A Decade of Development of Chromogenic Culture Media for Clinical Microbiology in an Era of Molecular Diagnostics

Candida spp.

A chromogenic medium for the identification and differentiation of pathogenic yeasts, CHROMagar Candida (CAC), was first reported in 1994 (2). As well as antibacterial agents, the medium incorporates two chromogenic substrates for the detection of β-hexosaminidase activity and phosphatase activity (18). The medium affords specific identification of Candida albicans/Candida dubliniensis, which form green colonies due to production of β-hexosaminidase, and Candida tropicalis, which forms blue colonies due to production of both enzymes. Other species of yeast form pink colonies due to phosphatase activity alone or produce neither of these enzymes and grow as white colonies. A range of commercially available chromogenic media has since been evaluated, including Albicans ID (19), CandiSelect (19), Candida ID (20), Candida diagnostic agar (21), Pourmedia Vi Candida (22), Chromogenic Candida agar (23), Brilliance Candida Agar (24) and HiCrome Candida differential agar (25). A common feature of these media is a chromogenic substrate for β-hexosaminidase to discriminate C. albicans/C. dubliniensis from other yeasts, and most include a second chromogenic substrate (usually to detect phosphatase or β-glucosidase) to provide further discrimination between species (18). The main advantage of such chromogenic agars is their ability to detect mixed cultures of yeasts due to the fact that different species frequently form colonies with different colors. Such mixtures of species may be indistinguishable and remain undetected as mixtures on conventional agars such as Sabouraud agar plus chloramphenicol (23, 26). This is important, as different species may have different susceptibilities to antifungal agents. While C. albicans/C. dubliniensis are usually susceptible to antifungal agents, chromogenic media may help to detect species with a higher likelihood of resistance to azoles and/or amphotericin B, including Candida krusei, Candida glabrata, Candida rugosa and Candida inconspicua (27).

There have been relatively few comparisons of different chromogenic agars using clinical specimens in the last decade. Ozcan et al. compared Oxoid Chromogenic Candida agar (OCCA) with CHROMagar Candida (CAC) and Sabouraud chloramphenicol agar (SCA) using 392 vaginal swabs. Yeasts were isolated from 161 samples, and 21 samples (13%) yielded a mixture of species on at least one medium (23). OCCA and CAC showed comparable sensitivity (96.9% versus 97.5%, respectively) for detection of positive samples, whereas the sensitivity of SCA was lower (91.9%). For the 21 polyfungal infections, 20 (95.2%) were detected using OCCA, compared with only 14 (66.7%) using CAC (P < 0.05). Sendid et al. compared CandiSelect 4 (CS4) with CAC using 1,549 clinical samples from a wide variety of sites (28). A total of 502 samples (32.4%) yielded one or more yeast species, including 37 samples (7.4%) that yielded more than one species. The sensitivities of CS4 and CAC were very similar (92.1 and 91.1%, respectively), with no false-positive results. CS4 was superior to CAC for presumptive identification of C. glabrata (80 versus 75%) and C. krusei (92 versus 83%) but was less effective for C. tropicalis (68 versus 76%).

Pseudomonas aeruginosa

P. aeruginosa is an important nosocomial pathogen and may also cause community-acquired infections, particularly in individuals with underlying disease. For example, in patients with cystic fibrosis, it is a common and important cause of respiratory tract infection. Laine et al. reported the first chromogenic medium designed specifically for the isolation of P. aeruginosa (PS-ID), which was subsequently commercialized as chromID Pseudomonas (11). The medium is notable as it is the first chromogenic medium to utilize a chromogenic substrate for peptidase activity. This substrate, β-alanyl pentylresorufamine, is hydrolyzed by a β-alanyl aminopeptidase produced by P. aeruginosa, resulting in the formation of purple colonies (29) (Fig. 1a). The medium was evaluated with 100 sputum samples from patients with cystic fibrosis and compared with Pseudomonas CN selective agar (CN). The recovery of P. aeruginosa was equivalent on both media (95.2%), but the positive predictive value of PS-ID (98.3%) was significantly higher than that of growth on CN (88.5%) for identification of P. aeruginosa (P < 0.05). Other species of Gram-negative bacteria were occasionally isolated as purple colonies on PS-ID, including Burkholderia cepacia complex (11). There are no other reports of this medium with clinical samples; however, Weiser et al. included chromID Pseudomonas in a comparison of five selective media that were challenged with 50 isolates of P. aeruginosa and 90 isolates belonging to closely related species (30). chromID Pseudomonas showed the highest specificity of the five media tested, but the authors reported that its sensitivity (95%) was negatively impacted by the large variation in color of P. aeruginosa colonies (including pink-brown and green, possibly due to interference from natural pigments of P. aeruginosa). In conclusion, there is a lack of any published studies with clinical samples that demonstrate a higher recovery of P. aeruginosa using chromogenic media.

Staphylococcus aureus

S. aureus is one of the most frequent and important human pathogens and is implicated in a range of infections, including superficial skin infections, abscesses, bacteremia, and food poisoning. It is frequently found colonizing the nose, throat, and skin without causing symptoms. The first chromogenic medium for the isolation of S. aureus, CHROMagar S. aureus, was reported in 2000 (4) and utilized a phosphatase substrate for detection of S. aureus as pink colonies (18). Since that first report, at least two other media have been made commercially available and evaluated with clinical samples, including S. aureus ID (31) (later commercialized as chromID S. aureus) and SaSelect (32). An alternative approach is utilized in chromID S. aureus, which relies upon production of α-glucosidase by S. aureus, resulting in the formation of green colonies. Each of these media has been reported to show sensitivity that is equivalent to or higher than that of conventional nonselective media (e.g., blood agar), and, due to the incorporation of selective agents for the inhibition of nonstaphylococci, they are particularly useful for specimens that yield a polymicrobial flora that includes Gram-negative bacteria. They also have high specificity (>90%) for detection of S. aureus, meaning that fewer confirmation tests are required when reading culture plates (4, 31–34). The sensitivity may be increased if incubation is extended to 48 h, particularly for CHROMagar S. aureus (31, 33), but this is offset by a small decrease in specificity due to other species forming colonies of the same color as S. aureus (31, 33, 34). As there is only one published “head-to-head” comparison of these chromogenic media with clinical samples (31), there are insufficient data to conclude whether any particular chromogenic medium is better than another.

Streptococcus agalactiae (Group B Streptococcus)

Infections caused by group B streptococci (GBS) are a leading cause of morbidity and mortality in newborn infants. Asymptomatic carriage of GBS in the maternal genitourinary tract may lead to colonization of the neonate, and in a small proportion of cases, this may lead to invasive disease. In an effort to reduce the burden of disease, authorities in many countries recommend universal screening of all pregnant women for vaginal/rectal colonization by GBS at 35 to 37 weeks of gestation (35). A widely used standard procedure involves overnight incubation of samples (i.e., vaginal/rectal swabs) in a selective enrichment broth followed by subculture onto blood-based culture media for investigation of typical hemolytic colonies (35). Granada medium is also widely used, and this medium allows GBS to grow as orange colonies under anaerobic conditions due to formation of a natural pigment (36).

The first chromogenic medium for GBS (a prototype of chromID Strepto B) was described in 2006 (6). The medium allows GBS to form red colonies based on production of phosphatase. Other species either form colorless colonies or hydrolyze additional chromogenic substrates (for esterase and β-cellobiosidase enzymes) to produce blue/green colonies (37). Since this first report, a large number of studies have evaluated a range of chromogenic media for detection of GBS. Table 2 summarizes a selection of eight of these studies (i.e., those with the largest number of positive samples). A number of studies have compared chromogenic media against selective blood-based agars (usually containing colistin and nalidixic acid) with or without the use of a selective enrichment broth (38–44). When both types of media were tested under the same conditions, chromogenic media showed a higher sensitivity than selective blood agars in all of these studies. Most studies show that the sensitivity of any culture medium for detection of GBS may be substantially improved by use of an enrichment broth (38, 39, 41, 44, 45). Chromogenic media have a potential advantage over Granada medium, as they do not require anaerobic incubation and have the ability to detect nonhemolytic strains of GBS that typically fail to produce pigment on Granada medium. Such strains are thought to account for up to 5% of invasive infections (46). Despite this, several studies have compared Granada medium with chromogenic agars (6, 40, 45–48), and overall there is no clear advantage of either in terms of sensitivity. Moreover, Granada medium invariably demonstrates 100% specificity and is arguably the only medium that can be used without the need to confirm the identification of suspect colonies of GBS (46).

Only a few studies have compared the performance of different chromogenic media for GBS in a head-to-head evaluation using clinical samples (46, 48, 49). No statistically significant advantage was found for any of the media tested, and sensitivity is likely to have been underestimated because they were not used in conjunction with an enrichment broth (38, 44, 45). Most studies conclude that chromogenic media for GBS are highly convenient tools that offer an increased sensitivity and specificity over conventional blood-based media. Further large studies would be needed to establish the superiority of any particular chromogenic medium, and the use of an enrichment broth should ideally be included in such studies.

Urinary Tract Pathogens

The first report of a chromogenic medium for diagnosis of urinary tract infections described an evaluation of CPS ID2 in 1995 (3). This medium exploited a substrate for β-glucuronidase to allow the specific identification of the most common urinary pathogen, Escherichia coli, as pink or red colonies. An additional substrate for β-glucosidase allows detection of enterococci as small green colonies and the Klebsiella-Enterobacter-Serratia (KES) group as larger green colonies. Finally, the inclusion of tryptophan and iron salts allows Proteeae (Proteus-Providencia-Morganella) to form brown colonies due to deaminase activity (50).

A range of other media has since been commercialized and evaluated with clinical samples, including CHROMagar Orientation (51), UriSelect medium (52), Rainbow Agar UTI medium (52), Chromogenic UTI medium (52), USA agar (53), Harlequin CLED (54), and Urichrom agar (55). A number of these media, including CHROMagar Orientation, UriSelect medium, and Chromogenic UTI medium, utilize a substrate for β-galactosidase for detection of E. coli (18). This allows the direct identification of a higher proportion of E. coli isolates, as approximately 99% of E. coli isolates produce β-galactosidase, compared with approximately 94% that produce β-glucuronidase (18). However, this advantage is offset by a small decrease in specificity due to the misidentification of a proportion of Citrobacter spp. as E. coli in some reports (56, 57). This small decrease in specificity can be largely eliminated by inclusion of a spot indole test, but this is laborious as it needs be applied to all colonies resembling E. coli (56).

In some studies, chromogenic media have been shown to provide a superior differentiation of mixed cultures due to the fact that different species may generate colonies with different colors and may not be easily differentiated on conventional agars. This can assist in the recognition of urine samples that may be contaminated, particularly compared with culture on cystine-lactose-electrolyte-deficient (CLED) agar (54, 56). However, this advantage is not apparent in other studies (58, 59). A consistent advantage of chromogenic media is their ability to identify E. coli and provide identification of other groups of species (KES and Proteeae). Several groups have shown how that can contribute to a decrease in workload for species identification and/or to cost savings to the laboratory (60–62); however, this may have no significant impact on the overall time taken to generate a final report (62).

Chromogenic media designed for detection of urinary tract infections are unique among chromogenic media as they do not contain antimicrobials as selective agents in order to cultivate as many species as possible. They can therefore potentially be used as single media for the culture of urine samples. In one of the largest reported studies, Aspevall et al. (58) evaluated four chromogenic media, i.e., Chromogenic UTI medium, CHROMagar Orientation from two commercial sources, and CPS ID2, alongside culture on CLED, blood agar, and MacConkey agar using 1,200 urine samples. Although incubation was extended for up to 48 h, this had a minimal impact on any of the test media. A total of 420 isolates deemed to be potentially significant were recovered from 379 urine samples at a count of ≥10⁴ CFU/ml. A total of 96% of these isolates were recovered on blood agar and also on CLED agar. The four chromogenic media recovered between 92 and 96% of isolates. The authors concluded that any of the chromogenic media studied could be used as a single medium for the isolation of uropathogens. The authors also reported that mixed urethral flora was easier to detect on blood agar due to better growth of fastidious species such as corynebacteria and alpha-hemolytic streptococci. They therefore advocated retaining blood agar as part of the urine culture workup, as isolation and discrimination of different Gram-positive bacterial species were found to be much easier with this medium. This has been noted by others; for example, Yarbrough et al. (62) noted the recovery of a smaller amount of periurethral flora (including lactobacilli and group B streptococci) on chromID CPS Elite than with culture on blood agar. This resulted in fewer reports of contaminated or insignificant growth. The authors concluded that chromID CPS Elite agar may be a feasible alternative to conventional media for isolation and identification of most common uropathogens in urine specimens.

Some brands of media have been subject to incremental improvements over the years and new generations of media have been commercialized. In the last decade, four published studies have compared either UriSelect 4 or CHROMagar Orientation with different generations of CPS media (CPS ID3, CPS ID4, and chromID CPS Elite). The studies revealed only minor differences in specificity between these media, with sensitivity reported to be broadly equivalent (57, 59, 63, 64). Data are lacking on the effectiveness of chromogenic media for recovery of some fastidious Gram-positive bacteria, some of which (e.g., Aerococcus urinae) have been implicated as human pathogens (63, 65).

Clostridium difficile

Over the last decade there has been renewed interest in the culture of stool samples for the isolation of C. difficile. One reason is the emergence of so-called hypervirulent strains that cause outbreaks of C. difficile infection (CDI) that are associated with an increased severity of disease and significant mortality (66). In order to track the spread of such strains, it is usually necessary to isolate them by culture and perform molecular typing. Culture also affords a very high sensitivity for detection of C. difficile and may therefore be useful in the diagnosis of CDI. In one large 7-year study, toxigenic culture resulted in the diagnosis of 355 cases of CDI that would have been missed using the fecal cytotoxin assay alone (67). For these reasons, isolation of C. difficile followed by demonstration of toxin or toxin genes (“cytotoxigenic culture”) is accepted by many as a “gold standard” for diagnosis of CDI (68).

The first chromogenic culture medium (IDCd) for isolation of C. difficile was reported in 2010 (13). The principle was based upon the ability of C. difficile to generate black colonies due to expression of β-glucosidase activity resulting in the hydrolysis of a chromogenic substrate (Fig. 1c). The authors reported that IDCd offered effective isolation of C. difficile within only 24 h of incubation with or without the use of alcohol shock treatment, in contrast to other selective media. IDCd was subsequently commercialized and marketed as chromID C. difficile. Since this first report, at least six published articles have reported evaluation data for chromID C. difficile in comparison with other media (69–74). Five of these studies are summarized in Table 3. In all cases, chromID C. difficile showed a sensitivity superior to that of comparator media and resulted in greater inhibition of other flora. In three of these studies, there was evidence that incubation of chromID C. difficile for 48 h improved the sensitivity of the medium, particularly for clinical samples with a low burden of C. difficile. In a further study, chromID C. difficile and Oxoid Clostridium difficile selective agar (CCFA) were used for the culture of 686 stool samples from 508 patients in four hospitals in Hong Kong, with incubation of both media for up to 72 h (74). C. difficile was isolated from 118 stool samples using chromID C. difficile, compared with 70 stool samples using CCFA (P < 0.001); however, the overall sensitivity of the two media was not reported.

The high sensitivity afforded by chromID C. difficile is due to a combination of high selectivity against unwanted bacteria and a strong propensity of the medium to stimulate germination of spores (13). These attributes were exploited by Hill et al., who demonstrated the superior sensitivity of chromID C. difficile for recovery of C. difficile from environmental surfaces (75). In a study with 496 samples from the hospital environment, the sensitivity of chromID C. difficile was 87.6%, compared with a sensitivity of 26.6% for cefoxitin-cycloserine-egg yolk agar plus lysozyme, a medium that has also been recommended for environmental screening (P < 0.0001) (76).

Despite the high sensitivity of chromID C. difficile, the medium also has some limitations. The chromogenic reaction is not specific for C. difficile, and black colonies may be produced by other anaerobic species, most notably Clostridium hathewayi (a species previously classified as Clostridium clostridioforme) (74). Colonies recovered on chromID C. difficile therefore require identification, and this can be readily achieved using MALDI-TOF MS (74). Alternatively, Park et al. proposed the use of a Gram stain plus a simple disk test for pyroglutamyl aminopeptidase, and this may be useful for laboratories without access to MALDI-TOF MS (77). Furthermore, a subset of C. difficile strains fails to generate black colonies due to the absence of the β-glucosidase gene, and this appears to be a consistent feature of strains of ribotype 023 (78, 79). In a UK study, the proportion of isolates failing to generate black or gray colonies within 48 h was reported to be 1.6% (13). Such isolates may still be detected on chromID C. difficile due to their characteristic colony shape, but care must be taken to ensure that such colonies are not overlooked.

It is worth emphasizing that culture of C. difficile alone has little predictive value for diagnosis of CDI without subsequent demonstration of the isolate's ability to produce cytotoxin. This can be directly demonstrated by testing culture supernatants on cell lines or may be inferred much more rapidly by testing colonies for toxin genes by PCR (80). Darkoh et al. claimed that they circumvented this problem with the report of a new chromogenic medium (the Cdifftox plate assay) on which toxigenic C. difficile isolates formed blue colonies, thus differentiating them from nontoxigenic isolates, which formed white colonies (81). This was achieved using 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside (X-Gal) for detection of the glycosyltransferase activity of toxins A and B. Despite its promise, I am not aware of any published evaluation data for this medium since it was first described in 2011 (81).

Campylobacter spp.

Campylobacter spp. are the most common bacterial cause of gastroenteritis (GI) in many countries, and infection is primarily due to ingestion of contaminated food. CASA (Fig. 1d) is a chromogenic medium initially designed for the isolation of Campylobacter spp. from food, but it has since been evaluated with stool samples from patients with suspected gastroenteritis. The chromogenic substrate used for detection of Campylobacter spp. is undisclosed by the manufacturer. Le Bars et al. compared CASA with two nonchromogenic agars (Karmali medium and Campylosel) for the isolation of Campylobacter spp. from 370 diarrheic stool samples (14). Cultures on all three media were incubated for up to 96 h in a microaerophilic atmosphere at two different temperatures, 37°C and 42°C. The sensitivity of CASA was equivalent to or slightly better than that shown by either of the other two media, but CASA was reported to be much more selective, which led to a reduction in the time required for processing colonies for confirmation of Campylobacter. Dalziel et al. reported the culture of 979 stool samples on CASA and modified charcoal-cefoperazone-deoxycholate agar (CCDA) with incubation of cultures at 42°C for 36 to 48 h in a microaerophilic atmosphere (82). The authors reported sensitivities of 100% and 84% for CASA and CCDA medium, respectively (P < 0.001), and also reported a higher selectivity of CASA. A limitation of CASA is the lack of a specific chromogenic reaction to indicate the presence of Campylobacter spp., as other species that grow on the medium also hydrolyze the chromogenic substrate to produce pink or red colonies, and some may therefore appear quite similar to Campylobacter spp.

Salmonella spp.

Salmonella spp. remain one of the most important causes of foodborne gastroenteritis, and chromogenic media designed for the specific isolation of Salmonella spp. have been available for at least 25 years. Rambach agar and SM-ID were the first two examples of such media (1), and both have been superseded by new generations of media. The principles of Salmonella detection exploited in these first generations of chromogenic media have been previously reviewed (50). It has been consistently demonstrated that chromogenic media do not offer a superior sensitivity to conventional agars such as xylose-lysine-deoxycholate (XLD) agar and Hektoen enteric agar (83, 84). Furthermore, in contrast to almost all chromogenic media, such conventional agars offer the opportunity to isolate both Salmonella and Shigella using a single culture medium. No single culture medium or combination of media can preclude the necessity for enrichment of stool samples, e.g., in selenite broth, which is essential for detection of low numbers of Salmonella and significantly enhances detection (84). The sole advantage of chromogenic media for Salmonella is the significantly higher specificity they afford compared to conventional media. This means that fewer confirmatory tests than with conventional media are required to investigate colonies of other species that may resemble Salmonella. This can result in cost savings for laboratories; e.g., in one study, a saving of EUR 2.7 (approximately US$3) per sample by inclusion of a chromogenic medium was projected (83).

Most of the chromogenic media for Salmonella that are currently marketed rely on detection of a C8-esterase enzyme produced by Salmonella that is detected by inclusion of a chromogen linked to caprylic (octanoic) acid. Substrates for β-glucosidase and/or β-galactosidase are included so that other coliforms generate a different color. Antibiotics such as cefsulodin and novobiocin may be included for the inhibition of Pseudomonas spp. and Proteus spp., respectively. Brilliance Salmonella agar incorporates a “suicide substrate” or “Inhibigen” that is hydrolyzed by E. coli to release a toxic product, thus inhibiting its growth (P. Druggan, 21 March 2002, patent application WO0222785). Only two published studies in the last decade have compared chromogenic media for the isolation of Salmonella (83, 84). The chromogenic media included were chromID Salmonella ELITE, BBL CHROMagar Salmonella, SM-ID2, and Brilliance Salmonella agar, and these were compared with conventional agars. Neither study reported any significant difference between any of the media with respect to sensitivity, but the specificity of chromogenic media was significantly higher than that afforded by conventional agars.

In summary, the use of a conventional agar (e.g., XLD agar) is appropriate for direct culture of stool samples, as it accommodates isolation of Shigella spp., and the evidence suggests that sensitivity is at least equivalent to that of chromogenic media for detection of Salmonella spp. The use of a chromogenic medium after enrichment in selenite broth is an attractive option to target Salmonella spp. with high specificity and consequently reduce the number of colonies requiring investigation. There is no clear advantage of any particular chromogenic medium for this purpose.

Shigella spp.

Shigella spp. produce few hydrolytic enzymes that provide useful differentiation from other bacteria, and that has restricted the development of chromogenic media for their detection. However, the observation that all species of Shigella produce β-ribosidase has provided one alternative means of detection (85). HardyCHROM SS is the only commercially available chromogenic medium allowing isolation of both Salmonella and Shigella that has been evaluated with clinical samples. Hinde et al. evaluated this medium with 400 stool samples in comparison with conventional MacConkey and Hektoen enteric agars (16). The authors reported a superior specificity for HardyCHROM SS that allowed fewer colonies requiring investigation, and they also reported a shorter time to detection. However, since only one isolate of Shigella spp. was recovered during the study, further studies with larger numbers of positive samples are essential.

Shiga Toxin-Producing Escherichia coli

Shiga toxin-producing E. coli (STEC) bacteria are a cause of foodborne gastroenteritis, hemorrhagic colitis, and hemolytic-uremic syndrome. Among STEC strains, Shiga toxin production was first associated with E. coli of serotype O157:H7, and it was noted that such strains, unlike most other E. coli strains, failed to ferment sorbitol. This led to the design of sorbitol MacConkey agar, which has become widely used by clinical laboratories. This situation has become much more complicated, and STEC can be found within many other serotypes that together may account for as much STEC-associated disease as O157:H7 (86). The fact that many of these additional serotypes typically ferment sorbitol has severely limited the effectiveness of sorbitol MacConkey agar for diagnosis of infection with STEC.

Recovery of STEC by culture is challenging due to a low density of organisms in some stool samples and the lack of consistent biochemical features among STEC serotypes (87). Despite this, since the first report of Rainbow agar O157 in 1998 (88), chromogenic media have been commercialized for the detection of STEC. Most of these media have been designed primarily for isolation of E. coli O157:H7, including CHROMagar O157, Colorex O157, and Rainbow O157, whereas CHROMagar STEC allows for detection of a wider range of STEC serotypes (89). Most of these media are based on similar principles, relying on an inability of E. coli O157 to produce acid from sorbitol and/or rhamnose and a lack of β-glucuronidase activity. A second chromogenic substrate (e.g., for α- or β-galactosidase) may be used to highlight the presence of E. coli O157:H7 among nonreactive background flora (18).

There have been six published evaluations of these media with clinical samples over the last decade. Grys et al. tested 204 stool samples using PCR for detection of toxin genes and by direct culture on CHROMagar O157; only four positive samples were found (all positive by PCR), and three of these were detected by culture (12). Hironven et al. tested 47 fecal samples from patients with hemorrhagic diarrhea by plating on CHROMagar STEC and using an immunochromatographic assay for O157 antigen and Shiga toxin. The chromogenic medium detected STEC in 16 positive samples, compared with only 14 detected by the immunoassay (90). Wylie et al. compared direct culture on CHROMagar STEC with a standard cytotoxin assay using 205 fecal samples (86). There were 14 positive samples, and the sensitivity and specificity of CHROMagar STEC were reported as 85.7% and 95.8%, respectively. Gouali et al. cultured 329 stool samples onto CHROMagar STEC and Drigalski agar after preenrichment of the samples in a nonselective broth. Colonies from Drigalski agar were harvested and tested by PCR for toxin genes (87). From 39 Shiga toxin-positive stool specimens, STEC was recovered as mauve colonies from 32 samples (sensitivity, 82.1%). Forty-eight isolates of E. coli that were not found to harbor toxin genes were recovered as mauve colonies on CHROMagar STEC.

McCallum et al. tested 282 fecal samples using PCR for toxin genes and culture on cefixime-tellurite-sorbitol MacConkey agar (CT-SMAC) and CHROMagar STEC (91). Only six positive samples were found, of which three were detected using CHROMagar STEC and only one detected using CT-SMAC, whereas all six were positive using PCR. Finally, Zeylas et al. tested 536 samples using direct inoculation of CHROMagar STEC and a PCR test for toxin genes following preenrichment in MacConkey broth (89). Thirteen samples were found to be positive, and all were detected using PCR. Eleven (84.6%) were detected by culture on CHROMagar STEC, and a further 68 false-positive colonies were recovered, giving a low positive predictive value of 13.9%.

The available evidence suggests that chromogenic media for the detection of STEC do not have sufficient sensitivity or specificity to replace methods that directly detect toxin or toxin genes in stool samples. Consequently most studies have concluded that the optimal use of such media is for the isolation of STEC from samples that are determined positive using more sensitive methods, e.g., PCR (86, 87, 89, 90).

Vibrio spp.

Media for the isolation of pathogenic Vibrio spp., which have been designed primarily for screening food samples, have been evaluated for use with clinical samples (15, 92). CHROMagar Vibrio and chromID Vibrio both allow for the isolation and differentiation of the two most important pathogenic species, Vibrio cholerae and Vibrio parahaemolyticus, and accommodate isolation of other Vibrio species. The chromogenic substrates used in these media are undisclosed by the manufacturers. CHROMagar Vibrio was compared with thiosulfate-citrate-bile salts-sucrose agar (TCBS) for the isolation of V. parahaemolyticus from 57 patients with suspected gastroenteritis who had recently ingested seafood. After enrichment in alkaline peptone water, five confirmed isolates of V. parahaemolyticus were recovered on CHROMagar Vibrio compared, with only one isolate on TCBS. There were no false positives on either medium (15). chromID Vibrio was evaluated against TCBS before and after enrichment in alkaline peptone water using 28 fecal samples and 66 artificially “spiked” fecal samples. There was equivalent sensitivity of the two media for isolation of both V. cholerae and V. parahaemolyticus, but the specificity of chromID Vibrio was reported to be twice that of TCBS (100% versus 50%) (92). Further studies are required with clinical samples, including studies in low-prevalence settings.

Yersinia enterocolitica

Y. enterocolitica is a foodborne pathogen and a cause of diarrhea and pseudoappendicitis. The most widely used conventional agar for detection of this pathogen in stool samples is cefsulodin-irgasan-novobiocin (CIN) agar, which utilizes mannitol fermentation as a biochemical indicator for Y. enterocolitica. CIN agar is effective, but its specificity is limited as a number of other species of Enterobacteriaceae are able to grow on the medium and ferment mannitol, thus also generating magenta colonies (e.g., Serratia spp., Providencia spp., Klebsiella oxytoca, and Citrobacter freundii). Weagant described the development of Yersinia enterocolitica chromogenic medium (YeCM) (93). This medium sought to improve on the specificity of CIN agar by utilizing cellobiose as the fermentable carbohydrate for detection of Y. enterocolitica and by including a chromogenic substrate for β-glucosidase, an enzyme that is produced by most other species of Enterobacteriaceae that are able to grow on CIN agar (93). As well as enabling differentiation of Y. enterocolitica from other species, this medium had the additional advantage that only pathogenic types of Y. enterocolitica would be detected (as nonpathogenic biovars produce β-glucosidase). There are no reports of the use of this medium with human clinical samples, although YeCM was used in a later study that examined the presence of Y. enterocolitica in 900 tonsil swabs from pigs (94).

Renaud et al. described the use of CHROMagar Yersinia (CAY) (Fig. 1e) for the isolation of Y. enterocolitica from 1,494 stool samples from hospitalized patients and used CIN agar as a comparator (17). Although the composition of this medium is undisclosed, the medium achieves outcomes very similar to those obtained with YeCM, suggesting the inclusion of a substrate for β-glucosidase to increase specificity (95). Six isolates of pathogenic Y. enterocolitica were successfully isolated using both CAY and CIN agar, but CAY showed a much higher specificity (99%) than CIN agar (90.4%), with only 14 false-positive isolates recovered on CAY (P < 0.001). In contrast, 137 isolates belonging to other species (predominantly C. freundii and Providencia spp.) grew as false-positive colonies on CIN agar, and there was no differentiation between the six pathogenic Y. enterocolitica isolates recovered and six additional nonpathogenic isolates of biovar 1A on CIN agar. Another chromogenic agar, Yersinia Enterocolitica Agar (YECA), has been described, but its composition is undisclosed and there are no data available for testing of human samples (96).

Methicillin-Resistant Staphylococcus aureus

MRSA strains exhibit resistance to most β-lactam antibiotics and frequently show resistance to other classes, such as macrolides and quinolones. MRSA is a significant nosocomial pathogen and has been implicated in numerous outbreaks of infection. Significant efforts have been made to control the spread of MRSA within the hospital environment, and some authorities advocate universal screening of patients to identify those who are asymptomatically colonized (97). A great deal of investment has therefore focused on optimal diagnostic methods for MRSA, including chromogenic media.

Chromogenic media for MRSA evolved from media designed for detection of all S. aureus by the simple inclusion of an antimicrobial that inhibits methicillin-susceptible isolates. The traditional agent of choice for this purpose was oxacillin, and this was incorporated into CHROMagar S. aureus to create the first chromogenic medium for detection of MRSA (5). Later studies demonstrated that cefoxitin was a superior option for selection of MRSA due to induction of methicillin resistance in isolates that showed heterogeneous expression of resistance (98). A range of chromogenic media has been commercialized and evaluated with clinical samples, including CHROMagar MRSA (99), chromID MRSA (100), MRSASelect (101), MRSA-screen (100), Brilliance MRSA (102), and Spectra MRSA (103). Such media are widely used; for example, in an external quality assessment exercise in 23 countries in Europe and in Israel, it was reported that 88% of laboratories utilized a chromogenic medium alone to screen for MRSA (104).

A survey of the literature since 2006 reveals at least 60 publications that have compared chromogenic media for MRSA with alternative methods using human clinical samples, and at least 25 of these include a head-to-head comparison of two or more chromogenic media. The conclusions of these studies are often conflicting, and an analysis of the data is further confounded by the fact that certain brands of media have been the subject of incremental improvements to produce newer generations of these media. When assessing the performance of chromogenic media for MRSA, it is prudent to analyze studies that are recent and include large numbers of clinical samples (ideally >1,000 samples).

There is a general consensus that chromogenic media show greater sensitivity than conventional agars such as mannitol salt agar plus oxacillin. The sensitivity of chromogenic agars is increased by incubating plates for 48 h, at the expense of a decrease in specificity. These observations are confirmed by a meta-analysis of 29 studies reported between 2004 and 2008 (105). Table 4 summarizes the findings of the five largest evaluations published since 2010 that compare two or more chromogenic media using clinical samples (106–110). From this selection of data and from the wider literature, it is difficult to conclude with any certainty that any particular chromogenic medium is superior or inferior to its competitors. A number of studies have shown conclusively that the use of broth enrichment prior to inoculation onto chromogenic agar can significantly increase the sensitivity of MRSA detection, and this is exemplified by two of the studies in Table 4 (109, 110).

Vancomycin-Resistant Enterococci

chromID VRE was the first chromogenic medium reported for the isolation of enterococci with acquired resistance to glycopeptides (e.g., vancomycin and teicoplanin) (8). In six studies with stool samples or rectal swabs, chromID VRE was compared with bile-esculin agars (with or without azide) supplemented with vancomycin (8, 111–115). Two of these studies reported a superior sensitivity of chromID VRE (8, 115), and the others reported equivalent sensitivity (111–114). A consistent advantage of chromID VRE was its ability to differentiate between Enterococcus faecalis and Enterococcus faecium. This is achieved by incorporating chromogenic substrates for detection of α-glucosidase and β-galactosidase (18). chromID VRE also showed a higher specificity, leading to a reduction in the number of confirmatory tests that were required for suspect colonies. Five other chromogenic media, AES VRE agar (115), Brilliance VRE (Fig. 2a) (116), CHROMagar VRE (117, 118), Spectra VRE (119–121), and VRESelect (122), have since been evaluated against bile-esculin agars (with or without azide) supplemented with vancomycin, in studies with stool samples or rectal swabs. In each of the eight cited studies, the chromogenic media offered higher sensitivity and specificity. All except AES VRE agar provided differentiation of E. faecalis from E. faecium.

Relatively few studies have included head-to-head comparisons of chromogenic agars for the isolation of VRE from clinical samples (115, 123–125). CHROMagar VRE and chromID VRE were evaluated with 259 stool samples after an overnight enrichment step and a 48-h incubation period (123). The authors reported an identical sensitivity of the two media (98.2%) and an equivalent specificity. Suwantarat et al. performed a large study that compared five chromogenic agars with bile-esculin-azide-vancomycin agar (BEAV) using 400 stool samples, of which 99 contained VRE (124). The chromogenic media comprised InTray Colorex VRE, chromID VRE, VRESelect, HardyCHROM VRE, and Spectra VRE. The authors reported a significantly higher sensitivity (89.9 to 93.9%) of all chromogenic agars than of BEAV (84.8%) and also an earlier time to detection. chromID VRE showed the highest sensitivity of the five chromogenic agars, but this was not statistically significant. Four of the chromogenic agars showed identical specificity (99.7%), whereas the specificity of InTray Colorex VRE was lower (98.3%). Differences were noted among the chromogenic media with respect to time to detection, the need for supplementary testing, and ease of color distinction among colonies, as well as non-VRE breakthrough growth. Finally, Gouliouris et al. compared chromID VRE and Brilliance VRE for the isolation of VRE from 295 stool samples from nursing home residents (125). They reported an equivalent sensitivity of the two chromogenic media and noted that the sensitivity of both media was significantly improved by incubation for 48 h and inclusion of a preenrichment step. The selectivity of Brilliance VRE was higher than that of chromID VRE.

In conclusion, chromogenic media for detection of VRE offer a sensitivity that is frequently reported as better than that of conventional media such as BEAV. Unlike BEAV, most chromogenic media allow differentiation and presumptive identification of the two dominant species of enterococci, and most reports note that less time is required for processing of colonies due to the higher selectivity/specificity of these media. There is no clear difference in sensitivity among most of the chromogenic agars reported here.

Extended-Spectrum-β-Lactamase-Producing Enterobacteriaceae

Extended-spectrum β-lactamases (ESBLs) have become globally disseminated within species of Enterobacteriaceae. These enzymes hydrolyze third-generation cephalosporins, typically conferring resistance to these agents, but are inhibited by clavulanic acid. Genes encoding ESBLs are frequently found on transmissible plasmids that often harbor other resistance determinants (e.g., encoding resistance to aminoglycosides). Prompt and accurate detection of ESBL producers can assist in guiding optimal antimicrobial therapy (126) and limiting nosocomial transmission (127). The first published report of a chromogenic medium for detection of ESBL producers described the evaluation of ESBL-Bx (a prototype of chromID ESBL). The authors compared this new medium with MacConkey agar supplemented with 2 μg/ml ceftazidime for the isolation of ESBL producers from 644 clinical samples (7). They reported a sensitivity of 97.7% for the chromogenic medium, compared with 84.1% for MacConkey agar plus ceftazidime. Since this first report, a range of chromogenic media has been made commercially available, and reports of their performance are summarized in Table 5 (128–134). The principles of these media have much in common. The media typically employ a combination of chromogenic substrates to detect β-galactosidase (or β-glucuronidase) production by E. coli and β-glucosidase production by the KES group. This allows detection and differentiation of the main species or groups associated with ESBL production. A cephalosporin is incorporated to inhibit susceptible strains, and other inhibitors may be included to inhibit the growth of Gram-positive bacteria and yeasts. Enterobacteriaceae with AmpC β-lactamase are frequently isolated on all of these media, and to a large degree this accounts for the relatively low positive predictive values shown in Table 5 (133).

Five of the studies in Table 5 include a direct comparison of two or more chromogenic media, and no significant difference in sensitivity was reported after 24 h of incubation in any of these studies (129, 131–134). The yield of ESBL producers may be increased significantly (particularly on chromID ESBL) by extending incubation up to 48 h (128, 134), and a preenrichment step may also significantly increase sensitivity (134). However, both of these strategies contribute to an even lower specificity (132, 134). In three of the studies in Table 5 the positive predictive value of chromogenic media was reported to be significantly higher than that of nonchromogenic comparators (128, 130, 132).

Carbapenemase-Producing Enterobacteriaceae

Perhaps the most significant development in clinical bacteriology over the last decade has been the global proliferation of Enterobacteriaceae with acquired carbapenemase enzymes (carbapenemase-producing Enterobacteriaceae [CPE]) (135). Available data suggest that the vast majority of carbapenemases in Enterobacteriaceae belong to one of the five major families: the IMP, NDM, and VIM metalloenzymes and the Klebsiella pneumoniae carbapenemase (KPC) and OXA-48-like enzymes (136). CPE are frequently resistant to virtually all β-lactam antibiotics (including carbapenems), and carbapenemase genes are transmissible via plasmids. Concomitant resistance to several other antimicrobial classes is common in CPE, dramatically reducing treatment options for infected patients. CPE are frequently implicated in nosocomial outbreaks, and fecal carriage of CPE is an important reservoir for subsequent transmission (137). In early 2009, the Centers for Disease Control and Prevention (CDC) and the Healthcare Infection Control Practices Advisory Committee recommended measures for active surveillance of Enterobacteriaceae with resistance to carbapenems in all acute-care facilities in the United States (138). A practical screening method was also recommended by the CDC, which involved inoculation of rectal swabs into 5 ml of Trypticase soy broth (TSB) along with a 10-μg ertapenem or meropenem disk (139). After incubation, the broth is subcultured to MacConkey agar, and lactose-fermenting colonies are investigated for carbapenemase production or carbapenem resistance. Detection of CPE is not straightforward, as not all carbapenemases confer clinical resistance to carbapenems and Enterobacteriaceae without carbapenemases may exhibit resistance to carbapenems via other mechanisms (140).

In 2008, Samra et al. evaluated CHROMagar KPC, the first chromogenic culture medium for detection of CPE (9). The medium was an adaptation of CHROMagar Orientation with additional selective agents for inhibition of Gram-positive bacteria and carbapenem-susceptible Gram-negative bacteria. CHROMagar KPC is also available as prepoured plates and is marketed as Colorex KPC (141). The authors compared culture on CHROMagar KPC with culture on MacConkey agar (plus carbapenem disks) and a PCR method for detection of bla_KPC genes. The sensitivity and specificity relative to PCR were 100% and 98.4%, respectively, for CHROMagar KPC and 92.7% and 95.9%, respectively, for MacConkey agar with carbapenem disks. Since this first report, a number of other chromogenic media have been made commercially available for detection of CPE. As with media for the detection of ESBL producers, the principles of these media have much in common, and they allow for the differentiation of the most relevant Enterobacteriaceae (i.e., E. coli and the KES group) as well as incorporating selective agents to inhibit Gram-positive bacteria and yeasts. The choice (and concentration) of antimicrobial(s) selected for inhibition of carbapenem-susceptible Enterobacteriaceae is the most critical factor that influences sensitivity and specificity, and the inclusion of antimicrobials other than carbapenems may have advantages over the inclusion of a carbapenem (S. Ghirardi, J. D. Perry, and G. Zambardi, 4 October 2012, international patent application WO2012131217; L. Devigne, S. Ghirardi, and B. Zambardi, 30 January 2014, international patent application WO2014016534). However, the ingredients of commercially available chromogenic media are typically undisclosed.

There have been a number of published studies that have examined the limit of detection of chromogenic agars for various types of CPE using challenge experiments with pure cultures (142–151). While such studies are useful, they cannot replicate the challenges that may be encountered when seeking to recover CPE from patient samples. Consequently, the values for sensitivity and specificity reported in evaluations with clinical samples are usually lower than those achieved with pure cultures (140). Table 6 summarizes the findings of the largest studies that have been performed with fecal samples (stool samples or rectal swabs) from patients (9, 152–159). In three early studies with CHROMagar KPC (9, 152, 153), sensitivity was equivalent to or higher than that of in-house preparations of MacConkey agar incorporating imipenem (or MacConkey agar with carbapenem disks). Later studies revealed that isolates producing NDM-1 carbapenemase may be inhibited by CHROMagar/Colorex KPC if the isolates have a relatively low MIC to meropenem (≤2 μg/ml) (141, 143). This emphasizes the importance of performing studies in different geographical areas where different carbapenemase enzymes may predominate. In a further example, chromID CARBA was proven to be effective in comparative studies of media in Greece (154, 156) and Pakistan (141, 160, 161) but had limited efficacy in Turkey (157) and Belgium (162), where OXA-48 is the dominant carbapenemase. To address this, the manufacturer offers a biplate (chromID CARBA SMART) that combines two media in a single petri dish: chromID CARBA and chromID OXA-48.

Brilliance CRE agar is marketed for the isolation of carbapenem-resistant Enterobacteriaceae (CRE) rather than CPE. This medium was evaluated in two simultaneous studies in two centers in Pakistan and showed a lower sensitivity and specificity than chromID CARBA for detection of CPE with NDM-1 carbapenemase (160, 161). The authors speculated that the stability of Brilliance CRE may have been compromised during transport of the culture media from the UK to Pakistan. However, in challenge experiments with pure cultures, a relatively low specificity of 71% was recorded for this medium due to the growth of AmpC- and/or ESBL-producing isolates (146). In another study with Brilliance CRE with patient samples (n = 77), the specificity of Brilliance CRE was lower than that of SuperCarba (86.6% versus 98.5%, respectively) (149). However, the sensitivity of Brilliance CRE was significantly better than that of chromID CARBA for detection of OXA-48 as shown by a study in Belgium (162). The sensitivity of both media was improved by preenrichment of rectal swabs in MacConkey broth (although this was not statistically significant), whereas prolonged incubation of media for 48 h showed no advantage.

SuperCarba medium is a nonchromogenic medium for isolation of CPE that incorporates a low concentration of ertapenem (0.5 μg/ml) in addition to cloxacillin in a zinc-supplemented Drigalski-lactose agar (163). Studies with bacterial isolates suggest a high sensitivity for detection of all types of CPE, but stability of the medium is limited to 1 week, and larger studies with patient samples are required. A chromogenic adaptation of this medium, CHROMagar mSuperCarba (also marketed as Colorex mSuperCarba) (Fig. 2d), has recently been made commercially available and shows a comparable sensitivity when tested with pure isolates of CPE (164). The medium, once prepared, has an improved shelf life of 1 month. Davies et al. (158) recently reported an evaluation of Colorex mSuperCarba with rectal swabs and reported a sensitivity equivalent to that of chromID CARBA and Brilliance CRE for recovery of CPE that mostly produced NDM carbapenemase (Table 6). García-Fernández et al. cultured 211 rectal swabs from distinct patients onto CHROMagar mSuperCarba and compared its performance with culture on chromID CARBA, chromID OXA-48, and chromID ESBL (165). CPE were detected in 61 samples (with OXA-48 reported as the dominant enzyme; n = 54), and the authors reported 100% sensitivity and specificity for CHROMagar mSuperCarba. The sensitivities of the comparator media were not reported.

In the three evaluations in Table 6 that included the CDC broth method, the sensitivity of chromogenic media was equivalent (154, 156) or better (157). Disadvantages of the CDC broth method include a longer time to detection (an extra day is required), lack of information regarding likely species identification, and the fact that CPE may not be detected if they fail to ferment lactose on MacConkey agar. A further disadvantage is that the CDC broth method is significantly more labor-intensive than use of a chromogenic medium, as shown by Mathers et al. (166).

It can be concluded that chromogenic media for CPE have clear advantages over the CDC broth method or the use of MacConkey agars supplemented with a carbapenem or used in conjunction with carbapenem disks. However, it is especially difficult to establish which, if any, chromogenic medium is optimal for detection of CPE due to the different types of carbapenemase that may be encountered and the dominance of particular types in certain geographical regions. In January 2016, Viau et al. published an extensive evaluation of all methods for the detection of CPE (140). Following a detailed statistical meta-analysis of published studies, the authors concluded that chromID CARBA and the nonchromogenic SuperCarba medium have excellent sensitivities for class A β-lactamases (e.g., KPC) that rival that of real-time PCR and that the CDC broth method was generally inferior to chromogenic media. For other media, and other carbapenemase types, there was insufficient evidence to draw firm conclusions.

Carbapenem-Resistant Acinetobacter spp.

Species of Acinetobacter, and especially Acinetobacter baumannii, are important nosocomial pathogens that have been responsible for outbreaks of infection among hospitalized patients. Strains that are resistant to carbapenem antibiotics are particularly problematic, as infections caused by such strains can be very difficult to treat (167). In 2009 CHROMagar Acinetobacter, the first chromogenic medium for detection of Acinetobacter spp., was described, and this medium can be used to detect all Acinetobacter spp. or it can be further supplemented to detect only isolates with resistance to carbapenems (10). The chromogenic substrate used for detection of Acinetobacter spp. is undisclosed by the manufacturer. The authors evaluated the supplemented version of CHROMagar Acinetobacter for detection of carbapenem-resistant strains from stool samples and perineal swabs of 70 patients and compared its performance to PCR following enrichment culture. The sensitivity and specificity of the chromogenic medium compared with the PCR assay were reported as 91.7% and 89.6%, respectively (10).

There have been few subsequent published evaluations of CHROMagar Acinetobacter with clinical samples for the isolation of carbapenem-resistant isolates. Song et al. used an updated version of the medium (Fig. 1f) for isolation of carbapenem-resistant Acinetobacter from 406 paired nasal and rectal swabs (168). They confirmed the high specificity of the medium for carbapenem-resistant Acinetobacter, but no comparator was used for assessment of sensitivity. In a study using spiked stool samples, CHROMagar Acinetobacter had a reported sensitivity of 86.5% and a specificity of 75% for detection of carbapenemase-producing A. baumannii (169).

One question that deserves further attention is whether a single chromogenic medium could be used by clinical laboratories to perform screening for both carbapenemase-producing Enterobacteriaceae and carbapenem-resistant Acinetobacter (170, 171). For example, Higgins et al. reported that a high proportion of carbapenem-resistant Acinetobacter strains with a diverse range of carbapenemase enzymes grow on chromID CARBA and that the medium effectively inhibits the growth of carbapenem-susceptible isolates (170). Zarakolu et al. evaluated CHROMagar Acinetobacter and chromID CARBA for the isolation of carbapenem-resistant Acinetobacter with 203 stool samples from patients at a hospital in Turkey (171). There was no significant difference between the two media, with sensitivities of 66% and 77%, respectively (P = 0.48) and comparable values for positive predictive value (62% and 68%). However, all isolates belonged to a single species (A. baumannii), and all produced OXA-23 carbapenemase. Further studies are therefore warranted for evaluation of CHROMagar Acinetobacter in different geographical areas, and it would be useful if further studies can establish whether separate media are necessary for screening for CPE and carbapenem-resistant Acinetobacter.