Urologic Oncology: Seminars and Original Investigations ] (2015) ∎∎∎–∎∎∎
Original article
Antibiotics may not decrease prostate-specific antigen levels or prevent unnecessary prostate biopsy in patients with moderately increased prostate-specific antigen levels: A meta-analysis Lu Yang, M.D., Yuchun Zhu, M.D., Zhuang Tang, M.D., Yongji Chen, M.D., Liang Gao, M.D., Liangren Liu, M.D., Ping Han, M.D., Xiang Li, M.D.*, Qiang Wei, M.D.* Department of Urology, West China Hospital, Sichuan University, Chengdu, China Received 2 December 2014; received in revised form 2 February 2015; accepted 2 February 2015
Abstract Objectives: To evaluate the effect of empiric antibiotics on decreasing prostate-specific antigen (PSA) levels and the possibility of avoiding unnecessary prostate biopsies (PBs). Materials and methods: A systematic search of PubMed, Embase, and the Cochrane Library was performed to identify all randomized controlled trials (RCTs) that compared effects of empiric antibiotics with no treatment or placebo on lowering PSA levels and minimizing unnecessary PBs in patients with moderately increased PSA levels. The Cochrane Collaboration Review Manager software (RevMan 5.1.4) was used for statistical analysis. Results: The inclusion criteria for the study were met by 6 RCTs (1 placebo controlled and 5 no treatment controlled) involving 656 patients. The synthesized data from these RCTs indicated that there were no significant differences between the antibiotic and control groups in the PSA levels after treatment (mean difference [MD] ¼ 0.15, 95% CI: 0.50 to 0.81, P ¼ 0.65], number of patients with decreased PSA levels after treatment (relative risk [RR] ¼ 1.22, 95% CI: 0.90–1.65, P ¼ 0.20], prostate-specific antigen density levels after treatment (MD ¼ 0.04, 95% CI: 0.15 to 0.07, P ¼ 0.47), f/t% PSA after treatment (MD ¼ 1.47, 95% CI: 4.65 to 1.71, P ¼ 0.37), number of patients with responsive PSA (RR ¼ 1.02, 95% CI: 0.58–1.81, P ¼ 0.94), and individual Pca-positive rate in these patients (RR ¼ 1.07, 95% CI: 0.53–2.16, P ¼ 0.86), and Pca-positive rates (RR ¼ 0.85, 95% CI: 0.48–1.50, P ¼ 0.57). However, the antibiotic group had a significant change in the net PSA decrease after treatment compared with the control group (MD ¼ 1.44, 95% CI: 0.70–2.17, P ¼ 0.0001). Conclusion: The use of empiric antibiotics may not significantly decrease PSA levels or avoid unnecessary PBs. r 2015 Elsevier Inc. All rights reserved.
Keywords: Antibiotic; Prostate biopsy; Prostate cancer; Prostate-specific antigen; Randomized controlled trial
1. Introduction
Lu Yang and Yuchun Zhu contributed equally to this work and should share the co–first author. This study was supported by the Prostate Cancer Foundation Young Investigator Award 2013, the National Natural Science Foundation of China (Grant No. 81300627, 81370855, and 30901484), and Program from Science and Technology Department of Sichuan Province (Grant No. 2013SZ0006 and 2014JY0219). * Corresponding authors. Tel.: þ86-189-8060-1425; fax: þ86-28-85422451. E-mail addresses: [email protected] (X. Li), weiqiang933@126. com (Q. Wei). http://dx.doi.org/10.1016/j.urolonc.2015.02.001 1078-1439/r 2015 Elsevier Inc. All rights reserved.
Prostate cancer (Pca), the most common cancer in men in Western countries, has been predicted and diagnosed at an early stage since prostate-specific antigen (PSA) testing began in the 1980s [1]. Worldwide, millions of prostate biopsies (PBs) are performed annually [2,3], most commonly because of elevated serum PSA levels. However, PSA levels can increase for several reasons, including trauma, ejaculation, and rectal and urethral procedures. In addition, numerous noncancer etiologies can cause elevated PSA levels, such as benign prostatic hyperplasia, inflammation, and infection [4].
2
L. Yang et al. / Urologic Oncology: Seminars and Original Investigations ] (2015) 1–8
A PSA range of 4 to 10 ng/ml is commonly recognized as the diagnostic “gray zone” [5]. Some estimates suggest that only 20% to 30% of men with PSA levels between 4 and 10 ng/ml actually have PCa [5]. There is a trend toward reducing the threshold PSA level to 2.5 ng/ml to detect more cases of Pca. Consistent with this, Gilbert et al. [6] reported a cancer detection rate of 27.48% in patients with a PSA level of 2.5 to 4 ng/ml by reviewing 26,316 biopsies. Therefore, how to best identify patients with Pca with PSA levels of 2.5 to 10 ng/ml and avoid unnecessary PB remains challenging. Antimicrobial treatment in patients with increased PSA levels and confirmed or suspected nonmalignancy may decrease the number of unnecessary PBs [5]. Therefore, a cost-effective and painless approach for decreasing the number of biopsies with negative findings may be available [7]. In clinical practice, many urologists use empiric antibiotic treatments, followed by a repeat PSA test in patients with moderately increased PSA. However, scientific evidence to support this approach is currently lacking [8,9]. Nevertheless, critics argue that not only is there is a lack of appropriate studies documenting that antibiotics can decrease PSA levels, but they are also no data to suggest that it is safe to defer biopsies in patients who exhibit a decrease in PSA levels after antibiotic treatment [10,11]. 2. Materials and methods 2.1. Search strategy We searched PubMed (1980 to May 2014), Embase (1980 to May 2014), the Cochrane Central Register of Controlled Trials, and the Cochrane Database of Systematic Reviews. The initial search process was designed to identify all relevant trials involving the terms “antibiotic,” “prostatespecific antigen”, “prostate biopsy”, and “prostate cancer” (and multiple synonyms for each term). Furthermore, the reference lists of relevant publications and conference proceedings from the American Urological Association, European Association of Urology, and Société Internationale d'Urologie between 2009 and 2014 were searched manually. Computer searches were supplemented with manual searches. Two authors (L.Y. and L.G.) screened all citations and abstracts isolated using this search strategy independently to identify potentially eligible studies.
symptoms, signs, or history of prostatitis; a prior prostatic surgery, biopsy, radiotherapy, or hormonal therapy; any abnormal findings in DRE; an acute infection of the urinary system, pyuria, and bacterial growth in asymptomatic patients; a recent history of instrumentation in the urinary system; previous use of a 5-α reductase inhibitor; a residual urine volume 4100 ml; and hypersensitivity to any medical ingredient in the antibiotics used. 2.3. Interventions and comparisons All patients with moderately increased PSA levels (generally between 2.5 [or 4] and 10 ng/ml) and normal findings on DRE were divided into an antibiotic group or untreated/placebo (control) group. Comparisons were made on the outcomes of antibiotic therapy or not in the decrease of PSA levels and the avoidance of unnecessary PBs. 2.4. Outcome measures The outcomes measured in this review included the following: the PSA level after treatment, net PSA decrease after treatment, number of patients with a decreased PSA after treatment, PSA density (PSAD, the ratio of serum total level to the prostate volume) level after treatment, f/t% PSA after treatment, number of patients with responsive PSA and individual rate of Pca-positive in these patients, and Pcapositive rates in antibiotic and control groups. The definition of responsive PSA was a decreased PSA level o4 or 2.5 ng/ml according to the individual RCTs or a reduction in PSA level of 450% after treatment. The net PSA decrease after treatment was defined as the initial PSArepeat PSA in both groups. 2.5. Data extraction Data were independently extracted by 2 authors (Y.J.C. and Z.T.) using a predesigned data extraction form. The extracted data included the data source, eligibility, methods, participant characteristics, interventions, and results. Both authors convened to synthesize their findings. The information was subsequently entered into RevMan 5.1.4. Any discrepancy among the extracted data was resolved by discussion, and any disagreements that persisted after discussion were resolved in consultation with another author (W.Q.).
2.2. Participants
2.6. Quality assessment
The inclusion criteria were as follows: male patients with moderately increased PSA levels (generally between 2.5 [or 4] and 10 ng/ml), who had a normal findings on digital rectal examination (DRE), who received antibiotics or no treatment/placebo for several weeks, and who then underwent a subsequent PSA test and PB. Patients with specific conditions were excluded, including those with any
The quality of included studies was assessed by 2 authors (L.L.R. and H.P.) according to the Cochrane Collaboration Reviewers' Handbook and the QUOROM guidelines [12,13]. The quality items assessed included the generation of randomization sequences, blinding method used, allocation concealment, description of patient withdrawals and dropouts, and intent-to-treat analysis.
L. Yang et al. / Urologic Oncology: Seminars and Original Investigations ] (2015) 1–8
2.7. Statistical analysis All meta-analyses were performed using RevMan 5.1.4. Dichotomous data and continuous data were presented as relative risk (RR) and mean difference (MD), respectively, with 95% CIs. Meta-analysis was performed using the fixed-effects method or the random-effects method. The fixed-effects method was used to combine the results when statistically significant heterogeneity was absent, whereas the random-effects method was applied when heterogeneity was present. Statistical heterogeneity among trials was evaluated using the I2 test with significance set at P o 0.05. Publication bias was evaluated using a funnel plot. Sensitivity analysis was performed if low-quality trials were included.
3
ciprofloxacin [14,15], 2 used ofloxacin [16,17], and 2 used levofloxacin [18,19]. The treatment duration was 2 to 4 weeks. Our review of funnel plots excluded publication bias. 3.2. Variations in levels of PSA and PSA derivatives after treatment 3.2.1. PSA levels after treatment PSA levels after treatment were reported in 3 studies with 314 patients [14,15,18]. In total, 173 patients were randomized to receive antibiotics, and 141 patients were randomized into the control group (Fig. 2A). There was no significant difference between the antibiotic and control groups (MD ¼ 0.15, 95% CI: 0.50–0.81, P ¼ 0.65).
3. Results 3.1. Description of studies In total, 1,175 reports were initially identified from the database and manual searches. After removing 344 records of redundant publications, reviews, and meta-analysis and scanning the titles and abstracts of 780 unrelated records, 1,124 reports were excluded from the study. After referring to the full texts and excluding 14 abstracts, 22 nonrandomized studies, and 9 studies irrelevant to the intervention of interest or outcome of interest, we identified and included 6 publications (randomized controlled trials [RCTs], including 1 study published after the abstract at meeting) that enrolled 656 patients in the review (Fig. 1) [14–19]. The characteristics and quality of the included studies are presented in Tables 1 and 2, and the search flow diagram is presented in Fig. 1. There were no significant differences in baseline information between the antibiotic and control groups. There were 5 studies containing 512 patients that were controlled using an untreated group without any blinding method applied [14–16,18,19], and 1 study involving 144 patients that was placebo controlled with doubleblinded method (Table 2) [17]. Additionally, 2 studies used
Fig. 1. Flow diagram of systematic review.
3.2.2. Net PSA decrease after treatment Data on the net PSA decrease were collected from 2 trials including 197 patients [15,16], of whom 115 and 82 were randomized to receive antibiotics and the control group, respectively (Fig. 2B). The net PSA decrease was significantly higher in the antibiotic group than in the control group (MD ¼ 1.44, 95% CI: 0.70–2.17, P ¼ 0.0001). 3.2.3. Number of patients with decreased PSA levels after treatment Two studies including 195 patients investigated the number of patients with decreased PSA levels after treatment [14,17]. There were 47 cases with decreased PSA levels among 87 patients randomized to receive antibiotics and 39 cases among 88 patients randomized to the control group (Fig. 2C). There was no significant difference between the antibiotic and control groups (RR ¼ 1.22, 95% CI: 0.90–1.65, P ¼ 0.20). 3.2.4. PSAD levels after treatment PSAD levels after treatment were reported in 2 studies including 237 patients [15,18]. There were 135 and 102 patients randomized to receive antibiotics and into the control group, respectively (Fig. 2D). There was no significant difference between the antibiotic and control groups (random-effects model; MD ¼ 0.04, 95% CI: 0.15 to 0.07, P ¼ 0.47). 3.2.5. f/t% PSA ratio after treatment The f/t% PSA ratio after treatment was reported in 2 studies that analyzed 237 patients [15,18]. There were 135 and 102 patients randomized to receive antibiotics and into the control group, respectively (Fig. 2E). There was no significant difference between the antibiotic and control groups (MD ¼ 1.47, 95% CI: 4.65 to 1.71, P ¼ 0.37).
L. Yang et al. / Urologic Oncology: Seminars and Original Investigations ] (2015) 1–8
4
Table 1 Characteristics of treatment groups in 6 randomized controlled trials Reference
Drug (6)/placebo (1) or no drug (5)
Inclusive criteria PSA range, ng/ml
DRE results
Eggener et al. 39/38 [14] Erol et al. [15] 65/32
42.5
Negative
44
ND
Saribacak et al. [16]
4–10
Negative
2.5–10
Negative
2.5–10
Negative
2.5–10
Negative
50/50
Stopiglia et al. 49/49 [17] Toktas et al. 70/70 [18] Ugurlu et al. [19]
72/72
Antibiotic types and durations
Inflammation diagnosed before inclusion
Repeat PSA and PSA derivatives
No
2 Weeks of ciprofloxacin, PSA 500 mg twice daily No Ciprofloxacin 500 mg PSA, fPSA, twice daily, and PSAD, and diclofenac 75 mg f/t% PSA for 2-3 weeks No 400 mg of ofloxacin PSA, fPSA, daily for 4 weeks PSAD, and f/t% PSA No 500-mg ciprofloxacin PSA for 28 days No 500-mg oral levofloxacin PSA, fPSA, once a day for 21 days PSAD, f/t% PSA According to EPS Levofloxacin 500 mg PSA, fPSA, results, divided into once a PSAD, and EPSþ and EPS day for 3 weeks f/t% PSA
Repeat prostate biopsy All patients All patients
All patients
All patients All patients
All patients
EPS ¼ expressed prostate secretion, ND ¼ not determined.
3.3. Responsive PSA and Pca-positive cases in patients with responsive PSA Subgroup analyses were performed that included patients with PSA levels that decreased to o4 ng/ml or o2.5 ng/ml and showed 450% reduction in PSA level.
3.3.1. Responsive PSA There were 5 studies that included 559 patients and investigated the number of patients with responsive PSA after intervention or control [14,16–19]. Patients whose PSA levels decreased to o4 ng/ml were reported in 2 studies with 177 patients [14,16]. Of 88 cases randomized to receive antibiotics, 16 exhibited decreased PSA levels o4 ng/ml in comparison with 16 of 89 patients randomized to the control group (Fig. 3). There was no significant difference between the antibiotic and control
groups (random-effects model; RR ¼ 0.93, 95% CI: 0.20– 4.40, P ¼ 0.93). Patients whose PSA level decreased to o2.5 ng/ml were reported in 4 studies of 387 patients [14,17–19]. In total, 28 of 193 cases randomized to receive antibiotics exhibited decreased PSA levels of o2.5 ng/ml in comparison with 23 of 194 patients randomized to the control group (Fig. 3). There was no significant difference between groups (random-effects model; RR ¼ 1.21, 95% CI: 0.68–2.14, P ¼ 0.52). There was 1 study that reported 1 case with a reduced PSA level 450% in 38 patients randomized to receive antibiotics in comparison with 4 of 39 patients randomized to the control group [14]. There was no significant difference between the antibiotic and control groups (P 4 0.05). Overall, there were no significant differences in cases of responsive PSA between the antibiotic and control groups (random-effects model; RR ¼ 1.02, 95% CI: 0.58–1.81, P ¼ 0.94; Fig. 3).
Table 2 Quality of included randomized controlled trials Reference
Drug (6)/placebo Generation of (1) or no drug (5) Randomization sequences
Allocation concealment
Blinding
Incomplete outcome data addressed
Free of selective reporting
Free of other bias
Country
Eggener et al. [14]
39/38
Low risk
Unclear risk
ND
Low risk
Low risk
Low risk
Erol et al. [15] Saribacak et al. [16] Stopiglia et al. [17] Toktas et al. [18] Ugurlu et al. [19]
65/32 50/50 49/49 70/70 72/72
Low risk Low risk Low risk Unclear risk Unclear risk
Low risk Unclear risk Unclear risk Unclear risk Low risk
ND ND Low risk ND ND
Low Low Low Low Low
Low Low Low Low Low
Low Low Low Low Low
US, Israel, and Canada Turkey Turkey Brazil Turkey Turkey
ND ¼ not determined.
risk risk risk risk risk
risk risk risk risk risk
risk risk risk risk risk
L. Yang et al. / Urologic Oncology: Seminars and Original Investigations ] (2015) 1–8
5
Fig. 2. (A) Repeat PSA levels in patients with and without antibiotics. (B) Net PSA decrease in patients with and without antibiotics. (C) Number of patients with decreased PSA levels after treatment or control. (D) Repeat PSAD levels in patients with and without antibiotics. (E) Repeat f/t% PSA ratio in patients with and without antibiotics. IV ¼ inverse variance; M-H ¼ Mantel-Haenszel; SD ¼ standard deviation. (Color version of figure is available online.)
3.3.2. Pca-positive rate in cases with responsive PSA There were 5 studies including 69 patients with responsive PSA subsequently investigated the Pca-positive rate in these patients after subjecting them to the intervention or control group [14,16–19].
The Pca-positive rate in patients with a decreased PSA level of o4 ng/ml was reported in 2 studies of 32 patients [14,16]. There were 2 cases of Pca diagnosed among 16 patients randomized to receive antibiotics in comparison with 5 of 16 patients randomized into the control group
Fig. 3. Patients with responsive PSA after treatment or control. (Color version of figure is available online.)
6
L. Yang et al. / Urologic Oncology: Seminars and Original Investigations ] (2015) 1–8
(Fig. 4). There was no significant difference in the Pcapositive rate between the antibiotic and control groups (RR ¼ 0.58, 95% CI: 0.13–2.62, P ¼ 0.48). The Pca-positive rate in patients with a decreased PSA level of o2.5 ng/ml was reported in 4 studies including 51 patients [14,17–19]. Of 28 patients randomized to receive antibiotics, 5 were diagnosed with Pca in comparison with 5 of 23 patients randomized to the control group (Fig. 4). There was no significant difference in the Pca-positive rate between the antibiotic and control groups (RR ¼ 1.05, 95% CI: 0.43–2.56, P ¼ 0.91). One study reported that the only case randomized to receive antibiotics was diagnosed with Pca in comparison with none of the 4 patients randomized to the control group [14]. There was no significant difference between the antibiotic and control groups (P 4 0.05). Overall, there were no significant differences in the Pcapositive rate among patients with responsive PSA in the antibiotic and control groups (RR ¼ 1.07, 95% CI: 0.53– 2.16, P ¼ 0.86; Fig. 4). 3.4. Pca rates in patients after treatment or control As the most noteworthy result for PB in this review, the Pca rates in patients after treatment or control were reported in 4 of the trials included [14,15,17,18]. There were 57 cases of diagnosed Pca among the 223 patients (25.6%) randomized to receive antibiotics in comparison with 55 of 189 patients (29.1%) randomized into the control group (Fig. 5). There was no significant difference in the
Pca-positive rate between the antibiotic and control groups (random-effects model; RR ¼ 0.85, 95% CI: 0.48–1.50, P ¼ 0.57).
4. Discussion Declines in PSA levels ranging from 7.1% to 43% after antibiotic treatment have been reported [20–22]. However, the strategy of empiric antibiotic administration lacks a strong base of evidence. In addition, there are no evidencebased data regarding differences between changes in PSA levels caused by antibiotic treatment and PSA fluctuations observed in normal untreated men [11]. For example, a significant degree of biological variation can be observed in the PSA levels in normal men, and a physiological fluctuation in the PSA level of 10% to 20% was reported in a screened population [23]. Therefore, it is unclear whether the observed changes in PSA levels were because of natural variation, regression toward the mean, or a true effect of the antibiotics. The use of empiric antibiotics for elevated PSA levels may result in side effects and a potentially increased risk of infection after biopsy. The use of antibiotics in this setting is associated with drug-related side effects, the promotion of microbial resistance, and an increased rate of sepsis after PB [24–26]. The outpatient use of fluoroquinolones was associated with an increased incidence of ciprofloxacinresistant and community-acquired Escherichia coli infections [25]. Most pertinent to patients with an elevated PSA
Fig. 4. Pca-positive rate in cases with responsive PSA after treatment or control. (Color version of figure is available online.)
L. Yang et al. / Urologic Oncology: Seminars and Original Investigations ] (2015) 1–8
7
Fig. 5. Pca rates in patients after treatment or control. (Color version of figure is available online.)
level, individuals who received 3 weeks of fluoroquinolonebased antibiotic treatment before biopsy had a significantly higher incidence of postbiopsy sepsis (5.4% vs. 1.7%, P o 0.05), and all patients in whom bacteria were identified in cultures harbored a fluoroquinolone-resistant organism [26]. Therefore, antibiotics should be used more responsibly in patients with elevated PSA levels. There are no data to confirm that it is safe to defer biopsy in patients who exhibited declined PSA values, even to levels of 2.5 or 4 ng/ml. Any decline in PSA levels, however small, is perceived by patients as a “successful” antimicrobial treatment and, despite appropriate counseling, several patients with decreased PSA levels are reluctant to undergo PB. In fact, several patients request a second course of “treatment for the elevated PSA” or decline further management. Such a false sense of security may affect the referring or treating urologist and the patient. In some previous studies, PBs were omitted if PSA levels decreased to o4 ng/ml [7]. However, 15% of men with PSA o4 ng/ml may have biopsy-detectable Pca [27]. In a study by Stopiglia et al. [17], there were 5 cases of Pca in individuals with a PSA level o2.5 ng/ml after antibiotic treatment. In the present meta-analysis, although some studies attempted to demonstrate that the use of empiric antibiotics could either decrease PSA levels [15,16] or reduce the possibility of unnecessary PB [16], most RCTs confirmed that the antibiotic approach was insufficient for decreasing PSA and detecting Pca [14,17–19]. This conclusion is supported by several lines of evidence in the present meta-analysis. First, although 2 studies reported a significant decrease in initial PSA–repeat PSA after antibiotic treatment [15,16], most data regarding the PSA levels after treatment, number of patients with decreased PSA levels after treatment, PSAD levels after treatment, and f/t% PSA after treatment suggested that there was no significant difference between antibiotic use and surveillance. Second, a responsive PSA (a PSA level that decreased to o4 or o2.5 ng/ml or a reduction in PSA level of 450% after treatment) and the number of Pca-positive patients with a responsive PSA were analyzed [14,16–19]. However, there were no obvious differences between the antibiotic and control groups with respect to either parameter. Therefore, when compared with surveillance, antibiotics did not reduce PSA levels or avoid unnecessary PB. Several patients with
responsive PSA in the antibiotic group were diagnosed with Pca; therefore, even if PSA levels were satisfactorily decreased after antibiotic use, it would not be safe to postpone or cancel PBs. Third, the total incidence of Pca was comparable between the antibiotic and control groups [14,15,17,18]. This suggests that treatment with antibiotics is not the most appropriate approach for decreasing PSA levels and minimizing the number of PBs. Furthermore, as suggested by Ugurlu et al. [19], empiric antibiotics could be used only in patients with definite prostatic inflammation or infection. However, the optimal method for identifying the target patients for the application of antibiotics and the safety of these patients with later-responsive PSA remain unclear. There are several limitations to the present study. First, there were only 2 studies that addressed the effects of positivity and negativity of Pca on the PSA levels after antibiotics [16,18]; therefore, no final conclusions regarding PSA variations could be made based on the pathologic results. Second, the number of participants in the enrolled studies was small. Owing to the limited number of clinical trials on this subject, some of the outcomes measured in this meta-analysis were only based on 2 studies. Third, the relationships among the PSA level, inflammation, and age are close and important; however, no information regarding age dependency was available in the included RCTs. Finally, we could not establish a correlation between histologic inflammation and PSA variations after antibiotic therapy based on the available data. Therefore, the conclusions should be considered with caution because the contributing studies varied in a number of factors, including the antibiotic regimen used, the duration of treatment, PSA cutoffs, and timing of PSA testing after antibiotic treatment.
5. Conclusion The results of this systematic review suggest that the use of empiric antibiotics in patients with moderately elevated PSA levels may neither decrease PSA levels significantly nor avoid unnecessary PBs. Therefore, it may not be safe to postpone or cancel PBs in patients who achieve a satisfactory PSA response to antibiotics.
8
L. Yang et al. / Urologic Oncology: Seminars and Original Investigations ] (2015) 1–8
References [1] Polascik TJ, Oesterling JE, Partin AW. Prostate specific antigen: a decade of discovery—what we have learned and where we are going. J Urol 1999;162:293–306. [2] Bostwick DG, Meiers I. Prostate biopsy and optimization of cancer yield. Eur Urol 2006;49:415–7. [3] Siegel R, Ward E, Brawley O, et al. Cancer statistics, 2011: the impact of eliminating socioeconomic and racial disparities on premature cancer deaths. CA Cancer J Clin 2011;61:212–36. [4] Stamey TA, Kabalin JN, Ferrari M. Prostate specific antigen in the diagnosis and treatment of adenocarcinoma of the prostate. III. Radiation treated patients. J Urol 1989;141:1084–7. [5] Raaijmakers R, Wildhagen MF, Ito K, et al. Prostate-specific antigen change in the European Randomized Study of Screening for Prostate Cancer, section Rotterdam. Urology 2004;63:316–20. [6] Gilbert SM, Cavallo CB, Kahane H, et al. Evidence suggesting PSA cutpoint of 2.5 ng/ml for prompting prostate biopsy: review of 26,316 biopsies. Urology 2005;65:549–53. [7] Bozeman CB, Carver BS, Eastham JA, et al. Treatment of chronic prostatitis lowers serum prostate specific antigen. J Urol 2002;167: 1723–1726. [8] Scardino PT. The responsible use of antibiotics for an elevated PSA level. Nat Clin Pract Urol 2007;4:1. [9] Baltaci S, Suer E, Haliloglu AH, et al. Effectiveness of antibiotics given to asymptomatic men for an increased prostate specific antigen. J Urol 2009;181:128–32. [10] Bunting PS, DeBoer G, Choo R, et al. Intraindividual variation of PSA, free PSA and complexed PSA in a cohort of patients with prostate cancer managed with watchful observation. Clin Biochem 2002;35:471–5. [11] Eastham JA, Riedel E, Scardino PT, et al. Variation of serum prostate-specific antigen levels: an evaluation of year-to-year fluctuations. J Am Med Assoc 2003;289:2695–700. [12] Higgins JPT, Green S (editors). Cochrane Handbook for Systematic Reviews of Interventions Version 5.1.4 [updated March 2011]. The Cochrane Collaboration, 2011. www.cochrane-handbook.org. [13] Moher D, Liberati A, Tetzlaff J, et al. Preferred reporting items for systematic reviews and meta-analyzes: the PRISMA statement. Br Med J 2009;339:b2535. [14] Eggener SE, Large MC, Gerber GS, et al. Empiric antibiotics for an elevated prostate-specific antigen (PSA) level: a randomised, prospective, controlled multi-institutional trial. BJU Int 2013;112: 925–929.
[15] Erol H, Beder N, Calişkan T, et al. Can the effect of antibiotherapy and anti-inflammatory therapy on serum PSA levels discriminate between benign and malign prostatic pathologies? Urol Int 2006;76:20–6. [16] Saribacak A, Yilmaz H, Ciftci S, et al. The role of empiric antibiotic treatment in preventing unnecessary prostate biopsies in asymptomatic patients with PSA levels between 4 and 10 ng/ml. Int J Clin Exp Med 2014;15:2230–5. [17] Stopiglia RM, Ferreira U, Silva MM Jr, et al. Prostate specific antigen decrease and prostate cancer diagnosis: antibiotic versus placebo prospective randomized clinical trial. J Urol 2010;183:940–4. [18] Toktas G, Demiray M, Erkan E, et al. The effect of antibiotherapy on prostate-specific antigen levels and prostate biopsy results in patients with levels 2.5 to 10 ng/mL. J Endourol 2013;27:1061–7. [19] Ugurlu O, Yaris M, Oztekin CV, et al. Impacts of antibiotic and antiinflammatory therapies on serum prostate-specific antigen levels in the presence of prostatic inflammation: a prospective randomized controlled trial. Urol Int 2010;84:185–90. [20] Kyung YS, Lee HC, Kim HJ. Changes in serum prostate-specific antigen after treatment with antibiotics in patients with lower urinary tract symptoms/benign prostatic hyperplasia with prostatitis. Int Neurourol J 2010;14:100–4. [21] Karazanashvili G, Managadze L. Prostate-specific antigen (PSA) value change after antibacterial therapy of prostate inflammation, as a diagnostic method for prostate cancer screening in cases of PSA value within 4–10 ng/ml and nonsuspicious results of digital rectal examination. Eur Urol 2001;39:538–43. [22] Serretta V, Catanese A, Daricello G, et al. PSA reduction (after antibiotics) permits to avoid or postpone prostate biopsy in selected patients. Prostate Cancer Prostatic Dis 2008;11:148–52. [23] Zhang L, Loblaw A, Klotz L. Modeling prostate specific antigen kinetics in patients on active surveillance. J Urol 2006;176:1392–7. [24] Akduman B, Akduman D, Tokgöz H, et al. Long-term fluoroquinolone use before the prostate biopsy may increase the risk of sepsis caused by resistant microorganisms. Urology 2011;78:250–5. [25] Kahlmeter G, Menday P, Cars O. Non-hospital antimicrobial usage and resistance in community-acquired Escherichia coli urinary tract infection. J Antimicrob Chemother 2003;52:1005–10. [26] Feliciano J, Teper E, Ferrandino M, et al. The incidence of fluoroquinolone resistant infections after prostate biopsy—are fluoroquinolones still effective prophylaxis? J Urol 2008;179:952–5. [27] Thompson IM, Pauler DK, Goodman PJ, et al. Prevalence of prostate cancer among men with a prostate-specific antigen level o or 4.0 ng per milliliter. N Engl J Med 2004;350:2239–46.
Original article
Antibiotics may not decrease prostate-specific antigen levels or prevent unnecessary prostate biopsy in patients with moderately increased prostate-specific antigen levels: A meta-analysis Lu Yang, M.D., Yuchun Zhu, M.D., Zhuang Tang, M.D., Yongji Chen, M.D., Liang Gao, M.D., Liangren Liu, M.D., Ping Han, M.D., Xiang Li, M.D.*, Qiang Wei, M.D.* Department of Urology, West China Hospital, Sichuan University, Chengdu, China Received 2 December 2014; received in revised form 2 February 2015; accepted 2 February 2015
Abstract Objectives: To evaluate the effect of empiric antibiotics on decreasing prostate-specific antigen (PSA) levels and the possibility of avoiding unnecessary prostate biopsies (PBs). Materials and methods: A systematic search of PubMed, Embase, and the Cochrane Library was performed to identify all randomized controlled trials (RCTs) that compared effects of empiric antibiotics with no treatment or placebo on lowering PSA levels and minimizing unnecessary PBs in patients with moderately increased PSA levels. The Cochrane Collaboration Review Manager software (RevMan 5.1.4) was used for statistical analysis. Results: The inclusion criteria for the study were met by 6 RCTs (1 placebo controlled and 5 no treatment controlled) involving 656 patients. The synthesized data from these RCTs indicated that there were no significant differences between the antibiotic and control groups in the PSA levels after treatment (mean difference [MD] ¼ 0.15, 95% CI: 0.50 to 0.81, P ¼ 0.65], number of patients with decreased PSA levels after treatment (relative risk [RR] ¼ 1.22, 95% CI: 0.90–1.65, P ¼ 0.20], prostate-specific antigen density levels after treatment (MD ¼ 0.04, 95% CI: 0.15 to 0.07, P ¼ 0.47), f/t% PSA after treatment (MD ¼ 1.47, 95% CI: 4.65 to 1.71, P ¼ 0.37), number of patients with responsive PSA (RR ¼ 1.02, 95% CI: 0.58–1.81, P ¼ 0.94), and individual Pca-positive rate in these patients (RR ¼ 1.07, 95% CI: 0.53–2.16, P ¼ 0.86), and Pca-positive rates (RR ¼ 0.85, 95% CI: 0.48–1.50, P ¼ 0.57). However, the antibiotic group had a significant change in the net PSA decrease after treatment compared with the control group (MD ¼ 1.44, 95% CI: 0.70–2.17, P ¼ 0.0001). Conclusion: The use of empiric antibiotics may not significantly decrease PSA levels or avoid unnecessary PBs. r 2015 Elsevier Inc. All rights reserved.
Keywords: Antibiotic; Prostate biopsy; Prostate cancer; Prostate-specific antigen; Randomized controlled trial
1. Introduction
Lu Yang and Yuchun Zhu contributed equally to this work and should share the co–first author. This study was supported by the Prostate Cancer Foundation Young Investigator Award 2013, the National Natural Science Foundation of China (Grant No. 81300627, 81370855, and 30901484), and Program from Science and Technology Department of Sichuan Province (Grant No. 2013SZ0006 and 2014JY0219). * Corresponding authors. Tel.: þ86-189-8060-1425; fax: þ86-28-85422451. E-mail addresses: [email protected] (X. Li), weiqiang933@126. com (Q. Wei). http://dx.doi.org/10.1016/j.urolonc.2015.02.001 1078-1439/r 2015 Elsevier Inc. All rights reserved.
Prostate cancer (Pca), the most common cancer in men in Western countries, has been predicted and diagnosed at an early stage since prostate-specific antigen (PSA) testing began in the 1980s [1]. Worldwide, millions of prostate biopsies (PBs) are performed annually [2,3], most commonly because of elevated serum PSA levels. However, PSA levels can increase for several reasons, including trauma, ejaculation, and rectal and urethral procedures. In addition, numerous noncancer etiologies can cause elevated PSA levels, such as benign prostatic hyperplasia, inflammation, and infection [4].
2
L. Yang et al. / Urologic Oncology: Seminars and Original Investigations ] (2015) 1–8
A PSA range of 4 to 10 ng/ml is commonly recognized as the diagnostic “gray zone” [5]. Some estimates suggest that only 20% to 30% of men with PSA levels between 4 and 10 ng/ml actually have PCa [5]. There is a trend toward reducing the threshold PSA level to 2.5 ng/ml to detect more cases of Pca. Consistent with this, Gilbert et al. [6] reported a cancer detection rate of 27.48% in patients with a PSA level of 2.5 to 4 ng/ml by reviewing 26,316 biopsies. Therefore, how to best identify patients with Pca with PSA levels of 2.5 to 10 ng/ml and avoid unnecessary PB remains challenging. Antimicrobial treatment in patients with increased PSA levels and confirmed or suspected nonmalignancy may decrease the number of unnecessary PBs [5]. Therefore, a cost-effective and painless approach for decreasing the number of biopsies with negative findings may be available [7]. In clinical practice, many urologists use empiric antibiotic treatments, followed by a repeat PSA test in patients with moderately increased PSA. However, scientific evidence to support this approach is currently lacking [8,9]. Nevertheless, critics argue that not only is there is a lack of appropriate studies documenting that antibiotics can decrease PSA levels, but they are also no data to suggest that it is safe to defer biopsies in patients who exhibit a decrease in PSA levels after antibiotic treatment [10,11]. 2. Materials and methods 2.1. Search strategy We searched PubMed (1980 to May 2014), Embase (1980 to May 2014), the Cochrane Central Register of Controlled Trials, and the Cochrane Database of Systematic Reviews. The initial search process was designed to identify all relevant trials involving the terms “antibiotic,” “prostatespecific antigen”, “prostate biopsy”, and “prostate cancer” (and multiple synonyms for each term). Furthermore, the reference lists of relevant publications and conference proceedings from the American Urological Association, European Association of Urology, and Société Internationale d'Urologie between 2009 and 2014 were searched manually. Computer searches were supplemented with manual searches. Two authors (L.Y. and L.G.) screened all citations and abstracts isolated using this search strategy independently to identify potentially eligible studies.
symptoms, signs, or history of prostatitis; a prior prostatic surgery, biopsy, radiotherapy, or hormonal therapy; any abnormal findings in DRE; an acute infection of the urinary system, pyuria, and bacterial growth in asymptomatic patients; a recent history of instrumentation in the urinary system; previous use of a 5-α reductase inhibitor; a residual urine volume 4100 ml; and hypersensitivity to any medical ingredient in the antibiotics used. 2.3. Interventions and comparisons All patients with moderately increased PSA levels (generally between 2.5 [or 4] and 10 ng/ml) and normal findings on DRE were divided into an antibiotic group or untreated/placebo (control) group. Comparisons were made on the outcomes of antibiotic therapy or not in the decrease of PSA levels and the avoidance of unnecessary PBs. 2.4. Outcome measures The outcomes measured in this review included the following: the PSA level after treatment, net PSA decrease after treatment, number of patients with a decreased PSA after treatment, PSA density (PSAD, the ratio of serum total level to the prostate volume) level after treatment, f/t% PSA after treatment, number of patients with responsive PSA and individual rate of Pca-positive in these patients, and Pcapositive rates in antibiotic and control groups. The definition of responsive PSA was a decreased PSA level o4 or 2.5 ng/ml according to the individual RCTs or a reduction in PSA level of 450% after treatment. The net PSA decrease after treatment was defined as the initial PSArepeat PSA in both groups. 2.5. Data extraction Data were independently extracted by 2 authors (Y.J.C. and Z.T.) using a predesigned data extraction form. The extracted data included the data source, eligibility, methods, participant characteristics, interventions, and results. Both authors convened to synthesize their findings. The information was subsequently entered into RevMan 5.1.4. Any discrepancy among the extracted data was resolved by discussion, and any disagreements that persisted after discussion were resolved in consultation with another author (W.Q.).
2.2. Participants
2.6. Quality assessment
The inclusion criteria were as follows: male patients with moderately increased PSA levels (generally between 2.5 [or 4] and 10 ng/ml), who had a normal findings on digital rectal examination (DRE), who received antibiotics or no treatment/placebo for several weeks, and who then underwent a subsequent PSA test and PB. Patients with specific conditions were excluded, including those with any
The quality of included studies was assessed by 2 authors (L.L.R. and H.P.) according to the Cochrane Collaboration Reviewers' Handbook and the QUOROM guidelines [12,13]. The quality items assessed included the generation of randomization sequences, blinding method used, allocation concealment, description of patient withdrawals and dropouts, and intent-to-treat analysis.
L. Yang et al. / Urologic Oncology: Seminars and Original Investigations ] (2015) 1–8
2.7. Statistical analysis All meta-analyses were performed using RevMan 5.1.4. Dichotomous data and continuous data were presented as relative risk (RR) and mean difference (MD), respectively, with 95% CIs. Meta-analysis was performed using the fixed-effects method or the random-effects method. The fixed-effects method was used to combine the results when statistically significant heterogeneity was absent, whereas the random-effects method was applied when heterogeneity was present. Statistical heterogeneity among trials was evaluated using the I2 test with significance set at P o 0.05. Publication bias was evaluated using a funnel plot. Sensitivity analysis was performed if low-quality trials were included.
3
ciprofloxacin [14,15], 2 used ofloxacin [16,17], and 2 used levofloxacin [18,19]. The treatment duration was 2 to 4 weeks. Our review of funnel plots excluded publication bias. 3.2. Variations in levels of PSA and PSA derivatives after treatment 3.2.1. PSA levels after treatment PSA levels after treatment were reported in 3 studies with 314 patients [14,15,18]. In total, 173 patients were randomized to receive antibiotics, and 141 patients were randomized into the control group (Fig. 2A). There was no significant difference between the antibiotic and control groups (MD ¼ 0.15, 95% CI: 0.50–0.81, P ¼ 0.65).
3. Results 3.1. Description of studies In total, 1,175 reports were initially identified from the database and manual searches. After removing 344 records of redundant publications, reviews, and meta-analysis and scanning the titles and abstracts of 780 unrelated records, 1,124 reports were excluded from the study. After referring to the full texts and excluding 14 abstracts, 22 nonrandomized studies, and 9 studies irrelevant to the intervention of interest or outcome of interest, we identified and included 6 publications (randomized controlled trials [RCTs], including 1 study published after the abstract at meeting) that enrolled 656 patients in the review (Fig. 1) [14–19]. The characteristics and quality of the included studies are presented in Tables 1 and 2, and the search flow diagram is presented in Fig. 1. There were no significant differences in baseline information between the antibiotic and control groups. There were 5 studies containing 512 patients that were controlled using an untreated group without any blinding method applied [14–16,18,19], and 1 study involving 144 patients that was placebo controlled with doubleblinded method (Table 2) [17]. Additionally, 2 studies used
Fig. 1. Flow diagram of systematic review.
3.2.2. Net PSA decrease after treatment Data on the net PSA decrease were collected from 2 trials including 197 patients [15,16], of whom 115 and 82 were randomized to receive antibiotics and the control group, respectively (Fig. 2B). The net PSA decrease was significantly higher in the antibiotic group than in the control group (MD ¼ 1.44, 95% CI: 0.70–2.17, P ¼ 0.0001). 3.2.3. Number of patients with decreased PSA levels after treatment Two studies including 195 patients investigated the number of patients with decreased PSA levels after treatment [14,17]. There were 47 cases with decreased PSA levels among 87 patients randomized to receive antibiotics and 39 cases among 88 patients randomized to the control group (Fig. 2C). There was no significant difference between the antibiotic and control groups (RR ¼ 1.22, 95% CI: 0.90–1.65, P ¼ 0.20). 3.2.4. PSAD levels after treatment PSAD levels after treatment were reported in 2 studies including 237 patients [15,18]. There were 135 and 102 patients randomized to receive antibiotics and into the control group, respectively (Fig. 2D). There was no significant difference between the antibiotic and control groups (random-effects model; MD ¼ 0.04, 95% CI: 0.15 to 0.07, P ¼ 0.47). 3.2.5. f/t% PSA ratio after treatment The f/t% PSA ratio after treatment was reported in 2 studies that analyzed 237 patients [15,18]. There were 135 and 102 patients randomized to receive antibiotics and into the control group, respectively (Fig. 2E). There was no significant difference between the antibiotic and control groups (MD ¼ 1.47, 95% CI: 4.65 to 1.71, P ¼ 0.37).
L. Yang et al. / Urologic Oncology: Seminars and Original Investigations ] (2015) 1–8
4
Table 1 Characteristics of treatment groups in 6 randomized controlled trials Reference
Drug (6)/placebo (1) or no drug (5)
Inclusive criteria PSA range, ng/ml
DRE results
Eggener et al. 39/38 [14] Erol et al. [15] 65/32
42.5
Negative
44
ND
Saribacak et al. [16]
4–10
Negative
2.5–10
Negative
2.5–10
Negative
2.5–10
Negative
50/50
Stopiglia et al. 49/49 [17] Toktas et al. 70/70 [18] Ugurlu et al. [19]
72/72
Antibiotic types and durations
Inflammation diagnosed before inclusion
Repeat PSA and PSA derivatives
No
2 Weeks of ciprofloxacin, PSA 500 mg twice daily No Ciprofloxacin 500 mg PSA, fPSA, twice daily, and PSAD, and diclofenac 75 mg f/t% PSA for 2-3 weeks No 400 mg of ofloxacin PSA, fPSA, daily for 4 weeks PSAD, and f/t% PSA No 500-mg ciprofloxacin PSA for 28 days No 500-mg oral levofloxacin PSA, fPSA, once a day for 21 days PSAD, f/t% PSA According to EPS Levofloxacin 500 mg PSA, fPSA, results, divided into once a PSAD, and EPSþ and EPS day for 3 weeks f/t% PSA
Repeat prostate biopsy All patients All patients
All patients
All patients All patients
All patients
EPS ¼ expressed prostate secretion, ND ¼ not determined.
3.3. Responsive PSA and Pca-positive cases in patients with responsive PSA Subgroup analyses were performed that included patients with PSA levels that decreased to o4 ng/ml or o2.5 ng/ml and showed 450% reduction in PSA level.
3.3.1. Responsive PSA There were 5 studies that included 559 patients and investigated the number of patients with responsive PSA after intervention or control [14,16–19]. Patients whose PSA levels decreased to o4 ng/ml were reported in 2 studies with 177 patients [14,16]. Of 88 cases randomized to receive antibiotics, 16 exhibited decreased PSA levels o4 ng/ml in comparison with 16 of 89 patients randomized to the control group (Fig. 3). There was no significant difference between the antibiotic and control
groups (random-effects model; RR ¼ 0.93, 95% CI: 0.20– 4.40, P ¼ 0.93). Patients whose PSA level decreased to o2.5 ng/ml were reported in 4 studies of 387 patients [14,17–19]. In total, 28 of 193 cases randomized to receive antibiotics exhibited decreased PSA levels of o2.5 ng/ml in comparison with 23 of 194 patients randomized to the control group (Fig. 3). There was no significant difference between groups (random-effects model; RR ¼ 1.21, 95% CI: 0.68–2.14, P ¼ 0.52). There was 1 study that reported 1 case with a reduced PSA level 450% in 38 patients randomized to receive antibiotics in comparison with 4 of 39 patients randomized to the control group [14]. There was no significant difference between the antibiotic and control groups (P 4 0.05). Overall, there were no significant differences in cases of responsive PSA between the antibiotic and control groups (random-effects model; RR ¼ 1.02, 95% CI: 0.58–1.81, P ¼ 0.94; Fig. 3).
Table 2 Quality of included randomized controlled trials Reference
Drug (6)/placebo Generation of (1) or no drug (5) Randomization sequences
Allocation concealment
Blinding
Incomplete outcome data addressed
Free of selective reporting
Free of other bias
Country
Eggener et al. [14]
39/38
Low risk
Unclear risk
ND
Low risk
Low risk
Low risk
Erol et al. [15] Saribacak et al. [16] Stopiglia et al. [17] Toktas et al. [18] Ugurlu et al. [19]
65/32 50/50 49/49 70/70 72/72
Low risk Low risk Low risk Unclear risk Unclear risk
Low risk Unclear risk Unclear risk Unclear risk Low risk
ND ND Low risk ND ND
Low Low Low Low Low
Low Low Low Low Low
Low Low Low Low Low
US, Israel, and Canada Turkey Turkey Brazil Turkey Turkey
ND ¼ not determined.
risk risk risk risk risk
risk risk risk risk risk
risk risk risk risk risk
L. Yang et al. / Urologic Oncology: Seminars and Original Investigations ] (2015) 1–8
5
Fig. 2. (A) Repeat PSA levels in patients with and without antibiotics. (B) Net PSA decrease in patients with and without antibiotics. (C) Number of patients with decreased PSA levels after treatment or control. (D) Repeat PSAD levels in patients with and without antibiotics. (E) Repeat f/t% PSA ratio in patients with and without antibiotics. IV ¼ inverse variance; M-H ¼ Mantel-Haenszel; SD ¼ standard deviation. (Color version of figure is available online.)
3.3.2. Pca-positive rate in cases with responsive PSA There were 5 studies including 69 patients with responsive PSA subsequently investigated the Pca-positive rate in these patients after subjecting them to the intervention or control group [14,16–19].
The Pca-positive rate in patients with a decreased PSA level of o4 ng/ml was reported in 2 studies of 32 patients [14,16]. There were 2 cases of Pca diagnosed among 16 patients randomized to receive antibiotics in comparison with 5 of 16 patients randomized into the control group
Fig. 3. Patients with responsive PSA after treatment or control. (Color version of figure is available online.)
6
L. Yang et al. / Urologic Oncology: Seminars and Original Investigations ] (2015) 1–8
(Fig. 4). There was no significant difference in the Pcapositive rate between the antibiotic and control groups (RR ¼ 0.58, 95% CI: 0.13–2.62, P ¼ 0.48). The Pca-positive rate in patients with a decreased PSA level of o2.5 ng/ml was reported in 4 studies including 51 patients [14,17–19]. Of 28 patients randomized to receive antibiotics, 5 were diagnosed with Pca in comparison with 5 of 23 patients randomized to the control group (Fig. 4). There was no significant difference in the Pca-positive rate between the antibiotic and control groups (RR ¼ 1.05, 95% CI: 0.43–2.56, P ¼ 0.91). One study reported that the only case randomized to receive antibiotics was diagnosed with Pca in comparison with none of the 4 patients randomized to the control group [14]. There was no significant difference between the antibiotic and control groups (P 4 0.05). Overall, there were no significant differences in the Pcapositive rate among patients with responsive PSA in the antibiotic and control groups (RR ¼ 1.07, 95% CI: 0.53– 2.16, P ¼ 0.86; Fig. 4). 3.4. Pca rates in patients after treatment or control As the most noteworthy result for PB in this review, the Pca rates in patients after treatment or control were reported in 4 of the trials included [14,15,17,18]. There were 57 cases of diagnosed Pca among the 223 patients (25.6%) randomized to receive antibiotics in comparison with 55 of 189 patients (29.1%) randomized into the control group (Fig. 5). There was no significant difference in the
Pca-positive rate between the antibiotic and control groups (random-effects model; RR ¼ 0.85, 95% CI: 0.48–1.50, P ¼ 0.57).
4. Discussion Declines in PSA levels ranging from 7.1% to 43% after antibiotic treatment have been reported [20–22]. However, the strategy of empiric antibiotic administration lacks a strong base of evidence. In addition, there are no evidencebased data regarding differences between changes in PSA levels caused by antibiotic treatment and PSA fluctuations observed in normal untreated men [11]. For example, a significant degree of biological variation can be observed in the PSA levels in normal men, and a physiological fluctuation in the PSA level of 10% to 20% was reported in a screened population [23]. Therefore, it is unclear whether the observed changes in PSA levels were because of natural variation, regression toward the mean, or a true effect of the antibiotics. The use of empiric antibiotics for elevated PSA levels may result in side effects and a potentially increased risk of infection after biopsy. The use of antibiotics in this setting is associated with drug-related side effects, the promotion of microbial resistance, and an increased rate of sepsis after PB [24–26]. The outpatient use of fluoroquinolones was associated with an increased incidence of ciprofloxacinresistant and community-acquired Escherichia coli infections [25]. Most pertinent to patients with an elevated PSA
Fig. 4. Pca-positive rate in cases with responsive PSA after treatment or control. (Color version of figure is available online.)
L. Yang et al. / Urologic Oncology: Seminars and Original Investigations ] (2015) 1–8
7
Fig. 5. Pca rates in patients after treatment or control. (Color version of figure is available online.)
level, individuals who received 3 weeks of fluoroquinolonebased antibiotic treatment before biopsy had a significantly higher incidence of postbiopsy sepsis (5.4% vs. 1.7%, P o 0.05), and all patients in whom bacteria were identified in cultures harbored a fluoroquinolone-resistant organism [26]. Therefore, antibiotics should be used more responsibly in patients with elevated PSA levels. There are no data to confirm that it is safe to defer biopsy in patients who exhibited declined PSA values, even to levels of 2.5 or 4 ng/ml. Any decline in PSA levels, however small, is perceived by patients as a “successful” antimicrobial treatment and, despite appropriate counseling, several patients with decreased PSA levels are reluctant to undergo PB. In fact, several patients request a second course of “treatment for the elevated PSA” or decline further management. Such a false sense of security may affect the referring or treating urologist and the patient. In some previous studies, PBs were omitted if PSA levels decreased to o4 ng/ml [7]. However, 15% of men with PSA o4 ng/ml may have biopsy-detectable Pca [27]. In a study by Stopiglia et al. [17], there were 5 cases of Pca in individuals with a PSA level o2.5 ng/ml after antibiotic treatment. In the present meta-analysis, although some studies attempted to demonstrate that the use of empiric antibiotics could either decrease PSA levels [15,16] or reduce the possibility of unnecessary PB [16], most RCTs confirmed that the antibiotic approach was insufficient for decreasing PSA and detecting Pca [14,17–19]. This conclusion is supported by several lines of evidence in the present meta-analysis. First, although 2 studies reported a significant decrease in initial PSA–repeat PSA after antibiotic treatment [15,16], most data regarding the PSA levels after treatment, number of patients with decreased PSA levels after treatment, PSAD levels after treatment, and f/t% PSA after treatment suggested that there was no significant difference between antibiotic use and surveillance. Second, a responsive PSA (a PSA level that decreased to o4 or o2.5 ng/ml or a reduction in PSA level of 450% after treatment) and the number of Pca-positive patients with a responsive PSA were analyzed [14,16–19]. However, there were no obvious differences between the antibiotic and control groups with respect to either parameter. Therefore, when compared with surveillance, antibiotics did not reduce PSA levels or avoid unnecessary PB. Several patients with
responsive PSA in the antibiotic group were diagnosed with Pca; therefore, even if PSA levels were satisfactorily decreased after antibiotic use, it would not be safe to postpone or cancel PBs. Third, the total incidence of Pca was comparable between the antibiotic and control groups [14,15,17,18]. This suggests that treatment with antibiotics is not the most appropriate approach for decreasing PSA levels and minimizing the number of PBs. Furthermore, as suggested by Ugurlu et al. [19], empiric antibiotics could be used only in patients with definite prostatic inflammation or infection. However, the optimal method for identifying the target patients for the application of antibiotics and the safety of these patients with later-responsive PSA remain unclear. There are several limitations to the present study. First, there were only 2 studies that addressed the effects of positivity and negativity of Pca on the PSA levels after antibiotics [16,18]; therefore, no final conclusions regarding PSA variations could be made based on the pathologic results. Second, the number of participants in the enrolled studies was small. Owing to the limited number of clinical trials on this subject, some of the outcomes measured in this meta-analysis were only based on 2 studies. Third, the relationships among the PSA level, inflammation, and age are close and important; however, no information regarding age dependency was available in the included RCTs. Finally, we could not establish a correlation between histologic inflammation and PSA variations after antibiotic therapy based on the available data. Therefore, the conclusions should be considered with caution because the contributing studies varied in a number of factors, including the antibiotic regimen used, the duration of treatment, PSA cutoffs, and timing of PSA testing after antibiotic treatment.
5. Conclusion The results of this systematic review suggest that the use of empiric antibiotics in patients with moderately elevated PSA levels may neither decrease PSA levels significantly nor avoid unnecessary PBs. Therefore, it may not be safe to postpone or cancel PBs in patients who achieve a satisfactory PSA response to antibiotics.
8
L. Yang et al. / Urologic Oncology: Seminars and Original Investigations ] (2015) 1–8
References [1] Polascik TJ, Oesterling JE, Partin AW. Prostate specific antigen: a decade of discovery—what we have learned and where we are going. J Urol 1999;162:293–306. [2] Bostwick DG, Meiers I. Prostate biopsy and optimization of cancer yield. Eur Urol 2006;49:415–7. [3] Siegel R, Ward E, Brawley O, et al. Cancer statistics, 2011: the impact of eliminating socioeconomic and racial disparities on premature cancer deaths. CA Cancer J Clin 2011;61:212–36. [4] Stamey TA, Kabalin JN, Ferrari M. Prostate specific antigen in the diagnosis and treatment of adenocarcinoma of the prostate. III. Radiation treated patients. J Urol 1989;141:1084–7. [5] Raaijmakers R, Wildhagen MF, Ito K, et al. Prostate-specific antigen change in the European Randomized Study of Screening for Prostate Cancer, section Rotterdam. Urology 2004;63:316–20. [6] Gilbert SM, Cavallo CB, Kahane H, et al. Evidence suggesting PSA cutpoint of 2.5 ng/ml for prompting prostate biopsy: review of 26,316 biopsies. Urology 2005;65:549–53. [7] Bozeman CB, Carver BS, Eastham JA, et al. Treatment of chronic prostatitis lowers serum prostate specific antigen. J Urol 2002;167: 1723–1726. [8] Scardino PT. The responsible use of antibiotics for an elevated PSA level. Nat Clin Pract Urol 2007;4:1. [9] Baltaci S, Suer E, Haliloglu AH, et al. Effectiveness of antibiotics given to asymptomatic men for an increased prostate specific antigen. J Urol 2009;181:128–32. [10] Bunting PS, DeBoer G, Choo R, et al. Intraindividual variation of PSA, free PSA and complexed PSA in a cohort of patients with prostate cancer managed with watchful observation. Clin Biochem 2002;35:471–5. [11] Eastham JA, Riedel E, Scardino PT, et al. Variation of serum prostate-specific antigen levels: an evaluation of year-to-year fluctuations. J Am Med Assoc 2003;289:2695–700. [12] Higgins JPT, Green S (editors). Cochrane Handbook for Systematic Reviews of Interventions Version 5.1.4 [updated March 2011]. The Cochrane Collaboration, 2011. www.cochrane-handbook.org. [13] Moher D, Liberati A, Tetzlaff J, et al. Preferred reporting items for systematic reviews and meta-analyzes: the PRISMA statement. Br Med J 2009;339:b2535. [14] Eggener SE, Large MC, Gerber GS, et al. Empiric antibiotics for an elevated prostate-specific antigen (PSA) level: a randomised, prospective, controlled multi-institutional trial. BJU Int 2013;112: 925–929.
[15] Erol H, Beder N, Calişkan T, et al. Can the effect of antibiotherapy and anti-inflammatory therapy on serum PSA levels discriminate between benign and malign prostatic pathologies? Urol Int 2006;76:20–6. [16] Saribacak A, Yilmaz H, Ciftci S, et al. The role of empiric antibiotic treatment in preventing unnecessary prostate biopsies in asymptomatic patients with PSA levels between 4 and 10 ng/ml. Int J Clin Exp Med 2014;15:2230–5. [17] Stopiglia RM, Ferreira U, Silva MM Jr, et al. Prostate specific antigen decrease and prostate cancer diagnosis: antibiotic versus placebo prospective randomized clinical trial. J Urol 2010;183:940–4. [18] Toktas G, Demiray M, Erkan E, et al. The effect of antibiotherapy on prostate-specific antigen levels and prostate biopsy results in patients with levels 2.5 to 10 ng/mL. J Endourol 2013;27:1061–7. [19] Ugurlu O, Yaris M, Oztekin CV, et al. Impacts of antibiotic and antiinflammatory therapies on serum prostate-specific antigen levels in the presence of prostatic inflammation: a prospective randomized controlled trial. Urol Int 2010;84:185–90. [20] Kyung YS, Lee HC, Kim HJ. Changes in serum prostate-specific antigen after treatment with antibiotics in patients with lower urinary tract symptoms/benign prostatic hyperplasia with prostatitis. Int Neurourol J 2010;14:100–4. [21] Karazanashvili G, Managadze L. Prostate-specific antigen (PSA) value change after antibacterial therapy of prostate inflammation, as a diagnostic method for prostate cancer screening in cases of PSA value within 4–10 ng/ml and nonsuspicious results of digital rectal examination. Eur Urol 2001;39:538–43. [22] Serretta V, Catanese A, Daricello G, et al. PSA reduction (after antibiotics) permits to avoid or postpone prostate biopsy in selected patients. Prostate Cancer Prostatic Dis 2008;11:148–52. [23] Zhang L, Loblaw A, Klotz L. Modeling prostate specific antigen kinetics in patients on active surveillance. J Urol 2006;176:1392–7. [24] Akduman B, Akduman D, Tokgöz H, et al. Long-term fluoroquinolone use before the prostate biopsy may increase the risk of sepsis caused by resistant microorganisms. Urology 2011;78:250–5. [25] Kahlmeter G, Menday P, Cars O. Non-hospital antimicrobial usage and resistance in community-acquired Escherichia coli urinary tract infection. J Antimicrob Chemother 2003;52:1005–10. [26] Feliciano J, Teper E, Ferrandino M, et al. The incidence of fluoroquinolone resistant infections after prostate biopsy—are fluoroquinolones still effective prophylaxis? J Urol 2008;179:952–5. [27] Thompson IM, Pauler DK, Goodman PJ, et al. Prevalence of prostate cancer among men with a prostate-specific antigen level o or 4.0 ng per milliliter. N Engl J Med 2004;350:2239–46.
