Empirical Antimicrobial Therapy for Bloodstream Infection Due to Methicillin‐Resistant Staphylococcus aureus: No Better than a Coin Toss
Background. Despite the high prevalence of methicillin‐resistant Staphylococcus aureus (MRSA) infection in the hospital, the proportion of patients with MRSA bacteremia who receive appropriate empirical therapy remains suboptimal.
Objective. To investigate the proportion of patients with MRSA bloodstream infection (BSI) who received appropriate empirical antibiotic therapy and to identify risk factors associated with receipt of appropriate empirical therapy.
Methods. We studied a cohort of patients from 10 hospitals. The primary outcome was the proportion of patients who received appropriate empirical antibiotic therapy for MRSA BSI. Appropriate therapy was defined as receipt of daptomycin, linezolid, quinupristin‐dalfopristin, or vancomycin within 1 calendar day after the first blood culture result positive for S. aureus (ie, before antimicrobial susceptibilities were known). Multivariable logistic regression was used to determine variables associated with receipt of appropriate empirical therapy.
Results. The study included 562 patients with MRSA BSI. The mean (±standard deviation) age of the patients was 
years, and 288 (51.2%) were male. Only 291 (51.8%) patients received appropriate empirical therapy. Patients were more likely to receive appropriate therapy if they required hemodialysis (odds ratio [OR], 1.36 [95% confidence interval {CI}, 1.00–1.85]), had undergone knee or hip arthroplasty (OR, 3.04 [95% CI, 1.21–7.6]), had a central venous catheter at admission (OR, 1.72 [95% CI, 1.01–2.93]), or had a McCabe score of 1 at admission (OR, 1.83 [95% CI, 1.16–2.83]). Bowel incontinence (OR, 0.41 [95% CI, 0.19–0.92]) and BSIs categorized as primary (OR, 0.41 [95% CI, 0.27–0.63]) were associated with a decreased likelihood of receiving appropriate empirical therapy.
Conclusions. Only half of patients with MRSA BSI received appropriate empirical therapy. Factors associated with receiving appropriate empirical antibiotics included the presence of a central venous catheter at admission and a history of joint arthroplasty. Surprisingly, prior MRSA infection was not predictive of receipt of appropriate antimicrobial therapy.
Received March 12, 2009; accepted May 20, 2009; electronically published October 5, 2009.
Methicillin‐resistant Staphylococcus aureus (MRSA) infections are responsible for a large proportion of infections in the community and in healthcare settings. Current estimates attribute 18,000 deaths per year to invasive community‐ and hospital‐acquired MRSA infections, in part because approximately three‐quarters of all invasive MRSA infections are bloodstream infections (BSIs).1 Among hospitalized patients, MRSA is the causative agent for up to 20% of all BSIs.2 It has been well documented that even with adequate antibiotic therapy, patients with MRSA bacteremia have greater morbidity and mortality and an increased length of stay than patients with methicillin‐susceptible S. aureus.3‐7
Rapid and appropriate antibiotic therapy clearly improves the outcomes of hospitalized patients with bacteremia.8 Few studies, however, have evaluated the appropriateness of empirical therapy prescribed for BSI due to MRSA. The few studies that have evaluated this have demonstrated that, despite the high prevalence of MRSA infection in the hospital, the proportion of patients with MRSA bacteremia who receive appropriate empirical therapy remains suboptimal. Roghmann demonstrated that patients with MRSA bacteremia were less likely to receive appropriate antibiotics during the first 48 hours after hospitalization than were patients with methicillin‐susceptible S. aureus bacteremia.6 Most recently, Kaye et al demonstrated that only 38.3% of patients with MRSA bacteremia received appropriate empirical antibiotic therapy at the time of admission.9 Identifying the factors that result in inappropriate empirical therapy for patients with MRSA BSI will facilitate efforts to change prescribing patterns and improve outcomes. This study was undertaken to investigate the proportion of patients with MRSA BSI who received appropriate empirical antibiotic therapy and to identify risk factors associated with receipt of appropriate empirical therapy.
Methods
Study Design and Hospital Settings
This was a nested case‐control study. An existing cohort was used that included patients with a history of invasive MRSA infections who were treated at Duke University Medical Center and at 9 other community hospitals that participated in the Duke Infection Control Outreach Network.9 Duke University Medical Center is a tertiary care hospital in Durham, North Carolina, with 750 beds, including 138 intensive care unit beds. It has 37,793 annual admissions. For the 9 hospitals affiliated with the Duke Infection Control Outreach Network, the mean number of annual admissions per hospital was 12,234 (median, 10,327 [range, 5,380–30,001]). The structure and function of the Duke Infection Control Outreach Network has been described elsewhere.10,11
During the study period, identical prospective surveillance methods for MRSA were used in the study hospitals. The patients from Duke University Medical Center who were included in this study were hospitalized during the period from 1999 through 2002. Data from patients admitted to the 9 network hospitals from 1999 through March 2003 were included. Informed consent was waived by the institutional review board at Duke University Medical Center because the data were deidentified.
Cohort Definition
Adult patients were included if they had a BSI due to MRSA. Only the initial episode of MRSA BSI was included for each patient. Patients with a history of BSI due to MRSA were excluded. Appropriate empirical therapy was defined as receipt of 1 or more of the following antimicrobial agents within 1 calendar day after the first blood culture result positive for S. aureus (ie, before antimicrobial susceptibilities were known): daptomycin, linezolid, quinupristin‐dalfopristin, or vancomycin.
Variables
Case patients were prospectively identified. Additional variables retrospectively collected from patients' medical records included demographic characteristics, comorbid conditions (including Charlson score),12 acute severity of illness (McCabe score),13 dates of hospitalization, time spent in the intensive care unit, presence of indwelling devices, functional status (evaluated according to the Katz criteria),14 surgical procedures performed, and antimicrobial therapy received in the hospital. BSIs were classified as community acquired, healthcare associated, or nosocomial according to previously published definitions.15 BSIs were defined according to definitions from the Centers for Disease Control and Prevention as primary (eg, catheter associated) or secondary (arising from an infection at another anatomic site).16 The date of diagnosis of BSI was defined as the date on which the first blood sample was obtained that subsequently had culture results positive for MRSA.
Data were also collected on mortality during the year after the date of infection, on the duration of hospitalization, and on readmissions to the hospital during the 90‐day period after infection. Outcome data were obtained from patients' medical records and the Social Security Death Index.17
Statistical Analysis
Data were collected by using Access (Microsoft). Statistical analyses were performed with SAS software, version 9.1 (SAS Institute). Bivariable analyses of continuous variables were performed by use of the Student t test or the Wilcoxon rank sum test. Dichotomous and ordinal variables were analyzed by use of the Fisher exact test or the χ2 test. Multivariable analysis was performed by using logistic regression for dichotomous variables. Two‐tailed P values of .05 or less were considered significant.
Multivariable models were constructed through a backward stepwise selection process by using variables that had a P value of .2 or less on bivariable analyses. If the β coefficient of model variables changed by 10% or more when a covariate was excluded from the model, then the covariate was retained in the final model. All P values were 2 sided.
Results
A total of 564 patients were enrolled in the study, and 2 were excluded because of incomplete antibiotic data. Of the 562 remaining patients, the mean (±standard deviation) age of the patients was 
years. Approximately one‐half of the patients were male (288 patients [51.2%]), and 294 (52.3%) were white (Table 1). Many patients with BSI due to MRSA were severely ill at the time of admission: 141 patients (25.1%) had a McCabe score of 1, which indicates an expected life span of less than 2 weeks. Furthermore, a majority of patients required assistance with activities of daily living; 250 (44.5%) required assistance with ambulation at the time of hospital admission, and 115 (20.5%) had bowel incontinence at the time of hospital admission. Two hundred seventy patients (48.0%) had a BSI that was categorized as nosocomial, and 376 (66.9%) had a BSI that was categorized as primary.
In bivariable analysis, the following 4 variables were associated with receipt of appropriate empirical therapy (Table 1): having undergone knee or hip arthroplasty (
), requiring hemodialysis (
), having an indwelling central venous catheter at the time of hospital admission (
), and being admitted to a tertiary care center (
). The following 2 variables were associated with receipt of inappropriate empirical therapy: bowel incontinence at the time of admission (
) and having a BSI categorized as primary (
).
In multivariate analysis, the following variables were independently associated with receipt of appropriate empirical antimicrobial therapy for MRSA infection (Table 2): prior knee or hip arthroplasty (odds ratio [OR], 3.04 [95% confidence interval {CI}, 1.21–7.60]), the presence of a central venous catheter at the time of admission (OR, 1.72 [95% CI, 1.01–2.93]), a McCabe score of 1 at admission (OR, 1.83 [95% CI, 1.16–2.83]), and requirement of hemodialysis (OR, 1.36 [95% CI, 1.00–1.85]). The following variables were associated with a decreased likelihood of receiving appropriate empirical MRSA therapy: bowel incontinence at admission (OR, 0.41 [95% CI, 0.19–0.92]) and a BSI categorized as primary (OR, 0.41 [95% CI, 0.27–0.63]). The following variables were retained in the final multivariate logistic regression model to control for the confounding effects of insurance status, Charlson score, and bladder continence at the time of hospital admission.
Discussion
To our knowledge, this is the largest study to examine risk factors for appropriate therapy for MRSA BSI. Patients who had a central venous catheter at the time of hospital admission were more likely to receive appropriate empirical therapy for MRSA infection than were patients without a central venous catheter. Given the known increased risk for MRSA infection associated with indwelling central venous catheters, including those that provide vascular access for hemodialysis, it is not surprising that these patients received appropriate empirical therapy at higher rates. Similarly, the increased risk of MRSA infection for patients who had undergone joint arthroplasty is well described,18‐20 which may explain the increased likelihood of receipt of appropriate empirical therapy for patients with prosthetic joints.
Of particular interest were variables that were independently associated with a decreased risk for receipt of appropriate therapy: bowel incontinence at the time of hospital admission and a BSI categorized as primary. It is possible that, among patients with bowel incontinence, in the presence of poor clinical status and poor hygiene, the treating physicians might have attributed the primary source of infection to the bowel or bladder rather than to the bloodstream. The fact that MRSA is an uncommon cause of bowel or bladder infection may have led the treating physicians not to provide empirical MRSA antimicrobial coverage to patients with bowel incontinence. Unfortunately, we were unable to explore this hypothesis in depth because of the retrospective nature of the study. We also explored the association between primary BSI (ie, catheter associated) and a decreased likelihood of receiving appropriate therapy. This association is particularly notable, since the presence of an indwelling central venous catheter at the time of admission was associated with an increased likelihood of receipt of appropriate therapy. Taken together, these data suggest that although catheters inserted at other institutions were appropriately recognized as potential infection sources, central venous catheters inserted during hospitalization were not accurately recognized as potential infection sources. Thus, there seems to be a 2‐fold opportunity here. We can prevent healthcare‐associated BSI due to MRSA by improving catheter insertion and care practices in the hospital. Moreover, we can improve the treatment of hospital‐acquired primary BSI due to MRSA by raising awareness of the risk of MRSA BSI associated with the placement of central venous catheters during hospitalization.
Interestingly, a previous MRSA infection was not associated with receipt of appropriate therapy. Although this study was underpowered to study the relationship between previous MRSA infection and appropriate therapy (fewer than 11% of patients had a history of MRSA infection), this finding suggests that healthcare workers need to be more cognizant of a patient’s MRSA history and of the increased risk for MRSA infection that is present among patients with previous MRSA infection or colonization.
This study is limited in that only hospitals in the southeastern United States were included. However, this study did include 10 hospitals (including a tertiary care center and several community hospitals) and is, to our knowledge, the largest published study to examine the appropriateness of treatment for BSI due to MRSA. Most of the data for this study were retrospectively collected, and biases related to this type of data collection pertain to these study data. However, cases of BSI were prospectively identified by infection control practitioners with the same standardized definitions at all study sites, which limits some types of biases, such as the misclassification bias.
BSI is associated with a greater than 2‐fold increase in mortality, even among patients who are relatively stable at the time of presentation at the hospital.21 BSI due to MRSA is of particular concern. A prior report by our group, which used the same cohort as this study, revealed a 90‐day post‐BSI mortality rate exceeding 33% and a duration of post‐BSI hospitalization of more than 17 days.9 Other studies have also demonstrated increased mortality and morbidity among patients with MRSA bacteremia, compared with other types of bacteremia.5,21,22 This study demonstrated that empirical treatment for MRSA bacteremia is surprisingly inappropriate and that, in many cases of primary BSI, the catheter is not recognized as a potential infection source. The results from this study can be used to improve the acute management of BSI due to MRSA, particularly among patients with catheter‐associated BSI.
Acknowledgments
Potential conflicts of interest. All authors report no conflicts of interest relevant to this article.
References
- 1. Klevens RM, Morrison MA, Nadle J, et al. Invasive methicillin‐resistant Staphylococcus aureus infections in the United States. JAMA 2007; 298:1763–1771.
- 2. Wisplinghoff H, Bischoff T, Tallent SM, Seifert H, Wenzel RP, Edmond MB. Nosocomial bloodstream infections in US hospitals: analysis of 24,179 cases from a prospective nationwide surveillance study. Clin Infect Dis 2004; 39:309–317.
- 3. 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; 26:166–174.
- 4. 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; 36:53–59.
- 5. Engemann JJ, Carmeli Y, Cosgrove SE, et al. Adverse clinical and economic outcomes attributable to methicillin resistance among patients with Staphylococcus aureus surgical site infection. Clin Infect Dis 2003; 36:592–598.
- 6. Roghmann MC. Predicting methicillin resistance and the effect of inadequate empiric therapy on survival in patients with Staphylococcus aureus bacteremia. Arch Intern Med 2000; 160:1001–1004.
- 7. Shurland S, Zhan M, Bradham DD, Roughmann MC. Comparison of mortality risk associated with bacteremia due to methicillin‐resistant and methicillin‐susceptible Staphylococcus aureus. Infect Control Hosp Epidemiol 2007; 28:273–279.
- 8. Leibovici L, Shraga I, Drucker M, Konigsberger H, Samra Z, Pitlik SD. The benefit of appropriate empirical antibiotic treatment in patients with bloodstream infection. J Intern Med 1998; 244:379–386.
- 9. Kaye KS, Anderson DJ, Choi Y, Link K, Thacker P, Sexton DJ. The deadly toll of invasive methicillin‐resistant Staphylococcus aureus infection in community hospitals. Clin Infect Dis 2008; 46:1568–1577.
- 10. Kaye KS, Schmidt K, Peiper C. The effect of increasing age on the risk of surgical site infection. J Infect Dis 2005; 191:1056–1062.
- 11. Kaye KS, Sloane R, Sexton DJ, Schmader KA. Risk factors for surgical site infections in older people. J Am Geriatr Soc 2006; 54:391–396.
- 12. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 1987; 40:373–383.
- 13. McCabe W. Gram‐negative bacteremia, I: etiology and ecology. Arch Intern Med 1962; 110:847–852.
- 14. Katz S, Ford AB, Moskowitz RW, Jackson BA, Jaffe MW. Studies of illness in the aged: the index of ADL: a standardized measure of biological and psychosocial function. JAMA 1963; 185:914–919.
- 15. Friedman ND, Kaye KS, Stout JE, et al. Health care–associated bloodstream infections in adults: a reason to change the accepted definition of community‐acquired infections. Ann Intern Med 2002; 137:791–797.
- 16. Garner JS, Jarvis WR, Emori TG, Horan TC, Hughes JM. CDC definitions for nosocomial infections, 1988. Am J Infect Control 1988; 16:128–140.
- 17. Social Security Index (SSDI). Available at: http://ssdi.rootsweb.com. Accessed December 15, 2007.
- 18. Edwards C, Counsell A, Boulton C, Moran CG. Early infection after hip fracture surgery: risk factors, costs and outcome. J Bone Joint Surg Br 2008; 90:770–777.
- 19. Kourbatova EV, Halvosa JS, King MD, Ray SM, White N, Blumberg HM. Emergence of community‐associated methicillin‐resistant Staphylococcus aureus USA 300 clone as a cause of health care‐associated infections among patients with prosthetic joint infections. Am J Infect Control 2005; 33:385–391.
- 20. Ridgeway S, Wilson J, Charlet A, Kefatos G, Pearson A, Coello R. Infection of the surgical site after arthroplasty of the hip. J Bone Joint Surg Br 2005; 87:844–850.
- 21. Kim PW, Perl TM, Keelaghan EF, et al. Risk of mortality with a bloodstream infection is higher in the less severely ill at admission. Am J Respir Crit Care Med 2005; 171:616–620.
- 22. Glenn KR, Craig A, Kainer MA. Thirty‐day and 180‐day case fatality rates among invasive methicillin‐resistant Staphylococcus aureus patients (Tennessee, 2004–2007). In: Program and abstracts of the International Conference on Emerging Infectious Diseases 2008; March 16–19, 2008; Atlanta, GA. Abstract 264.