INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

The prevalence of infections due to enterococci has increased dramatically in the past decade. The latest National Nosocomial Infection Study reported that enterococci were the second most common organisms recovered from cultures and that the prevalence of vancomycin-resistant Enterococcus faecium (VREF) infection increased from 0.3% in 1989 to 7.9% in 1993 (13). The trend observed at our medical center has mirrored the data from the Centers for Disease Control and Prevention, but with a more dramatic rise; VREF isolates were first documented at the Detroit Medical Center in 1991 and now account for approximately 40% of enterococcal isolates. Because VREF is also commonly resistant to all other currently available antibiotics, alternative agents are being aggressively pursued.

Quinupristin-dalfopristin (Q-D; RP 59500; Synercid) is a new water-soluble, semisynthetic antibiotic that is synthesized from natural streptogramin antibiotics and that is combined in a 30:70 ratio. Quinupristin (RP 57669) is derived from pristinamycin IA and is a group B streptogramin (belonging to the macrolide, lincosamide, and group B streptogramin [MLS_B] antibiotic family). Dalfopristin (RP 54469) is derived from pristinamycin IIA and is a group A streptogramin (4, 14). The compounds act synergistically via inhibition of the early and late stages of protein synthesis (16) and display either bacteriostatic or bactericidal activity against enterococci (including VREF), with less intrinsic activity against Enterococcus faecalis than E. faecium (10, 17, 25, 33, 35, 37). Q-D is currently available under a compassionate-use protocol for the treatment of VREF infections; preliminary results indicate that 65.4% of bacteriologically evaluable patients achieved either cure or improvement of their infection (29, 34).

Study of the pharmacodynamic parameters of antimicrobial agents has allowed the optimization of therapy for serious infections (19). Analysis of such parameters as the effects of concentration and exposure time on killing, the area under the concentration-time curve (AUC) relative to the MIC or the minimal bactericidal concentration (MBC) (AUC/MIC or AUC/MBC ratios), and the postantibiotic effect (PAE) has aided in the development of once-daily aminoglycoside regimens that take advantage of the concentration-dependent killing and the long PAEs of these drugs, as well as continuous-infusion regimens for -lactams and vancomycin that optimize the time above the MICs of these agents (6, 26). Tolerance is another pharmacodynamic phenomenon in which the MBC for the organism is markedly elevated relative to the MIC (ratio, 32). This tolerance is often associated with dramatically decreased bactericidal activity and, in the case of -lactam antibiotics, has been hypothesized to be related to deficiencies in autolysin production by the microorganisms (1).

A number of investigations have described the pharmacodynamic properties of Q-D, but most have concentrated on staphylococci. In particular, many studies have evaluated the effects of constitutive MLS_B resistance (conferring resistance to the quinupristin component) on Q-D killing activity. Boswell and colleagues (9) reported that a Staphylococcus aureus strain for which the Q-D MBC was increased displayed phenotypic tolerance and that in time-kill curve studies Q-D killing activity was decreased compared to that for a strain for which the Q-D MBC was not increased. In a separate investigation (8), those investigators reported minimal and nonsignificant trends toward shorter PAEs for S. aureus strains for which MBCs were increased. Fantin et al. (23) reported that Q-D was bactericidal against only 54% of S. aureus strains with constitutive MLS_B resistance, whereas it was bactericidal against 100% of MLS_B-susceptible strains in in vitro time-kill curve studies. Additionally, Q-D (with a targeted combined peak concentration in serum of 6 µg/ml) was significantly more active against MLS_B-susceptible S. aureus than against constitutive MLS_B-resistant S. aureus in a rabbit endocarditis model. In that study, the AUC for Q-D divided by the MIC of quinupristin was the only significant correlate with in vivo activity. Entenza et al. (21) also showed that Q-D had significantly less activity against constitutive MLS_B-resistant S. aureus in an in vivo rat model of endocarditis and reported that killing activity could be improved by increasing the exposure time of the dalfopristin component.

Because few data that substantiate the relationships mentioned above exist for enterococci, we chose to investigate the activity of Q-D against strains of VREF with different Q-D susceptibility profiles. Our objectives were to develop in vitro correlations between Q-D, Quinupristin, and/or dalfopristin susceptibility parameters and antibacterial activity as measured by reductions in inocula, rates of killing, and time to 99.9% killing in time-kill curves studies and by in vitro PAE assessment.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Bacterial strains. Twenty-two clinical isolates of VREF obtained from either blood, wound, or intra-abdominal specimens were used in all susceptibility tests and in the time-kill curve experiments. VREF isolates for which Q-D MBCs ranged from 0.25 to 64 µg/ml were used to determine Q-D PAEs. All VREF isolates were obtained from either the Detroit Receiving Hospital (Detroit, Mich.) or the William Beaumont Hospital (Royal Oak, Mich.).

Antimicrobial agents. Analytical-grade Q-D (lots CB063235 and 9609410), quinupristin (lot P94122V), and dalfopristin (lot P95094) powders were obtained from Rhone-Poulenc Rorer (Collegeville, Pa., and Cedex, France).

Media. Mueller-Hinton broth (Difco, Detroit, Mich.) supplemented with calcium (25 mg/liter) and magnesium (12.5 mg/liter) (SMHB) was used for all microdilution susceptibility tests, time-kill curve experiments, and PAE determinations. Tryptic soy agar (TSA; Difco) plates were used for the plating of experimental samples to determine bacterial densities.

Susceptibility testing. The microdilution MICs and MBCs of each drug were determined with a standard inoculum of 5 × 10⁵ CFU/ml according to the guidelines of the National Committee for Clinical Laboratory Standards (30). The MICs and MBCs were determined in quadruplicate on at least 2 different days, and modal MICs and MBCs were used.

Time-kill curve experiments. All time-kill curve experiments were performed in duplicate with a starting inoculum of 10⁶ CFU/ml. Two to three colonies from an overnight growth of VREF on TSA plates were added to normal saline and were adjusted as necessary to produce a 0.5 McFarland suspension of organisms. This suspension was diluted 1:10 with SMHB, and 0.8 ml was added to 7.2 ml of SMHB to provide the desired starting inoculum. On the basis of the results of previous studies (22), Q-D was added to provide a concentration of 6 µg/ml; this concentration also approximates the lower range of peak concentrations of Q-D in serum obtained in humans (7). As an additional assessment of the effect of concentration on Q-D killing activity, we also evaluated the effects of multiples of the MIC from 1/2 to 48 times the MIC for three VREF isolates for which Q-D MBCs were 0.25, 4, and 64 µg/ml, respectively. Samples (100 µl) were taken at 0 (inoculum control), 2, 4, 8, and 24 h, serially diluted with cold normal saline, and plated in triplicate on TSA plates for determination of the numbers of CFU per milliliter. For situations in which use of the first dilution was necessary for bacterial enumeration, samples were placed on a 0.45-µm-pore-size polysulfone filter (Gelman Sciences, Ann Arbor, Mich.) and washed with cold normal saline, and the filter was then applied aseptically to a TSA plate to minimize the potential effects of antibiotic carryover. Using these methods, we have previously determined our reliable limits of detection to be 100 CFU/ml (32).

PAE. The PAE was determined by methods described by Craig and Gudmundsson (20). The organisms were grown overnight in 10 ml of SMHB to achieve a concentration of ~10⁸ to 10⁹ CFU/ml. The overnight growth was diluted 1:100 with prewarmed SMHB and was incubated for 3 h to allow growth of ~1 log₁₀ CFU/ml and also to achieve the exponential growth phase. One milliliter of this suspension was added to 9 ml of prewarmed SMHB, and Q-D was added to provide a concentration of four times the MIC. Test tubes were sampled just prior to and 1 h after antibiotic addition. After 1 h of exposure to Q-D, the samples were diluted 1:1,000 with prewarmed SMHB to allow for effective removal of Q-D. This 1/400 times the MIC of Q-D did not have any effect on bacterial growth, as determined in preliminary growth experiments. Samples were taken immediately after this dilution and every hour thereafter for up to 7 h. The PAE was calculated by the following equation: PAE = T  C, where T was the time to achieve 1-log₁₀ CFU/ml growth for the antibiotic-exposed sample and C was the time to achieve 1-log₁₀ CFU/ml growth for the untreated control sample. For all PAE experiments, T and C were determined either by linear regression (if R was 0.95) or by visual inspection of the regrowth curve. Each PAE experiment was performed in duplicate to ensure reproducibility.

Statistical analyses. The change in inoculum over 24 h (expressed as the change in the log₁₀ numbers of CFU per milliliter or as a percent reduction in the inoculum from that at the baseline), the time to 99.9% killing, and the rate of killing were determined from the plots of the log₁₀ numbers of CFU per milliliter versus time. The time to achieve 99.9% killing was determined either by linear regression (if was R 0.95) or by visual inspection. The rate of killing was defined as the slope of the killing curve from the start of the experiment to the time of maximal reduction in the log₁₀ numbers of CFU per milliliter. The change in inoculum at 24 h and the killing rates were compared between groups by analysis of variance with Tukey's test for multiple comparisons. The relationships between the Q-D, quinupristin, and dalfopristin susceptibility profiles, pharmacodynamic parameters such as the concentration of Q-D (C_Q-D)/MIC, C_Q-D/MBC, and MIC/MBC ratios and experimental outcomes were evaluated by using stepwise multivariate linear regression analysis. For all comparisons, a P value of 0.05 indicated statistical significance. All statistical analyses were performed with SPSS Statistical Software (release 6.1.3; SPSS, Inc., Chicago, Ill.).

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Susceptibility testing. A summary of the susceptibility profiles for the 22 isolates of VREF is shown in Table 1. The Q-D, quinupristin, and dalfopristin MICs and MBCs ranged from 0.125 to 1 and 0.25 to 64, 8 to 512 and >512, and 2 to 8 and 8 to 512 µg/ml, respectively. For all isolates Q-D MICs were below the expected peak concentrations of Q-D in serum and were under the currently proposed susceptibility breakpoint of 4 µg/ml (5, 34). However, for only 9 of the 22 isolates (41%) were Q-D MBCs less than or equal to the expected peak serum Q-D concentration of 8 to 10 µg/ml (7). There were no significant correlations between the MICs and MBCs of the individual components and the MICs and MBCs of Q-D.

Time-kill curves. Figures 1 and 2 summarize the time-kill curve profiles for the various VREF isolates grouped either by Q-D MBC or by quinupristin MIC. Significant differences were observed between the log₁₀ CFU per milliliter counts at 24 h for all killing curves versus those for the growth controls. Clear visual differences in killing profiles were observed when the results were grouped either by the Q-D MBC or by the quinupristin MIC. Both figures show that activity against VREF appeared to be inversely correlated with the susceptibility parameter evaluated, but the killing curves grouped by quinupristin MIC had much more variability, as evidenced by the larger standard deviations compared with those for the groupings by Q-D MBC. There were significant differences in the killing curve rates between organisms which had Q-D MBCs of 0.25, 0.5, 2, and 4 µg/ml and those with Q-D MBCs of 8 µg/ml (P < 0.05). No noticeable trends in killing curve profiles were noted when the results were grouped by dalfopristin MIC or MBC (graphs not shown). Q-D appeared to have concentration-dependent activity against VREF 580 and VREF 579 (Q-D MICs and MBCs, 0.125 and 0.25 µg/ml and 0.5 and 4 µg/ml, colony counts at 24 h were 2.5 and 2.0 log₁₀ CFU/ml lower at 12 and 2 times the MIC, respectively. In contrast, colony counts at 24 h were only 1 log₁₀ CFU/ml lower at 12 and 2 times the MIC, respectively. In contrast, colony counts at 24 h were only 1 log₁₀ CFU/ml lower for 12 times the MIC than for 2 times the MIC for VREF 595 (Q-D MIC and MBC, 0.5 and 64 µg/ml, respectively).

View larger version (21K): [in this window] [in a new window]   FIG. 1.   Time-kill curves for VREF grouped by Q-D MBCs. GC, growth control.

View larger version (18K): [in this window] [in a new window]   FIG. 2.   Time-kill curves for VREF grouped by quinupristin MICs. GC, growth control.

PAEs. Results from the PAE experiments are presented in Table 2. The PAEs ranged from 0.2 to 3.2 h, with statistically significant differences in PAEs observed over the range of Q-D MBCs tested.

Evaluation of parameters predictive of activity against VREF. By multivariate linear regression analysis, the percent change in the inoculum over 24 h was significantly correlated with the Q-D concentration-to-MBC ratio (R = 0.58; P = 0.02) and was inversely correlated with the quinupristin MIC (R = 0.74; P = 0.008) and the Q-D MBC/MIC ratio (R = 0.68; P = 0.003).

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

The determination of pharmacodynamic properties has become an integral part of the assessment of new antimicrobial agents, and the analysis of older antimicrobial agents has helped to modify previously standardized doses to optimize killing effects on bacteria (6, 19, 26). Since the pharmacodynamics of Q-D against VREF have been incompletely studied and because of the ongoing use of Q-D on a compassionate-use basis for VREF infections, we felt that it was important to better characterize the impact of different Q-D susceptibility profiles first on its in vitro activity.

We found a relatively narrow range of Q-D MICs (0.125 to 1 µg/ml) and a very wide range of Q-D MBCs (0.25 to 64 µg/ml) for our 22 clinical isolates of VREF. For all of our isolates Q-D MICs were at or below the currently proposed sensitivity breakpoint guideline of 1.0 µg/ml (5, 36), but for the majority of strains (59%) Q-D MBCs were 8 µg/ml, which are near the expected peak concentrations in serum obtained from the currently used dose of 7.5 mg/kg of body weight (7). Because Q-D is rapidly eliminated (half-lives of approximately 1 and 0.5 h for quinupristin and dalfopristin, respectively), bactericidal concentrations against VREF would not be likely for significant lengths of time if the current 8- to 12-h dosing intervals were used. Our results were similar to those reported for 20 strains of enterococci, the species of which were not determined (17), and for 92 strains of vancomycin-sensitive E. faecium (VSEF) and 14 strains of VREF, in which Q-D MBCs were 1 µg/ml for 81% of VSEF isolates and Q-D MBCs were >64 µg/ml 57% of VREF isolates (37). As well, our results parallel those for strains of S. aureus with different MLS_B resistance phenotypes, for which Q-D MICs remain relatively constant but for which variations in Q-D MBCs are greater (11, 27, 28, 37). We found a noticeable but nonsignificant trend toward lower Q-D MICs and MBCs for VREF isolates for which quinupristin MICs were lower. Since the vast majority of the isolates were MLS_B resistant, this probably affected our inability to achieve statistically significant correlations.

The results from our killing curve experiments indicated that weak and inconsistent bactericidal activity can be expected against VREF in vivo. We detected bactericidal activity only against the VREF isolates for which MBCs were 4 µg/ml. As discussed previously, due to the in vivo peak concentrations of approximately 10 µg/ml and the rapid elimination of Q-D, bactericidal drug concentrations would not be expected to persist throughout the dosing interval.

Significant relationships existed between the percent reduction in the initial inoculum and the quinupristin MIC, Q-D MBC/MIC ratio, and the Q-D concentration-to-MBC ratio (10, 17, 35). It is intuitive that the closer that a given drug concentration is to being at or above the MBC for an organism, the higher the likelihood of bactericidal activity. Since irreversible binding of Q-D to the ribosome occurs, the actual amounts of Q-D delivered to the bacterial ribosomes (as opposed to the time of persistence at the site) should have the most influence on its antimicrobial activity (2).

The correlations of the reduction in the inoculum and the killing rate with the ratio of the Q-D concentration to the MBC, coupled with the results of the killing curve experiments with multiples of the MICs, suggest concentration-dependent killing activity for Q-D against VREF for which MBCs are 8 µg/ml. These findings are not completely consistent with previously reported data. Hill et al. (25) tested doubling concentrations of Q-D against four VREF isolates and reported slow bactericidal activity with no apparent effect of concentration on the killing profiles. No Q-D MBCs or quinupristin MICs were given, but the killing profiles for their organisms were similar to those for our VREF isolates for which Q-D MBCs were 8 µg/ml (25). For these strains, the Q-D concentration (relative to the Q-D MIC) had a lesser effect on killing compared to the effect on our VREF isolates for which Q-D MBCs were 4 µg/ml. Other investigators have performed killing curve experiments with enterococci (the species were not determined) and reported no difference in killing activity at 8 h over a range of one to eight times the MIC (11). These results likely include data for both E. faecium and E. faecalis, which makes comparison with our results difficult. Caron et al. (12) reported bactericidal activity against seven VREF isolates for which quinupristin MICs were 4 to 8 µg/ml and Q-D MBCs were 1 µg/ml but only bacteriostatic activity against 23 VREF isolates for which quinupristin MICs and Q-D MBCs were 64 µg/ml. These results agree with the correlations that we observed between the quinupristin MIC, Q-D MBC, and activity against VREF.

Interestingly, it appears that these correlations hold true for both E. faecium and S. aureus. The killing rate and the level of reduction of the inoculum were both lower for a strain of S. aureus for which the Q-D MBC was elevated (16 µg/ml) compared those for a strain for which the MBC was 1 µg/ml, while Q-D had bactericidal activity against 6 of 6 erythromycin-susceptible strains of S. aureus but only 7 of 13 strains of S. aureus with constitutive resistance to erythromycin (quinupristin) (9, 23). A reduction in the inoculum in the killing curve studies was less for three strains for which Q-D MBCs were 16, 8, and 2 µg/ml, respectively, than for two strains for which MBCs were 1 µg/ml, and these trends in killing were also observed in rabbit models of endocarditis (21, 23).

Many articles have stated that Q-D possesses a substantial PAE against most gram-positive bacteria (including enterococci), but actual published data are extremely limited (10, 24). In vivo PAEs (which also include the effects of sub-MICs of Q-D and host factors) have been reported to be as long as 9 to 10 h against Streptococcus pneumoniae and S. aureus (18). A mean in vitro PAE of 2.4 h was reported for both 3 strains of enterococci and 10 strains of S. aureus at the MIC of Q-D and increased to >5.5 h at both 3 and 10 times the MIC, and these were independent of erythromycin susceptibility status (11). Q-D at concentrations of 0.5, 1, and 2 times the MIC produced PAEs of 3, 4.5, and 8.75 h, respectively, against one VREF isolate with an unknown susceptibility profile (3). In contrast to these reports, we found a range of appreciably shorter Q-D PAEs (from 0.2 to 3.2 h) for our clinical strains of VREF at a Q-D concentration of 4 times the MIC. The reasons for the reduced PAEs for our isolates are unclear, but unreported differences in quinupristin or Q-D susceptibility profiles between the strains are a likely explanation.

The effects of concentration, exposure time, and susceptibility on the Q-D PAE have been studied more extensively for S. aureus. The Q-D PAE has been reported to be longer against MLS_B-sensitive and inducibly MLS_B-resistant strains of S. aureus and S. epidermidis than against constitutive MLS_B-resistant strains (15, 31). In another report, the PAEs in pooled human serum were similar for six strains of S. aureus for which MBCs were elevated, whereas they were not for six strains for which MBCs were not elevated, but significant variations were noted within the groups (8). The differences in the PAEs for our VREF strains appeared to be related to the Q-D MBC and quinupristin MIC but were best correlated with the Q-D concentration-to-MBC ratio. For VREF isolates for which MBCs were 16 µg/ml, the PAE was minimal (1 h). Because the Q-D MBCs for the majority of our isolates were at or above 16 µg/ml, reliance on the Q-D PAE to prolong dosing intervals might not be prudent.

In conclusion, we observed correlations between the in vitro activity of Q-D against VREF and the pharmacodynamic parameters that incorporate either the quinupristin MIC or the Q-D concentration relative to the MBC. The quinupristin MIC may be the most useful parameter for predicting pharmacodynamic activity against both E. faecium and S. aureus due to the relative ease and rapidity of determining quinupristin MICs in the clinical laboratory compared to the ease and rapidity of determining quinupristin MICs in the clinical laboratory compared to the ease and rapidity of MBCs. Analyses of treatment successes and failures in the VREF compassionate-use program that incorporate the Q-D PAEs, quinupristin MICs, and/or Q-D concentration-to-MBC ratios for the recovered organisms could allow additional insights into Q-D pharmacodynamics against VREF.