Colonization with multidrug-resistant organisms (MDROs) is common within nursing homes; the rates often surpass those in acute care.Reference Cassone and Mody1 Knowledge of colonization risk factors can allow for the implementation of timely, cost-effective interventions. Several MDRO colonization risk factors have been well characterized, including antibiotic exposure, use of indwelling devices, functional disability, and longer hospital stay.Reference Cassone and Mody1 However, the influence of demographic characteristics, markedly sex and race, has seldom been investigated, and results have been contradictory.Reference Humphreys, Fitzpatick and Harvey2-Reference Graham, Lin and Larson4
In this study, we investigated whether sex and race influence individual MDRO prevalence using data collected from a prospective cohort study of newly admitted nursing home patients. Specifically, we compared the overall and body-site–specific prevalence of methicillin-resistant Staphylococcus aureus (MRSA), vancomycin-resistant enterococci (VRE), and resistant gram-negative bacilli (R-GNB) in 6 nursing homes in southeastern Michigan.
Methods
Study design and patient population
We conducted secondary analysis exploiting data from a larger prospective cohort study of newly admitted patients in 6 nursing homes in southeastern Michigan.Reference Mody, Foxman and Bradley5 MDRO colonization was assessed by swabbing (BactiSwab, Remel, Lenexa, KS) patient hands, nares, oropharynx, groin, and perianal area on enrollment, days 14 and 30, and monthly up to 6 months or discharge. MRSA, VRE, and R-GNB were identified using previously described methods.Reference Mody, Foxman and Bradley5 For this study, we included patients for whom all body sites were screened at least once (519 patients; 1,388 visits). Hispanic patients (n = 6) and Asian patients (n = 2) were excluded from the analyses due to low numbers.
Statistical analysis
We assessed overall and body site-specific MDRO prevalence, stratifying by race (non-Hispanic white and black) and sex. Prevalence was defined as colonization at any time during the patient’s stay. Multivariable logistic regression assessed race and sex as risk factors for MDRO prevalence after adjusting for the following factors: age; physical self-maintenance score [PSMS, a functionality score ranging from 6 (independent) to 30 (dependent)]Reference Lawton and Brody6; Charlson comorbidity indexReference Charlson, Szatrowski, Peterson and Gold7; use of indwelling devices (feeding tube or urinary catheter) at enrollment; 30-day history of antibiotics at enrollment; prolonged previous hospitalization (>14 days); and total number of sampling visits. To understand whether subgroup differences in MDRO prevalence exist, we stratified each race by sex. For subgroup analysis, multivariable regression was used to assess whether subgroups had significantly higher odds of MDRO prevalence when compared to white females. All regression analyses were clustered by facility. Analyses were performed using SAS version 9.4 software (SAS Institute, Cary, NC).
Results
Demographics
Of the 651 patients enrolled, 519 (79.7%) had all body sites screened and were thus eligible for analysis. Among them, the mean age was 74.6 years (standard deviation [SD], 12.1); 56.5% were female, with no significant differences between facilities, and 60.3% were white. The mean PSMS score was 13.9 (SD, 4.4); the mean Charlson comorbidity index was 2.5 (SD, 2.0); 9.8% of patients had indwelling devices on enrollment; and the mean number of sampling visits was 2.7 (SD, 1.7). Significant differences in race distribution across facilities were identified (χ2, 365.0; P < .001), such that independence between race and facility-associated risks could not be established.
Multidrug-resistant organism prevalence
Overall, 74.2% were colonized with an MDRO at some point during their stay: 23.5% with MRSA, 49.5% with VRE, and 52.8% with R-GNB. Furthermore, 59% were colonized on enrollment: 15.2% with MRSA, 35.5% with VRE, and 34.3% with R-GNB. Among 155 patients followed after testing negative for all MDROs on enrollment, 79 (51.0%) acquired an MDRO during their stay.
Assessment of race and sex as risk factors for multidrug-resistant organism prevalence
Although no significant race differences could be established (Fig. 1), we found interesting sex differences in MDRO colonization. Male patients had 1.48 times higher odds of MDRO colonization than female patients (95% confidence interval [CI], 1.14–1.93). Body site-specific sex-differences in MDRO prevalence were identified. Specifically, male sex was an independent risk factor for higher odds of hand contamination (odds ratio [OR], 2.11; 95% CI, 1.48–2.99), nares colonization (OR, 1.41; 95% CI, 1.04–1.90), oropharynx colonization (OR, 1.53; 95% CI, 1.05–2.25), and groin colonization (OR, 1.38; 95% CI, 1.05–1.82) with MDROs.
MDRO colonization at any time of 519 nursing home patients, stratified by sex (A) and by race (B). Adjusted odds ratio not reported for VRE and R-GNB colonization of the nares as quasi-complete separation of data points occurs. Referent group for A is non-Hispanic white patients while the referent group in B is female patients. MDRO, multidrug-resistant organism; MRSA, methicillin-resistant Staphylococcus aureus; R-GNB, resistant gram-negative bacilli; VRE, vancomycin-resistant enterococci. Asterisk indicates statistical significance at P < .05.
Male patients had 1.34 times higher odds of colonization with MRSA than female patients (95% CI, 1.05–1.71). This difference was not driven by colonization at a specific body site. Furthermore, male patients had 1.84 times significantly higher odds of VRE colonization than female patients (95% CI, 1.69–2.01). Body-site–specific sex differences were identified. Specifically, male sex was an independent risk factor for hand contamination (OR, 2.47; 95% CI, 1.75–3.48) and perianal colonization (OR, 1.32; 95% CI, 1.06–1.65) with VRE. There were no significant sex differences in R-GNB prevalence in this population.
Subgroup analysis of race stratified by sex
To further our research objective, we evaluated whether significant differences in MDRO prevalence existed between subgroups. Black males were most often colonized with MDROs (78.1%), followed by white males (74.6%), black females (73.6%), and white females (72.1%). Multivariable analysis identified subgroup differences in MDRO prevalence appreciable at patient hands, nares, and oropharynx (Table 1).
Subgroup Analysis of MDRO Colonization at Any Time Among 519 Patients in 6 Nursing Homes
MDRO, multidrug-resistant organism; MRSA, methicillin-resistant Staphylococcus aureus; OR, odds ratio; R-GNB, resistant gram-negative bacilli; VRE, vancomycin-resistant enterococci; bold indicates statistical significance at P < .05.
a Multivariate model components: age, subgroup indicator variables (black male, black female, and white male), Charlson comorbidity index, functional status on enrolment, baseline presence of indwelling devices (feeding tube or urinary catheter), 30-day history of antibiotic exposure on enrolment, prolonged hospitalization (>14 d) prior to nursing home stay, and total number of sampling visits.
Discussion
In this study, we utilized multiple body screening to investigate whether race and sex, 2 readily available demographic characteristics, can be used to infer MDRO colonization risk in a heterogeneous nursing home population with a high prevalence of MRSA, VRE, and R-GNB. We have demonstrated that male patients have significantly higher odds of MDRO colonization than female patients and that these sex differences may only be appreciable at specific body sites and for specific MDROs.
The role of patient hands in highlighting sex differences is of significant interest. Patient hands have only recently received attention as sites for targeted MDRO surveillance and intervention.Reference Cao, Min, Lansing, Foxman and Mody8 Hand sampling is well accepted, and assessment of their role in MDRO transmission is important in the development of simple and effective interventions.
Literature reporting sex differences in MDRO prevalence has been relatively scarce and often contradictory. Male sex has been postulated as a risk factor for MRSA carriage,Reference Humphreys, Fitzpatick and Harvey2 VRE carriage,Reference Cassone and Mody1 and R-GNB carriage.Reference Graham, Lin and Larson4 On the other hand, a recent national prevalence survey highlighted contradictory results for MRSA, including a US-based study of >4,000 carriers failing to report a statistically significant difference. Such inconsistencies may be related to heterogeneity in population and sampling methods.Reference Humphreys, Fitzpatick and Harvey2 Possible explanations for higher MDRO colonization in male patients range from sexually dimorphic immune responseReference Grice and Segre9 to skin microenvironment and microbiomeReference Grice and Segre9 to differences in hand use and hygiene.Reference Fierer, Hamady, Lauber and Knight10
No race differences in MDRO prevalence were identified. However, we identify instances in which black male patients had significantly higher odds of colonization than white female patients at specific body sites, especially the hands. We recommend further research on whether race can infer colonization risk.
This study had several limitations. First, exclusion of patients without all body sites collected may have introduced selection bias. Second, our study is representative of 6 Michigan nursing homes; studies involving diverse areas may provide a more generalizable picture. Third, race distribution of patients across facilities was not homogeneous; some facilities had a larger black population. Furthermore, we did not investigate new acquisition trends. Building on the results of our pilot study, we hope to see future prospective comprehensive epidemiologic studies to understand these differences.
In conclusion, we have demonstrated that sex differences in MDRO prevalence exist in nursing homes, however, only for certain body sites and for specific MDROs. Future research is needed to understand these differences, particularly as they pertain to MDRO transmission, and to inform prevention strategies in an era of monumental increases in MDRO prevalence.
Acknowledgments
We thank the patients and nursing homes that participated in this study.
Financial support
This work was supported by the National Institutes of Health (grant nos. RO1 AG041780 and K24 AG050685 to L.M.); the University of Michigan Claude D. Pepper Older Americans Independence Center (REC Scholarship, and pilot grant no. AG024824 to M.C.); the Michigan Institute for Clinical and Health Research (grant number UL1TR002240 to L.M.); and the National Institute on Aging (grant no. P30 AG024824 to L.M.). L.M. is also supported by the Geriatrics Research, Education and Clinical Centers, Veterans’ Affairs Ann Arbor Healthcare System.
Conflicts of interest
The authors report no conflicts of interest relevant to this article.
