Clinical-Use-Associated Decrease in Susceptibility of Vancomycin-Resistant Enterococcus faecium to Linezolid: a Comparison with Quinupristin-Dalfopristin

Abstract

The susceptibility of 135 vancomycin-resistant Enterococcus faecium bacteremic isolates to linezolid and quinupristin-dalfopristin was determined. All were susceptible to linezolid, while 88% were susceptible to quinupristin-dalfopristin prior to the clinical use of the drugs at our hospital. More than 6 months after their clinical use, a decrease in susceptibility was noted for only linezolid at 83%. This was related in part to a single G2576U gene mutation in domain V of the 23S rRNA gene.

The last decade was marked by the emergence of vancomycin-resistant enterococci (VRE), which are associated with serious bloodstream infections involving critically ill and immunocompromised patients (3, 4). Given the multidrug-resistant profile of VRE organisms, treatment options based on conventional antimicrobial agents were limited during most of the 1990s (11). At the turn of the century, two novel antimicrobial agents, quinupristin-dalfopristin (Q-D) and linezolid, were approved as therapeutic options for vancomycin-resistant Enterococcus faecium (VREF) (1, 6, 10).

In this study, we examined susceptibility trends for linezolid and Q-D associated with their clinical use. In addition, isolates with decreased linezolid susceptibility were screened for the presence of the point mutation most commonly associated with linezolid resistance.

(This material was presented in part as an abstract at the 41st Annual Meeting of the Infectious Diseases Society of America, 9 to 12 October 2003, San Diego, Calif.)

Between January 1996 and January 2002, 153 VRE blood isolates in cancer patients were obtained from the clinical microbiology laboratory at the M. D. Anderson Cancer Center. The susceptibility testing included only the 135 VREF isolates. Q-D was placed on our hospital formulary and introduced into clinical use in September 1999, and linezolid was added in May 2000. The susceptibility patterns of the two drugs were compared during three phase periods: (i) preuse, that is, before the use of Q-D or linezolid; (ii) early use, which included the first 6 months immediately after the introduction of either antibiotic into clinical use at our hospital; and (iii) postuse, which was the period following at least 6 months of antibiotic use in the hospital, up to February 2002. The amount of each antibiotic used during each time period was calculated in terms of the total number of adult defined daily doses (DDDs). The single adult DDD is the assumed average maintenance dose of a drug used per day for its main indication in adults, which is a technical unit of measurement and does not necessarily reflect the actual amount or dose used. We obtained from the pharmacy the total quantity of each of the two drugs dispensed during the entire study period to calculate the DDD of each drug used. We used 1,200 mg/day as the daily dose of linezolid and 1,500 mg/day as that of Q-D.

Susceptibility testing was performed by using a microtiter broth dilution method that was in accordance with the guidelines established by the National Committee for Clinical Laboratory Standards. The MIC cutoff points for linezolid were based on the package insert (≤2 mg/liter, susceptible; 4 mg/liter, intermediate; ≥ 8 mg/liter, resistant). Similarly, the MIC cutoff points for Q-D were as follows: ≤1 mg/liter, susceptible; 2 mg/liter, intermediate; ≥4 mg/liter, resistant.

Pulsed-field gel electrophoresis (PFGE) was performed as previously described (12) on all of the VREF bacteremic isolates. VREF isolates that were indistinguishable by PFGE were considered to belong to the same clone. Isolates were considered closely related if their PFGE patterns differed by only two or three bands. VREF strains that demonstrated decreased susceptibility to linezolid (intermediate or resistant) were studied for the G2576U mutation in domain V of the 23S rRNA as previously described (8, 9). We examined clones to determine the number of genes that have the mutation. We used a technique that was previously published and that was shown to produce results that correlated well with data obtained from restriction digestion (8, 9). Internal fragments of the 23S rRNA genes were amplified by PCR, ligated to a multicopy vector, and transformed into Escherichia coli DH10B. Inserts from 24 transformants were amplified with forward and reverse primers, and the products were digested with MaeI. The percentage of transformants digested with this enzyme was used to determine the number of the six 23S rRNA genes with the G2576U mutation as previously reported (9).

Most (135 of 153, 88%) of the VRE bacteremic isolates were VREF. There were 18 (12%) vancomycin-resistant E. faecalis bacteremic isolates, all of which were susceptible to linezolid and none of which were susceptible to Q-D. Nine (50%) of the vancomycin-resistant E. faecalis isolates had intermediate susceptibility to Q-D, while the other nine were resistant.

During the early use period, the DDDs of Q-D and linezolid were 230 and 1,045, respectively. However, during the postuse period, the DDDs of Q-D and linezolid were 1,270 and 3,000, respectively (Fig. 1).

FIG. 1.

Clinical use (in DDD) of linezolid and Q-D in the first 6 months after their introduction (early use) and more than 6 months after their introduction into the clinical service (postuse).

During the preuse period, all (100%) of the VREF bacteremic clones (115 isolates classified into 76 clones) were susceptible to linezolid, whereas during the postuse period, the susceptibility of VREF decreased significantly to 83% (10 of 12 clones).

In addition, during the preuse period, the susceptibility of VREF to Q-D was 88% (59 of 67 clones susceptible), which was lower than that of linezolid (100% susceptibility to linezolid). During the postuse period, more than 6 months after the introduction of the two antibiotics, the susceptibility to Q-D was 96% (22 of 23 susceptible isolates) while the susceptibility to linezolid was 83% (10 of 12 susceptible isolates) (Fig. 2).

FIG. 2.

Percent susceptibility of nonidentical VRE strains to linezolid or Q-D before its clinical use (preuse) and more than 6 months after their introduction into the clinical service (postuse).

Of the 12 VREF clones identified during the linezolid postuse period, 2 showed decreased (intermediate) susceptibility to linezolid (MIC, 4 mg/liter). When studied for the G2576U mutation, one of the two clones showed a mutation involving one of the six 23S rRNA genes in domain V. The other clone did not appear to have the G2576U mutation in any of the six 23S rRNA genes or any other mutation in the amplified region (9). The two clones with intermediate susceptibility to linezolid were associated with nosocomial bacteremia during the postuse period, whereby the bloodstream infection occurred more than 7 days after hospital admission.

Several factors could have contributed to the wider use of linezolid: (i) the broader spectrum of linezolid effectiveness against vancomycin-resistant E. faecalis and VREF during the preuse period (Fig. 2), (ii) its ease of administration (oral availability), and (iii) its better safety profile (1, 6, 10). However, the heavy use of linezolid was subsequently associated with the development of decreased VREF susceptibility to this antibiotic.

Gonzalez et al. reported decreased susceptibility of VREF to linezolid associated with protracted treatment with this antibiotic in five patients, four of whom were transplant patients (5). In the present study, the two VREF bacteremic clones with decreased susceptibility to linezolid were identified as nosocomial strains that were closely related to VREF strains that were susceptible to linezolid when identified in the linezolid preuse period.

The fact that the frequency of VREF susceptibility to Q-D did not decrease with its clinical use could be related to its limited use at our hospital, possibly due to adverse events (14). In addition, Q-D is an antibiotic made of the synergistic combination of two streptogramins, 70% dalfopristin, or streptogramin A, and 30% quinupristin, or streptogramin B. Dalfopristin binds to the ribosomal 50S site, changing its conformation and increasing the ability of quinupristin to bind to the bacterial ribosome. Although resistance to either of the two streptogramin components could develop separately, resistance to both the A and B components may result in a higher level of resistance (B. Bozdogan and R. Leclercq, 39th Intersci. Conf. Antimicrob. Agents Chemother., abstr. 841, 1999). Studies have associated the emergence of VREF resistance to Q-D with extensive use of virginiamycin as an animal feed antibiotic rather than hospital use (15, 16). However, emergence of resistance to Q-D during therapy has also occurred in the clinical setting (2, 5, 7, 17).

The single G2576U mutation involving one of the six 23S rRNA genes associated with decreased susceptibility to linezolid is consistent with previous reports (9, 13). Marshall et al. recently demonstrated a direct positive correlation between the number of 23S rRNA genes with a G2576U mutation and the level of resistance of VREF to linezolid in clinical isolates (9). Hence, this single gene mutation associated with decreased (intermediate) susceptibility to linezolid could be the prelude to higher levels of oxazolidinone resistance associated with more extensive clinical use of this antibiotic, selecting for the G2576U mutation in domain V of the 23S rRNA gene. This single gene mutation was detected in one organism. However, one cannot rule out the possibility that the decreased susceptibility of the other organism could be secondary to gene conversion associated with prolonged selective antibiotic pressure.