High Staphylococcus aureus Colonization Prevalence among Patients with Skin and Soft Tissue Infections and Controls in an Urban Emergency Department

Abstract

Staphylococcus aureus is a commensal species that can also be a formidable pathogen. In the United States, an epidemic of community-acquired methicillin-resistant Staphylococcus aureus (MRSA) infections has been occurring for the last 15 years. In the context of a study in which we identified patients with skin and soft tissue infections (SSTIs) and randomized them to receive one of two antimicrobial treatment regimens, we assessed S. aureus colonization in the nares, throat, and perianal skin on the day of enrollment and 40 days after therapy. We compared the prevalence of colonization between the SSTI patients and an uninfected control population. A total of 144 subjects and 130 controls, predominantly African American, participated in this study, and 116 returned for a 40-day follow-up visit. Of the SSTI patients, 76% were colonized with S. aureus at enrollment, as were 65% of the controls. Patients were more likely than the controls to be colonized with USA300 MRSA (62/144 [43.1%] versus 11/130 [8.5%], respectively; P < 0.001). The nares were not the most common site of colonization. The colonization prevalence diminished somewhat after antibiotic treatment but remained high. The isolates that colonized the controls were more likely than those in the patients to be methicillin-susceptible S. aureus (MSSA) (74/84 [88.1%] versus 56/106 [52.8%], respectively; P < 0.001). In conclusion, the prevalence of S. aureus colonization among SSTI patients was high and often involved USA300 MRSA. The prevalence diminished somewhat with antimicrobial therapy but remained high at the 40-day follow-up visit. Control subjects were also colonized at a high prevalence but most often with a genetic background not associated with a clinical infection in this study. S. aureus is a commensal species and a pathogen. Plans for decolonization or eradication should take this distinction into account.

INTRODUCTION

Staphylococcus aureus is a commensal species that can be a formidable pathogen. It has been traditionally held that the primary site of colonization with the species is the nasopharynx (1) and repeatedly documented that 25% to 40% of the population is colonized in the nasopharynx by S. aureus. Colonization is asymptomatic, but colonized individuals may develop an active infection and do so more frequently than uncolonized individuals (2). The pathophysiological processes that result in infection among colonized individuals are poorly understood.

Recently, the notion that the nasopharynx is the primary site of colonization has been challenged by the finding of a higher prevalence of colonization at other body sites, notably the oropharynx and even the inguinal area (3, 4). Indeed, many individuals with colonization in the throat or on the skin over the inguinal area have no colonization in the nasopharynx (5–8). Small studies in the United States and elsewhere have demonstrated that S. aureus, and specifically the methicillin-resistant S. aureus (MRSA) genetic background USA300 can be found elsewhere on the body (reviewed in reference 4).

In the United States, an epidemic of community-acquired MRSA infections has occurred in the last 15 years (9). This epidemic has been largely propagated by the S. aureus genetic background USA300, named after its banding pattern when subjected to pulsed-field gel electrophoresis. Strains belonging to this background have been responsible for the upsurge of skin and soft tissue infections (SSTIs) that have crowded emergency rooms in the United States (10).

The epidemiology of USA300 MRSA has been studied. The National Health and Nutrition Examination Survey (NHANES) (11) conducted by the Centers for Disease Control and Prevention (CDC) revealed that nasal MRSA colonization in the noninstitutionalized U.S. population increased to 1.5% in 2003–2004 from 0.9% in 2001–2002. This increased but low prevalence in the general population stands in contrast to the higher colonization prevalences found by others in certain populations (12, 13). Estimates of prevalence have varied by anatomic site cultured to assess colonization, the study setting, and the methodology used for culture processing and sampling.

Given that USA300 MRSA is the most important cause of S. aureus SSTIs in the United States, it is curious that the nasal colonization prevalence is relatively low despite the frequent recovery of isolates from this staphylococcal genetic background from infections. The relationship between colonization and infection when S. aureus of other genetic backgrounds colonizes the nasopharynx or elsewhere is also unclear.

The relationship between colonization and infection is important. Among patients with S. aureus bacteremia, an organism of identical genetic background may be found in the nasopharynx prior to onset of infection. This implies that colonization precedes clinical infection, at least sometimes (14).

In this study, we sought to define the prevalence and genotypic spectrum of S. aureus colonization at three body sites among patients who sought care in the emergency department (ED) at the University of Chicago Medical Center (UCMC) for uncomplicated SSTIs. These patients were enrolled in a clinical trial to compare the efficacy of two antibiotics in treating these infections (21). At one site (Chicago), we compared the findings among the patients to a control population who sought care in the ED in the same period for noninfectious diagnoses. We evaluated the effect of antibiotics on S. aureus carriage by reevaluating colonization among the SSTI population after antibiotic treatment was completed.

MATERIALS AND METHODS

UCMC is an academic medical center located on the south side of Chicago, Illinois, and serves the surrounding inner-city population and patients referred for tertiary care. Emergency care is provided to pediatric and adult patients. Nearly 90,000 patients seek care in the ED annually.

From October 2009 to July 2011, children and adults, aged 6 months to 84 years, seeking care in the ED for an uncomplicated SSTI were screened Monday through Friday from 9 a.m. to 9 p.m. for possible enrollment in the study. Patients with an SSTI who were otherwise healthy and whose abscess was >5 cm (>3 cm for children 6 to 11 months of age, >4 cm for children 12 months to 8 years of age) or who had multiple abscesses, purulent cellulitis, nonpurulent cellulitis, or a combination of abscess and cellulitis were subjected to incision and drainage if an abscess was present. Purulent material was cultured (21).

At enrollment, informed consent was obtained to define asymptomatic colonization. Each consenting enrollee provided a nasal, oropharyngeal, and perianal or inguinal culture using a dry rayon-tipped applicator (CultureSwab, BD Diagnostic Systems).

We initially offered patients a choice of a perianal culture or a culture of the skin over the inguinal ligament. Data reflecting the cultures of these two sites were pooled as there was no significant difference in S. aureus colonization prevalence between the two sites (data not shown).

A total of 144 patients with SSTIs were enrolled. Risk factors for health care-associated MRSA infections were present among 24 (16.6%) patients, including 19 (13.2%) who were hospitalized in the previous year and 5 (3.5%) others who had undergone a surgical procedure during that time. Patients with an SSTI were prescribed a course of clindamycin or trimethoprim-sulfamethoxazole (TMP-SMX) therapy in a blinded fashion for 10 days. The patients were asked to return for a follow-up visit 40 days after the initial presentation to ensure that the original lesion was healed and to assess any new lesions (21). At this follow-up visit, the three colonization cultures were repeated.

The 130 control patients sought care in the ED during the same time frame for a noninfectious complaint. Antibiotics were not prescribed. Twenty-two (16.9%) control patients also had risk factors for health care-associated MRSA infections; 17 (13.1%) had been hospitalized in the previous 12 months, 3 (2.3%) had surgery during this period, and 2 (1.5%) had a nursing home stay. The control group included patients with an age, gender, and self-reported ethnicity similar to those of the patients with SSTIs (Table 1), and they were cultured identically.

TABLE 1.

Demographics

Demographic characteristic	No. (%) of patients with SSTIs (n = 144)	No. (%) of controls (n = 130)
Age
6–23 mo	0 (0)	0 (0)
2–4 yr	13 (9)	1 (1)
5–12 yr	41 (28)	10 (8)
13–17 yr	13 (9)	14 (11)
18–25 yr	22 (15)	19 (15)
26–40 yr	37 (26)	49 (38)
41–60 yr	17 (12)	28 (22)
61–84 yr	1 (1)	9 (7)
Gender
Male	53 (37)	65 (50)
Female	91 (63)	65 (50)
Race/ethnicity
African American	140 (97)	120 (92)
White	4 (3)	10 (8)

Nearly all of the patients and controls were African American, reflecting the population served by our ED (Table 1). Few patients and therefore few controls were >40 years of age.

Culture swabs were transported to the University of Chicago MRSA Research Laboratory and incubated in Trypticase soy broth with 7% sodium chloride overnight at 37°C. The next day, an aliquot of the broth was plated onto BBL CHROMagar S. aureus medium (BD Diagnostic Systems) and incubated for 24 h at 37°C. One colony was picked from the plate, and an isolate was confirmed to be S. aureus by typical colony morphology, positive catalase, and a Staphaureux test (Remel). S. aureus isolates were stored at −70°C until further evaluation.

Genomic DNA was extracted from each isolate using the Qiagen DNeasy blood and tissue kit following the manufacturer's instructions and modified by incubation with lysostaphin in a resuspension buffer (at 37°C for 30 min) to facilitate S. aureus lysis (15). S. aureus speciation was confirmed using a PCR assay specific for spa (encoding protein A). S. aureus isolates were characterized by multilocus sequence typing (MLST) (16) to determine the genetic background and by typing of the SCCmec element, the mobile genetic element that carries mecA (17), with type assignments using published guidelines (18). Detection of the genes encoding the Panton-Valentine leukocidin (PVL) was performed as described (19). Two S. aureus isolates were considered indistinguishable if they shared an MLST and SCCmec type and were concordant with respect to the presence or absence of the PVL genetic determinants. Based on a previous investigation demonstrating that the presence of both the PVL genes and the arcA gene, found on the arginine catabolic mobile element (ACME), is highly concordant with the USA300 MRSA genetic background assessed by pulsed-field gel electrophoresis (20), isolates with these characteristics were considered USA300 MRSA. USA300 isolates that lacked mecA and an SCCmec element were termed USA300 methicillin-susceptible S. aureus (MSSA).

RESULTS

Of the 144 patients with SSTIs, 102 (71%) had a single lesion. Of patients with a single lesion, 47 (46%) had an abscess, 19 (19%) had purulent cellulitis, and 36 (35%) had nonpurulent cellulitis.

The remaining 42 patients (29%) had multiple lesions; 21 (50%) had multiple abscesses, 11 (26%) had multiple cellulitis lesions (in 6, ≥1 lesion was purulent), and 10 (24%) had at least one abscess and one cellulitis lesion.

Eight of the 36 patients with nonpurulent cellulitis had a single lesion present at the enrollment visit. Two days later, an abscess was noted at the same site; incision and drainage were performed, and the abscess material was cultured. These patients were considered to have had an abscess at enrollment.

Thus, a single abscess was considered to have been present at the enrollment visit in 55 (54%) patients, purulent cellulitis in 19 (19%), and nonpurulent cellulitis in 28 (27%).

Culture and molecular characteristics of the infecting isolates.

Of the 55 patients with a single abscess, culture of the abscess material in 41 (75%) yielded S. aureus, and 35 of the 41 isolates (85%) were USA300 MRSA.

Of the 47 patients with a single cellulitis lesion that did not develop into an abscess, 19 (40%) were purulent and cultured. Fifteen (79%) yielded S. aureus, and 11 of the 15 (73%) were USA300 MRSA. Of the remaining 4, 1 was PVL positive (PVL⁺), arcA negative (arcA⁻), ST8, and SCCmec type IV, and 3 were MSSA. Two of these were closely related (PVL⁺, arcA⁻, and ST8). The other was PVL⁻, arcA⁻, and ST20.

Of the 21 patients with multiple abscesses at the enrollment visit, the lesions of 18 (86%) yielded S. aureus. Fifteen (83%) grew MRSA from a culture of each infective site; 14 were USA300 MRSA and 1 was closely related (PVL⁺, arcA⁻, SCCmec type IV, ST8). Two patients with multiple abscesses had MSSA isolated from each; both were ST8 and arcA⁻, and one was PVL⁺. The other patient had a USA300 MRSA isolate and an ST30 MSSA isolate from the various lesions. The former was PVL⁺ and arcA⁺, and the latter was PVL⁻ and arcA⁻.

Of the 11 patients with multiple cellulitis lesions at the enrollment visit, 6 (55%) had at least one purulent cellulitis lesion, and all 6 grew at least one S. aureus isolate. Four (67%) had USA300 MRSA. One patient had two S. aureus genotypes when the isolates from different lesions were tested: one was MSSA, PVL⁻, arcA⁻, and ST8, and the other was MSSA, PVL⁻, arcA⁻, and ST88. One patient had a MRSA isolate with SCCmec IV, PVL⁺, arcA⁻, and ST22 isolated from purulent cellulitis.

Of the 10 patients who had ≥1 abscess and ≥1 area of cellulitis at enrollment, 9 had S. aureus isolated from ≥1 lesion. Eight grew MRSA (all USA300), and 1 grew a closely related MSSA strain (PVL⁺, arcA⁻, ST8).

The 144 patients were stratified into those who had lesions from which S. aureus was isolated (n = 89, 62%) and those who had lesions from which S. aureus was not isolated (n = 55, 38%). Of the 89 patients who had S. aureus isolates, 78 (88%) had MRSA isolates while 10 (11%) had MSSA isolates; 1 patient had MRSA and MSSA.

USA300 MRSA was isolated from the infective lesion in 72 of 144 patients (50%). The lesions of 4 additional patients had MRSA with SCCmec IV, PVL⁺, and ST8 isolated but not termed USA300 because they lacked arcA.

Colonization of SSTI patients at the enrollment visit.

Of the 144 SSTI enrollees, 8 had ≥1 colonization isolate unavailable for typing or were missing clinical information. Thus, the denominators below differ somewhat to reflect the exclusion of some or all of these isolates.

As mentioned, 89 (62%) of 144 enrollees had an S. aureus isolate from their infective lesion. A total of 76% (108/143) of SSTI enrollees were colonized with S. aureus at ≥1 site at the enrollment visit (Table 2). The perianal area (81/143, 57%) was the most common site of colonization, followed by the throat (74/143, 52%). The nose was the least commonly colonized site of the three (56/143, 39%) (Table 2), and the observed colonization prevalence would have been 39% if we had cultured this site alone.

TABLE 2.

Prevalence of S. aureus colonization in patients and controls

Site and isolate type	No. (%) of patients at enrollment (n = 144)	No. (%) of patients at 40-day follow-up (n = 116)	No. (%) of controls (n = 130)
At ≥1 site	108/143a (76)	63/116 (54)	84/130 (65)
MRSA	68/106g,i (64)	26/59h,j (44)	19/84k (23)
MSSA	56/106g,i (53)	40/59h,j (68)	74/84k (88)
Nose	56/143a (39)	27/116 (23)	42/130 (32)
MRSA	34/55b (62)	11/27 (41)	6/40e (15)
MSSA	21/55b (38)	16/27 (59)	34/40e (85)
Throat	74/143a (52)	26/116 (22)	68/130 (52)
MRSA	30/71c (42)	6/26 (23)	8/64f (13)
MSSA	41/71c (58)	20/26 (77)	56/64f (87)
Perianal/groin	81/143a (57)	29/116 (25)	46/130 (35)
MRSA	52/76d (68)	16/29 (55)	11/43c (26)
MSSA	24/76d (32)	13/29 (45)	32/43c (74)

^aOne case (and corresponding nose, throat, and perianal swab) not available for analysis.

^bOne isolate not available for typing.

^cThree isolates not available for typing.

^dFive isolates not available for typing.

^eTwo isolates not available for typing.

^fFour isolates not available for typing.

^gTwo cases with isolates that were unable to be typed.

^hFour cases with isolates that were unable to be typed.

ⁱEighteen patients had at least one MRSA and one MSSA isolate.

^jSeven cases had at least one MRSA and one MSSA isolate.

^kNine cases had at least one MRSA and one MSSA isolate.

Of the 89 enrollees with culture-proven S. aureus infection, 52 (58.4%) were colonized with an isolate that had an identical genetic background to the S. aureus isolate obtained from the infective lesion. These were uniformly USA300 MRSA or a closely related strain (48 were USA300 MRSA isolates; 2 were ST8 arcA⁻ PVL⁺ MRSA isolates; and 2 were ST8 arcA⁻ PVL⁻ MSSA isolates).

Sixty-eight (64%) of 106 of the SSTI patients were colonized with MRSA isolates, and 56 (53%) were colonized with MSSA isolates. Eighteen of the 106 (17%) were colonized with ≥1 MSSA isolate and ≥1 MRSA isolate (Table 2).

Among the 202 colonizing isolates from the 106 S. aureus-colonized patients, 116 (57%) were MRSA and were isolated from 68 colonized patients. Nearly all were ST8 (114/116, 98%). Eighty-six (43%) were MSSA and were isolated from 56 colonized patients. About a quarter (20/86, 23%) of these MSSA isolates were ST8. Twenty-six genetic backgrounds were represented in the MSSA strain pool from the colonized patients with an SSTI, with ST5, ST15, and ST97 being the most prevalent.

Of the 106 patients with an SSTI and colonized with S. aureus, 26 (25%) were colonized by ≥1 S. aureus strain type. We underestimated the percentage of individuals with polyclonal colonization; our method of polyclonal colonization detection was insensitive because only one colony from each cultured site was selected. Polyclonal colonization by S. aureus is common, especially when multiple sites are assessed.

Thus, at enrollment, S. aureus colonization is frequent and often at multiple sites. Colonization by USA300 MRSA, the most common infecting genotype, was frequent, but many other genotypes were also often found. Many (39/56, 70%) of the SSTI patients at enrollment were colonized with an MSSA strain that did not cause an SSTI in our population.

Colonization of the SSTI patients at the 40-day follow-up visit.

At the 40-day follow-up visit, 116 (81%) of the 144 patients returned. The S. aureus colonization prevalence among returning SSTI patients decreased (P = 0.0003), although 54% were still colonized despite receiving a 10-day course of clindamycin or TMP-SMX (Table 2). The rate decreased significantly at all sites sampled. The perianal region was again the most frequently colonized site (29/116, 25%), followed by the nose (27/116, 23%) and the throat (26/116, 22%) (Table 2). Had we cultured only the nose, many colonized individuals would not have been detected; the documented colonization prevalence would have been much lower than that documented by culturing three sites (Table 2).

Multiple-site colonization was common at enrollment, when most (75/108, 69.4%) colonized patients had S. aureus isolated from >1 site. However, after 40 days and a course of clindamycin or TMP-SMX, the prevalence of multiple-site colonization decreased to 20 (31.7%) of 63, and many colonized individuals had S. aureus isolated from only 1 site (Fig. 1).

FIG 1.

Venn diagram showing colonization with S. aureus of the nose, throat, and perianal skin among enrolled patients with SSTIs and controls who sought care in the ED for presumably noninfectious complaints. The numbers in the diagrams reflect the number of patients and/or controls in each category. (A) Colonization at the enrollment visit. Of 143 SSTI patients, 108 (75.6%) were colonized with S. aureus; 28 patients (19.6%) were colonized at all 3 sites. Of the 81 patients (56.6%) colonized in the perianal area, 16 (11.2%) were colonized at that site alone. Of the 74 subjects (51.7%) colonized in the throat, 13 (9%) were colonized at that site alone. Of the 56 subjects (39.2%) colonized in the nares, only 4 (2.8%) were colonized at that site alone. (B) Colonization at the 40-day visit. A total of 116 SSTI patients (81.1%) returned at the 40-day visit, of whom 63 (54.3%) were colonized with S aureus. Only 4 (3.4%) were colonized at all 3 sites. Of the 32 patients (27.6%) colonized in the perianal area, 17 (14.7%) were colonized at that site alone. Of the 27 patients (23.3%) colonized in the throat, 13 (11.2%) were colonized at that site alone. Of the 28 subjects (24.1%) colonized in the nares, 13 (11.2%) were colonized at that site alone. (C) Colonization of control subjects. Of the 130 control subjects, 84 (65%) were colonized; 16 (12.3%) were colonized at all 3 sites. Of the 43 control patients (33.1%) who were colonized in the perianal area, 9 (6.9%) were colonized at that site alone. Of the 64 control subjects (49.2%) colonized in the throat, 23 (17.7%) were colonized at that site alone. Of the 40 subjects (30.8%) colonized in the nares, 5 (3.8%) were colonized at that site alone.

At the 40-day visit, the 63 patients who were colonized with S. aureus had 82 isolates recovered, a significant (P < 0.001) decline from the prevalence at enrollment. The prevalence of both MRSA (33 [40%] of the 82 isolates from 26 patients) and MSSA (49 [60%] of the 82 isolates from 40 patients) significantly declined. Ten (20.4%) of the 49 MSSA isolates were ST8. The other MSSA strains came from 19 genetic backgrounds.

USA300 MRSA colonization also declined but in proportion to the decrease observed with other staphylococcal backgrounds. Of those returning at 40 days, 18 (37%) of 49 were colonized with USA300 MRSA, and 11 (28%) of 40 were colonized with isolates from another staphylococcal background that remained identical to the colonizing isolate at enrollment (P = 0.355).

Among the 72 patients who had a USA300 MRSA S. aureus-infecting isolate, 60 (83.3%) were still colonized with S. aureus at 40 days, mostly (48/60, 80%) with USA300 MRSA.

Acquisition of S. aureus carriage.

At the enrollment visit, 46 patients were colonized with S. aureus of a genetic background other than USA300 MRSA. Forty (87.0%) of these patients returned for the 40-day visit. Of these, 9 patients had acquired S. aureus colonization (4 USA300 MRSA, 5 another genetic background) of a background different from that documented at enrollment. Eleven patients were colonized with S. aureus of an identical background to that at enrollment, and 20 patients were no longer colonized. Thus, acquisition and loss of carriage were common.

Between enrollment and the 40-day follow-up visits, acquisition of S. aureus carriage occurred at a similar prevalence in patients with culture-positive SSTIs and culture-negative SSTIs. Of the 72 patients with a USA300 MRSA infection isolate at enrollment, 4 (20%) of 20 uncolonized patients were colonized at the 40-day visit; the 3 patients whose isolates were available for typing had ≥1 MSSA-colonizing isolate. Of those lacking an S. aureus infection isolate at enrollment (55/144, 38%), 4 (25%) of 16 uncolonized patients were colonized at the 40-day visit with MSSA isolates of different genetic backgrounds.

Thirty-nine patients (81.3%) who had a USA300 infection isolate and were colonized at enrollment with a USA300 MRSA isolate returned for the 40-day visit; 26 (66.7%) were still colonized with S. aureus, and 14 (53.8%; P = 0.013 compared with enrollment) were still colonized with a USA300 MRSA isolate.

Colonization among control subjects.

Of 130 control subjects, 84 (64.6%) were colonized with S. aureus, a prevalence lower than that found among the SSTI patients at enrollment (P = 0.049) (Table 2). Thirty-seven colonized control subjects (44%) had S. aureus isolated at only 1 site, 31 (37%) at 2 sites, and 16 (19%) at all 3 sites sampled (Fig. 1). The throat (68/130, 52%) was the most frequently colonized site, followed by the perianal region (46/130, 35%) and the nose (42/130, 32%). Importantly, had we only sampled the nose for colonization, the 32% prevalence would have greatly underestimated the overall prevalence of colonization.

Control subjects were colonized with a USA300 MRSA isolate less frequently than subjects with an SSTI (P < 0.001) (Table 2). Most often, control subjects were colonized with an MSSA isolate with a genetic background that was not detected among infection isolates. Thirty-five genetic backgrounds were represented among the colonizing S. aureus isolates obtained from the controls.

Selected isolate characteristics.

MRSA isolates colonizing patients at the enrollment and 40-day visits almost always carried the genes for the PVL toxin and arcA (Table 3). MSSA strains, surprisingly, often had these genetic features as well. Of interest is the persistence of MRSA and MSSA isolates that contained these toxin genes and arcA, our arginine catabolic mobile element (ACME) proxy, although at a lower percentage, at the 40-day visit. Isolates obtained from control subjects had these traits but less frequently.

TABLE 3.

Selected characteristics of S. aureus patients and control colonizing isolates

Characteristic	No. (%) of isolates at enrollment		No. (%) of isolates at 40-day follow-up		No. (%) of isolates in control group
	MRSA	MSSA	MRSA	MSSA	MRSA	MSSA
PVL positive	66 (97.1)	27 (48.2)	26 (100)	12 (30.0)	14 (73.7)	12 (16.2)
arcA positive	62 (91.2)	21 (37.5)	23 (88.5)	8 (20.0)	13 (68.4)	11 (14.9)

Genetic backgrounds that caused infection.

Eight genetic backgrounds were isolated from an SSTI: 4 backgrounds were MRSA (USA300, n = 72 patients; other ST8, n = 4; ST1, n = 1; ST22, n = 1), and 4 backgrounds were MSSA (USA300, n = 1 patient; other ST8, n = 8; ST20, n = 1; ST30 n = 1).

At the enrollment visit, 77 (73%) of 106 colonized SSTI patients were colonized with an S. aureus strain whose genetic background was one of the 8 found in an infection, a prevalence that significantly exceeded that found in controls. Twenty-nine (27%) were colonized with a strain that was not recovered from any patient with an SSTI. Of those that returned at 40 days and were colonized, 35 (59%) of 59 were colonized with an S. aureus strain whose genetic background was found in an infection, a prevalence insignificantly lower than that found at enrollment (P = 0.08). Forty-one percent were colonized with an isolate that was not recovered from an SSTI in this study. Thus, antibiotic therapy of the infectious lesion did not eradicate colonization by strains that caused an SSTI.

For the controls, 35 (42%) of 84 colonized individuals were colonized with an S. aureus strain whose genetic background was found in an infection, but 49 (58%) were colonized with an S. aureus isolate that was not recovered from an SSTI. Thus, the frequent colonization of controls occurred less often with an organism that caused an SSTI.

Summary of results.

Among the patients with an SSTI, the nasal S. aureus colonization prevalence was 39% (Table 2). However, 76% of patients with an SSTI were colonized at ≥1 site, and many were colonized at all 3 cultured sites. After treatment, colonization at extranasal sites among nasal carriers decreased. Also, colonization prevalence in the throat and the perianal region decreased somewhat, as did multisite colonization (Fig. 1). The site of colonization that was most resistant to decolonization after antibiotic treatment was the throat. After antibiotic treatment, 20 (28%) of 71 patients remained colonized in the throat, 19 (25%) of 76 remained colonized in the perianal region, and 13 (24%) of 55 remained colonized in the nose. These eradication rates by site did not differ significantly.

The nasal colonization prevalence was lower among the control subjects than among the subjects with an SSTI but was still 32% (Fig. 1; Table 2). The throat was most heavily colonized, and the nose was colonized least often. The S. aureus colonization prevalence among control subjects, while high, was lower in the throat and perianal region than that in the cases at enrollment. Most control subjects were colonized at ≥1 other site, and many were colonized at all 3 cultured sites (Fig. 1).

Comparing the strain type of colonization isolates among SSTI patients and controls, profound differences were noted in that the controls were colonized much less frequently with MRSA isolates (Table 2). Control isolates also had arcA or PVL genes much less frequently (data not shown).

DISCUSSION

The prevalence of S. aureus colonization among patients with SSTIs and controls attending the ED at the University of Chicago Hospital when three sites were cultured was much higher than we had anticipated. A colonization prevalence of 25% to 40% is usually assumed, although often only the nose is sampled and a broth enrichment step like the one we employed is not used. The perianal skin was the tested site with the highest prevalence of S. aureus colonization among patients with an SSTI. The high prevalences of colonization at this site and in the throat suggest that colonization studies that sample only the nares or only 1-2 sites may obtain data yielding an incomplete picture.

Among the controls, the throat was the most common site of colonization, and the perianal region was the second most common. The nares were colonized in 32% of patients, a prevalence resembling that found by others. Thus, culture surveys of the nares alone greatly underestimate the prevalence of S. aureus colonization. We would not have detected S. aureus colonization in more than half of this control population had we sampled only the nasopharynx.

Our data thus do not support the notion that the nose is the primary site of S. aureus colonization. Rather, S. aureus was found to colonize frequently in each of the sites we cultured among the patients with SSTIs and among the controls. This observation was independent of background and not a function of USA300 colonization at extranasal sites.

Furthermore, the prevalence of S. aureus colonization among control patients attending the ED for noninfectious complaints was also high but lower than that found in the patients with SSTIs, and the prevalence of USA300 MRSA colonization was much lower. These data support the notion that S. aureus has a unique relationship with its human host. Most colonizing S. aureus isolates are commensals, members of the flora, and besides USA300 MRSA (a predominant cause of an SSTI) and a few other genetic backgrounds, most staphylococcal backgrounds seldom cause disease but are identified in asymptomatic colonization studies. This is true despite the relatively high prevalence of USA300 MRSA colonization we found among these control patients.

Following antibiotic treatment of the patients with SSTIs, the prevalence of S. aureus colonization decreased, as did that of USA300 MRSA, but both remained high. The notion that treatment of an infection, such as the presenting SSTI, with antibiotics eliminates asymptomatic carriage of S. aureus, including USA300 MRSA, is incorrect. If colonization is a prerequisite and a risk factor for infection, these data may contribute to our understanding of the high recurrence rate among patients with an SSTI.

We did not design our study to detect the extent of polyclonal colonization. The traditional assumption is that people are colonized with a single strain of S. aureus. In our study, only 1 isolate was selected for typing and storage from each cultured site of each colonized patient or control. Thus, the identification of polyclonal colonization required ≥2 S. aureus isolates of different genetic background to be present at ≥2 of the 3 sites that were sampled. Despite this, we found that 25% of the people were polyclonally colonized. The prevalence of polyclonal colonization is therefore probably much higher; its definition will require a different study design.

The prevalence of colonization in patients and controls was high. Colonization loss and the acquisition of strains with new genetic backgrounds were frequent, as was polyclonal colonization. These data paint a complex picture of the interaction between S. aureus and a human host. Certainly, infected people are often colonized with staphylococcal infections that can cause an SSTI, but we also isolated many colonizing strains that did not cause disease. After antibiotic therapy at the 40-day follow-up visit, there was less colonization of all S. aureus strains, including USA300 MRSA. However, the prevalences of all kinds of S. aureus colonization at the follow-up visit, although lower, were still high. We did not do additional follow-up to see if the prevalence rates in patients with an SSTI eventually reached the levels found in the general population.

In summary, our findings challenge several common wisdoms about S. aureus colonization. The nares are not the most common site of carriage among patients with an SSTI or among the control patients who attended the ED. Many carriers are colonized with ≥1 strain type. Antibiotic treatment does not usually clear S. aureus carriage, including the carriage of S. aureus from more pathogenic genetic backgrounds. Most S. aureus isolates are commensals and colonize asymptomatically and are not an apparent threat to human health. A great deal more needs to be learned about the relationship between colonizing strains and infection.