Mortality Due to Vancomycin-Resistant Enterococcal Bacteremia versus Vancomycin-Susceptible Enterococcal Bacteremia: An Ecological Analysis

15 NOVEMBER Correspondence Mortality Due to Vancomycin-Resistant Enterococcal Bacteremia versus VancomycinSusceptible Enterococcal Bacteremia: An Ecological Analysis Sir—DiazGranados et al. [1] estimated the impact of vancomycin resistance on mortality associated with enterococcal bacteremia. Using a meta-analytic method to aggregate the results of 9 studies, they demonstrated that there is a significantly higher mortality risk for patient groups with vancomycin-resistant enterococcal (VRE) bacteremia than for patient groups with vancomycin-susceptible enterococcal (VSE) bacteremia. They had limited their analysis to 9 studies that had adjusted for severity of illness, 7 of which used APACHE II scores for this. However, the length of hospital stay (LOS) before the onset of bacteremia was significantly greater for the patients with VRE bacteremia than for the patients with VSE bacteremia in 16 studies that provide these data: median LOS for patient groups with VSE bacteremia was 14 days, and median LOS for patient groups with VRE bacteremia was 27 days (P p .0005, Wilcoxon rank sum test) (table 1). Is the use of APACHE II scores to adjust for severity of illness valid for patients at different days of their stay in the hospital? Patients residing in the hospital for longer periods of time are not only more sick, but they are more likely to have been exposed to multiple courses of antibiotics and more likely to acquire drug-resistant enterococcal species as a result. With this question in mind, I show the relationship between group mortality rates and group LOS values before bacteremia for each group in the 14 studies that included this data (figure 1). I have included a more recently published study [14]. LOS information was not available for all studies cited by DiazGranados et al. [1]. In linear regression, the group mortality per day of hospital stay increases by 0.89% (P ! .001) when weighted for study size and by 0.71% (P p .035) when unweighted for study size. This suggests that, for enterococcal bacteremia of either type, the group mortality rate increases with the group LOS value. Although this is an ecological relationship based on group average data, it does suggest a systematic bias between the 2 groups that adjustment by APACHE II score may not adequately control for. Of interest is the relationship between the group mortality rate and the mean LOS prior to bacteremia onset among patient groups; this relationship is similar to the relationship seen in studies of the association of methicillin-resistant Staphy- lococcus aureus bacteremia and methicillin-susceptible Staphylococcus aureus bacteremia with group mortality risk [18]. Similar analytic issues were identified in a meta-analysis performed to estimate the magnitude of the association between antecedent vancomycin use and the acquisition of VRE bacteremia [19]. The association between antecedent vancomycin use and risk of acquisition of VRE bacteremia is stronger but significantly more heterogeneous in a meta-analysis of studies that used patient groups with VSE as controls than in a meta-analysis of studies that used a differently selected reference group. Moreover, studies that failed to adjust for LOS prior to the acquisition of bacteremia also found effect measures that were stronger but significantly more heterogeneous than those found in a meta- Table 1. Group mean length of hospital stay (LOS) prior to onset of enterococcal bacteremia in 16 studies. Mean LOS in days, by patient group VSE bacteremia VRE bacteremia DiazGranados and Jernigan [2] Edmond et al. [3] Garbutt et al. [4] 15 NA 6 22 24 21 Gleason et al. [5] Lautenbach et al. [6] Linden et al. [7] 10 11 24 29 23 43 Lodise et al. [8] Lucas et al. [9] 10 17 12 24 Mainous et al. [10] Montecalvo et al. [11] Peset et al. [12] 12 NA 24 28 26 41 Shay et al. [13] Song et al. [14] Stosor et al. [15] NA NA 17 27 20 34 Stroud et al. [16] Vergis et al. [17] NA 3 33 17 Reference NOTE. NA, not available; VSE, vancomycin-sensitive enterococci; VRE, vancomycin-resistant enterococi. CORRESPONDENCE • CID 2005:41 (15 November) • 1541 Figure 1. The relationship between length of hospital stay (LOS) before the onset of bacteremia and mortality rates for groups of patients with either vancomycin-resistant enterococcal bacteremia (filled circles) or vancomycin-resistant enterococcal bacteremia (open circles) from 16 studies that reported LOS data. The groups’ sizes (!35, 36–54, 55–124, and 1125 patients) are represented by the increasing size of the symbols. analysis using only studies that attempted to account for LOS [19]. The difference that DiazGranados et al. [1] describe could be confounded by the difference in LOS between the groups of patients with VRE bacteremia and the groups with VSE bacteremia, irrespective of any adjustment for APACHE II score. Acknowledgments Potential conflicts of interest. J.C.H.: no conflicts. James C. Hurley Division of Medicine and University of Melbourne Rural Clinical School, Ballarat Health Services, Ballarat, Victoria, Australia References 1. DiazGranados CA, Zimmer SM, Klein M, Jernigan JA. Comparison of mortality associated with vancomycin-resistant and vancomycinsusceptible enterococcal bloodstream infections: a meta-analysis. Clin Infect Dis 2005; 41:327–33. 2. DiazGranados CA, Jernigan JA. Impact of vancomycin resistance on mortality among neutropenic patients with enterococcal blood stream infections. J Infect Dis 2005; 191: 588–95. 3. Edmond MB, Ober JF, Dawson JD, Weinbaum DL, Wenzel RP. Vancomycin-resistant enterococcal bacteremia: natural history and at- tributable mortality. Clin Infect Dis 1996; 23: 1234–9. 4. Garbutt JM, Ventrapragada M, Littenberg B, Mundy LM. Association between resistance to vancomycin and death in cases of Enterococcus faecium bacteremia. Clin Infect Dis 2000; 30: 466–72. 5. Gleason T, Crabtree T, Pelletier S, et al. Prediction of poorer prognosis by infection with antibiotic-resistant gram-positive cocci than infection with antibiotic sensitive strains. Arch Surg 1999; 134:1033–40. 6. Lautenbach E, Bilker WB, Brennan PJ. Enterococcal bacteremia: risk factors for vancomycin resistance and predictors of mortality. Infect Control Hosp Epidemiol 1999; 20: 318–23. 7. Linden PK, Pasculle AW, Manez R, et al. Differences in outcomes for patients with bacteremia due to vancomycin-resistant Enterococcus faecium or vancomycin-susceptible E. faecium. Clin Infect Dis 1996; 22:663–70. 8. Lodise TP, McKinnon PS, Tam VH, Rybak MJ. Clinical outcomes for patients with bacteremia caused by vancomycin-resistant Enterococcus in a level 1 trauma center. Clin Infect Dis 2002; 34:922–9. 9. Lucas GM, Lechtzin N, Puryear DW, Yau LL, Flexner CW, Moore RD. Vancomycin-resistant and vancomycin-susceptible enterococcal bacteremia: comparison of clinical features and outcomes. Clin Infect Dis 1998; 26: 1127–33. 10. Mainous MR, Lipsett PA, O’Brien M. Enterococcal bacteremia in the surgical (...truncated)