Ciprofloxacin resistance in the faecal carriage of patients undergoing transrectal ultrasound guided prostate biopsy

Authors


  • S.T. and J.M. contributed equally to this work.

Abstract

What's known on the subject? and What does the study add?

  • Transrectal ultrasound guided prostate biopsies (TRUSBx) are associated with a spectrum of complications, including most significantly infection, which affects up to 5% of patients. In the most severe cases, infection leads to sepsis, a life-threatening complication. Escherichia coli is the primary responsible pathogen. Although antibiotic prophylaxis with fluoroquinolones is routinely used, there is evidence that the infection rate after TRUSBx is increasing, and this appears to be due to an increasing prevalence of ciprofloxacin-resistant rectal flora.
  • This is the largest prospective clinical trial to date analysing the rectal flora of men undergoing prostate biopsies. We determined the microbial and antibiotic sensitivity profiles from 849 patients. Ciprofloxacin-resistant Gram-negative organisms were identified in the rectal flora of 19.0% of men. Furthermore, fluoroquinolone use within 6 months preceding a TRUSBx and the presence of a prosthetic heart valve were significant predictors of ciprofloxacin resistance on rectal swab. Determining the prevalence of rectal fluoroquinolone resistance has important implications in evaluation of the suitability of prophylactic regimens. Antimicrobial profiles derived from rectal swabs pre-biopsy may prove useful in guiding targeted antibiotic prophylaxis.

Objectives

  • To establish the prevalence of ciprofloxacin-resistant bacteria in patients undergoing transrectal ultrasound guided prostate biopsies (TRUSBx) and to determine whether this predicts subsequent infectious complications.
  • To identify risk factors for harbouring ciprofloxacin-resistant flora.

Patients and Methods

  • Any patient undergoing a TRUSBx from 2009 to 2011 was eligible for enrolment in this prospective study.
  • Pre-biopsy rectal and urine cultures and post-biopsy urine cultures were obtained and antimicrobial susceptibility was determined.
  • Univariate and multivariate analyses were performed to identify independent patient risk factors associated with ciprofloxacin-resistant rectal flora.

Results

  • A total of 865 patients underwent TRUSBx, of whom 19.0% were found to have ciprofloxacin-resistant Gram-negative coliforms.
  • Escherichia coli was the most prevalent Gram-negative rectal isolate (80.9%) and accounted for 90.6% of ciprofloxacin resistance.
  • Patient characteristics that conferred an increased risk of harbouring ciprofloxacin-resistant organisms included a history of a heart valve replacement (P < 0.05) and ciprofloxacin use in the past 3 months (P < 0.05).
  • Infectious complications were observed in 3.6% (n = 31) of the patient population and 48% of these patients grew ciprofloxacin-resistant organisms on the pre-biopsy rectal swab (P < 0.001).

Conclusions

  • Antimicrobial resistance to ciprofloxacin in the rectal flora was common, particularly in patients with recent ciprofloxacin use and a heart valve replacement.
  • Despite a significant correlation between those patients who developed infections and the detection of ciprofloxacin-resistant organisms, only 9.0% (n = 15) of the total group with ciprofloxacin resistance developed an infectious complication.
  • Future studies will need to evaluate the cost effectiveness and clinical utility of a pre-biopsy rectal culture in targeting antibiotic prophylaxis.
Abbreviations
TRUSBx

transrectal ultrasound guided prostate biopsy

MIC

minimum inhibitory concentration

Introduction

Despite routine antibiotic prophylaxis, infectious complications are a continued threat following transrectal ultrasound guided prostate biopsies (TRUSBx). It is theorized that bacterial flora harboured in the rectum are introduced into the genitourinary system or systemically into the bloodstream following perforation of the rectal mucosa with the TRUSBx needle [1]. Sepsis is the most serious and potentially life threatening complication. Escherichia coli is the primary organism worldwide causing infection in patients receiving a TRUSBx [2-4].

The American Urologic Association Best Practice Policy Statement on Urologic Surgery Antimicrobial Prophylaxis currently recommends a single dose of a fluoroquinolone before TRUSBx as a first-line agent to reduce infectious complications [5, 6]. Despite the proven efficacy of these first-line agents, there has been an increasing rise in post-biopsy infectious complications [7]. For example, a Canadian retrospective study evaluated 75 000 Canadian men who underwent TRUSBx between 1996 and 2005 and determined that infection-related hospital admissions had gone up fourfold in this time period (P < 0.001) [7]. The increase in hospital admissions is thought to be a consequence of an increasing prevalence of ciprofloxacin-resistant rectal flora [1, 4, 8].

Ciprofloxacin resistance is a concern because it is an independent risk factor for mortality in patients with E. coli bacteraemia [9]. If bacterial growth on a rectal swab culture proves to be a significant predictor of clinical infection, then determining the prevalence of rectal fluoroquinolone resistance may have important implications in the evaluation of the suitability of prophylactic regimes. The antimicrobial profile of the rectal swabs may also be useful in aiding the empirical treatment of patients presenting to hospital with post-biopsy infectious complications.

In this prospective trial we sought to investigate the antimicrobial resistance profiles of the rectal flora of patients undergoing a TRUSBx, define populations at risk for harbouring ciprofloxacin-resistant organisms and evaluate the utility of a pre-biopsy rectal swab in predicting clinical infection. Here we report the microbiological growth and patterns of antibiotic resistance in isolates from rectal cultures taken immediately prior to TRUSBx.

Subjects and Methods

Study Setting and Patient Population

The protocol for the study was approved by the University of British Columbia's Clinical Ethics Review Committee and was carried out at Vancouver General Hospital. During the period from December 2009 to April 2011, 1069 men were screened for participation. Any patient for whom a TRUSBx was ordered according to standard clinical reasoning and judgement was considered eligible for this study. In all, 229 patients (26.5%) were undergoing repeat biopsy on active surveillance for low risk prostate cancer. At the initiation of the study the exclusion criteria were as follows: (i) inability to provide consent; (ii) allergy to ciprofloxacin; (iii) ciprofloxacin use for any reason in the past 3 months; (iv) history of UTI within the past 3 months; (v) history of prior UTI or sepsis after TRUSBx; and (vi) allergy to iodine or povidone-iodine. Criteria (iii), (iv) and (v) were eliminated after the first 450 patients because of the low number of observed infectious complications. The criteria were initially set to ensure a homogeneous study cohort, but they were found to be overly restrictive and eliminated the patients at higher risk for infectious complications.

Study Design

This was a prospective randomized clinical trial designed to test the efficacy of rectal cleansing with povidone-iodine prior to TRUSBx in order to reduce the rate of subsequent infectious complications. A secondary endpoint in the study was to evaluate the rectal and urine flora in these patients before and after biopsy. Here we report on this secondary endpoint only and will discuss the primary endpoint separately.

Following provision of informed consent, the patients were provided with instructions for a 3-day course of ciprofloxacin prophylaxis (1000 mg extended release ciprofloxacin once daily) commencing the day prior to biopsy and the use of a Fleet® enema at home approximately 2 h before biopsy. In addition, the men were asked to provide midstream urine samples more than 48 h before and 48 h after the biopsy.

Rectal cultures were procured immediately prior to prostate biopsy using cotton-tipped culture swabs in standard collection systems without enrichment and were transferred to the microbiology laboratory for processing within 60 min. Specimens were handled according to usual microbiological procedures. All specimens were plated to 5% sheep blood and MacConkey agars and incubated 18–20 h in ambient air at 35 °C. After this incubation period the plates were evaluated for growth. Plates without growth after the full incubation period were reported as ‘no growth’. Plates exhibiting only mixed Gram-positive organisms by visual inspection and lack of growth on MacConkey agar were reported as ‘mixed Gram-positive growth, no Gram-negative organisms present’. If Gram-negative organisms were present, up to three different morphotypes were selected for further identification and susceptibility testing by microbroth dilution using a commercial automated identification and susceptibility system (Phoenix, BD Diagnostics, Franklin Lakes, NJ, USA). Interpretation of susceptibility and resistance was based on Clinical Laboratory Standards Institute guidelines, with susceptible defined as ciprofloxacin minimum inhibitory concentration (MIC) ≤1 μg/mL and resistance as a ciprofloxacin MIC ≥4 μg/mL [10].

In order to test our primary endpoint (see above) patients were randomized to undergo TRUSBx either with or without rectal cleansing. The rectal cleansing was administered via DRE, which was performed on all subjects. In the group receiving rectal cleansing (treatment group), the physician performing the DRE used a gauze soaked with povidone-iodine over the index finger to wipe back and forth across the prostate at least five times from one lateral margin to the other.

All prostate biopsies were performed transrectally with an 18-gauge Tru-Cut® needle under ultrasound guidance after local infiltration of bupivacaine. The number of cores taken corresponded to prostate volume: 12 cores for >50 mL, 10 cores for 35–50 mL and eight cores for <35 mL.

Following the biopsy the patients were asked to measure and record their oral temperature every 6 h for 48 h after the procedure. The patients were contacted by telephone 7 days after the procedure and a post-biopsy questionnaire was performed to screen for infectious complications including fever (temperature ≥38.0 °C), symptoms of UTI and presentation to hospital or physician for any reason.

Variables analysed as possible predictors of a ciprofloxacin-resistant rectal culture included demographic characteristics, comorbid medical conditions (diabetes, hypertension, haemorrhoids, coagulopathy, prior pelvic radiation, immunosuppression), antibiotic use in the past 3 months, recent UTI and prostate-specific factors (PSA, previous biopsies, prior prostate cancer and prostate volume).

Data Analysis

Descriptive statistics were computed pre-biopsy for patient demographic characteristics and clinical pre-procedural factors. Univariate analysis was performed using an unpaired t test for continuous variables and Pearson's exact two-tailed χ2 test for categorical variables. Variables determined to have a P < 0.05 in univariate analysis were further evaluated in multivariate analysis using logistic regression to assess for independent risk factors associated with ciprofloxacin-resistant faecal carriage.

Results

Patient Characteristics

Of the 1069 men initially screened for inclusion in our study, 65 patients were exluded based upon the above exclusion criteria, 118 refused and 21 dropped out. A total of 865 patients with a mean age of 65.2 (range 37–87) years were enrolled and underwent prostate biopsy. Patient demographic and clinical characteristics are presented in Table 1.

Table 1. Pre-biopsy patient characteristics categorized by rectal swab profile.
 TotalCiprofloxacin-resistant Gram-negative organismsAll other growthP
Total (%)MeanRangeNo. (%)MeanRangeNo. (%)MeanRange
  1. Positive urine culture >100 CFU/m of E. coli. *Paired t test. Two-tailed Fisher's exact test.
Age86565.237–8716666.246–8068465.037–870.168*
PSA8337.90.1–6271627.70.2–576718.00.1–6270.585*
Prostate volume (mL)83942.64–93416543.04–16967442.55–9340.652*
Rectal swabs taken849 (98.2)  166 (19.5)  684 (80.5)   
Medical comorbidities          
Hypertensive262 (31.6)  62 (37.3)  200 (29.2)  0.049
Diabetic76 (9.0)  20 (12.0)  56 (8.2)  0.089
Easy bleeder73 (8.6)  20 (12.0)  53 (7.7)  0.089
Suppressed immunity11 (1.3)  4 (2.4)  8 (1.2)  0.800
Haemorrhoids157 (18.5)  34 (20.5)  123 (18.0)  0.584
Diagnosed prostate cancer162 (19.1)  34 (20.5)  128 (18.7)  0.584
Prior pelvic radiation10 (1.2)  1 (0.6)  9 (1.3)  0.696
Heart valve replacement16 (1.9)  8 (4.8)  8 (1.2)  0.006
Prostatitis19 (2.2)  5 (3.0)  14 (2.0)  0.394
UTI in past 3 months21 (2.5)  9 (5.4)  12 (1.7)  0.011
Ciprofloxacin use in past 3 months23 (2.7)  11 (6.6)  12 (1.7)  0.002
Other antibiotic use in past 3 months73 (8.6)  21 (12.7)  52 (7.6)  0.044
Previous biopsy221 (26.0)  45 (27.1)  176 (25.7)  0.621

Microbiological Characteristics

Rectal swabs were adequately obtained in 849 patients (98.2%). Quantitative analysis of the rectal swab cultures revealed 47.7% (n = 405) isolated mixed Gram-positive growth, 3.5% (n = 30) no growth and 48.8% (n = 414) Gram-negative growth (Table 2). E. coli was the predominant Gram-negative faecal isolate (80.9%, n = 399), followed by Klebsiella (7.71%, n = 38) and Pseudomonas species (4.26%, n = 21) (Table 2).

Table 2. Rectal swab cultures and resistance profiles.
Growth patternTotal rectal swabs, n (% of 849)Gram-negative strain frequency, n (% of 493)Ciprofloxacin resistance, n (% of 161)ESBL, n (% of 39)
No growth30 (3.5%)
Isolated mixed Gram-positive growth405 (47.7%)
Gram-negative coliforms414 (48.8%)493 (100%)161 (100%)39 (100%)
E. coli365 (43.0%)399 (80.9%)145 (90.1%)37 (94.9%)
Klebsiella. spp37 (4.4%)38 (7.7%)3 (1.9%)2 (5.1%)
Pseudomonas spp.21 (2.5%)21 (4.3%)7 (4.4%)0 (0%)
Citrobacter spp7 (0.82%)7 (1.42%)0 (0%)0 (0%)
Stenotrophomonas maltophilia6 (0.71%)6 (0.12%)2 (1.3%)0 (0%)
Enterobacter spp.5 (0.59%)5 (1.01%)0 (0%)0 (0%)
Morganelli morganii4 (0.47%)4 (0.8%)2 (1.3%)0 (0%)
Proteus spp.4 (0.47%)4 (0.8%)0 (0%)0 (0%)
Other Enterobacteriaceae9 (1.1%)9 (1.8%)0 (0%)0 (0%)
Acinetobacter spp.3 (0.35%)3 (0.6%)0 (0%)0 (0%)
Comamonas spp.2 (0.24%)2 (0.4%)2 (1.3%)0 (0%)

Ciprofloxacin-resistant Gram-negative coliforms were detected in 19.0% (n = 161) of the patient population (Table 2). The ciprofloxacin resistance rates among the isolates containing E. coli and Klebsiella were 36.3% and 7.9% respectively (Table 2). Global antibiotic resistance profiles are presented in Fig. 1 for the Gram-negative coliforms procured. High levels of resistance were seen with other popular antibiotics including ampicillin (49.4%), trimethoprim-sulfamethoxazole (23.7%) and the first generation cephalosporins, cephalothin/cephalexin (18.8%). Resistance to the aminoglycosides gentamicin and tobramycin remained low (9%–11%). Coliforms expressing extended spectrum β-lactamase activity, conferring resistance to the extended spectrum cephalosporins such as ceftriaxone, were identified in 4.6% of the patient population. Resistance to ceftriaxone and ceftazidime was 8.1% and 1.4% respectively. The susceptibility to the carbapenem, imipenem and the combination piperacillin/tazobactam was excellent, exhibiting 99.98% and 99.9% susceptibility respectively.

Figure 1.

Antimicrobial resistance of Gram-negative coliforms (n = 493).

Pre- and post-biopsy urine cultures are depicted in Table 3. Only 118 patients (13.6%) delivered pre-biopsy urines, and asymptomatic Gram-negative bacteriuria was present in two patients (1.7%). Post-biopsy urine cultures were collected within 48 h following the procedure in 646 patients. No growth was seen in 90% of patients (n = 580) (Table 3). There were 15 cases of Gram-negative bacteriuria and five (33%) of these exhibited ciprofloxacin resistance (P < 0.001).

Table 3. Pre- and post-biopsy urine cultures.
  1. Urine cultures were obtained within 48 h before and after biopsy. Mixed Gram-positive growth is labelled ‘contamination’ and was not speciated.
Pre-biopsy urine cultures, N (% of 118) 
‘Contamination’28 (24%)
Gram-negative growth2 (1.7%)
Gram-positive growth4 (3.4%)
No growth84 (71.2%)
Post-biopsy urine cultures, N (% of 646) 
‘Contamination’48 (7.4%)
Gram-negative growth15 (2.3%)
Gram-positive growth3 (0.4%)
No growth580 (90%)

Risk Factors for Ciprofloxacin-Resistant Faecal Carriage

Table 1 illustrates the univariate analysis conducted to isolate potential independent risk factors associated with harbouring ciprofloxacin-resistant carriage. P < 0.05 was considered to be statistically significant. Patient characteristics that conferred an increased risk of harbouring ciprofloxacin-resistant organisms on univariate regression included (i) a history of a heart valve replacement (P < 0.001); (ii) a diagnosis of hypertension (P < 0.05); (iii) ciprofloxacin use in the past 3 months (P < 0.005); (iv) antibiotic use excluding ciprofloxacin in the past 3 months (P < 0.05); and (v) a history of UTI in the past 3 months (P < 0.05). These variables were subsequently incorporated into a multivariate regression (Table 4) to confirm their independent effects on ciprofloxacin resistance. Ciprofloxacin use in the past 3 months and the presence of heart valve replacement remained statistically significant in multivariate regression (P < 0.05). Patients with a recent UTI exhibited a trend for increased ciprofloxacin resistance; however, the difference was not statistically significant (P = 0.148). Age, haemorrhoids, medication-related immunosuppression, hypertension, anticoagulation, prior prostate cancer and pelvic radiation did not increase the risk of ciprofloxacin-resistant rectal flora (P > 0.05). No statistical association of prostatic factors including PSA, prostate volume, number of prostate cores and positive biopsy (benign vs cancer) to ciprofloxacin-resistant faecal carriage was observed.

Table 4. Multivariate analysis examining independent patient risk factors for harbouring ciprofloxacin-resistant faecal carriage.
Risk factorsAdjusted odds ratio95% CIP
Hypertension   
No1.00  
Yes1.330.895–1.9580.155
Heart valve replacement   
No1.00  
Yes3.0651.022–9.0890.0413
UTI in 3 months preceding TRUSBx   
No1.00  
Yes2.0050.762–5.0970.148
Ciprofloxacin use in 3 months preceding TRUSBx   
No1.00  
Yes2.8051.044–7.2950.036

Infectious Complications

Infectious complications were observed in 3.6% (n = 31) of the patient population: 18 patients (2.1%) suffered transient fever, two patients (0.2%) suffered an isolated UTI and 11 patients (1.3%) experienced sepsis (defined according to the American College of Chest Physicians and Critical Care Conference) [11]. Of the patients who developed sepsis or UTI, 77% (n = 11) harboured ciprofloxacin-resistant faecal carriage.

Discussion

This study is the largest prospective trial to date assessing fluoroquinolone resistance in the rectal flora of patients receiving a TRUSBx. We have demonstrated that ciprofloxacin-resistant Gram-negative organisms can be frequently (19.0%) identified in the rectal flora of men prior to TRUSBx. Infectious complications occur more frequently in men with ciprofloxacin-resistant Gram-negative rectal flora. However, only 15 of 161 men (9.3%) harbouring resistant organisms developed an infectious complication.

Four studies have evaluated fluoroquinolone resistance in the rectal flora of patients undergoing TRUSBx and have determined a prevalence of harbouring ciprofloxacin-resistant organisms ranging from 10.6% to as high as 22% (Table 5) [1, 4, 12, 13]. The variation in fluoroquinolone resistance between studies may reflect regional and temporal differences in antimicrobial susceptibilities [8]. Methodological differences between studies are also relevant. Although our rate of resistance is relatively high, several factors may have led us to underestimate the detection of ciprofloxacin coliforms. First, no rectal coliforms were detected in 30 patients (3.5%), which is probably due to errors in specimen collection, prior use of a prophylactic Fleet® enema, ciprofloxacin prophylaxis and the limits of laboratory detection. Second, detection of resistant coliforms would have been enhanced by selective media. Third, up to three coliform colonies per patient were selected for speciation and antimicrobial susceptibility testing, but it is conceivable that a small minority of patients could harbour more than three coliforms so that a ciprofloxacin-resistant organism was missed. Last, we would anticipate higher rates of fluoroquinolone-resistant organisms had we not had strict inclusion criteria for the first half of the study that excluded some patients at higher risk for infection.

Table 5. Literature on faecal carriage of patients undergoing TRUSBx.
 LocationYearsNFluoroquinolone resistanceInfection
  1. aWith targeted prophylaxis.
Steensels et al. 2001 [12]Belgium2003–200923622.0%3.0%
Batura et al. 2010 [1]UK2007–200844510.6%1.8%
Taylor et al. 2012 [13]USA2010–201111219.6%0%a
Liss et al. 2011 [4]USA2009–201013622.0%3.6%

Recent reports suggest that the rate of infectious complications after TRUSBx is increasing [8, 14] and that this increase is directly related to an increasing prevalence of ciprofloxacin-resistant organisms [7, 8, 11, 14, 15]. It is thought that the widespread use of fluoroquinolones selects for this resistance [1, 4], and statistical models have predicted that by 2013 the rate of fluoroquinolone-resistant E. coli will be as high as 45% in populations with high fluoroquinolone usage [16]. One case–control study retrospectively analysed data from 5798 patients undergoing TRUSBx at a Canadian hospital [14]. The authors determined that the incidence of sepsis post-TRUSBx increased from 0.52 per 100 patients between 2002 and 2009 to 2.15 per 100 patients between 2010 and 2011. Ciprofloxacin resistance assessed in blood cultures of patients with sepsis and urine cultures of patients with UTIs increased significantly in the same time period.

We determined that the use of ciprofloxacin in the 3 months prior to TRUSBx independently predicts the presence of ciprofloxacin-resistant faecal carriage. This is consistent with previous reports [16-19]. Yagci et al. [17] found that the prevalence of fluoroquinolone-resistant E. coli strains in faecal flora increased with the duration of fluoroquinolone therapy. The duration of time an individual will harbour fluoroquinolone-resistant faecal carriage after discontinuing antibiotics, however, is unknown. We also determined that a history of heart valve replacement is a risk factor for harbouring ciprofloxacin-resistant faecal carriage (P < 0.05). This, to our knowledge, has not been previously reported. One could speculate that patients with prosthetic valves have had significant antibiotic exposure for prophylaxis and treatment of endocarditis causing changes in their gut flora, but this will require verification in other studies.

As a consequence of the increasing rates of ciprofloxacin-resistant infections, a number of groups have evaluated the use of alternative antibiotics for empirical prophylaxis, particularly in areas with endemic fluoroquinolone resistance [20]. Antimicrobial profiles derived from pre-biopsy rectal swabs may prove useful in guiding regionally specific modifications of universal antibiotic prophylaxis. Based on the antimicrobial profile assessed in the rectal swabs of our patient population (who were all pre-treated with ciprofloxacin), the aminoglycosides or the third-generation cephalosporins have potential to exhibit reasonable efficacy as additional or alternative prophylactic agents. Batura et al. [20] supported the use of aminoglycosides as a prophylactic option in conjunction with ciprofloxacin, revealing a significantly lower rate of infection and bacteraemia (P = 0.002). The major limitation of aminoglycosides is the need for parenteral administration. Stenseels et al. [12] promote the use of ceftriaxone (third-generation cephalosporin) as an alternative prophylactic regime. At our centre, isolates with extended spectrum β-lactamase activity were identified in 4.6% of our patient population, which remains low. Despite near perfect susceptibility to imipenemen or piperacillin/tazobactam in our population, these agents are unsuitable for universal prophylaxis. They should be reserved for the management of more serious infectious complications so that new resistance patterns are not promoted.

An alternative to changing the choice of antibiotic for universal prophylaxis is to tailor prophylaxis to microbial resistance patterns on routine pre-biopsy rectal swabs. Taylor et al. [13] were able to eliminate infectious complications in 112 consecutive patients managed in this manner, compared with nine infectious complications among 345 patients given traditional empirical prophylaxis (P = 0.12). One potential limitation of this strategy would be cost, since we observed a clinically relevant infectious complication in only a small fraction of patients with ciprofloxacin-resistant faecal carriage. However, a cost–benefit analysis in the study by Taylor et al. determined that tailored prophylaxis led to highly significant cost savings.

Acknowledgements

This study was funded by a Team Grant from the Vancouver Coastal Health Research Institute. We thank the technicians in the Radiology and Microbiology Departments, especially Gary Franco and Matilda Nasariya, for their assistance in this study. We are also grateful to the team of radiologists (C. Zwirewich, A. Harris, J. Buckley and S. Chang) who conducted the biopsies, and the urologists who enrolled their patients (M. Gleave, A. So, M. McLoughlin, B. Chew, R. Paterson, M. Nigro, H. Fenster and C. Nguan).

Conflict of Interest

None declared.

Ancillary