Evaluation of Surrogate Disk Tests for Detection of Ciprofloxacin and Levofloxacin Resistance in Clinical Isolates of Salmonella enterica

Abstract

Detection of fluoroquinolone resistance in Salmonella enterica has become increasingly difficult due to evolving resistance mechanisms to this antimicrobial class in this organism. We evaluated two quinolone disks and five fluoroquinolone disks for their ability to act as a surrogate agent for the detection of fluoroquinolone resistance in a collection of 136 S. enterica isolates, including 111 with intermediate or resistant ciprofloxacin MICs mediated by a variety of resistance mechanisms. Ciprofloxacin, ofloxacin, and pefloxacin disks detected all isolates resistant to ciprofloxacin (0% very major error) and yielded false resistance (major error) in 8, 4, and 12% of susceptible isolates, respectively. Ciprofloxacin and pefloxacin provided clearer differentiation of susceptible and resistant isolates.

INTRODUCTION

Fluoroquinolones are an important class of antimicrobial agents for the treatment of typhoid fever (1). However, defining an optimal method for the detection of fluoroquinolone resistance in Salmonella enterica that is practical for laboratories in developing countries, where typhoid fever is endemic, has proven difficult. The Clinical and Laboratory Standards Institute (CLSI) has published several revisions to the M100 document recommendations for the detection of fluoroquinolone resistance in Salmonella over the past 3 years. These changes were driven by the following: (i) the recognition that several emerging fluoroquinolone resistance mechanisms are not detected by traditional phenotypic methods; (ii) reevaluation of fluoroquinolone pharmacokinetics and pharmacodynamics; and (iii) several excellent reports documenting fluoroquinolone treatment failures in patients infected with an isolate with low-level fluoroquinolone resistance (2).

The most common mechanism of fluoroquinolone resistance in clinical isolates of S. enterica is mutation to the quinolone resistance-determining regions (QRDR) of the gyrA gene. This mutation is associated with ciprofloxacin MICs above those found in wild-type strains (MIC > 0.06 μg/ml) and nalidixic acid resistance (MIC ≥ 32 μg/ml). Less frequently, fluoroquinolone resistance can occur following mutation to the QRDR of gyrB or topoisomerase genes parC and parE or by acquisition of plasmid-mediated quinolone resistance genes, such as qnr or aac(6′)-Ib-cr (2). Resistance via these latter pathways confers elevated MICs, but not resistance, to nalidixic acid, which was recommended as an indicator of fluoroquinolone resistance for S. enterica by the CLSI prior to 2012.

In order to resolve this testing issue, the CLSI revised the ciprofloxacin, levofloxacin, and ofloxacin MIC breakpoints for Salmonella to better reflect the wild-type distribution of Salmonella MICs to these antimicrobials (Table 1) (3). The European Committee on Antimicrobial Susceptibility Testing (EUCAST) independently published Salmonella-specific MIC breakpoints for ciprofloxacin, and the U.S. Food and Drug Administration (FDA) revised the ciprofloxacin MIC breakpoints to match those of the CLSI, but for S. enterica serovar Typhi alone (Table 1). CLSI established Salmonella disk diffusion breakpoints for ciprofloxacin (3), but levofloxacin and ofloxacin disk breakpoints have not yet been established, although these have been proposed by Sjolund-Karlsson and colleagues (4, 5).

TABLE 1.

Current CLSI, EUCAST, and FDA fluoroquinolone breakpoints for Salmonella spp.

Antimicrobial agent and source of breakpoint criteria	Disk diffusion diameter (mm) for categorya:			MIC (μg/ml) for category:a
	S	I	R	S	I	R
Ciprofloxacin
CLSI	≥31	21–30	≤20	≤0.06	0.12–0.5	≥1.0
EUCASTb				≤0.06		>0.06
FDA (S. Typhi only)	≥31	21–30	≤20	≤0.06	0.12–0.5	≥1.0
Levofloxacin, c CLSI				≤0.12	0.25–1.0	≥2.0
Ofloxacin, d CLSI				≤0.12	0.25–1.0	≥2.0
Surrogate/screening agents for detection of fluoroquinolone resistance
Nalidixic acid, CLSI	≥19	14–18	≤13	≤16		≥32
Pefloxacin
CLSI	≥24		≤23
EUCAST	≥24		<24

^aS, susceptible; I, intermediate; R, resistant.

^bA 5-μg pefloxacin disk was used to detect resistance in Salmonella spp.

^cDisk diffusion zone diameters proposed by Sjolund-Karlsson et al. (4): ≥28 mm, susceptible; 19 to 27 mm, intermediate; ≤18 mm, resistant.

^dDisk diffusion zone diameters proposed by Sjolund-Karlsson et al. (4): ≥25 mm, susceptible; 16 to 24 mm, intermediate; ≤15 mm, resistant.

Several laboratories both in developing countries, where enteric fever is endemic and fluoroquinolone resistance is prevalent but MIC testing is infrequently performed, and developed countries, where this disease is less common, have noted difficulties in interpreting ciprofloxacin disk results (2). In order to address this concern, both EUCAST and CLSI now recommend as an option the use of a 5-μg pefloxacin disk test as a surrogate marker by which to detect fluoroquinolone resistance in Salmonella (Table 1). Studies performed in Europe and the United States and presented to the CLSI Microbiology Subcommittee on Antimicrobial Susceptibility Testing in July 2014 demonstrated pefloxacin zone diameters of ≥24 mm to be an excellent indicator of wild-type isolates (i.e., those without fluoroquinolone resistance mechanisms). This report summarizes the studies performed in the United States to identify pefloxacin as a surrogate agent for the detection of fluoroquinolone resistance in S. enterica. Performance of five other quinolone/fluoroquinolone and nalidixic acid disks were also evaluated for their abilities to detect fluoroquinolone resistance in Salmonella spp.

MATERIALS AND METHODS

Isolates.

A collection of 136 S. enterica isolates, including 29 S. enterica serovar Typhi isolates, 2 S. enterica serovar Paratyphi A isolates, 1 S. enterica serovar Paratyphi B isolate, and 104 S. enterica isolates of nontyphoid serovars were included in these studies (Table 2). Fluoroquinolone resistance mechanisms in these isolates were predetermined by molecular methods, using PCR for the specific targets as in the accompanying article by Skov et al. (6). Twenty-four isolates had no resistance genes, 37 isolates harbored a qnr gene, 1 isolate had an aac(6′)-Ib-cr gene, and 45 isolates had mutations in the QRDR region of the gyrA topoisomerase gene. An additional 29 isolates were not characterized for resistance mechanism but were S. enterica serovar Typhi with the ciprofloxacin-intermediate, nalidixic acid-resistant phenotype that is typically associated with mutation to the QRDR. Twenty-five isolates were ciprofloxacin susceptible (18%), 91 (65%) were ciprofloxacin intermediate, and 20 (18%) were ciprofloxacin resistant (5) (Fig. 1). The isolates with ciprofloxacin-intermediate MICs are the same isolates evaluated in the accompanying article by Skov and colleagues (6).

TABLE 2.

S. enterica isolates included in this study

No. of isolates (% typhoidal)	Resistance mechanism	Ciprofloxacin MIC (μg/ml)			Nalidixic acid MIC (μg/ml)
		Range	50%	90%	Range	50%	90%
1 (0)	aac(6′)-Ib-cr	1.0	1.0	1.0	32	32	32
37 (0)	qnr	0.12–1.0	0.5	1.0	4.0–32	32	32
45 (0)	QRDR mutation	0.06–0.5	0.25	0.25	>128	>128	>128
29 (90)	Not characterized	0.12–16	0.5	16	128– >128	>128	>128
24 (25)	None	≤0.008–0.06	0.015	0.03	2–16	4	16

FIG 1.

Distribution of ciprofloxacin MICs among S. enterica isolates studied. Isolates with MICs in the intermediate (0.12 to 0.5 μg/ml) range by CLSI M100-S25 breakpoints are indicated by the shaded box.

Susceptibility testing.

Disk diffusion was performed according to CLSI standards (3). Briefly, a suspension of the test organism equivalent to a 0.5 McFarland standard was prepared in saline from three to five well-isolated colonies of each isolate grown overnight on a blood agar plate (BBL; Becton Dickinson [BD], Sparks, MD). Using a swab, the organism was inoculated onto a Mueller-Hinton agar (MHA) II plate (BBL), and disks were applied. Disks tested include five fluoroquinolones (all from BD), ciprofloxacin (5 μg), enoxacin (10 μg), levofloxacin (5 μg), norfloxacin (10 μg), and ofloxacin (5 μg), and two quinolones, nalidixic acid (BD) (30 μg) and pefloxacin (Oxoid; Thermo Fisher, Reinach, Switzerland) (5 μg). The plates were incubated for 18 h at 35°C, and zones were read visually, using reflected light, by two independent readers. If discrete colonies were present within the zone of inhibition, testing was repeated. If colonies within the zone were observed a second time, the colony-free zone was measured. Using the same inoculum suspension, isolates were tested by broth microdilution (BMD) for ciprofloxacin and levofloxacin MICs on panels prepared in-house according to CLSI standards using cation-adjusted Mueller-Hinton broth (Difco, BD). Ciprofloxacin and levofloxacin concentrations ranged from 0.008 μg/ml to 16 μg/ml. Quality control was assessed using Escherichia coli ATCC 25922, Pseudomonas aeruginosa ATCC 27853, Staphylococcus aureus ATCC 29213, and Enterococcus faecalis ATCC 29212 (MIC only). The quality control results for all MIC and disk diffusion tests were within acceptable quality control ranges (3).

The precision of pefloxacin disk results across three manufacturers of disk was evaluated, using 72 isolates with ciprofloxacin MICs in the intermediate range, as these isolates are the most problematic for laboratory testing (2). Disk diffusion was performed as described above, and 5-μg pefloxacin disks from three manufacturers (BD, Mast [Merseyside, United Kingdom], and Oxoid) were placed on the same BBL MHA plate. The plates were incubated for 18 h and read as described above.

Data analysis.

Disk diffusion results were evaluated to determine whether any of the seven disks tested could serve as a surrogate agent for the detection of fluoroquinolone resistance in S. enterica. For the purpose of this study, fluoroquinolone resistance was defined as not susceptible to ciprofloxacin (MIC ≥ 0.12 μg/ml), using the current CLSI breakpoints for Salmonella (3). The definition for the CLSI intermediate category implies clinical efficacy when a higher than normal dosage of a drug can be used (3); however, isolates with intermediate MICs (0.12 to 0.5 μg/ml) were lumped together with those demonstrating resistant MICs, because there are no data to indicate that ciprofloxacin monotherapy is efficacious against isolates with ciprofloxacin MICs in the intermediate range (0.12 to 0.5 μg/ml), regardless of dose (2). Therefore, the breakpoint for each of the seven “surrogate” agents was defined as the zone diameter value that most reliably differentiated Salmonella isolates that were susceptible versus not susceptible to ciprofloxacin, based on the assumption that ciprofloxacin MICs determined by BMD could reliably differentiate fluoroquinolone-susceptible from non-fluoroquinolone-susceptible isolates. The performance of three disks (ciprofloxacin, nalidixic acid, and pefloxacin) to act as surrogate agents for levofloxacin MIC was also evaluated in the same manner. Very major errors (VME; i.e., number of isolates that were falsely not susceptible/total number of non-ciprofloxacin-susceptible isolates) and major errors (ME; i.e., number of isolates that were falsely not susceptible/total number of ciprofloxacin-susceptible isolates) were calculated. Overall categorical agreement (susceptible or not susceptible) was also calculated.

In addition, VME, ME, and minor errors (mE; i.e., one result is intermediate, and the other is either susceptible or resistant) were calculated in the traditional manner, using the current Enterobacteriaceae susceptible, intermediate, and resistant breakpoints for each disk surrogate agent, with the exception of ofloxacin and levofloxacin, where proposed Salmonella disk diffusion breakpoints were used (4), and pefloxacin, for which no intermediate breakpoint exists.

RESULTS

Evaluation of a surrogate disk based on ciprofloxacin MIC results.

Scattergrams for each of the seven disks tested, compared to ciprofloxacin MICs, are presented in the figures in the supplemental material. VME and ME calculations, using a “susceptible” versus a “not susceptible” definition as described in Materials and Methods, are presented in Table 3. No VME were noted for ciprofloxacin (Table 3), and 8% ME were observed. It should be noted that only 25 ciprofloxacin-susceptible isolates were tested and that 8% ME represents two isolates. One ME was for an isolate with a QRDR mutation but with a ciprofloxacin MIC of 0.06 μg/ml and inhibition zone diameter of 27 mm. This is the only isolate in the study where a ciprofloxacin-susceptible MIC was found in an isolate with a fluoroquinolone resistance mechanism. The second ME was for an isolate with no resistance mechanism with a ciprofloxacin MIC of 0.03 μg/ml and inhibition zone diameter of 30 mm.

TABLE 3.

Performance of surrogate disks for the differentiation of ciprofloxacin-susceptible versus non-ciprofloxacin-susceptible MICs in Salmonella spp.^a

Surrogate agent and not susceptible breakpointb	Performance (%)c
	VME	ME	CA
Ciprofloxacin, ≤30 mm	0.0	8.0	98.5
Levofloxacin, ≤27 mmd	0.0	4.0	99.3
Ofloxacin, ≤24 mmd	0.0	4.0	99.3
Nalidixic acid, ≤18 mm	1.8	20.0	94.9
Enoxacin
≤17 mm	53.1	0.0	56.6
≤23 mme	0.0	4.0	99.3
Norfloxacin
≤16 mm	90.9	0.0	25.7
≤27 mme	0.9	8.0	97.8
Pefloxacin, ≤23 mm	0.0	12.0	97.8

^aA total of 111 isolates were ciprofloxacin intermediate or resistant; 25 isolates were ciprofloxacin susceptible.

^bCurrent CLSI not susceptible disk breakpoints for Enterobacteriaceae (nalidixic acid, enoxacin, and norfloxacin) or Salmonella (ciprofloxacin) were used.

^cVME, very major errors; ME, major errors; CA, categorical agreement.

^dIntermediate (corresponds to not susceptible) breakpoint proposed by Sjolund-Karlsson et al. (4).

^eAlternative not susceptible breakpoint, proposed in this study.

When evaluated in the traditional manner to include intermediate breakpoints, 0 VME, 0 ME, and 11.0% mE (15 isolates) were observed between ciprofloxacin disk and ciprofloxacin MIC results. The majority of these mE (n = 11) were in isolates with a ciprofloxacin MIC of 1.0 μg/ml and ciprofloxacin disk diffusion zones in the intermediate range (21 to 30 mm [see Fig. S1 in the supplemental material]).

The ofloxacin and levofloxacin disks were evaluated using proposed not susceptible disk breakpoints of ≤24 mm and ≤27 mm (4), respectively. No VME and one ME (4.0%) were noted for both ofloxacin and levofloxacin (Table 3). Both ME were for the same isolate with a ciprofloxacin MIC of 0.03 μg/ml and ofloxacin inhibition zone diameter of 24 mm and levofloxacin inhibition zone diameter of 27 mm; this isolate had no resistance mechanism. When ofloxacin and levofloxacin disk results were evaluated against ciprofloxacin MIC to include the intermediate category, 11.0% and 11.8% mE were noted, respectively (Table 4). As was the case for ciprofloxacin, the majority of these mE were for isolates that gave intermediate result by the ofloxacin or levofloxacin disk but a ciprofloxacin-resistant MIC of 1.0 μg/ml (see Fig. S3 and S4 in the supplemental material).

TABLE 4.

Performance of surrogate disks compared to ciprofloxacin-susceptible, -intermediate, and -resistant MICs in Salmonella spp.^a

Surrogate disk	Performance (%)b
	VME	ME	mE	CA
Ciprofloxacin	0.0	0.0	11.0	89.0
Levofloxacin	0.0	0.0	11.8	88.2
Ofloxacin	0.0	0.0	11.0	89.0
Nalidixic acid	0.0	4.0	66.2	33.1
Enoxacin	25.0	0.0	47.8	48.5
Norfloxacin	50.0	0.0	72.1	20.6
Pefloxacin	0.0	12.0		97.8

^aTwenty isolates were ciprofloxacin resistant, 91 isolates were ciprofloxacin intermediate, and 25 isolates were ciprofloxacin susceptible.

^bVME, very major errors; ME, major errors; mE, minor errors; CA, categorical agreement.

Enoxacin and norfloxacin disks demonstrated 53.1% and 90.9% VME, respectively, when evaluated using CLSI Enterobacteriaceae susceptible versus not susceptible breakpoints for these antimicrobials; no ME were noted (Table 3). If susceptible, intermediate, and resistant breakpoints were considered for enoxacin and norfloxacin disks, the number of VME was 25% and 50%, respectively (Table 4). However, a Salmonella-specific not susceptible breakpoint could be assigned to enoxacin and norfloxacin at ≤23 and ≤27 mm, respectively. These Salmonella-specific breakpoints result in VME rates of 0.0% (enoxacin) and 0.9%, (norfloxacin) and ME rates of 4.0% and 8.0% for these two surrogate disks (Table 3).

By the current CLSI resistant breakpoints, nalidixic acid testing does not detect all mechanisms of fluoroquinolone resistance in S. enterica (3). However, we interpreted nalidixic acid disk results as susceptible versus not susceptible (versus ciprofloxacin MIC) to determine whether this approach would improve performance of nalidixic acid as a surrogate agent for detection of fluoroquinolone resistance. When data were analyzed in this manner, 1.8% (n = 2) VME and 20% (n = 5) ME were noted (Table 3). Both VME were in isolates with qnr resistance mechanisms and ciprofloxacin MICs of 0.12 to 0.5 μg/ml. In contrast, by the current CLSI nalidixic acid-susceptible, -intermediate, and -resistant disk breakpoints, 66.2% mE were observed (Table 4). Of these mE, 84/90 (78%) were in isolates with a ciprofloxacin-intermediate MIC and a nalidixic acid-resistant zone, which is the phenotype of the QRDR fluoroquinolone resistance mechanism, although 29 of the 90 mE were in isolates with plasmid-mediated quinolone resistance mechanisms (not shown). For several isolates with ciprofloxacin-intermediate MICs, discrete colonies within the nalidixic acid zone of inhibition were observed, but these did not impede reading of nalidixic acid zones, as even when these colonies were ignored, zone measurements were interpreted as resistant (Fig. 2A).

FIG 2.

Representative disk diffusion result for nalidixic acid (A) and pefloxacin (B) demonstrating colonies within the zone of inhibition.

Pefloxacin disk results, interpreted by CLSI and EUCAST susceptible versus not susceptible breakpoints for Salmonella, yielded no VME and 12% (n = 3) ME compared to ciprofloxacin MICs (Tables 3 and 4). One of the ME was in the isolate with a QRDR mutation and ciprofloxacin-susceptible MIC of 0.06 μg/ml and inhibition zone diameter of 21 mm. Similar to what was seen with nalidixic acid, for 47% of isolates with ciprofloxacin MICs that were not susceptible, discrete colonies were observed within the pefloxacin zone of inhibition (Fig. 2B). We did not observe colonies within the zone of inhibition for any of the fluoroquinolones evaluated for this collection of isolates.

When applying susceptible versus not susceptible breakpoints, performance of four of the seven surrogate disks (ciprofloxacin, pefloxacin, ofloxacin, and levofloxacin) was equivalent (Table 3). In order to determine whether one disk would better separate ciprofloxacin-susceptible from non-ciprofloxacin-susceptible isolates (based on ciprofloxacin MICs), the percentage of isolates with inhibition zone diameters or zone sizes near the proposed breakpoint for each agent, arbitrarily chosen as 3 mm on either side of the breakpoint, was calculated (shown in figures in the supplemental material). Twenty-six percent of isolates included in the study had a zone size within 3 mm of the ofloxacin breakpoint (zone sizes of 22 to 27 mm), and 30% of the isolates had a zone size within 3 mm of the levofloxacin breakpoint (25 to 30 mm). In contrast, 16.9% of isolates had zone sizes within 3 mm of the ciprofloxacin breakpoint (28 to 33 mm) and only 15.4% of isolates had a zone size within 3 mm of the pefloxacin breakpoint (21 to 26 mm).

Evaluation of a surrogate disk based on levofloxacin MIC results.

No Salmonella levofloxacin disk breakpoints have been formally established, and yet levofloxacin is commonly used in many U.S. hospitals as the formulary fluoroquinolone. Performance of the proposed levofloxacin disk breakpoints for S. enterica is reported elsewhere (4, 5). We therefore assessed the performance of ciprofloxacin, nalidixic acid, and pefloxacin disks as potential surrogate agents for the detection of levofloxacin-susceptible versus non-levofloxacin-susceptible MICs. Ofloxacin, enoxacin, and norfloxacin were not included in this analysis due to poor performance with ciprofloxacin. Categorical agreement between the ciprofloxacin disk and levofloxacin MIC was 97.8%. One VME was noted (0.9% of 112 non-levofloxacin-susceptible isolates [Table 5]) for an S. enterica serovar Typhi with a ciprofloxacin-susceptible disk zone of 31 mm but a levofloxacin-intermediate MIC of 0.25 μg/ml. This isolate had a ciprofloxacin-susceptible MIC of 0.06 μg/ml and a mutation in the QRDR. Two ME were noted (8.7%) in isolates with levofloxacin MICs of 0.12 μg/ml but ciprofloxacin disk zones of 28 and 30 mm. A categorical agreement of 95.6% was found between nalidixic acid disks and levofloxacin MICs, which included 3 VME (2.7% of 112 non-levofloxacin-susceptible isolates) and 3 ME (13% of 24 levofloxacin-susceptible isolates) (Table 5). Two of the VME were for isolates with qnr and levofloxacin MICs of 0.25 μg/ml; the third was for an S. enterica serovar Typhi with no fluoroquinolone resistance mechanism detected and a ciprofloxacin MIC of 0.06 μg/ml but a levofloxacin MIC of 0.25 μg/ml. Categorical agreement was 100% for pefloxacin disk compared to levofloxacin MIC (Table 5).

TABLE 5.

Performance of surrogate disks for the differentiation of levofloxacin-susceptible versus non-levofloxacin-susceptible MICs in Salmonella spp.^a

Surrogate disk	Performance (%)b
	VME	ME	CA
Ciprofloxacin	0.9	8.7	97.8
Nalidixic acid	2.7	13.0	95.6
Pefloxacin	0.0	0.0	100

^aTwenty isolates were levofloxacin resistant, 92 isolates were levofloxacin intermediate, and 24 isolates were levofloxacin susceptible.

^bVME, very major errors; ME, major errors; CA, categorical agreement.

Performance of pefloxacin disks from three manufacturers.

As pefloxacin disks are not available in the United States and have not been evaluated by CLSI prior to inclusion in CLSI document M100-S25 (7) for detection of Salmonella fluoroquinolone resistance, we evaluated the performance of pefloxacin from three manufacturers using a subset of 72 isolates. While BD and Mast disks yielded equivalent zones, zone sizes obtained by Oxoid disks were on average 1.9 mm (standard deviation, 0.96) and 1.8 mm (standard deviation, 0.66) larger than zones determined using BD and Mast disks (P < 0.0001). However, this difference did not change the pefloxacin-susceptible versus non-pefloxacin-susceptible interpretation for any of the 72 isolates tested.

DISCUSSION

This study evaluated the performance of seven surrogate disks for the detection of fluoroquinolone resistance in S. enterica. We demonstrated excellent performance for ciprofloxacin and pefloxacin disks at differentiating ciprofloxacin-susceptible versus non-ciprofloxacin-susceptible S. enterica. ME rates were higher than the typically accepted rate of <3% when testing a random selection of organisms; however, it should be noted that ME rates are inflated in the present study due to the inclusion of few (n = 25) susceptible isolates. Levofloxacin and ofloxacin disks also performed well, with few VME and ME. However, many isolates had zone sizes near the breakpoint for these two agents (see figures in the supplemental material), which was of concern because disk diffusion is generally accepted as precise to only ±2 mm. This variability is related to disk lot and manufacturer, medium lot and manufacturer, incubation temperature and time, and reader proficiency, among other factors. A major limitation of the present study is that only one manufacturer of MHA (BD) was used. To this point, studies performed at the EUCAST Development Laboratory, Växjö, Sweden, and Staten Serum Institute for EUCAST during their evaluation of pefloxacin disks noted a much larger proportion of isolates with zones within 1 to 3 mm of the ciprofloxacin breakpoint when using three manufacturers of MHA (6). As a result, EUCAST does not recommend ciprofloxacin as a surrogate agent for the detection of fluoroquinolone resistance in S. enterica. We also evaluated enoxacin and norfloxacin as potential surrogate agents, but these did not perform well, with an unacceptable number of VME (Table 3).

It is important to note that no surrogate agent will detect all fluoroquinolone resistance mechanisms. For example, isolates that acquire aac(6′)-Ib-cr as the sole fluoroquinolone resistance determinant will test susceptible to pefloxacin, nalidixic acid, ofloxacin, and levofloxacin, among others (8). This is because aac(6′)-Ib-cr encodes an enzyme that acetylates the amino nitrogen on the R₇ piperazinyl substituent found on some, but not all, quinolones, including ciprofloxacin, norfloxacin, and enoxacin (8, 9). This resistance mechanism remains rare and may be coupled with other fluoroquinolone resistance mechanisms, as was the case for the one isolate tested in the present study (R. Skov, unpublished observations). Similarly, we observed an S. enterica serovar Typhi isolate with documented mutation to the QRDR that repeatedly tested susceptible to ciprofloxacin by broth microdilution. Nalidixic acid did not detect qnr mutations in 5.4% of the 37 isolates with this resistance mechanism studied here.

Both nalidixic acid and pefloxacin are currently included in the CLSI M100-S25 document (7) as surrogate agents for the detection of fluoroquinolone resistance in Salmonella spp. Laboratories need not test both agents but rather evaluate whether surrogate disks should be used (i.e., cannot perform ciprofloxacin MIC) and which surrogate disk best meets their needs. Laboratories that have noted difficulties in reading ciprofloxacin disks and are in countries where pefloxacin disks are available may opt to use pefloxacin disks as a surrogate agent for the detection of ciprofloxacin resistance. We noted several colonies within the zone of inhibition for both pefloxacin and nalidixic acid; these should not be ignored, as these isolates all had ciprofloxacin-intermediate MICs. The most common resistance mechanism for typhoidal Salmonella remains the QRDR mutation in gyrA, which is readily detected by nalidixic acid disk testing (2). As such, laboratories in countries where typhoid fever is endemic and that have noted success with nalidixic acid may opt to continue to test with this agent. If this approach is taken, interpretation of nalidixic acid disk zones as susceptible versus not susceptible (rather than susceptible, intermediate, and resistant) would be recommended, because some isolates will yield ciprofloxacin MICs in the intermediate range and nalidixic acid-resistant zones, the classic phenotype of isolates with QRDR mutation (2) (Tables 3 and 4). However, laboratories that choose to continue testing with nalidixic acid must be cognizant that this disk will over-call resistance, as demonstrated by the 20% ME rate (five isolates). Furthermore, laboratories testing nontyphoidal Salmonella isolates wherein plasmid-mediated quinolone resistance is more common must be aware that nalidixic acid will not detect such resistance mechanisms.

Laboratories in the United States and other developed countries may consider performing an MIC test for ciprofloxacin or levofloxacin (depending on the agent to be prescribed). However, no commercial MIC susceptibility test panels produced in the United States contain ciprofloxacin or levofloxacin concentrations low enough to allow use of the current CLSI susceptible breakpoints. We recently demonstrated that Etest performs well compared to BMD at differentiating ciprofloxacin- and levofloxacin-susceptible isolates versus isolates not susceptible to these agents (5). However, U.S. laboratories that desire to use Etest for S. enterica with the current CLSI Salmonella breakpoints would need to perform a verification study prior to implementation for clinical testing, as this product has not been cleared by the FDA for use with these CLSI breakpoints. Since S. enterica isolates that require susceptibility testing (e.g., typhoid serovars or isolates from extraintestinal specimens) are infrequently encountered in most U.S. clinical laboratories, investing the resources to perform such studies is likely not cost-effective. Pefloxacin disks are not available in the United States and are unlikely to become available in the near future. Pefloxacin has not been cleared by the FDA for therapeutic use, which precludes sale of pefloxacin disks without appropriate clinical trial data supporting the performance of this test. The ciprofloxacin disk test may be a suitable alternative for differentiating ciprofloxacin- and levofloxacin-susceptible isolates versus isolates not susceptible to these agents for S. enterica for U.S. laboratories if MIC tests are not available. If a laboratory opts to use a surrogate drug (Table 3), the results should be reported as susceptible or resistant to ciprofloxacin or levofloxacin on the patient report. Results from the surrogate agent tested (other than ciprofloxacin or levofloxacin) should not be reported, as this could imply that the surrogate might be used for therapy. Reporting results of surrogate agents is not recommended for other surrogate agents (e.g., cefoxitin for detection of mecA-mediated resistance in Staphylococcus aureus [2]).