Cutaneous Microenvironment of Human Immunodeficiency Virus (HIV)-Seropositive and HIV-Seronegative Individuals, with Special Reference to Staphylococcus aureus Colonization

Abstract

A cross-sectional quantitative study of cutaneous bacterial and yeast flora at seven body sites in 99 human immunodeficiency virus-seropositive and 50 seronegative military personnel was performed. Statistically significant differences in carriage rates were only observed for Staphylococcus aureus on the foreheads of seropositive individuals. Seronegative individuals demonstrated staphylococcal carriage rates 1.3 to 2 times as great as those of historical controls (defined as healthy individuals not receiving any medications) at five of six body sites. We conclude that seropositive military personnel do not exhibit statistically significant elevations in densities and carriage rates of the microorganisms examined (except Staphylococcus aureus), relative to seronegative individuals. Seropositive individuals may be predisposed to staphylococcal carriage. The elevated staphylococcal carriage rates of military personnel undergoing basic training warrants a formal evaluation of the impact of training exercises on cutaneous flora. The information gained may serve to limit the spread of infection during training exercises and battlefield conditions.

Perturbations in skin flora due to systemic disease are frequently seen. Antibiotic-resistant group JK coryneform bacteria are more commonly found on the skin of patients with leukemia than on that of healthy individuals (20). Injection drug users, diabetics, individuals with dermatologic disorders, and dialysis-dependent chronic renal failure patients all have elevated nasal Staphylococcus aureus carriage rates relative to healthy controls (7, 16, 27, 37).

The present study was performed to investigate the effect of human immunodeficiency virus (HIV) infection on skin flora. The relationship between HIV infection and cutaneous S. aureus carriage was of particular interest due to the high incidence and morbidity associated with S. aureus pneumonia, sepsis, and soft tissue abscesses in patients with AIDS (18). Furthermore, the possibility that high S. aureus carriage rates could accelerate HIV disease progression warranted an examination of the cutaneous density of this pathogen in early viral infection.

Numerous earlier studies have sought to evaluate S. aureus skin colonization. Investigators have focused on prevalence rates. Surprisingly, little quantitative data on the cutaneous carriage of this or other potential pathogens in HIV-seropositive (henceforth referred to as seropositive) and HIV-seronegative (henceforth referred to as seronegative) individuals is available (3, 5, 8, 11, 14, 34, 40; S. Kravcik, B. Toye, L. Gilbert, N. Hawley-Foss, and D. W. Cameron, 3rd Conf. Retroviruses Opportun. Infect., abstr. 188, p. 87, 1996). This study provides such data.

MATERIALS AND METHODS

Quantitative bacteriological cultures were obtained from seven body sites (i.e., the scalp, forehead, cheek, nose, axilla, toe webs, and perineum). Fifty healthy seronegative military personnel from Fort Dix Training Center and 99 seropositive patients from Walter Reed (WR) Army Medical Center and National Naval Medical Center were cultured. Both groups had no contact with hospitalized patients or other risk factors for S. aureus carriage (such as injection drug use, diabetes mellitus, renal failure, and dermatologic disease). The length of military service prior to enrollment in the study was not a recorded parameter; as such, it is uncertain how recently the personnel had received standard immunizations (which may be a risk factor for S. aureus colonization, secondary to skin penetration by needle).

Seropositive patients were classified according to the WR Staging Classification System (33). Individuals fulfilling criteria for WR stages I and II had more than 400 TH cells/mm³ in their peripheral blood; those categorized to a more advanced WR stage had fewer TH cells in serum. WR stage I encompassed asymptomatic individuals, stage encompassed II those with lymphadenopathy, stage III encompassed those with lymphadenopathy whose TH cell count dipped below 400 cells/mm³, stage IV encompassed those partially anergic, stage V encompassed those completely anergic, and stage VI encompassed those with opportunistic infections and oral candidiasis. For analytical purposes, we consolidated persons in WR stages I to V; these individuals would represent one range of immunosuppression, while WR stage VI would identify those suffering from an even greater degree of immune system impairment.

All of the seropositive patients were outpatients. They were seen regularly at the HIV clinic and had free access to all retroviral medications and therapies for prophylaxis against opportunistic infections. Patients with TH cell counts below 200 cells/mm³ received prophylactic therapy against Pneumocystis carinii pneumonia with aerosolized pentamidine or trimethoprim-sulfamethoxazole. The majority of patients on the latter medications were classified in WR stage VI.

Cultures were obtained using sterile swabs moistened with 1 ml of nonionic detergent (0.1% buffered Tween 80) (Difco Labs, Detroit, Mich.) for the anterior nares, toe interspace, and 5-cn² areas of other body sites, by a modification of the detergent scrub method of Williamson et al. (46). Tween 80 (0.1% buffered is not inhibitory for bacteria; moistening swabs in this detergent aids in the optimal recovery of flora from (oftentimes) dry cutaneous surfaces. The duration of scrubbing was 1 min. The anterior nares were sampled as follows. A sterile cotton-tipped swab was moistened in a culture tube containing 2 ml of 0.1% buffered Tween 80. The swab was wrung out within the tube, swirled inside the anterior nares for five clockwise and five counterclockwise rotations, reintroduced into the culture tube, and again wrung out. The axilla sample was obtained in the most superior aspect of the midaxillary line.

Swabs were deposited in tubes containing 2 ml of detergent fluid (0.1% buffered Tween 80) serially diluted in 10-fold steps. Forty μl per dilution was then drop plated onto the following media, obtained from BBL (Cockeysville, Md.): (i) mannitol salt agar for selective isolation of S. aureus; (ii) Trypticase soy agar with 5% sheep blood and 0.5% yeast extract (B agar), a general nutrient medium for the isolation of lipophilic diphtheroids growing as small colonies; (iii) B agar with 0.5% Tween 80 (B80 agar), a standard nutrient medium for lipophilic and large-colony diphtheroids; (iv) B80 agar with 100 μg of phosphomycin per ml for selective isolation of diphtheroids; (v) B80 with 100 μg of fosfomycin and 100 μg of ticarcillin for selective isolation of Corynebacterium jeikeium (group JK) (42); (vi) phenylethyl alcohol agar with 5% sheep blood for selective isolation of gram-positive rods; (vii) MacConkey agar for selective isolation of gram-negative rods; (viii) Mycosel agar for selective isolation of dermatophytes; (ix) Sabouraud dextrose agar with 0.005% chloramphenicol for selective isolation of fungi; and (x) Schaedler agar (19, 20). All antibiotics were obtained from Sigma (St. Louis, Mo.).

Plates were incubated at 37°C for 72 h under aerobic conditions and then stored at room temperature for 4 days to enhance pigment production and colony morphology. Schaedler agar plates were treated in the same manner but under anaerobic conditions. Representative colonies were Gram stained and identified by using the following API systems (Analytab Products, Plainview, N.Y.) (19, 20): API Staph-Ident system, API 20E enteric system, and API 20L clinical yeast system. Corynebacteria were identified as described by McGinley et al. (23), and other organisms were identified by standard techniques (26, 35). In this manner, qualitative and quantitative determinations were performed for the following organisms: Staphylococcus spp. (including S. aureus), Kytococcus sedentarius, Rhodococcus equi (including C. jeikeium (group JK), Corynebacterium lipophylicus, Corynebacterium xerosus, and Corynebacterium minutissimum), Brevibacterium epidermitis, Propionobacterium spp. (including Propionobacterium acnes and Propionobacterium granulosum), Pseudomonas aeruginosa, Proteus mirabilis and other gram-negatives bacteria, and Candida albicans and other yeasts, dermatophytes, and fungi.

Repeated-measures analysis of variance was used to compare microorganism densities, expressed as log (base 10) CFUs, in seropositive and seronegative individuals. S. aureus carriage densities were dichotomized prior to analysis around a log of 5 so as to control for the frequency of transient carriage of this organism in those over 20 years of age (estimated to be between 60 and 90%) (4, 16, 37, 43). The data were further assessed using row × column χ² analysis. No differences were apparent between the repeated-measures analysis of variance and the χ² analysis. A P value of less than 0.05 was considered significant.

RESULTS

Descriptive data on subjects are shown in Table 1. Carriage rates and densities for S. aureus and other microorganisms by site and carriage rates by site in historical controls are summarized in Table 2. The data for historical controls are based upon studies (published as recently as October 1999) that excluded all individuals diagnosed with a disease and those taking medications. Most people included in these investigations were university students and manual or office employees not exposed to the hospital setting (30, 44). The average age and age distribution of persons included in these studies closely approximate those of the individuals included in the present investigation. The culture techniques employed also closely resemble those utilized here. As such, the referenced studies provide a valuable perspective for assessing the data gathered in the present investigation.

TABLE 1.

Patient descriptions

Characteristic	No. (%) with HIV status		WR stage
	Negative	Positive	I	II	III	IV	V	VI
No. of patients	50	99	11	20	18	14	19	17
Median age (yr) (range)	27 (17–45)	30 (22–54)	29 (23–48)	29 (23–49)	32.5 (26–50)	36 (22–48)	30 (24–54)	33 (26–52)
Ethnicity
Caucasian	24 (48)	50 (50.5)	6	9	8	10	8	9
African-American	18 (36)	45 (45.5)	5	11	8	3	10	8
Hispanic	7 (14)	3 (3)	0	0	1	1	1	0
Asian-American	1 (2)	0 (0)	0	0	0	0	0	0
Filipino	0 (0)	1 (1)	0	0	1	0	0	0
Gender
Female	0 (0)	6 (6)	1	0	0	1	3	1
Male	50 (100)	93 (94)	10	20	18	13	16	16

TABLE 2.

Microoganism carriage rates and densities in HIV-negative and positive military personnel

Organism(s)	Patientsa	Scalp		Forehead		Cheek
		% Colonized	Mean densityb (SD)	% Colonized	Mean density (SD)	% Colonized	Mean density (SD)
S. aureus	HIV−	36	2.7 (0.8)	26	2.3 (0.5)	28	3.3 (1.0)
	WR I–V	45	3.3 (1.5)	30	2.8 (1.1)	44	3.2 (1.1)
	WR VI	41	3.6 (1.3)	52	3.7 (2.4)	47	3.2 (1.0)
	Historical controls (reference no.)	7 (7), 20 (45)f	NAe	6 (25)	NA	8 (25)	NA
C. jeikeium (group JK)	HIV	2	3.0d	2	2.7d	2	2.5d
	WR I–V	0		0		0	0
	WR VI	6	4.9d	6	3.9d	6	3.8d
P. aeruginosa	HIV−	0		0		0	0
	WR I–V	4	4.9 (2.1)	2	5.7 (0.1)	2	5.1 (3.5)
	WR VI	0		0		0
Other gram-negative bacteriac	HIV−	16	2.8 (1.0)	6	2.5 (0.5)	10	2.4 (0.6)
	WR I–V	13	3.8 (2.0)	11	3.6 (2.0)	20	3.3 (1.7)
	WR VI	24	3.7 (1.1)	12	3.4 (0.6)	12	3.7 (0.9)
C. albicans	HIV−	0		0		0
	WR I–V	0		0		0
	WR VI	0		0		0
Other yeasts	HIV−	6	2.7 (1.6)	2	2.4d	2	1.7d
	WR I–V	6	1.9 (0.4)	1	1.6d	5	2.3 (0.6)6
	WR VI	6	3.1d	0		6	4.1d
Dermatophytes	HIV−	0		0		0
	WR I–V	0		0		0
	WR VI	0		0		0
Nose		Axilla		Perineum		Toe web
% Colonized	Mean density (SD)	% Colonized	Mean density (SD)	% Colonized	Mean density (SD)	% Colonized	Mean density (SD)
48	4.5 (1.2)	16	3.9 (1.0)	52	5.2 (1.2)	46	5.7 (1.4)
56	4.1 (1.5)	32	3.8 (1.3)	55	5.7 (1.3)	56	6.2 (1.4)
65	4.6 (1.3)	41	3.8 (1.5)	71	5.1 (1.2)	65	5.6 (1.6)
10–45 (1), 16 (7), 21 (9), 22 (29), 27 (42), 22–47 (24), 33 (12), 34 (32), 37 (16), 38 (30), 40 (4), 40–44 (44), 51 (39), 53 (38)	NA	2 (25), 8 (45), 8–12 (43), 9 (2)	NA	6 (25), 12 (2), 16–26 (43), 18 (45), 22 (20)	NA	5 (25), 11 (2), 16 (43)	NA
30	5.1 (1.2)	12	4.6 (1.2)	36	4.8 (1.8)	36	5.5 (1.9)
0		7	3.2 (1.3)	28	3.7 (1.7)	20	4.6 (2.2)
6	4.8d	6	2.4d	47	5.3 (2.2)	41	5.0 (2.7)
0		0		0		4	5.0 (2.2)
2	3.4 (2.5)	2	5.6 (1.4)	5	4.2 (1.7)	9	6.0 (2.4)
6	1.9d	0		0		0
32	3.5 (0.9)	28	3.5 (1.3)	40	4.2 (1.6)	34	3.1 (1.3)
38	4.1 (1.4)	15	4.2 (1.9)	43	4.0 (1.6)	29	3.8 (1.7)
6	2.64d	12	3.4 (1.0)	47	3.5 (1.7)	24	3.3 (0.6)
0		0		0		0
10	3.8 (0.8)	1	2.6d	5	2.3 (1.0)	2	1.9 (0.4)
0		0		0		0
2	2.2d	0		20	2.8 (0.7)	22	2.6 (0.9)
6	2.7 (1.5)	5	1.9 (0.2)	17	2.0 (0.4)	34	2.4 (0.7)
0		0		24	3.1 (0.7)	24	3.6 (1.8)
0		0		0		0
0		0		0		0
0		0		0		0

^aSee Materials and Methods and reference 33 for a description of the WR staging classification system. HIV⁻, HIV negative. 

^bDensity data are expressed as log (base 10). 

^cThe density of P. mirabilis is included. 

^dThe standard deviation is not calculated because only one density measurement is available. 

^eNA, data not available. 

^f7 (7) refers to the head alone; 20 (45) refers to the neck alone. The following refer to forearms and hands; 5–40 (43), 11 (42), 14–40 (44), 16–18 (4), 17 (21), 33 (29), 40 (45). 

Seropositive individuals did not demonstrate elevated S. aureus carriage rates and densities relative to seronegative controls by a χ² analysis grouping all body sites together. A χ² analysis by site did reveal a statistically significant elevation in the cutaneous density of this organism on the forehead of seropositive individuals as compared to controls (0.012 [Fisher's exact test]). Although overall statistical significance was not reached, S. aureus carriage rates for seropositive personnel (both WR stages I to V and WR stage VI) exceeded the high rates observed in seronegative individuals at all body sites examined. In some locations such as the axilla, this difference is striking (41% for WR stage VI versus 16% for seronegative individuals). In relation to historical controls, the colonization rates seen in WR stage VI are most dramatic. The latter group exhibits carriage rates that are at least three times as large as those of historical controls for all sites examined except the nose.

χ² analysis did not identify statistically significant differences in microbial densities or carriage rates for any other cultured organism. The P value for the exact χ² test for C. albicans was 0.2075. Of note, C. albicans carriage was observed exclusively in seropositive hosts; curiously, however, candida carriage was not observed in WR stage VI patients. The rates of C. albicans colonization at the axilla, toe interspaces, perineum, and nose were 1, 2, 5, and 10%, respectively, in seropositive individuals. This prevalence is rather striking given the infrequency with which this organism colonizes normal skin (31). These colonization rates exceed estimates of candidal proliferation in healthy adult populations as reported in the literature. Previous investigators have detected C. albicans in only 0.02 to 1% of toe interspaces and 0.06% of perineum cultures from healthy adult populations (1, 22).

DISCUSSION

HIV infection is associated with skin flora alterations. One study has compared carriage rates and densities of S. aureus and C. albicans in 19 seropositive and 19 seronegative patients (8). To our knowledge, however, a systematic analysis across many organisms and body sites has not been undertaken. Factors warranting such an examination include the high incidence of infectious disorders in seropositive individuals (18) and the (at least theoretical) potential for microorganism-driven HIV-disease acceleration. Colonization may accelerate immune system deterioration via superantigen activation of T cells and stimulation of apoptosis of Vβ-specific and cross-reactive T cells (13).

Data on the incidence of S. aureus colonization are not consistent. Some investigators find higher rates of nasal and perineal colonization in seropositive individuals versus healthy controls (3, 5, 8, 11, 16, 27). A study comparing the nasal carriage rates of seropositive subjects to those with chronic diseases and healthy hospital staff revealed incidences of 44, 31, and 23%, respectively (40). One analysis showed that rates rise with the degree of immunosuppression (40); another demonstrated that staphylococcal adherence to cells is enhanced in HIV and that the degree of enhancement correlates with severity of infection. Members of our group have previously found that S. aureus cutaneous infection rates rise by WR stage (34). Other investigators present conflicting data (16). A study of 301 seropositive patients found no association between nasal carriage rates of bacteria (including S. aureus) and CD4 counts (19). Another study found identical rates in homosexual seropositive men, asymptomatic individuals, and patients with AIDS or AIDS-related complex (5).

Quantitative techniques are utilized infrequently. A small study of 38 patients quantified microbial isolates at four body sites, while another categorized growth patterns into four groups (i.e., absent, rare, numerous discrete, and confluent colonies) (8, 14). The present study provides a more extensive quantitative estimation of staphylococcal colonization. The advantage of a quantitative analysis is that it allows one to discriminate between transient and nontransient carriage. Microbial densities need to surpass a stringent 5-log threshold before carriage is classified as persistent.

S. aureus colonization rates in seropositive subjects exceeded those in seronegative subjects at all sites examined, though significance was reached only at the forehead. Seropositive individuals demonstrated higher densities at the scalp and forehead, though significance was not achieved. The S. aureus carriage rate in seronegatives was dramatically greater at all sites evaluated (except the nose) than that seen in historical controls (Table 2). Most (82%) controls were colonized at one or more sites, versus historical estimates of 10 to 15%.

Reasons why the present study did not discern a significant difference in S. aureus carriage rates and densities between groups include limited sample size, the abnormally high rates seen in controls, and the use of antimicrobial agents for P. carinii pneumonia prophylaxis in seropositive individuals. The study design did not select for seronegative subjects who received aerosolized pentamidine or trimethoprim-sulfamethoxazole for other conditions; only a fraction of seropositive patients (roughly 17 patients) were under treatment with these agents.

The work environment of seronegative military personnel may account for their elevated S. aureus carriage rates. They participated in training camp exercises, whereas seropositive personnel were office based. The daily physical activity of seronegative individuals (including prolonged contact with water) resulted in minor skin injuries which expose bacterial skin adhesion molecules that facilitate S. aureus colonization (14, 16). Heat, perspiration, and frequent showering may have released these microorganisms from deep skin reservoirs (15), and close living quarters promoted aerosolized propagation of S. aureus (4, 32, 37, 43, 44).

Significant differences were not observed for other organisms. With respect to fungal colonization, our data contrast with those of a study that found significantly elevated C. albicans carriage rates in 40 asymptomatic seropositive former drug addicts relative to seronegative individuals (10). Patients with advanced disease (WR stage VI), despite having an increased risk of systemic fungal infections, surprisingly did not demonstrate cutaneous carriage. WR stage VI patients may have received systemic antifungals for candidiasis (thus altering carriage patterns), though records to this effect are absent.

Our data on dermatophytes confirm those of other investigators who failed to demonstrate elevated densities in seropositive subjects (10, 17, 36). The absence of dermatophytes from seronegative subjects in this study is unusual but not unprecedented. Dermatophyte prevalence in the toe webs, nail plates, and feet of seronegative individuals ranges from 0 to 32% in earlier studies (21, 36, 43). Lack of dermatophytes in seropositive subjects' toewebs in our study does contrast with earlier data. Dermatophytes have been isolated from 37 to 54% of toe webs, nail plates, and feet of seropositive subjects (17, 28, 36). As above, use of antifungals may explain the absence of dermatophytes in our seropositive subjects; most seronegative subjects, however, did not receive such therapy.

The fact that the S. aureus carriage rates of seropositive military personnel exceeded the high rates of controls at all sites suggests that HIV infection may be associated with cutaneous proliferation of this pathogen. The risk of infection and (potentially) of superantigen-mediated disease acceleration may warrant a review of ways to limit its presence on skin. Furthermore, the high rates seen in healthy military personnel at basic training are striking. Formal evaluation of the factors mediating S. aureus dissemination in this group may be justified to address infection during future training and deployment exercises.