Single-Dose Antibiotic Prophylaxis in Core Prostate Biopsy: Impact of Timing and Identification of Risk Factors

Transcription

1 european urology 50 (2006) available at journal homepage: Infections Single-Dose Antibiotic Prophylaxis in Core Prostate Biopsy: Impact of Timing and Identification of Risk Factors Sandra Lindstedt a, Ulla Lindström a, Eva Ljunggren b, Björn Wullt b, Magnus Grabe a, * a Department of Urology, Malmö University Hospital, University of Lund, Malmö, Sweden b Department of Urology, Lund University Hospital, University of Lund, Lund, Sweden Article info Article history: Accepted May 3, 2006 Published online ahead of print on May 16, 2006 Keywords: Prostate biopsy Urinary tract infection Genitourinary infection Antibiotic prophylaxis Infection surveillance Abstract Objectives: To assess the level of infectious complications and the impact of timing of a single, prophylactic, oral dose of ciprofloxacin 750 mg given either 2 hours before or in conjunction with ultrasound-guided core biopsy of the prostate in men without recognised risk factors and to analyse potential risk factors. Methods: All men undergoing prostate biopsy for elevated prostate specific antigen or clinical suspected prostate cancer were enrolled in an open, comparative prospective study. Excluded were men with recognised risk factors for infective complications. Two end points were chosen: febrile genitourinary infection and the results of postbiopsy urine culture. Results: A total of 1322 prostate biopsy occasions were made in 1157 men. Twelve (0.9%) cases of febrile genitourinary infections were recorded, two of which had proven sepsis. Administrating the drug 2 hours before or at the time of biopsy ( p > 0.5) showed no statistical difference. Eight of 12 patients were shown to have prebiopsy undisclosed risk factors. Four percent developed postbiopsy, asymptomatic, significant bacteriuria. In addition, three (27%) men with prebiopsy unrecognised bacteriuria, who were accidentally enrolled, developed febrile genitourinary infection; one had proven sepsis. Conclusions: A single high-dose of oral ciprofloxacin 750 mg can be administered in direct conjunction with prostate biopsy to men without recognised risk factors, keeping the infection rate at approximately 1%. Bacteriuria before biopsy is a major risk factor for infective complications. Attention given to recognising individual risk factors would reduce the risk of infection further. # 2006 European Association of Urology. Published by Elsevier B.V. All rights reserved. * Corresponding author. Department of Urology, Malmö University Hospital, SE Malmö, Sweden. Tel ; Fax: address: magnus.grabe@skane.se (M. Grabe) /$ see back matter # 2006 European Association of Urology. Published by Elsevier B.V. All rights reserved. doi: /j.eururo

2 european urology 50 (2006) Introduction Transrectal ultrasound-guided core prostate biopsy is a key event in the diagnosis of prostate cancer [1]. Transient side events such as local pain, haematuria, haematospermia, dysuria and rectal bleeding are reported in a large number of patients [2 4]. Bacteriuria is seen in 20 53% and bacteraemia in as many as 73% of patients [5,6]. Fever associated with genitourinary symptoms is described in 3 10% or more, turning into septicaemia in less than 5% [5 8]. Antimicrobial agents lower the incidence of postbiopsy infectious complications [8 12]. The timing, duration of the regimen and the route of administration remain controversial [13]. In most studies, a regimen of at least 3 days has been used [14]. In a Canadian trial, 1 and 3 days of a fluoroquinolone showed no difference [15]. Although administering the antimicrobial agent 1 2 two hours before intervention generally is recommended to reach adequate dosage at the time of the procedure [16], studies on timing for antibiotic prophylaxis in urology are limited. To study the impact of timing, we opted to compare the logistics in praxis in two nearby university hospitals: the Lund University Hospital, where the antibiotic was given orally 2 hours before biopsy, and the Malmö University Hospital, where a single high dose of a fluoroquinolone was administered in direct conjunction with the procedure. In addition, we intended to analyse potential risk factors for genitourinary infection (GUI). 2. Material and methods 2.1. Study design The study was open, prospective and comparative. All men undergoing transrectal ultrasound-guided core biopsy at the departments of urology of the nearby university hospitals of Lund and Malmö for elevation of prostate specific antigen (PSA) or clinical signs suggesting prostate cancer were enrolled in a surveillance programme with the use of a specially designed database [17]. Excluded were men with identified bacteriuria, positive urine dipstick test for bacteriuria, indwelling catheter treatment for acute urinary retention (AUR), history of repeated urinary tract infection (UTI), prostatitis or febrile genitourinary infection (GUI), hypersensitivity to ciprofloxacin and artificial heart valves, and men on steroid medication. Written consent was required Antibiotic regimen We opted for a single high dose of oral ciprofloxacin 750 mg (Ciproxin TM ; Bayer Sweden AB, Göteborg, Sweden). In the Lund protocol, the patients were given the tablet according to the general recommendations of administering the drug 2 hours before procedure. The drug was mailed to the patients along with appropriate instructions. At biopsy, the patients had to confirm intake of the drug. In the Malmö protocol, the patients were given the tablet just before biopsy under the control of the staff Population From May 2003 to June 2004, a total of 1588 prostate biopsies, 916 in Lund and 672 in Malmö, were performed in 1369 men. Out of these, 125 biopsy occasions in Lund and 141 in Malmö were excluded from the trial because of the above-mentioned risk factors. Five patients in Lund with positive urine culture (>10 5 CFU/ml) and one man in Malmö with a positive dipstick test and eventually proven positive urine culture (>10 5 CFU/ ml) were receiving antibiotic prophylaxis erroneously according to the study protocols. Also, five patients in Malmö with negative dipstick tests were shown to have bacterial growth >10 5 CFU/ml. Those 11 patients are included in this report and discussed separately, because they are of special interest for risk factors for infective complications Biopsy technique A standardised sextant technique, as described originally by Hodge [18], was used. In case of rebiopsy, the number of biopsies usually was increased to eight or 12 samples [19]. An ultrasound scanner (Falcon type 2101; B-K Medical; DK-2730 Herlev) equipped with a 5 10 MHz 8808 probe was used. The biopsy needles were 18 G 20 cm (sample length, 1.9 cm) from Bard Magnum Biopsy instrument system (cat no. MG1522, Bard Ltd, UK). No special bowel preparation was used, and local anaesthesia was not applied End points Two end points were chosen to study the efficacy of the two antibiotic prophylaxis administration modalities: (1) febrile GUI and (2) the result of the urine culture taken after biopsy. A febrile GUI was defined as fever above 38 8C accompanied by at least one symptom of the lower urinary tract (urgency, frequency, dysuria or suprapubic tenderness), with or without positive urine culture, and requiring additional antibiotic treatment, which is in accordance with the definition of the Centers for Disease Control and Prevention (CDC) [20]. Patients with the above-mentioned signs accompanied by systemic symptoms such as body temperature > 39 8C, shivering, sweating and general prostration, and combined with laboratory markers of systemic infection were hospitalised for suspected septicaemia. Positive blood culture was required for defining sepsis Urine and blood cultures In the Lund protocol, a urine culture was taken 1 week before biopsy; patients with bacteriuria >10 4 CFU/ml were excluded. In the Malmö protocol, the urine culture was collected the day of biopsy at the time a dipstick test for bacteriuria was taken. Patients with positive dipstick tests were excluded. In both settings, a urine culture was taken 2 4 weeks after biopsy.

3 834 european urology 50 (2006) Table 1 Details of 15 febrile genitourinary infections after 1322 ultrasound-guided transrectal core biopsy procedures Pat Site Age (y) Type of infection a Urine culture Blood culture Possible underlying risk factors for the event Febrile GUI in men with negative urine culture before biopsy (n = 1311) 1 L 57 GUI + >39 8C No growth No growth None identified; unclear fever 2 L 67 GUI + >39 8C Klebsiella pneumoniae No growth History of GUI, AUR, indwelling catheter 3 L 71 GUI + sepsis No growth E coli AUR, indwelling catheter 4 L 55 GUI n.d. n.d. None identified 5 L 66 GUI Pseudomonas aeruginosa n.d. History of UTI and pyelonephritis; earlier culture showed Pseudomonas sp growth 6 L 67 GUI + >39 8C No growth b No growth b History of prostatitis with PSA > 100; rebiopsy 7 M 55 GUI + sepsis E coli E coli None identified 8 M 66 GUI + >39 8C No growth No growth History of prostatitis, LUTS and previous Enterococcus faecalis infection. Control culture showed Enterococcus faecalis growth after postbiopsy treatment. 9 M 64 GUI E coli n.d. Rebiopsy; quinolone-resistant E coli 10 M 58 GUI E coli n.d. Rebiopsy; quinolone-resistant E coli 11 M 58 GUI + >39 8C No growth b No growth b History of prostatitis 12 M 64 GUI + >39 8C No growth No growth None identified Febrile GUI in men with unidentified bacteriuria before biopsy (n = 11) 13 L 68 GUI + >39 8C No growth No growth Bacteriuria before biopsy with Enterococcus sp; negative culture because of ongoing antibiotic treatment 14 L 81 GUI + >39 8C E coli n.d. Bacteriuria before biopsy 15 M 73 GUI + sepsis E coli E coli Bacteriuria; advanced CVD and secondary neurologic disorder; history of instrumentation and UTI; quinolone-sensitive E coli at biopsy, resistant strain afterwards AUR = acute urinary retention; CVD = cardiovascular disease; GUI = genitourinary infection; L = Lund; LUTS = lower urinary tract symptoms; M = Malmö; n.d. = not detected; PSA = prostate specific antigen; UTI = urinary tract infection. a GUI + >398 identify patients with high fever, shivering and systemic symptoms; GUI + sepsis identify patients with high fever, shivering, systemic symptoms and positive blood culture. b Treatment initiated before urine sample collection. Further urine and blood cultures were taken as required. All patients were informed personally and received written instructions on how to act in case of signs of infection, fever and shivering after biopsy. calculated originally on the basis of a reduction of infection rates from 5% to 2.5% in the group receiving the antibiotic 2 hours before biopsy. An interim assessment was scheduled after 1 year Urine dipstick test Study approval A nitrite dipstick was performed for screening for bacteriuria with uropathogens, because most of those bacteria reduce nitrate to nitrite [21]. Combur 5 Test 1 (Roche Diagnostics, CH- Basel) dipstick was used in this study. The leucocyte esterase test for the detection of pyuria was not registered in this study. The ethics committees of the University of Lund and the Swedish Drug Agency approved the study. 3. Results 2.8. Data recording The patients were asked to complete a questionnaire 2 weeks after biopsy to be handed over to the investigation nurses. The questionnaire included information on side-effects, febrile GUI, need for consultation with a physician or emergency ward and additional antibiotic treatment related to the procedure Statistics The Fisher exact test was chosen to compare events in the two groups, with the use of a two-tailed p value <0.05 for significance. A sample size of 1600 biopsy occasions was A total of 1161 patients underwent 1322 prostate biopsies (average, 1.14 per patient), 791 in Lund and 531 in Malmö. Haematuria after biopsy occurred in 46% of the biopsy occasions and dysuria in 15% Febrile GUI in men with negative urine culture before biopsy In all, 12 (0.9%) patients developed febrile GUI, six (0.8%) in Lund and six (1.1%) in Malmö (Table 1). Five of the 12 were hospitalised for a total of 12 days for severe infection, two of whom had positive blood

4 Nowadays, most men undergoing prostate biopsy usually are healthy otherwise. The procedure by it self, however, is not harmless. The most serious and even life-threatening complication is febrile GUI with septicaemia. Antibiotic prophylaxis has been shown to reduce the postbiopsy infection rate [8 12]. We opted for an open comparative study with a continuous prospective registration of all biopsy occasions in a specially designed quality registry [17] to follow our daily practice and to collect a large number of cases rather than a smaller number of patients in a randomised control study. Interesteuropean urology 50 (2006) culture, giving a rate of septicaemia of 0.17%. No intensive care was required. The rate of postbiopsy GUI did not differ ( p > 0.5) between the centres. The additional cost for hospitalisation and medication was estimated at about 6000 Euro. The history of all 12 patients with postbiopsy GUI was examined. Interestingly, as many as eight of them had one or more undetected or ignored risk factor. Five patients had a history of febrile GUI and/ or prostatitis, two had bladder outlet obstruction with AUR, and three of the patients underwent a second biopsy. Bacterial growth was detected in the urine in five of 12 (42%) of the cases. Escherichia coli was detected in both cases of sepsis (Table 1) Febrile GUI in men with unidentified bacteriuria before biopsy In the Lund protocol group, five patients with positive urine culture (two with E coli, two with Proteus sp and one with Enterococcus faecalis) were included erroneously and given single-dose antibiotic prophylaxis. Two of them developed a febrile GUI with clinical signs compatible with sepsis (Table 1). In the Malmö group, five patients had a negative dipstick test and were enrolled correctly in the study, but they were shown eventually to have significant bacterial growth before biopsy: two E coli, and one each of Enterococcus faecalis, Pseudomonas sp and Staphylococcus sp. One patient with an advanced vascular disorder and secondary neurologic symptoms developed septicaemia. One patient with positive dipstick and positive urine culture (Enterococcus sp) was enrolled erroneously in Malmö, without complications. Thus, three of 11 (27%) men with undisclosed or ignored significant bacteriuria at the time of biopsy developed a clinically severe infection, compared with a baseline rate of 1% in patients with sterile urine ( p < ) Bacterial growth after prostate biopsy The distribution of bacteria in 1042 urine cultures obtained 2 4 weeks after biopsy in patients without registered GUI is given in Table 2. The large majority (91%) were negative. Bacterial growth >10 5 CFU/ml was identified in 44 (4.3%) patients, whereas bacterial growth between 10 4 and 10 5 CFU/ml was seen in 49 (4.8%) cultures Validity of the dipstick test A total of 681 validated urine cultures were obtained after dipstick tests for bacteriuria taken before prostate biopsy in the protocol patients (Table 3). Table 2 Distribution and frequency of bacteria recovered in 1042 urine cultures (665 in Lund and 377 in Malmö) 2 4 weeks after biopsy in patients without registered GUI [n (%)] of all cases) Bacterial strains Only two false-positive tests were disclosed. On the other hand, five patients had a negative dipstick test in the presence of bacterial growth >10 5 CFU/ml, including E coli, Enterococcus sp and Proteus sp. The predictive value of the test for negative urine culture was close to 1.0, whereas a false-negative test was observed in half of the unrecognised asymptomatic patients with bacterial growth >10 5 CFU/ml. The dipstick test did not reveal low bacterial growth. 4. Discussion Bacterial growth (CFU/ml) > Escherichia coli 6 (0.6%) 7 (0.7%) Enterobacter, Klebsiella, 9 (0.9%) 12 (1.2%) Pseudomonas and Proteus spp Enterococcus spp 12 (1.1%) 9 (0.9%) Streptococcus group B 9 (0.9%) 15 (1.4%) Other gram-positive strains 8 (0.8%) 6 (0.6%) Total number of strains 44 (4.3%) 49 (4.8%) Table 3 Validity of the nitrite dipstick test for bacteriuria Bacterial growth (CFU/ml) Urine culture (n) Positive Nitrite test Negative Pd a > b Between c Negative (<10 4 ) d a Predictive value for bacterial growth. b Predictive value for significant bacterial growth. Urine culture showed growth of E coli, Enterococcus sp and Proteus sp. c Predictive value for low count bacterial growth. d Predictive value for sterile urine.

5 836 We conclude that a single oral dose of ciprofloxacin 750 mg, given in direct conjunction with ultrasoundeuropean urology 50 (2006) ingly, one centre was the other s control, because there were no consultations about the results during the study period. It is generally considered important to achieve an optimal drug concentration on the target site at the time of the procedure, thus administrating the antibiotic 1 2 hours before any procedure [16]. In addition, the choice of an antimicrobial agent is based on the pharmacokinetic proprieties. Fluoroquinolones, which cover most of the expected gramnegative and gram-positive strains causing GUI, reach high and long-lasting concentrations in the urine and the prostate [22,23]. Ciprofloxacin with its performing penetration, theoretically, is an appropriate drug for prophylaxis in prostate biopsy. We opted for a high dose of the antibiotic, to our knowledge not yet assessed, to ensure a drug concentration above the minimal inhibitory concentration (MIC) of the causative pathogens for a reasonable period of time, at least hours [24]. Administrating antibiotic prophylaxis 2 hours before this outpatient procedure implies appropriate planning and logistics as well as good patient compliance. Therefore, it would be practical to administer a single dose at the time of biopsy. In this study, we could demonstrate that, with a single dose of ciprofloxacin 750 mg given either 2 hours before or at the time of biopsy, the rate of severe febrile GUI could be kept as low as 1% in men with no prebiopsy identified risk factors, which is among the lowest rates reported in the literature [2,4,7 12]. The results were fully comparable between the two centres, confirming the validity of the results and the absence of impact of timing. Giving the drug at the time of biopsy ensures the highest possible patient compliance. Interestingly, eight of 12 of the patients had at least one assumed risk factor in terms of history of GUI, prostatitis, proven bacterial growth in an earlier urine culture, AUR or repeated prostate biopsy unrecognised before biopsy. Identifying those patients would reduce the number of infective complications further. There is an important ecologic advantage in reducing antibiotic prophylaxis to a single dose, because resistance to antibiotics is increasing and because of the known direct relation between resistance development and amount of drug used in the community [25,26]. Eleven patients with significant bacteriuria before biopsy were correctly or accidentally given antibiotic prophylaxis according to the protocols. Three (27%) developed a GUI with high fever, and one had proven septicaemia. This rate is dramatically higher, compared with the baseline rate in patients with sterile urine, and, thus, identifies bacteriuria even asymptomatic as a striking risk factor in prostate biopsy. This rate is of the same order of magnitude as the rate found for bacteriuria as a risk factor for GUI in conjunction with transurethral resection of the prostate [27]. Urine culture is the optimal method to identify patients at risk, but it is logistically and economically questionable whether it is justified to take a urine culture before prostate biopsy in all patients, at a cost in our settings of approximately 8000 Euro per 1000 cultures. We could show that a simple dipstick test for identification of bacteriuria with uropathogens is a relevant test to confirm the absence of bacterial growth with only two false-positive tests. However, the test paradoxically could be weaker in detecting bacterial growth >10 5 CFU/ml in asymptomatic patients, because a false-negative result was displayed in half of the cases. The dipstick test also will not reveal low bacterial growth, the importance of which still has to be investigated. On the basis of these figures and the use of a dipstick as the only prebiopsy screening for bacteriuria, failure to display significant bacterial growth would occur in some 1 2% of all patients without recognised risk factors. This figure is very close to our assumption that approximately 2 4% of the patients without obvious or known risk factors have asymptomatic bacteriuria. This observation underlines how important it is to evaluate prebiopsy patient-related risk factors, such as previous history of GUI infection, prostatitis, episodes of bacteriuria and obstruction of the lower urinary tract, to minimise the rate of infective complications. A prebiopsy urine culture is reasonable in this category of patients and in unclear cases. Bacteriuria is suggested as a marker for UTI and GUI after most procedures within the genitourinary tract. In our series of 1042 urine culture obtained 2 4 weeks after prostate biopsy from patients without reported infectious complications, we found that 9% of patients had high or low bacterial growth. In addition, only 42% of the clinical cases of GUI had a microbiologically proven infection. Consequently, a urine culture obviously is not an indicator for infection after prostate biopsy. This observation underlines our dilemma as for the possibility to define an infection within the genitourinary tract and the difficulty to create a valid definition for studies ahead. 5. Conclusions

6 european urology 50 (2006) guided transrectal biopsy of the prostate, is a costeffective regimen in our environment, giving a postprocedure GUI rate of 1%, including cases with sepsis. Prebiopsy bacteriuria dramatically increases the rate of infection. A dipstick test is relevant for screening for bacteriuria, although a small number of false-negative tests will fail to detect patients with unrecognised bacteriuria before biopsy. Further efforts to reduce infective complications can be achieved by a careful patient history to rule out any known or assumed risk factors. Conflicts of interest A grant from Bayer Sweden AB (Göteborg, Sweden) supported the development of the surveillance database and the present study. The company generously supplied the drug during the trial period. References [1] Aus G, Abbou CC, Bolla M, et al. EAU Guidelines on Prostate Cancer. Euro Urol 2005;48: [2] Raaijmakers R, Kirkels WJ, Roobol MJ, Wildhagen MF, Schröder FH. Complication rates and risk factors of 5802 transrectal ultrasound-guided sextant biopsies of the prostate within a population-based screening program. Urology 2002;60: [3] Crundwell MC, Cooke PW, Wallace DM. Patients tolerance of transrectal ultrasound-guided prostatic biopsy: an audit of 104 cases. BJU Int 1999;83: [4] Djavan B, Waldert M, Zlotta A, et al. Safety and morbidity of first and repeat transrectal ultrasound guided prostate needle biopies: results of a prospective European prostate cancer detection study. J Urol 2001;166: [5] Thompson PM, Talbot RW, Packham DM, Dulake C. Transrectal biopsy of the prostate and bacteraemia. Br J Surg 1980;67: [6] Lindert KA, Kabalin JN, Terris MK. Bacteraemia and bacteriuria after transrectal ultrasound guided prostate biopsy. J Urol 2000;164: [7] Enlund AL, Varenhorst E. Morbidity of ultrasound-guided transrectal core biopsy of the prostate without prophylactic antibiotic therapy. A prospective study in 415 cases. Br J Urol 1997;79: [8] Crawford ED, Haynes AL, Story MW, Borden TA. Prevention of urinary tract infection following transrectal prostate biopsy. J Urol 1982;127: [9] Kapoor DA, Klimberg IW, Malek GH, et al. Single-dose oral ciprofloxacin versus placebo for prophylaxis during transrectal prostate biopsy. Urology 1998;52: [10] Isen K, Kupeli B, Sinik Z, Sozen S, Bozkirli I. Antibiotic prophylaxis for transrectal biopsy of the prostate: a prospective randomized study of the prophylactic use of a single dose oral fluoroquinolone versus trimethoprimsulfamethoxazole. Int Urol Nephrol 1999;31: [11] Aron M, Rajeev TP, Gupta NP. Antibiotic prophylaxis for transrectal needle biopsy of the prostate: a randomised controlled study. BJU Int 2000;85: [12] Aus G, Ahlgren G, Bergdahl S, Hugosson J. Infection after transrectal core biopsy of the prostate risk factors and antibiotic prophylaxis. Br J Urol 1996;77: [13] Taylor HM, Bingham JB. The use of prophylactic antibiotics in ultrasound-guided transrectal prostate biopsy. Clin Radiol 1997;52: [14] Webb NR, Woo HH. Antibiotic prophylaxis for prostate biopsy. BJU Int 2002;89: [15] Sabbagh R, McCormack M, Peloquin F, et al. A prospective randomized trial of 1-day versus 3-day antibiotic prophylaxis for transrectal ultrasound guided prostate biopsy. Can J Urol 2004;11: [16] Bergamini TM, Polk Jr HC. The importance of tissue antibiotic activity in the prevention of operative wound infection. J Antimicrob Chemother 1989;23: [17] Grabe M, Bjartell A, Wullt B. PROQUR: a tool for quality control, epidemiological surveillance, patient follow-up and clinical research activities related to prostate cancer. Acta Oncologica 2005;44: [18] Hodge KK, McNeal JE, Terris MK, Stamey TA. Randome systematic versus directed ultrasound guided transrectal biopsies of the prostate. J Urol 1989;142:71 5. [19] Matlaga BR, Eskew LA, McCullough DL. Prostate biopsy: indications and technique. J Urol 2003;169:12 9. [20] Horan TC, Gaynes RP. Surveillance of nosocomial infections. In: Mayhall CG, editor. Hospital epidemiology and infection control. 3rd ed. Philadelphia: Lippincott, Williams & Wilkins; p [21] Devillé WL, Yzermans JC, van Duijn NP, et al. The urine dipstick test useful to rule out infections. A meta-analysis of the accuracy [review]. BMC Urol 2004;4:4. [22] Naber KG, Adam D, Kees F. In vitro activity and concentrations in serum, urine, prostatic secretion and adenoma tissue of ofloxacin in urological patients. Drugs 1987; 34: [23] Naber KG. Use of quinolones in urinary tract infections and prostatitis. Rev Infect Dis 1989;11:S [24] Wagenlehner FM, Naber KG. Antimicrobial treatment of prostatitis. Expert Rev Anti Infect Ther 2003;1: [25] Goossens H, Ferech M, van der Stichele R, Elseviers M, for the ESAC Project Group. Outpatient antibiotic use in Europe and association with resistance: a cross national database study. Lancet 2005;365: [26] Kahlmeter G. An international survey of the antimicrobial susceptibility of pathogens from uncomplicated urinary tract infections: the ECO SENS Project. J Antimicrob Chemother 2003;51: [27] Grabe M, Hellsten S. Bacteriuria, a risk factor in men with bladder outflow obstruction. In: Kass EH, Svanborg-Edén C, editors. Host-parasite interaction in urinary tract infection. Chicago, Boston: University of Chicago Press; p