Introduction

There is growing concern that vancomycin may provide suboptimal therapy for severe methicillin-resistant Staphylococcus aureus (MRSA) infections [1, 2]. Although there are conflicting results [3], the recent minimum inhibitory concentration (MIC) creep in vancomycin has prompted guidelines to recommend a vancomycin target trough of 15–20 μg/ml [4]. A reduction in the efficacy of vancomycin against MRSA strains with a high MIC (1–2 μg/ml) has been described in observational studies [5, 6], suggesting that subtle changes in the MIC may explain the clinical failure. Soriano et al. [7] described that independent predictors of mortality in multivariate analysis in patients with MRSA bacteremia included the receipt of empirical vancomycin and having an isolate with a vancomycin MIC of 2 μg/ml (odds ratio [OR] 6.39). The purpose of the present study was to evaluate further the clinical characteristics of the 2 μg/ml strains in patients with bacteremia, compared with the characteristics in those with other infection sites.

Methods

Blood culture was performed using BacT-ALERT 3D (Sysmex-biomerieux, Tokyo, Japan). Isolates were identified and susceptibility determined according to standard techniques, using the Auto-Scan W/A and Comb Panel (Siemens). Methicillin susceptibility was determined by microdilution using Clinical Laboratory Standards Institute methods [8]. One hundred and twenty-eight episodes of MRSA bacteremia were recorded between January 2005 and June 2008 at a single center, and the isolated strains were stored at −20°C. To identify MRSA strains of vancomycin MIC 2 μg/ml, a standard E test was performed according to the instructions in the manufacturer’s manual (Sysmex-biomerieux). Briefly, several colonies were suspended in Mueller-Hinton broth (Nissui Pharmaceutical, Tokyo, Japan) to obtain an inoculum equivalent to a 0.5 McFarland standard. One hundred microliters of inoculum was evenly streaked onto a 90-mm BHIA (BBL; Becton Dickinson, Mountain View, MD, USA) plate and allowed to dry. Vancomycin E test strips (Sysmex-biomerieux) were applied to the surface of the agar, and the plates were incubated at 35°C in air for 24 h. Zones were read at complete inhibition, with care, to visualize hazy growth and microcolonies [9].

To compare with the clinical features of MRSA vancomycin MIC 2 μg/ml strains isolated from bacteremia, we evaluated 631 MRSA strains isolated from infection sites other than blood in the same period. These MRSA strains included 227 strains from respiratory tract infections, 205 strains from skin and soft tissue infections (incisional surgical site infections; 148 strains), 92 strains from intraabdominal infections, 58 strains from urinary tract infections, and 49 strains from other infections.

The following data were obtained as risk factors for mortality in all patients: age, sex, preexisting comorbidities, need for mechanical ventilation, need for hemodialysis, intensive care unit (ICU) stay, receipt of prior glycopeptide, receipt of prior anti-pseudomonas antibiotics, implant infection, current administration of steroids for >30 days, current administration of immunosuppressants, current administration of anticancer drugs, source of bacteremia, history of hospitalization within 6 months, presence of shock when blood samples were obtained for culture, use of linezolid (1st or 2nd choice), initial appropriate therapy (anti-MRSA drug use) within 24 h, and MRSA bacteremia with strains of vancomycin MIC 2 μg/ml. Comorbidity was defined as a disease or therapy that could predispose patients to infection, such as the following; diabetes, liver cirrhosis, renal failure, alcoholism, active neoplastic disease, receipt of a solid organ or bone marrow transplant, neutropenia, chronic obstructive pulmonary disease, or severe cardiac disease with systematic heart failure.

Empirical treatment was defined as an antibiotic administered before the definitive microbiological testing result was obtained, and it was considered to be appropriate when the isolated MRSA strain was sensitive. The mortality and clinical efficacy of vancomycin or linezolid for MRSA bacteremia caused by the strain of vancomycin MIC 2 μg/ml were compared with those for MRSA bacteremia caused by strains of vancomycin MIC ≤1 μg/ml. Mortality was defined as death from any cause at 28 days. The dosage of vancomycin was adjusted to obtain a trough serum concentration of 15–20 μg/ml after 3 days of administration. The study concerning MRSA bacteremia was approved by the ethics committee at Hyogo College of Medicine. Informed consent was not required because all procedures were routine.

Statistical analysis was performed as follows; categorical variables were compared by the χ2 test with Yates’ correction or Fisher’s exact test when necessary, by using Microsoft Excel 2003. The variables selected by univariate analysis (p < 0.2) were subjected to multivariate analysis using a stepwise logistic regression model. SPSS version 16.0 (Tokyo, Japan) was used to perform these analyses. The level of statistical significance was set at p < 0.05.

Results

The MIC of vancomycin was 2.0 μg/ml for 41 strains (32.0%), 1 μg/ml for 85 strains (66.4%), and 0.5 μg/ml for 2 strains (1.6%) in MRSA isolated from blood cultures; however, in MRSA isolated from other materials, 2 μg/ml MIC strains accounted for only 8.9%, and there was a significant difference (p < 0.001, Table 1). The geometric mean MIC for vancomycin in MRSA isolated from blood culture was higher than that in MRSA from other materials (blood 1.235 vs. other materials 0.952, 1.3-fold).

Table 1 Distribution of vancomycin minimum inhibitory concentration (MIC) in methicillin-resistant Staphylococcus aureus (MRSA) isolated from blood samples and other materials
Full size table

Table 2 shows the primary diseases in patients with episodes of MRSA bacteremia according to the vancomycin MIC. There was no significant difference between groups. As the source of bacteremia, the incidence of pneumonia was significantly higher in patients with strains of 2 μg/ml MIC than in those with strains of ≤1 μg/ml MIC (Table 3). In patients with MRSA infections other than bacteremia, the percentages of each infection site were similar in patients infected with 2 μg/ml strains and in those infected with ≤1 μg/ml strains (Table 4).

Table 2 Primary diseases in patients with episodes of MRSA bacteremia
Full size table
Table 3 Sources of infections in patients with episodes of MRSA bacteremia
Full size table
Table 4 Infections in patients with episodes of MRSA other than bacteremia: ≤1 μg/ml MIC strains versus 2 μg/ml MIC strains
Full size table

In the patients with MRSA bacteremia, and in the patients with infections other than bacteremia, prior vancomycin use did not correlate with the isolation of 2 μg/ml strains in either of the groups (Tables 5, 6). Antibiotics selected as 1st line therapy for patients with MRSA bacteremia were comparable in patients with 2 μg/ml MIC strains and those with ≤1 μg/ml MIC strains (Table 7). In bacteremia patients administered glycopeptides as 1st line therapy, the clinical efficacy for strains with a vancomycin MIC of 2 μg/ml was significantly lower than that for strains with a vancomycin MIC of ≤1 μg/ml (30.0 vs. 78.8%, p < 0.001) (Table 8). In the analysis of the efficacy of glycopeptides as 1st line therapy for infections other than bacteremia, there was no significant difference between infections with 2 and ≤1 μg/ml MIC strains (p = 0.109) (Table 9).

Table 5 Influence of prior antibiotic use on the isolation of ≤1 μg/ml MIC strains and 2 μg/ml MIC strains in patients with bacteremia
Full size table
Table 6 Influence of prior antibiotic use on the isolation of ≤1 μg/ml MIC strains and 2 μg/ml MIC strains in patients with infections other than bacteremia
Full size table
Table 7 Antibiotics selected as 1st line therapy for patients with MRSA bacteremia
Full size table
Table 8 Clinical efficacy of antibiotic therapy in patients with episodes of MRSA bacteremia
Full size table
Table 9 Clinical efficacy of antibiotic therapy in patients with MRSA infections other than bacteremia
Full size table

In the analysis of each infection site, the efficacy of glycopeptides as 1st line therapy was 71.0% in infections with MIC ≤1 μg/ml strains and 53.3% in infections with MIC 2 μg/ml strains in the pulmonary tract (p = 0.253), 91.2 and 83.3% in skin and soft tissue, 80.0 and 71.0% in the intraabdominal region, 81.1 and 100% in the urinary tract, and 63.3 and 50.0% in other areas. Infections other than pulmonary tract infections were not submitted to statistical analysis because of the low numbers in the data. The efficacy of linezolid was comparable in infections other than bacteremia (80.1 and 73.7%, respectively, Table 9).

Mortality was documented in 65.8 and 19.5% of episodes due to strains with vancomycin MICs of 2 and ≤1 μg/ml, respectively, in patients with bacteremia (p < 0.001). In patients with MRSA infections other than bacteremia, mortality was 45/575 (7.8%) in those with ≤1 μg/ml MIC strains and 6/56 (10.7%) in those with 2 μg/ml MIC strains (p = 0.617).

To elucidate the role of vancomycin MIC in mortality, univariate and multivariate analyses were performed, using death at 28 days as a dependent variable. In univariate analysis, episodes involving strains with a 2 μg/ml MIC, liver cirrhosis, ICU stay, and pneumonia as a source were associated with mortality. Independent predictors associated with mortality in the multivariate analysis included having an isolate with a vancomycin MIC of 2 μg/ml, liver cirrhosis, and ICU stay, and initial appropriate therapy lowered the risk (odds ratio [OR] 0.22; 95% confidence interval [CI] 0.06–0.91) (Table 10).

Table 10 Factors independently associated with mortality of patients with episodes of MRSA bacteremia; logistic regression model
Full size table

Discussion

The incidence of strains with vancomycin MIC 2 μg/ml has varied considerably among several studies [3, 10–12]. As reasons for this inconsistency, an increase of these strains could be caused by either the extensive use of vancomycin or the dissemination of clones with less sensitivity to vancomycin. Moise et al. [13] demonstrated that MRSA bloodstream isolates from patients recently treated with vancomycin showed reduced susceptibility and increased tolerance to vancomycin in vitro. In the present study, however, vancomycin use did not affect the isolation of 2 μg/ml strains. It is important to clearly distinguish MIC creep from the increased occurrence of specific epidemic clones with elevated vancomycin MIC values. From another point of view, the proportion of 2 μg/ml strains among strains isolated from bacteremia in the present study was significantly higher than that in strains isolated from other infections (32.0 vs. 8.9%, p < 0.001). As an explanation, refractory infections caused by 2 μg/ml strains might subsequently develop systemic infections, such as bacteremia. Interestingly, the incidence of pneumonia as a source of infection in bacteremia due to 2 μg/ml strains was significantly higher than that in bacteremia due to ≤1 μg/ml strains. Only 20–30% of the vancomycin serum concentration is achieved in lung tissue, thus requiring even higher doses to treat pneumonia.

As reported by other investigators [2, 5, 7], the present study has suggested that the vancomycin MIC has an important impact on antibiotic efficacy and on mortality in patients with bacteremia due to MRSA. In patients administered glycopeptides, there was a significant difference in clinical efficacy between infections caused by MIC ≤1 μg/ml strains and those caused by MIC 2 μg/ml strains. In the analysis of infection sites other than blood, however, there was no significant difference in clinical efficacy. Because of the low number of 2 μg/ml strains in infections other than bacteremia, insignificant difference of the efficacy of glycopeptides between pulmonary infections caused by MIC ≤1 μg/ml strains and MIC 2 μg/ml strains may not be worth evaluating. Mortality was significantly higher in patients with MRSA bacteremia infected with 2 μg/ml strains than in those infected with ≤1 μg/ml strains; however, there was no significant difference in mortality from infections other than bacteremia. Factors independently associated with mortality in patients with bacteremia in a regression model included bacteremia with MRSA strains with a vancomycin MIC of 2 μg/ml, liver cirrhosis, and ICU stay, and initial appropriate therapy lowered the risk.

Considering the current problems with vancomycin treatment, it is necessary to clarify whether new anti-staphylococcal agents, such as linezolid, daptomycin, or tigecycline, are superior to vancomycin when the strain has a vancomycin MIC >1 μg/ml. Thickening of the cell wall observed in vancomycin-intermediate S. aureus prevents the passage of vancomycin through a thickened peptidoglycan layer [14]. Camargo et al. [15] observed a strong correlation between the presence of a thickened cell wall and daptomycin resistance. In contrast, Mitsutaka et al. [16] demonstrated that the linezolid MIC showed no correlation with VCM resistance and they suspected that linezolid would be effective to treat infections that are poorly responsive to vancomycin. In our study, the clinical efficacy of linezolid use did not reveal a significant difference according to the MIC value in patients with bacteremia and other infections. As linezolid was mainly used as 2nd line therapy in bacteremia in our study, further study is necessary to confirm its efficacy in the treatment of MRSA infections with a vancomycin MIC of ≥2 μg/ml.

In conclusion, several characteristics such as a higher incidence than at other infection sites, a high incidence of pneumonia as a source of infection, a low success rate of vancomycin therapy, and poor prognosis were confirmed in 2 μg/ml MIC MRSA isolated from bacteremia; however, such features were not apparent in infections other than bacteremia.