Prostate cancer is the most frequently diagnosed tumor in elderly men and is the sixth leading cause of cancer death among men worldwide [1]. With the widespread popularity of prostate-specific antigen (PSA) testing, an ultrasound-guided transrectal prostate biopsy has become one of the most common urological procedures. Many men are now undergoing transrectal biopsy owing to increasing awareness of prostate cancer. Thus, with the increased incidence of prostate biopsy, there has been a concomitant increase in complications of infection. Clinically significant infectious complications include urinary tract infection (UTI), acute bacterial prostatitis, epididymo-orchitis, and sepsis.

According to the European Association of Urology guideline, oral or intravenous antibiotics are essential before biopsy. Quinolones are the drug of choice, with ciprofloxacin superior to ofloxacin [2]. Resistance to quinolones is on the rise, however, owing to their increased use. Infectious complications are a continued threat following transrectal ultrasound (TRUS)-guided prostate biopsies. The incidence of serious complications requiring hospital admission ranges from 0.1% to 2.5%, most often secondary to UTI, fever, or sepsis [3]. Recently, quinolone resistance has increased rapidly worldwide and is especially high in Korea. Quinolone resistance of Escherichia coli in Korean hospitals had already exceeded 30% in 2004 [4].

Increasing core number during prostate biopsy is a recent trend to achieve a higher detection rate of prostate cancer. In 2007, an average of 8 cores was performed in our hospital; in 2012, however, the number increased to 12 cores for prostate biopsy. These factors may also increase the risk of infection. Since 2010, postbiopsy febrile complications have increased in Chung-Ang University Hospital. Accordingly, we evaluated the incidence of infection after prostate biopsy and additionally identified risk factors retrospectively.

From 2007 to 2012, 1,195 men underwent TRUS-guided prostate biopsy. The mean age of the patients was 66.3±9.8 years and their mean PSA level was 12.5 ng/mL. The mean prostate volume was 50.6±21.4 mL. All men in the biopsy group were Asian. The most common comorbidity was hypertension (41.9%). Among the 1,195 subjects, 409 (34.2%) were diagnosed with prostate cancer (Table 1).

The incidence of prostate biopsy and the number of cores taken during biopsy gradually increased. Prostate biopsy at our institution increased from 127 cases in 2007 to 253 cases in 2012. An average of 8.43 cores was taken in 2007, and this value increased to 11.74 cores in 2012. Quinolone was widely used before prostate biopsy.

Overall, 39 men (3.1%) were hospitalized within 7 days after biopsy owing to febrile UTI. They were treated by use of third-generation cephalosporin. There were no mortalities.

Table 2 illustrates the univariate analyses conducted to identify potential independent risk factors associated with febrile UTI. There was no significant relationship between the development of febrile UTI and patient age, serum total PSA, prostate volume, hypertension, diabetes, or BMI.

Of the 409 patients who had adenocarcinoma on biopsy, only 7 patients (1.71%) developed febrile UTI. This suggested that prostate cancer was not responsible for the infectious complication.

In the univariate analysis, complications were directly correlated with the number of needle cores in patients who underwent <12 vs. >12 cores (p=0.024). However, the multivariate logistic regression analysis did not show any significant correlations between febrile UTI and the number of cores after adjustment for confounders (p=0.365) (Table 3).

Of 210 patients with a history of diabetes, only 4 patients (1.9%) developed febrile UTI. Patients with BMI>25 kg/m² had an increase in frequency of febrile UTI (5.2%). BMI>25 kg/m² was a definite risk factor for the development of febrile UTI after prostate biopsy in univariate (p=0.004) and multivariate (p=0.011) analysis.

The overall incidence of infectious complications was 3.1%. When stratified by year, the incidence of infectious complications was 5 times higher in 2010-2012 than in 2007-2009, with statistical significance in the univariate (p=0.014) and multivariate (p=0.035) analysis. Fluoroquinolone resistance was more common in the latter period. To predict the probability of infection on the basis of the multivariate analysis result, in 2010, actual infection was higher than expected. This coincided with the quinolone resistance that occurred in our hospital (Fig. 1).

Of the 39 patients with febrile UTI, 25 patients (67.6%) had positive urine or blood cultures. E. coli was the most frequent pathogen (n=20, 80%), followed by Enterococcus faecalis (n=2) and Pseudomonas aeruginosa, Proteus mirabilis, and Burkholderia cepacia (n=1). Quinolone resistance was observed in 70% of patients who developed febrile UTI. Resistance patterns were most commonly seen with ampicillin (85.7%), piperacillin (75%), fluoroquinolone (70%), cephalothin (57.1%), and amoxicillin/clavulanic acid (33.3%). However, 90% or greater susceptibility was demonstrated for amikacin, imipenem, piperacillin/tazobactam, cefazolin, tigecycline, and aztreonam. Table 4 shows a more comprehensive list of susceptibility patterns for bacteria in urine and blood.

The efficacy of fluoroquinolone was initially hypothesized to result from its potent activity against a large spectrum of clinically relevant pathogens, especially E. coli; additionally, excellent prostatic tissue levels were obtained after oral administration. This assumption was further confirmed by the high rates of success when fluoroquinolone was used prior to biopsy procedures [5,6].

However, the routine prophylactic application of fluoroquinolones has posed a problem. Our results indicated that the incidence of infection was similar to that reported in other studies. Recently, the incidence of quinolone resistance has increased further with an increased ratio of quinolone-resistant E. coli as well. This suggests that the increased prevalence of quinolone-resistant E. coli is the cause of infection. Many recent reports have shown that fluoroquinolone-resistant infection after prostate biopsy is on the rise [7].

The likely reason for increased quinolone resistance could be increased use. Since the 1980s, the broad-spectrum fluoroquinolone ciprofloxacin has been the antibiotic of choice for the treatment of UTIs, mostly because of its potent activity against a large spectrum of clinically relevant pathogens. Between 2008 and 2012, fluoroquinolone resistance gradually increased from 33% to 85%. This phenomenon has prompted an ongoing search for improved prophylactic antibiotic protocols.

South Korea is a country with especially high quinolone usage, with daily defined doses much higher than in other countries [8]. It would be particularly useful to reduce the use of fluoroquinolone. Any move toward a better antibiotic regimen should be based on comparative prospective randomized studies. Perhaps the use of less common antibiotics should be reserved for prophylaxis and treatment of infection in the context of antibiotic cycling. These practices would decrease the likelihood of introducing resistance mechanisms against them. Furthermore, no study to date, including the present study, has demonstrated that extended dosing as a prophylactic regimen is superior to single dosing.

It may be important to consider the replacement of antibiotics or the use of individual customized antibiotics. The choice of antibiotics is crucial in determining the outcome. Duplessis et al. [9] and Taylor et al. [10] proposed that rectal cultures be obtained before TRUS-guided biopsy to identify fluoroquinolone-resistant flora to facilitate targeted antibiotic prophylaxis and decrease the overall cost of care. This would complicate the adherence to a specific treatment protocol. Resistance profiles could vary among isolates indicating that a single treatment algorithm would not work for all bacteria. Thus, follow-up of culture results and the customization of each patient's treatment regimen are indicated. Additionally, a regular review of regional and local resistance patterns is also recommended, because these reveal resistance patterns of various bacterial species and could guide the practitioner toward better treatment options.

According to our data, amikacin, imipenem, piperacillin/tazobactam, cefazolin, and tigecycline were useful. Compared with fluoroquinolones, imipenem had demonstrably much lower resistance rates. However, the inclusion of such agents in a routine prebiopsy preparation protocol will involve greater expense and the obligation of injections. Also, imipenem use must be carefully considered to avoid creating resistance to these agents. Carignan et al. [11] recommended the use of amikacin or carbapenem.

An analysis of the resistance demonstrated that penicillins had the highest level of resistance. We observed high levels of ampicillin resistance (85.7%), similar to other reports. Ampicillin should not be considered as a first-line antibiotic for treatment of postbiopsy febrile UTI.

Some studies identified diabetes or increased core number as the cause of postbiopsy febrile UTI [12,13], unlike the current findings. Our study shows obesity as a risk factor for infection, but not diabetes. In our opinion, patients diagnosed with diabetes are getting close to glucose control, whereas patients identified as obese could include those with glucose intolerance or undiagnosed diabetes. In addition, obese patients are known to be vulnerable to infection [14].

In 14 of the 39 cases (32.4%), no bacteria were isolated from either blood or urine. However, antibiotic treatment resulted in the improvement of clinical symptoms and leukocytosis. Negative cultures in the setting emphasized the importance of taking multiple blood cultures from different sites.

Our study had several important limitations. First, the sample size of the groups with infection was limited because of the low incidence of such complications in general. Second, it was a retrospective nonrandomized study using data derived from an Institutional Review Board-approved database and the medical records of the enrolled patients. The impetus to initiate an investigation on febrile UTI was determined solely by the patient's symptoms and not by a standardized collection of urine or blood cultures for all patients after biopsy. Thus, there may have been a risk of missing cases of asymptomatic bacteriuria, although the clinical implications of this would be limited. Third, the initial design of the study did not consider a previous history of quinolone use.

The increase in UTI was due to quinolone-resistant bacteria. Therefore, alternative prophylactic antibiotics for prostate biopsy must be researched further. The physician is compelled to consider several factors and customize antibiotic prophylaxis on an individual basis. Postbiopsy infections after standard fluoroquinolone prophylaxis are suggested to be treated with regimens such as imipenem, piperacillin/tazobactam, or cefazolin until culture findings are available to guide therapy.