Activity of Eravacycline against Escherichia coli Clinical Isolates Collected from U.S. Veterans in 2011 in Relation to Coresistance Phenotype and Sequence Type 131 Genotype

Abstract

Eravacycline is a novel broad-spectrum fluorocycline with potent Gram-negative activity, including for multidrug-resistant strains. Among 472 Escherichia coli clinical isolates from 24 Veterans Affairs medical centers (in 2011), divided equally as susceptible versus resistant to fluoroquinolones, broth microdilution eravacycline MICs were distributed unimodally, ranging from 0.03 to 1.0 μg/ml (MIC₅₀ of 0.125 μg/ml, MIC₉₀ of 0.25 μg/ml). Eravacycline MICs were ∼2-fold higher among fluoroquinolone-resistant, gentamicin-resistant, multidrug-resistant, and sequence type 131 (ST131) isolates (P < 0.01 for each comparison).

TEXT

Sequence type 131 (ST131) of Escherichia coli causes most antimicrobial-resistant E. coli infections in the United States today, especially among veterans (1, 2). It is closely associated with fluoroquinolone resistance, thanks to its H30 subclone (1–5), and, with extended-spectrum beta-lactamase (ESBL) production, thanks to the H30 subclone's H30Rx subclone (6–8).

Eravacycline is a novel, broad-spectrum fluorocycline with potent Gram-negative activity, including against multidrug-resistant (MDR) strains (9–15). In phase 2 and 3 clinical trials for complicated intra-abdominal infections, in which E. coli predominated, eravacycline was as effective as ertapenem (11, 16). Eravacycline activity against ST131 is undefined (9). Here, for E. coli clinical isolates from Veterans Affairs medical centers (VAMCs) across the United States (1), we determined MICs to eravacycline and 8 comparator agents, and we compared susceptibility results with ST131, H30, and H30Rx genotype.

Isolates.

Sets of ∼20 each of consecutive E. coli clinical isolates were collected in 2011 from 24 widely distributed Veterans Affairs medical centers, as ∼10 each of fluoroquinolone-susceptible and -resistant isolates per center (1). ST131 and its H30 and H30Rx subclones were detected by lineage-specific PCR assays (4, 8).

Susceptibility testing.

Broth microdilution MICs were determined for eravacycline, minocycline, tigecycline, levofloxacin, piperacillin-tazobactam, gentamicin, ceftazidime, ceftriaxone, and meropenem using 2015 Clinical and Laboratory Standards Institute (CLSI)-specified procedures, reference strains, and breakpoints (17, 18). The median of 2 to 5 total MIC determinations was used. Categorical susceptibility to noneravacycline comparator agents was defined per CLSI criteria (18). The new susceptibility data, plus existing data for other agents (1), were used to classify isolates as MDR based on nonsusceptibility to ≥1 member each of ≥3 drug classes. Extended-spectrum beta-lactamase (ESBL) production was determined by a standard double-disk method, using cefotaxime and ceftazidime, with and without clavulanate (17, 18).

Statistical analysis.

Comparisons of proportions and MIC values were tested using a chi-square test and the Mann-Whitney U test, respectively.

Eravacycline and comparator agent MICs versus fluoroquinolone phenotype.

Among the 472 study isolates, eravacycline MICs were distributed unimodally from 0.03 to 1.0 mg/liter (MIC₅₀ of 0.125 mg/liter, MIC₉₀ of 0.25 mg/liter) (Fig. 1). They were statistically significantly but only 2-fold higher among fluoroquinolone-resistant than fluoroquinolone-susceptible isolates (Table 1; Fig. 1). Comparator agent MICs followed the same pattern, being significantly higher, although only slightly so, among fluoroquinolone-resistant than fluoroquinolone-susceptible isolates (Table 1). Categorical susceptibility exhibited a similar trend, being less common among fluoroquinolone-resistant than fluoroquinolone-susceptible isolates.

FIG 1.

Distribution of eravacycline MICs among 472 Escherichia coli isolates from U.S. veterans, overall and in relation to fluoroquinolone (FQ) phenotype. Isolates were considered FQ resistant if the ciprofloxacin MIC was ≥2 mg/liter.

TABLE 1.

MICs for eravacycline and 8 other antimicrobial agents among 472 Escherichia coli isolates from veterans (in 2011)

Drug	MIC (mg/liter) of:							No. (column %) susceptible of:
	FQ-Sa isolates (n = 234)			FQ-Rb isolates (n = 238)			P valuec	FQ-S isolatesa (n = 234)	FQ-R isolatesb (n = 238)	P valuec
	Range	MIC50	MIC90	Range	MIC50	MIC90
Eravacycline	0.03 to 0.50	0.13	0.25	0.03 to 1.0	0.25	0.50	<0.001	Undefined	Undefined	n.a.d
Minocycline	0.25 to 256	1	8	0.50 to 256	2	8	<0.001	204 (87)	199 (84)
Piperacillin-tazobactame	≤1 to > 128	2	8	≤1 to > 128	2	32	<0.001	222 (95)	204 (86)	0.001
Gentamicin	0.06 to > 16	2	2	0.25 to > 16	2	> 16	<0.001	227 (97)	157 (66)	<0.001
Levofloxacin	≤0.06 to 1	0.32	0.25	4 to >8	>8	>8	<0.001	234 (100)	0 (0)	<0.001
Ceftazidime	≤0.13 to > 16	0.50	1	≤0.13 to > 16	1	2	<0.001	231 (99)	219 (92)	<0.001
Ceftriaxone	≤0.03 to >4	0.06	0.25	≤0.03 to >4	0.13	1	<0.001	230 (98)	215 (90)	<0.001
Meropenem	≤0.03 to 0.13	<0.03	<0.03	≤0.03 to 2	≤0.03	≤0.03	0.02	234 (100)	237 (99.6)
Tigecycline	0.01 to 1	0.50	0.50	0.25 to 4	0.50	1	<0.001	234 (100)	237 (99.6)

^aFQ-S, fluoroquinolone susceptible.

^bFQ-R, fluoroquinolone resistant.

^cP values (by a two-tailed Mann-Whitney U test [MIC data] or by Fisher's exact test [% susceptible]) for comparisons of fluoroquinolone-susceptible versus fluoroquinolone-resistant isolates are shown where the P value is <0.05.

^dn.a., not applicable (no data to compare).

^eMIC data shown are for piperacillin. The tazobactam concentration was 4 mg/liter throughout.

Eravacycline MICs versus resistance to comparator agents.

Eravacycline MICs varied significantly in relation to resistance to comparator agents only for fluoroquinolones (Table 1) and gentamicin (MIC₅₀ for gentamicin-susceptible versus gentamicin-resistant isolates, 0.125 mg/liter versus 0.25 mg/liter; P = 0.003). The difference for gentamicin disappeared with stratification by fluoroquinolone phenotype (not shown), consistent with it reflecting the association of gentamicin resistance and fluoroquinolone resistance (Table 1).

Eravacycline MICs versus MDR and ESBL status.

By class, the prevalence of resistance declined as follows: penicillins (57%), fluoroquinolones (50%), folate inhibitors (31%), cephalosporins (22%), aminoglycosides (19%), tetracyclines (15%), nitrofurans (7%), carbapenems (0.4%), and glycylcyclines (0.2%). Overall, 41% of isolates qualified as MDR. Eravacycline MICs were approximately 2-fold higher among MDR isolates (MIC₅₀ of 0.25 mg/liter; range, 0.06 to 1.0 mg/liter) than non-MDR isolates (MIC₅₀ of 0.125 mg/liter; range, 0.03 to 1.0 mg/liter) (P < 0.001).

Of the 28 ceftriaxone and/or ceftazidime-resistant isolates, 19 (68%) were ESBL producers. Eravacycline MICs were similar among these 19 isolates (MIC₅₀ of 0.125 mg/liter, MIC₉₀ of 0.50 mg/liter) and the remaining nine cephalosporin-resistant isolates (MIC₅₀ of 0.25 mg/liter, MIC₉₀ of 0.50 mg/liter).

ST131.

ST131 and its H30 subclone accounted for over three-fourths of fluoroquinolone-resistant isolates but for only 2% to 6% of fluoroquinolone-susceptible isolates; the H30Rx subclone was restricted to and accounted for 15% of fluoroquinolone-resistant isolates (Table 2). These lineages were associated with slightly (by approximately 2-fold) but statistically significantly higher eravacycline MICs than other isolates (Table 3). These associations weakened or disappeared when the comparisons were stratified by fluoroquinolone phenotype (Table 4, left-hand P value column), suggesting that they resulted mostly from the association of ST131 with fluoroquinolone resistance (Table 2), which in turn was associated with higher eravacycline MICs (Table 1). Accordingly, associations of the eravacycline MIC with fluoroquinolone resistance persisted despite stratification for ST131, H30, and H30Rx status (Table 4, right-hand P value column).

TABLE 2.

Prevalence of ST131 and its clonal subsets in relation to fluoroquinolone resistance among 472 Escherichia coli isolates from U.S. veterans

Clonal subset	Prevalence of subset (no. [column %]) in:		P valuea
	Fluoroquinolone-susceptible isolates (n = 234)	Fluoroquinolone-resistant isolates (n = 238)
ST131 (n = 201)	14 (6)	187 (79)	<0.001
H30 (n = 185)	5 (2)	180 (76)	<0.001
H30Rx (n = 35)	0 (0)	35 (15)	<0.001

^aP values are by Fisher's exact test (two-tailed).

TABLE 3.

Association of eravacycline MIC with ST131 subsets

ST131 subset	Specific group	No. of isolates	MIC (mg/liter) data			P valuea for ST131 subset vs others
			Range	MIC50	MIC90
ST131	ST131	201	0.03–1.0	0.25	0.50	<0.001
	Others	271	0.03–1.0	0.13	0.25
H30	H30	185	0.03–1.0	0.25	0.50	<0.001
	Others	287	0.03–1.0	0.13	0.25
H30Rx	H30Rx	35	0.06–1.0	0.25	0.50	0.001
	Others	437	0.03–1.0	0.13	0.25

^aP values are by the Mann-Whitney U test (two-tailed).

TABLE 4.

Association of eravacycline MIC with ST131 genotype, stratified by fluoroquinolone phenotype

Fluoroquinolone phenotype (no.)	ST131 subset	Group	ST131 subset members and MIC data				P valuea
			No. of isolates	MIC range, mg/liter	MIC50, mg/liter	MIC90, mg/liter	ST131 subset vs others	Susceptible vs resistant isolates
Susceptible (234)	ST131	ST131	14	0.06 to 0.50	0.13b	0.50	0.04	0.17
		Others	220	0.03 to 0.50	0.13b	0.25		<0.001
	H30	H30	5	0.06 to 0.25	0.13	0.25	0.78	0.06
		Others	229	0.03 to 0.50	0.13	0.25		<0.001
	H30Rx	H30Rx	0	n.a.c	n.a.	n.a.	n.a.	n.a.
		Others	234	0.03 to 0.50	0.13	0.25		<0.001
Resistant (238)	ST131	ST131	187	0.03 to 1.0	0.25	0.50	0.27
		Others	51	0.03 to 1.0	0.13	0.50
	H30	H30	180	0.03 to 1.0	0.25	0.50	0.20
		Others	58	0.06 to 1.0	0.25	0.50
	H30Rx	H30Rx	35	0.06 to 1.0	0.25	0.50	0.28
		Others	203	0.03 to 1.0	0.13	0.50

^aP values (by two-tailed Mann-Whitney U test) are for comparison of ST131 group members versus others (same fluoroquinolone phenotype) or of fluoroquinolone-susceptible versus fluoroquinolone-resistant isolates (same ST131 genotype). Bold font represents significant difference (P < 0.05).

^bMean MIC (fluoroquinolone-susceptible isolates): ST131, 0.21 mg/liter; others, 0.14 mg/liter.

^cn.a., not applicable (no fluoroquinolone-susceptible H30Rx isolates).

Comment.

Our findings support four main conclusions. First, eravacycline was broadly active, with MICs of ≤1.0 mg/liter irrespective of coresistance phenotype and ST131 status. Second, fluoroquinolone resistance was associated with higher MICs for all agents tested, although, for eravacycline, this difference was small. Third, eravacycline MICs varied with additional coresistance only for gentamicin and MDR status, likely due to associations with fluoroquinolone resistance. Fourth, although ST131 and its clonal subsets exhibited higher eravacycline MICs than other isolates, these differences also were small and likely reflected fluoroquinolone resistance. Thus, eravacycline should prove useful for treating multidrug-resistant E. coli infections among veterans, regardless of coresistance and ST131 genotype.

The present eravacycline MICs for E. coli from veterans are consistent with previously reported values from nonveterans (9, 11), suggesting that they are widely generalizable. Their unimodal distribution indicates the absence of a naturally resistant subpopulation that could be selected for by widespread eravacycline use. Whether such a resistant subpopulation might emerge with future extensive use remains to be seen.

The increment in eravacycline MICs observed among fluoroquinolone-resistant isolates was statistically significant but small (∼2-fold), so it is of doubtful clinical importance. Conceivably, a common resistance mechanism (e.g., specific efflux pumps) may reduce susceptibility to both fluoroquinolones and eravacycline. The similar minor (∼2-fold) but statistically significant eravacycline MIC elevations noted among gentamicin-resistant and MDR isolates disappeared with stratification by fluoroquinolone resistance, suggesting that they resulted from confounding by fluoroquinolone phenotype, consistent with the strong observed associations of gentamicin resistance and MDR status with fluoroquinolone resistance.

The ST131 isolates, and specifically those representing the H30 and H30Rx subclones, had slightly (∼2×) but significantly higher eravacycline MICs than other isolates. These differences likewise disappeared with stratification by fluoroquinolone phenotype. Thus, the overall higher eravacycline MICs of ST131 isolates likely reflected the higher prevalence of fluoroquinolone resistance in ST131 isolates, not some distinctive aspect of the ST131 clonal background. The small size of these ST131-associated MIC differences suggests that eravacycline should be comparably effective for treating E. coli infections regardless of the ST131 status of a strain.

This study is limited by its in vitro nature and the paucity of isolates from sources other than urine. Study strengths include the geographically dispersed strain collection, the comparisons with ST131 genotype, and the comparisons with key coresistance phenotypes, including MDR status and ESBL production.

Summary.

Among clinical E. coli isolates from U.S. veterans (in 2011), eravacycline MICs largely mirrored those reported previously for nonveteran patient populations (9, 11), with a symmetrical unimodal distribution centered around 0.125 to 0.25 mg/liter. Eravacycline MICs were statistically significantly higher, but only slightly so, in association with fluoroquinolone, gentamicin, and multidrug resistance and among ST131, H30, and H30Rx isolates, all of which were associations that tracked with fluoroquinolone resistance. Thus, eravacycline should be equally useful for treating extraintestinal E. coli infections in veterans regardless of coresistance phenotype and ST131 status.