|Home | About | Journals | Submit | Contact Us | Français|
We compared recovery of Staphylococcus aureus and methicillin-resistant S. aureus (MRSA) from nasal and groin swab specimens of 600 HIV-infected outpatients by selective and nonselective direct plating and broth enrichment. Swabs were collected at baseline, 6-month, and 12-month visits and cultured by direct plating to mannitol salt agar (MSA) and CHROMagar MRSA (CM) and overnight broth enrichment with subculture to MSA (broth). MRSA isolates were characterized by pulsed-field gel electrophoresis (PFGE), staphylococcal cassette chromosome mec (SCCmec) typing, and PCR for the Panton-Valentine leukocidin. At each visit, 13 to 15% of patients were colonized with MRSA and 30 to 33% were colonized with methicillin-susceptible S. aureus (MSSA). Broth, CM, and MSA detected 95%, 82%, and 76% of MRSA-positive specimens, respectively. MRSA recovery was significantly higher from broth than CM (P ≤ 0.001) or MSA (P ≤ 0.001); there was no significant difference in recovery between MSA and CM. MSSA recovery also increased significantly when using broth than when using MSA (P ≤ 0.001). Among specimens collected from the groin, broth, CM, and MSA detected 88%, 54%, and 49% of the MRSA-positive isolates, respectively. Broth enrichment had a greater impact on recovery of MRSA from the groin than from the nose compared to both CM (P ≤ 0.001) and MSA (P ≤ 0.001). Overall, 19% of MRSA-colonized patients would have been missed with nasal swab specimen culture only. USA500/Iberian and USA300 were the most common MRSA strains recovered, and USA300 was more likely than other strain types to be recovered from the groin than from the nose (P = 0.05).
Methicillin-resistant Staphylococcus aureus (MRSA) is an important pathogen in both health care and community settings. According to the National Healthcare Safety Network (NHSN) annual update, S. aureus accounts for 15% of health care-acquired infections in the United States; 50 to 60% of these S. aureus isolates are MRSA (20). S. aureus is known to colonize the nares of approximately 30 to 35% of healthy persons, and estimates for MRSA colonization range from 1 to 9%, depending on the study group (11, 12, 18, 26, 27, 42). S. aureus colonization is associated with an increased risk of subsequent staphylococcal infection in patients in intensive care units and following hospitalization or surgery (22, 44, 53).
For more than a decade, community-associated MRSA (CA-MRSA) has been the leading cause of skin and soft tissue infections among healthy individuals and selected groups, including athletes, intravenous drug users, inmates, and men who have sex with men (1–5, 24, 29, 32, 38). The predominant pulsed-field gel electrophoresis (PFGE) type associated with CA-MRSA in the United States is USA300 (35), which typically contains the Panton-Valentine leukocidin (PVL) toxin and staphylococcal cassette chromosome mec (SCCmec) type IVa (SCCmec IVa) (13, 25, 52). MRSA carriage is increasing among persons in the community (18, 27), and recent reports have noted an increased isolation of USA300 S. aureus in health care settings (17, 21, 23, 39, 47).
In 2003, the Society for Healthcare Epidemiology of America (SHEA) recommended the collection of samples for surveillance cultures at hospital admission for patients at high risk for MRSA carriage (40). Although active surveillance for MRSA is recommended, there are no standard methods for the recovery of MRSA from surveillance cultures. The anterior nares are considered the primary reservoir for S. aureus colonization; however, several studies have suggested that CA-MRSA may preferentially colonize other body sites, including throat, tonsils, rectum, and groin; this has been noted in specific groups, including infants, children, and HIV-infected individuals (6, 7, 14, 15, 37, 42, 51).
As part of a study of MRSA colonization in HIV-infected outpatients (45), we compared recovery of MRSA and methicillin-susceptible S. aureus (MSSA) from nasal and groin swab specimens taken at three different collection times over a 1-year period and cultured with selective and nonselective direct plating methods and a broth enrichment method. MRSA isolates were characterized by PFGE, SCCmec typing, and PCR for PVL.
Nasal and groin swab specimens for culture of S. aureus were collected from a study population of 600 HIV-infected outpatients at the Atlanta, GA, Veterans Affairs Medical Center (VAMC). Swabs were collected at the enrollment (n = 600), 6-month (n = 502), and 12-month (n = 427) visits between September 2007 and June 2009. Sterile rayon swabs with liquid Stuart's transport medium (Becton Dickinson, Sparks, MD) were used to swab the anterior nares (collected by clinic personnel) and the groin (self-collected from the skin folds between the thigh and the genital area) and then stored at 4°C for up to 7 days. Swabs were then plated directly to mannitol salt agar (MSA; Becton Dickinson) and CHROMagar MRSA (CM; Becton Dickinson), before being placed in Trypticase soy broth containing 6.5% sodium chloride (broth; Becton Dickinson); all cultures were incubated overnight at 35°C. Following overnight incubation, broths were subcultured to MSA and incubated as described above.
CM plates were examined at 24 and 48 h for mauve-colored colonies, which were subcultured on Trypticase soy agar with 5% sheep blood (blood agar plates [BAPs]; Becton Dickinson). MSA plates were examined at 48 h for gold or yellow colonies, which were subcultured on BAPs. Presumptive S. aureus isolates from BAPs were examined for morphology consistent with S. aureus; identification was confirmed by a positive Staphaurex latex agglutination test (Remel, Lenexa, KS). All S. aureus isolates were frozen at −70°C until further characterization.
Growth of any staphylococci on direct plating to MSA was used as an indicator of specimen integrity and S. aureus viability. Potential variation in bacterial growth from swabs stored at 4°C for up to 7 days was assessed by comparing the distribution of cultures positive for S. aureus among swabs stored for 1 to 3 days with the distribution of those stored for 4 to 7 days prior to culture as described above.
Methicillin resistance was determined using the cefoxitin disk-diffusion test following Clinical and Laboratory Standards Institute (CLSI) guidelines (9). MRSA isolates were genotyped by PFGE using SmaI (New England BioLabs, Beverly, MA) digestion as described previously (35). PFGE patterns were analyzed with BioNumerics software (version 5.10; Applied Maths, Austin, TX) and were assigned to USA pulsed-field types using Dice coefficients and 80% relatedness. USA500, Iberian, and Archaic PFGE types were grouped together as USA500/Iberian because they are closely related and difficult to separate by PFGE. The SCCmec type and the presence of PVL were determined for all isolates using PCR, as described previously (16, 31). For the purposes of this paper, USA300 was defined as an isolate with a USA300 PFGE pattern that was PVL positive and contained SCCmec IVa.
A patient was defined to be colonized for the purpose of prevalence if S. aureus was detected at either body site at each collection or at any collection for the overall colonization prevalence (Table 1). For the determination of overall prevalence, each patient could be counted only once. Sensitivity of culture methods was based on aggregate data from all collection periods. The number of positive culture results per body site for the method evaluated was compared to the number of positive results from that body site detected by any method (Table 2). Prevalence of MRSA strain types was determined by counting all unique PFGE types isolated per sample type from each collection period (Table 3). Sampling methods were compared by the dependent Z test for proportion. Distribution of strain types by body site was evaluated using the chi-square test. The null hypothesis was rejected at P values of ≤0.05. The potential impact of storage of the swabs at 4°C on recovery of S. aureus was examined using a chi-square test comparing swabs stored for 1 to 3 days and those stored for 4 to 7 days prior to culture.
MRSA colonization at any site averaged 13.7% per collection, for an overall MRSA colonization prevalence of 21.3% of subjects over the course of the 1-year study (Table 1). MSSA colonization averaged 31.7% at each collection time, with 46.5% of patients colonized over the course of the study. As expected, S. aureus carriage was higher in the nose than in the groin, although inclusion of culture of specimens from the groin site enhanced detection of both MRSA and MSSA compared to nasal specimen culture alone. The frequency of nasal and groin carriage did not vary between collections for MRSA or MSSA. Nasal carriage prevalence of MRSA at each collection averaged 11.0%, compared to 8.2% groin carriage, and MSSA nasal carriage prevalence averaged 27.2%, compared to 18.3% groin carriage. Overall, MRSA and MSSA prevalence rates were 16.8% and 40.8%, respectively, for the nares, and 15.2% and 30.5%, respectively, for the groin. Inclusion of culture of specimens from the groin site increased the number of colonized patients detected by 26.7% for MRSA and 13.8% for MSSA. Seventeen patients (2.8% of patients enrolled; 6.1% of MRSA-positive patients) were cocolonized with MSSA and MRSA in the nares, groin, or both.
A total of 296 MRSA and 698 MSSA isolates were recovered from among the 1,529 nasal and groin swab specimen cultures performed (Table 2). The broth method detected 272 of 296 (91.9%) MRSA-positive specimens, direct MSA detected 201 (67.9%), and CM detected 216 (73.0%) (Table 2). Broth was more sensitive than CM or MSA for groin swab specimen cultures (P < 0.0001) and for overall MRSA recovery (P < 0.0001). When limited to only nares swab specimen cultures, there was no significant difference in MRSA recovery between the three methods. There was also no significant difference in MRSA recovery between the CM and MSA direct plating methods; this was true for both culture sites and overall MRSA recovery. MSSA recovery was significantly increased when using broth than when using MSA direct plating for both the nares and groin swab specimens and for overall MSSA recovery (P = <0.0001).
Storage of swabs at 4°C for up to 7 days prior to culture did not appear to have a negative impact on the recovery of S. aureus, as there was no change in recovery among those stored for 1 to 3 days and those stored for 4 to 7 days when plated directly to MSA (P = 0.11) or incubated in broth prior to plating on MSA (P = 0.79). Surprisingly, there was a higher odds of recovery for S. aureus among swabs stored for a longer period before plating to CM (P = 0.046).
Three PFGE types predominated, accounting for approximately 95% of all MRSA isolates recovered (Table 3); these included USA500/Iberian (53%), USA300 (34%), and USA100 (6.4%). Although the differences at individual collections were not significantly different, overall, USA300 was more likely than other MRSA strain types to be recovered from the groin than from the nose (P < 0.05). Twelve of 48 (25%) patients colonized with only USA300 in the groin would have been missed with nares swab specimen culture alone. Although the difference in recovery between the nose and the groin was not statistically significant for USA500/Iberian isolates, 10 of 69 (14.5%) patients colonized with only USA500/Iberian type in the groin would not have been detected with nares swab specimen culture alone. Six patients were colonized with more than one MRSA strain; these were detected on the basis of different colony morphologies on MSA, and the results were confirmed by PFGE.
Although the prevalence of MRSA colonization among healthy persons in the U.S. population is low (0.8 to 1.5%) (18, 27), some reports suggest that it is increasing (18, 27). The prevalence of MRSA colonization in our HIV-infected study population was higher at all three study visits (range, 13 to 15%) than the reported rates in the community; the cumulative MRSA prevalence was 21%. The reported prevalence of MRSA colonization in HIV-infected adults varies widely and has been reported to be as low as 2 to 4% in Boston, MA, and Nebraska (34, 51) and as high as 17% in Atlanta and New York City (21, 50). This variation may be due to factors associated with HIV status or local MRSA colonization pressure, but is also likely to reflect the number and source of the swab specimens taken and the method used for culture, as well as the natural dynamics of human carriage of S. aureus (26, 45).
Numerous clinical laboratory studies have compared chromogenic or selective media for the recovery of MRSA, but most used isolates from culture collections or cultures of specimens taken from clinical infections or from populations with high MRSA colonization rates (19, 33, 46, 50). It is difficult to compare the results of these studies to MRSA recovery from surveillance samples, which are often from populations with low MRSA prevalence.
We found that broth enrichment was more sensitive than direct plating to MSA or CM for detection of MRSA and MSSA colonization. Overall, broth enrichment increased sensitivity 19% compared to the use of CM and 24% compared to the use of MSA for recovery of MRSA; the MSSA recovery rate increased 29% with broth enrichment compared to direct plating on MSA. Safdar et al. evaluated 32 different microbiological techniques, including broth enrichment, for detection of MRSA nasal carriage in hospitalized patients and also found an increase of 7 to 14% in sensitivity when broth enrichment was compared to direct plating (48). Other studies comparing broth enrichment to direct plating for MRSA recovery have reported a wide variation in sensitivity, ranging from a decrease of 4% to an increase of 20% (10, 30, 41, 43). The dramatic differences in recovery from broth enrichment that we observed may reflect low numbers of MRSA present in groin swab samples which could be missed by direct plating, as the increase in sensitivity between direct plating and broth enrichment was most noticeable from groin swab samples. Among the 19 MRSA isolates identified with CM but not detected with broth enrichment in our study, 14 (74%) occurred when patients were cocolonized with MRSA and MSSA. We hypothesize that in these instances, small amounts of MRSA could be if missed if abundant MSSA is present in the original sample.
Although MSA requires up to 48 h of incubation and some experience for optimal use, it allows one to detect both MSSA and MRSA, and variations in the color and intensity produced by different S. aureus strains on the same plate can be distinguished. CM is approximately four times the cost of MSA, but the result is easy to interpret and can be read at 24 h. A limitation of our study is the lack of specificity data for the recovery of MRSA from the three methods that we employed. The primary goal of our study was to identify all S. aureus carriage. Because of this, we erred on the side of oversampling by subculturing all suspicious isolates grown on CM and MSA and using a latex agglutination test to rule out coagulase-negative staphylococci. We found that the most sensitive method for recovery of MRSA and MSSA was broth enrichment and then plating to MSA, but broth enrichment and plating to CM or use of MRSA-specific broth enrichment might be the most sensitive for recovery of MRSA only. Only one swab did not yield any Staphylococcus species when plated directly to MSA (data not shown), and we observed no decreased S. aureus yield when swabs were stored under refrigeration for up to 7 days. Thus, weekly batch culturing of swabs might be a practical approach for detecting S. aureus colonization in large-scale or multisite surveillance studies.
The addition of a groin swab specimen culture in our study dramatically enhanced recovery of both MRSA and MSSA. Other studies have reported increased sensitivity when multiple body sites were sampled. Several studies have demonstrated that culture of samples from additional body sites, such as the throat, axilla, perineum, or groin, in addition to the nares, increased the sensitivity of S. aureus detection, with improvements ranging from 5 to 25% (6, 28, 37). However, others have found that the addition of a specimen from a second body site had little impact on the overall sensitivity of MRSA recovery (8). Although other colonization studies have demonstrated a dramatic increase in MRSA detection when throat swabs were cultured in addition to the nasal swabs, we specifically sought to address the impact of carriage site on MRSA strain types and thought that strains carried in the nose and throat were likely to be the same. Inclusion of a groin swab specimen culture in our population not only increased the overall recovery of MRSA but also specifically increased recovery of PVL-positive USA300 strains. Overall, 41.7% of MRSA strains recovered from the groin were USA300, compared to 29.0% from the nose. Similar findings have been reported; Lautenbach et al. found that a larger percentage of CA-MRSA was isolated from the groin and perineum than the nares, throat, and axilla (28), and a recent study by Faden and colleagues found an association between rectal, but not nasal, MRSA carriage and S. aureus skin and soft tissue infection in children (15). At 11.5% prevalence, the USA500/Iberian type was the most frequently isolated strain in our population, representing 48.0% of the MRSA isolates recovered from the groin and 57.4% from the nares. This is higher than the reported prevalence of USA500 among HIV-infected patients in Dallas, TX, and New York City (range, 4.8 to 5.6%) (8, 49) but lower than the reported prevalence (19%) among general patients screened at admission to another Atlanta-area hospital (21), and both USA300 and USA500 colonization has been associated with HIV infection (36).
In summary, the rate of MRSA colonization among our cohort of HIV-infected outpatients was higher than that which has been reported for the healthy U.S. population and for other groups of HIV-infected individuals. Broth enrichment significantly increased detection of MRSA and MSSA compared to direct plating, and we saw no difference in sensitivity of direct plating to CM or MSA. Inclusion of a groin swab specimen also increased MRSA detection and specifically increased detection of PVL-positive USA300 strains. Thus, the impact on both strain diversity and overall sensitivity should be considered when selecting body sites for MRSA sampling.
The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention.
Published ahead of print on 12 October 2011.