Systematic Review and Meta-Analysis of the Epidemiology of Vancomycin-Intermediate and Heterogeneous Vancomycin-Intermediate Staphylococcus aureus Isolates

Background Vancomycin-intermediate Staphylococcus aureus (VISA) and heterogeneous VISA (hVISA) are associated with vancomycin treatment failure, and are becoming an increasing public health problem. Therefore, we undertook this study of 91 published studies and made subgroup comparisons of hVISA/VISA incidence in different study years, locations, and types of clinical samples. We also analyzed the genetic backgrounds of these strains. Methods A systematic literature review of relevant articles published in PubMed and EMBASE from January 1997 to August 2014 was conducted. We selected and assessed journal articles reporting the prevalence rates of hVISA/VISA. Results The pooled prevalence of hVISA was 6.05% in 99,042 methicillin-resistant S. aureus (MRSA) strains and that of VISA was 3.01% in 68,792 MRSA strains. The prevalence of hVISA was 4.68% before 2006, 5.38% in 2006–2009, and 7.01% in 2010–2014. VISA prevalence was 2.05%, 2.63%, and 7.93%, respectively. In a subgroup analysis of different isolation locations, the prevalence of hVISA strains was 6.81% in Asia and 5.60% in Europe/America, and that of VISA was 3.42% and 2.75%, respectively. The frequencies of hVISA isolated from blood culture samples and from all clinical samples were 9.81% and 4.68%, respectively, and those of VISA were 2.00% and 3.07%, respectively. The most prevalent genotype was staphylococcal cassette chromosome mec (SCCmec) II, which accounted for 48.16% and 37.74% of hVISA and VISA, respectively. Sequence Type (ST) 239 was most prevalent. Conclusion The prevalence of hVISA/VISA has been increasing in recent years, but has been grossly underestimated. Its incidence is higher in Asia than in Europe/America. hVISA is isolated from blood culture samples more often than from other samples. These strains are highly prevalent in epidemic MRSA strains. This study clarifies the epidemiology of hVISA/VISA and indicates that the detection of these strains and the control of nosocomial infections must be strengthened.


Introduction
Staphylococcus aureus, one of the major nosocomial and community-acquired pathogens, causes a variety of clinical problems, including infections of the skin and soft tissues [1]. Since the 1960s, the prevalence of methicillin-resistant Staphylococcus aureus (MRSA) has increased at a dramatic rate [2,3], and is associated with higher rates of morbidity and mortality than methicillin-susceptible S. aureus (MSSA) [4].
Glycopeptides, such as vancomycin, are popular and effective antimicrobial drugs for treating MRSA. Unfortunately, vancomycin-intermediate S. aureus (VISA) and heterogeneous VISA (hVISA) have been reported since 1997. hVISA is an S. aureus isolate with a minimum inhibitory concentration (MIC) for vancomycin within the susceptible range when tested with routine methods, but in which a proportion of the cell population is within the vancomycinintermediate range [5]. hVISA/VISA infections are commonly associated with persistent infections, prolonged bacteremia, and/or prolonged hospitalization [6][7][8][9]. Today, there is growing concern that hVISA and VISA are becoming prevalent worldwide [10][11][12].
In recent years, there have been many reports from single medical centers or individual countries of the incidence of hVISA/VISA, but few systematic reviews or meta-analyses on their prevalence. The review by Liu et al. on the epidemiology of hVISA/VISA was published over 10 years ago [13]. Another meta-analysis, by Van Hal et al., selectively analyzed the clinical significance and outcomes of hVISA [9]. In this systematic review and meta-analysis, we pooled the published studies that have reported the prevalence of hVISA/VISA, and made subgroup comparisons of the incidence of hVISA/VISA in different years, locations, and types of clinical samples. We also analyzed the genetic backgrounds of these strains. The results of this study will help to clarify the epidemiology of hVISA/VISA and will advance the control and management of these drug-resistant isolates.

Search strategy
Two independent examiners (S.S.Z. and X.X.S.) performed a systematic literature review of potentially relevant studies pertaining to VISA and hVISA. The studies were identified in the PubMed and EMBASE databases from articles published between January 1997 and August 2014. The following terms and connectors were used in the search strategy: (1)

Selection of studies
Studies identified in the literature search were checked by title and abstract. The papers with relevant abstracts were examined in full. The criteria for the inclusion and exclusion of the studies were established by the investigators before the literature was reviewed. The inclusion criteria were as follows: (1) studies that were original articles, short communications, correspondence, or letters that provided sufficient original data about the prevalence of hVISA/ VISA; (2) studies in which all MRSA strains were randomly selected; (3) studies that used normative and publicly accepted detection methods for hVISA/VISA; and (4) studies that were published in English. The exclusion criteria were: (1) studies that contained duplicate data or were overlapping articles; (2) reviews and conference abstracts; and (3) articles that included fewer than 10 cases.

Data extraction
Two authors independently ascertained the characteristics of each study, including the first author's surname, year of publication, continent, country, study years, isolate source, detection method, hVISA frequency, VISA frequency, and genotypes. When there was disagreement, the relevant paper was reviewed and the differences were resolved by consensus.

Assessment of study quality
The studies were assessed for quality, and only high-quality studies were included in the analysis. The criteria for high-quality studies were (1) that they provided basic data that included the study period and area, total number of isolates tested, and number of hVISA/VISA isolated; and (2) that they used dilution methods or E-test to detect VISA, and population analysis profile-area under the curve (PAP-AUC), macromethod Etest (MET), or screening agar to detect hVISA. When two studies overlapped, the more recent and larger study was included in the analysis. If one article included more than one study period, it was divided into several independent studies.

Statistical analysis
Statistical analyses were performed with STATA version 12.0. The data were pooled using the fixed-effects model (FEM) [14] and the random-effects model (REM) [15]. Statistical heterogeneity was assessed using the Cochran Q and I 2 statistical methods [16]. P < 0.1 was considered statistically significant. For all analyses, the results of FEM are presented only when there was no heterogeneity between the studies. Otherwise, the results of REM are presented. Freeman-Tukey arcsine transformations were performed to stabilize the variances, and after the metaanalysis, we transformed the summary estimates and the confidence interval (CI) boundaries back to proportions using the sine function [17].

Prevalence of hVISA/VISA in different study periods
To analyze the trends in the changes in hVISA/VISA prevalence in recent years, we performed a subgroup analysis of the prevalence of these two types of strains according to the study year. Three periods (before 2006, 2006-2009, and 2010-2014) were designated. Some studies that did not conform to the periods (e.g., reported for 2003-2007) were not included in this analysis.
It can be seen from Table 2

Discussion
The infections caused by MRSA are problematic because they entail high mortality and only limited antimicrobial drugs are available for their treatment [109]. Vancomycin has generally been the first drug of choice for the treatment of MRSA infections [110]. However, studies have reported that the treatment failure rate for vancomycin is increasing.  . hVISA and VISA are thought to be among the primary causes of treatment failure. However, in 76 studies (including 99,042 strains) chosen for our analysis, the prevalence of hVISA was only 6.05%, and the prevalence of VISA was only 3.01% in 68,792 strains in 38 studies. Therefore, we speculate that the incidence of hVISA/VISA was underestimated, possibly because of the resistance mechanisms and biological characteristics of these strains. Unlike MRSA and vancomycinresistant S. aureus (VRSA), the genetic backgrounds associated with hVISA/VISA remain unclear, and a molecular biological method to detect these strains is not yet available. The growth rates of hVISA/VISA are also slow [113,114]; hence, conventional methods, such as the Kirby-Bauer and instrument-based methods, do not produce accurate results. The PAP-AUC method is considered the gold standard technique for detecting hVISA. However, this method is time-consuming, cumbersome, and unsuitable for clinical laboratories [69], so a significant number of strains may have been missed. Therefore, there is an urgent need for a convenient and effective method with which to detect these strains.
To analyze the trends in the prevalence of hVISA/VISA in recent years, we divided the study period into three periods: before 2006, 2006-2009, and 2010-2014. The first period used the initial resistance breakpoint (vancomycin MIC of 8-16 μg/ml) and the two later periods used the present resistance breakpoint (vancomycin MIC of 4-8 μg/ml). Our study suggests that the prevalence of hVISA/VISA has been increasing in recent years. We consider that the more frequent use of vancomycin for MRSA infections is responsible for this situation because the high prevalence of hVISA/VISA reflects the level of vancomycin use [115,116]. The inappropriate management of drug-resistant strains has accelerated the spread of hVISA/VISA, and the change in the vancomycin-resistance breakpoint has also contributed to the increase in the prevalence rate.
Since the first reports of hVISA/VISA, the occurrence rates of these strains have varied throughout the world: the incidence of hVISA was 6.81% in Asia and 5.60% in Europe/America, and that of VISA was 3.42% and 2.75%, respectively. Current evidence supports the proposition that hVISA/VISA is more endemic in Asian countries than in Europe/America. Several factors may account for this situation. First, most countries in Europe and America are developed, with high public hygiene standards and scrupulous antimicrobial treatments [69,101,102]. Second, the control of nosocomial infections is more successful in European and American countries [41,95]. Third, Asia is the most populous region of the world, which can create an environment amenable to microbial transmission. The pooled prevalence rate for hVISA in mainland China was 15.78% [58, 71,83,84], and in India, the pooled prevalence rates for of hVISA and VISA were 12.41% and 15.09%, respectively [40, 51, 52, 85,99]. Fourth, because far more MRSA infections occur in Asian countries [117], vancomycin has been used more frequently for their treatment. Therefore, it is not surprising that hVISA and VISA are more common in Asia than elsewhere.
Previous studies have demonstrated that hVISA/VISA are prevalent among bacteremic specimens, and that these strains can persist in the bloodstream for a long time [19]. Our analysis confirms that hVISA is more common in blood-borne MRSAs, consistent with previous opinion. However, the prevalence of VISA was not obviously higher among isolates from blood culture samples than other samples. The reason for this is unclear, but this result suggests that not only blood culture samples but all clinical samples should be given attention.
The mecA gene, which is located within the SCCmec element, is the specific genetic mechanism of methicillin resistance [118]. Many epidemiological studies have demonstrated that community-associated MRSA (CA-MRSA) can be distinguished from hospital-acquired MRSA (HA-MRSA) by the type of the SCCmec element present. The most common SCCmec types in CA-MRSA strains are SCCmec IV and V, whereas SCCmec I, II, and III predominate in HA-MRSA strains [119]. The results of our pooled analysis show that SCCmec II and III were the most prevalent molecular types among the VISA strains. Previous studies have demonstrated limited vancomycin-resistance potential in SCCmec IV MRSA clones [120]. However, we found that the prevalence of SCCmec IV in hVISA was similar to that of SCCmec III. This phenomenon suggests that hVISA is not limited to typical "hospital" clones of S. aureus.
MLST is a powerful and highly discriminatory method for analyzing the population structures and epidemiology of S. aureus [121]. Our study demonstrates that ST239 and ST5 are the most epidemic genotypes of hVISA/VISA. ST239 and ST5 are two international HA-MRSA lineages prevalent in Asia, South America, and Eastern Europe [122,123]. ST239 MRSA strains are typically resistant to many classes of antibiotics, including β-lactams, fluoroquinolones, aminoglycosides, and macrolide antibiotics. Our results strongly suggest that hVISA and VISA are highly prevalent among international epidemic MRSA strains. Moreover, the genetic backgrounds of these strains are complex, and many ST types are dispersed among hVISA/VISA isolates, including ST59, ST72, and ST900 [70,72,74].
The present study had several limitations. Genetic information was available in only 27% (25/91) of the studies we reviewed, which could have led to publication bias and influenced our results. There was also considerable heterogeneity between studies because they differed in various study variables, such as the patient populations examined, testing methodologies used, study durations, previous vancomycin therapy, and concomitant illnesses. These confounding factors could not be circumvented with subgroup analyses. As in previous meta-analyses in which unexplained heterogeneity was identified, we accommodated this condition by using REM, in which the effects underlying the results of different studies are assumed to be drawn from a normal distribution [124]. However, this heterogeneity could not be balanced out by REM alone, so that the stability of the final results must have been affected by the heterogeneity of the sample.
In summary, the results of our study suggest that the prevalence rates of hVISA/VISA have increased in recent years. Our data also supports the view that hVISA/VISA are more prevalent in Asian countries than in Europe/America. Our study confirms that hVISA strains are more common in blood-borne MRSA than in other MRSA. Finally, the most epidemic genotypes of hVISA/VISA are SCCmec II and SCCmec III on SCCmec typing, and ST239 and ST5 on MLST typing, which are predominant among the HA-MRSA strains. However, the incidence of hVISA/VISA is grossly underestimated. Therefore, the detection of hVISA/VISA must be strengthened, especially in samples from patients with bacteremic HA-MRSA infections, and the use of vancomycin and nosocomial infections must be urgently and strictly controlled, particularly in Asian hospitals.

dL A, H N, T JF, J-G ML, T G, B A, et al. Control And Outcome Of A Large Outbreak Of Colonization
And