Antibiotic prophylaxis and complications following prostate biopsies – a systematic review

During the past three decades, the epidemiology of prostate cancer (PCa) has changed dramatically and, excluding non-melanoma skin cancer, PCa is now the most common male cancer diagnosis in most Western countries [1]. The 4-8% annual increase in PCa incidence observed in high-resource countries is mainly attributed to a growing awareness of PCa and the widespread use of prostate-specific antigen (PSA) testing. Prostate biopsy remains the mainstay in the diagnosis of PCa, and systematic transrectal ultrasound-guided biopsies (TRUS-gb) of the prostate is the gold standard technique. The procedure may be associated with mild or serious adverse events, most importantly infectious complications [2-4], but no international consensus exists on the use of antibiotics prior to and after TRUS-gb. In Denmark (DK), approximately 10,000 TRUS-guided biopsy sets are performed annually, but the frequency of complications and infections remains unknown. We systematically reviewed the current evidence of post-biopsy complications and the use of prophylactic antibiotics and discussed the use of antibiotic regimens in DK.

METHODS

This review was drafted in accordance with the Prisma Guidelines [5]. Three different searches in the PubMed, Embase and Cochrane databases were performed using the search strings:

1: biopsy AND prostate AND (prophylaxis OR prophylactic) AND antibiotic

2: biopsy AND prostate AND infection AND bacteria

3: biopsy AND prostate AND resistance.

In PubMed, the limits set were “adult: 19+ years”, and “humans”. In Embase, limits were “humans”, “adult 18 to 64” and “Aged 65+ years” and publication year 2005-2015.

In the PubMed and Cochrane databases, publication year was limited to the period from 2005 to the present was used. Only randomised controlled trials (RCT) reported in English and full-text articles were included, but prospective trials originating from Europe were screened in order to further examine regimens and recommendations of countries in relative proximity to DK. Finally, we identified the literature originating from DK.

Based on titles, all relevant abstracts were read, and the selected articles were evaluated for eligibility.

A flow chart is presented in Figure 1. All retrospective studies, reviews and case reports were excluded, unless originating from DK, as were papers deemed irrelevant because of their subject, e.g. papers on other cancer forms than PCa.

All Danish urology departments performing TRUS-gb were identified and contacted to obtain information about biopsy technique and use of prophylactic antibiotic strategy, including antibiotic dose and duration used. For information about development of resistant bacterial strains in DK, we reviewed the DANMAP, a Danish programme for surveillance of antibiotic consumption and bacterial resistance in humans.

RESULTS

We identified 19 eligible trials: ten randomised trials, seven prospective trials and two retrospective trials. Four of the randomised trials originated from Europe, two from Asia, one from North America, two from

South America and one from Canada. Two retrospective

studies from DK were identified. A total of 295 papers were excluded; 18 were not available in English, 275 were irrelevant (e.g. case reports) and two papers were not available in full text. The study characteristics are listed in Table 1.

Post-biopsy complications were described in all of the included trials. Non-infectious complications, regardless of study type, included haematuria (14.3-62.3%), haemospermia (17.8-59.1%) and rectal bleeding (4.8-21%) [6-13]. The post-biopsy infectious complications ranged from post-bioptic positive urine cultures (0.9-18.1%), urinary tract infection (UTI) (0.8-9.3%), prostatitis (0.67-0.7%), epididymitis (0.4%) and dysuria (0.4-13.3%), to more severe complications such as fever (0.3-8.6%), need for hospitalisation (0.4-3.1%), sepsis (0.17-5.7%), septic shock (0.2-0.45%) and death (0.2%) [6, 7, 9-22].

Type of antibiotic

Five trials comparing different prophylactic antibiotic regimens were identified. In these regimens, the administration of antibiotics was initiated between 30 minutes and 24 hours prior to biopsy.

All of the five trials were RCTs. The first study randomised patients to either tosufloxacin or levofloxacin orally, both in a two-day administration [9]. The second study randomised patients into three groups receiving either a single dose of ceftriaxone i.m., a single dose of ciprofloxacin (CFLX) orally or a three-day course of CFLX orally [11]. The third study randomised patients into four arms receiving CFLX or levofloxacin as either single doses or as a three-day regimen [8]. The fourth study randomised patients to receive CFLX with or without the addition of a periprostatic injection of cephalosporin [12]. None of the studies demonstrated any differences in the infectious or non-infectious complication rates following TRUS-gb. However, the fifth study showed that quinolones are superior to chloramphenicol in reducing infectious complications [22]. The trial randomised patients in a four-armed study comparing oral CFLX as either a single-dose or a three-day regimen to chloramphenicol and norfloxacin orally, both as three-day regimens. The study found a significant reduction in the risk of post-biopsy infections favouring CFLX both as single-dose and as three-day regimen compared with chloramphenicol (p = 0.0003) and norfloxacin (p = 0.03).

Route and timing of administration of prophylaxis

No studies have compared intravenous to oral administration of prophylactic antibiotics prior to TRUS-gb. We identified three studies investigating the importance of the timing of the initial dose of antibiotic. A prospective trial evaluated two different antibiotic strategies during two consecutive years. In the first year, the first dose of antibiotic was administered immediately before the biopsy procedure, and in the second year the antibiotic was administered 24 hours before biopsy. The study concluded that administration 24 hours before biopsies resulted in a relative risk reduction for post-biopsy infection of 55% compared with antibiotics just prior to the procedure [15].

Another study randomised patients to oral levofloxacin 500 mg given either 30-60 minutes before TRUS-gb or immediately after the bioptic procedure, but found no differences in infectious complications [7]. Finally,

a prospective trial compared a single dose of oral CFLX 750 mg given either two hours before or in conjunction with the biopsy procedure, but failed to demonstrate a significant impact on the risk of infectious complications [13].

Duration of prophylaxis

Several studies have compared the duration of antibiotics during TRUS-gb administered either as a single dose, a one-day regimen or a three-day regimen. One study [14] compared a single dose of 400 mg norfloxacin with a regimen of 400 mg immediately before the biopsies followed by three days of 400 mg norfloxacin twice

daily. No statistically significant difference in the risk of complications (p = 0.07) was demonstrated. An additional six studies compared different durations of antibiotic prophylaxis for TRUS-gb procedures, but none were able to show a benefit of continuing prophylaxis for more than a single dose [6, 10, 11, 16, 22] or a three-day [8] regimen.

Isolated bacteria

Bacteria isolated from blood or urine were Escherichia coli in 30-100% of the affected cases, Klebsiella spp. in 4-15% and Pseudomonas spp. in up to 14%, but also

Enterococcus, Enterobacter, Staphylococcus aureus,

Proteus mirabilis and Streptococcus were found. One

trial reported infections to be caused by Gram-negative species (e.g. E.coli, Enterobacter and Klebsiella) in 81% of cases, but did not name the isolated bacteria strains [5, 8-21].

Resistance/sensitivity of bacteria

Several the bacterial strains isolated from the blood and/or urine of patients admitted with complications following TRUS-gb were found to be resistant to the most commonly used antibiotics, i.e. quinolones, cephalosporins, aminopenicillins (i.e. ampicillin, amoxicillin) and aminoglycosides [9, 10, 12, 15, 17-21]. Bacterial sensitivity towards the most commonly used antibiotics was demonstrated in several studies, including cephalosporins, carbapenem/imipenem/meropenem, and amikacin [9, 12, 15, 17, 20, 21]. In one study, rectal swabs were performed immediately prior to the prostate biopsy. Among 236 rectal cultures obtained, 58 contained bacteria resistant to CFLX; of these, 52 were E. coli [21].

Another trial also obtained rectal swabs immediately

before the biopsy procedure and found that 19% of patients were colonised with extended spectrum beta-

lactamase (ESBL)-producing E. coli (ESBL-PE) [18].

Risk factors

Several risk factors were associated with post-biopsy infections, including diabetes, a history of prostatitis, previous biopsies, fluoroquinolone consumption within two months prior to biopsy, pre-biopsy bacteriuria and

faecal carriage of fluoroquinolone resistant E. coli [13, 16, 18, 21]. One trial demonstrated that faecal carriage of ESBL-PE was associated with symptoms of UTI without microbiological evidence [18]. However, no association between faecal ESBL carriage and post-procedure symptomatic UTI was found. Also, prior use of quinolones or other antibiotics, as well as diabetes, were associated with a risk of carrying ESBL-PE before biopsy

in multivariate analysis [18]. A second trial found that chronic prostatitis and fluoroquinolone consumption within six months before biopsy were risk factors associated with faecal carriage of fluoroquinolone-resistant E. coli strains [21]. A further two trials found no correlation between infection and either the number of core samples taken or the procedure being a re-biopsy [16, 21].

Two Danish reports on prophylactic antibiotics and post-biopsy infectious complications have been published. A retrospective analysis demonstrated a post-

biopsy sepsis incidence of 0.91% (n = 4) among 438

patients who underwent TRUS-gb biopsies in the 2009-2011 period. The antibiotic prophylaxis used was 500 mg of CFLX and 500 mg of metronidazole taken before, six hours after, and the day after the procedure. E. coli with resistance towards CFLX was described as the causative pathogen in patients who experienced a post-bioptic septic complication [23].

The second study considered patients who underwent TRUS-gb at a single hospital in DK in the 2010-2013 period, and who were administered three different prophylactic antibiotic regimens, namely CFLX or the combination of pivmecillinam and amoxicillin/clavulanic acid for either one or three days [24]. Overall, the admission rates due to infectious complications were 1.8%, and the most commonly isolated pathogens were E. coli, Klebsiella pneumoniae and Enterococcus faecalis. Of these, 12.1-55.3% were found to harbour resistance towards CFLX and 12.1-21.1% towards cephalosporins. Among the E. coli isolates, up to 15.8% were ESBL-PE. When changing the prophylaxis from CFLX to the pivmecillinam and amoxicillin/clavulanic acid combination, a reduction in CFLX resistance from 55.3 to 12.1% was seen.

Prophylactic antibiotic strategies in Denmark

We identified 19 departments of urology in DK that perform TRUS-gb (Table 2). All facilities reported taking out 10-12 needle cores per biopsy set, and all had a standard prophylactic regimen prescribed to patients before biopsy. The most prevalent regimen was monotherapy with CFLX, which was reported by 13 facilities. CFLX in combination with metronidazole and amoxicillin + clavulanic acid in combination with mecillinam was reported by three and two facilities, respectively. There was much variation in the duration of prophylaxis, ranging from a single dose to a three-day regimen.

DISCUSSION

Complications following biopsy of the prostate range from mild and self-limiting to life-threatening conditions requiring prompt medical attention. Antibiotic prophylaxis is necessary to reduce the risk of serious infectious complications, but the evidence supporting specific regimens remains scarce.

The most commonly reported non-infectious complications following TRUS-gb were haematuria and haemospermia, both with occurrence rates reaching 60%. Rectal bleeding was also a frequent complication seen in up to 20% of patients. UTI and dysuria were the most common infectious complications with occurrence rates reaching 9.3% and 13.3%, respectively, whereas sepsis, septic shock and death were reported in rates of up to 5.7%, 0.45% and 0.2%, respectively. The differences in the reported incidence of complications may be attributed to different follow-up regimens, different disease definitions and different methods of data collection.

Risk factors associated with septic complications have not been rigorously defined, although comorbidity and previous use of antibiotics seem to be predisposing risk factors. Only quinolones and chloramphenicol have been tested in a randomised trial.

The fact that CFLX is superior to chloramphenicol and norfloxacin may not seem surprising when reviewing drug pharmacokinetics. At least 50-70% of CFLX is

excreted in the urine as unmetabolised drug, which is twice as much as for norfloxacin and chloramphenicol. Chloramphenicol is not active against a number of Gram-negative strains, some E. coli strains are even spontaneously resistant to the drug, and a relatively high concentration of chloramphenicol is needed to achieve a minimum inhibitory concentration in E. coli bacteria [22]. Interestingly, only one study reported a difference in rates of infectious complications between quinolone-treated groups [22]. Also, no trials were able to demonstrate that administration of several doses of antibiotics is superior to a single dose and thus prolonged use of antibiotics following TRUS-gb remains controversial [6, 8, 10, 11, 16, 22]. This is in line with the findings by Zani et al from 2011 [25]. It remains unknown whether a single dose compared with several doses is associated with an increase in bacterial resistance mechanisms, which could potentially influence the risk of infectious complications if patients are scheduled for re-biopsy procedures. Consequently, current guidelines on antibiotic prophylaxis in association with TRUS-gb must primarily be based on empirical evidence and knowledge about rectal bacterial culture as well as the typical response to known antibiotics.

E. coli, Klebsiella spp. and Enterobacter spp. – all Gram-negative, facultatively anaerob bacteria – seem to be the most commonly isolated bacteria in symptomatic patients, and an antibiotic strategy should target these species. However, since anaerob culturing of urine or blood was not performed systematically, the prevalence of strict anaerob bacteria may be underreported. The study performing rectal swabs prior to the prostate biopsy demonstrated that almost one fourth of the patients were carriers of CFLX-resistant bacteria [21]. Thus, for patients with increased risk of post-bioptic infections (i.e. diabetes, chronic prostatitis and prior use of antibiotics), performing a rectal swab that allows for diagnosis of resistant bacterial strains may prove valuable in creating a targeted antibiotic strategy.

Recent studies suggest that an increase in infectious complications after prostate biopsies has occurred over time [26]. In Canada, an increased risk of hospitalisation 30 days after TRUS-gb was demonstrated between 1996 and 2005 (odds ratio = 3.74; 95% confidence interval (CI): 2.0-7.0; p < 0.0001) [27]. Similarly, an increase in the frequency of hospitalisation due to infectious complications after TRUS-gb of the prostate over a study

period from 1991 to 2007 (p for trend 0.001) has been reported [28]. Repeat biopsies during active surveillance have been argued to increase the risk of post-biopsy infectious complications. One trial found the odds of infectious complications after TRUS-gb to increase by 1.33 with each additional previous biopsy (95% CI: 1.01-1.74; p = 0.041) [29]. On the other hand, previous biopsies have also been found not to be related to a higher risk of post-bioptic complications [21].

A range of different measures have been applied to the TRUS-gb procedure in an attempt to minimise the number and severity of post-bioptic complications, i.e. cleansing of the biopsy needle between each biopsy [30], pre-bioptic assessment of urine bacterial culture [31], the transperineal biopsy technique [32] and the use of enemas with or without an added antibiotic. The evidence for an improved outcome for these regimens, however, remains low.

The Danish perspective

According to Danish guidelines, population-based PSA screening is not recommended and the PSA test is reserved for symptomatic patients, patients who have findings on clinical examination that raise suspicion of PCa, or patients with at least two close relatives who have suffered from early-onset Pca [33]. Current Danish guidelines advocate a set of ten biopsies preceded by the administration of prophylactic, but otherwise unspecified, antibiotics [34].

During the past 20 years, DK has witnessed a dramatic increase in the incidence of prostate cancer with an expected, corresponding increase in the number of TRUS-gb performed each year and, accordingly, an increased use of prophylactic antibiotics. According to DANMAP, a monitoring programme of the antimicrobial resistance in DK, the overall use of antimicrobial agents in Danish patients has increased by 20% since 2004, but the use of fluoroquinolones has increased by more than 40%. At the same time, CFLX-resistance in E. coli isolates from urine has increased from 3% to 12% among hospitalised patients [35]. Whether this is associated with the use of fluoroquinolones when performing TRUS-gb is unknown. With CFLX monotherapy being the most commonly used prophylaxis for TRUS-gb in DK, the increase in CFLX-resistant E. coli raises concern whether a concomitant increase in post-biopsy infectious complications may follow.

Local clinical guidelines on the use of antibiotics with TRUS-gb are not uniform in DK and there seems to be considerable variation among the drugs of choice even within hospitals of close proximity. CFLX remains the most widely used prophylactic antibiotic for TRUS-gb in DK, but also metronidazole, amoxicillin/clavulanic acid and mecillinam are commonly used. Whether the risk of complications varies between institutions is unknown.

It has been shown that different types of antibiotics penetrate differently into the prostate tissue, and the mode of administration is also important for the concentration of the antibiotic that is reached in the prostatic tissue [36-38]. International guidelines, such as the European Association of Urology Guidelines, recommend oral or intravenous administration of a quinolone prior to biopsy [39]. The American Urological Association recommends oral fluoroquinolones or intravenous or

intramuscular administration of cephalosporins or trimethoprim-sulfamethoxazole as first-choice prophylactic antibiotic [40]. If, however, a rectal swab is performed, the antimicrobial prophylaxis can be adjusted towards the growth result [40], which, according to

previous studies appears highly efficient in reducing

infectious complications [41, 42]. Minimising the antimicrobial spectrum, the dose and the duration of the antibiotic are important when taking into account the ecological consequences of antimicrobial use (e.g. the development of resistant bacterial strains). Furthermore, with regard to optimal use of antibiotic, the peak serum concentration of CFLX is reached 1-3 hours after oral intake, which may therefore possibly be considered the ideal time for biopsy [6].

CONCLUSION

Evidence supporting a specific antibiotic prophylaxis strategy for TRUS-gb is scarce. Furthermore, the frequent use of antibiotics in the population in general, the surfacing of many resistant bacteria and the fact that many men undergo re-biopsies means that is seems difficult to introduce a uniform, prophylactic antibiotic regimen for TRUS-gb. Patients who develop post-biopsy sepsis often harbour bacteria with complex resistance patterns; thus, the available microbiological expertise should be applied when initiating therapy.

Correspondence: Nina Klemann. E-mail: ninakhp@gmail.com

Accepted: 15 November 2016

Conflicts of interest: Disclosure forms provided by the authors are available with the full text of this article at www.danmedj.dk