Novel Selective Medium for Isolation of Staphylococcus lugdunensis from Wound Specimens

Abstract

We compared a novel selective Staphylococcus lugdunensis (SSL) medium with routine media (blood and chocolate agars) for the detection of S. lugdunensis in 990 clinical specimens (from tissue, pus, or wound swabs). Significantly more S. lugdunensis isolates were detected on SSL medium (34/990) than on routine medium (7/990) (P = 0.001, McNemar's test).

TEXT

While Staphylococcus lugdunensis is a coagulase-negative staphylococcus (CoNS), infections caused by this organism are similar in type and severity to those caused by Staphylococcus aureus (1, 2). One study reported that S. lugdunensis was found in 13% of cutaneous abscesses and 4.5% of pustules but in only ≤1% of infections other than abscesses and pustules, such as impetigo, unspecified wound infections, and incisional surgical wound infections (3). The rates of S. lugdunensis detection in wound specimens were also affected by laboratory protocols, choice of culture medium, and methods for the identification of CoNS (3–5). Although progress has been made in the detection and identification of this organism (3, 5, 6), the recognition of its presence in a plate with mixed growth of commensals or other pathogens remains challenging. Notably, one study reported that S. lugdunensis was isolated as part of mixed flora in 60% of the wound specimens tested (7).

In an attempt to improve the detection of S. lugdunensis in wound specimens, a novel medium was designed. This selective S. lugdunensis (SSL) medium comprises (per liter) 54.2 g Giolitti-Cantoni broth, 45 g sodium chloride, 10 g l-ornithine monohydrochloride, 15 g agar bacteriological (Oxoid agar no. 1), 1 g potassium nitrate, 0.1 g deferoxamine mesylate (Sigma), 0.01 g uracil, 0.005 g pyridoxal hydrochloride, 0.005 g hemin, and 0.0005 g vitamin K1. Bromocresol purple (0.02 g/liter) and phenol red (0.01 g/liter) were added as pH indicators, and the final pH was titrated to 6.8 with hydrochloric acid. We assumed that the ornithine decarboxylase (ODC) of S. lugdunensis would elevate the pH of the medium and change the color of the indicator to purple, thereby facilitating the recognition of this organism in mixed growth, whereas sodium chloride and deferoxamine mesylate inhibit the growth of, respectively, common salt-intolerant bacteria and Staphylococcus epidermidis (for which ODC activity is present in a minority of strains) (1, 5, 8).

The selectivity of the SSL medium was challenged with bacteria that are commonly found in wound specimens. In total, 208 isolates, including 110 S. lugdunensis clinical isolates (each from a different patient), 58 other staphylococcus isolates, and 40 isolates from other bacteria were plated onto SSL medium and incubated anaerobically for 48 h (Table 1). The collection included 34 ATCC strains, and 174 were clinical isolates which had been identified at the species level (6, 9). All S. lugdunensis cultures grew after 24 h, and easy recognition of the typical grayish colonies with purple halos (i.e., discoloration of the medium surrounding the colonies) often required incubation for 48 h (Fig. 1). For the majority of the other bacteria, growth was either completely inhibited or poor, with only pinpoint colonies after 48 h of incubation. Several staphylococci (S. aureus, S. arlettae, S. haemolyticus, S. sciuri, S. simulans, and S. xylosus) grew large colonies (with 2- to 3-mm diameters) in SSL medium after 48 h, but no purple halos were observed around the colonies. A total of 32 S. epidermidis isolates were tested (of which, 11 isolates were ODC positive). All of them grew as a fine haze in the primary inoculum, with pinpoint colonies in the streaked-out sectors. While there was some purple discoloration over the fine haze of confluent growth for the ODC-positive isolates, no purple halos were observed around the isolated pinpoint colonies. Vibrio parahaemolyticus, which can grow under strict anaerobic conditions (10), was the only other bacterium that produced colonies with purple halos (Table 1). However, the colonies were bluish green and were readily distinguished from the grayish colonies of S. lugdunensis.

TABLE 1.

Growth characteristics of Staphylococcus lugdunensis and other bacteria in SSL medium

Isolate type and organism (no. of isolates)	No. of isolates	Growth on SSL agar after 48 h (no. of isolates)			Purple halos surrounding colonies on SSL agar (no. of isolates)
		Good	Poorf	Absent	Yes	No
Reference strains (34)a
Acinetobacter baumannii	2			2
Enterobacteriaceaeb	13		13			13
Enterococcus speciesc	4		4			4
Lactobacillus rhamnosus	1			1
Kocuria rosea	1			1
Pseudomonas aeruginosa	1			1
Brevundimonas diminuta	1			1
Staphylococcus aureus	3	3				3
Staphylococcus lugdunensis	1	1			1
Staphylococcus epidermidis	1		1			1
Staphylococcus hominis	1		1			1
Staphylococcus saprophyticus	1		1			1
Staphylococcus sciuri	1	1				1
Staphylococcus simulans	1	1				1
Stenotrophomonas maltophilia	1			1
Vibrio parahaemolyticus	1	1			1g
Clinical isolates (174)
Aeromonas hydrophila	1			1
Enterobacteriaceaed	3		1	2		1
Enterococcus speciese	2		2			2
Flavobacterium meningosepticum	1			1
Pseudomonas aeruginosa	1			1
Staphylococcus aureus	3	3				3
Staphylococcus lugdunensis	109	109			109
Staphylococcus arlettae	1	1				1
Staphylococcus epidermidis	31		31			31
Staphylococcus equorum	2		1	1		1
Staphylococcus haemolyticus	5	4	1			5
Staphylococcus kloosii	1			1
Staphylococcus saprophyticus	2		2			2
Staphylococcus sciuri	4	3	1			4
Staphylococcus xylosus	1	1				1
Streptococcus bovis	1			1
Vibrio alginolyticus	1	1				1
Vibrio parahaemolyticus	5		1	4		1
Total	208	129	60	19	111	78

^aSee Table S2 in the supplemental material for designations of the reference strains.

^bCitrobacter freundii (n = 1), Enterobacter cloacae (n = 3), Escherichia coli (n = 2), Klebsiella oxytoca (n = 1), Klebsiella pneumoniae (n = 2), Proteus mirabilis (n = 1), Proteus vulgaris (n = 1), Providencia stuartii (n = 1), and Salmonella enterica serovar Typhimurium (n = 1).

^cEnterococcus casseliflavus (n = 1), Enterococcus durans (n = 1), and Enterococcus faecalis (n = 2).

^dEscherichia coli (n = 1), Plesiomonas shigelloides (n = 1), and Serratia marcescens (n = 1).

^eEnterococcus faecium (n = 1) and Enterococcus gallinarum (n = 1).

^fPoor growth was defined by pinpoint-sized colonies after 48 h of incubation.

^gColony size was smaller than that of S. lugdunensis, and colony color was bluish green, while for S. lugdunensis it was gray.

FIG 1.

Wound culture on a plate of selective Staphylococcus lugdunensis (SSL) medium after 48 h of anaerobic incubation at 35°C. S. lugdunensis (black arrows) appears in large grayish colonies with purple discoloration of the surrounding medium because of ornithine decarboxylase activity, whereas other staphylococci, such as S. aureus (white arrows), appear in smaller colonies without purple discoloration of the medium.

To investigate the quantitative recovery of S. lugdunensis in the SSL medium, four strains of the organism were grown overnight and suspended in 0.9% normal saline to a density of 0.5 McFarland standard. The suspensions were serially diluted by the Miles and Misra method and plated onto SSL and blood agar (BA) plates. After 48 h of incubation, no significant differences in colony counts were found between the two media.

Finally, the performance of SSL medium in the detection of S. lugdunensis in clinical specimens was prospectively examined. From September 2013 to February 2014, 990 consecutive wound specimens (188 pus, 201 tissue, and 601 wound swabs) sent to two clinical laboratories (776 specimens to laboratory A and 214 specimens to laboratory B) for bacteriological investigation were screened for the presence of S. lugdunensis. In the two laboratories, wound specimens were routinely inoculated onto three culture plates (horse BA and chocolate agar for aerobic culture and a neomycin BA plate for anaerobic culture). Besides the routine setup, all specimens were inoculated onto one SSL plate and incubated anaerobically. Routinely set up plates were processed by the on-duty laboratory technicians in the usual manner, while the SSL plates were followed by a research student. The results of the cultures from routine processing and the SSL plates were kept blinded until the final analysis. The inoculated plates were incubated at 35°C for 24 h. When there was insufficient or no growth, the plates were reincubated for another 24 h. Colonies on the routine plates suggestive of S. lugdunensis (Eikenella corrodens-like odor, colony pleomorphism, and large β-hemolysis) and the SSL plates (grayish colonies with purple halos) were picked. Identification to the species level was achieved by slide agglutination (Staphaurex Plus, Remel) and tube coagulase, ODC, and pyrrolidonyl aminopeptidase testing, and confirmation was obtained by PCR (6, 11). Overall, S. lugdunensis was detected in 36 specimens (Table 2), of which 5 were detected by both routine and SSL plates, 2 were detected by routine plates only, and 29 were detected by SSL plates only (P < 0.001, McNemar's test). In most specimens (94.4%, 34/36), S. lugdunensis was detected as part of mixed growth (see Table S1 in the supplemental material). The sensitivities of the routine and SSL plates were 19.4% (7/36) and 94.4% (34/36), respectively. Although the presence of S. lugdunensis on the SSL plates was suspected in some cases after 24 h, incubation for 48 h was usually required for development of the typical colony morphology. In the two specimens with S. lugdunensis that were missed by SSL plates, only a few colonies were found in the routine plates. Inoculation problems may explain the missed results. Reasons for the inability of the routine medium to detect S. lugdunensis in 29 specimens included overgrowth by other competing bacteria, atypical colony morphology of S. lugdunensis, and the presence of >1 different CoNS in the growth. A lack of awareness by some technicians reading the routine plates may have also been a factor (5), especially when the SSL plates were followed by a trained student. Anaerobic incubation of the SSL plates is essential for the specific detection of ODC activity. False-positive reactions in the ODC test in aerobic incubation are well recognized (12). Moreover, anaerobic incubation is required to suppress the growth of S. epidermidis. Despite the addition of deferoxamine, we noted that some S. epidermidis isolates had large colonies in the SSL medium after 24 to 48 h of aerobic incubation.

TABLE 2.

Recovery of S. lugdunensis from SSL plates compared to that from conventional cultures

S. lugdunensis detected in conventional culture	S. lugdunensis detected in SSL plates	No. of specimens
Yes	Yes	5
Yes	No	2
No	Yes	29
No	No	954
Total		990

P < 0.001 (McNemar's test) for comparison of S. lugdunensis detection by the two methods

In conclusion, SSL medium supports the growth of S. lugdunensis, facilitates the differential recognition of the organism in specimens with mixed growth of other bacteria (especially when there were other staphylococci), and is significantly more sensitive than routine media for the primary isolation of this organism. SSL medium can be added to, but cannot replace, routine media for the isolation of other skin and soft tissue pathogens.