|Home | About | Journals | Submit | Contact Us | Français|
Among 40 Escherichia coli urine isolates from renal transplant recipients (Galveston, TX, 2003 to 2005), sequence type ST131 (O25:H4) was highly prevalent (representing 35% of isolates overall and 60% of fluoroquinolone-resistant isolates), virulent appearing, antimicrobial resistant (but extended-spectrum-cephalosporin susceptible), and associated with black race. Pulsotypes were diverse; some were linked to other locales. ST131 emerged significantly during the study period. These findings suggest that E. coli ST131 may constitute an important new multidrug-resistant threat to renal transplant recipients.
Renal transplant recipients are at increased risk for urinary tract infection (UTI), which can cause local and systemic manifestations and precipitate acute allograft injury (AAI) (14-17). Escherichia coli is the most common cause of UTI among renal transplant recipients (14-17). The increasing prevalence of resistance to trimethoprim-sulfamethoxazole (TMP-SMX) and other antimicrobials in E. coli threatens the utility of posttransplant TMP-SMX prophylaxis and makes acute UTI therapy more challenging.
A recent survey of E. coli urine isolates from renal transplant recipients at the University of Texas, Galveston, TX (2003 to 2005), identified a high prevalence of serogroup O25 (15), which was associated with levofloxacin resistance (J. Rice, unpublished data). Since O25 and fluoroquinolone (FQ) resistance characterize an emerging clonal group of pathogenic E. coli designated sequence type ST131 (3, 12), which is associated with the CTX-M-15 extended-spectrum beta-lactamase, we assessed the prevalence, characteristics, and clinical correlates of ST131 within this distinctive population.
From April 2003 to June 2005, renal transplant recipients at the University Texas Medical Branch, Galveston, TX, with E. coli bacteriuria were identified during clinic visits or hospital admissions (15). The main inclusion criterion was E. coli bacteriuria (i.e., ≥105 CFU/ml), if detected at >10 days posttransplant (to exclude perioperative bacteriuria). Forty E. coli urine isolates (one per patient) met inclusion criteria and became the study population. The experimentation guidelines of the authors' institutions were followed in the conduct of clinical research.
Demographic data, underlying host conditions, and clinical manifestations were ascertained by interview and record review. Asymptomatic bacteriuria, cystitis, and acute pyelonephritis were defined clinically; AAI was defined by a ≥20% serum creatinine increase (15).
Susceptibility to ampicillin, cefazolin, ceftriaxone, levofloxacin, gentamicin, nitrofurantoin, and TMP-SMX was assessed by the broth microdilution method per Clinical Laboratory Standards Institute specifications. Intermediate susceptibility was considered resistance. The resistance score was the number of agents to which an isolate exhibited resistance. Levofloxacin-resistant isolates were considered fluoroquinolone resistant (FQ-R).
E. coli phylogenetic groups (A, B1, B2, and D), the O25b rfb variant, blaCTX-M-15, and 55 virulence genes were detected by PCR (8). ST131 status was defined by ST131-specific single-nucleotide polymorphisms (SNPs) in gyrB and mdh, plus random amplified polymorphic DNA (RAPD) profiling (8). XbaI pulsed-field gel electrophoresis (PFGE) profiles were analyzed using BioNumerics (Applied Maths). Profiles ≥94% similar to the index profile for an established PFGE type within a private database (255 ST131 isolates and 125 PFGE types; J. R. Johnson) were assigned to that PFGE type; others defined a novel PFGE type.
O antigens were determined at the E. coli Reference Center at Pennsylvania State University by using 174 O-specific antisera, with selective confirmation by restriction analysis of O antigen gene clusters. H types were determined by restriction analysis of fliC amplicons (10).
Comparisons of proportions and continuous variables were tested using Fisher's exact test and the Mann-Whitney U test, respectively (both 2-tailed). Univariable and multivariable logistic regression analyses were used to identify predictors of ST131 status. Principal coordinate analysis (PCoA), a multidimensional scaling method analogous to principle components analysis (8), was used to collapse the molecular data set for simplified between-group comparisons of ST131 versus non-ST131 isolates. Groups were compared on the first 2 PCoA axes by using a 2-tailed t test.
The 40 renal transplant recipients were predominantly female (72%); the ages ranged from 20 to 68 years (median, 29.5 years). The racial/ethnic distribution was 55% Caucasian, 23% black, and 23% Hispanic. The UTI episodes occurred from 10 days to 145 months (median, 19 months) posttransplant. Most subjects (98%) were receiving prednisone plus calcineurin inhibitors. Fifteen (38%) had a previous UTI, 32 (82%) used prophylactic TMP-SMX during the first 6 months posttransplant, and 17 (44%) had received an FQ within 6 months before the UTI episode. At the index encounter, 30 subjects (75%) had relevant clinical manifestations, including cystitis (13%), pyelonephritis (39%), fever (43%), tachycardia (25%), any systemic symptom (49%), allograft tenderness (8%), and AAI (43%).
The 40 E. coli isolates exhibited by-agent resistance prevalences as follows: ampicillin, 75%; TMP-SMX, 68%; levofloxacin, 50%; cefazolin, 48%; gentamicin, 23%; and nitrofurantoin and ceftriaxone, 3% each. Isolates were resistant to from 0 to 5 of the 7 agents (median, 3 agents).
The isolates derived from phylogenetic groups B2 (65%), D (20%), A (10%), and B1 (5%). According to SNP-based PCR and RAPD analyses, 14 (54%) of the group B2 isolates represented ST131. Thus, ST131 accounted for 35% of isolates, versus only 30% for the rest of group B2 and ≤20% each for groups A, B1, and D.
The PFGE profiles of the 14 ST131 isolates formed a discrete cluster (Fig. (Fig.1),1), with 9 different PFGE types (1 to 3 isolates each). Three of these 9 types were already assigned within a private ST131-associated PFGE library, as the 1st-, 3rd-, and 7th-most-prevalent types. These 3 established types accounted for 6 study isolates and, in the reference library, included representatives from 14 locales on 3 continents (not shown). The remaining 6 (novel) PFGE types accounted for 1 or 2 isolates each (8 isolates total).
The 14 ST131 isolates all contained the O25b rfb allele; 13 expressed the O25 antigen (versus 0 non-ST131 isolates) (P < 0.001). All 14 exhibited flagellar type H4 (versus 3 non-ST131 isolates) (P < 0.001).
Compared with non-ST131 isolates, ST131 isolates exhibited significantly greater prevalences of resistance to levofloxacin (86%, versus 31%) (P = 0.002), cefazolin (71%, versus 35%) (P = 0.046), and gentamicin (43%, versus 12%) (P = 0.044), with a similar trend for ampicillin (93%, versus 65%) (P = 0.07). Accordingly, they exhibited significantly higher resistance scores (median, 4.0, versus 2.0) (P = 0.001). ST131's proportional contributions to resistance, by agent, were 67% for gentamicin, 60% for levofloxacin, 53% for cefazolin, 43% for ampicillin, and 41% for TMP-SMX. Notably, all 14 ST131 isolates were ceftriaxone susceptible and blaCTX-M-15 negative.
In a PCoA that incorporated virulence genes and phylogenetic groups, the ST131 and non-ST131 isolates were well separated on the first 2 axes (P = 0.001 and P < 0.001, respectively), which captured 69.4% of the total variance (Fig. (Fig.2).2). Compared with non-ST131 isolates, the ST131 isolates exhibited borderline or significantly greater prevalences of the F10 papA allele, iha, sat, fyuA, iutA, the K5 capsule gene, traT, ompT, and malX (Table (Table1).1). Conversely, they exhibited borderline or significantly lower prevalences of non-F10 pap elements, sfa and/or foc, hra, hlyD, cnf1, vat, pic, ireA, clbB, and clbN (Table (Table1).1). Similar patterns emerged in comparisons limited to the 20 FQ-resistant isolates, with all statistically significant differences favoring ST131 (Table (Table22).
Overall, the aggregate virulence scores were similar for ST131 and non-ST131 isolates (median, 9.0 per group). However, FQ-susceptible non-ST131 isolates exhibited high scores, FQ-resistant non-ST131 isolates low scores, and ST131 isolates intermediate scores (Fig. (Fig.3,3, left panel). Likewise, when the ST131 isolates (group B2) were compared for virulence scores with the non-ST131 isolates, stratified by group B2 status, the group B2, non-ST131 isolates exhibited the highest scores; the non-B2, non-ST131 isolates the lowest scores; and the ST131 isolates intermediate scores (Fig. (Fig.3,3, middle panel). In contrast, for resistance scores, the ST131 isolates exhibited the highest scores; the non-B2, non-ST131 isolates intermediate scores; and the group B2, non-ST131 isolates the lowest scores (Fig. (Fig.3,3, right panel). Thus, the ST131 isolates superseded the comparison groups for combined rank according to virulence score and resistance score.
ST131 occurred in 6 (67%) of 9 black subjects, versus 8 (26%) (P = 0.04) of 31 Hispanics and Caucasians (who exhibited similar ST131 prevalences, so the groups were combined for analysis). Likewise, among subjects with FQ-resistant isolates, ST131 occurred all 6 (100%) black subjects, versus 6 (43%) (P = 0.04) of 14 others. According to univariable logistic regression analysis, black race (but not other demographic or clinical variables) and resistance to levofloxacin, gentamicin, and cefazolin all significantly predicted ST131 status (Table (Table3).3). When black race, gentamicin resistance, and levofloxacin resistance were used together as predictors in a multivariable model, each exhibited an odds ratio similar to its univariable value, with borderline or significant P values (Table (Table3).3). ST131 status was not significantly associated with clinical manifestations, amount of time since transplant, or recent (within 6 months) FQ use. ST131 isolates occurred significantly later in the study period than non-ST131 isolates (median, day 600 [range, day 343 to day 807], versus day 460 [range, day 0 to day 789]) (P = 0.036), suggesting emergence over time.
This analysis of 40 E. coli urine isolates from renal transplant recipients (2003 to 2005) yielded several novel findings regarding E. coli ST131 (O25:H4). First, ST131 was remarkably prevalent (35% of isolates overall and 60% of FQ-R isolates). Second, ST131 isolates were more extensively antimicrobial resistant than others, especially those from group B2, yet had virulence scores similar to those of other isolates and significantly higher than those of non-group B2 isolates. Third, all ST131 isolates were ceftriaxone susceptible and lacked blaCTX-M-15. Fourth, ST131 was associated with black race. Fifth, the ST131 isolates' clonal diversity obliged explanations other than local transmission for ST131's prominence. Sixth, ST131 emerged significantly during the study period.
The 35% overall prevalence of ST131 is among the highest reported for a single E. coli clonal group in any extraintestinal E. coli collection. Likewise, the 60% prevalence of ST131 among FQ-R isolates may be the highest reported for a single E. coli clonal group within any antimicrobial-resistant population. The basis for this remarkable prominence requires explanation.
Contributing to ST131's prominence might be ST131's comparatively robust virulence and resistance profiles. Historically, antimicrobial-resistant E. coli isolates have had fewer virulence traits than susceptible E. coli isolates (6, 7). ST131 defies this paradigm by combining extensive resistance with an extensive virulence gene repertoire, which appears to be fairly homogeneous across locales and clinical sources. Thus, ST131 may have both virulence and resistance advantages over other resistant strains and yet have a substantial resistance advantage (and only a slight virulence disadvantage) over susceptible strains. The extensive multiresistance of the ST131 isolates would predictably complicate therapy, especially with oral agents, although nitrofurantoin and fosfomycin (and, for beta-lactamase-negative strains, oral cephalosporins) are usually active.
ST131's association with black race was unexpected. However, previous epidemiological studies have identified Hispanic ethnicity and Asian race as correlates with E. coli clonal group A and TMP-SMX resistance (1, 4, 13). It is possible that certain exposures, behaviors, or other factors associated with black race caused predisposition to bacteriuria with ST131. Confirmation of this association and attention to race and ethnicity in future epidemiological studies of drug-resistant E. coli are needed.
Against localized point-source spread as the basis for ST131's observed high prevalence, the ST131 isolates' PFGE profiles were diverse, with no 2 profiles being indistinguishable. Moreover, 3 of 9 ST131-associated PFGE types (including 2 with multiple isolates) had been identified previously in other U.S. and international locales, suggesting ongoing widespread dissemination of these lineages. Future studies should explore mechanisms for the dissemination of ST131, including possible host-to-host or food-borne transmission, with or without an environmental component.
The absence of ceftriaxone resistance and blaCTX-M-15 among the ST131 isolates contrasts with initial reports focusing on CTX-M-15 and extended-spectrum-cephalosporin resistance (3, 5, 12). However, several recent studies from Europe and Canada suggest that ST131 is usually susceptible to extended-spectrum cephalosporins (2, 8, 9, 11). This seeming conflict may derive from the initial focus on CTX-M-15-mediated cephalosporin resistance, whereas from a broader, population-based perspective, other phenotypes account for a greater proportion of ST131 isolates.
In summary, ST131 accounted for 35% of 40 E. coli urine isolates (60%, if FQ-R) from renal transplant recipients. ST131 isolates were more extensively antimicrobial resistant than other isolates, had fairly high virulence scores, and were associated with black race. The diversity of ST131 PFGE profiles argued against local clonal spread, whereas temporal trends suggested ongoing emergence of ST131. These findings implicate ST131 as an important emerging multidrug-resistant uropathogen among renal transplant recipients.
This material is based upon work supported by the Office of Research and Development, Medical Research Service, Department of Veterans Affairs.
Dave Prentiss (VA Medical Center) prepared the figures.
J. R. Johnson has received grants, consultancies, and/or honoraria from Merck, Bayer, Ortho-McNeil, Wyeth-Ayerst, Rochester Medical, and Procter & Gamble. The other authors have no conflicts of interest.
Published ahead of print on 16 November 2009.