Clinical Implications of Varying Degrees of Vancomycin Susceptibility in Methicillin-Resistant Staphylococcus aureus Bacteremia

Mitchell J. Schwaber, Sharon B. Wright, Yehuda Carmeli, Lata Venkataraman, Paola C. DeGirolami, Aneta Gramatikova, Trish M. Perl, George Sakoulas, Howard S. Gold


Emerging Infectious Diseases. 2003;9(6) 

In This Article


First reported in a clinical specimen from Japan in 1997,[2] VISA strains (MIC 8-16 mg/L)[17] have now been isolated in numerous countries around the world[3,24,25] and have been identified in several patients in the United States.[7,19,26] These patients have in common extensive prior exposure to vancomycin, and their clinical courses are notable for a suboptimal response to this agent.[1,7] Since VISA isolates are presumed to spread as do vancomycin-susceptible strains of S. aureus, their appearance has prompted the issuance of guidelines for their identification and control of transmission.[27,28]

Shortly after the report of the first VISA isolate was published, Hiramatsu et al. reported an S. aureus isolate exhibiting heteroresistance to vancomycin.[6] They found this phenotype of vancomycin heteroresistance to be widespread in Japanese university hospitals. In the ensuing 5 years, reports emerged of heteroresistant S. aureus in numerous countries.[6,7,8,9,10,11,12,13,14,15,16]

Several areas of uncertainty have marked the subject of vancomycin heteroresistance since it was first described. First, no standardized definition exists, and investigators have defined it by using a variety of criteria.[6,8,11,16,29,30] Second, the clinical significance of heteroresistance remains unclear. Although existing evidence supports the hypothesis that heteroresistant isolates have a greater likelihood of developing homogeneous intermediate resistance than do susceptible strains,[6] and other data suggest an association between isolation of VISA and an adverse outcome,[7] few studies have examined whether specific risk factors exist for infection with hetero-VISA, and whether such infection is associated with adverse outcomes. Those studies that have examined risk factors have had small sample sizes,[10] inadequate generalizability,[8] or a method of control selection that did not allow for direct comparison between patients with hetero-VISA and those infected with S. aureus isolates that were homogeneously susceptible to vancomycin.[13] No study before ours has explored the clinical implications of varying degrees of vancomycin susceptibility in MRSA isolates that do not qualify as hetero-VISA.

This study addresses the prevalence of vancomycin heteroresistance among bloodstream MRSA isolates at two large, urban teaching hospitals and explores clinical correlates of bacteremia with isolates exhibiting heterogeneously reduced vancomycin susceptibility, as defined by growth on vancomycin screening agar (vancomycin 4 mg/L). We chose a screening method to encompass as broad an array of potentially heteroresistant isolates as possible. By subjecting subclones to agar dilution, however, we were deliberately conservative, in order to determine whether our isolates fulfilled Hiramatsu's criteria for vancomycin heteroresistance in as unambiguous a manner as possible.[6] Although none of our isolates were characterized as hetero-VISA under these strict criteria, over 40% had subpopulations capable of growth on our vancomycin-containing screening media. Such findings are plausible because S. aureus strains are known to differ in their propensity to develop reduced susceptibility to vancomycin.[31] It is possible that some or all of the isolates from our cases are potential precursors of truly heteroresistant isolates (hetero-VISA), which may in turn be forerunners of VISA.[6,18,32]

Howe et al. have criticized the method of screening for vancomycin heteroresistance in the presence of the antibiotic, followed by MIC testing of subclones, as being poorly reproducible and potentially selecting for, rather than simply detecting, resistant mutants.[29] Regarding reproducibility, we repeated screening for 12 clinical isolates (6 with initially positive results, and 6 with negative results) at different times, using freshly made media each time. Ten (83%) of these had concordant results for the two tests (growth within 48 h on screening media with or without NaCl was the criterion for a positive result). Regarding the possibility of selection, even if this method does select for subclones that grow on screening media, its ability to do so may represent detection of a strain-specific phenomenon that could be of clinical importance if it occurs in vivo during vancomycin therapy.

Having identified the positive screening phenotype among our isolates, we sought to uncover clinical predictors of bacteremia with isolates exhibiting this phenotype, as well as to determine whether such bacteremia was associated with adverse outcomes. Our results were negative on both counts. The results were rendered more robust by remaining essentially unchanged even as we varied the definition of a positive screening result and performed an institution-based subgroup analysis. More stringent definitions of positivity other than those we explored (for example, a requirement of growth on non-salt-containing media), may have led us to undiscovered clinical differences between case-patients and controls.

These negative results can be interpreted in a number of ways. First, our study may have lacked statistical power to detect small differences between the groups in predictors and outcomes. Given the degree to which most of the p values deviate from statistical significance, however, analysis of a substantially larger cohort would be required to disprove such a claim. Moreover, our isolate cohort was highly clonal, as a single type accounted for 69% of isolates and the two most prominent types accounted for 84%. This degree of clonality among pathogenic isolates of MRSA may bias any attempted comparison of clinical features among the bacteremic patients.[33]

Another possible explanation for our failure to detect outcome differences between case-patients and controls is the small degree of vancomycin exposure among all patients in the cohort before blood was drawn for culture. Only 17% (10 case-patients and 14 controls) of patients in the case and control groups received vancomycin before their blood was cultured, and the average amount of time that vancomycin was used by case-patients and controls before culture was approximately 1 day. Clinically, this finding is understandable: Most patients in the cohort had no reason to receive vancomycin before the growth of MRSA. Still, as pharmacologic data were not uniformly available before the patient's hospitalization, we were able to focus on receipt of vancomycin during the index admission only, thereby perhaps limiting our ability to distinguish between study groups based on vancomycin exposure.

No difference may actually exist between bacteremia with VSSA strains that exhibit heterogeneously reduced susceptibility to vancomycin, as defined by growth on vancomycin 4 mg/L screening agar, and homogeneously susceptible isolates. Such a conclusion would not rule out a clinical difference associated with hetero-VISA bacteremia; because our cohort contained no such cases, we cannot draw any conclusions regarding this question. Also, we may have not detected a difference in outcomes because we did not focus on the relevant ones. If, for example, heterogeneously reduced vancomycin susceptibility comes at the expense of a certain degree of virulence, as suggested by Burnie et al. on the basis of a mouse model,[34] then we would not necessarily expect our case-patients to have more hospital deaths or even a greater length of stay in the hospital. We might, however, expect to find several years from now that case-patients will have a greater number of recurrent MRSA infections than will controls, or longer durations of bacteremia during such infections, due to failure of vancomycin to eradicate the organism effectively.

Our results show that despite a subtle phenotypic difference in the MRSA isolates of a large minority of the patients in our cohort, these patients were no different than the reference group with respect to clinical characteristics, antimicrobial use and other hospital exposures, and clinical outcomes. These results add weight to assertions that clinical microbiology laboratories need not routinely screen for vancomycin heteroresistance in S. aureus isolates with vancomycin MICs in the susceptible range.[1,7] Additional studies with larger cohorts and a longer period of follow-up are needed to validate these findings, determine whether they apply to infection with true hetero-VISA, and evaluate outcomes suggestive of noneradicated, indolent infection.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.