European Urology

1. Introduction

In 2008, prostate cancer was diagnosed in 340 000 men and caused nearly 71 000 deaths across the European Union [1]. In the United States, it is estimated that >1 million ultrasound-guided prostate needle biopsies (PNBs) are performed annually [2]. Complications include hematuria, hemospermia, bacteremia, and urinary tract infections (UTIs) [3]. Fluoroquinolones (FQs) are the most frequently used antibacterial prophylaxis before transrectal prostate biopsy. They are attractive because of their high bioavailability, antimicrobial spectrum, and high concentrations in prostatic tissues. Randomized controlled trials have demonstrated their efficacy in decreasing the risk of sepsis after PNB [4] and [5], and ciprofloxacin for 1 d is currently recommended by European, US, and Canadian expert committees [6], [7], and [8].

Lately, increasing post-PNB infectious complication rates have been observed in Europe and North America [9], [10], [11], and [12]. Some of these prior studies have mentioned the potential impact of increasing FQ resistance on antibacterial treatment prior to biopsy, but they were descriptive, lacked a control group [10] and [11], did not evaluate risk factors [11], or lacked a microbiologic definition of the outcomes [9] and [12]. Consequently, no mitigating strategy has been proposed.

To delineate the respective contribution of antimicrobial resistance, patient-related risk factors, and procedure-related risk factors in the emergence of post-PNB infections, we conducted a case–control study of patients diagnosed with post–prostate biopsy infections between 2002 and 2011 in a secondary- and tertiary-care hospital in Canada.

2. Patients and methods

3. Results

3.1. Characteristics of cases and incidence

We identified 5798 PNBs during the study period, following which we documented 48 cases of postprocedure infections (42% with bacteremia). Median age of cases was 66.7 yr (interquartile range [IQR]: 61.8–72.0 yr). Incidence was 0.71% in 2002–2003, 0.65% in 2004–2005, 0.15% in 2006–2007, and 0.69% in 2008–2009, but it increased to 2.15% in 2010–2011 (Fig. 1). Cases occurred a median of 3 d (IQR: 1–7 d) after PNB. Thirty-two patients (66.7%) needed to be hospitalized for a median duration of 4 d (IQR: 3–5 d). Of these 32 patients, 5 were admitted to the ICU (duration, range: 2–6 d), and one needed vasopressors. No patient died within 30 d of his infection.

Fig. 1 Incidence of infectious complications after transrectal ultrasound–guided prostate biopsy: Centre Hospitalier Universitaire de Sherbrooke, 2002–2011.

Escherichia coli was the most frequent pathogen (n = 36; 75%) among cases, followed by Enterobacter and Citrobacter species (n = 3; 6% each). Gram-positive cocci accounted for only 10% of cases. Among the 42 cases with aerobic or facultative Gram-negative rods (GNRs), the following susceptibility rates were observed: ciprofloxacin, 20 of 42 cases (48%); trimethoprim-sulfamethoxazole, 31 of 41 cases (76%); gentamicin, 32 of 41 cases (78%): ceftriaxone, 26 of 28 cases (93%); and meropenem, 26 of 26 cases (100%).

Among the 46 cases for whom it was possible to identify the treatment received (for 2 patients, emergency room notes did not mention any treatment plan), 27 patients (59%) were treated with an initial regimen that included ciprofloxacin. Ciprofloxacin-resistant bacteria were isolated in all 5 patients who did not receive an adequate antibacterial treatment ≤24 h after presentation.

4. Discussion

The incidence of post-PNB infections has increased in several centers in North America and Europe [9] and [12], and we now document a similar trend in another region of Canada. While the underlying causes have not yet been delineated, the geographic spread of this problem suggests the presence of some systemic factors.

We found no evidence that deficiencies in infection control could explain this increasing risk of post-PNB infections. Cases did not cluster; the risk did not vary according to the facility at which the biopsy was performed or with the rank of cases within the daily program. Some authors have suggested that a higher number of needle core samples may translate into a higher incidence of infection [9]. The current practice in our center is to obtain 12 needle core samples, and we found no difference in the number of needle core samples between cases and controls. In univariate analyses, post-PNB infections were associated with the procedure being carried out by urologist 3, who, between 2002 and 2005, performed biopsies simultaneously with cystoscopy without screening for asymptomatic bacteriuria. This practice was abandoned in 2005.

Having a biopsy performed in 2010–2011, recent hospitalization, diabetes mellitus, and COPD were independent risk factors for a post-PNB infection. However, the number of patients with diabetes, COPD, and recent hospitalization was limited, especially in 2010–2011, and cannot account for temporal changes in risk. While the prevalence of diabetes mellitus has increased in our population [15], as elsewhere in North America, this fact would not by itself suffice to explain the substantial increase in incidence during 2010–2011. Furthermore, the prevalence of COPD has not changed, the number of hospital admissions at our center is stable, and performance of the procedure in 2010–2011 remained associated with the outcome even after adjustment for these other independent risk factors.

Several arguments suggest that the higher risk of infection in 2010–2011 was driven by an increase in ciprofloxacin resistance in E coli and a concomitant decrease in the efficacy of ciprofloxacin prophylaxis in patients undergoing PNB. First, over the last decade we have experienced an increase in the prevalence of resistance to ciprofloxacin among E coli strains isolated from patients with UTIs in general, presumably reflecting changes in the strains that colonize the patients’ gastrointestinal tracts [16]. Second, the recent increase in post-PNB infections was more pronounced for ciprofloxacin-resistant than for ciprofloxacin-susceptible pathogens (Fig. 1), and there was a commensurate decrease in ciprofloxacin susceptibility among pathogens isolated from patients with post-PNB infections in 2010–2011 (Table 3). Third, even among E coli isolates deemed susceptible, the MIC increased during the latter period. Taken together, these results suggest that ciprofloxacin resistance was indeed the main factor responsible for the recent increase in incidence of postbiopsy infections in our center. Prospective studies would also make it possible to determine whether an FQ-resistant sequence type 131 E coli played a role in the rise of postbiopsy sepsis, as in other centers [17].

Recent hospitalization is a risk factor for community-onset febrile UTIs caused by ciprofloxacin-resistant bacteria [18] and was recently shown to be a risk factor for post-PNB infections in England [19]. This risk factor most likely represents acquired colonization with pathogens with decreased sensitivity to FQ. Diabetic patients have an increased propensity to develop infections in general, and hyperglycemia-related impairment of the immune response may lead to an increase in postbiopsy infections. In vitro studies have shown that glycosuria enhances the growth of uropathogenic E coli[20], and diabetes was an independent risk factor for fever after PNB in a recent post hoc analysis of a European randomized trial [12]. Nicotine use delays primary healing and increases the risk of some infections [6], which might explain the association with COPD. COPD and diabetes might also represent residual confounding. We tried to assess previous antibacterial exposure, but the retrospective nature of our work and lack of access to some outpatient antibacterial prescriptions raise the possibility that these two factors, and especially COPD, may represent surrogate markers for outpatient antibiotic prescription and colonization with FQ-resistant GNRs.

How should this increased risk of post-PNB infections be addressed? In the context of increasing ciprofloxacin resistance and creeping MICs, the current recommendation of using a ciprofloxacin prophylaxis of duration ≤24 h needs to be reexamined [8]. In patients colonized with bacteria harboring a ciprofloxacin MIC of 0.5–2 μg/ml (ciprofloxacin sensitive/intermediate), prostatic concentrations following a single dose of ciprofloxacin may fall below these MICs 24–48 h after the procedure [21], when there is still an iatrogenic epithelial break.

An alternative would be to give all patients a single dose of an antibiotic toward which E coli and other GNRs are less likely to be resistant, for example, amikacin. This drug achieves concentrations in the prostate that are lower than quinolones, but amikacin is bactericidal and has urine concentrations that exceed peak plasma levels 25- to 100-fold [22]. The drug may therefore be effective against the urine and bloodstream isolates seen in our patients. Amikacin was also recently shown to reduce post-PNB infections in addition to supplying an FQ-based prophylaxis [23], as did gentamicin in another study [24]. A single dose would not lead to significant nephrotoxicity or ototoxicity. However, we emphasize that antibiotic prophylaxis for PNB needs to be tailored to local resistance patterns. For example, isolates with CTX-M β-lactamases frequently exhibit cross-resistance to amikacin [25]. In centers at which extended-spectrum β-lactamase strains are highly prevalent [26], carbapenems could be considered for prophylaxis—especially ertapenem, given its long half-life. Such alternate prophylaxis should be considered especially for patients who have received an FQ in the last year, who are more likely to be colonized with FQ-resistant E coli[16].

Another approach would be to screen for ciprofloxacin-resistant GNRs in rectal swabs prior to a transrectal prostate biopsy (using selective media), followed by a course of a customized prophylaxis regimen [27] and [28]. However, this procedure would represent a substantial burden for clinical microbiology laboratories and would fail to detect ciprofloxacin-sensitive isolates with borderline MICs, which accounted for half of our post-PNB infections in 2010–2011. Another problem is that because of the relative instability of ciprofloxacin in broths or agars, there is no commercially available ciprofloxacin-containing media. Screening for asymptomatic bacteriuria would presumably be of little help, as demonstrated by others [29], given that most causal pathogens probably originate from the rectum.

Our study also provided hints for the improved management of patients presenting with sepsis after a PNB, a serious infection whose treatment required hospital admission in two-thirds of cases and ICU admission in 15% of cases. The use of ciprofloxacin as an empirical antimicrobial was inappropriate given the frequency of resistant strains. Empirical treatment should include either a third-generation cephalosporin or a carbapenem (if extended-spectrum β-lactamase strains are a concern) and vancomycin for Gram-positive coverage.

Our study had limitations. First, some patients who lived outside Sherbrooke may have presented to local hospitals when developing post-PNB infections, causing an underestimation in the true frequency of this complication. Second, the measurement of antibacterials ordered by physicians working outside the hospital was suboptimal and may have led to residual confounding. Third, the exact prescription for antibacterial prophylaxis (a single dose compared with a 3-d regimen) was generally not documented in the progress notes.

5. Conclusions

We found that emerging resistance to ciprofloxacin is the most likely cause of the increasing risk of infectious complications after PNB. Creeping ciprofloxacin MICs may even contribute to the increase in cases caused by FQ-susceptible strains. Novel approaches for antibacterial prophylaxis need to be designed and evaluated.

Author contributions: Alex Carignan had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Carignan, Valiquette, Pépin.

Acquisition of data: Lapointe, Roussy.

Analysis and interpretation of data: Carignan, Pépin.

Drafting of the manuscript: Carignan.

Critical revision of the manuscript for important intellectual content: Carignan, Valiquette, Sabbagh, Roussy, Lapointe, Pépin.

Statistical analysis: Carignan.

Obtaining funding: Carignan, Pépin.

Administrative, technical, or material support: None.

Supervision: Carignan.

Other (specify): None.

Financial disclosures: Alex Carignan certifies that all conflicts of interest, including specific financial interests and relationships and affiliations relevant to the subject matter or materials discussed in the manuscript (eg, employment/affiliation, grants or funding, consultancies, honoraria, stock ownership or options, expert testimony, royalties, or patents filed, received, or pending), are the following: Dr. Pepin has served on advisory boards for Pfizer, Wyeth, Ortho, Merck, Acambis, Iroko, and The Medicines Company. Dr. Valiquette has served on advisory boards for Oryx, Iroko, Abbott, and Wyeth and has received compensation to conduct clinical trials involving antibacterials from Genzyme, Wyeth, Pfizer, BioCryst, Trius, Cempra, Optimer, and Arpida. Dr. Carignan has served on advisory boards for Merck, Pfizer, and Astellas Pharma. Dr. Sabbagh has served on advisory boards for GSK, Watson, and Amgen. Other authors: no reported conflicts of interest.

Funding/Support and role of the sponsor: None.

Acknowledgment statement: We would like to thank Isabelle Alarie and Julie Perron for their editorial input and important intellectual content and Cynthia Grenier for her support in conducting the study.