Obesity as a Determinant of Staphylococcus aureus Colonization Among Inmates in Maximum-Security Prisons in New York State

Abstract

Obesity increases a person's susceptibility to a variety of infections, including Staphylococcus aureus infections, which is an important cause of morbidity in correctional settings. Using a cross-sectional design, we assessed the association between obesity and S. aureus colonization, a risk factor for subsequent infection, in New York State maximum-security prisons (2011–2013). Anterior nares and oropharyngeal cultures were collected. Structured interviews and medical records were used to collect demographic, behavioral, and medical data. Body mass index (BMI; weight (kg)/height (m²)) was categorized as 18.5–24.9, 25–29.9, 30–34.9, or ≥35. The association between BMI and S. aureus colonization was assessed using log-binomial regression. Thirty-eight percent of 638 female inmates and 26% of 794 male inmates had a BMI of 30 or higher. More than 40% of inmates were colonized. Female inmates with a BMI of 25–29.9 (prevalence ratio (PR) = 1.37, 95% confidence interval (CI): 1.06, 1.76), 30–34.9 (PR = 1.52, 95% CI: 1.17, 1.98), or ≥35 (PR = 1.49, 95% CI: 1.13, 1.96) had a higher likelihood of colonization than did those with a BMI of 18.5–24.9 after we controlled for age, educational level, smoking status, diabetes status, and presence of human immunodeficiency virus. Colonization was higher among male inmates with a BMI of 30–34.9 (PR = 1.27, 95% CI: 1.01, 1.61). Our findings demonstrate an association between BMI and S. aureus colonization among female prisoners. Potential contributory biologic and behavioral factors should be explored.

Obesity has become a worldwide epidemic, with approximately 1.9 billion and 600 million adults worldwide classified as overweight and obese, respectively (1). The influence of obesity on noncommunicable diseases is well established (2, 3), and evidence is accumulating of its potential role in the transmission of communicable diseases (4–7). Obese individuals are reportedly more susceptible to community-associated pneumonia (8), wound infections (9), bacteremia (5), and sepsis (10). Staphylococcus aureus is an important cause of these conditions (11–13), and more than 80% of S. aureus isolates causing bacteremia in clinical settings are identical to the colonizing strain, indicating that S. aureus colonization plays an important role in subsequent infection (14). With the global burden of obesity, it is important to assess how excess adiposity may influence communicable diseases such as S. aureus infections.

Results from animal and human studies have suggested that obesity negatively affects the immune response (15, 16). There is also evidence that the composition of commensal organisms at different body sites differs between obese and normal-weight individuals, which could affect the pathogenicity of these organisms (17, 18). S. aureus is a commensal organism that colonizes the skin and mucosal surfaces of humans, and the anterior nares are considered its ecological niche (13). Among noninstitutionalized adults in the United States, 25% to 30% have colonization of the anterior nares, and 0.8% to 1.5% are colonized with methicillin-resistant S. aureus (19). Determinants of S. aureus colonization are largely unknown and are presumably dependent on both bacterial and host factors (20). Obesity as a host determinant of colonization has been reported among patients (21) and noninstitutionalized adults (19, 22). However, findings among obese men have been inconsistent (19, 22), and whether an association between obesity and S. aureus colonization at other mucosal sites, such as the oropharynx, exists has not been examined (23). Furthermore, the association has not been characterized among incarcerated persons, a population with high rates of S. aureus infections and methicillin-resistant S. aureus colonization and infection (24, 25). Because colonization is a risk factor for subsequent infection with the colonizing strain (14), the objective of the present study was to assess the association between body mass index (BMI) and S. aureus colonization among inmates in New York State maximum-security prisons. We hypothesized that higher BMIs would be associated with higher rates of colonization with S. aureus.

METHODS

Study population and setting

We used data from a cross-sectional study funded by the National Institutes of Health (Risk Factors for Spread of Staphylococcus aureus in Prisons) in 2 New York State maximum-security prisons, Bedford Hills Correctional Facility (women's prison) and Sing Sing Correctional Facility (men's prison). Detailed information on study methodology and sampling has been published previously (26, 27). Briefly, eligible participants were inmates aged 16 years or older who were capable of providing informed consent. Interested inmates spoke with an interviewer in a private room and were enrolled if informed consent was provided. The institutional review boards of Columbia University Medical Center and the New York State Department of Corrections and Community Supervision approved the study.

Data collection

The present study includes data collected from March 2011 to December 2013. Five trained interviewers used a structured questionnaire to collect demographic, behavioral, and medical history information from inmates. During the interview, the anterior nares and oropharynges of participants were swabbed, and culture swabs were later processed to assess asymptomatic colonization with S. aureus. In addition, interviewers extracted relevant data from inmate medical records, specifically the Health Services System problem list and medical provider's progress notes. All medical conditions identified or reported at screening upon an inmate's entry into the prison system and those identified during incarceration are included in the Health Services System problem list or provider's progress notes.

Outcome assessment

S. aureus characterization was carried out as previously described (28). Briefly, anterior nares and oropharynx specimens were incubated overnight at 35°C in enrichment broth (6% sodium chloride–supplemented tryptic soy broth). Aliquots of incubated broth were subsequently plated onto mannitol salt agar plates and incubated for 48 hours at 37°C. Individual colonies identified as possibly being S. aureus on the salt agar plates were streaked onto sheep blood agar plates, and StaphAurex (Remel Europe Ltd., Dartford, England) was used for confirmation (29). Confirmed isolates were further characterized using DNA sequence analysis of the protein A gene variable repeat region and compared using Ridom StaphType software, version 2.2.1 (Ridom GmbH, Würzburg, Germany) (30).

Exposure assessment

Although medical records data were available, self-reported height and weight were used to calculate BMI because recorded height and weight could have reflected values at prison entry rather than current height and weight. BMI was calculated as weight in kilograms divided by height in meters squared and used to categorize inmates as normal weight (BMI 18.5–24.9), overweight (BMI 25–29.9), obese (BMI 30–34.9), and severely obese (BMI ≥35).

Potential confounders

Demographic, behavioral, and health-related covariates were assessed. Specifically, self-reported age (<35, 35–49, or ≥50 years), sex, race/ethnicity (non-Hispanic white, non-Hispanic black, Hispanic, or other), educational level (less than high school, high school graduate, or some college/college graduate), and smoking status (yes or no) were measured. In addition, participants' medical charts were reviewed for comorbid medical conditions such as diabetes, human immunodeficiency virus, history of infection, and chronic skin conditions (dermatitis, psoriasis, or eczema). Last, length of time spent in prison (≤6 months, 7–12 months, 1.1–2 years, or >2 years) was also assessed.

Statistical analysis

Because there were so few underweight individuals, 7 inmates with a BMI less than 18.5 were excluded from the analysis. Means and standard errors were calculated for continuous variables, and frequencies (percentages) were calculated for categorical variables. Bivariate analyses were conducted using χ² test or t tests, as appropriate. Multivariable analysis using log-binomial regression was conducted to assess the association between BMI category and S. aureus colonization. Age, smoking status, diabetes status, and presence of human immunodeficiency virus were included in the analysis as potential confounders based on a priori knowledge of their associations with BMI category and S. aureus colonization. In addition, covariates that were associated with both BMI category and S. aureus colonization with a 10% level of significance in bivariate analysis and that changed the β coefficient associated with BMI category by 10% or more were retained in the final model. Effect measure modification was examined using cross-product terms, and the −2 log-likelihood estimates of the full (with the interaction terms) and reduced (without the interaction terms) models were compared. Because the existing literature suggests that the association between obesity and S. aureus colonization varies by sex (22) and because of the established biological differences in men and women related to adiposity (31), the final models were stratified by sex. Two multivariable models are presented: one stratified by sex alone and a second stratified by both sex and age in order to assess whether the association between obesity and S. aureus colonization also varied by age. In addition, because the nature of the association between BMI and S. aureus colonization has not yet been established, we explored the shape of the relationship using generalized additive models stratified by sex. Generalized additive models are an extension of generalized linear models that allow for the flexible specification of the relationship between BMI as a continuous variable and colonization with S. aureus.

Sensitivity analyses were conducted to examine how robust the findings were to different degrees of misclassification of BMI category. Based on the literature, we assumed perfect specificity and that any misclassification resulted from underreporting of weight or overreporting of height (32). We also assumed that misclassification occurred only between adjacent cells. Monte Carlo simulation studies with a uniform distribution and bootstrapping methods were then conducted based on assumptions in which 10% or 15% of the study population was randomly selected and BMI category changed to what was presumed to be the “true” BMI category. All statistical analyses were performed using SAS, version 9.3 (SAS Institute, Inc., Cary, North Carolina), and the manuscript was organized in a manner compliant with the Strengthening the Reporting of Observational Studies in Epidemiology statement (33).

RESULTS

Participation rates at both facilities were high (93.6% of male inmates and 80.7% of female inmates). Characteristics of the total sample stratified by BMI are presented in Table 1. The mean ages of the participating 638 women and 794 men were 36.8 years and 37.8 years, respectively. Approximately 27% of female inmates and 39% of male inmates had less than a high school education. Forty-three percent of female inmates were non-Hispanic black, 31% were non-Hispanic white, 17% were Hispanic, and 9% were in the “other” category. Non-Hispanic black inmates accounted for 55% of male inmates; 28% were Hispanic, 13% were non-Hispanic white, and 5% were in the “other” category. Obesity, defined by a BMI of 30 or higher, was prevalent among both male and female inmates. Approximately 38% of female inmates and 26% of male inmates were obese. The prevalence of obesity and severe obesity was highest among non-Hispanic black women (46%) and Hispanic women (39%). Twenty-seven percent of Hispanic male inmates and 27% of non-Hispanic black male inmates were either obese or severely obese. Prison inmates also had a high prevalence of S. aureus colonization, with 43% of female inmates and 44% of male inmates being colonized. Exclusive colonization of the anterior nares ranged from 9% among normal-weight female inmates to 18.5% among those who were severely obese. Among male inmates, 8.7% of normal-weight inmates were colonized compared with 17.6% of severely obese inmates. No discernable trend in oropharyngeal colonization was observed.

Table 1.

Characteristics of Inmates in New York State Maximum-Security Prisons, by Body Mass Index^a Category, 2011–2013

Characteristic	BMIa								Total
	18.5–24.9		25.0–29.9		30.0–34.9		≥35.0
	No.	%	No.	%	No.	%	No.	%	No.	%
	Female Inmates
No. of inmates	189	29.6	207	32.4	134	21.0	108	16.9	638	100
Age, yearsb	35.7 (11.1)		36.2 (11.6)		37.6 (10.9)		39.1 (11.2)		36.8 (11.3)
Race/ethnicity
Non-Hispanic white	81	42.9	65	31.4	28	20.9	27	25.0	201	31.5
Non-Hispanic black	58	30.7	90	43.4	72	53.7	55	50.9	275	43.1
Hispanic	28	14.8	38	18.4	20	14.9	22	20.4	108	16.9
Otherc	22	11.6	14	6.8	14	10.4	4	3.7	54	8.5
Educational level
Some high school	52	27.5	46	22.2	38	28.4	38	35.2	174	27.3
High school graduate	72	38.1	68	32.9	44	32.8	39	36.1	223	34.9
Some college or college graduate	65	34.4	93	44.9	52	38.8	31	28.7	241	37.8
Current smoker	131	69.3	129	62.3	87	64.9	74	68.5	421	66.0
Diabetic	7	3.7	19	9.2	15	11.2	22	20.4	63	9.9
HIV positive	9	4.8	8	3.9	6	4.5	6	5.6	29	4.5
Colonized with S. aureus	63	33.3	92	44.4	67	50.0	54	50.0	276	43.3
Nose exclusively	17	9.0	21	10.1	30	22.4	20	18.5	88	13.8
Throat exclusively	31	16.4	40	19.3	21	15.7	19	17.6	111	17.4
Both sites	15	7.9	31	15.0	16	11.9	15	13.9	77	12.1
	Male Inmates
No. of inmates	172	21.7	414	52.1	157	19.8	51	6.4	794	100
Age, yearsb	36.8 (11.6)		37.9 (10.2)		38.6 (9.8)		38.5 (9.6)		37.8 (10.4)
Race/ethnicity
Non-Hispanic white	20	11.6	54	13.0	18	11.5	8	15.7	100	12.6
Non-Hispanic black	92	53.5	226	54.6	92	58.6	25	49.0	435	54.8
Hispanic	48	27.9	111	26.8	44	28.0	16	31.4	219	27.6
Otherc	12	7.0	23	5.6	3	1.9	2	3.9	40	5.0
Educational level
Some high school	65	37.8	154	37.2	66	42.0	22	43.1	312	39.3
High school graduate	67	38.9	166	40.1	60	38.2	19	37.3	307	38.7
Some college or college graduate	40	23.3	94	22.7	31	19.7	10	19.6	175	22.0
Current smoker	107	62.2	228	55.1	93	59.2	27	52.9	455	57.3
Diabetic	4	2.3	12	2.9	11	7.0	10	19.6	37	4.7
HIV positive	6	3.5	9	2.2	3	1.9	1	2.0	19	2.4
Colonized with S. aureus	69	40.1	183	44.2	76	48.4	24	47.1	352	44.3
Nose exclusively	15	8.7	43	10.4	24	15.3	9	17.6	91	11.5
Throat exclusively	38	22.1	86	21.0	31	19.7	11	21.6	166	20.9
Both sites	16	9.3	54	13.0	21	13.4	4	7.8	95	12.0

Abbreviations: BMI, body mass index; HIV, human immunodeficiency virus; S. aureus, Staphylococcus aureus.

^a Weight (kg)/height (m)².

^b Values are expressed as mean (standard deviation).

^c “Other” category includes Asians, Native Americans, Pacific Islanders, and persons with 2 or more racial/ethnic groups.

In bivariate analyses, BMI category was significantly associated with S. aureus colonization of the nose or throat among female inmates (P < 0.01) but not among male inmates (P = 0.36) (Table 2). The prevalence of colonization was 33.3% among female inmates with a BMI of 18.5–24.9, 44.4% among female inmates with a BMI of 25–29.9, and 50.0% for female inmates with a BMI of 30 or higher. A similar but less pronounced pattern was observed among male inmates: Colonization rates were 40.1% in male inmates with a BMI of 18.5–24.9, 44.2% in those with a BMI of 25–29.9, 48.4% in those with a BMI of 30–34.9, and 47.1% in those with a BMI of 35 or higher.

Table 2.

Characteristics of Inmates in New York State Maximum-Security Prisons, by Staphylococcus aureus Colonization,^a 2011–2013

Characteristic	Female Inmates (n = 638)					Male Inmates (n = 794)
	Colonized		Not Colonized		P Value	Colonized		Not Colonized		P Value
	No.	%	No.	%		No.	%	No.	%
Total	276	43.3	362	56.7		352	44.3	442	55.7
Age, years					0.77					<0.01
<34	136	49.3	178	49.2		191	54.3	167	37.8
35–49	104	37.7	130	35.0		122	34.7	205	46.4
≥50	36	13.0	54	14.9		39	11.1	70	15.8
Race/ethnicity					0.50					0.94
Non-Hispanic white	82	29.7	119	32.9		47	13.4	53	12.0
Non-Hispanic black	117	42.4	158	43.6		192	54.5	243	55.0
Hispanic	49	17.8	59	16.3		95	27.0	124	28.0
Otherb	28	10.1	26	7.2		18	5.1	22	5.0
Educational level					0.10					0.18
Some high school	84	30.4	90	24.9		131	37.2	176	39.8
High school graduate	100	36.2	123	34.0		133	37.8	179	40.5
Some college or college graduate	92	33.3	149	41.2		88	25.0	87	19.7
Current smoker	190	68.8	231	63.8	0.18	187	53.1	268	60.6	0.03
Diabetic	31	11.2	32	8.8	0.31	19	5.4	18	4.1	0.38
HIV positive	13	4.7	16	4.4	0.86	8	2.3	11	2.5	0.85
History of infection	64	23.2	51	14.1	<0.01	57	16.2	67	15.2	0.69
Incarceration durationc					0.43					0.35
≤6 months	83	32.3	107	32.0		86	25.2	97	23.7
7–12 months	28	10.9	46	13.8		17	5.0	18	4.4
1.1–2 years	39	15.2	60	18.0		33	9.7	27	6.6
>2 years	107	41.6	121	36.2		205	60.1	267	65.3
BMI categoryd					<0.01					0.36
≤24.9	63	22.8	126	34.8		69	19.6	103	23.3
25.0–29.9	92	33.3	115	31.8		183	52.0	231	52.3
30.0–34.9	67	24.3	67	18.5		76	21.6	81	18.3
≥35.0	54	19.6	54	14.9		24	6.8	27	6.1

Abbreviations: BMI, body mass index; HIV, human immunodeficiency virus.

^a Colonization of the anterior nares, oropharynx, or both.

^b “Other” category includes Asians, Native Americans, Pacific Islanders, and persons with 2 or more racial/ethnic groups.

^c Incarceration durations were available for 591 female inmates and 750 male inmates.

^d Weight (kg)/height (m)².

In the multivariable analysis that was adjusted for age, educational level, smoking status, diabetes status, and presence of human immunodeficiency virus, higher BMI category was associated with higher prevalence of S. aureus colonization among female inmates (Table 3). Compared with female inmates with a BMI of 18.5–24.9, those with a BMI of 25–29.9 were 1.37 (95% confidence interval (CI): 1.06, 1.76) times more likely to be colonized with S. aureus. The prevalence ratios increased to 1.52 (95% CI: 1.17, 1.98) and 1.49 (95% CI: 1.13, 1.96) among female inmates with BMIs of 30–34.9 and 35 or higher, respectively. Among male inmates, a trend of higher S. aureus colonization was observed, but it was statistically significant only among those with a BMI of 30–34.9 (prevalence ratio = 1.27, 95% CI: 1.01, 1.61). Sensitivity analysis assuming 15% misclassification did not change the results among female inmates, but among male inmates, the observed association with BMI 30.0–34.9 was no longer statistically significant (Table 3). Multivariable regression further stratified by age category demonstrated a significant association between BMI and colonization among female inmates 35–49 years of age (Table 4). The prevalence of S. aureus colonization among female inmates in this age group with a BMI of 25–29.9 was 1.74 (95% CI: 1.08, 2.80) times that of female inmates with a BMI of 18.5–24.9. The prevalence ratios were 2.20 (95% CI: 1.48, 3.81) and 1.84 (95% CI: 1.13, 2.97) among female inmates with BMIs of 30–34.9 and 35 or higher, respectively.

Table 3.

Estimated Prevalence Ratios of Staphylococcus aureus Colonization of the Anterior Nares, Oropharynx, or Both Among Inmates in New York State Maximum Security Prisons by Body Mass Index Category, 2011–2013

Sex and BMI Categorya	Crude PR	95% CI	Adjusted PRb	95% CI	Adjusted PRc	95% CI
Female inmates (n = 638)
18.5–24.9	1.00	Referent	1.00	Referent	1.00	Referent
25.0–29.9	1.33	1.04, 1.72	1.37	1.06, 1.76	1.31	1.04, 1.58
30.0–34.9	1.50	1.15, 1.95	1.52	1.17, 1.98	1.48	1.21, 1.76
≥35.0	1.50	1.14, 1.97	1.49	1.13, 1.96	1.49	1.21, 1.77
Male inmates (n = 794)
18.5–24.9	1.00	Referent	1.00	Referent	1.00	Referent
25.0–29.9	1.10	0.89, 1.36	1.12	0.91, 1.38	1.12	0.90, 1.34
30.0–34.9	1.21	0.95, 1.54	1.27	1.01, 1.61	1.21	0.98, 1.46
≥35.0	1.17	0.83, 1.65	1.22	0.88, 1.71	1.24	0.94, 1.54

Abbreviations: BMI, body mass index; CI, confidence interval; PR, prevalence ratio.

^a Weight (kg)/height (m)².

^b Model was adjusted for age, educational level, smoking status, diabetes status, and human immunodeficiency virus status.

^c Model represents sensitivity analysis with perfect specificity and 85% sensitivity, resulting in a 15% misclassification between adjacent BMI categories. The model was adjusted for age, educational level, smoking status, diabetes status, and human immunodeficiency virus status.

Table 4.

Estimated Prevalence Ratios for Staphylococcus aureus Colonization^a by Body Mass Index Category and Age Category in Inmates in New York State Maximum-Security Prisons, 2011–2013

Age and BMI Categoryb	Female Inmates (n = 638)		Male Inmates (n = 794)
	PRc	95% CI	PRc	95% CI
<35 years
18.5–24.9	1.00	Referent	1.00	Referent
25.0–29.9	1.16	0.83, 1.60	1.21	0.93, 1.58
30.0–34.9	1.07	0.73, 1.58	1.24	0.89, 1.72
≥35	1.30	0.89, 1.91	0.39	0.93, 2.06
35–49 years
18.5–24.9	1.00	Referent	1.00	Referent
25.0–29.9	1.74	1.08, 2.80	1.01	0.67, 1.52
30.0–34.9	2.20	1.39, 3.48	1.38	0.90, 2.11
≥35	1.84	1.13, 2.97	1.20	0.67, 2.17
≥50 years
18.5–24.9	1.00	Referent	1.00	Referent
25.0–29.9	1.61	(0.71, 3.67)	0.94	0.49, 1.79
30.0–34.9	1.97	(0.89, 4.40)	0.79	0.28, 2.17
≥35	1.54	(0.61, 3.89)	0.23	0.03, 1.62

Abbreviations: BMI, body mass index; CI, confidence interval; PR, prevalence ratio.

^a Colonization of the anterior nares, oropharynx, or both.

^b Weight (kg)/height (m)².

^c Models were adjusted for educational level, smoking status, diabetes status, and human immunodeficiency virus status.

The continuous relationship between BMI and the adjusted predicted probability of being colonized with S. aureus in both male and female inmates is depicted in Figure 1. For female inmates, the association had a distinct threshold effect characteristic of a quadratic association. Among male inmates, however, the association between BMI and S. aureus colonization had an inverted-U shape that was much more variable than that observed among the women. For female inmates, both the linear (P < 0.01) and nonlinear (P = 0.03) relationships were statistically significant; however, for male inmates, only the nonlinear (P = 0.02) relationship was significant (α = 0.05).

Figure 1.

The relationship between body mass index (weight (kg)/height (m²)) and the probability of being colonized with Staphylococcus aureus among male and female inmates in New York state maximum-security prisons, 2011–2013. A) Predicted probability of colonization with S. aureus by body mass index among female inmates after adjustment for age, educational level, smoking status, diabetes status, and presence of human immunodeficiency virus. Both the linear (P < 0.01) and nonlinear (P = 0.03) relationships were statistically significant. B) Predicted probability of colonization with S. aureus by body mass index among male inmates. Only the nonlinear (P = 0.02) relationship was significant (α = 0.05). The significant P values of the nonlinear curves suggest that the relationship between body mass index and the probability of S. aureus colonization is nonlinear for both men and women.

DISCUSSION

We observed significantly higher rates of colonization with S. aureus in female inmates with BMIs in the higher categories. This association appeared to have a threshold effect; the prevalence of S. aureus colonization was similar among obese and severely obese women, and both prevalences were significantly higher than the prevalence observed among normal-weight women. Among male inmates, however, the association was observed in those who were obese but not in those who were overweight or severely obese. These findings partially support our hypothesis that excess adiposity might serve as a host determinant of S. aureus colonization.

The association between BMI and S. aureus colonization among noninstitutionalized adults has been previously reported. Gorwitz et al. (19) used data from a US population-based survey (National Health and Nutrition Examination Survey) and observed a modest association in both male and female participants. Although their findings are consistent with what we report here, they used only anterior nares specimens to assess S. aureus colonization, thereby possibly underestimating the overall burden of S. aureus carriage. Furthermore, data from the National Health and Nutrition Examination Survey exclude incarcerated populations, which we show have a high burden of both obesity and S. aureus colonization. In another population-based study, Olsen et al. (22) assessed S. aureus colonization of the anterior nares in 3,878 men and women aged 30–87 years in Tromsø, Norway. They reported an association between BMI and S. aureus colonization only in premenopausal women.

Our study contributes to the existing literature because we assessed a population with a high burden of obesity and S. aureus colonization in which the association has not been examined. In addition, we included an oropharyngeal assessment, which suggested that the association of obesity with S. aureus colonization might be site specific; only nasal colonization rates were consistently higher in participants in the higher BMI categories. Future studies should be done to assess this association at various sites, particularly the groin or axillae, in an effort to understand site-specific factors that might mechanistically explain the observed association. We further add to the existing literature by stratifying the analysis by age and including men and women younger than 30 years of age. In doing so, we demonstrate that in our population the association was restricted to women of premenopausal age, specifically 35–49 years of age, which is consistent with findings from the study in Tromsø (22).

Various findings from the published literature suggest that a causal relationship between excess adiposity and susceptibility to microbial organisms, and specifically colonization with S. aureus, is biologically plausible. S. aureus colonization has been hypothesized to initiate a subclinical immune response that is insufficient to clear the organism from the site of colonization (34). Antimicrobial peptides such as human β-defensin 1–3 and human neutrophil peptide 1–3 can be induced or constitutively expressed in keratinocytes, which line the anterior nasal epithelium, and play a role in innate immune defenses against bacterial colonization (35). Antimicrobial peptide activity is reportedly significantly attenuated in nasal secretions of persistent carriers of S. aureus compared with intermittent carriers or noncarriers, suggesting that these peptides may play a role in clearance of the organism (34).

Leptin, a key adipokine that influences appetite and modulates the immune response, is now known to have immunomodulatory effects (36). It has also been shown to influence antimicrobial peptide secretions (37) and could serve as the link between obesity and S. aureus colonization. Chronically elevated levels of leptin, as is found in persons who are obese, can induce central leptin resistance, which may interfere with its role in antimicrobial defense (38). Furthermore, leptin has been implicated in the thymic aging observed in obese individuals that results in reduced T-cell diversity and immune surveillance and is also associated with chronic inflammation (15, 38). The mechanisms by which obesity influences human susceptibility to microbial organisms have yet to be elucidated, and implications of chronically elevated leptin levels for immune function merit further investigation.

We found that the association between obesity and S. aureus varied by sex and age. Altered levels of circulating hormones have been observed in obese persons (15), which may explain some of the age- and sex-related differences we observed. Specifically, a sexual dimorphism in body fat distribution typically results in higher amounts of subcutaneous adipose tissue in premenopausal women than in men with similar BMIs (31). Subcutaneous adipose tissue expresses higher levels of leptin (31), which could explain sex- and age-related differences in S. aureus colonization. Factors such as circulating sex hormones, which are altered differently in obese men compared with women, could also play a role (15). The variation by sex could also be explained by the precision with which BMI approximates excess adiposity in men and women. For a given BMI, women have greater amounts of total body fat than do men (39), which suggests that BMI might not capture excess adiposity as accurately in men as it does in women. Furthermore, BMI misclassifies individuals with high muscle mass, such as athletes, as overweight or obese (40). More granular measures of excess adiposity might clarify an association between obesity and S. aureus colonization among men.

Another possible biological explanation for the observed results is that obesity may perturb the microbiome of the anterior nares, rendering them more conducive to colonization by S. aureus. Evidence is accumulating that the human microbiome is influenced by and influences different pathological states (18, 41). Frank et al. (42) observed that the microbiota of the anterior nares of hospitalized patients are substantially different from those of healthy adults and hypothesized that underlying host pathology might be associated with shifts in the microbial composition of different anatomical sites, including the anterior nares. Other studies have also suggested that the interaction between site-specific microbiota and tissue-specific immune responses could hinder or propagate the proliferation of pathological species such as S. aureus (43, 44). Furthermore, bacterial interference of S. aureus with other species has been reported, which suggests that the ecology of the anterior nares and oropharynx plays a role in S. aureus colonization (45, 46). Given the increasing evidence implicating the gut microbiome in local and systemic immunity and the microbiome of other mucosal sites in tissue-specific immune responses (47–49), it is plausible that a change in the microenvironment of the anterior nares could explain the trend observed in our data of more common colonization of the anterior nares in obese individual than in individuals of normal weight.

The present study has several limitations. Because of the cross-sectional design, we cannot establish temporality. It is unlikely, however, that colonization of the anterior nares by S. aureus leads to obesity, although there may be an unmeasured factor driving the association. In addition, because the height and weight in medical records may have been recorded at prison entry and therefore not been current, we used self-reported height and weight data to calculate BMI, which could thus be subject to misclassification. However, inmates did not know their colonization status, and it is doubtful that colonized individuals were more likely to report heights and weights that would include them in higher BMI categories, so any misclassification is likely nondifferential. Sensitivity studies suggest that our results in female inmates were robust to 15% misclassification of BMI. Moreover, as previous research has shown, BMI may not precisely capture excess adiposity in male inmates compared with female inmates, which could have masked our ability to detect an association (39, 40). Furthermore, the type of adipose tissue—visceral adipose tissue as opposed to subcutaneous adipose tissue—might play an important role, and the prevalence of each differs by sex. However, we were unable to assess whether tissue type had any influence. Last, it is possible that there was residual confounding from undiagnosed disease, such as diabetes or human immunodeficiency virus.

Despite these limitations, the present study has several strengths. The study was conducted in a distinct population at high risk of S. aureus and methicillin-resistant S. aureus infection (50). This population is also relatively large, with considerable variability in BMI, which enabled us to make robust estimates of the associations. To our knowledge, our study was the first in which an oropharyngeal specimen was included when assessing the influence of BMI, and in doing so we demonstrated that the association was pronounced in the anterior nares and may therefore be site specific. Another mucosal site where this association might prove particularly important is the groin, where obesity has been shown to influence the microbial composition (51) and S. aureus colonization has been shown to play a role in subsequent infection (52).

In conclusion, our results indicate that higher adiposity is associated with higher prevalence of S. aureus colonization among incarcerated women. Approximately 26% of female inmates and 24% of male inmates were colonized in the anterior nares, which is comparable to published estimates in the US general population (19) but lower than what has been reported in other correctional settings (53). If obese individuals have a higher risk of colonization with organisms that may become pathogenic, then the global burden of obesity might play an important role in the transmission of communicable diseases. Understanding the consequences of obesity for S. aureus colonization and infection may improve our ability to identify at-risk groups and further tailor prevention efforts within relevant communities.