Abstract
BACKGROUND
Determining risk factors for acquisition of methicillin-resistant Staphylococcus aureus (MRSA) in hospitals is important for defining infection-control measures that may lead to fewer hospital-acquired infections.
OBJECTIVE
To determine patient-associated risk factors for acquisition of MRSA in a tertiary care hospital with the goal of identifying modifiable risk factors.
METHODS
A retrospective matched case-control study was performed. Case patients who acquired MRSA during hospitalization and 2 matched control patients were selected among inpatients admitted to target units during the period from 2001 through 2008. The odds of exposure to potential risk factors were compared between case patients and control patients, using matched univariate conditional logistic regression. A single multivariate conditional logistic regression model identifying independent patient-specific risk factors was generated.
RESULTS
A total of 451 case patients and 866 control patients were analyzed. Factors positively associated with MRSA acquisition were as follows: target unit stay before index culture; primary diagnosis of respiratory disease, digestive tract disease, injury or trauma, or other diagnosis compared with cardiocirculatory disease; peripheral vascular disease; mechanical ventilation with pneumonia; ventricular shunting or ventriculostomy; and ciprofloxacin use. Factors associated with decreased risk were receipt of a solid-organ transplant and use of penicillins, cephalosporins, rifamycins, daptomycin or linezolid, and proton pump inhibitors.
CONCLUSION
Among the factors associated with increased risk, few are modifiable. Patients with at-risk conditions could be targeted for intensive surveillance to detect acquisition sooner. The association of MRSA acquisition with target unit exposure argues for rigorous application of hand hygiene, appropriate barriers, environmental control, and strict aseptic technique for all procedures performed on such patients. Our findings support focusing efforts to prevent MRSA transmission and restriction of ciprofloxacin use.
Methicillin-resistant Staphylococcus aureus (MRSA) is a dominant hospital pathogen in the United States and worldwide.1–8 MRSA colonization increases the risk of infection and contributes to healthcare-associated transmission.9,10 Determining risk factors for MRSA acquisition in hospitals is important for defining infection-control measures, which may lead to reduction in MRSA transmission and ultimately fewer MRSA hospital-acquired infections (HAIs), reduced mortality from those infections, and subsequent reduction in unnecessary healthcare expenditures. Individual patient risk factors for MRSA acquisition that have been identified in previous studies include length of stay (LOS), the presence of open wounds, the presence of a tracheostomy or nasoenteric feeding tube, perioperative hemodialysis or apheresis, high Omega score (composite of common hospital procedures), high severity of illness score, the presence of an MRSA-positive roommate, dependency in activities of daily living, and fluoroquinolone use.11–18 Most of these studies investigated specific populations and used small numbers of subjects, and not all facilities reporting risk factors for MRSA colonization perform active surveillance for MRSA, making it impossible to determine the timing of acquisition.
Surveillance data obtained at the University of Pittsburgh Medical Center (UPMC)–Presbyterian University Hospital (PUH) during the period from 2005 through 2008 revealed an average rate of MRSA acquisition in the hospital of less than 2%. Although this is relatively low, hospitals now face the challenge of aiming for zero acquisition. A retrospective matched case-control study was conducted to determine patient-associated risk factors for MRSA acquisition at UPMC-PUH with the goal of identifying modifiable risk factors.
METHODS
Subjects and Setting
In 2001, UPMC-PUH, a 745-bed tertiary care teaching facility, implemented an MRSA prevention bundle in a target unit and then gradually increased the number of target units that participated. The bundle included MRSA active surveillance testing, hand hygiene, real-time notification of incident MRSA isolates, use of contact and droplet precautions, use of clean or dedicated equipment, enhanced environmental cleaning, and electronic flagging of patients. In 2003, UPMC-PUH implemented restriction of the use of certain antibiotics, including fluoroquinolones, carbapenems, line-zolid, tigecycline, and daptomycin. Active surveillance testing was accomplished by inoculating nasal swab samples onto BBL CHROMagar MRSA (BD Diagnostic Systems). Target units are defined as areas housing patient populations with increased risk of developing multidrug-resistant organism colonization and/or HAIs and include all intensive care units (medical, surgical, cardiothoracic, neurosurgical/neurology, solid-organ transplant, coronary care, and trauma), the orthopedic unit, and medical step-down areas. Samples for culture are obtained from all patients admitted to target units at admission, weekly, and on discharge from the target area. Target patients, defined as at-risk patients admitted from other healthcare facilities, undergo active surveillance testing at hospital admission. Compliance with process measures is regularly monitored.
Adult patients admitted to target units during the period from January 2001 through December 2008 with initial negative results at MRSA surveillance culture were potential subjects. Case patients were defined as patients who underwent MRSA conversion and fulfilled inclusion criteria: the presence of at least 1 prior negative result at MRSA active surveillance testing culture from a sample obtained at hospital admission and any subsequent MRSA-positive culture result (from a nasal or clinical/nonnasal sample) from a sample collected in a target unit 3 or more days after hospitalization. Inclusion criteria for control patients included the presence of at least 1 negative result at active surveillance testing culture from a nasal sample collected 3 or more days after hospitalization. Potential case patients and control patients were excluded if they had a concurrent or prior MRSA-positive culture result or history, and control patients were also excluded if they had a subsequent MRSA-positive culture result during the study period. Two unique control patients per case patient were chosen when possible, matched by (1) date of the index negative culture result within 7 days of the case patient’s index positive culture result; (2) hospital unit at time of culture sampling; and (3) minimum LOS, ie, the control patient’s hospital stay needed to be at least as long as the number of days between the case patient’s last negative culture result and the index MRSA-positive culture result. Index culture was defined as the first positive culture result for case patients and the corresponding negative culture result for control patients. Case patients were excluded if a matching control patient could not be identified.
Data from a subset of case patients (subset 1) whose negative culture results and subsequent MRSA-positive culture results were obtained during a contiguous target unit stay and whose positive culture results were obtained 3 or more days after admission to a target unit were compared with data from control patients and analyzed. Data from another subset of case patients (subset 2) whose converter status was based on a subsequent positive result at active surveillance testing culture, excluding case patients identified only on the basis of nonnasal clinical culture results, were similarly compared with data from control patients and analyzed. These subset analyses served to confirm results of the larger data set and to reduce detection bias in determining conversion from negative to positive MRSA carrier status.
Data Collection
The study was approved by the institutional review board of the University of Pittsburgh. Data were obtained from the UPMC Medical Archival Retrieval System and supplemented with data from an internal infection control database. The Medical Archival Retrieval System is the repository of information forwarded from the health system’s electronic clinical, administrative, and financial databases. It is indexed on every word in the medical record and is capable of recovering information on all encounters for a given patient that occurred between specified dates.19
To ensure patient confidentiality, all data were deidentified with use of an honest broker system. Data collection was performed using computerized algorithms that did not discriminate between case patients and control patients. Variables collected included demographic data; International Classification of Diseases, Ninth Revision, Clinical Modification diagnosis and procedure codes; comorbid conditions and devices present at admission; institution of contact isolation for other multidrug-resistant organisms; medical procedures and major surgical procedures performed during hospitalization; laboratory detection of other multidrug-resistant organisms; and use of oral and intravenous antibiotics, mupirocin administered topically, H2 blockers, proton pump inhibitors, and vasopressors during hospitalization. The Charlson comorbidity index was calculated, and laboratory values that reflected the patient’s clinical condition around the time of index culture sampling were collected.
Statistical Analysis
The odds of exposure to potential risk factors for MRSA acquisition were compared for case patients and control patients using univariate conditional logistic regression analysis for both binary and continuous data, conditioning on matched sets of subjects. Matched odds ratios (ORs) for categorical variables reflect the odds of exposure to a variable for case patients compared with control patients, whereas matched ORs for continuous variables reflect the increase in matched OR for every unit increase in the value of the variable. A single multivariate conditional logistic regression model to identify independent patient-specific risk factors significantly associated with MRSA acquisition was generated using stepwise regression methods on variables that yielded a P value of less than .10 on univariate analysis, using a cutoff P value of less than .05. Goodness of fit was assessed by examining plots of Pearson and deviance residuals and influence measures. Selected biologically plausible interactions between covariates found significant on multivariate analysis were tested. Similar statistical methods were applied to subsets. Prevalence rates of MRSA-positive nasal screening culture results among UPMC patients who had chronic kidney disease, required hemodialysis, and/or had received solid-organ transplants were calculated to investigate univariate and multivariate analysis results. All analyses were performed with SAS 9.2 software (SAS Institute).
RESULTS
A total of 474 case patients were identified. Twenty-three case patients could not be matched to control patients and were excluded, resulting in 451 case patients and 866 matched control patients. Subset 1 included 359 case patients and 599 matched control patients. Subset 2 included 336 case patients and 651 matched control patients. The MRSA acquisition rate (proportion of patients who underwent MRSA conversion on the basis of nasal culture results) among patients screened with active surveillance testing during the 8-year study period was 353 of 31,448 (1.12%), decreasing from 91 of 4,312 (2.11%) during the first half of the study period to 262 of 27,136 (0.97%) during the second half of the study period. There were no outbreaks of MRSA acquisition or MRSA HAI, and MRSA HAI rates declined during the study period from 196 MRSA HAIs in 183,540 patient-days (1.07 cases per 1,000 patient-days) in 2001 to 65 MRSA HAIs in 224,458 patient-days (0.29 cases per 1,000 patient-days) in 2008.
Overall, 20 variables were identified as significant on univariate analysis (Tables 1–3). Exposure to a target unit prior to index culture sampling was the only admission characteristic significantly associated with MRSA acquisition (Table 1). Among the potential risk factors and antibiotic use examined (Tables 2 and 3), the results of univariate analysis indicated that pneumonia, use of mechanical ventilation with or without pneumonia, ventricular shunting or ventriculostomy, and H2 blocker use were significantly associated with MRSA acquisition; ciprofloxacin use was borderline significant (P = .054).
TABLE 1.
Univariate Analysis of Demographic and Admission Characteristics of Patients Who Acquired Methicillin-Resistant Staphylococcus aureus during Hospitalization and Matched Control Patients
| Variable | Case patients (n = 451) | Control patients (n = 866) | OR (95% CI) | P |
|---|---|---|---|---|
| Age, years, mean ± SD | 60.2 ± 17.5 | 59.2 ± 16.4 | 1.00 (1.00–1.01) | .35 |
| Male sex | 250 (55.4) | 484 (55.9) | 0.95 (0.75–1.20) | .67 |
| Race | .98 | |||
| White | 330 (73.2) | 639 (73.8) | 1.00 | |
| Black | 42 (9.3) | 79 (9.1) | 1.02 (0.69–1.52) | |
| Other or unknown | 79 (17.5) | 148 (17.1) | 1.02 (0.75–1.41) | |
| Emergent or urgent admission | 401 (88.9) | 767 (88.6) | 1.06 (0.73–1.55) | .75 |
| Admission from long-term care facility | 41 (9.1) | 51 (5.9) | 1.52 (0.99–2.35) | .06 |
| Admitting service | .52 | |||
| Medical | 260 (57.6) | 520 (60.0) | 1.00 | |
| Surgical | 183 (40.6) | 329 (38.0) | 1.18 (0.88–1.58) | |
| Other | 8 (1.8) | 17 (2.0) | 0.92 (0.38–2.24) | |
| LOS, days, median (IQR) | 30 (17–50) | 32 (20–54) | 1.00 (1.00–1.00) | .51 |
| Prior to index culture | 12 (6–24) | 13 (8–24) | 1.00 (1.00–1.00) | .98 |
| In target unit | 21 (12–35) | 21 (12–34) | 1.00 (1.00–1.00) | .29 |
| Target unit stay prior to index culture | 447 (99.1) | 821 (94.8) | 6.27 (2.23–17.61) | <.001 |
| LOS in target unit prior to index culture, days, median (IQR) | 9 (5–18) | 10 (5–19) | 1.00 (1.00–1.00) | .65 |
| Other UPMC hospitalization during past 6 months | 100 (22.2) | 241 (27.8) | 0.73 (0.56–0.95) | .02 |
| Substance abuse or dependence | ||||
| Alcohol | 42 (9.3) | 95 (11.0) | 0.84 (0.57–1.23) | .37 |
| Drug | 23 (5.1) | 40 (4.6) | 1.16 (0.68–1.99) | .58 |
| Tobacco | 35 (7.8) | 72 (8.3) | 0.89 (0.58–1.38) | .61 |
| Primary diagnosisa | .08 | |||
| Cardiocirculatory disease | 51 (11.3) | 135 (15.6) | 1.00 | |
| Respiratory disease | 100 (22.2) | 146 (16.9) | 2.00 (1.26–3.17) | |
| Digestive tract disease | 48 (10.7) | 86 (9.9) | 1.59 (0.94–2.68) | |
| Cerebrovascular disease | 38 (8.4) | 71 (8.2) | 1.64 (0.87–3.06) | |
| Injury or trauma | 98 (21.8) | 186 (21.5) | 1.54 (0.97–2.45) | |
| Other diagnosisb | 115 (25.6) | 242 (27.9) | 1.38 (0.89–2.15) | |
| Age-adjusted Charlson comorbidity index, mean ± SD | 3.8 ± 2.9 | 4.0 ± 3.0 | 0.97 (0.93–1.01) | .13 |
| Total point value assigned for abnormal laboratory values,c median (IQR) | 5 (4–6) | 5 (4–6) | 0.95 (0.87–1.03) | .18 |
| Contact isolation for other target organism | 161 (35.7) | 339 (39.1) | 0.83 (0.64–1.07) | .14 |
| Discharge disposition | .15 | |||
| Alive | 286 (63.4) | 539 (62.2) | 1.00 | |
| Deceased | 112 (24.8) | 193 (22.3) | 1.10 (0.83–1.46) | |
| Unknown | 53 (11.8) | 134 (15.5) | 0.74 (0.52–1.06) | |
NOTE. Data are no. (%) of patients unless otherwise indicated. CI, confidence interval; IQR, interquartile range; LOS, length of stay; OR, odds ratio; SD, standard deviation; UPMC, University of Pittsburgh Medical Center.
Excludes 1 patient whose primary diagnosis was unknown.
Classified under hematology/oncology, infectious diseases, obstetrics/gynecology, rheumatology, dermatology, endocrinology, neurology, genitourinary, or psychiatry.
One point given for the presence of each of the following 9 laboratory values: white blood cell count of at least 15 or less than 3 ×109 cells/L, hematocrit level less than 30% or at least 46%, albumin level less than 3.5 g/dL, bilirubin level 2 mg/dL or more, creatinine level less than 0.6 or at least 1.4 mg/dL, bicarbonate level less than 22 or at least 32 meq/L, sodium level less than 130 or at least 150 mmol/L, potassium level less than 3.5 or at least 5.5 meq/L, and CD4+ count less than 200 cells/mm3. The most extreme laboratory value within 7 days of index culture was used and designated abnormal on the basis of hospital reference ranges and clinical cutoff values.
TABLE 3.
Univariate Analysis of Antibiotic Use during Hospital Stay for Patients Who Acquired Methicillin-Resistant Staphylococcus aureus during Hospitalization and Matched Control Patients
| Variable | Case patients (n = 451) | Control patients (n = 866) | OR (95% CI) | P |
|---|---|---|---|---|
| Antibiotics administered intravenously or orally | ||||
| Any antibiotic | 432 (95.8) | 840 (97.0) | 0.73 (0.40–1.32) | .30 |
| Penicillins | 239 (53.0) | 531 (61.3) | 0.70 (0.55–0.89) | .003 |
| Cephalosporins | 301 (66.7) | 622 (71.8) | 0.77 (0.59–1.00) | .048 |
| Carbapenems | 41 (9.1) | 108 (12.5) | 0.67 (0.46–1.00) | .048 |
| Vancomycin | 307 (68.1) | 595 (68.7) | 0.96 (0.74–1.24) | .76 |
| Daptomycin or linezolid | 35 (7.8) | 102 (11.8) | 0.62 (0.41–0.94) | .02 |
| Aminoglycosides | 62 (13.7) | 145 (16.7) | 0.78 (0.57–1.08) | .14 |
| Ciprofloxacin | 106 (23.5) | 166 (19.2) | 1.32 (0.99–1.76) | .054 |
| Levofloxacin or moxifloxacin | 39 (8.6) | 91 (10.5) | 0.77 (0.49–1.20) | .25 |
| Rifamycins | 11 (2.4) | 53 (6.1) | 0.32 (0.16–0.67) | .003 |
| Trimethoprim-sulfamethoxazole | 71 (15.7) | 156 (18.0) | 0.82 (0.59–1.13) | .22 |
| Metronidazole | 211 (46.8) | 429 (49.5) | 0.89 (0.70–1.13) | .33 |
| Clindamycin | 22 (4.9) | 47 (5.4) | 0.89 (0.53–1.50) | .66 |
| Macrolides | 89 (19.7) | 155 (17.9) | 1.13 (0.84–1.51) | .43 |
| Mupirocin administered topically | 22 (4.9) | 48 (5.5) | 0.88 (0.50–1.56) | .67 |
NOTE. Data are no. (%) of patients unless otherwise indicated. CI, confidence interval; OR, odds ratio.
TABLE 2.
Univariate Analysis of Potential Risk Factors for Hospital Acquisition of Methicillin-Resistant Staphylococcus aureus
| Variable | Case patients (n = 451) | Control patients (n = 866) | OR (95% CI) | P |
|---|---|---|---|---|
| Ischemic heart disease | 184 (40.8) | 383 (44.2) | 0.86 (0.68–1.09) | .21 |
| Peripheral vascular disease | 53 (11.8) | 77 (8.9) | 1.40 (0.96–2.04) | .08 |
| Chronic kidney disease | 93 (20.6) | 217 (25.1) | 0.76 (0.58–1.01) | .06 |
| Acute renal failure | 74 (16.4) | 163 (18.8) | 0.84 (0.62–1.14) | .28 |
| Requirement of hemodialysis | 92 (20.4) | 245 (28.3) | 0.63 (0.47–0.84) | .001 |
| Venous stasis or decubitus ulcer | 76 (16.9) | 117 (13.5) | 1.30 (0.95–1.79) | .10 |
| Immunosuppressiona | 163 (36.1) | 368 (42.5) | 0.74 (0.58–0.96) | .02 |
| Receipt of solid-organ transplant | 37 (8.2) | 126 (14.5) | 0.43 (0.28–0.68) | <.001 |
| Prior to current stay | 18 (4.0) | 76 (8.8) | 0.38 (0.21–0.67) | <.001 |
| During current stay | 19 (4.2) | 50 (5.8) | 0.67 (0.37–1.23) | .20 |
| Substantial steroid use | 125 (27.7) | 280 (32.3) | 0.79 (0.61–1.03) | .08 |
| Pneumonia | 269 (59.6) | 422 (48.7) | 1.62 (1.27–2.07) | <.001 |
| Without mechanical ventilation | 6 (1.3) | 20 (2.3) | 0.58 (0.23–1.45) | .25 |
| Mechanical ventilation | 415 (92.0) | 766 (88.5) | 1.86 (1.15–3.00) | .01 |
| With pneumonia | 263 (58.3) | 402 (46.4) | 1.69 (1.33–2.16) | <.001 |
| Without pneumonia | 152 (33.7) | 364 (42.0) | 0.71 (0.56–0.90) | .005 |
| Central venous catheterization | 443 (98.2) | 858 (99.1) | 0.52 (0.18–1.46) | .21 |
| Nasoenteric or ostomy tube feeding | 218 (48.3) | 379 (43.8) | 1.24 (0.98–1.58) | .08 |
| Upper gastrointestinal endoscopy | 50 (11.1) | 133 (15.4) | 0.61 (0.44–0.85) | .004 |
| Cardiac catheterization or contrast angiography | 92 (20.4) | 208 (24.0) | 0.80 (0.59–1.08) | .14 |
| Placement of invasive vascular deviceb | 38 (8.4) | 116 (13.4) | 0.57 (0.38–0.85) | .007 |
| Intra-aortic balloon pump | 12 (2.7) | 38 (4.4) | 0.55 (0.27–1.13) | .10 |
| Ventricular assist device | 14 (3.1) | 53 (6.1) | 0.49 (0.27–0.90) | .02 |
| Neurosurgery | 51 (11.3) | 78 (9.0) | 1.45 (0.94–2.24) | .09 |
| Ventricular shunting/ventriculostomy | 32 (7.1) | 37 (4.3) | 1.98 (1.14–3.43) | .02 |
| Cardiothoracic surgery | 80 (17.7) | 160 (18.5) | 0.96 (0.64–1.44) | .84 |
| Extracorporeal membrane oxygenation | 36 (8.0) | 86 (9.9) | 0.67 (0.38–1.18) | .17 |
| Gastrointestinal surgery | 67 (14.9) | 171 (19.7) | 0.67 (0.48–0.93) | .02 |
| Colon surgery | 13 (2.9) | 41 (4.7) | 0.58 (0.31–1.10) | .09 |
| Total number of major surgical procedures, median (IQR) | 1 (0–2) | 1 (0–2) | 0.93 (0.82–1.04) | .20 |
| Vasopressor use | 268 (59.4) | 538 (62.1) | 0.88 (0.68–1.14) | .34 |
| Proton pump inhibitor use | 224 (49.7) | 514 (59.4) | 0.63 (0.49–0.80) | <.001 |
| H2 blocker use | 407 (90.2) | 748 (86.4) | 1.55 (1.05–2.28) | .03 |
| Other target organism, positive culture or toxin test result | 145 (32.2) | 312 (36.0) | 0.82 (0.64–1.06) | .14 |
| Vancomycin-resistant Enterococcus | 128 (28.4) | 278 (32.1) | 0.80 (0.62–1.05) | .10 |
| Multidrug-resistant Acinetobacter | 2 (0.4) | 14 (1.6) | 0.24 (0.05–1.10) | .07 |
| Clostridium difficile | 30 (6.7) | 59 (6.8) | 0.95 (0.60–1.52) | .84 |
NOTE. Data are no. (%) of patients unless otherwise indicated. Other variables with P ≥ .20 not shown: congestive heart failure; cerebrovascular disease; intracranial hemorrhage; diabetes mellitus; chronic obstructive pulmonary disease; sepsis; any infection; C. difficile–associated disease; cirrhosis; malignancy; human immunodeficiency virus/AIDS; presence of intracardiac pacemaker or defibrillator, tracheostomy, or gastrointestinal stoma; peripheral venous or arterial and Foley catheterization; placement of subcutaneous intravenous port; transesophageal echocardiography; use of positive airway pressure ventilation; bronchoscopy; chest tube insertion; thoracentesis; percutaneous coronary intervention or stenting; lumbar puncture; blood transfusion; head and neck surgery; orthopedic surgery; vascular surgery; and other surgical subcategories. CI, confidence interval; IQR, interquartile range; OR, odds ratio.
Also includes immunosuppression from pharmacotherapy or radiotherapy, malignancy, and immunodeficiency syndromes.
Also includes coronary stenting and intracardiac pacemaker or defibrillator placement.
Factors present at admission, including hospitalization in a UPMC facility during the past 6 months (Table 1), immunosuppression (all causes), and receipt of a solid-organ transplant prior to admission, and some procedures performed during the hospital stay, including hemodialysis, placement of an invasive vascular device, placement of a ventricular assist device, upper gastrointestinal endoscopy, and gastrointestinal surgery (Table 2), were significantly associated with decreased risk of MRSA acquisition. However, most of these factors did not enter the final model. Medication use variables, including use of proton pump inhibitors (Table 2), penicillins, cephalosporins, carbapenems, daptomycin or linezolid, and rifamycins, were significantly associated with decreased risk (Table 3).
On multivariate analysis, target unit stay prior to index culture sampling had the greatest association with MRSA acquisition (Table 4). Other risk factors for MRSA acquisition were a primary diagnosis of respiratory disease, digestive tract disease, injury or trauma, or other diagnosis, compared with cardiocirculatory disease; peripheral vascular disease; mechanical ventilation with pneumonia; ventricular shunting and/or ventriculostomy; and ciprofloxacin use. Receipt of a solid-organ transplant and use of penicillins, daptomycin or linezolid, rifamycins, cephalosporins, or proton pump inhibitors were protective. No interaction variables were significant. Elimination of extreme or influential observations did not result in any substantial changes in the main model, and thus, all observations were retained.
TABLE 4.
Multivariate Model for Predicting Acquisition of Methicillin-Resistant Staphylococcus aureus
| Variable | OR (95% CI) | P |
|---|---|---|
| Target unit stay prior to index culture | 4.95 (1.71–14.34) | .003 |
| Primary diagnosis | .02 | |
| Respiratory disease vs cardiocirculatory disease | 2.25 (1.36–3.72) | |
| Digestive tract disease vs cardiocirculatory disease | 2.39 (1.35–4.25) | |
| Injury or trauma vs cardiocirculatory disease | 1.90 (1.14–3.18) | |
| Cerebrovascular disease vs cardiocirculatory diseasea | 1.51 (0.75–3.02) | |
| Other diagnosis vs cardiocirculatory disease | 1.64 (1.02–2.62) | |
| Peripheral vascular disease | 1.54 (1.01–2.33) | .04 |
| Receipt of solid-organ transplant | 0.33 (0.20–0.54) | <.001 |
| Receipt of mechanical ventilation with pneumonia | 1.75 (1.34–2.29) | <.001 |
| Ventricular shunting or ventriculostomyb | 2.20 (1.23–3.94) | .008 |
| Ciprofloxacin use | 1.57 (1.13–2.17) | .007 |
| Penicillin use | 0.66 (0.51–0.89) | .002 |
| Daptomycin or linezolid use | 0.56 (0.35–0.89) | .01 |
| Rifamycin use | 0.38 (0.17–0.82) | .01 |
| Cephalosporin use | 0.73 (0.55–0.98) | .03 |
| Proton pump inhibitor use | 0.68 (0.52–0.89) | .006 |
NOTE. CI, confidence interval; OR, odds ratio.
The CI for this primary diagnosis category crossed 1.00, but the category is retained in the model as part of a categorical variable.
Among all patients who underwent this procedure, 37 of 69 (53.6%) had an intracranial hemorrhage.
For subset 1 (contiguous target unit stay only), the final model had only 2 variables, mechanical ventilation with pneumonia (adjusted OR, 1.79 [95% confidence interval {CI}, 1.33–2.41]; P < .001) and ventricular shunting and/or ventriculostomy (adjusted OR, 1.97 [95% CI, 1.03–3.77]; P = .04), that were associated with MRSA acquisition. Except for rifamycin and daptomycin or linezolid use, protective factors were similar to those in the main model and had comparable adjusted ORs and P values (data not shown). Carbapenem use (adjusted OR, 0.58 [95% CI, 0.36–0.95]; P = .03) and placement of an invasive vascular device (adjusted OR, 0.60 [95% CI, 0.36–0.98]; P = .04) were additional protective factors.
In the final model for subset 2 (active surveillance testing culture only), 4 factors were associated with MRSA acquisition, including target unit stay prior to index culture (adjusted OR, 4.65 [95% CI, 1.37–15.76]; P = .01), mechanical ventilation with pneumonia (adjusted OR, 1.67 [95% CI, 1.23–2.28]; P = .001), ventricular shunting and/or ventriculostomy (adjusted OR, 2.26 [95% CI, 1.12–4.55]; P = .02), and ciprofloxacin use (adjusted OR, 1.70 [95% CI, 1.17–2.49]; P = .006). Protective factors were likewise similar to those in the main model (data not shown). Additional protective factors identified were cardiac catheterization and/or contrast angiography (adjusted OR, 0.61 [95% CI, 0.40–0.91]; P = .02) and placement of an invasive vascular device (adjusted OR, 0.58 [95% CI, 0.34–0.99]; P = .047).
During the study period, the prevalence of MRSA-positive nasal culture results among MRSA-screened patients at UPMC with a diagnosis of chronic kidney disease or requirement of hemodialysis was 345 (20.2%) of 1,705, compared with 3,350 (11.3%) of 29,745 among other patients (adjusted OR, 1.98 [95% CI, 1.74–2.25]; P < .001 [adjusted for receipt of solid-organ transplant]); and their first nasal screening culture result was positive more often than that of patients without chronic kidney disease and who did not require hemodialysis (242 [14.2%] of 1,705 vs 2,451 [8.2%] of 29,745; adjusted OR, 1.88 [95% CI, 1.63–2.18]; P < .001). However, the prevalence of MRSA-positive nasal culture results among MRSA-screened patients who had received solid-organ transplants was not significantly different from that among patients who had not (318 [13.7%] of 2,317 vs 3,377 [11.6%] of 29,133; adjusted OR, 1.04 [95% CI, 0.91–1.18]; P = .56 [adjusted for chronic kidney disease or requirement of dialysis]).
DISCUSSION
In general, few modifiable risk factors for acquisition of MRSA were identified in this large retrospective study. The association of certain primary diagnoses and peripheral vascular disease with MRSA acquisition could be reflective of disease mechanisms, underlying health status of the patients, the long-term nature of illnesses in these diagnostic categories, or the types of procedures performed for these patients.
The association of mechanical ventilation with pneumonia and MRSA acquisition is not surprising, because mechanical ventilation is a known risk factor for hospital-acquired pneumonia and MRSA is a common pathogen in ventilator-associated pneumonia.20,21 The link between mechanical ventilation and pneumonia may be reflective of either development of hospital-acquired MRSA pneumonia among patients receiving ventilation or pneumonia promoting MRSA acquisition among patients receiving ventilation. Mechanical ventilation with pneumonia is plausible as a risk factor, because the endotracheal tube bypasses normal upper respiratory tract defenses against aspiration and prevents coughing and clearing of secretions. Respiratory tract damage, disruption of respiratory flora, and increased secretions due to pneumonia result in a favorable environment for MRSA. In addition, intubation, suctioning, and other airway manipulation could serve as a portal of entry for MRSA from the hands or equipment of healthcare workers.
Ventricular shunting and/or ventriculostomy was the only surgical procedure associated with acquisition of MRSA. Its significance may reflect the health status of patients who undergo the procedure or the indications for the procedure. These patients often undergo multiple manipulations for drainage of cerebrospinal fluid, which might explain the increased risk of MRSA acquisition. However, ventricular shunting was performed for only 32 (7.1%) of case patients.
Requirement of hemodialysis, which is a known risk factor for MRSA infection and colonization, was not found to be a risk factor for MRSA acquisition.3 On univariate analysis, the requirement of hemodialysis seemed protective and chronic kidney disease trended toward being protective (Table 2). Patients with a diagnosis of chronic kidney disease or who required hemodialysis had significantly higher prevalence of MRSA-positive nasal cultures, as well as significantly higher rates of incident positive results on initial nasal screening, compared with the general hospital population. This greatly reduced the number of patients requiring hemodialysis who were eligible as case patients in our study, because we looked at converters only during hospitalization, which could account for the apparent protective effect. A healthy survivor effect among patients who had repeated healthcare exposure but were still eligible for the study may also explain why patients who received hemodialysis during hospitalization were not at higher risk for MRSA acquisition. Our data suggest that MRSA acquisition among patients with chronic kidney disease or who require hemodialysis occurs before hospitalization, such as in outpatient dialysis centers.
Ciprofloxacin use during the hospital stay was a significant risk factor for MRSA acquisition in the main model and was confirmed in the subset 2 analysis. This is consistent with previously published findings.11,22,23 Use of fluoroquinolones as a class and of the respiratory fluoroquinolones levofloxacin and moxifloxacin were not significant risk factors. Variations in antistaphylococcal and anti-MRSA activity may explain this differential finding within the class.
Previous studies of group-level antibiotic use have shown the use of β-lactam antibiotics to be a risk factor for MRSA colonization, which is in contrast to our findings, but prevalent colonization should be distinguished from acquisition and patient-level use should be distinguished from group-level use.11,24 Patient-level data in 1 study showed narrow-spectrum penicillins to be protective against MRSA colonization, which is consistent with our results.23 The effect of combinations of penicillins with anti-MRSA antibiotics on MRSA acquisition is a potential mechanism or confounder because of the high proportion of vancomycin use (approximately 68%) among our patient population (Table 3). Rifamycin and daptomycin or linezolid use were also associated with a decreased risk. It is plausible that, because these agents are active against MRSA, recovery of MRSA on active surveillance testing could have been suppressed. However, anti-MRSA antibiotics would generally not be used for controlling MRSA acquisition. Additional analysis of antibiotic use is planned to further explore study findings.
Receipt of a solid-organ transplant seemed to be protective. This may be confounded by antibiotic use among transplant recipients, because several antibiotic classes were shown to be protective. The similar prevalence of MRSA-positive nasal screening culture results among the populations of UPMC patients who had or had not received solid-organ transplants is consistent with study findings. The mechanisms behind these findings are unclear and may be explored in future studies.
The use of proton pump inhibitors was also found to be associated with decreased risk of MRSA acquisition. This is difficult to explain in light of the association of the use of proton pump inhibitors with pneumonia and infection with other healthcare-associated pathogens, such as Clostridium difficile.25–28
The cardiac procedures that were found to be significant on subset analyses are consistent with the finding in the main model of comparatively lower risk of MRSA acquisition among patients with a primary diagnosis of cardiocirculatory disease, who are more likely to undergo these procedures. This finding may be reflective of the underlying good health status of subjects who undergo some of these procedures. Only a small number of patients (less than 7%) required intra-aortic balloon pump and ventricular assist device support. The emergence of these protective factors only on subset analyses may reflect differences in the method of identification of case patients and control patients in the subsets.
As a result of the matching method, it can be presumed that case patients and control patients were exposed to similar group-level antibiotic use, MRSA colonization pressure, care workload, and unit location, and thus these could not be studied as risk factors.11,29 Similarly, LOS parameters would be reflective of matching accuracy and could not be studied as risk factors. Because subjects were chosen on the basis of results from cultures performed using samples obtained at a target unit, the resulting adjusted ORs for the dichotomous variables indicating prior exposure to a target unit may not truly reflect the odds of exposure in a general population of MRSA converters to a certain type of unit. Because target units housed populations that were at higher risk for MRSA HAIs than were populations in other units, this finding is plausible but should be interpreted with caution. Perhaps this was due to higher MRSA burden and/or performance of more invasive procedures in target units than in nontarget units.
The limitations of our study include its retrospective nature and the method of data capture by means of electronic extraction alone. The sensitivity of MRSA detection also depended on the culture sensitivity. However, these potential biases would unlikely be different between the case and control groups. As a result of multiple comparisons among many variables, the P value of less than .05 may be overly sensitive but was felt to be appropriate because of the study’s exploratory nature. Because of study design, residual confounding may have occurred and causation cannot be determined.
We report one of the largest single-center studies of patient-associated risk factors for MRSA acquisition performed in a tertiary care hospital. Our data revealed a number of risk factors, most not modifiable but some of which could be targeted for more intensive active surveillance testing in hopes of detecting MRSA acquisition as it occurs so that precautions could be implemented sooner. Increased risk for MRSA acquisition seems to be associated with underlying primary diagnoses and exposure to conditions in target units. The significance of ciprofloxacin use as a risk factor for MRSA acquisition supports restricting its use. Overall, few medical or surgical procedures studied as possible risk factors for MRSA acquisition yielded a significant result. The use of the MRSA prevention bundle may have helped to safeguard patients undergoing procedures who might not otherwise have been protected, because this program identifies colonized patients and requires implementation of barrier precautions. However, the significance of target unit exposure indicates that there is room for improvement in curtailing MRSA spread, particularly in these settings. MRSA reduction efforts should be less focused on modifying specific patient-associated risk factors but rather on preventing transmission, with emphasis on rigorous application of basic infection control strategies, such as hand hygiene before and after patient or environmental contact, habitual cleaning of equipment, and intensified environmental control, as well as assiduous use of and removal of barriers as appropriate. These principles and strict aseptic techniques should be applied to the care of all patients undergoing medical and surgical procedures. Because these procedures are often medically necessary, it would be impossible to reduce their frequency. Instead, consistent application of infection-control measures might decrease the risk for MRSA acquisition during these procedures, as well as during patient stays in high-risk units.
Acknowledgments
We thank Lloyd G. Clarke, clinical systems analyst in the Division of Infectious Diseases, Jamie L. Gray, clinical systems analyst in the Department of Infection Control, University of Pittsburgh Medical Center, and Mary G. Krauland, postdoctoral fellow in the Infectious Diseases Epidemiology Research Unit, Graduate School of Public Health and School of Medicine, for their contribution to this study.
Footnotes
Potential conflicts of interest. All authors report no conflicts of interest relevant to this article.
References
- 1.Boyce JM. Methicillin-resistant Staphylococcus aureus: a continuing infection control challenge. Eur J Clin Microbiol Infect Dis. 1994;13(1):45–49. doi: 10.1007/BF02026126. [DOI] [PubMed] [Google Scholar]
- 2.Klevens RM, Edwards JR, Tenover FC, McDonald LC, Horan T, Gaynes R. Changes in the epidemiology of methicillin-resistant Staphylococcus aureus in intensive care units in US hospitals, 1992–2003. Clin Infect Dis. 2006;42(3):389–391. doi: 10.1086/499367. [DOI] [PubMed] [Google Scholar]
- 3.Klevens RM, Morrison MA, Nadle J, et al. Invasive methicillin-resistant Staphylococcus aureus infections in the United States. JAMA. 2007;298(15):1763–1771. doi: 10.1001/jama.298.15.1763. [DOI] [PubMed] [Google Scholar]
- 4.Boyce JM. Understanding and controlling methicillin-resistant Staphylococcus aureus infections. Infect Control Hosp Epidemiol. 2002;23(9):485–487. doi: 10.1086/502092. [DOI] [PubMed] [Google Scholar]
- 5.Boyce JM. Increasing prevalence of methicillin-resistant Staphylococcus aureus in the United States. Infect Control Hosp Epidemiol. 1990;11(12):639–642. doi: 10.1086/646114. [DOI] [PubMed] [Google Scholar]
- 6.Blot SI, Vandewoude KH, Hoste EA, Colardyn FA. Outcome and attributable mortality in critically ill patients with bacteremia involving methicillin-susceptible and methicillin-resistant Staphylococcus aureus. Arch Intern Med. 2002;162(19):2229–2235. doi: 10.1001/archinte.162.19.2229. [DOI] [PubMed] [Google Scholar]
- 7.Cosgrove SE, Sakoulas G, Perencevich EN, Schwaber MJ, Karchmer AW, Carmeli Y. Comparison of mortality associated with methicillin-resistant and methicillin-susceptible Staphylococcus aureus bacteremia: a meta-analysis. Clin Infect Dis. 2003;36(1):53–59. doi: 10.1086/345476. [DOI] [PubMed] [Google Scholar]
- 8.Cosgrove SE, Qi Y, Kaye KS, Harbarth S, Karchmer AW, Carmeli Y. The impact of methicillin resistance in Staphylococcus aureus bacteremia on patient outcomes: mortality, length of stay, and hospital charges. Infect Control Hosp Epidemiol. 2005;26(2):166–174. doi: 10.1086/502522. [DOI] [PubMed] [Google Scholar]
- 9.Davis KA, Stewart JJ, Crouch HK, Florez CE, Hospenthal DR. Methicillin-resistant Staphylococcus aureus (MRSA) nares colonization at hospital admission and its effect on subsequent MRSA infection. Clin Infect Dis. 2004;39(6):776–782. doi: 10.1086/422997. [DOI] [PubMed] [Google Scholar]
- 10.von Eiff C, Becker K, Machka K, Stammer H, Peters G. Nasal carriage as a source of Staphylococcus aureus bacteremia. N Engl J Med. 2001;344(1):11–16. doi: 10.1056/NEJM200101043440102. [DOI] [PubMed] [Google Scholar]
- 11.Muller A, Mauny F, Talon D, Donnan PT, Harbarth S, Bertrand X. Effect of individual- and group-level antibiotic exposure on MRSA isolation: a multilevel analysis. J Antimicrob Chemother. 2006;58(4):878–881. doi: 10.1093/jac/dkl343. [DOI] [PubMed] [Google Scholar]
- 12.Warren DK, Guth RM, Coopersmith CM, Merz LR, Zack JE, Fraser VJ. Epidemiology of methicillin-resistant Staphylococcus aureus colonization in a surgical intensive care unit. Infect Control Hosp Epidemiol. 2006;27(10):1032–1040. doi: 10.1086/507919. [DOI] [PubMed] [Google Scholar]
- 13.Fishbain JT, Lee JC, Nguyen HD, et al. Nosocomial transmission of methicillin-resistant Staphylococcus aureus: a blinded study to establish baseline acquisition rates. Infect Control Hosp Epidemiol. 2003;24(6):415–421. doi: 10.1086/502224. [DOI] [PubMed] [Google Scholar]
- 14.Marshall C, Wolfe R, Kossmann T, Wesselingh S, Harrington G, Spelman D. Risk factors for acquisition of methicillin-resistant Staphylococcus aureus (MRSA) by trauma patients in the intensive care unit. J Hosp Infect. 2004;57(3):245–252. doi: 10.1016/j.jhin.2004.03.024. [DOI] [PubMed] [Google Scholar]
- 15.Hashimoto M, Sugawara Y, Tamura S, et al. Acquisition of methicillin-resistant Staphylococcus aureus after living donor liver transplantation: a retrospective cohort study. BMC Infect Dis. 2008;8:155. doi: 10.1186/1471-2334-8-155. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Moore C, Dhaliwal J, Tong A, et al. Risk factors for methicillin-resistant Staphylococcus aureus (MRSA) acquisition in roommate contacts of patients colonized or infected with MRSA in an acute-care hospital. Infect Control Hosp Epidemiol. 2008;29(7):600–606. doi: 10.1086/588567. [DOI] [PubMed] [Google Scholar]
- 17.Ibelings MM, Bruining HA. Methicillin-resistant Staphylococcus aureus: acquisition and risk of death in patients in the intensive care unit. Eur J Surg. 1998;164(6):411–418. doi: 10.1080/110241598750004210. [DOI] [PubMed] [Google Scholar]
- 18.Rioux C, Armand-Lefevre L, Guerinot W, Andremont A, Lucet JC. Acquisition of methicillin-resistant Staphylococcus aureus in the acute care setting: incidence and risk factors. Infect Control Hosp Epidemiol. 2007;28(6):733–736. doi: 10.1086/516664. [DOI] [PubMed] [Google Scholar]
- 19.Yount RJ, Vries JK, Councill CD. The Medical Archival System: an information retrieval system based on distributed parallel processing. Inf Process Manag. 1991;27:379–389. [Google Scholar]
- 20.Cook DJ, Kollef MH. Risk factors for ICU-acquired pneumonia. JAMA. 1998;279(20):1605–1606. doi: 10.1001/jama.279.20.1605. [DOI] [PubMed] [Google Scholar]
- 21.Rubinstein E, Kollef MH, Nathwani D. Pneumonia caused by methicillin-resistant Staphylococcus aureus. Clin Infect Dis. 2008;46(suppl 5):S378–S385. doi: 10.1086/533594. [DOI] [PubMed] [Google Scholar]
- 22.Weber SG, Gold HS, Hooper DC, Karchmer AW, Carmeli Y. Fluoroquinolones and the risk for methicillin-resistant Staphylococcus aureus in hospitalized patients. Emerg Infect Dis. 2003;9(11):1415–1422. doi: 10.3201/eid0911.030284. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.LeBlanc L, Pepin J, Toulouse K, et al. Fluoroquinolones and risk for methicillin-resistant Staphylococcus aureus, Canada. Emerg Infect Dis. 2006;12(9):1398–1405. doi: 10.3201/eid1209.060397. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Tacconelli E, De Angelis G, Cataldo MA, Pozzi E, Cauda R. Does antibiotic exposure increase the risk of methicillin-resistant Staphylococcus aureus (MRSA) isolation? A systematic review and meta-analysis. J Antimicrob Chemother. 2008;61(1):26–38. doi: 10.1093/jac/dkm416. [DOI] [PubMed] [Google Scholar]
- 25.Laheij RJ, Sturkenboom MC, Hassing RJ, Dieleman J, Stricker BH, Jansen JB. Risk of community-acquired pneumonia and use of gastric acid-suppressive drugs. JAMA. 2004;292(16):1955–1960. doi: 10.1001/jama.292.16.1955. [DOI] [PubMed] [Google Scholar]
- 26.Aseeri M, Schroeder T, Kramer J, Zackula R. Gastric acid suppression by proton pump inhibitors as a risk factor for Clostridium difficile-associated diarrhea in hospitalized patients. Am J Gastroenterol. 2008;103(9):2308–2313. doi: 10.1111/j.1572-0241.2008.01975.x. [DOI] [PubMed] [Google Scholar]
- 27.Cordonnier C, Buzyn A, Leverger G, et al. Epidemiology and risk factors for gram-positive coccal infections in neutropenia: toward a more targeted antibiotic strategy. Clin Infect Dis. 2003;36(2):149–158. doi: 10.1086/345435. [DOI] [PubMed] [Google Scholar]
- 28.Owens RC, Jr, Donskey CJ, Gaynes RP, Loo VG, Muto CA. Antimicrobial-associated risk factors for Clostridium difficile infection. Clin Infect Dis. 2008;46(suppl 1):S19–S31. doi: 10.1086/521859. [DOI] [PubMed] [Google Scholar]
- 29.Merrer J, Santoli F, Appere de Vecchi C, Tran B, De Jonghe B, Outin H. “Colonization pressure” and risk of acquisition of methicillin-resistant Staphylococcus aureus in a medical intensive care unit. Infect Control Hosp Epidemiol. 2000;21(11):718–723. doi: 10.1086/501721. [DOI] [PubMed] [Google Scholar]