Decreased Invasive Capacity of Quinolone- Resistant Escherichia coli in Patients with Urinary Tract Infections

Transcription

1 MAJOR ARTICLE Decreased Invasive Capacity of Quinolone- Resistant Escherichia coli in Patients with Urinary Tract Infections María Velasco, 1 Juan P. Horcajada, 1 José Mensa, 1 Antonio Moreno-Martinez, 1 Jordi Vila, 2 José A. Martinez, 1 Joaquim Ruiz, 2 Margarita Barranco, 2 Gloria Roig, 3 and Eladio Soriano 1 1 Infectious Diseases Unit and 2 Microbiology Department, Hospital Clinic, and 3 Centro de Atención Primaria Manso, Barcelona, Spain Quinolone-resistant (QR) Escherichia coli may have lower invasive capacity than does quinolone-susceptible E. coli. To evaluate this, we prospectively collected data regarding all cases of E. coli invasive urinary tract infections (IUTI) in 669 adults admitted to the Infectious Diseases Unit of our hospital during a 3-year period, as well as 10,950 patients with cystitis or asymptomatic bacteriuria who presented to the outpatient clinic during a 1-year period. QR E. coli was isolated in 20% of patients with cystitis, compared with 8% of those with IUTI ( P!.05). The proportion of E. coli isolates that were quinolone resistant was similar in patients with bacteremic and nonbacteremic IUTI. The factors of urinary manipulation and structural abnormalities were independently associated with the presence of quinolone resistance. Old age was the only variable independently associated with blood invasion. QR E. coli is less likely to produce invasive disease (pyelonephritis and prostatitis) than is quinolone-susceptible E. coli. However, once pyelonephritis or prostatitis have developed, there is no difference in the incidence of bacteremia. Quinolones, a relatively recently developed group of antimicrobial agents, are very useful in treating a variety of infections. The frequent use of these agents for multiple conditions, such as urinary tract infections (UTIs) and prophylaxis in patients with cirrhosis and hematological disorders, has been associated with the appearance of antibiotic resistance [1]. In particular, quinolone resistance in Escherichia coli is becoming a clinical problem in some areas, and several reports have appeared [2]. In Spain, quinolone resistance among bacteria infecting the urinary tract has reached an incidence of 20% 30% in recent years [3]. A number of mech- Received 12 March 2001; revised 5 June 2001; electronically published 10 October Financial support: Fundación Máximo Soriano. Reprints or correspondence: Dr. María Velasco, Servicio de Infecciones, Hospital Clinic, C/Villarroel 170, Barcelona, Spain (velasco@medicina.ub.es). Clinical Infectious Diseases 2001; 33: by the Infectious Diseases Society of America. All rights reserved /2001/ $03.00 anisms for this resistance are known, including mutations in the topoisomerase II and IV genes and decreased intrabacterial accumulation of quinolone, mostly secondary to the overexpression of efflux pumps or decreases in cell wall permeability [4]. The acquisition of resistance may be associated with phenotypic changes in bacteria [5, 6]. Some of these changes include the loss of virulence factors that might inhibit the invasion of renal and prostatic parenchyma by E. coli [6]. However, this potential reduction in the invasive capacity of quinolone-resistant (QR) E. coli has received little attention in clinical studies. In this study, we analyzed whether quinolone resistance in isolates of E. coli is associated with a different pattern of invasive capacity in UTIs. PATIENTS AND METHODS We prospectively collected data regarding all cases of urinary parenchymatous infection, including cases of 1682 CID 2001:33 (15 November) Velasco et al.

2 pyelonephritis, prostatitis, and febrile UTIs, in adults who were admitted to, or who first visited as outpatients, the Infectious Diseases Unit in the Hospital Clinic, Barcelona, Spain (a facility with 900 beds), from January 1997 to January All patients had community-acquired UTIs. Patients with nosocomial infections (defined as infections that were noted 48 h after admission or within 4 weeks after a previous discharge) were excluded from the study. In addition, we reviewed the results of all the urine cultures that were performed in an outpatient clinic from the same geographical area during 1999 that indicated cystitis or asymptomatic bacteriuria. This clinic serves a population of 800,000 persons, and its microbiology laboratory received 22,279 urine samples in Patients were evaluated according to an established protocol determined by some of the authors (M.V., J.P.H., and A.M.M.). All of the patients included in the study from the outpatient clinic had presented with pyuria and were mostly seen for urinary symptoms. Acute pyelonephritis was clinically defined as an armpit temperature of 138 C (fever), pyuria, and lumbar tenderness. Prostatitis was defined by the presence of fever, pyuria, and prostatic tenderness. Febrile UTIs were defined by the presence of fever and pyuria in the absence of other sources of infection. Patients with febrile UTIs who did not have lumbar or prostate tenderness probably had prostatitis or pyelonephritis, although we were not able to define the source of infection clinically. Cystitis was considered when dysuria and pyuria were present without fever. Parenchymatous urinary infections included cases of pyelonephritis, prostatitis, and febrile UTIs, and were considered to be invasive UTIs (IUTIs). Urine samples were collected using the clean-catch midstream method. Urine samples were spread on McConkey and cistine-leucine electrolyte agar-deficient agar plates for quantification and incubated for 48 h. Positive urine cultures were defined by a bacterial growth 10 5 cfu/ml. Identification was performed by means of standard methods, and susceptibility testing was performed using a microdilution method according to the guidelines established by the National Committee for Clinical Laboratory Standards [7]. Resistance to ciprofloxacin was defined as an MIC of 12 mg/l. Patients with negative urine culture results at admission were excluded from the study. An initial urine culture, 2 blood cultures, and a blood count were processed for all patients. Urine cultures were repeated after 48 h of treatment and 1 week and 1 month after therapy finished. Blood cultures were processed using an automatic infrared device system (BACTEC 9240; Becton Dickinson Diagnostic Instrument Systems) for 5 days. A patient was considered to have bacteremia when the same microorganism was isolated from the urine and blood specimens in 1 of the blood cultures. All of the patients were followed up for 4 weeks after therapy was completed. For the purpose of this study, urologic manipulation was defined as use of any urological instrumentation. Structural abnormalities included prostatic hypertrophy, bladder diverticuli, cystocele, urethral strictures, congenital abnormalities, and renal cysts, as well as functional abnormalities, such as neurogenic bladder and vesicoureteral reflux. An episode of IUTI at any time in the patient s life was considered to be a previous IUTI. Antibiotic use was taken into account if antibiotics were used within 1 month before the present episode of IUTI. The following host factors were included in the analysis: age, sex, and comorbid conditions (i.e., cirrhosis, diabetes, corticosteroid treatment, and kidney failure). Table 1. Quinolone resistance in Escherichia coli leading to cystitis and invasive urinary tract infection (IUTI). Total No. (%) of strains Susceptible Resistant Cultures that indicated non-iuti cystitis 10, (80) 2180 (20) Cultures that indicated UTI!.001 Patients with pyelonephritis, prostatits, or febrile UTI a 615 (92) 54 (8) Female 453 (92.7) 35 (7.3) Male 162 (90) 19 (10) Pyelonephritis, no. (no. female/no. male) 516 (461/55) Prostatitis, no. (no. female/no. male) b 83 ( /83) Febrile UTI, no. (no. female/no. male) a 70 (19/51) No. (%) of patients with bacteremia 213 (31.8) 199 (93.4) 14 (6.6).41 No. (%) of patients without bacteremia 456 (68.2) 416 (91.2) 40 (8.8) NOTE. UTI, urinary tract infection. a Febrile UTI indicates febrile urinary tract infections, which were defined by the presence of fever and pyuria in the absence of other sources of infection. b Prostatitis not possible in female patients. P Quinolone-Resistant Uropathogenic E. coli CID 2001:33 (15 November) 1683

3 Table 2. Univariate and multivariate analysis of the association between different clinical characteristics and the presence of quinolone-resistant Escherichia coli. Presence of characteristic Multivariate analysis Absence of characteristic P a OR (95% CI) P Underlying conditions Diabetes (n p 38) 5 (13.2) 49 (7.8).22 Cirrhosis (n p 13) 2 (15.4) 52 (7.9).28 Kidney failure (n p 11) 1 (9.1) 53 (8.1).607 Immunosuppressive treatment (n p 10) 0 (0) 54 (8.2) 1 Urological conditions Structural abnormalities (n p 82) b 17 (20.7) 37 (6.3) ( ).025 Previous urological instrumentation (n p 14) 8 (57.1) 46 (7)! ( ).0009 Indwelling urinary catheter (n p 10) 4 (40) 50 (7.6).006 NS Previous antibiotic use (n p 85) 15 (17.6) 39 (6.7).002 Previous IUTI (n p 128) 10 (7.8) 44 (8.1) 1 Age of patients with IUTI, mean years SD d ! ( ).0004 c Female IUTI (n p 480) 34 (7) 19 (10).27 NOTE. Data are no. (%) of subjects with characteristic, unless otherwise noted, and results are expressed as proportions of quinolone resistance in the presence or absence of the clinical condition. Clinical variables associated with quinolone-resistant bacteria ( P!.1) were included in the multivariate analysis. IUTI, invasive urinary tract infection. a Determined by means of univariate analysis. b Probably related to prostatic syndrome, because, when the analysis was done separately by sex, the urologic disease was no longer significant in women, and in men, it was substituted in the model by the presence of prostatic syndrome (OR, 4.8; 95% CI, ; P p.0044). c For decades. d Mean age SD for all patients was All patients with IUTIs received an initial iv dose of an antibiotic and continued to receive iv therapy until clinical defervescence. When urine culture results became available, therapy was modified appropriately according to the findings of an antibiogram. Blood cultures were not repeated if the clinical evolution was satisfactory. Cure was defined as the clinical resolution of symptoms and a negative urine culture result on chemotherapy and during the follow-up period. Results are expressed as means SDs and incidences among the study population. Categorical data were compared using the x 2 or Fisher s exact test. Mean ages were compared using the Student s t test. Variables associated with quinolone resistance or bacteremia in the univariate analysis ( P!.1) were included in a binary logistic regression analysis with quinolone resistance and bacteremia as separate dependent variables. The level of significance was set at P!.05. Statistical analysis was performed using SPSS software for Windows, version 8 (SPSS), and results are expressed as CIs and ORs. RESULTS During the study period, 1075 patients with parenchymatous IUTIs were admitted to the hospital, and 16,102 urine samples that revealed cystitis or asymptomatic bacteriuria were collected in the outpatient clinic. Of the 1075 admitted patients, 718 had positive urine culture results and were included in the study. E. coli was isolated from 669 (93.2%) of the 718 patients. The clinical distribution of the patients is shown in table 1. In patients with parenchymatous IUTIs, the mean serum C-reactive protein concentration was mg/dl, and mean leukocyte count was 13, cells/cm 3. Bacteremia was present in 213 patients (31.8%). In the outpatient clinic, 12,882 urine cultures (80%) were positive, and E. coli was isolated in 10,950 (85%) of them. QR E. coli was isolated from 2180 (20%) of the patients with cystitis, compared with 54 (8%) of the patients with parenchymatous IUTIs ( P!.05). In contrast, the rate of resistance was similar in cases of bacteremic and nonbacteremic IUTI (6.6% vs. 8.8%, respectively; P p.41). The rate of quinolone resistance was stable during the 3 years of the study. In patients with IUTIs, QR E. coli was associated with structural abnormalities, urological instrumentation, previous IUTI, previous antibiotic use, and old age by univariate analysis (table 2). To further assess the independent value of the previous associations, we performed logistic regression analysis with quinolone resistance as the dependent variable. Urinary manipulation (OR, 7.67; 95% CI, ; P p.0009) and structural abnormalities (OR, 2.24; 95% CI, ; P p.025) were independently associated with the presence of quinolone resistance. In contrast, previous antibiotic use did not reach statistical significance (table 2) CID 2001:33 (15 November) Velasco et al.

4 Table 3. Univariate and multivariate analysis of the association between different clinical characteristics and the presence of blood invasion. Presence of characteristic Multivariate analysis Absence of characteristic P a OR (95% CI) P Underlying conditions Diabetes (n p 38) 20 (52.6) 199 (31.6).011 NS Cirrhosis (n p 13) 5 (38.5) 216 (32.9).539 Kidney failure (n p 11) 4 (36.4) 213 (32.4).754 Immunosuppressive treatment (n p 10) 5 (50) 212 (32.2).307 Urological conditions Structural abnormalities (n p 82) 29 (35.4) 187 (31.9).442 Previous urological instrumentation (n p 14) 5 (35.7) 211 (32.3).766 Indwelling urinary catheter (n p 10) 4 (40) 213 (32.3).482 Previous antibiotic use (n p 85) 33 (38.8) 184 (31.5).214 Quinolone-resistant E. coli IUTI (n p 54) 14 (25.9) 202 (32.9).442 Previous IUTI (n p 128) 40 (31.3) 177 (32.7).916 Age of patients with IUTI, mean years SD b ! ( )!.0001 c Female IUTI (n p 480) 159 (33.1) 58 (30.6).577 NOTE. Data are no. (%) of subjects with characteristic, unless otherwise noted, and results are expressed as proportions of blood invasion in the presence or absence of the clinical condition. Clinical variables associated with blood invasion ( P!.1) were included in the multivariate analysis. IUTI, invasive urinary tract infection. a Determined by means of univariate analysis. b Mean age SD for all patients was c For decades. Bacteremia was associated with old age and diabetes by univariate analysis. By logistic regression analysis, old age was the only variable significantly to be associated with bacteremia, which possibly indicates that diabetes is more prevalent in older people (table 3). DISCUSSION E. coli remains quinolone susceptible (QS) in many places around the world [8, 9]. In contrast, in Spain, according to recently published data, the rates of QR E. coli in urine are 25% [6, 10]. In this study, we have assessed the difference in the prevalence of QR E. coli in cases of invasive and noninvasive UTIs, and we have found an unexpectedly low rate of resistance in parenchymatous IUTIs. This apparent discrepancy is probably the result of a different origin of urine samples: previously published rates in Spain refer to the global rate of quinolone resistance in all urine isolates, whereas our data discriminate between uncomplicated cystitis and parenchymatous urinary infections. Thus, invasive infections (e.g., pyelonephritis, prostatitis, and febrile UTIs) have a lower resistance rate. The American Association of Infectious Diseases recommends empiric treatment with quinolone in UTIs in areas where the prevalence of resistance to trimethoprim and trimethoprim-sulfamethoxazole among uropathogens is high [11]. Because the incidence of trimethoprim-sulfamethoxazole resistance in Spain is up to 30%, quinolone could still be an acceptable empirical treatment for IUTIs in Spain. Our data suggest that uropathogenic QR E. coli was less likely to produce invasive disease than is QS E. coli. These results are in agreement with a recent clinical work that found a lower proportion of QR E. coli in cases of pyelonephritis, renal abscesses, and bacteremia [6]. However, the occurrence of bacteremia was not assessed separately from other IUTI. In our work, we have found that once renal or prostatic parenchyma had been invaded, there was no difference in the incidence of bacteremia between QR E. coli and QS E. coli. It appears that quinolone resistance impairs the capacity of E. coli to invade local tissue of the kidney and prostate, but not to produce bacteremia once local invasion has taken place. It is well known that bacterial isolates recovered from patients with pyelonephritis more frequently exhibit virulence factors, such as adhesions, than do isolates recovered from patients with cystitis [12, 13]. Some reports have shown that exposure to quinolone antibiotics results in the reduced production of certain factors that contribute to the virulence of bacteria [5, 14]. Indeed, there are data regarding disruption of the regulatory mechanisms that control bacterial morphology and of decreased type 1 fimbriae production after quinolone exposition [5]. This could underlie the lower probability of pyelonephritis with QR E. coli. Preliminary results from our laboratory show a significant Quinolone-Resistant Uropathogenic E. coli CID 2001:33 (15 November) 1685

5 reduction of virulence factors in QR E. coli producing pyelonephritis as compared with QS E. coli [15]. In our series, previous use of antibiotics or urologic structural abnormalities was associated with QR E. coli IUTI by means of univariate analysis. Of interest, according to logistic regression analysis, the variables significantly associated with the presence of QR IUTI were a previous urologic manipulation and structural abnormalities, which suggests that the physical disruption of mucosae and host factor alterations may increase invasiveness of QR E. coli. The association with increased age may reflect the presence of age-related incontinence or other minor urologic abnormalities that are not considered to be structural abnormalities. It is known that uropathogenic E. coli isolates recovered from compromised patients or after bladder instrumentation often lack some virulence factors [16]. Our data are in agreement with a potential lower intrinsic virulence of QR E. coli, thus requiring additional local factors favoring invasiveness. In addition to local invasiveness, other factors seem to modulate the ability of E. coli to produce systemic disease, as assessed by positive blood cultures. Otto et al. [17] found a higher probability of blood invasion in a cohort of patients who did not have complicating factors when E. coli possessed P type fimbriae. However, when they analyzed patients with complicating factors, they did not find an increased frequency of blood invasion produced by E. coli that possessed such fimbriae [17]. Underlying conditions, immunosuppression, and old age are known to act in favor of blood invasion [18 20]. In our series of IUTI, according to multivariate analysis, quinolone resistance was not associated with an increased or decreased frequency of bacteremia. In contrast, host factors (particularly old age) were significantly associated with positive blood cultures, which suggests that once local invasion has occurred, host factors dominate, regardless of the presence of QR bacteria. In summary, our data suggest that QR E. coli is less able to cause IUTI, particularly in the absence of urological intervention. The potential implications for an adequate empiric therapy of UTIs deserves further evaluation. On the other hand, once local IUTI occurs, host factors (particularly old age) are the major risk factors for systemic infection, without a significant effect of the presence of QR bacteria. References 1. Peña C, Albareda J, Pallares R, et al. Relationship between quinolone use and emergence of ciprofloxacin-resistant Escherichia coli in bloodstream infections. Antimicrob Agents Chemother 1995; 39: Goettsch W, Pelt W, Nagelkerke N, Hendrix M, et al. Increasing resistance to fluoroquinolones in Escherichia coli from urinary tract infections in The Netherlands. J Antimicrob Chemother 2000; 46: Garau J, Xercavins M, Rodriguez-Carballeira M, et al. Emergence and dissemination of quinolone-resistant Escherichia coli in the community. Antimicrob Agents Chemother 1999; 43: Vila J, Ruiz J, Navia MM. Molecular bases of quinolone resistance acquisition in gram-negative bacteria. In: S.G. Pandaly, ed. Recent Res Devel Antimicrob Agents Chemother. India: Reseach Signpost, 1999: Bagel S, Heisig P, Wiedemann B. Fluoroquinolone resistance of Escherichia coli frequently associated with decreased expression of type 1 fimbriae [abstract c-37]. In: Program and abstracts of the 37th Interscience Conference on Antimicrobial Agents and Chemotherapy (Toronto). Washington, DC: American Society for Microbiology, 1997: Blázquez R, Menasalvas A, Carpena I, et al. Invasive disease caused by ciprofloxacin-resistant uropathogenic Escherichia coli. Eur J Clin Microbiol Infect Dis 1999; 18: National Committee for Clinical Laboratory Standards. Methods for dilution antimicrobial susceptibility tests for bacteria that grow aerobically, 5th ed. Approved standard M7-A5. Wayne, PA: National Committee for Clinical Laboratory Standards; Gupta K, Scholes D, Stamm W. Increasing prevalence of antimicrobial resistance among uropathogens causing acute uncomplicated cystitis in women. JAMA 1999; 281: Lepelletier D, Caroff N, Reynaud A, et al. Escherichia coli: epidemiology and analysis of risk factors for infections caused by resistant strains. Clin Infect Dis 1999; 29: Pardo P, Moreno R, Sabater S, et al. Aspectos epidemiológicos de la resistencia de Escherichia coli a ciprofloxacino en un hospital general. Rev Esp Quimioterap 2000; 13: Warren JW, Abrutyn E, Hebel JR, et al. Guidelines for antimicrobial treatment of uncomplicated acute bacterial cystitis and acute pyelonephritis in women. Clin Infect Dis 1999; 29: Blanco M, Blanco JE, Alonso MP, et al. Virulence factors and O groups of Escherichia coli isolates from patients with acute pyelonephritis, cystitis and asymptomatic bacteriuria. Eur J Epidemiol 1996; 12: Svanborg C, Godaly G. Bacterial virulence in urinary tract infection. Infect Dis Clin North Am 1997; 11: Martínez-Martínez L, Fernández F, Perea E. Relationship between haemolysis production and resistance to fluoroquinolones among clinical isolates of Escherichia coli. J Antimicrob Chemother 1999; 43: VelascoM, Horcajada J, Moreno-Martinez A. et al. Invasive capacity of quinolone resistant uropathogenic E. coli [abstract 1752]. In: Program and abstracts of the 40th Interscience Conference of Antimicrobial Agents and Chemotherapy (Toronto). Washington, DC: American Society for Microbiology, 2000: Johnson J, Roberts P, Stamm W. P fimbriae and other virulence factors in Escherichia coli urosepsis: association with patient s characteristics. J Infect Dis 1987; 156: Otto G, Sandberg T, Arklund B, et al. Virulence factors and pap genotype in Escherichia coli isolates from women with acute pyelonephritis, with or without bacteremia. Clin Infect Dis 1993; 17: Hekker T, Groeneveld A, Simoons-Smit A, et al. Role of bacterial virulence factors and host factors in the outcome of Escherichia coli bacteraemia. Eur J Clin Microbiol Infect Dis 2000; 19: Ikäheimo R, Siitonen A, Kärkkäinen U, et al. Community-acquired pyelonephritis in adults: characteristic of E. coli isolates in bacteremic and non-bacteremic patients. Scand J Infect Dis 1994; 26: Horcajada J, Martinez J, Moreno-Martinez A, et al. Factores predictivos de la presencia de bacteriemia en los varones con infección urinaria. Med Clin (Barc) 1999; 112: CID 2001:33 (15 November) Velasco et al.