INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Identification of organisms of the family Enterobacteriaceae constitutes a considerable proportion of the workload in many clinical microbiology laboratories. The primary goal in a diagnostic microbiologiy laboratory is the rapid identification of a pathogen. The traditional methodology for routine detection of enteric pathogens nearly always employs a combination of at least three different media in order to increase the sensitivity and the specificity of the detection and identification method. This is due to the fact that if one medium is highly inhibitory to some members of the Enterobacteriaceae, e.g., Escherichia coli, there is a simultaneous loss of sensitivity for fastidious pathogens such as the Shigella spp. (21). At the Statens Serum Institut (SSI; Copenhagen, Denmark), a medium that combines differentiating and selective properties without loss of the ability to detect has been developed. For more than 20 years, the SSI enteric medium alone has been used for detection of enteric pathogens (with the exception of Campylobacter spp. and anaerobic bacteria) in our reference laboratory, where 75,000 to 80,000 stool samples are investigated each year. Although the SSI enteric medium has been used succesfully for such a long time in typical clinical microbiology laboratories in Denmark, it has never been evaluated in tests designed for that purpose. The aims of this study were to evaluate the SSI enteric medium and to compare its ability to detect and identify enteric pathogens to those of other standard enteric media. Semiquantitative growth, the ability of the medium to produce distinctive biochemical reactions, and the detection limits in stools were evaluated. Also, quantitative growth on SSI medium was compared with that on other enteric media.

(This work was presented at the 98th General Meeting of the American Society for Microbiology, Atlanta, Ga., 17 to 21 May 1998 [12a].)

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains. A total of 275 different strains, representing 28 genera, were used for the evaluation of SSI enteric medium. The strains were isolated from clinical specimens at the SSI, and they were chosen to include the whole spectrum of gram-negative rods (n = 264) with the addition of a number of gram-positive bacteria (n = 11) as controls to examine the selectivity of the medium (Table 1). The bacteria were identified according to routine procedures. Only major pathogenic species were evaluated for their detection limits in stools and compared for their quantitative growth on the different enteric media. The 275 strains represented a pool from which the bacteria for each investigation were selected.

SSI enteric medium. The SSI enteric medium is available both as agar plates and granulated powder. The selective action on SSI enteric medium is caused by the combination of a detergent (sodium dodecyl benzene sulfonate) and sodium deoxycholate, on which gram-positive bacteria, gram-negative cocci, and certain gram-negative rods (e.g., Pasteurella and Haemophilus) are excluded from growth. Also, the detergent is an efficient inhibitor of the swarming of nearly all Proteus strains. Upon inoculation, the medium produces four differentiating reactions.

(i) H₂S reaction. Due to the presence of sodium thiosulfate and ferric citrate in the agar, H₂S-positive bacteria produce a distinct ferrosulfide precipitate located centrally and deeply in the more anaerobic parts of the colony.

(ii) Lactose fermentation. Fermentation is manifested by acid production made visible by using as an indicator the combination of deoxycholate and neutral red. The acid production is confined to the colony and its immediate surroundings due to the reaction with deoxycholate, which yields deoxycholic acid. Deoxycholic acid precipitates and becomes moderately colored by neutral red. The confinement of the reaction to the colony makes recognition of adjacent lactose-negative colonies possible.

(iii) Phenylalanine deaminase reaction. Phenylalanine deaminase-active bacteria are recognized. With ferric citrate in an acid environment, the conversion of L-phenylalanine to phenylpyruvate is visible as a brown, diffusible pigment in the SSI enteric medium.

(iv) Indole reaction. This reaction is included as a differentiating characteristic in a pure subculture from a suspect colony on the primary medium. Deamination of tryptophan produces a volatile chemical, indole, which may be detected by placing a piece of filter paper inside the lid of the inverted petri dish that has been wetted with the reagents of Ehrlich's indole test, i.e., p-dimethylaminobenzaldehyde, ethyl alcohol, and hydrogen chloride.

View larger version (103K): [in this window] [in a new window]   FIG. 1.   Colony appearance of four different genera of the Enterobacteriaceae on SSI enteric medium. SSI enteric medium yields six differentiating biochemical reactions that allow direct identification of a range of enteric pathogens (see the text). In this figure, only the four major reactions are illustrated, with different bacteria. (A) Shigella sonnei, with the characteristic rough transformations, causes the surfaces and the edges of the colonies to become serrated. (B) E. coli and Shigella sonnei. E. coli organisms appear as red colonies due to lactose production, and there is no rough transformation. (C) E. coli and Y. enterocolitica. Yersinia spp. appear as tiny round translucent colonies which resemble pearls on a string (indicated by the arrow). They are easily recognized compared to the much larger, red colonies of E. coli. (D) H2S-positive Salmonella spp. H2-positive organisms appear with a black precipitate located centrally and deeply in the anaerobic parts of the colony.

(i) Metallic sheen reaction. This reaction is characterized by a metallic sheen on top of the black center in H₂S-producing Salmonella (except in Salmonella typhi). The cause of this metallic sheen is largely unknown.

(ii) Rough transformation. Due to rather high concentrations of divalent ions (Mg⁺⁺ and Ca⁺⁺), a rough transformation is regularly induced in Shigella sonnei, affecting the surfaces and the edges of the colonies. The resulting increased spreading of the colonies makes recognition possible also when they are located among colonies of other bacteria.

Other media. (i) XLD agar. Xylose-lysine-desoxycholate (XLD) agar (Oxoid, Basingstoke, United Kingdom) was originally formulated by Taylor for the isolation and identification of Shigella spp. from stool specimens (17). It has since been found also to be a satisfactory medium for Salmonella spp. (8). XLD agar contains deoxycholate as the selective agent and identifies the bacteria according to lactose, sucrose, and xylose fermentation; H₂S production; and lysine decarboxylation.

(ii) HE agar. Hektoen enteric (HE) agar (Oxoid) was developed by King and Metzger (10) to grow shigellae as readily as other pathogens while inhibiting normal intestinal flora. HE agar contains bile salts and may contain novobiocin as selective agent; it identifies the bacteria according to lactose, sucrose, and salicin fermentation and H₂S production.

(iii) SS agar. Salmonella-Shigella (SS) agar (Becton Dickinson, Basel, Switzerland) is a differential selective medium for the isolation of Salmonella and some Shigella species from clinical speciments. It contains brilliant green and bile salts as selective agents and identifies bacteria according to lactose fermentation and H₂S production.

(iv) CIN agar. Cefsulodin-irgasan-novobiocin (CIN) agar (Oxoid) was developed by Schiemann (15) for the isolation of Yersinia enterocolitica. It contains bile salts and antibiotics as selective agents and identifies the bacteria by mannitol fermentation.

Confirmatory reactions. H₂S production was investigated by the addition of several colonies into a ferrocloride-peptone-gelatin medium (SSI). After incubation at 22°C for a maximum of 4 days, H₂S-positive bacteria were black or grey along the stab line (12). We tested for phenylalanine deaminase by the addition of several colonies into a culture medium containing L-phenylalanine (SSI) followed by incubation at 35°C for 2 to 4 h and the addition of a few drops of reagent (FeCl₃ · 6H₂O and 5 N HCl). Phenylalanine deaminase-active bacteria give an immediate and strong dark green coloring of the culture medium (2). Lactose and indole fermentation was confirmed by using standard procedures. The indole reaction was performed after xylene extraction.

Growth and reactivity of gram-negative rods on the SSI enteric medium. Bacteria were grown overnight on 5% blood agar (SSI). Several colonies were streaked on the SSI enteric medium and incubated overnight at 35°C. The plates were examined for the colony characteristics, size, and color of the test organisms, and the number of colonies on the medium was determined. Isolates showing unexpected reactions on the SSI enteric medium were retested further in confirmatory reactions by using the standard procedures described earlier.

Semiquantitative growth. A total of 59 bacterial strains were tested; 48 represented the relevant gram-negative organisms, and 11 were gram-positive controls. Bacteria of each strain were suspended in saline and were adjusted to approximately 2 × 10⁸ CFU/ml as measured by optical density (540 nm) using a colorimeter (Ciba-Corning Colorimeter 254; Ciba Corning Diagnostics Ltd., Suffolk, United Kingdom). The bacterial suspensions were diluted in saline to produce 10¹ to 10⁷ suspensions, and 200 µl of each (including the undiluted bacterial suspension) was plated onto 5% blood agar (SSI) and the SSI enteric medium. The bacteria were categorized as having the same growth on both media, less growth on SSI enteric medium than on 5% blood agar, or no growth. Same growth indicates that the number of colonies on the SSI enteric medium was between 10 and 120% of the number on the 5% blood agar, whereas less growth was between 0.1 and 10%.

Detection limit for enteric pathogens in simulated stool cultures. The most common enteric pathogens (a total of 28 strains) were selected for this experiment. The bacteria in question were suspended in saline, adjusted to approximately 2 × 10⁸ CFU/ml and diluted in saline from 10¹ to 10⁷. The diluted bacteria were mixed with normal feces (e.g., fecal samples with no growth of enteric pathogens) by suspending 10 µl of feces into 2 ml of phosphate-buffered saline (pH 7.38) and adding 5 µl of each dilution of bacteria. The suspension was mixed, and 10 µl was plated on SSI enteric medium and incubated at 35°C overnight. The highest dilution resulting in growth on the medium was determined.

Comparison with other enteric media. A total of 125 strains belonging to the major pathogenic species were tested on XLD, CIN, HE, and SS agar, and the results were compared with those obtained with SSI enteric medium. The bacteria in question were suspended and diluted so that an inoculum of ~200 CFU was applied. Each of the five media were inoculated on the same day by the same technician. After incubation overnight at 35°C the number of colonies was counted.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Growth and reactivity of gram-negative rods on the SSI enteric medium. All tested bacteria (n = 248) grew on SSI enteric medium and very few (10 of 248 [4%]) showed discrepant biochemical reactions (H₂S production, lactose fermentation, phenylalanine deaminase production, and indole production) (Table 3). Five strains (Klebsiella pneumoniae, Providencia rettgeri, and three Enterobacter cloacae strains) gave a false-negative reaction for lactose fermentation. Three other strains (Citrobacter freundii, Proteus vulgaris, and Proteus mirabilis) did not produce H₂S, and one strain (Alcaligenes piechaudii) did not produce phenylalanine deaminase. However, all of these strains expressed the expected reactions in confirmatory tests.

Semiquantitative growth. Most gram-negative rods (42 of 47 tested [92%]) grew with little or no inhibition on SSI enteric medium compared to growth on 5% blood agar (Table 4). Three Vibrio spp. (one each of V. vulnificus, V. cholera O139, and V. parahaemolyticus) and an Aeromonas sobria strain showed reduced growth (10- to 1,000-fold). The gram-positive bacteria did not grow on SSI enteric medium.

Detection limit for enteric pathogens in simulated stool cultures. Most enteropathogens (16 of 28 [57%]) had a detection limit of 50 bacteria/ml of feces except for Vibrio spp. and some Shigella spp. (Table 5). Of the five Vibrio spp., one was not detected (V. cholera non-O1), two had a detection limit of 5,000 bacteria/ml of feces (V. vulnificus and V. cholera O139), and two had a detection limit of 50,000 CFU/ml of feces (V. parahaemolyticus and V. cholera O1). The strains with the high detection limits were the same as the ones showing reduced growth in the semiquantitative growth experiment. Of the four Shigella spp. tested, one had a detection limit of only 5 CFU/ml of feces (S. sonnei) and three had a detection limit of 500 CFU/ml of feces (two S. flexneri and S. boydii).

Growth on SSI enteric medium compared to that on standard enteric media. Growth on SSI enteric medium was as good as or better than that on other standard enteric media (Table 6). This was particularly evident for Shigella spp. and enteropathogenic E. coli. Possible exceptions were Vibrio spp., which grew slightly better on HE agar than on all other media. The SSI enteric medium also detected Yersinia spp. When an inoculum of ~200 CFU was applied to all five media and the quantitative growth rates were compared, SSI enteric medium showed >50 CFU in 75% of the tested bacteria compared to 48% for XLD medium, 75% for HE medium, and 71% for SS agar (Table 6).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

The SSI enteric medium evaluated in this study provided the most effective universal medium for the recovery of enteric pathogens. Since we could not evaluate all of the many different types of media available for the isolation of enteric pathogens, we decided to include in this study only those widely used in typical clinical bacteriology laboratories.

XLD agar performed second best, although it had a lower sensitivity towards Shigella spp. and E. coli than did the SSI enteric medium. Like Taylor and Schelhart (21), Altwegg et al. (1), and Desmond and Janda (3), we found XLD agar to be superior to HE agar and SS agar, although HE agar supported growth of Vibrio spp. best of all tested media. This may explain why some investigators have reported HE agar to be equal to or slightly superior to XLD agar (3, 4, 8). Even though this study showed reduced growth and high detection limits of some Vibrio spp. on the SSI enteric medium, the medium detected V. cholerae O1-positive stools in a 1-year prospective household-based diarrhea study in Guinea-Bissau (13). The SSI enteric medium also proved its utility during a V. cholerae epidemic in September 1987 in Guinea-Bissau, although the sensitivity of detecting V. cholerae was increased with the additional use of thiosulfate-citrate-bile-sucrose (TCBS) agar (12b). All cases of cholera diagnosed in the Department of Gastrointestinal Infection at the SSI during the past 20 years have been diagnosed as fast as or faster on the SSI enteric medium than on the traditional TCBS medium after preenrichment in peptone broth. This is probably due to the extremely high numbers of organisms present in stools of patients with cholera.

Some investigators have questioned the efficiency of CIN agar for its use in detection of Yersinia enterocolitica (6, 7, 9). As the SSI enteric medium allows growth of all enteric pathogens, including Yersinia spp., CIN agar could be removed from the list of standard plating agars in the typical laboratory and thereby diminish workload and cost to the laboratory. Yersinia spp. grow as easily recognizable tiny round translucent colonies which resemble pearls on a string on SSI enteric medium.

The performance of any medium is dependent upon the skill and training of the technician and his or her familiarity with the medium. To fully benefit from the capacity to directly identify enteric pathogens on SSI enteric medium, the recognition of the metallic-sheen reaction and the rough transformation reaction in particular requires practical experience.

In conclusion, SSI enteric medium is selective for gram-negative bacteria and allows growth of all aerobic enteric pathogens. The medium compares favorably to similar media, and its ability to detect Yersinia spp. would limit the number of media needed in a typical laboratory. Also, the direct identification of enteric pathogens on the SSI enteric medium may provide a more rapid diagnosis.