Adding amikacin to fluoroquinolone-based antimicrobial prophylaxis reduces prostate biopsy infection rates

Authors

  • Deepak Batura,

    1. Departments of Urology and *Microbiology, Northwick Park Hospital, London, and Statistics, Modelling and Bioinformatics Department, Centre of Infections, Health Protection Agency, UK
    Search for more papers by this author
  • G. Gopal Rao,

    1. Departments of Urology and *Microbiology, Northwick Park Hospital, London, and Statistics, Modelling and Bioinformatics Department, Centre of Infections, Health Protection Agency, UK
    Search for more papers by this author
  • Peder Bo Nielsen,

    1. Departments of Urology and *Microbiology, Northwick Park Hospital, London, and Statistics, Modelling and Bioinformatics Department, Centre of Infections, Health Protection Agency, UK
    Search for more papers by this author
  • André Charlett

    1. Departments of Urology and *Microbiology, Northwick Park Hospital, London, and Statistics, Modelling and Bioinformatics Department, Centre of Infections, Health Protection Agency, UK
    Search for more papers by this author

Mr Deepak Batura, Department of Urology, Northwick Park Hospital, Watford Road, Harrow, Middlesex HA1 3UJ, UK. e-mail: Deepakbatura@gmail.com

Abstract

Study Type – Therapy (case series)

Level of Evidence 4

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

Fluoroquinolones are almost universally used as antimicrobial prophylaxis for prevention of infections after transrectal ultrasound guided prostate biopsy. However, many recent studies have identified an alarming trend of increasing fluoroquinolone resistance leading to a rise in infections after biopsy. This indicates the unsuitability of ciprofloxacin as appropriate antibiotic prophylaxis.

In this study, we have demonstrated a significant reduction in post biopsy infections by adding amikacin to fluoroquinolone based prophylaxis.

OBJECTIVE

• To examine the efficacy of adding amikacin to fluoroquinolone-based antimicrobial prophylaxis in preventing transrectal ultrasonography-guided prostate biopsy (TGB) associated infections.

PATIENTS AND METHODS

• Infections after TGB were compared before adding amikacin to antimicrobial prophylaxis (2006) with those that occurred after adding amikacin to the prophylaxis (2007 and 2008).

• During both periods antimicrobial prophylaxis consisted of ciprofloxacin, co-amoxiclav and metronidazole except after August 2008 when co-amoxiclav was discontinued.

• Amikacin was added to the prophylaxis protocol in the period 2007 and 2008.

RESULTS

• Before adding amikacin 11 of 281 (3.9%) patients developed infections after TGB (seven urinary tract infections (UTIs) and seven bacteraemias) whereas after adding amikacin six UTIs and two bacteraemias occurred in 590 (1.4%) patients.

• In both periods, most of the strains causing the infections were ciprofloxacin resistant (2006: 13 of 14; 2007 and 2008: seven of eight).

• Overall, there is strong statistical evidence that the total infection rate was significantly reduced after the inclusion of amikacin into the prostate biopsy prophylaxis regimen.

• In 2007 and 2008 when amikacin was included in prophylaxis, the bacteraemia rate was reduced to just over one-tenth of the rate in 2006 before introducing amikacin (P= 0.002).

• Although just failing to reach the conventional significance level of 0.05, the evidence suggests a reduction in UTI by an estimated 60% after adding amikacin.

CONCLUSION

• We conclude that adding amikacin to fluoroquinolone-based antimicrobial prophylaxis in areas with high fluoroquinolone resistance confers significant benefit in preventing infections after TGB.

Abbreviations
Est. RR

estimate of the infection rate ratio

TGB

TRUS-guided prostate biopsy

INTRODUCTION

TRUS-guided prostate biopsies (TGB) are a common and increasingly performed procedure in urological practice. However, these procedures can be complicated by infections, e.g. UTI and even life-threatening bacteraemia. Hence, antimicrobial prophylaxis is routinely used to prevent these complications [1]. Ciprofloxacin is widely prescribed as prophylaxis because of its good penetration into the prostatic cytosol, ease of use and safety profile [2]. Unfortunately, many recent studies have identified an alarming trend of increasing resistance to fluoroquinolones worldwide, limiting their suitability as appropriate prophylaxis [3–6]. In our practice in North-west London, we have also seen an increase in the number of urinary and blood stream infections with extended spectrum β-lactamase producing coliforms (ESBLs), 90% of which were resistant to ciprofloxacin and 50% resistant to co-amoxiclav. However, there was negligible resistance to amikacin [7].

We noted a high incidence of infections after TGB in 2006 when using ciprofloxacin-based antimicrobial prophylaxis. Most of these infections were due to ciprofloxacin-resistant organisms, which were sensitive to amikacin.

Our clinical experience together with the worldwide trend of increasing ciprofloxacin resistance in Gram-negative bacilli prompted us to review our antimicrobial prophylaxis protocol [7,8].

The aim of the present study was to examine if the addition of amikacin to ciprofloxacin-based antimicrobial prophylaxis reduces infections after TGB.

PATIENTS AND METHODS

This observational, before and after study, using an interrupted time series design was conducted on patients attending the urology clinic for TGB during the period 2006–2008 at the Northwick Park Hospital, a teaching hospital in North-west London. Infections after TGB in 2006 were compared with infections in 2007 and 2008.

During 2006, antimicrobial prophylaxis consisted of ciprofloxacin (500 mg, twice daily) orally starting 2 days before the biopsy and continuing for 5 days after the procedure, parenteral co-amoxiclav 1.2 g immediately before the procedure followed by oral co-amoxiclav (625 mg, three times daily) for 5 days after TGB. At the end of the procedure, a metronidazole suppository (500 mg) was inserted. During the period January 2007 to August 2008, in addition to ciprofloxacin, co-amoxiclav (625 mg, three times daily) was given orally starting 2 days before and continuing for 5 days after TGB. Use of i.v. co-amoxiclav was discontinued. Co-amoxiclav was discontinued altogether after August 2008. Due to the high number of infections in 2006, amikacin was added to the prophylaxis protocol in the period 2007 and 2008. This was given i.v. (1 g) immediately before TGB.

Our laboratory information system was used to identify patients who presented to us with evidence of bacteraemia or significant bacteriuria (>105/mL) within 30 days of TGB and the antimicrobial-resistance pattern of the strains were noted.

Infection rates after TGB between the two periods were compared using the chi-square test of association. Estimates of the infection rate ratio (Est. RR) and their 95% CI were obtained from the Poisson estimating equation model. The outcome variable was the numbers of infections after TGB each month, the intervention was included as an independent variable in the model, and the logarithm of the number of TGBs each month was included as an offset. The correlation structure was assumed to be first-order autoregressive. This model was fitted to each of the four outcomes: UTIs, bacteraemias, infected patients and all infections after TGB.

Alternative regression and autoregressive integrated moving average (ARIMA) models were also used to obtain estimates of the reduction in infections after TGB after the implementation of amikacin prophylaxis. These were performed to ensure that results similar to those presented using the generalized estimating equation model were obtained.

RESULTS

In all, 871 patients underwent TGBs between January 2006 and December 2008. The age distribution of patients in the periods before and after starting amikacin prophylaxis was similar: the mean (sd; range) age before amikacin was 70.3 (8.8; 41–91) years and with amikacin was 68.7 (7.7; 47–88) years. Similarly, the age distribution of patients developing infections in the two periods was comparable: the mean (sd; range) age before amikacin was 68.2 (11.7; 46–84) years and with amikacin was 68 (9.9; 54–82) years. The mean (range) TGB to infection interval before amikacin was 13.1 (2–28) days for UTI and 4 (2–14) days for bacteraemia. With amikacin, the mean (range) interval was 15.1 (4–26) days and 3.4 days for UTI and bacteraemia, respectively. The differences between these intervals are not statistically significant. In both periods, most of the strains causing the infections were ciprofloxacin resistant. In the before amikacin period 13/14 (93%, 95% CI 66–99.8) strains were ciprofloxacin resistant and in the with amikacin period seven of eight (88%, 95% CI 47–99.7%) were resistant. The details of our patents with infections after TGB are given in Table 1. In all, 19 patients had 22 episodes of infection after TGB. The occurrence of infections during the various prophylaxis regimes are shown in Table 2. Figure 1 is a time series of the proportion of infections after TGB in each month of the study and shows that the infection rate declined over time.

Table 1.  Characteristics of patients with infections after TGBs
Sample noSourceBiopsy to infection interval, daysOrganism isolatedAntimicrobial sensitivity
AmikacinCiprofloxacinCo-amoxiclav
  1. NT, not tested; S, sensitive; R, resisitant.

2006      
  1Urine2.5ColiformsNTRS
  2Urine20.2ColiformsNTRS
  3Urine1.9ColiformsNTRS
  4Urine21.4ColiformsNTRS
  5Urine15.0ColiformNTRS
  6Urine27.8ColiformsNTRS
  7Urine2.9ColiformsNTRS
  8Blood2.5E. coliSRS
  9Blood2.5E. coliSRS
 10Blood2.6E. coliSRS
 11Blood14.4E. coliSRS
 12Blood1.7E. coliSSR
 13Blood2.6E. coliSRS
 14Blood2.0E. coliSRS
2007 and 2008      
 1Urine10.6ColiformsNTRS
 2Urine23.7ColiformsNTRS
 3Urine3.7ColiformsNTRS
 4Urine19.4ColiformsNTRS
 5Urine25.6PseudomonasNTSNT
 6Urine7.5ColiformsNTRS
 7Blood3.4E. coliSRR
 8Blood3.4E. coliSRR
Table 2.  Description of antimicrobial prophylaxis and crude infection rates
Period and prophylaxis usedNo. of patientsBacteraemias, n (%)UTIs, n (%)Total infections, n (%)Infected patients, n (%)
  1. Cip, ciprofloxacin; Aug, co-amoxiclav; Mz, metranidazole; Amik, amikacin.

January–December 2006 (Cip + Aug + Mz)2817 (2.5)7 (2.5)14 (5.0)11 (3.9)
January 2007–August 2008 ((Cip + Aug + Mz + Amik)4842 (0.4)6 (1.2)8 (1.7)8 (1.7)
September 2008–December 2008 (Cip + Mz + Amik)1060000
Total8719 (1.0)13 (1.5)22 (2.5)19 (2.2)
Figure 1.

Monthly incidence rate of infections per 1000 TGBs.

The period during which co-amoxiclav was discontinued was relatively short (September to December 2008), and hence affects the statistical power to detect any clinically important reduction in infection rate. Therefore, we only compared the two periods before and after inclusion of amikacin. There is strong statistical evidence that the inclusion of amikacin into the TGB prophylaxis regimen reduces the blood stream infection rate after TGB. The chi-square test statistics (1 d.f.) for blood stream infections was 8.62 (P= 0.003), showing statistically significant reduction of bloodstream infections from 2.5% before inclusion of amikacin to 0.3% when amikacin was included. The estimated bloodstream infection rate was reduced to just over one-tenth in 2007/2008 as compared with 2006 after introducing amikacin (Est. RR 0.13, 95% CI 0.04–0.48, P= 0.002).

For UTI, the chi-square result was 2.81 (P= 0.09), indicating that the reduction in UTI from 2.5% before to 1.0% after inclusion of amikacin did not reach conventional levels of statistical significance. Although just failing to reach conventional significance level of 0.05, the evidence suggests a reduction in UTIs by an estimated 60% after adding amikacin (Est. RR 0.40, 95% CI 0.16–1.03, P= 0.06).

The chi-square test statistic for the total number of infected patients is 5.83 (P= 0.02). As both blood and urinary infections were reduced, the total number of infected patients were also significantly reduced once amikacin was included in the prophylaxis regimen (Est. RR 0.34, 95% CI 0.14–0.83, P= 0.02).

Results were also obtained for all infections combined. The fact that bloodstream infections are more likely to occur in those patients who also have urine infection complicates the statistical analysis. If these dependencies are ignored the estimated reduction in infection rate ratio is 0.27 (95% CI 0.13–0.57, P= 0.001). Similar results were obtained when considering these dependencies in a random effects Poisson regression model, with an Est. RR of 0.27 (95% CI 0.10–0.71, P= 0.008). The results of the generalized estimating equations assuming first-order temporal autocorrelation are presented in Table 3.

Table 3.  Results of Poisson generalised estimating equation
Prophylaxis regimenEst. RR (95% CI)P
Blood stream infection:  
 Before amikacinReference 
 Amikacin added0.13 (0.04–0.48)0.002
Urine infection:  
 Before amikacinReference 
 Amikacin added0.40 (0.16–1.03)0.06
Infected patients:  
 Before amikacinReference 
 Amikacin added0.34 (0.14–0.83)0.02
Total infections:  
 Before amikacinReference 
 Amikacin added0.27 (0.13–0.57)0.001

DISCUSSION

Current evidence indicates that infection rates after TGBs are reduced considerably by antimicrobial prophylaxis [1]. However, there is no consensus on the choice of the most appropriate antimicrobial and aminoglycosides, doxycycline and fluoroquinolones have been used with no clear evidence in support of any particular regime [9]. Based on its pharmacological profile, its high bioavailability in the prostate, its relative safety and ease to use, ciprofloxacin has emerged as the most common antimicrobial used for TGB prophylaxis [2]. Unfortunately, antimicrobial resistance to ciprofloxacin in systemic E. coli infections in the UK has been found to be increasing annually (1% in 1994 to 23% in 2006) [8]. Indeed, prevalence studies have also highlighted an emerging trend of increasing ciprofloxacin resistance in urology practice with implications for prophylaxis and treatment [3]. This trend is also reflected in data from our local population where a high level of ciprofloxacin resistance was detected in patients developing bacteraemia secondary to UTI [7]. Hence, it is not surprising that this study has shown a similarly high level of ciprofloxacin resistance in patients developing infections after TGB. This implies the need for urological services to be aware of the local prevalence of antimicrobial resistance in intestinal flora of patients undergoing TGB [10].

The favourable characteristics of ciprofloxacin in the treatment of prostate and UTIs are well recognised but the high level of resistance necessitates a re-evaluation of the use of fluoroquinolones as prophylaxis in the prevention of TGB infections. Despite the development of infections after TGBs with ciprofloxacin-resistant organisms, we are hesitant to suggest that fluoroquinolones be avoided altogether as it is difficult to estimate what infections ciprofloxacin may have prevented. A recent study reported a reduction in infections after TGB by the addition of gentamicin to ciprofloxacin compared with ciprofloxacin alone for antimicrobial prophylaxis [11]. Furthermore, another study has shown that an aminoglycoside (gentamicin) was inferior to ciprofloxacin in the prevention of infections after TGB [12].

The AUA and European Association of Urology guidelines for antibacterial prophylaxis for TGBs recommend fluoroquinolones as agents of first choice [13,14]. The AUA guidelines also recommend aminoglycosides or aztreonam with metranidazole or clindamicin as alternatives to ciprofloxacin [13]. However, they do not recommend a specific aminoglycoside. In the UK, the Prostate Cancer Risk Management Programme of 2006 recommends that antibiotic prophylaxis must be provided to all patients undergoing TGBs and a minimum of one antibiotic should be used such as an oral fluoroquinolone and the use of a second antibiotic against anaerobic bacteria should be considered [15]. The National Institute of Clinical Excellence prostate cancer guidelines of February 2008 endorse these recommendations [16].

We chose amikacin as an additional aminoglycoside agent because very little resistance to amikacin was detected in urine and bloodstream isolates from specimens received in our laboratory. In general, resistance to amikacin is less common than to gentamicin because only one enzyme (aminoglycoside 6′-n-acetyl transferase) produced by bacteria has been identified that can modify amikacin and render the bacterium resistant as compared with at least six enzymes that can modify gentamicin [17]. Moreover, a high concentration of amikacin is found in the prostatic tissue after one dose [18].

There was a significant reduction in number of bacteraemia after TGB but the reduction in UTIs although considerable (60%) just failed to reach a statistically significant level. It is conceivable that some of the UTIs may have been acquired after cessation of the prophylaxis because the average time of onset of UTI was 15 days after TGB and thus we could not confidently differentiate infections acquired because of TGB from those acquired after the procedure. While the addition of amikacin significantly reduced infections after TGB, not all infections with ciprofloxacin-resistant organisms were prevented. The reason for this is unclear. Similarly, we did not investigate the occurrence of allergic reactions or other adverse side-effects to amikacin and are therefore unable to comment on the occurrence of side-effects.

We conclude that adding amikacin to fluoroquinolone-based antimicrobial prophylaxis significantly reduces infections after TGB, particularly in populations where ciprofloxacin resistance is common. However, further studies would be helpful in confirming our findings.

ACKNOWLEDGEMENTS

The authors acknowledge all urological surgeons of North-west London Hospitals NHS Trust, Dr A Chambers, Consultant Radiologist, Dr R Wall and Dr S Shafi, Consultant Microbiologists for devising and implementing the antimicrobial prophylaxis protocols. We also thank the microbiology laboratory staff for their support.

CONFLICT OF INTEREST

None declared.

Ancillary