Prevalence and Characterization of Invasive Isolates of Streptococcus pyogenes with Reduced Susceptibility to Fluoroquinolones

Abstract

Fluoroquinolone susceptibility testing was performed on invasive group A streptococcus isolates from 1992-1993 and 2003 from Ontario, Canada. None were nonsusceptible to levofloxacin. Two of 153 (1.3%) from 1992-1993 and 7 of 160 (4.4%) from 2003 had a levofloxacin MIC of 2 μg/ml; all nine had parC mutations, and eight were serotype M6.

Ciprofloxacin and levofloxacin have only moderate microbiologic activity against group A streptococci (GAS), with MICs at which 90% of the isolates tested are inhibited (MIC₉₀s) of 1.0 μg/ml and 0.5 μg/ml, respectively (2). Previous reports of resistance have, however, been rare. In a nationwide, multicenter susceptibility surveillance study in Spain by Pérez-Trallero et al. 70 (3.4%) of 2,039 isolates of Streptococcus pyogenes had a ciprofloxacin MIC of ≥4 μg/ml and were found equally in children and adults (7). There have been only three reports of clinical failures in association with resistance (8, 9, 12).

The goal of this study was to determine whether there has been an increase in the frequency of GAS with elevated MICs to fluoroquinolones from an ongoing population-based surveillance program in Ontario, Canada, and to describe the clinical and molecular characteristics of such isolates.

A detailed description of the methodology of our population-based surveillance has been published previously (4). In summary from 1 January 1992 until 31 December 2003 all microbiology laboratories in Ontario (population approximately 12.3 million on 1 January 2004) reported and sent sterile-site GAS isolates to a central study office (Mount Sinai Hospital/Toronto Medical Laboratories, Toronto, Canada). Isolates were confirmed as GAS by standard techniques. A total of 160 isolates from 1992-1993 and 2003 were randomly selected from the database representing 49.7 and 51.6% of the total pool of isolates from the two time periods, respectively. In vitro susceptibility testing was performed by broth microdilution by standard NCCLS methodology (6). Isolates with a levofloxacin MIC of ≥1 μg/ml were examined for active efflux as described previously (3). Strains for which there was a fourfold or greater decrease in the ciprofloxacin MIC in the presence of reserpine were considered in this study to be positive for reserpine-inhibited efflux. Typing was performed by precipitation at the National Centre for Streptococcus in Edmonton, Alberta, Canada (5, 10).

The quinolone resistance-determining regions (QRDRs) of parC of all strains with a levofloxacin MIC of ≥1 μg/ml and 10 GAS isolates from both 1992-1993 and 2003 with lower levofloxacin MICs of ≤0.5 μg/ml were amplified by PCR and sequenced as previously described (3). We analyzed the QRDR of gyrA for all strains with a levofloxacin MIC of ≥2 μg/ml. Primers were based on previously published sequences (12). DNA sequencing was performed by ABI Prism Big Dye terminator cycle sequencing with the ABI 377 automated sequencer (PE Applied Biosystems, Mississauga, Ontario, Canada). Nucleotide and amino acid sequence comparisons were performed with the multiple-alignment sequence function of Vector NTI Suite software (InfoMax Inc., Bethesda, Md.). The GenBank accession numbers for the wild-type sequences used for comparison purposes were AF220946 for parC and AF220945 for gyrA. We preformed a phylogenetic tree calculation based on a sequence distance method and utilized a neighbor-joining algorithm with the deduced amino acid sequences of the QRDR of parC to determine relatedness (11).

As the initial analysis of levofloxacin MICs demonstrated that the majority of GAS isolates with a levofloxacin MIC of 2 μg/ml were serotype M6, a further sample of all M6 isolates within the database was also tested for in vitro antimicrobial susceptibility.

Of the 320 selected isolates, 153 of 160 isolates from 1992-1993 and all isolates from 2003 were viable. The isolates from the two time periods were similar with respect to age, site of isolation, outcome, and frequency of streptococcal toxic shock syndrome in infected patients. There were, however, significantly more females in 2003 (45.8% versus 57.5%: P ≤ 0.05), as well significant variations in the M types of isolates from the two populations. Compared to 2003, the sample from 1992-1993 had significantly fewer isolates with the M types 1, 5, and 6 and significantly more isolates with M types 22 and 28 or that were nontypeable.

Of the 153 isolates from 1992-1993, 14 (9.2%) isolates had a levofloxacin MIC of 1 μg/ml, and 2 (1.3%) isolates had a levofloxacin MIC of 2 μg/ml. In the sample from 2003 there were 15 (9.4%) isolates with a levofloxacin MIC of 1 μg/ml and 7 (4.4%) with a levofloxacin MIC of 2 μg/ml. There were no isolates in either year considered nonsusceptible by NCCLS interpretive criteria (levofloxacin MIC of ≥4 μg/ml). There was no significant difference between samples in the number of isolates with a levofloxacin MIC of ≥1 μg/ml or MIC of 2 μg/ml (P ≥ 0.05). Of the nine isolates with a levofloxacin MIC of 2 μg/ml the MICs to other fluoroquinolones measured were as follows: all had a ciprofloxacin MIC of 2 μg/ml, all had a gatifloxacin MIC of 0.5 μg/ml, and the moxifloxacin MICs ranged from 0.25 to 0.5 μg/ml. Among the 29 isolates with a levofloxacin MIC of 1 μg/ml the ciprofloxacin MIC range was 0.5 to 2.0 μg/ml (MIC₉₀ = 2 μg/ml), the gatifloxacin MIC range was 0.25 to 0.5 μg/ml (MIC₉₀ = 0.5 μg/ml), and the moxifloxacin MIC range was 0.12 to 0.5 μg/ml (MIC₉₀ = 0.5 μg/ml). None of the 38 isolates with a levofloxacin MIC of ≥1 μg/ml had a change in their ciprofloxacin MIC after repeat susceptibility testing in the presence of reserpine.

The demographic and clinical characteristics of isolates with a levofloxacin MIC of 2 μg/ml (n = 9) and ≤1 μg/ml (n = 304) were similar other than differences in frequency of specific M types. There was no difference between groups with respect to site of isolation, age of patients, or outcomes. There were more isolates with an M serotype 6 in the group with a levofloxacin MIC of 2 μg/ml (88.9% versus 0.3%, P ≤ 0.0001) and more isolates with an M type of 1 (0% versus 37.2%, P ≤ 0.05) in the group with a levofloxacin MIC of ≤1 μg/ml.

The parC mutations identified from isolates with different levofloxacin MICs are displayed in Table 1. Of the nine isolates with levofloxacin MICs of 2 μg/ml all had a parC mutation, compared to only 16 (32.7%) of 49 isolates in the group with levofloxacin MICs of ≤1 μg/ml (P ≤ 0.001). There were no gyrA mutations identified in isolates with a levofloxacin MIC of 2 μg/ml. The S-79-A mutation occurred in 100% of isolates with a levofloxacin MIC of 2 μg/ml, compared to 4.1% of isolates with a levofloxacin MIC of ≤1 μg/ml (P ≤ 0.001). Of the nine isolates with a levofloxacin MIC of 2 μg/ml, eight (88.9%) were M6, and of the nine M6 isolates only one had a MIC of <2 μg/ml. This isolate had a levofloxacin MIC of 1 μg/ml and possessed the S-79-A mutation. There was no significant difference in the frequency of other parC mutations between isolates with levofloxacin MICs of 2 μg/ml and isolates with a levofloxacin MIC of ≤1 μg/ml (P > 0.05). The mutations other than the S-79-A mutation are not likely relevant mutations affecting the fluoroquinolone MIC. The base pair homology of the sequenced QRDR of the parC gene for the 58 isolates sampled revealed that all nine M6 isolates are very closely genetically related even though they were not clustered by hospital, region, or time of year.

TABLE 1.

parC mutations of isolates with different levofloxacin MICs

Levofloxacin MIC (μg/ml) (no. tested)	parC mutation(s)	No. (%) with mutation(s)
2 (9)	S(Ser)-79-A(Ala)	9 (100)
1 (29)	S(Ser)-79-A(Ala)	2 (6.9)
	D(Asp)-91-N(Asn), S(Ser)-140-P(Pro)	2 (6.9)
	S(Ser)-140-P(Pro)	1 (3.4)
	S(Ser)-80-P(Pro), D(Asp)-91-N(Asn)	1 (3.4)
	A(Ala)-121-V(Val)	1 (3.4)
≤0.5 (20)	A(Ala)-121-V(Val)	6 (30)
	D(Asp)-91-N(Asn)	2 (10)
	D(Asp)-91-N(Asn), S(Ser)-140-P(Pro)	1 (5)

There have been 100 episodes of invasive GAS due to M6 isolates from 1 January 1992 to 1 January 2004. Ninety-two of the 100 M6 isolates had in vitro susceptibility testing done; nine (9.8%) had a levofloxacin MIC of 1 μg/ml, and 83 (90.2%) had a levofloxacin MIC of 2 μg/ml. In 2003 there were 310 cases of invasive GAS in Ontario, Canada. The M type was available for 296 (95.4%) isolates, of which 14 (4.7%) were M6. The annual incidence of invasive GAS disease due to M6 isolates is demonstrated in Fig. 1.

FIG. 1.

Annual incidence of all invasive GAS infection due to M6 isolates (bars) compared with annual incidence of invasive GAS due to all M types (line, secondary axis) in Ontario, Canada, from 1992 through 2003.

Molecular analyses of the fluoroquinolone-nonsusceptible isolates from recent case reports all demonstrate an amino acid substitution at position 79 of parC. Two of the isolates had a serine-to-phenylalanine substitution, while one had a serine-to-tyrosine substitution. Molecular analysis of an isolate from Spain with a levofloxacin MIC of 2 μg/ml demonstrated the identical S-79-A mutation as in our M6 clone (1).

The emergence of GAS isolates with elevated levofloxacin MICs was not associated with the same risk factors for fluoroquinolone resistance as seen in pneumococci. There was no association with respiratory isolates or age greater than 65, where the majority of fluoroquinolone use occurs. Of the nine isolates with an elevated levofloxacin MIC none were from individuals greater than 65 years of age and three were from individuals less than 10 years of age, where fluoroquinolones are used very infrequently (data not shown). The annual incidence of invasive infection due to M6 isolates had also varied over time despite constantly increasing fluoroquinolone use in Ontario.

Despite the frequent use of fluoroquinolones to treat respiratory tract infections in Ontario, we have not seen the emergence of resistance among GAS. The majority of isolates with reduced susceptibility to the fluoroquinolones can be accounted for by a stable clone that first emerged in 1992-1993 and remains present at low levels today.