtained by reagent rental rather than capital purchase. Additionally, the test can be performed outside a molecular laboratory. However, as in our prior study of a different toxin gene PCR kit, BD GeneOhm, samples positive by PCR and negative by cytotoxicity assay did not clearly indicate disease, yet their detection by PCR will lead to both contact isolation and treatment. The costs and consequences of identifying, isolating, and treating C. difficile carriers, either with or without significant diarrhea, should be a high priority for further study.