Abstract
Background: Previous data have demonstrated the clinical importance of vancomycin MIC values in Staphylococcus aureus bacteraemia (SAB); however, the impact of vancomycin tolerance (VT) is unknown.
Objectives: To compare the frequency of clinical failure between patients with VT and non-VT isolates in SAB.
Methods: This was a retrospective cohort study of patients with SAB, excluding treatment <48 h or polymicrobial bacteraemia. The primary outcome was clinical failure (composite of 30 day mortality, non-resolving signs and symptoms, and 60 day recurrence). Vancomycin MIC and MBC were determined by broth microdilution. The association between VT (MBC/MIC ≥32) and clinical failure was evaluated by multivariable Poisson regression.
Results: Of the 225 patients, 26.7% had VT isolates. VT was associated with clinical failure (48.0% overall) in unadjusted analysis [68.3% (n = 41/60) versus 40.6% (n = 67/165); P < 0.001] and this relationship persisted in multivariable analysis (adjusted risk ratio, 1.74; 95% CI, 1.36-2.24; P < 0.001). The association between VT and clinical failure was also consistent within strata of methicillin susceptibility [methicillin susceptible (n = 125, risk ratio, 1.67; 95% CI, 1.20-2.32; P = 0.002); methicillin resistant (n = 100, risk ratio, 1.69; 95% CI, 1.14-2.51; P = 0.010)]. Among methicillin-susceptible SAB cases treated with β-lactam therapy, VT remained associated with clinical failure (risk ratio, 1.77; 95% CI, 1.19-2.61; P = 0.004).
Conclusions: VT was associated with clinical failure in SAB, irrespective of methicillin susceptibility or definitive treatment. VT may decrease the effectiveness of cell-wall-active therapy or be a surrogate marker of some other pathogen-specific factor associated with poor outcomes. Future research should evaluate if bactericidal non-cell-wall-active agents improve outcomes in VT SAB.
Introduction
Staphylococcus aureus is implicated in a variety of invasive infections, including MSSA bacteraemia (MSSA-B) and MRSA bacteraemia (MRSA-B).1,2 Vancomycin has been the mainstay of MRSA treatment for a half-century, yet treatment failure is common even when the corresponding isolate tests susceptible.3–7 Moreover, the increasing prevalence of elevated vancomycin MIC values has presented new treatment challenges.6,8,9 In a recent meta-analysis, vancomycin MIC ≥1.5 mg/L was associated with treatment failure in MRSA-B.10 Elevated vancomycin MIC is not unique to MRSA-B, however, and has also been independently associated with β-lactam treatment failure in MSSA-B.11,12 A theory within the infectious diseases community is that elevated vancomycin MIC may simply be a marker for some other pathogen-specific factor or combination of factors that alters fitness and leads to poorer outcomes.11
Previous studies have demonstrated an advantage of antimicrobial regimens featuring a bactericidal agent in the treatment of S. aureus bacteraemia (SAB).5,13,14 Reduced bactericidal activity of vancomycin has been associated with poorer outcomes in MRSA-B, including longer duration of bacteraemia and vancomycin treatment failure.5,15 One measure of reduced bactericidal activity is antibiotic tolerance, in which there is a large dissociation between MIC and MBC values (MBC/MIC ratio ≥32).16 In a study of nine hospitals, vancomycin tolerance (VT) was found in 20% of MRSA isolates, although the prevalence was as high as 43% in some institutions.17 Vancomycin tolerance is also observed in MSSA isolates and may be even more prevalent in these infections.18 The literature is scant regarding the relationship between VT and clinical outcomes in SAB and the clinical impact of VT has yet to be fully elucidated. Therefore, the objective of this study was to evaluate the relationship between VT and clinical failure in SAB.
Patients and methods
Study population and data sources
This was a retrospective cohort study of hospitalized patients at the University of Kansas Hospital (Kansas City, Kansas, USA), a tertiary care academic medical centre. All adult patients with a positive blood culture for S. aureus from 1 September 2012 through 30 June 2014 were eligible for inclusion. Exclusion criteria were: (i) antistaphylococcal therapy for <48 h; or (ii) polymicrobial bacteraemia at onset. For patients with multiple episodes of SAB, only the first episode and isolate were included.
Clinical data were collected primarily by retrospective chart review. Variables collected included demographics, ICU admission status at time of initial culture, setting of bacteraemia onset (community acquired versus hospital acquired), comorbidities, antimicrobial treatment data, previous hospitalizations, previous vancomycin exposure within 30 days, vancomycin concentrations, laboratory values and vital signs. Additional mortality (Social Security Death Index), laboratory and vital signs data were collected via the Healthcare Enterprise Repository for Ontological Narration (HERON), a clinical database available at our institution.19 Bacteraemia was considered hospital-acquired if all elements of infection were first present ≥48 h after admission. The associated focus of SAB was determined as documented by a treating physician and categorized according to mortality risk.20 Immunosuppression was defined as neutropenia, leukopenia, chronic steroid use (equivalent to ≥20 mg prednisone for ≥14 days) or antineoplastic use. Acute kidney injury was defined according to RIFLE criteria.21 Baseline sepsis and septic shock (within 24 h of positive blood culture) were defined according to Surviving Sepsis Campaign guidelines.22
Outcome measures
The primary outcome was clinical failure (adapted from previous studies), defined as a composite of: (i) 30 day all-cause mortality; (ii) non-resolving signs and symptoms of bacteraemia (body temperature ≥38 °C, white blood cell count ≥12000/μL or persistent positive blood cultures) for ≥5 days while on antimicrobial therapy; and (iii) recurrent bacteraemia within 60 days of the index SAB episode.6,23 A threshold of 5 days was chosen a priori based on evidence that persistent SAB at that time point is associated with worse clinical outcomes.24
Secondary outcomes were duration of bacteraemia, SAB persistence ≥3 days and hospital length of stay (LOS). Duration of bacteraemia was defined as the time from the first positive S. aureus blood culture until the first negative blood culture. Hospital LOS was defined as the date of first positive S. aureus blood culture until date of discharge.
Microbiological analysis
Clinical S. aureus blood isolates from SAB cases were stored at −70 °C and passed three consecutive days on tryptic soy agar to ensure uniform metabolic activity prior to microbiological testing. Vancomycin and oxacillin MICs were determined by manual broth microdilution (BMD, 0.125-64 mg/L) at a standard inoculum according to CLSI (formerly NCCLS) guidelines.25 Vancomycin MIC was also determined by Etest according to manufacturer recommendations (bioMérieux, Marcy-l’Étoile, France). Etest MICs were read independently by 2 study personnel and discrepancies were resolved by joint re-inspection. In cases of equivocal results, Etest MICs were repeated. A vancomycin MIC of 2 mg/L by BMD or ≥1.5 mg/L by Etest was classified as elevated. The vancomycin and oxacillin MBC was determined by BMD per CLSI recommendations.16 Briefly, a 100 μL aliquot of each well with no visible growth after 24 h of incubation at 35 °C was streaked onto tryptic soy agar, allowed to visibly dry at room temperature and cross-streaked to account for antibiotic carryover.16,26 The MBC was defined as the lowest concentration of drug with ≥99.9% killing at 24 h. Clinical isolates with an MBC/MIC ≥32 by BMD were classified as tolerant. Microbiological analyses were performed in duplicate to ensure reproducibility.
Statistical analysis
Categorical variables were compared by χ2 test or two-tailed Fisher’s exact test and continuous variables were compared by Student’s t-test or Mann-Whitney U-test. The MBC/MIC ratio was assessed as both a continuous and dichotomous (≥32) measure. As a continuous measure, classification and regression tree (CART) analyses were performed to identify the MBC/MIC threshold associated with clinical failure. Additionally, the relationship between increasing MBC/MIC ratio and probability of clinical failure was assessed using the Cochran-Armitage trend test. For studies of common outcomes (>10%), logistic regression overestimates relative measures of association. To overcome this, we attempted to use log-binomial regression. Because of limited population size, the model did not converge. Given that the majority of study outcomes occurred at a fixed time, Poisson regression with robust variance estimates, with either VT or the CART-derived MBC/MIC threshold as the exposure of interest, was used to validly estimate the adjusted risk ratios (RRs) and provide conservative confidence intervals relative to other types of regression models.27 Variables associated with VT or clinical failure in the bivariate analyses (P < 0.2) were eligible for model entry. The most parsimonious models were identified using a backward stepwise approach. Variables which altered the point estimate for either VT or the CART-derived MBC/MIC threshold associated with clinical failure ≥10% were considered potential confounders and retained in the final models. To assess for effect measure modification, stratified analyses were performed. Specific stratifying variables examined were methicillin susceptibility, antibiotic treatment (vancomycin versus β-lactam) and vancomycin troughs. To further control for potential confounding, analyses were conducted in which methicillin resistance and first vancomycin trough concentrations <15 mg/L were forced into multivariable models. Statistical analyses were performed using SAS (version 9.2, SAS Institute Inc., Cary, NC, USA) and Salford Systems (San Diego, CA, USA) with a two-sided P < 0.05 considered statistically significant.
Ethics
The University of Kansas Medical Center institutional review board approved this study with a waiver of informed consent (protocol number 00000310).
Results
Patient and isolate summary
A total of 225 patients met study criteria and were included in the final analysis. The majority of SAB cases were MSSA-B [55.6% (n = 125/225) versus 44.4% (n = 100/225) MRSA-B]. The sources of these infections were skin/soft tissue (n = 53), osteoarticular (n = 39), pneumonia (n = 38), intravenous/haemodialysis catheter (n = 36), unknown (n = 28), cardiac (n = 21), genitourinary (n = 7), intra-abdominal (n = 2) and CNS (n = 1). VT was observed in 60 of the 225 (26.7%) corresponding S. aureus clinical isolates tested. Baseline characteristics of patients with or without VT clinical isolate were compared (Table 1). As shown, there were no statistically significant differences observed between the two groups, including the frequency of infectious diseases consultation. The distribution of vancomycin MIC by BMD and corresponding MBC/MIC ratios are displayed in Table 2. Only 48.9% (n = 110/225) of the isolates analysed had equal vancomycin MIC and MBC values. Vancomycin MIC ≥1.5 mg/L by Etest was observed in 156 of 225 clinical isolates (69.3%) and occurred independently of VT (P = 0.601).
Table 1.
Patient characteristics in SAB according to vancomycin tolerance
| Characteristic | Vancomycin tolerant (n=60) | Vancomycin non-tolerant (n=165) | P |
|---|---|---|---|
| Age (years), mean±SD | 56.8±16.2 | 58.7±14.4 | 0.396a |
| age ≥65 years, n (%) | 19 (31.7) | 57 (34.5) | 0.686 |
| Previous vancomycin exposure within 30 daysb, n (%) | 5 (13.5) | 9 (7.3) | 0.236 |
| Female gender, n (%) | 28 (46.7) | 56 (33.9) | 0.087 |
| Methicillin resistance, n (%) | 24 (40.0) | 76 (46.1) | 0.418 |
| Elevated vancomycin MIC, n (%)c | 40 (66.7) | 116 (70.3) | 0.601 |
| vancomycin Etest MIC=2 mg/L, n (%) | 6 (10.0) | 12 (7.3) | 0.505 |
| Hospital acquired, n (%) | 19 (31.7) | 41 (24.8) | 0.306 |
| ICU, n (%) | 16 (26.7) | 53 (32.1) | 0.433 |
| Infectious diseases physician consultation, n (%) | 47 (78.3) | 129 (78.2) | 0.981 |
| Sepsis, n (%) | 38 (63.3) | 104 (63.0) | 0.967 |
| Septic shock, n (%) | 11 (18.3) | 29 (17.6) | 0.895 |
| Immunosuppression, n (%) | 17 (28.3) | 54 (32.7) | 0.531 |
| High-risk focus, n (%)d | 14 (23.3) | 48 (29.1) | 0.393 |
| S. aureus endocarditis, n (%) | 6 (10.0) | 18 (10.9) | 0.845 |
| Intermediate-risk focus, n (%)d | 36 (60.0) | 84 (50.3) | 0.227 |
| Low-risk focus, n (%)d | 10 (16.7) | 33 (20.0) | 0.574 |
| S. aureus bacteriuria, n (%) | 4 (6.7) | 14 (8.5) | 0.786e |
| Diabetes mellitus, n (%) | 26 (43.3) | 72 (43.6) | 0.968 |
| Alcoholism, n (%) | 3 (5.0) | 17 (10.3) | 0.293e |
| Cirrhosis, n (%) | 5 (8.3) | 15 (9.1) | 0.860 |
| Congestive heart failure, n (%) | 9 (15.0) | 18 (10.9) | 0.404 |
| Malignancy, n (%) | 13 (21.7) | 36 (21.8) | 0.981 |
| Acute kidney injury, n (%) | 15 (25.0) | 44 (26.7) | 0.802 |
| Haemodialysis, n (%) | 12 (20.0) | 37 (22.4) | 0.697 |
| Chronic obstructive pulmonary disease, n (%) | 7 (11.7) | 20 (12.1) | 0.926 |
| Mechanical ventilation, n (%) | 6 (10.0) | 25 (15.2) | 0.321 |
| Pitt bacteraemia score, median (IQR) | 1 (0–3) | 2 (0–3) | 0.449 |
| Charlson comorbidity index, median (IQR) | 6 (4–8) | 6 (4–9) | 0.330 |
| APACHE II, median (IQR) | 10 (6–16) | 11 (7–17) | 0.484 |
Calculated by Student’s t-test; all other continuous variables compared by Mann-Whitney U-test.
Proportions among those with previous hospitalization data available (vancomycin tolerant, n = 37; vancomycin non-tolerant, n = 124)
Etest MIC ≥1.5 mg/L and/or BMD MIC = 2 mg/L.
High-risk foci: cardiac, pneumonia, abdominal, CNS; intermediate-risk foci: osteoarticular, soft-tissue, unknown; low-risk foci: intravenous/haemodialysis catheter, genitourinary.
Calculated by Fisher’s exact test; all other categorical variables compared by χ2 test.
Table 2.
Vancomycin MIC and MBC/MIC ratios by broth microdilution in SAB
| Vancomycin MIC (mg/L) | No. isolates (%) (N=225) | Vancomycin MBC/MIC (n=MRSA) |
|||||
|---|---|---|---|---|---|---|---|
| 1 | 2 | 4 | 8 | 16 | ≥32a | ||
| 0.25 | 2 (0.9) | – | 1 (0) | – | – | – | 1 (1) |
| 0.5 | 74 (32.9) | 23 (7) | 20 (8) | 1 (0) | 5 (3) | 1 (0) | 24 (6) |
| 1 | 146 (64.9) | 85 (41) | 8 (5) | 4 (1) | 4 (2) | 11 (7) | 34 (16) |
| 2 | 3 (1.3) | 2 (2) | – | – | – | – | 1 (1) |
Among the 60 vancomycin-tolerant isolates, 54 had vancomycin MBC values ≥32 mg/L.
Clinical outcomes
Clinical failure was common and occurred in 48.0% (n = 108/225) of SAB cases [MSSA-B, 49.6% (n = 62/125); MRSA-B, 46.0% (n = 46/100); P = 0.591]. Vancomycin tolerance was significantly associated with SAB clinical failure in unadjusted analysis [68.3% (n = 41/60) versus 40.6% (n = 67/165); P < 0.001]. A summary of all clinical outcomes evaluated is included in Table 3. Overall, 30 day mortality was observed in 13.3% of cases (n = 30/225). The median duration of SAB was 2 days (IQR, 1-4 days). The CART-derived MBC/MIC breakpoint for clinical failure was 48. There were 43 isolates (19.1%) with an MBC/MIC ≥ 48. Factors associated with clinical failure (P < 0.2) in bivariate analysis included VT (P < 0.001), vancomycin MBC/MIC ≥ 48 (P < 0.001), vancomycin exposure within preceding 30 days (P = 0.001), sepsis (P = 0.007), septic shock (P < 0.001), hospital-acquired infection (P = 0.048), high-risk focus (P = 0.049), ICU admission (P = 0.017), cirrhosis (P = 0.193), alcoholism (P = 0.091) and vancomycin Etest = 2 mg/L (P = 0.098). Elevated vancomycin MIC (≥1.5 mg/L by Etest or 2 mg/L by BMD) was not significantly associated with clinical failure (P = 0.586). In multivariable analysis, the association between clinical failure and both VT and the CART-derived MBC/MIC breakpoint persisted (Table 4). Proportions of clinical failure by vancomycin MBC/MIC ratio were compared and a trend of more frequent clinical failure with increasing MBC/MIC ratio was observed (P = 0.001; Figure 1).
Table 3.
Comparison of clinical outcomes by vancomycin tolerance in SAB
| Outcome | Vancomycin tolerant (n=60) | Vancomycin non-tolerant (n=165) | P |
|---|---|---|---|
| Clinical failure, n (%) | 41 (68.3) | 67 (40.6) | <0.001 |
| 30 day all-cause mortality, n (%) | 6 (10.0) | 24 (14.5) | 0.375 |
| persistent SAB signs/symptomsa, n (%) | 36 (60.0) | 46 (27.9) | <0.001 |
| 60 day SAB recurrence, n (%) | 2 (3.3) | 3 (1.8) | 0.611 |
| Hospital LOS (days), median (IQR) | 11 (7–17) | 11 (6–20) | 0.780 |
| Bacteraemia duration (days), median (IQR) | 2 (1–3) | 2 (1–4) | 0.504 |
| persistence ≥3 daysb, n (%) | 27 (48.2) | 56 (38.4) | 0.202 |
Categorical variables compared by χ2 test.
Continuous variables compared by Mann-Whitney U-test.
Body temperature ≥38°C, white blood cell count ≥12000/μL or persistent positive blood cultures for ≥ 5 days while on antimicrobial therapy.
Analysis includes only patients with ≥1 follow-up blood culture drawn [vancomycin tolerant (n = 56); vancomycin non-tolerant (n = 146)].
Table 4.
Multivariable Poisson regression models of factors associated with clinical failure in SAB
| Factor (N=225) | Adjusted RR (95% CI) | P |
|---|---|---|
| Model 1 | ||
| vancomycin tolerance | 1.74 (1.36–2.24) | <0.001 |
| ICU admission | 1.35 (1.04–1.76) | 0.026 |
| high-risk focusa | 1.26 (0.96–1.67) | 0.102 |
| Model 2 | ||
| vancomycin MBC/MIC ≥48b | 1.81 (1.42–2.31) | <0.001 |
| ICU admission | 1.33 (1.02–1.74) | 0.033 |
| high-risk focusa | 1.23 (0.93–1.63) | 0.148 |
Variables eligible for inclusion in multivariable models: vancomycin Etest MIC = 2 mg/L, hospital-acquired infection, ICU admission, sepsis, septic shock, vancomycin exposure within preceding 30 days, focus of infection, alcoholism, cirrhosis, female gender.
High-risk foci: cardiac, pneumonia, abdominal, CNS.
CART analysis breakpoint.
Figure 1.
Percentages of clinical failure in SAB according to vancomycin MBC/MIC ratio. MBC/MIC ratio categories with n < 5 are not depicted. Trend tested by Cochran-Armitage test.
Methicillin resistance
Vancomycin tolerance remained significantly associated with clinical failure within strata of methicillin susceptibility [MSSA (n = 125, unadjusted RR, 1.67; 95% CI, 1.20-2.32; P = 0.002); MRSA (n = 100, unadjusted RR, 1.69; 95% CI, 1.14-2.51; P = 0.010); Figure 2 and Table 5]. In addition, after adjusting for methicillin resistance and other confounders in multivariable Poisson regression the relationship between VT and clinical failure persisted [Table S1 (available as Supplementary data at JAC Online), models 1 and 2].
Figure 2.
Percentages of clinical failure in SAB according to vancomycin tolerance, stratified by methicillin resistance. Reported P values are unadjusted.
Table 5.
Analysis of clinical failure by vancomycin tolerance stratified for methicillin resistance, antibiotic choice and vancomycin trough concentrations
| Vancomycin tolerant | Vancomycin non-tolerant | RR (95% CI) | P | |
|---|---|---|---|---|
| Methicillin susceptibility | ||||
| MRSA (n=100) | 16/24 (66.7) | 30/76 (39.5) | 1.69 (1.14–2.51) | 0.010 |
| MSSA (n=125) | 25/36 (69.4) | 37/89 (41.6) | 1.67 (1.20–2.32) | 0.002 |
| Empirical therapy | ||||
| vancomycin (n=209) | 38/56 (67.9) | 65/153 (42.6) | 1.79 (1.19–2.68) | 0.001 |
| MRSA (n=94) | 14/22 (63.6) | 29/72 (40.3) | 1.64 (1.01–2.95) | 0.046 |
| MSSA (n=115) | 24/34 (70.6) | 36/81 (44.4) | 1.89 (1.08–3.29) | 0.010 |
| Definitive therapy | ||||
| Overall | ||||
| vancomycin (n=119) | 19/29 (65.5) | 40/90 (44.4) | 1.43 (1.01–2.02) | 0.048 |
| MRSA (n=100) | ||||
| vancomycin (n=89) | 14/22 (63.6) | 29/67 (43.3) | 1.56 (0.86–2.82) | 0.097 |
| MSSA (n=125) | ||||
| vancomycin (n=30) | 5/7 (71.4) | 11/23 (47.8) | 1.83 (0.53–6.28) | 0.399 |
| non-VAN (n=95)a | 20/29 (69.0) | 26/66 (39.4) | 1.65 (1.11–1.86) | 0.008 |
| β-lactam (n=93)b | 20/29 (69.0) | 25/64 (39.1) | 1.77 (1.19–2.61) | 0.004 |
| Vancomycin trough concentrations | ||||
| Empirical (n=126) | ||||
| <15 mg/L (n=32) | 5/9 (41.7) | 7/23 (30.4) | 1.56 (0.72–3.41) | 0.187 |
| ≥15 mg/L (n=94) | 16/26 (61.5) | 37/68 (54.4) | 1.19 (0.68–2.06) | 0.533 |
| Definitive (n=80) | ||||
| <15 mg/L (n=14) | 2/3 (66.7) | 4/11 (36.3) | 1.91 (0.36–10.56) | 0.538 |
| ≥15 mg/L (n=66) | 10/15 (66.7) | 26/51 (51.0) | 1.47 (0.68–3.17) | 0.283 |
Non-VAN, non-vancomycin treatment [includes β-lactam and daptomycin (n = 2)].
Includes nafcillin, ampicillin/sulbactam, piperacillin/tazobactam, cefazolin, ceftriaxone, cefepime.
Antibiotic treatment
All of the patients in this study received appropriate empirical treatment within 24 h of positive S. aureus blood culture. Among patients receiving empirical vancomycin (n = 209), VT was significantly associated with clinical failure, which persisted when further stratified by methicillin resistance (Table 5). Among patients receiving vancomycin as definitive therapy (n = 119), VT was significantly associated with clinical failure; however, this relationship did not persist after further stratification by methicillin resistance (Table 5). Clinical failure among patients with MRSA-B treated with definitive vancomycin was 48.3% (n = 43/89).
In the MSSA cohort (n = 125), 62 patients (49.6%) were treated with a penicillin derivative (nafcillin, piperacillin/tazobactam or ampicillin/sulbactam), 31 (24.8%) with a cephalosporin (cefazolin, ceftriaxone or cefepime), 2 (1.6%) with daptomycin and 30 (24%) with vancomycin as definitive therapy. Among those treated with definitive β-lactam therapy (n = 93), the majority received empirical therapy with either vancomycin plus a broad-spectrum β-lactam or β-lactam monotherapy. Only 25 patients received vancomycin monotherapy prior to definitive β-lactam therapy. Among these patients, there was no difference in median time to definitive β-lactam therapy between VT and vancomycin non-tolerant groups [50.6 h (IQR, 27.8-79.7 h) versus 56.9 h (IQR, 37.0-66.2 h); P = 0.839] and median time to definitive therapy was not associated with clinical failure [61.0 h (IQR, 37.9-74.1 h) versus 54.3 h (IQR, 12.7-66.0 h); P = 0.312].
The association between VT and clinical failure remained in MSSA patients treated with definitive β-lactam therapy (n = 93) or non-vancomycin definitive therapy (n = 95; Table 5). Interestingly, β-lactam tolerance (measured in vitro using oxacillin) was not associated with clinical failure in definitive β-lactam therapy MSSA patients (RR, 1.08; 95% CI, 0.79-1.46; P = 0.614).
Vancomycin trough concentrations
Vancomycin trough concentrations were available in 126/209 (60.3%) cases of empirical vancomycin treatment [MSSA-B, n = 64/115 (55.7%); MRSA-B, n = 62/94 (65.9%)]. Suboptimal concentrations (<15 mg/L) were observed in 32/126 (25.4%) cases [MSSA-B, n = 24/64 (37.5%); MRSA-B, n = 8/62 (12.9%)] and were not significantly associated with clinical failure among these patients [59.4% (n = 19/32) versus 48.9% (n = 46/94); P = 0.307]. Within strata of therapeutic and suboptimal vancomycin concentrations, no association was observed between VT and clinical failure (Table 5). After adjusting for other confounding variables, including suboptimal vancomycin trough concentrations, both VT and the CART-derived MBC/MIC breakpoint remained significantly associated with clinical failure (Table S1, models 3 and 4).
Vancomycin trough concentrations were available in 80/119 (67.2%) cases receiving vancomycin definitive therapy [MSSA-B, n = 19/30 (63.3%); MRSA-B, n = 61/89 (68.5%)]. Suboptimal concentrations (<15 mg/L) were observed in 14/80 (17.5%) cases [MSSA-B, n = 6/19 (31.6%); MRSA-B, n = 8/61 (13.1%)]. Suboptimal vancomycin concentrations were not associated with clinical failure in the vancomycin definitive treatment subgroup [42.9% (n = 6/14) versus 54.5% (n = 36/66); P = 0.426] and did not modify the relationship between VT and clinical failure.
Discussion
In this study, we found an independent association between VT and clinical failure in SAB that occurred irrespective of methicillin susceptibility. This association also persisted within strata of vancomycin treatment and β-lactam treatment of MSSA-B.
While previous researchers have evaluated the impact of bactericidal activity in SAB, data supporting the clinical relevance of this activity as measured by MBC/MIC ratio are limited.13,28–30 The most recent study evaluated S. aureus endocarditis (n = 62) and was underpowered to detect an association between VT and clinical outcomes.29 Data on the influence of β-lactam tolerance on clinical outcomes in MSSA infections are conflicting as well. One study failed to find an association between oxacillin tolerance and clinical outcomes, while another found nafcillin tolerance was associated with increased duration of clinical symptoms of infection.14,29 This is similar to our findings, in that clinical signs and symptoms also persisted significantly longer when VT was observed in vitro. We did not, however, find an association between oxacillin tolerance and clinical failure.
A potential explanation for decreased killing at higher vancomycin concentrations relative to the MIC lies in the “paradoxical effect”, in which S. aureus is killed more slowly at higher concentrations of antibiotic than at concentrations slightly above the MIC.31 This has been observed with vancomycin versus MRSA and represents a potential mechanism of VT.32 At clinical vancomycin concentrations (12 mg/L) vancomycin can bind and block access of murein hydrolases to substrates, leading to inhibition of the cell wall autolytic system in some S. aureus strains.32 More recent evidence demonstrates that vancomycin exposure leads to a doubling in cell wall thickness among VT MRSA strains.33 If tolerance can be induced at vancomycin concentrations that would be observed in vivo, it is plausible that this may have clinical implications.34,35 The observed increase in clinical failure associated with the VT phenotype described in the present study appears to support this hypothesis.
Although VT was associated with increased clinical failure in this cohort, a difference in 30 day mortality was not observed. Rather, the point estimate for mortality was lower when VT was observed in vitro. Similarly, heteroresistant vancomycin-intermediate S. aureus (hVISA) infection has been independently associated with persistent bacteraemia despite decreased mortality in SAB.2,36 This phenomenon appears to be due to alterations in the accessory gene regulator (agr) controlling for virulence in S. aureus.23 The agr group II genotype has been associated with reduced vancomycin bactericidal activity.37 Although we did not perform agr genotyping or hVISA screening in this study, the lack of an observed mortality increase despite high rates of clinical failure suggest a possible interplay between agr genotype, virulence, heteroresistance and VT that warrants further exploration.
Conflicting results from studies examining the relationship between elevated vancomycin MIC and clinical outcomes in SAB may be partially explained by unmeasured phenotypic variation or reduced vancomycin bactericidal activity.37,38 In the present study, the observation that VT – not elevated vancomycin MIC – was significantly associated with clinical failure in SAB is a novel finding that supports this hypothesis. As no association between VT and elevated MIC was observed in this cohort, these microbiological characteristics appear to be caused by distinct mechanisms. It is plausible that VT may have served as an unrecognized variable associated with treatment failure in previous studies. As VT was not significantly associated with any corresponding patient-specific factors, its presence appears to be due to some other microbiological characteristic that warrants further characterization.
We observed VT in 26.7% of clinical isolates, consistent with percentages at other institutions.17 The observation that VT is prevalent among MSSA isolates has also been previously described.18,29 A novel finding from the present study was the association between VT and clinical failure even among β-lactam-treated patients with MSSA-B. As vancomycin and β-lactams both act at similar sites in the bacterial cell wall, it is biologically plausible that changes resulting in decreased vancomycin activity may also lead to lower effectiveness of other cell-wall-active agents. Our findings, combined with previous data demonstrating an association between elevated vancomycin MIC and mortality in patients receiving antistaphylococcal penicillin therapy for MSSA-B, suggest vancomycin susceptibility may importantly influence the effectiveness of cell-wall-active agents.11 Intriguingly, baseline oxacillin tolerance was not associated with an increased risk of clinical failure among patients with MSSA-B. Since the majority of patients were treated empirically with vancomycin, it is unclear if the first few days of therapy are the most important, if vancomycin exposure leads to some inducible oxacillin tolerance or if VT is a marker for some unknown factor leading to poorer outcomes.
Vancomycin exposure within the preceding 30 days has been associated with a decreased in vitro killing by vancomycin among MRSA blood isolates.39,40 In the present study, VT clinical isolates were more common among those with previous vancomycin exposure, but did not reach statistical significance. Previous vancomycin exposure was significantly associated with clinical failure in univariable analysis, but was not a confounder in multivariable analyses. As the number of patients with known vancomycin exposure was limited, the relationship between previous vancomycin exposure, reduced vancomycin bactericidal activity and clinical outcomes warrants future investigation.
The present study has a number of limitations. This study featured patients with mostly intermediate-risk foci, a relatively short duration of culture-positive infection and a low overall mortality rate (13.3% versus approximately 20% in the previously published literature).2 Thus, the external validity of these results should be assessed within the context of the patient population studied. The definition of clinical failure included a composite of signs and symptoms of infection within a timeframe relevant to the prognosis of patients with SAB; however, data on potentially important biomarkers such as C-reactive protein (CRP) were not collected as they are not commonly utilized in the management of SAB at our institution.2,24 Despite a more conservative clinical failure definition, the observed clinical failure rate in this study was consistent with previous literature.23 We attempted to adjust for the potential influence of antibiotic choice on clinical failure within strata of MSSA-B and MRSA-B in stratified analyses, but the limited sample in these subgroups left us underpowered to assess for potential differences in outcomes. Within these strata with sparse subjects (e.g., vancomycin in the MSSA-B and MRSA-B strata), the relationship between VT and clinical failure remains unknown. The inability to detect a relationship between vancomycin trough concentrations and clinical failure in this cohort could also be due to limited sample size. A significantly larger sample than was available would be required to power these subanalyses. Additionally, we were unable to capture the duration of treatment completed upon hospital discharge. Finally, although bacterial isolates were passed three times post-freezing, we cannot exclude the possibility that the bacterial population distribution may have shifted during the freezing process.
While this study found VT to be a novel predictor of clinical failure in SAB, it is important to note this is likely one component of a complex relationship of clinical and microbiological factors influencing outcomes in SAB. Future research should prospectively evaluate if a relationship between VT and clinical failure exists when bactericidal alternatives to cell-wall-active agents are used for SAB. This will help to elucidate if VT is a clinical predictor for failure with β-lactam agents or rather a surrogate for some unknown virulence factor. If alternative treatments positively modify this relationship, clinical microbiology laboratories should consider expanded testing services to include MBC so that patients with VT can be readily identified or – at a minimum – performing these tests on a representative sample of clinical isolates to understand the institution-specific distribution of VT. These actions should aid clinicians in strategically selecting antimicrobial therapy with the highest probability of success in patients with SAB.
In summary, VT, but not elevated vancomycin MIC, was significantly associated with clinical failure in a population of primarily intermediate-risk SAB. This association persisted regardless of methicillin susceptibility or antibiotic treatment and after adjusting for confounding factors in multivariable regression. This finding adds to the growing body of evidence underscoring the importance of bactericidal activity in SAB. Due to the observed association between VT and clinical failure among patients with MSSA-B treated with a β-lactam, it appears that VT may serve as an important marker for failure of cell-wall-active antimicrobial therapy. Alternatives to cell-wall-active antimicrobials may be warranted in cases of SAB caused by VT isolates. Further research to determine the optimal treatment of VT SAB is needed.
Supplementary Material
Acknowledgements
This study was presented in part at the Translational Science Meeting, Washington, DC, USA, 2014 (Oral Session D1) and the Fifty-fourth Interscience Conference on Antimicrobial Agents and Chemotherapy, Washington, DC, USA, 2014 (Abstract # K1900).
We are grateful to Tamara McMahon for her advice and assistance with the HERON database and to Dr Kenneth Lamp for his thoughtful review.
Funding
This work was supported by the National Institutes of Health (TL1TR000120-03 to N. S. B. and UL1TR000001 to the HERON database team).
Transparency declarations
None to declare.
Supplementary data
Table S1 is available as Supplementary data at JAC Online (http://jac.oxfordjournals.org/).
References
- 1. Lowy FD. Staphylococcus aureus infections. N Engl J Med 1998; 339: 520–32. [DOI] [PubMed] [Google Scholar]
- 2. van Hal SJ, Jensen SO, Vaska VL. et al. Predictors of mortality in Staphylococcus aureus bacteremia. Clin Microbiol Rev 2012; 25: 362–86. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Tenover FC, Moellering RC., Jr. The rationale for revising the Clinical and Laboratory Standards Institute vancomycin minimal inhibitory concentration interpretive criteria for Staphylococcus aureus. Clin Infect Dis 2007; 44: 1208–15. [DOI] [PubMed] [Google Scholar]
- 4. Rybak MJ, Lomaestro BM, Rotschafer JC. et al. Therapeutic monitoring of vancomycin in adults: summary of consensus recommendations from the American Society of Health-System Pharmacists, the Infectious Diseases Society of America, and the Society of Infectious Diseases Pharmacists. Pharmacotherapy 2009; 29: 1275–9. [DOI] [PubMed] [Google Scholar]
- 5. Sakoulas G, Moise-Broder PA, Schentag J. et al. Relationship of MIC and bactericidal activity to efficacy of vancomycin for treatment of methicillin-resistant Staphylococcus aureus bacteremia. J Clin Microbiol 2004; 42: 2398–402. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Kullar R, Davis SL, Levine DP. et al. Impact of vancomycin exposure on outcomes in patients with methicillin-resistant Staphylococcus aureus bacteremia: support for consensus guidelines suggested targets. Clin Infect Dis 2011; 52: 975–81. [DOI] [PubMed] [Google Scholar]
- 7. Moravvej Z, Estaji F, Askari E. et al. Update on the global number of vancomycin-resistant Staphylococcus aureus (VRSA) strains. Int J Antimicrob Agents 2013; 42: 370–1. [DOI] [PubMed] [Google Scholar]
- 8. Brink AJ. Does resistance in severe infections caused by methicillin-resistant Staphylococcus aureus give you the ‘creeps’? Curr Opin Crit Care 2012; 18: 451–9. [DOI] [PubMed] [Google Scholar]
- 9. Howden BP, Ward PB, Charles PG. et al. Treatment outcomes for serious infections caused by methicillin-resistant Staphylococcus aureus with reduced vancomycin susceptibility. Clin Infect Dis 2004; 38: 521–8. [DOI] [PubMed] [Google Scholar]
- 10. van Hal SJ, Lodise TP, Paterson DL. The clinical significance of vancomycin minimum inhibitory concentration in Staphylococcus aureus infections: a systematic review and meta-analysis. Clin Infect Dis 2012; 54: 755–71. [DOI] [PubMed] [Google Scholar]
- 11. Holmes NE, Turnidge JD, Munckhof WJ. et al. Antibiotic choice may not explain poorer outcomes in patients with Staphylococcus aureus bacteremia and high vancomycin minimum inhibitory concentrations. J Infect Dis 2011; 204: 340–7. [DOI] [PubMed] [Google Scholar]
- 12. Holmes NE, Turnidge JD, Munckhof WJ. et al. Vancomycin minimum inhibitory concentration, host comorbidities and mortality in Staphylococcus aureus bacteraemia. Clin Microbiol Infect 2013; 19: 1163–8. [DOI] [PubMed] [Google Scholar]
- 13. Denny AE, Peterson LR, Gerding DN. et al. Serious staphylococcal infections with strains tolerant to bactericidal antibiotics. Arch Intern Med 1979; 139: 1026–31. [PubMed] [Google Scholar]
- 14. Rahal JJ, Jr, Chan YK, Johnson G. Relationship of staphylococcal tolerance, teichoic acid antibody, and serum bactericidal activity to therapeutic outcome in Staphylococcus aureus bacteremia. Am J Med 1986; 81: 43–52. [DOI] [PubMed] [Google Scholar]
- 15. Moise PA, Sakoulas G, Forrest A. et al. Vancomycin in vitro bactericidal activity and its relationship to efficacy in clearance of methicillin-resistant Staphylococcus aureus bacteremia. Antimicrob Agents Chemother 2007; 51: 2582–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. National Committee for Clinical Laboratory Standards. Methods for Determining Bactericidal Activity of Antimicrobial Agents: Approved Guidline M26-A. NCCLS, Wayne, PA, USA, 1999. [Google Scholar]
- 17. Sader HS, Jones RN, Rossi KL. et al. Occurrence of vancomycin-tolerant and heterogeneous vancomycin-intermediate strains (hVISA) among Staphylococcus aureus causing bloodstream infections in nine USA hospitals. J Antimicrob Chemother 2009; 64: 1024–8. [DOI] [PubMed] [Google Scholar]
- 18. Appleman MD, Citron DM. Efficacy of vancomycin and daptomycin against Staphylococcus aureus isolates collected over 29 years. Diagn Microbiol Infect Dis 2010; 66: 441–4. [DOI] [PubMed] [Google Scholar]
- 19. Waitman LR, Warren JJ, Manos EL. et al. Expressing observations from electronic medical record flowsheets in an i2b2 based clinical data repository to support research and quality improvement. AMIA Annu Symp Proc 2011; 2011: 1454–63. [PMC free article] [PubMed] [Google Scholar]
- 20. Soriano A, Martinez JA, Mensa J. et al. Pathogenic significance of methicillin resistance for patients with Staphylococcus aureus bacteremia. Clin Infect Dis 2000; 30: 368–73. [DOI] [PubMed] [Google Scholar]
- 21. Hoste EA, Clermont G, Kersten A. et al. RIFLE criteria for acute kidney injury are associated with hospital mortality in critically ill patients: a cohort analysis. Crit Care 2006; 10: R73. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Dellinger RP, Levy MM, Rhodes A. et al. Surviving sepsis campaign: international guidelines for management of severe sepsis and septic shock: 2012. Crit Care Med 2013; 41: 580–637. [DOI] [PubMed] [Google Scholar]
- 23. Moise-Broder PA, Sakoulas G, Eliopoulos GM. et al. Accessory gene regulator group II polymorphism in methicillin-resistant Staphylococcus aureus is predictive of failure of vancomycin therapy. Clin Infect Dis 2004; 38: 1700–5. [DOI] [PubMed] [Google Scholar]
- 24. Kullar R, McKinnell JA, Sakoulas G. Avoiding the perfect storm: the biologic and clinical case for reevaluating the 7-day expectation for methicillin-resistant Staphylococcus aureus bacteremia before switching therapy. Clin Infect Dis 2014; 59: 1455–61. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Clinical and Laboratory Standards Institute. Methods for Dilution Antimicrobial Susceptibility Tests for Bacteria that Grow Aerobically—Ninth Edition: Approved Standard M07-A9. Clinical and Laboratory Standards Institute, Wayne, PA, USA, 2012. [Google Scholar]
- 26. Pelletier LL, Jr, Baker CB. Oxacillin, cephalothin, and vancomycin tube macrodilution MBC result reproducibility and equivalence to MIC results for methicillin-susceptible and reputedly tolerant Staphylococcus aureus isolates. Antimicrob Agents Chemother 1988; 32: 374–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. McNutt LA, Wu C, Xue X. et al. Estimating the relative risk in cohort studies and clinical trials of common outcomes. Am J Epidemiol 2003; 157: 940–3. [DOI] [PubMed] [Google Scholar]
- 28. May J, Shannon K, King A. et al. Glycopeptide tolerance in Staphylococcus aureus. J Antimicrob Chemother 1998; 42: 189–97. [DOI] [PubMed] [Google Scholar]
- 29. Pasticci MB, Moretti A, Stagni G. et al. Bactericidal activity of oxacillin and glycopeptides against Staphylococcus aureus in patients with endocarditis: looking for a relationship between tolerance and outcome. Ann Clin Microbiol Antimicrob 2011; 10: 26. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Rajashekaraiah KR, Rice T, Rao VS. et al. Clinical significance of tolerant strains of Staphylococcus aureus in patients with endocarditis. Ann Intern Med 1980; 93: 796–801. [DOI] [PubMed] [Google Scholar]
- 31. Eagle H, Fleischman R, Musselman AD. The bactericidal action of penicillin in vivo: the participation of the host, and the slow recovery of the surviving organisms. Ann Intern Med 1950; 33: 544–71. [DOI] [PubMed] [Google Scholar]
- 32. Sieradzki K, Tomasz A. Inhibition of the autolytic system by vancomycin causes mimicry of vancomycin-intermediate Staphylococcus aureus-type resistance, cell concentration dependence of the MIC, and antibiotic tolerance in vancomycin-susceptible S. aureus. Antimicrob Agents Chemother 2006; 50: 527–33. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Cazares-Dominguez V, Cruz-Cordova A, Ochoa SA. et al. Vancomycin tolerant, methicillin-resistant Staphylococcus aureus reveals the effects of vancomycin on cell wall thickening. PLoS One 2015; 10: e0118791. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Rodvold KA, Blum RA, Fischer JH. et al. Vancomycin pharmacokinetics in patients with various degrees of renal function. Antimicrob Agents Chemother 1988; 32: 848–52. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Patel N, Pai MP, Rodvold KA. et al. Vancomycin: we can't get there from here. Clin Infect Dis 2011; 52: 969–74. [DOI] [PubMed] [Google Scholar]
- 36. van Hal SJ, Jones M, Gosbell IB. et al. Vancomycin heteroresistance is associated with reduced mortality in ST239 methicillin-resistant Staphylococcus aureus blood stream infections. PLoS One 2011; 6: e21217. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Rose WE, Fallon M, Moran JJ. et al. Vancomycin tolerance in methicillin-resistant Staphylococcus aureus: influence of vancomycin, daptomycin, and telavancin on differential resistance gene expression. Antimicrob Agents Chemother 2012; 56: 4422–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Honda H, Doern CD, Michael-Dunne W., Jr et al. The impact of vancomycin susceptibility on treatment outcomes among patients with methicillin resistant Staphylococcus aureus bacteremia. BMC Infect Dis 2011; 11: 335. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Britt NS, Patel N, Horvat RT. et al. Vancomycin 24-hour area under the curve/minimum bactericidal concentration ratio as a novel predictor of mortality in methicillin-resistant Staphylococcus aureus bacteremia. Antimicrob Agents Chemother 2016; 60: 3070–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Moise PA, Smyth DS, El-Fawal N. et al. Microbiological effects of prior vancomycin use in patients with methicillin-resistant Staphylococcus aureus bacteraemia. J Antimicrob Chemother 2008; 61: 85–90. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.