|Home | About | Journals | Submit | Contact Us | Français|
Vancomycin MIC creep has been reported by some institutions but not confirmed in large surveillance studies. We evaluated the possible occurrence of MIC creep when testing vancomycin and daptomycin against methicillin (oxacillin)-resistant Staphylococcus aureus (MRSA) by using precise incremental reference MIC methods. Nine hospitals (one in each U.S. census region) randomly selected bloodstream MRSA strains (target, 40/year) from 2002 to 2006. MICs were determined by the reference broth microdilution method using incremental dilutions (eight for each log2 dilution step). Isolates for which vancomycin MICs were >1 μg/ml were typed by pulsed-field gel electrophoresis (PFGE). The vancomycin MIC mode was either 0.625 μg/ml (for eight hospitals) or 0.813 μg/ml (for one hospital), and vancomycin MIC results for 72.9% of strains were between 0.563 and 0.688 μg/ml. No yearly variation in the central tendency of vancomycin MICs for the wild-type population in any medical center was observed; however, when data were analyzed by the geometric mean statistic, vancomycin MIC increases (at three sites) and declines (at three sites) were observed. The daptomycin MIC mode varied from 0.156 μg/ml (2003 to 2005) to 0.219 μg/ml (2002 and 2006), and MIC results for 83.5% (80.3 to 89.2% in each of the centers) of isolates fell between these values. Among PFGE-typed strains, 43 of 55 (78%; from seven hospitals) showed a pattern consistent with that of the USA100 clone, which was represented by all strains from two hospitals and 64 to 88% of strains from five other medical centers; only one strain (2%) was USA300. In conclusion, the perception of MIC creep may vary according to the methods used to analyze the data. Geometric mean MIC data revealed a possible, very-low-level MIC creep at three of nine sites over the 5-year period, which was not evident using modal MICs or the data from all nine hospitals (+0.02 μg/ml). The occurrence of isolates for which the vancomycin MIC was >1 μg/ml was very unusual, with no increased trend, but these organisms were usually clonal (USA100).
Studies reporting vancomycin MIC creep among methicillin (oxacillin)-resistant Staphylococcus aureus (MRSA) strains have produced conflicting results. While some institutions have reported MIC increases, large multicenter surveillance studies have not demonstrated significant trends toward higher vancomycin MIC results (1, 15, 17). Several factors may be responsible for these discrepancies, including the sensitivity of the antimicrobial susceptibility testing method used for detecting MIC variations and the statistical analysis applied to the MIC data.
The word “creep” can be defined as a “gradual and unnoticed movement or shift.” Thus, vancomycin MIC creep should be recognized as a gradual increase in the central tendency of the vancomycin MIC for the dominant wild-type population, and it is said to be caused by long-term and/or extensive use of vancomycin. Since susceptibility testing methods usually evaluate the MIC on a base 2 logarithmic scale (by using doubling dilutions), this gradual shift may remain unnoticed for some time until the mode shifts to the next tested dilution. Some studies have applied agar stable-gradient tests, such as the Etest (AB Biodisk, Solna, Sweden), since this method includes an intermediate concentration between doubling dilutions and, consequently, may be more sensitive for the detection of MIC creep (6, 12).
The published studies of vancomycin MIC creep have analyzed the data by determining the proportion of strains for which vancomycin MICs are above a certain value (usually 1 or 1.5 μg/ml) and/or comparing the vancomycin MIC geometric means over a defined period of time (1, 12, 15, 17). However, an increase in the frequency of isolates for which vancomycin MICs are elevated does not necessarily indicate an increase in the central tendency of the vancomycin MIC for the wild-type population. Furthermore, an increase in the proportion of strains for which the vancomycin MIC is >1 or >1.5 μg/ml may be caused by either extensive use of vancomycin or, more likely when dealing with S. aureus, the dissemination of a clone or clones with less susceptibility to vancomycin (4, 9, 14). Thus, it is important to clearly distinguish MIC creep from an increased occurrence of specified epidemic clones for which vancomycin MICs are elevated.
In this work, we systematically evaluated the possible MIC creep when testing vancomycin and daptomycin against MRSA in a large, comprehensive, multicenter study (including nine geographically diverse hospitals) by using precise reference broth microdilution methods and testing discrete/incremental MIC dilution steps.
(This study was presented in part at the 48th Annual Interscience Conference on Antimicrobial Agents and Chemotherapy [ICAAC]-Infectious Diseases Society of America [IDSA] 46th Annual Meeting, Washington, DC, 2008.)
Nine hospitals located in generally large urban areas (one from each of the U.S. census regions) and having established high vancomycin use were selected to participate in the study. Each medical center contributed 200 MRSA strains collected consecutively from patients with bloodstream infections (1 strain per patient) from 2002 through 2006 (target, 40 randomly selected isolates per year per center). The participant centers were New England Medical Center, Boston, MA; New York Hospital Queens, New York, NY; Ochsner Clinic Foundation, New Orleans, LA; the University of Colorado Denver, Aurora, CO; the University of Nebraska Medical Center, Omaha, NE; the University of Washington, Seattle, WA; the University of Alabama at Birmingham, Birmingham, AL; Wayne State University, Detroit, MI; and the Medical University of South Carolina, Charleston, SC.
MICs of vancomycin and daptomycin were determined by the reference broth microdilution method (using frozen-form panels), with appropriate medium variations (50 mg/liter of calcium) for the testing of daptomycin (2, 3). Precise incremental dilutions were tested as follows: (i) for vancomycin, 36 dilution steps ranging from 64 to 0.06 μg/ml (64, 32, 16, 8, 4, 3, 2.5, 2, 1.875, 1.75, 1.625, 1.5, 1.375, 1.25, 1.125, 1, 0.938, 0.875, 0.813, 0.75, 0.688, 0.625, 0.563, 0.5, 0.469, 0.438, 0.406, 0.375, 0.344, 0.313, 0.281, 0.25, 0.188, 0.12, 0.094, and 0.06 μg/ml), and (ii) for daptomycin, 20 dilution steps ranging from 16 to 0.06 μg/ml (16, 8, 4, 2, 1.5, 1, 0.875, 0.75, 0.625, 0.5, 0.438, 0.375, 0.313, 0.25, 0.219, 0.188, 0.156, 0.12, 0.094, and 0.06 μg/ml). The Clinical and Laboratory Standards Institute broth microdilution methods were used throughout, with all quality control results for recommended strains being within published guidelines (2, 3).
Isolates for which the vancomycin MIC was >1 μg/ml were epidemiologically typed by pulsed-field gel electrophoresis (PFGE). Gel pattern analysis was carried out using the GelCompar II software (Applied Math, Kortrijk, Belgium), and the PFGE patterns obtained during this study were compared to those of the USA clones, such as USA100, USA300-0114, USA700, and USA1100 (10, 16). Representatives of the USA clones were kindly provided by the Network on Antimicrobial Resistance in S. aureus (NARSA; www.narsa.net). Percent similarities were identified with a dendrogram derived by the unweighted-pair group method using arithmetic averages and based on Dice coefficients. Band position tolerance and optimization were set at 2.3% and 0.5%, respectively.
The overall vancomycin MIC mode was 0.625 μg/ml (Table (Table1).1). The MIC mode for eight of nine centers was 0.625 μg/ml, and 72.9% (65.0 to 86.5% in each of the eight centers) of the vancomycin MICs were between 0.563 and 0.688 μg/ml (only a 0.125-μg/ml range). For the remaining medical center (site 9), the vancomycin MIC mode was 0.813 μg/ml (Table (Table1).1). No yearly variation in the central tendency of the vancomycin MIC for the wild-type population in any medical center was observed. Even with the use of precise concentrations, the vancomycin MIC mode remained very stable during the study period. Analyses of standard vancomycin MIC doubling dilutions (i.e., values rounded to the next CLSI standard doubling dilution) did not reveal any changes at the studied medical centers over the tested interval (2002 to 2006). All modes and MIC90s were 1 μg/ml for the medical centers in each year.
Interestingly, when analyzed by the geometric mean statistic, vancomycin MICs showed increases in three medical centers (for example, site 3) but a decline in three other centers (for example, site 7) (Fig. (Fig.1).1). In the medical center with the most prominent MIC creep when yearly geometric mean results were analyzed, the geometric mean increased only 0.120 μg/ml (from 0.933 μg/ml in 2002 to 1.053 μg/ml in 2006) if the rounded log2 MIC was applied and only 0.076 μg/ml (from 0.631 μg/ml in 2002 to 0.707 μg/ml in 2006) if precise discrete MIC results were used. Furthermore, in one medical center (site 8), different trends in the geometric mean were obtained depending on if the rounded or incremental MICs were used in the structural analysis. A slight increase (0.052 μg/ml over 5 years) was observed if rounded values were used, while a very minor decrease (0.032 μg/ml over the same period) was noted if precise discrete MICs were applied (Fig. (Fig.11).
Daptomycin MIC modes varied from 0.156 μg/ml (2003 to 2005) to 0.219 μg/ml (2002 and 2006), and MIC results for 83.5% (80.3 to 89.2% in each of the centers) of MRSA isolates fell between these values (Table (Table2).2). No yearly tendency toward an increase or decrease in daptomycin MICs was observed when data from the medical centers were combined (Table (Table2)2) or analyzed separately (data not shown).
The overall frequency (for all years combined) of MRSA isolates for which the vancomycin MIC was >1 μg/ml varied from ≤1.5% in medical centers 4 (0.0%), 2 (1.0%), and 3 (1.5%) to the highest rate of 5.0% in medical centers 1 and 7 (Table (Table1).1). Only one isolate for which the vancomycin MIC result was >2 μg/ml (2.5 μg/ml, which rounds to 4 μg/ml) was identified during this investigation. When results from all medical centers were analyzed together, the frequency of isolates for which vancomycin MIC results were >1 μg/ml varied from 1.6% in 2003 to 3.8% in 2004 (3.1% overall; 55 strains) and no clear trend toward greater resistance during the study period was observed (Table (Table1).1). Although there was no yearly tendency toward increase or decrease in vancomycin potency at any medical center, some variation in some hospitals over the years was observed. The frequencies of isolates for which vancomycin MICs were >1 μg/ml were higher in three medical centers (4, 5, and 6) in the last 2 to 3 years of the study period (Table (Table1).1). On the other hand, the frequencies of isolates for which vancomycin MICs were elevated were lower in two other participant sites (1 and 7) in the last 2 years of the study, and no significant differences between the first and the last years in the remaining medical centers could be observed (Table (Table11).
Among 55 MRSA isolates for which vancomycin MICs were high (>1 μg/ml), 43 (78.2%) showed a PFGE pattern consistent with that of USA100, also called New York/Japan clone, and this pattern was designated A in the present study (Fig. (Fig.22 and Table Table3).3). The A pattern isolates were observed in seven of eight hospitals, and 24 subtypes were also identified (Table (Table3).3). Isolates with PFGE pattern A and its subtypes represented all strains from two hospitals and 63.6 to 87.5% of strains from the other five medical centers where this pattern was observed. The remaining 12 strains (from five medical centers) exhibited eight distinct PFGE patterns (with 1 to 2 strains in each group). Only one strain (2%) was related to the emerging community-acquired MRSA USA300 clone (Table (Table33).
Studies evaluating the occurrence of vancomycin MIC creep have shown conflicting results (1, 6, 7, 15, 17). While some studies report that MIC creep could not be detected (1, 6), others report the occurrence of MIC creep based on the finding of an increased frequency of isolates for which vancomycin MICs are elevated (15, 17). However, these studies fail to demonstrate increases in the central tendency of vancomycin MIC results for the S. aureus populations investigated. Furthermore, strains for which vancomycin MIC results were elevated were not typed to detect the possible influence of clonal dissemination.
Using discrete incremental dilutions, we could not detect substantial MIC increase between the years of 2002 and 2006 for either vancomycin or daptomycin when monitoring 1,800 bacteremic MRSA strains from nine different U.S. medical centers. The results of the present study also indicate that the perception of vancomycin MIC creep may vary according to the methods used to analyze the data. Geometric mean MIC data revealed that three of nine sites had a slight MIC creep, which was not evident using modal MIC data. Furthermore, the highest MIC increase was 0.12 μg/ml over the 5-year period, or 0.024 μg/ml per year. Since the modal MIC was 0.625 μg/ml, it would require 15 years to move the vancomycin MIC by one log2 dilution step.
One of the most important findings of this study was the clonality among isolates for which vancomycin MICs were elevated (>1 μg/ml). Although the prevalence of strains for which vancomycin MICs were >1 μg/ml was very low, with no apparent increase, the occurrence of MRSA strains for which vancomycin MIC results were high was clearly related to clonal occurrences varying over time. Almost 80% of strains showed a unique PFGE pattern (corresponding to USA100). The USA100 pattern is the major PFGE pattern found among MRSA isolates from health care-associated infections throughout the United States. Furthermore, USA100 has been related to many vancomycin-intermediate S. aureus isolates and some vancomycin-resistant S. aureus isolates reported in Michigan and Pennsylvania (10).
The principle concern about the occurrence of MIC creep is the gradual loss of vancomycin activity and, consequently, the compromise of clinical utility (5, 8, 11, 13). However, the occurrence of MIC creep appears to be limited, and when it occurs, the MIC increase is a fraction of a log2 dilution step and emerges very slowly (over more than a decade). On the other hand, false perceptions of vancomycin MIC creep (increase in the central tendency of MICs for the wild-type population) appear to be caused by escalating occurrences of strains, usually clonal, for which vancomycin MICs are elevated (>1 μg/ml). Rapid and extensive spread of these MRSA clones has been well documented in the medical literature (4, 9, 10, 14, 16), and the spread of MRSA clones for which vancomycin MICs are elevated within and between medical centers would have a much greater impact on vancomycin clinical utility than the occurrence of true MIC creep.
In summary, vancomycin MIC creep during the period from 2002 to 2006 in the nine hospitals evaluated could not be demonstrated. The results presented here also demonstrated that increases in the vancomycin geometric mean MIC or increases in the frequency of strains for which vancomycin MICs were elevated were most influenced by documented clonal dissemination of strains that were less susceptible to vancomycin. We consider the geometric mean MIC to be an inappropriate tool to evaluate the occurrence of so-called vancomycin MIC creep.
We acknowledge the National Institutes of Health/National Institute of Allergy and Infectious Diseases and the NARSA program for kindly providing isolates for this study.
This study was funded by a research grant from Cubist Pharmaceuticals, Inc.
Published ahead of print on 27 July 2009.