Yield of Vancomycin-Resistant Enterococci and Multidrug-Resistant Enterobacteriaceae from Stools Submitted for Clostridium difficile Testing Compared to Results from a Focused Surveillance Program

Abstract

It has been suggested that a method of performing surveillance for vancomycin-resistant enterococci (VRE) is to screen specimens submitted for Clostridium difficile testing. We compared this approach to our focused surveillance program of high-risk units during October 1997 to compare the yield of VRE and multidrug-resistant Enterobacteriaceae (MDRE) with both methods. Of the stools submitted for C. difficile testing, 14% were positive for VRE or MDRE, whereas rectal swabs from routine surveillance yielded 11% VRE- or MDRE-positive results. Although stools submitted for C. difficile testing resulted in a higher percentage of positive cultures, 14 VRE- and 2 MDRE-positive patients from our high-risk population were missed because many patients had no stool submitted for C. difficile testing. Therefore, while screening stools submitted for C. difficile testing cannot replace our focused surveillance program, it appears advantageous to assess these stools at various intervals to detect new patient reservoirs of drug-resistant organisms that may benefit from routine surveillance cultures.

Within the last decade, the prevalence of colonization and infection due to vancomycin-resistant enterococci (VRE) has increased in health care institutions across the United States (1). In 1995, the Centers for Disease Control and Prevention recommended that in hospitals where VRE have not yet been detected, periodic screening should be performed, with the microbiology laboratory deemed “a first line of defense” against the spread of these organisms (3). However, many hospitals lack and/or have not yet prioritized the resources necessary to perform surveillance for VRE in addition to other infection control activities.

Culture of stools submitted for Clostridium difficile detection has been proposed as one alternative to actual collection of rectal swabs as a means of VRE surveillance (4, 12; A. L. Leber, J. F. Hindler, E. O. Kato, D. A. Bruckner, and D. A. Pegues, Abstr. 4th Decennial Int. Conf. Nosocomial Healthc.-Assoc. Infect., abstr. P-T2-44; 2000; M. T. Seville, R. B. Carey, and J. P. O'Keefe, Abstr. 35th Intersci. Conf. Antimicrob. Agents Chemother., abstr. J-75, 1995). This approach provides a suitable specimen (stool) for detecting VRE, as well as screening populations of patients exposed to many risk factors identified with VRE infection (5). We tested this approach by screening stools submitted for detection of C. difficile in October 1997 and compared the recovery of VRE by this approach to that of our formal surveillance program, which targets patients especially at risk for acquiring VRE infection. Additionally, we tested both methods as a screen for the presence of multidrug-resistant Enterobacteriaceae (MDRE).

As part of the routine focused surveillance program, patients in the hematology-oncology, bone marrow and solid-organ transplant, surgical intensive-care, and medical intensive-care units are screened weekly for the presence of VRE and MDRE by the collection and culture of rectal swabs. Patients transferred from other health care institutions where VRE is known to be endemic are also screened upon admission to our hospital. Perirectal swabs are collected from neutropenic patients. To perform the C. difficile screen, all stool specimens accepted for C. difficile culture and toxin testing during the month of October 1997 were used for the study regardless of patient location.

To culture specimens, the surveillance swab and stool specimens were plated to screening agar within 8 h of receipt. The screening agar was prepared in our laboratory. Our VRE screening medium was purposely formulated to allow the growth and selection of both VRE and MDRE. The agar consists of a tryptic soy base with 5% sheep cells plus the following antibiotics as previously described (T. Zembower, D. Peters, D. Dressel, G. Noskin, R. Thompson, and L. Peterson, Prog. Abstr. 35th Annu. Meet. Infect. Dis. Soc. Am., abstr. 743, 1997): vancomycin, 6 μg/ml; amphotericin B, 2 μg/ml; ceftazidime, 2 μg/ml; and clindamycin, 1 μl (VACC medium). This formulation of agar allows the growth of VRE and MDRE while inhibiting most other flora. The screen plates were incubated at 35°C in 5 to 10% CO₂ and examined at 24 and 48 h. Organisms found growing on the VACC medium were Gram stained. The gram-positive catalase-negative cocci and gram-negative oxidase-negative bacilli were identified to the species level by traditional biochemical tests using an agar replicator method (2). Susceptibility testing was performed by the agar dilution method according to NCCLS recommendations (8). We considered an organism to be a member of the MDRE if it was resistant to aztreonam and/or ceftazidime (MIC, >16 μg/ml) and was a member of the family Enterobacteriaceae. VRE were defined as Enterococcus faecium or Enterococcus faecalis isolates resistant to vancomycin (MIC, >6 μg/ml).

One hundred sixty-four stool samples were submitted for C. difficile testing in October 1997 (Table 1). Twenty-three stool specimens (14%) were positive for VRE or MDRE, with two patients harboring both types of organism. Seventeen (10.4%) samples were culture positive for VRE, and 8 (4.9%) samples were positive for MDRE. Of the 17 VRE cultured from stools submitted for C. difficile testing, only 9 had been collected from patients on high-risk units. Of the eight MDRE cultured from stools submitted for C. difficile testing, four were collected from patients on high-risk units. In total, only 79 of 164 (48%) stool samples submitted for C. difficile testing came from high-risk units. However, by screening these stool samples in addition to routine focused surveillance, nine additional patients with VRE- or MDRE-positive cultures were detected (Table 2). Seven of the nine positive patients had never been positive with that drug-resistant organism before, and two were known positive patients from clinical cultures. In the same period, 290 focused surveillance cultures were collected on high-risk units. Thirty-two rectal-swab samples (11%) were positive for VRE or MDRE. Twenty-eight (9.7%) were positive for VRE, and nine (3.1%) were positive for MDRE. Eight patients had cultures that were positive by both methods. Three patients with positive focused surveillance cultures had positive clinical cultures, and five patients with positive C. difficile screens had positive clinical cultures. However, a majority of the patients (29 of 35 [83%]) were found to be harboring drug-resistant organisms by means of one of the screening methods and not through clinical culture. During the study month, our C. difficile positive test result rate was 12.5%.

TABLE 1.

Results of specimen testing from C. difficile samples and routine focused surveillance

Test subject	No. of positive resultsa
	Specimens from routine surveillance	Specimens from stool submitted for C. difficile testing	Both methods
Total no. of tests performed	290	164
Cultures with VRE	28 (9.7)	17 (10.4)
Cultures with MDRE	9 (3.1)	8 (4.9)
Total positive cultures for VRE or MDRE	32 (11)b	23 (14)c
Total no. of patients tested	212	123
Patients with VRE	20 (9.4)	13 (10.6)	6
Patients with MDRE	7 (3.3)	7 (5.7)	4
Total patients positive for VRE or MDRE	24 (11.3)d	17 (13.8)d	8

^aPercent of total in parentheses. 

^bFive cultures harbored both VRE and MDRE. 

^cTwo cultures harbored both VRE and MDRE. 

^dThree patients harbored both VRE and MDRE. 

TABLE 2.

Detection of VRE and MDRE from samples submitted for C. difficile testing according to patient location

Organism and source	No. detected
	High-risk units (total tests = 79)	Non-high-risk units (total tests = 84)
VRE
Positive cultures	9	8
Positive patients	5	8
MDRE
Positive cultures	4	4
Positive patients	3	4

Several studies have been performed to determine the risk factors associated with the acquisition of VRE. They indicate that prior antibiotic exposure is a risk factor for VRE acquisition (4, 5, 7). Because antibiotic exposure also plays a role in the development of C. difficile disease and the epidemiology is similar to that of VRE disease, it has been suggested that screening stool samples submitted for C. difficile culture and/or toxin testing may be an effective way to perform screening for VRE. Three centers have published results examining the prevalence of VRE in stools submitted for C. difficile screening. These centers were able to detect appreciable numbers of VRE by this method, with rates of 19.8, 16.5, and 12% (12; Leber et al., Abstr. 4th Decennial Int. Conf. Nosocomial Healthc.-Assoc. Infect.; Seville et al., Abstr. 35th Intersci. Conf. Antimicrob. Agents Chemother.). This is even higher than the 10.4% we found when screening C. difficile samples (Table 1). This would suggest that testing stools submitted for C. difficile analysis may be an alternative to active surveillance consisting of culturing patients specifically for VRE.

However, studies have suggested that other risk factors, such as the severity of illness and increased length of stay, can predispose patients to VRE colonization (7, 10). These factors are associated with patients who are likely to be found in a high-risk group, such as those we routinely survey by actively collecting rectal or perirectal swabs to detect important nosocomial pathogens. In our study, if we had focused our surveillance only on the specimens submitted for C. difficile testing, we would have screened only 79 high-risk patients as opposed to the 290 covered by our routine focused surveillance program and would have missed 14 VRE- and 2 MDRE-positive patients from this susceptible patient population. However, 2 of the 16 patients found to be positive only by focused surveillance also had positive clinical cultures, so technically 13 VRE- and 1 MDRE-positive patients would have been missed if only the C. difficile screen had been performed.

We believe it is prudent to use whatever means are available to detect asymptomatic colonized patients in high-risk units (intensive care, transplant, etc.) in order to prevent further spread of drug-resistant organisms. Preventing colonization of patients in high-risk units by drug-resistant bacteria is important because colonization in this group can lead to infections with these organisms, often with adverse consequences (9).

An active, focused surveillance program of weekly cultures on high-risk units can enhance infection control efforts by early detection and containment of patients harboring VRE and MDRE (6). This can actually reduce overall, institution-wide health care costs by helping to prevent nosocomial infections and avoiding the excess cost to the hospital associated with them. An example of this was reported by Stosor and colleagues, who demonstrated that the cost of a hospital stay for a patient with VRE bacteremia is nearly 1.5 times that of a patient with a vancomycin-susceptible enterococcal bacteremia (13). Additionally, studies at our institution have shown that the added annual direct cost of a comprehensive surveillance program ($200,000) is minimal compared to the overall health care dollars saved ($2,000,000) when health care-associated infections are prevented (6).

Because our screening medium is designed to detect MDRE and VRE, we were able to assess stools collected for C. difficile testing for the recovery of MDRE as well as VRE (Table 1). The occurrence of MDRE infection or colonization is less frequent than that of VRE infection or colonization at our hospital. However, as was seen with VRE, C. difficile stool testing recovered more MDRE isolates than routine focused surveillance alone. Together, both screening methods recovered MDRE isolates from 10 patients from our institution in this 1-month period. This highlights the fact that MDRE do colonize high-risk patients and that they are important to consider when screening for drug-resistant bacteria. Clinical microbiology laboratories should begin to identify and report members of the family Enterobacteriaceae as MDRE, primarily to alert physicians so that appropriate antibiotic therapy can be administered if an infection develops but also to monitor for these types of organisms as an infection control risk (11).

In conclusion, the use of stools submitted for C. difficile testing to detect drug-resistant organisms has been shown to yield an appreciable number of VRE and MDRE isolates. This may be a useful method to establish some idea of the prevalence of VRE or MDRE, depending on the culture medium utilized. However, C. difficile stool testing showed a likelihood of surveying only a portion of our high-risk patients, and therefore it could not replace our current method of VRE surveillance. Importantly, in light of the ability of C. difficile stool testing to enhance overall detection of important nosocomial pathogens, it may be advantageous to supplement routine surveillance with C. difficile specimen testing at various intervals to determine new reservoirs of VRE or MDRE that may need ongoing, active surveillance.