Nasal Swabs Collected Routinely To Screen for Colonization by Methicillin-Resistant Staphylococcus aureus in Intensive Care Units Are a Sensitive Screening Test for the Organism in Clinical Cultures

Abstract

Background:

Many hospitals screen patients for methicillin-resistant Staphylococcus aureus (MRSA) on admission to the intensive care unit (ICU). We hypothesized that this screening information could be used to assist with empiric antibiotic decisions.

Methods:

The medical records of patients admitted to a university-affiliated community hospital as well as a tertiary-care university hospital were reviewed. Patients admitted to the ICU were screened for MRSA colonization with a nasal swab that was analyzed with either chromogenic medium (hospital 1) or polymerase chain reaction (PCR) (hospital 2). The results of the nasal swab were compared with clinical culture results.

Results:

There were 141 patients, and 167 cultures were obtained. The majority of the cultures (70%) were performed on sputum specimens in an effort to diagnose pneumonia. The remaining cultures were performed on blood (10.1%), incisions (21.5%), and urine (3.4%). The overall sensitivity of nasal swab results was 69.5%. However, the sensitivity was significantly higher for nasal swab screening performed within six days of clinical cultures compared with screening performed seven days or more before cultures were obtained. (79% vs. 46%; p < 0.0001). Sensitivity also differed significantly depending on the surveillance method, being significantly higher among patients screened with PCR within six days of developing an infection than in patients screened with chromogenic medium (88% vs. 65.5%; p = 0.006).

Conclusion:

Screening with PCR analysis of nasal swab specimens is a highly sensitive test for MRSA in clinical cultures. Clinicians may be able to use the swab results to tailor more appropriate empiric antimicrobial regimens. The results with chromogenic medium screening are markedly poorer, which suggests that clinicians should view them with caution.

Antibiotic resistance has been a clinical challenge since shortly after the introduction of antibiotics [1]. Methicillin became available for clinical use in 1959. The first evidence of methicillin-resistant Staphylococcus aureus (MRSA) was noted in 1961, and the first major hospital outbreak was described in 1968. Since that time, the incidence of MRSA has increased markedly. Such strains are endemic in many health centers, including intensive care units (ICUs), burn units, and long-term care facilities [2]. Greater than 50% of S. aureus strains in many ICUs have acquired methicillin resistance [3]. Infections caused by MRSA have been associated with increased morbidity and mortality rates, longer hospital stays, and greater hospital costs than methicillin-susceptible S. aureus (MSSA) infections [2].

A number of risk factors for MRSA infection and carriage have been characterized. Certain co-morbidities, such as diabetes mellitus and chronic obstructive pulmonary disease, increase the risk of an MRSA infection [4]. Patient demographics, such as residence at a long-term care facility, may increase the risk of MRSA carriage [5]. Previous exposures to antibiotics, as well as current or recent hospitalization, also increase the risk of MRSA colonization. In order to characterize the extent of MRSA prevalence fully, many hospitals have initiated active surveillance [6,7]. These programs generally utilize specimens obtained from the anterior nares to screen for clinically silent MRSA colonization.

To date, hospitals have used MRSA screening primarily for epidemiologic purposes and determination of the need for patient isolation. Although patients colonized with MRSA are at a higher risk of developing infections, there has been minimal evaluation of the concordance between nasal swab and subsequent culture results. We hypothesized that patients with negative screening for MRSA by nasal swabs are at low risk of growing MRSA in clinical cultures. Cumulative use of antibiotics is a major driving force behind the development of resistance. Accordingly, the ability to discern a subgroup of patients who do not require MRSA coverage would be an important clinical advance.

We conducted a study with the following objectives: (1) To evaluate the sensitivity of nasal swab screening in identifying patients at low risk of developing MRSA in clinical cultures; and (2) to determine the risk factors associated with discordance between nasal swabs and culture results.

Patients and Methods

The medical records of patients admitted to a university-affiliated community hospital as well as a tertiary-care university hospital were reviewed. Patients who had a clinical culture positive for MRSA and a nasal swab screening conducted during the admission were included in the study. The study population focused on patients admitted to the medical or surgical ICUs (University of Minnesota Medical Center, UMMC) and patients admitted to the medical/surgical and trauma/neuro ICUs (North Memorial Medical Center). Information regarding patient demographics and hospital stay was noted. We recorded which ICU treated the patient and what service was primarily responsible for the patient's care during the hospital stay. In addition, the severity of the illness was assessed by gathering data on hospital diagnoses, including respiratory dysfunction, renal failure necessitating dialysis, and use of vassopressor drugs. Hospitalization in the past three months was documented. We also documented which patients had been in the hospital for at least three days prior to the nasal swab being obtained.

University of Minnesota Medical Center screening protocol

At the University of Minnesota, patients admitted to the medical and surgical ICUs have been screened for MRSA since October 2007. A nasal swab is collected on admission to the ICU, and polymerase chain reaction (PCR) analysis is conducted for MRSA (Cepheid, Sunnyvale, CA). The results typically are available within 2–4 h. Patients are not routinely re-screened during an ICU admission, although they are if they are readmitted to the ICU at a later date.

North Memorial Medical Center screening protocol

Patients admitted to the medical/surgical ICU or to the trauma/neuro ICU at North Memorial Medical center have been screened for MRSA since January 2007. Patients are screened on admission only. A swab from the anterior nares is cultured in BBL CHROMagar Staph medium (Microbiology International, Paris, France). Specimens containing MRSA grow mauve when cultured in this medium. The results generally are available within 24–48 h.

Data analysis

All patients who were included in the study had both a clinical culture that was positive for MRSA and a nasal swab completed during the same hospitalization. The results of the nasal swab were compared with the culture results. Nasal swab results were categorized as true-positive (TP) if they were positive for MRSA or false-negative (FN) if they were negative for MRSA (Table 1). Sensitivity was calculated by the standard formula (Table 2). The representation of TP and FN nasal swabs relates only to the concordance or discordance of nasal swab results with clinical culture results and does not speak to the precision by which a positive nasal swab actually denotes colonization.

Table 1.

Calculation of Sensitivity for Methicillin-Resistant Staphylococcus aureus infection

Nasal swab	Positive clinical cultures
Positive	True-positive (TP)
Negative	False-negative (FN)

Table 2.

Sensitivity Profile ^a of Polymerase Chain Reaction vs. Chromogenic Medium at Different Intervals Between Nasal Swab Screening and Clinical Culture

Days between swab and culture	PCR	Chromogenic medium
0^b	78.9 (24)	72.2 (18)
1	85.7 (8)	83.3 (6)
2	100 (6)	85.7 (7)
3	100 (10)	25.0 (4)
4	50.0 (2)	100 (1)
5	100 (4)	50.0 (2)
6	100 (4)	100 (1)
7	0 (1)	–
> 7	60.0 (37)	25.0 (16)

Sensitivity was calculated as TP/(TP + FN)(see Table 1) at each day interval and are reported as percent with sample size in parentheses.

Day 0 represents nasal swab screening the same day as clinical culture results.

The number of days separating the nasal swab from the culture also was noted. We evaluated the risk factors for MRSA and whether this affected the concordance rate with nasal swabs.

Statistical analysis was carried out with Analyse-it software (Leeds, United Kingdom). Categorical data were evaluated with chi-square tests, and continuous data were analyzed with t-tests.

Results

There were 80 (57.7%) male and 61 female patients with a mean age of 59.7 years. The mean hospital length of stay was 19.3 days, with an ICU length of stay of 11.75 days. Patients spent an average of 8.63 days on a ventilator. The hospital mortality rate was 20.1%.

The patients were divided evenly between the hospitals. One-half the patients (50.3%) thus were screened with PCR, whereas the remaining patients were screened with chromogenic medium. Patients also were divided between different ICUs, with 41% being admitted to the medical ICU, 47.8% to a surgical or trauma ICU, and 11.2% to both types of ICUs during their stays.

There were 167 clinical cultures performed. Thus, there were more cultures than patients, as some patients had a positive MRSA culture at more than one site or more than one time. The majority of the cultures (70%) were performed on sputum specimens in an effort to diagnose pneumonia. The remaining cultures were performed on blood (10.1%), wounds (21.5%), and urine (3.4%). Nasal swab screening was accomplished in 70% of the patients at UMMC and 90% of the patients at North Memorial. Screening was positive in 12% of the patients at the former and 6% of the patients at the latter.

We evaluated the sensitivy of nasal swab screening to stratify the risk of finding MRSA in clinical cultures. The overall sensitivity of the swab results was 69.5%. The sensitivity was significantly higher for screening performed within six days of clinical cultures compared with screening performed seven days or more before cultures were obtained (79% vs. 46%; p < 0.0001) (Table 2). Sensitivity also differed significantly according to the surveillance method (Table 3). The sensitivity in patients screened with PCR within six days of developing an infection was significantly higher than that for patients screened with chromogenic medium (90% vs. 69%; p = 0.006). The superiority of PCR also was noted among patients who had a culture performed at least one week after the screening was conducted, although the sensitivity of each test was low (58.6% vs. 27.8%; p = 0.006).

Table 3.

Sensitivity of Nasal Swab Screening Completed within 6 Days of Clinical Cultures ^a

Site	Polymerase chain reaction	Chromogenic media	P value
Overall	90 (58)	69 (61)	0.006
Sputum	93 (41)	68 (41)	0.005
Wound	70 (10)	62 (13)	0.67
Blood	100 (4)	83 (6)	0.38
Urine	100 (3)	100 (1)	1

Data are presented as percentage with sample size in parenthesis.

In addition, we examined how recent hospitalization, a risk factor for MRSA colonization, affected sensitivity (Table 4). We defined recent hospitalization as within the previous three months or for more than three days prior to cultures being obtained. There was no difference in overall sensitivity in patients with the risk factor of recent hospitalization compared with patients without this risk factor (72% vs. 69%; p = 0.72). However, of those with a risk factor for colonization, the sensitivity was significantly higher for patients screened with PCR than for patients in whom chromogenic medium was used (75.6% vs. 57.1%; p = 0.02). A similar trend was observed in patients without the risk factor of recent hospitalization, as PCR sensitivity was higher than that of chromogenic medium (93% vs. 58%; p = 0.02).

Table 4.

Sensitivity of Nasal Swabs to Clinical Culture Results in Different Clinical Scenarios

	PCR (%)	Chromogenic medium (%)	P value
Culture < 6 days from time of screen	90	69	0.006
Culture ≥ 7 days from the time of screen	59	27.7	0.006
No healthcare exposure	93	58	0.02
Healthcare exposure	76	57	0.02
No healthcare exposure and culture < 6 days from time of screen	100	61	0.016
Healthcare exposure and culture < 6 days from time of screen	87	72	0.07

Discussion

The incidence of multi-drug-resistant organisms continues to increase in ICUs worldwide, which has complicated the treatment of infectious diseases. Specifically, the incidence of MRSA among critically ill patients has increased significantly over the past three decades. This crisis in drug-resistant organisms mandates new methods of evaluating and treating these infections. Our results demonstrate that commonly collected epidemiologic information can be used in a clinical fashion to tailor empiric antibiotic therapy.

Physicians who treat infections in patients in ICUs must deal with competing principles. Overutilization of antibiotics clearly has led to persistent and progressive resistance of bacteria to commonly used regimens. Underutilization of antibiotics, however, can lead to significant complications in individual patients. The importance of rapid and broad empiric antimicrobial management among patients with severe infections has been well documented [8]. As a result, published guidelines have recommended providing empiric broad antimicrobial coverage to patients with suspected severe infections and subsequently tailoring the coverage on the basis of culture results. Our study was designed to determine if the results of nasal swab studies could provide additional information that could direct empiric antibiotic use among critically ill patients.

In order to combat the spread of MRSA, many hospitals have instituted active surveillance [2]. In these programs, nasal swab specimens are obtained from patients on admission to the ICU. Patients found to be colonized with MRSA are placed in isolation precautions to limit the spread of the organism to other patients. A few hospitals also have proceeded with attempts at pharmacologic decolonization [9 –11]. Multiple studies have demonstrated that patients colonized with MRSA have more subsequent complications than do patients colonized with MSSA. The odds ratio of infection among patients colonized with MRSA has been reported to be > 4 [12]. Although infection risks and epidemiologic information have been well documented in previous studies, the ability of a nasal swab to predict subsequent clinical cultures has not been evaluated. Our data demonstrate that in certain clinical scenarios, negative swab results are highly associated with the absence of MRSA in clinical cultures.

The overall sensitivity of nasal swab screening was only 70%. In other words, patients with a clinical culture of MRSA had a negative nasal swab (false-negative result) 30% of the time. These results are altered significantly, however, according to the type of test used to screen the patients. There are three primary methods utilized to screen for MRSA colonization: Formal culture, chromogenic medium culture, and PCR [13]. Chromogenic medium culture involves incubating a nasal swab specimen in specialized agar, which changes color in the presence of MRSA. Polymerase chain reaction is an automated test that evaluates genes associated with MRSA. In our study, screening with PCR had a markedly greater sensitivity than screening with chromogenic medium. As a result, clinicians who practice in a facility that uses chromogenic medium as a screen for MRSA colonization should not use the information clinically. In contrast, the strong performance of PCR suggests that clinicians can consider these results when determining an empiric antibiotic regimen. Patients who subsequently developed a positive clinical culture for MRSA rarely had a PCR-negative nasal swab.

We looked at patient characteristics to determine if the sensitivity changed with the clinical situation. The sensitivity of nasal swab screening was significantly higher among patients who had a positive clinical culture within six days of being screened with a nasal swab. This suggests that patients who developed an MRSA infection more than one week after admission to the ICU may have acquired the organism during the admission. In such circumstances, admission surveillance cannot predict subsequent culture results and should not be used in clinical decision making. In view of our results, surveillance cultures need to be performed at least weekly.

We also evaluated patients according to their exposure to the healthcare system. The sensitivity of nasal swab screening was similar among patients who had recent exposure compared with patients who had no recent exposure. This finding suggests that the clinical utility of nasal swabs is not altered by the length of stay in the hospital prior to the test and that it is equally useful in chronically ill patients, who may have frequent encounters with the healthcare system. Regardless of the subpopulations we analyzed, chromogenic medium testing was inferior to PCR analysis.

An advantage of using PCR screening information in clinical decision making is the rapid availability of the results (within 2 h). Additionally, nasal swab screens often are performed prior to the onset of clinical infection. If there is suspicion of infection within six days of performing a nasal swab PCR, empiric antibiotics could be chosen on the basis of the screening results. Chromogenic medium cultures also are rapid, with results generally available within 24 h. Nonetheless, the poor sensitivity of this test prevents it from have more than epidemiologic utility.

Previous studies evaluating infection risks subsequent to nasal colonization have utilized a different sequence of analysis [12,14 –16]. These studies have evaluated patients with a positive nasal swab and reported their risk of subsequent infection. Such studies have shown elegantly that infection risks are higher among patients with MRSA colonization. Our study goal differed from that in these reports in that we specifically evaluated a group of patients having clinical cultures for MRSA. We then evaluated the sensitivity of a screening tool to identify the highest risk of MRSA in clinical cultures.

Previous studies also evaluated the utility of surveillance cultures in developing empiric antibiotic regimens. Such cultures have been studied primarily in ventilator-associated pneumonia (VAP). Papadomichelakis et al. found that in 82% of VAP episodes, the pathogens implicated were predicted accurately by surveillance cultures from the respiratory or gastrointestinal tract [17]. In a study by Robicsek et al., 68% of patients with culture-proved MRSA disease were simultaneously colonized with MRSA in the nares [14]. The advantage of our study is that we evaluated a screening tool available in many hospitals. Active surveillance culture programs require an additional level of laboratory support, and the results often are not available for three to five days. Our approach of using PCR as a screening tool to select empiric antimicrobial coverage can be accomplished more rapidly and can utilize existing hospital programs rather than demanding an additional screening tool.

There are limitations of the study worth noting. First, it was a retrospective analysis rather than a prospectively conducted protocol. Nonetheless, such retrospective data are needed to formulate evidence-based clinical pathways. Second, although the sensivity of nasal swab PCR was high, it was not 100%. There were a few patients with a negative nasal swab and a culture positive for MRSA. As such, each clinician must determine the severity of the infection, as well as the importance of universal appropriate initial treatment, when selecting empiric antimicrobial therapy. Third, although the policy at each hospital was to obtain nasal swab screening on admission to the ICU, this was not accomplished universally. As a result, there could have been a selection bias of patients who received screening. Fourth, we did not look at patients with MSSA infections. As a result, we cannot comment on this population. We chose to omit them, as our focus was to determine if certain populations could forego empiric MRSA coverage on the basis of nasal swab results.

Another potential criticism of this study is our use of sensitivity to evaluate the efficacy of nasal swabs as a screening modality. Other options are reporting the positive and negative predictive values, the accuracy, the odds ratio, or the specificity. We believe that sensitivity was the optimum metric for this study, and we chose it for several reasons. Nasal swab testing is essentially a screening tool, and sensitivity is the appropriate measure when the primary event of interest is the likelihood of a false-negative result. A highly sensitive test accomplishes this goal, as such a test has a low rate of false-negative results. A large negative predictive value has the benefit of minimizing false-negative tests because NPP = TN/(TN+FN). The disadvantage is that any given screening test will have drastically different positive and negative predictive values with various degrees of prevalence. In our study, we could have altered the predictive values just by altering the prevalence of MSSA patients in our cohort. Additionally, prevalence data apply only to individual ICUs, so predictive values would not be generalizable to any other ICU. For these reasons, we did not include information regarding positive or negative predictive values. The same line of analysis prevents us from reporting odds ratios. We do not have data on clinical cultures that were MRSA-negative, so we cannot report the odds ratio or specificity. However, we believe that reporting the sensitivity of the screening test provides the information necessary to interpret our results.

In conclusion, PCR analysis of nasal swab specimens is a highly sensitive screening test for MRSA. Clinicians may be able to use these results to tailor more appropriate empiric antimicrobial regimens. The results with chromogenic medium screening are markedly less reliable, which suggests that clinicians should view these results with caution. Prospective studies with a clearly defined empiric antibiotic pathway that considers the results of the PCR screen are now warranted. The results of PCR screening appear to be most valid early in the ICU course. As such, a potential protocol would be to obtain PCR screening at the time of all clinical cultures and tailor empiric antibiotics on the basis of these results.

Footnotes

Author Disclosure Statement

No conflicting financial interests exist.

Presented at the Thirtieth Annual Meeting of the Surgical Infection Society, Las Vegas, Nevada, April 18–20, 2010.

References

Wenzel

. The antibiotic pipeline—Challenges, costs, and values. N Engl J Med, 2004; 351:523–526.

Warren

, Guth

, Coopersmith

et al. Epidemiology of methicillin-resistant Staphylococcus aureus colonization in a surgical intensive care unit. Infect Control Hosp Epidemiol, 2006; 27:1032–1040.

Lee

, Park

, Oh

et al. Comparison of protocols for surveillance of methicillin-resistant Staphylococcus aureus (MRSA): Medical staff vs ICU patients. Ann Clin Lab Sci, 2007; 37:248–250.

Troillet

, Carmeli

, Samore

et al. Carriage of methicillin-resistant Staphylococcus aureus at hospital admission. Infect Control Hosp Epidemiol, 1998; 19:181–185.

Lucet

, Chevret

, Durand-Zaleski

et al. Prevalence and risk factors for carriage of methicillin-resistant Staphylococcus aureus at admission to the intensive care unit: Results of a multicenter study. Arch Intern Med, 2003; 163:181–188.

Warren

, Guth

, Coopersmith

et al. Impact of a methicillin-resistant Staphylococcus aureus active surveillance program on contact precaution utilization in a surgical intensive care unit. Crit Care Med, 2007; 35:430–434.

Ridenour

, Wong

, Call

et al. Duration of colonization with methicillin-resistant Staphylococcus aureus among patients in the intensive care unit: Implications for intervention. Infect Control Hosp Epidemiol, 2006; 27:271–278.

Kollef

, Sherman

, Ward

et al. Inadequate antimicrobial treatment of infections: A risk factor for hospital mortality among critically ill patients. Chest, 1999; 115:462–474.

Ammerlaan

, Kluytmans

, Wertheim

et al. Eradication of methicillin-resistant Staphylococcus aureus carriage: A systematic review. Clin Infect Dis, 2009; 48:922–930.

10.

McConeghy

, Mikolich

, LaPlante

. Agents for the decolonization of methicillin-resistant Staphylococcus aureus. Pharmacotherapy, 2009; 29:263–280.

11.

Coates

, Bax

, Coates

. Nasal decolonization of Staphylococcus aureus with mupirocin: Strengths, weaknesses and future prospects. J Antimicrob Chemother, 2009; 64:9–15.

12.

Safdar

, Bradley

. The risk of infection after nasal colonization with Staphylococcus aureus. Am J Med, 2008; 121:310–315.

13.

Paule

, Mehta

, Hacek

et al. Chromogenic media vs real-time PCR for nasal surveillance of methicillin-resistant Staphylococcus aureus: Impact on detection of MRSA-positive persons. Am J Clin Pathol, 2009; 131:532–539.

14.

Robicsek

, Suseno

, Beaumont

et al. Prediction of methicillin-resistant Staphylococcus aureus involvement in disease sites by concomitant nasal sampling. J Clin Microbiol, 2008; 46:588–592.

15.

Davis

, Stewart

, Crouch

et al. Methicillin-resistant Staphylococcus aureus (MRSA) nares colonization at hospital admission and its effect on subsequent MRSA infection. Clin Infect Dis, 2004; 39:776–782.

16.

Garrouste-Orgeas

, Timsit

, Kallel

et al. Colonization with methicillin-resistant Staphylococcus aureus in ICU patients: Morbidity, mortality, and glycopeptide use. Infect Control Hosp Epidemiol, 2001; 22:687–692.

17.

Papadomichelakis

, Kontopidou

, Antoniadou

et al. Screening for resistant gram-negative microorganisms to guide empiric therapy of subsequent infection. Intensive Care Med, 2008; 34:2169–2175.