Comparison of Five Chromogenic Media for Recovery of Vancomycin-Resistant Enterococci from Fecal Samples

Abstract

Five chromogenic agars, evaluated using 400 stool specimens, were found to be superior in sensitivity (range, 89.9 to 93.9%) to bile esculin azide agar with vancomycin (BEAV) agar (84.8%) for detecting vancomycin-resistant enterococci (VRE), and the results were available 24 to 48 h sooner. The time to detection, need for supplemental testing, color distinction, and breakthrough of non-VRE organisms vary among the chromogenic media tested and may factor into the decision to use a particular medium.

TEXT

Vancomycin-resistant enterococci (VRE) are important causes of nosocomial infections and are associated with increasing mortality among specific patient groups, such as those with hematologic malignancy (1, 2). The rapid and reliable identification of antibiotic-resistant organisms is critical for patient management and infection control measures in both infected and colonized individuals (1, 3). The CDC recommends that institutions with moderate to high rates of VRE perform active surveillance using stool or rectal swab specimens (4). Bile esculin azide agar with vancomycin (BEAV) has been used as a traditional screening method in many clinical laboratories for detecting VRE in rectal swabs or stool specimens. Confirmation of VRE using this medium requires a minimum of 72 h. Chromogenic agars have been developed to decrease this turnaround time. These media are both selective and often differential for detecting VR Enterococcus faecium and VR Enterococcus faecalis. Many of these chromogenic media can be read and finalized within 24 h, with few or no additional supplemental tests needed (5–11). This dramatic decrease in the turnaround time for VRE detection allows more timely implementation of proper infection control measures, potentially preventing nosocomial transmission. However, the published performance characteristics for VRE identification among the available chromogenic media are limited. Therefore, we compared five different commercially available chromogenic VRE media to the traditional BEAV agar and a bile esculin azide broth with vancomycin (BEAV broth) for the detection of VRE in stool specimens.

Deidentified stool specimens from patients who provided samples for Clostridium difficile testing were used for this evaluation. One microliter of remnant stool was inoculated and aseptically streaked for isolation onto each of the five chromogenic media: the InTray Colorex VRE (BioMed Diagnostics, White City, OR), chromID VRE (bioMérieux, Marcy l'Étoile, France), VRESelect (Bio-Rad, Marnes-la-Coquette, France), HardyCHROM VRE (Hardy Diagnostics, Santa Maria, CA), and Spectra VRE (Remel, Lenexa, KS), along with BEAV agar (6 μg/ml vancomycin) (Enterococcosel; BD Diagnostics, Sparks, MD) and BEAV broth (6 μg/ml vancomycin) (BEAV broth; Hardy Diagnostics). At the time of this study, BEAV agar was routinely used for VRE surveillance at the Johns Hopkins Hospital. For the purposes of this study, BEAV broth was added in order to enhance VRE detection performance. The media and BEAV broths and all chromogenic media were incubated in the dark, in non-CO₂, at 35 to 37°C. The identification of the enterococci from the chromogenic agars was based on the observation of appropriately colored colonies and supplemental testing (Gram stain, catalase, and l-pyrrolidonyl-β-naphthylamide enzyme [PYR]) if required, as per the manufacturers' instructions (Table 1). Vancomycin susceptibility tests were performed using Etest (AB bioMérieux, Durham, NC) for the VRE isolates recovered from each medium.

TABLE 1.

Comparison of VRE detection media in specimen type, identification, and supplemental testing

Medium (manufacturer)	Specimen type	Vancomycin concn (μg)	FDA approved	Incubation time (h)	Chromogenic color identification	Supplemental testinga
Enterococcosel (BEAV) (BD Diagnostics)	Rectal swab and stool	6	No	48	Black, no species differentiation	Gram stain, PYR, susceptibility testing from nonselective medium
InTray Colorex VRE (BioMed)	Not applicable	Unknown	Environmental and food microbiology only	18–24	Pink to mauve, no differentiation between E. faecium and E. faecalis	None
chromID (bioMérieux)	Stool	8	Yes	48, read at 24 & 48	E. faecium→violet, E. faecalis→blue to green	Gram stain, catalase, susceptibility testing from nonselective medium
VRESelect (Bio-Rad)	Rectal swabs and stool	8	Yes	24–28	E. faecium→pink, E. faecalis→blue	Catalase for E. faecalis only; if negative, perform susceptibility testing from nonselective medium
HardyCHROM VRE (Hardy Diagnostics)	Rectal swabs and stool	10	No	48, read at 24 & 48	E. faecium→dark blue, E. faecalis→dark red	PYR, susceptibility testing from nonselective medium
Spectra VRE (Remel)	Rectal swabs and stool	6	Yes	24	E. faecium→navy blue to pink, E. faecalis→light blue	None

^aA positive result for VRE chromogenic agar that did not match the results of the BEAV broth was assessed by subculturing organisms to nonselective medium, followed by reidentification and repeat susceptibility using BD Phoenix and Etest.

At 24 h, BEAV broths with black color development were subcultured to 5% sheep blood agar and incubated for an additional 18 to 24 h. Catalase-negative PYR-positive colonies were sent to the BD Phoenix automated microbiology system (BD Diagnostics, Sparks, MD) for identification to the species level and susceptibility testing. The vancomycin susceptibility results were reported using current Clinical and Laboratory Standards Institute interpretive guidelines (vancomycin resistant, MIC > 32 μg/ml) (12). A positive result for a chromogenic agar that did not match the results of the BEAV broth was assessed by subculturing the organisms from each chromogenic medium and BEAV broth to 5% sheep blood agar again for reidentification and repeat susceptibility, as described above, using BD Phoenix and Etest. If the identification of the colored colony did not match the Phoenix identification, it was considered an incorrect identification. Appropriately colored colonies growing directly on chromogenic medium identified as VRE and the results of the subculture from positive BEAV plates and broth served as a combined true-positive gold standard for this study after the resolution of discrepancies by repeat testing.

Of the 400 specimens tested, four were excluded from the analysis because they grew only VRE Enterococcus raffinosus, an organism for which none of the chromogenic agars have FDA approval. Of the 396 remaining specimens, 297 (75%) were negative for VRE. The performance characteristics of each medium, BEAV agar, and BEAV broth were compared using combined data (Table 2). Overall, the sensitivities of all of the chromogenic media (89.9 to 94.9%) were higher than that of the BEAV agar (84.8%). chromID had the highest sensitivity (94.9%). All of the chromogenic media recovered significantly more VRE than did the BEAV agar. There was no statistical difference in VRE detection performance among the five chromogenic media (Table 2).

TABLE 2.

Comparison of VRE detection on various media from 396 specimens (combined data)

Medium	No. of resultsa				Performance (% [95% CI])b
	TP	FP	TN	FN	Sensitivity	Specificity	PPV	NPV
BEAV broth	97	0	297	2	98.0 (94.7–98.0)	100 (98.9–100)	100 (96.7–100)	99.3 (98.3–99.3)
Enterococcosel (BEAV)	84	0	297	15	84.8 (80.9–84.8)	100 (98.7–100)	100 (95.3–100)	95.2 (93.9–95.2)
InTray Colorex VRE	91	5	292	8	91.9 (86.9–94.8)	98.3 (96.6–99.3)	94.8 (89.6–97.9)	97.3 (95.7–98.3)
chromID	94	1	296	5	94.9 (91.0–95.9)	99.7 (98.3–100)	98.9 (94.8–99.9)	98.3 (97.0–98.7)
VRESelect	91	1	296	8	91.9 (87.8–92.9)	99.7 (98.3–100)	98.9 (94.4–99.9)	97.4 (96.0–97.7)
HardyCHROM VRE	89	1	296	10	89.9 (85.6–90.9)	99.7 (98.2–100)	98.9 (94.2–99.9)	96.7 (95.4–97.0)
Spectra VRE	93	1	296	6	93.9 (89.9–94.9)	99.7 (98.3–100)	98.9 (94.7–99.9)	98.0 (96.7–98.3)

^aA true positive (TP) is defined as an appropriately colored colony growing directly on chromogenic medium, BEAV agar, or the subculture from BEAV broth that was identified by supplemental tests or BD Phoenix as E. faecium or E. faecalis and resistant by Etest. A false negative (FN) is defined as the absence of an isolate on a particular agar that was identified as VRE on one or multiple media. FP, false positive; TN, true negative.

^bCI, confidence interval; PPV, positive predictive value; NPV, negative predictive value.

A total of 99 of 396 specimens (25%) were positive for 117 VRE isolates, from which 92 isolates (79%) were VR E. faecium and 17 (15%) were VR E. faecalis. Several cultures grew more than one VRE isolate, and the detection rate for the mixed cultures varied by medium (Table 3). The VRE species recovered, misidentified VRE species, and breakthrough of non-VRE isolates as detected on each chromogenic agar and BEAV agar and broth are depicted in Table 3. Very few isolates of non-VRE enterococci and species other than enterococci (Staphylococcus aureus, normal enteric flora, and yeast) broke through on the chromogenic media (Table 3).

TABLE 3.

Species identification, misidentified VRE isolates, and non-VRE isolates recovered from 99 stool specimens positive for VRE

Medium	Total isolates	Mixed culturesa	No. of VRE isolatesb			Species (no.) identified forc:
			E. faecium	E. faecalis	E. raffinosus	Misidentified VRE species	Non-VRE isolates
BEAV broth	115	10	90	17	8	NA	E. gallinarum (11), E. casseliflavus (6), E. raffinosus (2)
BEAV	83	NA	NA	NA	NA	NA	E. gallinarum (7), E. casseliflavus (4)
InTray Colorex VRE	88	NA	NA	NA	NA	NA	E. gallinarum (2), yeast (2), normal enteric flora (1)
chromID	100	6	86	14	0	E. faecalis E. faecium, (2), missing additional species (1)	E. faecalis (non-VRE)
VRESelect	99	8	83	16	0	E. faecalis E. faecium (1), E. faecium E. faecalis, (1)	Normal enteric flora (1)
HardyCHROM VRE	93	10	81	12	0	E. faecalis E. faecium (3), E. faecium E. faecalis (2), missing additional species (4)	S. aureus (1), normal enteric flora (1)
Spectra VRE	114	10	95	19	0	E. faecalis E. faecium (3), E. faecium E. faecalis (1)	E. raffinosus (1)

^aMixed cultures were the specimens that were positive for more than one VRE isolate. NA, not applicable.

^bNumber of VRE and non-VRE isolates recovered from chromogenic media and BEAV broth identified by BD Phoenix and Etest.

^cSee Materials and Methods for definition of misidentified VRE isolates. In parentheses are the correct VRE identifications and number of isolates. Some media missed additional VRE isolates compared to other agars.

Among the chromogenic media, the InTray Colorex VRE had the least desirable performance, as it missed 26 VRE isolates and had more breakthrough than the other chromogenic media (Table 3). BEAV broth and BEAV agar identified more isolates of vanC-carrying enterococci (Enterococcus gallinarum and Enterococcus casseliflavus) than any of the chromogenic media. Likewise, 8 isolates of vancomycin-resistant E. raffinosus were recovered from the BEAV broth. None of these grew on any of the chromogenic media (Table 3).

To our knowledge, this is the first comprehensive study to compare five available chromogenic media for VRE detection. The high sensitivity and specificity observed by the chromogenic agars in our study are comparable to those found in previous studies that examined their individual performances compared to those of standard BEAV agar and BEAV broth (5–11). Our study is useful as a novel comparison of these 5 chromogenic media, both to each other and to a commonly used reference method. At the time of our study, the FDA-approved selective media for VRE detection included chromID VRE, VRESelect, and Spectra VRE. Since 2013, the FDA has also approved the Brilliance VRE agar (Oxoid, Ottawa, Canada).

There are differences among the chromogenic media with respect to need for supplemental testing; however, some are minimal and would have little impact on turnaround time. Among the FDA-approved media, the VRESelect and Spectra VRE media yielded results 24 to 48 h sooner than BEAV agar and BEAV broth. The chromID agar was more sensitive than BEAV agar and is capable of differentiating between E. faecalis and E. faecium; however, it needs longer incubation than VRESelect and Spectra VRE. The color interpretation of the positive results might be slightly more difficult with Spectra VRE than with VRESelect. The InTray Colorex VRE and HardyCHROM VRE have not been cleared by the FDA. Color interpretation of the InTray Colorex VRE is easy, but it cannot differentiate between E. faecalis and E. faecium. The sensitivity for VRE detection with the HardyCHROM VRE is the lowest of the other chromogenic media. However, based on the confidence intervals, the sensitivity of this medium is significantly higher than that of the standard BEAV agar and not different from the other chromogenic media. The observed differences in the performance of VRE recovery might be impacted by different vancomycin concentrations among the chromogenic media, as well as by the amount of VRE organism growth. A previous study reported that chromogenic media were negative if the growth of VRE was <10 colonies/plate (6). The quantity of VRE isolates in a patient specimen might be lower than the detection capabilities of the different media. In our study, when specimens contained two different species of VRE, all chromogenic media were capable of detecting at least one of the VRE species. Therefore, the failure to detect additional VRE species would not have a critical impact on infection control decision making.

The misidentifications in our study were similar to those reported in previous studies (6–9). Peterson et al. (8), using sequencing analysis to compare VRE species identification with phenotypic results (Vitek 2, bioMérieux, Marcy l'Étoile, France), also reported misidentification of species (2 of 110 isolates) by the Spectra VRE. These misidentified VRE isolates might be from technical errors when trying to interpret colony color. Other important factors in chromogenic medium selection might be the costs of the medium, laboratory workflow, and technician processing time. In order to prevent false-positive results or the breakthrough of non-VRE isolates, it is very important to perform the cultures as per the manufacturer's specific instructions, such as adherence to 24-h incubation time and incubation of the media in the dark to maintain stability.

Our study has some limitations. We did not perform sequencing analysis to confirm phenotypic and biochemical testing for VRE species identification. We enhanced the VRE recovery using BEAV broth, and we confirmed identification using the BD Phoenix. We performed VRE detection only with stool specimens. A previous study showed comparable VRE detection rates using rectal swab and stool specimens (13). At our institution, after we selected the new media for VRE detection, we performed a validation of the rectal swabs for VRE detection, and the results were comparable to those obtained using stool specimens (data not shown). Since the completion of our study, other VRE detection methods have been developed, including a new FDA-cleared medium, Brilliance VRE (Oxoid, Ottawa, Canada) and additional molecular assays. We did not have the opportunity to evaluate Brilliance VRE, since it was not available at the time we initiated the study. A subsequent study showed good performance of this new chromogenic agar (10).

In summary, this study demonstrates that chromogenic media are superior in sensitivity to BEAV agar for VRE detection but are less sensitive than enrichment broth. All of the five evaluated chromogenic media yielded results 24 to 48 h sooner than BEAV agar and BEAV broth. There are differences among the chromogenic agars with respect to time to detection, need for supplemental testing, and ease of color distinction among colonies, as well as non-VRE breakthrough that may factor into a laboratory's decision to use a particular agar.