1377Vol. 11 | No. 02 | March- April 2014 |U R O LO G Y J O U R N A L 1 Department of Urology, Nan- fang Hospital, Southern Medical University, Guangzhou, China. 2 Section 5 Department of Internal Medicine, Guilin TCM Hospital of China, Guilin, China. Yongtong Zhu,1 Chunyan Wang,2 Xiang Pang,1 Fei Li,1 Wei Chen,1 Wanlong Tan1 Antibiotics Are Not Beneficial in the Management of Category III Prostatitis: A Meta-Analysis Corresponding Author: Wan-long Tan, MD Department of Urology, Nanfang Hospital Affiliated to Southern Medical University, Guangzhou 510515, China. Tel: +86 020 61641762 E-mail: tanwanlong@gmail.com Received October 2012 Accepted March 2013 Purpose:‎To‎determine‎whether‎antibiotics‎are‎beneficial‎in‎the‎management‎of‎category‎III‎ prostatitis. Materials and Methods:‎The‎PubMed,‎Medline‎and‎Embase‎databases‎were‎searched‎for‎all‎ published‎documents‎from‎January‎1,‎1965‎to‎September‎1,‎2012‎without‎language‎restriction.‎ The‎randomized‎controlled‎trials‎that‎mentioned‎comparable‎groups‎of‎antibiotics‎treatment‎ versus‎placebo‎or‎other‎control‎group‎for‎patients‎with‎category‎III‎prostatitis‎were‎included‎ based‎on‎specific‎criteria.‎The‎quality‎of‎studies‎was‎assessed‎by‎the‎modified‎Jadad‎scale,‎and‎ Revman‎5.0‎software‎was‎used‎for‎data‎syntheses‎and‎analysis. Results:‎Seven‎studies‎which‎met‎the‎selection‎criteria‎were‎included‎in‎this‎review.‎All‎of‎ them‎were‎high‎quality‎according‎to‎the‎modified‎Jadad‎scale.‎A‎random‎effect‎model‎was‎ap- plied‎because‎of‎the‎high‎heterogeneity.‎The‎meta-analysis‎showed‎that‎summary‎association‎ between‎category‎III‎prostatitis‎and‎antibiotics‎were‎not‎statistically‎significant.‎ Conclusion:‎Our‎meta-analysis‎reveals‎that‎antibiotics‎are‎not‎beneficial‎in‎the‎management‎of‎ category‎III‎prostatitis.‎Therefore,‎we‎may‎reduce‎the‎usage‎of‎antibiotics‎in‎such‎a‎population. Keywords:‎prostatitis;‎drug‎therapy;‎treatment‎failure;‎classification;‎treatment‎outcome;‎meta- analysis. REVIEW ARTICLE 1378 | Review Article INTRODUCTION Antibiotics‎are‎one‎of‎the‎most‎common‎treat-ments‎employed‎by‎urologists‎for‎patients‎pre-senting‎ with‎ prostatitis,‎ regardless‎ of‎ culture‎ results.‎More‎than‎90%‎of‎prostatitis‎cases‎are‎category‎ III‎prostatitis‎which‎is‎not‎associated‎with‎a‎significant‎ bacteriuria.‎Whereas,‎it‎is‎a‎condition‎referred‎to‎chronic‎ prostatitis/chronic‎pelvic‎pain‎syndrome‎(CP/CPPS).(1,2) Subgroups‎of‎CPPS‎are‎inflammatory‎CPPS‎(IIIA)‎where‎ leukocytes‎ are‎ found‎ in‎ the‎ expressed‎ prostatic‎ secre- tions,‎and‎non-inflammatory‎CPPS‎(IIIB).(3) According to‎the‎summary‎of‎National‎Institutes‎of‎Health‎(NIH),‎ patients‎with‎category‎III‎prostatitis‎are‎advised‎to‎take‎ antimicrobial‎agents‎for‎3-6‎weeks‎as‎the‎first-line‎treat- ment,(3,4)‎ which‎ response‎ to‎ the‎ anti-infective‎ therapy.‎ One‎ systematic‎ review‎ published‎ by‎Thakkinstian‎ and‎ colleagues(5) suggested that antibiotics appeared to be beneficial‎for‎patients‎with‎CP/CPPS‎and‎most‎appropri- ate‎for‎therapy‎of‎CP/CPPS. Nevertheless,‎as‎the‎diagnosis‎of‎category‎III‎prostatitis‎ demands‎ the‎exclusion‎of‎ infection,‎ the‎reason‎for‎ the‎ response‎associated‎with‎antibiotics‎is‎not‎immediately‎ clear.(6)‎ Some‎ people‎ suggested‎ that‎ there‎ was‎ a‎ poor‎ benefit‎after‎antibiotic‎therapy,(7)‎Nickel‎and‎colleagues(8)‎ found‎ that‎ the‎ levofloxacin‎ therapeutic‎ effect‎ in‎ men‎ diagnosed‎with‎CP‎was‎not‎significantly‎different‎from‎ placebo.‎ Then‎ DeRose‎ and‎ colleagues(9)‎ and‎ Kim‎ and‎ colleagues(10)‎also‎demonstrated‎that‎antibiotics‎did‎not‎ markedly‎reduce‎the‎symptoms‎in‎men‎with‎CPPS. Since‎the‎diagnostic‎criteria‎and‎treatment‎were‎not‎uni- fied,‎many‎early‎cases‎led‎to‎CP,‎and‎the‎overuse‎of‎anti- biotics caused bacterial resistance.(11) In order to strictly control‎ the‎ clinical‎ applying‎ indications‎ for‎ fluoroqui- nolone‎and‎strengthen‎management,‎our‎country‎execut- ed‎the‎Clinical‎Use‎of‎Antibiotics‎Guiding‎Principle‎in‎ 2004‎ and‎ the‎Clinical‎Use‎ of‎Antibiotics‎Management‎ Approach‎in‎August‎1,‎2012.‎Therefore,‎the‎purpose‎of‎ this‎study‎is‎to‎assess‎whether‎antibiotics‎are‎effective‎in‎ treating category III prostatitis by synthesizing the data from‎ all‎ related‎ available‎ randomized‎ controlled‎ trials‎ (RCTs). There‎ are‎ several‎ systematic‎ reviews‎ discuss‎ the‎ rela- tionship‎between‎therapeutic‎intervention‎and‎CP/CPPS.‎ Two‎studies(5,6)‎only‎included‎three‎RCTs‎to‎discuss‎the‎ relationship‎between‎antibiotics‎and‎CP/CPPS,‎one‎oth- ers(12)‎included‎two‎RCTs.‎Regarding‎our‎emphasis‎and‎ the‎ inconsistent‎ findings‎ on‎ the‎ relationship‎ between‎ antibiotics and category III prostatitis, we conducted an updated‎meta-analysis‎of‎RCTs‎on‎this‎subject.‎Our‎goal‎ was‎to‎determine‎whether‎antibiotics‎are‎associated‎with‎ the‎management‎of‎category‎III‎prostatitis. MATERIALS AND METHODS Literature Search We‎conducted‎a‎systematic‎literature‎search‎in‎the‎Pub- Med,‎ Medline‎ and‎ Embase‎ databases‎ to‎ identify‎ the‎ eligible‎studies‎before‎September‎1,‎2012.‎The‎follow- ing‎terms‎were‎used‎in‎the‎primary‎search:‎(randomized‎ controlled‎trial‎[pt]‎OR‎controlled‎clinical‎trial‎[pt]‎OR‎ randomized‎[tiab]‎OR‎placebo‎[tiab]‎OR‎clinical‎ trials‎ as‎topic‎[mesh:‎noexp]‎OR‎randomly‎[tiab]‎OR‎trial‎[ti])‎ NOT‎(animals‎[mh]‎NOT‎humans‎[mh])‎AND‎(prosta- titis)‎AND‎(antibiotics‎OR‎*xacin‎OR‎antibacterial‎OR‎ antimicro*).‎The‎search‎was‎focused‎on‎human‎studies,‎ without‎language‎restriction.‎In‎addition,‎we‎checked‎the‎ relevant‎review‎articles‎and‎their‎references‎to‎identify‎all‎ available‎literature‎that‎may‎not‎have‎been‎included‎in‎the‎ database‎results.‎The‎search‎following‎chat‎is‎presented‎ in‎Figure‎1.‎ Inclusion and Exclusion Criteria The‎study‎was‎included‎if‎it‎met‎the‎following‎criteria:‎ (1)‎it‎was‎an‎original‎RCT;‎(2)‎the‎disease‎has‎been‎clear- ly‎defined‎as‎category‎III‎prostatitis‎or‎CP/CPPS;‎(3)‎the‎ paper had a conclusion about the association between antibiotics‎and‎category‎III‎prostatitis‎and‎(4)‎the‎study‎ had‎provided‎enough‎information‎to‎estimate‎the‎effect‎ sizes.‎The‎exclusion‎criteria‎were:‎ (1)‎duplicate‎study;‎ (2)‎review‎paper;‎(3)‎systematic‎review;‎(4)‎abstract/title‎ only;‎(5)‎non‎category‎III‎prostatitis‎study;‎(6)‎other‎inter- ventions;‎(7)‎non‎comparative‎study‎and‎(8)‎non‎interest‎ outcome. Data Extraction Two‎reviewers‎independently‎extracted‎the‎information‎ from‎the‎eligible‎studies‎according‎to‎the‎inclusion‎and‎ 1379Vol. 11 | No. 02 | March- April 2014 |U R O LO G Y J O U R N A L Antibiotics and Category III Prostatitis | Zhu et al exclusion‎ criteria.‎ Disagreement‎ was‎ resolved‎ through‎ discussion.‎The‎collected‎data‎included‎first‎author,‎pub- lication‎year,‎study‎design,‎duration‎of‎therapy,‎age,‎inter- vention,‎sample‎size‎and‎the‎outcome‎data. Outcome Measures The‎ following‎ variables‎ were‎ examined:‎ chronic‎ pros- tatitis‎symptom‎index‎(CPSI)‎score,‎which‎include‎pain‎ score,‎voiding‎score,‎quality‎of‎life‎(QoL)‎score‎and‎total‎ score,‎at‎the‎baseline,‎at‎the‎end‎of‎study‎and‎the‎change‎ from‎the‎baseline‎to‎the‎end. Study Quality Evaluation Two‎ reviewers‎ graded‎ each‎ study‎ independently‎ using‎ the‎modified‎Jadad‎scale(13)‎(Table‎1).‎The‎score‎for‎each‎ article‎can‎range‎from‎0‎(lowest‎quality)‎ to‎8‎(highest‎ quality).‎Scores‎of‎4-8‎represent‎good‎to‎excellent‎(high‎ quality),‎whereas,‎0‎to‎3‎represent‎poor‎or‎low‎quality. Statistical Analysis All‎analyses‎were‎performed‎in‎Review‎Manager‎5‎statis- tical‎software.‎The‎continuous‎data‎were‎summarized‎as‎ the‎weighted‎mean‎differences‎(WMD)‎with‎the‎standard‎ deviations‎(SD).‎ If‎the‎study‎only‎reported‎the‎median,‎the‎range‎of‎con- tinuous‎data‎and‎the‎size‎of‎the‎trial,‎we‎used‎previous‎ formula(14) to‎translate‎these‎data‎to‎WMD‎and‎SD.‎When‎ it‎was‎desirable‎to‎combine‎two‎reported‎subgroup‎into‎a‎ single‎group,‎we‎used‎the‎formula‎reported‎in‎Table‎7.7.a‎ of‎the‎Cochrane‎Hand‎book‎5.0.2‎to‎combine‎them.‎If‎ there‎was‎a‎lack‎of‎WMD‎and‎SD‎of‎the‎changes‎from‎ baseline,‎while‎the‎baseline‎and‎final‎WMD‎and‎SD‎were‎ known,‎we‎imputed‎SD‎for‎ the‎changes‎from‎baseline‎ using‎the‎formula‎reported‎in‎16.1.3.2‎of‎the‎Cochrane‎ Handbook‎5.0.2.‎When‎considering‎the‎sensitivity‎analy- sis,‎the‎value‎of‎Corr‎was‎imputed‎as‎0.5.‎We‎used‎the‎I2‎ statistic to assess the statistical heterogeneity between the trials.‎If‎a‎heterogeneity‎of‎(I2‎>‎50%)‎existed‎we‎used‎ the‎random‎effect‎model‎to‎perform‎the‎meta-analysis.‎ Otherwise,‎the‎fixed‎effect‎model‎was‎used.‎The‎signifi- cance‎of‎the‎overall‎effect‎was‎tested‎with‎Fisher’s‎z-test,‎ P‎<‎.05‎was‎considered‎as‎a‎significant‎level.‎All‎results‎ representing‎the‎effect‎size‎were‎stated‎with‎95%‎con- fidence‎intervals‎(CI).‎The‎sensitivity‎analysis‎was‎per- formed‎by‎excluding‎low‎quality‎studies‎to‎assess‎if‎the‎ results‎were‎significant.‎If‎the‎score‎of‎trials‎were‎more‎ than‎5,‎the‎publication‎bias‎of‎the‎study‎was‎assessed‎by‎ a‎funnel‎plot. RESULTS Study Selection, Characteristics and Quality Table 1. The modified Jadad scale. Eight items Yes No Not Described Was the research described as randomized? 1 0 ----- Was the approach of randomization appropriate?* 1 -1 0 Was the research described as blinding? 1 0 ----- Was the approach of blinding appropriate? 1 -1 0 Was there a presentation of withdrawals and dropouts? 1 0 ----- Was there a presentation of the inclusion/exclusion criteria? 1 0 ----- Was the approach used to assess adverse effects described? 1 0 ----- Was the approach of statistical analysis described? 1 0 ----- * Double-blind got 1 score, single-blind got 0.5 score. 1380 | The‎study‎selection‎flow‎is‎described‎in‎Figure‎1.‎A‎total‎ of‎seven‎RCTs‎with‎539‎men(8-10,15-18) were included in our‎analysis.‎Of‎study‎subjects‎267‎were‎randomized‎to‎ an‎experimental‎group‎and‎the‎remaining‎272‎men‎were‎ assigned‎to‎a‎controlled‎group.‎One‎trial(15)‎enrolled pa- tients‎classified‎IIIA‎and‎IIIB,‎the‎remaining‎studies‎en- rolled‎patients‎with‎CPPS,‎all‎of‎ them‎were‎belong‎to‎ category‎III‎prostatitis.‎The‎mean‎age‎ranged‎from‎17‎to‎ 78‎years‎while‎this‎information‎was‎not‎provided‎in‎one‎ trial.(15)‎The‎mean‎duration‎of‎treatment‎ranged‎from‎6‎ weeks‎to‎12‎weeks.‎The‎intervention‎in‎the‎trials‎included‎ levofloxacin,(8,15,16)‎mepartricin,(10)‎ciprofloxacin(9,17) and tetracycline.(18)‎The‎management‎of‎control‎groups‎were‎ broadly‎classified‎into‎two‎methods:‎placebo(8,10,17,18)‎and others.(9,15-17)‎Two‎articles(15,17)‎had‎two‎sets‎of‎data,‎we‎ calculated‎them‎with‎divided‎and‎combined‎data,‎respec- tively.‎All‎ included‎ studies‎ had‎ the‎ high‎ quality‎ score‎ of‎the‎modified‎Jadad‎scale.‎The‎characteristics‎and‎the‎ quality‎of‎all‎included‎studies‎are‎presented‎in‎Table‎2. Meta-Analysis for the Change of Total Score of CPSI All‎the‎trials‎evaluated‎the‎effect‎of‎interventions‎on‎to- tal‎score‎of‎CPSI.‎Among‎these‎RCTs,‎three‎trials(8,10,18)‎ compared‎placebo‎with‎antibiotics,‎three‎trials(9,15,16)‎com- pared‎α-blocker‎with‎α-blocker‎plus‎antibiotics,‎and‎the‎ Review Article Table 2. Characteristics and quality of studies. Study Patient Intervention No. of Subjects Withdrawn Duration of Therapy Age Modified Jadad Scale Nickel et al., 2003(8) CP/CPPS Levofloxacin 45 12 weeks 56.0 (39-77) 8 Placebo 35 56.2 (36-78) DeRose et al., 2004(10) CP/CPPS Mepartricin 15 60 days 34.75 ± 6.69 5 Placebo 15 36.5 ± 7.54 Alexander et al., 2004(17) CP/CPPS Placebo 45 (4) 6 weeks 42.6 ± 12.0 8 Ciprofloxacin 42 (7) 45.9 ± 11.7 Tamsulosin 45 (4) 45.3 ± 9.7 Ciprofloxacin + Tamsulosin 42 (7) 44.5 ± 11.4 Jeong et al., 2008(16) CP/CPPS Doxazosin 26 6 weeks 41.5 (23-56) 4 Levofloxacin + Doxazosin 29 41.1 (27-60) Ye et al., 2008(15) IIIA Tamsulosin 21 45 days No mentioned 5 Tamsulosin + Levofloxacin 21 IIIB Tamsulosin 21 Tamsulosin + Levofloxacin 21 Zhou et al., 2008(18) CPPS Tetracycline 24 12 weeks 29-50 (all) 4 Placebo 24 Kim et al., 2011(9) CP/CPPS Tamsulosin 40 12 weeks 45.7 5 Tamsulosin + Ciprofloxacin 28 46.1 Keys: CP, chronic prostatitis; CPPS, chronic pelvic pain syndrome. 1381Vol. 11 | No. 02 | March- April 2014 |U R O LO G Y J O U R N A L other one study(17)‎included‎all‎the‎management‎methods.‎ One‎trial(17) got‎the‎change‎from‎the‎baseline‎directly;‎the‎ remaining‎trials‎reported‎the‎mean‎scores‎at‎follow-up‎ and‎got‎the‎change‎via‎the‎formula.‎ Our‎quantitative‎accumulation‎analysis‎with‎the‎random- effect‎model‎(I2‎>‎75%)‎revealed‎that‎patients‎using‎an- tibiotics‎had‎a‎greater‎reduce‎in‎total‎score‎from‎baseline‎ when‎compared‎with‎the‎control‎group‎(P‎=‎.01;‎Figure‎ 2A).‎However,‎ the‎subgroup‎analysis‎showed‎ that‎ this‎ difference‎was‎not‎statistically‎significant‎both‎in‎the‎an- tibiotics‎group‎versus‎the‎placebo‎group‎(P =‎.05)‎and‎in‎ the‎α-blocker‎group‎versus‎the‎α-blocker‎plus‎antibiotics‎ group (P‎=‎.13).‎The‎sensitivity‎analysis‎was‎not‎done‎ necessarily‎because‎of‎the‎median‎or‎the‎high‎quality‎of‎ all‎the‎seven‎studies.‎There‎was‎a‎potential‎publication‎ bias‎in‎our‎analysis‎according‎to‎the‎funnel‎plot‎presented‎ in‎Figure‎3. Considering‎a‎longer‎treatment‎interval‎may‎have‎a‎more‎ positive‎role‎in‎the‎effect‎of‎the‎symptom‎scores‎improv- ing,‎we‎conducted‎a‎subgroup‎analysis‎(Table‎3)‎based‎on‎ the‎treatment‎duration‎by‎12‎weeks,‎which‎was‎chosen‎as‎ the‎dividing‎line.‎The‎analysis‎result‎based‎on‎the‎eligi- bility‎data‎showed‎that‎antibiotics‎were‎not‎beneficial‎in‎ the‎management‎of‎category‎III‎prostatitis‎when‎the‎treat- ment‎duration‎was‎more‎than‎12‎weeks.‎ Meta-Analysis for the Change of Pain Score, Voiding Score and QoL Score of CPSI Excluded‎the‎study‎by‎Zhou,‎2008‎as‎the‎data‎were‎in- complete,‎there‎were‎six‎trials(8-10,15-17) (491‎men)‎which‎ had‎ evaluated‎ the‎ effect‎ of‎ the‎ interventions‎ on‎ pain‎ score,‎voiding‎score‎and‎QoL‎score‎of‎CPSI. With‎a‎random-effect‎model‎(I2‎>‎50%),‎the‎pooled‎anal- ysis‎revealed‎that‎patients‎using‎antibiotics‎had‎a‎greater‎ reduce‎in‎pain‎score‎when‎compared‎to‎the‎control‎group‎ (P‎=‎.02;‎Figure‎2B).‎The‎subgroup‎analyses‎showed‎the‎ same‎difference‎when‎comparing‎placebo‎to‎antibiotics‎ (P‎=‎.04),‎however,‎the‎difference‎between‎α-blocker‎and‎ α-blocker‎plus‎antibiotics‎was‎not‎statistically‎significant‎ (P‎=‎.16). We‎ used‎ a‎ random-effect‎ model‎ to‎ estimate‎ the‎ void- ing‎score‎because‎of‎the‎huge‎heterogeneity‎between‎the‎ studies‎(I2‎>‎75%).‎A‎quantitative‎accumulation‎revealed‎ Antibiotics and Category III Prostatitis | Zhu et al Table 3. Subgrouping based on the treatment duration. Variables No. of Studies Antibiotics/Control WMD 95% CI P I2 < 12 weeks Antibiotics vs. placebo 2 57/60 -4.18 -6.55 -1.81 .00 81% Antibiotics + α-blocker vs. α-blocker 2 84/87 -5.34 -7.06 -3.62 .00 97% >12 weeks Antibiotics vs. placebo 1 45/35 -2.50 -6.48 -1.48 .22 NA Antibiotics + α-blocker vs. α-blocker 1 28/40 -0.07 -2.47 -2.33 .95 NA Keys: CI, confidence interval; NA, not applicable; WMD, weighted mean differences. 1382 | Review Article that‎patients‎using‎antibiotics‎had‎no‎significantly‎greater‎ reduce‎in‎voiding‎score‎when‎compared‎to‎the‎control‎ group (P‎=‎.10;‎Figure‎2C).‎The‎results‎of‎the‎subgroup‎ analyses‎were‎the‎same‎both‎in‎placebo‎versus‎antibiotics‎ (P =‎.08)‎and‎in‎α-blocker‎versus‎α-blocker‎plus‎antibiot- ics (P‎=‎.21).‎ With‎a‎potential‎heterogeneity‎(I2‎>‎50%),‎the‎random- effect‎model‎analysis‎revealed‎that‎patients‎using‎antibi- otics‎had‎a‎significantly‎greater‎reduce‎in‎QoL‎score‎when‎ compared‎to‎the‎control‎group‎(Figure‎2D).‎The‎subgroup‎ analyses‎showed‎that‎the‎difference‎was‎also‎exist‎when‎ comparing‎α-blocker‎to‎α-blocker‎plus‎antibiotics‎(P = .04),‎whereas‎there‎was‎no‎significant‎difference‎in‎the‎ subgroup‎of‎placebo‎versus‎antibiotics‎(P‎=‎.10). DISCUSSION Category‎ III‎ prostatitis‎ is‎ the‎ most‎ common‎ urologic‎ diseases.‎ The‎ medicine‎ treatment‎ contains‎ antibiotics,‎ α-blockers,‎anti-inflammatory‎analgesics,‎and‎so‎on.‎In‎ our‎study,‎there‎were‎four‎interventions‎in‎the‎treatment‎ of‎category‎III‎prostatitis:‎placebo,‎antibiotics,‎α-blockers‎ and‎ α-blockers‎ plus‎ antibiotics.‎ In‎ the‎ same‎ baseline‎ level‎of‎CPSI,‎ the‎difference‎of‎ the‎score‎reduction‎in‎ the‎α-blockers‎group‎and‎the‎α-blockers‎plus‎antibiotics‎ group‎reflected‎the‎role‎of‎antibiotics‎in‎the‎treatment,‎ although‎α-blockers‎also‎play‎a‎role.‎So‎these‎four‎inter- ventions‎were‎divided‎into‎two‎subgroups‎in‎our‎meta- analysis. In‎this‎meta-analysis,‎the‎quantitative‎accumulation‎re- sults‎showed‎that‎antibiotics‎had‎a‎significantly‎greater‎ role‎in‎the‎reduction‎in‎total‎score,‎pain‎score‎and‎QoL‎ score,‎were‎same‎with‎the‎former‎recommendation‎that‎ antibiotics‎are‎useful‎in‎category‎III‎prostatitis.(8) How- ever,‎the‎analysis‎of‎the‎subgroup‎showed‎the‎opposite‎ result;‎ that‎ summary‎ association‎ between‎ category‎ III‎ prostatitis‎ and‎ antibiotics‎ were‎ not‎ statistically‎ signifi- cant,‎ especially‎ in‎ total‎ score‎ and‎ voiding‎ score.‎ Our‎ findings‎ revealed‎ that‎ antibiotics‎ are‎ not‎ beneficial‎ in‎ the‎management‎of‎category‎III‎prostatitis.‎In‎part‎of‎the‎ sub-score‎analysis,‎antibiotics‎had‎a‎significantly‎greater‎ reduce‎in‎pain‎and‎QoL‎score,‎but‎not‎in‎voiding‎score.‎ Such‎result‎met‎the‎point‎of‎view‎that‎the‎antibiotics‎had‎ anti-inflammatory‎and‎analgesic‎properties.(19) Antibiot- ics,‎especially‎the‎fluoroquinolones,‎had‎been‎proven‎to‎ influence‎the‎cytokine‎activity.‎For‎example,‎levofloxa- cin‎had‎an‎immunomodulatory‎function‎on‎the‎cytokine‎ production‎not‎relying‎on‎the‎antimicrobial‎activity;(20)‎ cotrimoxazole‎was‎prescribed‎for‎the‎anti-inflammatory‎ or‎ immunosuppressive‎effects‎on‎ the‎noninfectious‎ ill- ness,(21)‎and‎so‎on.‎The‎subgroup‎analysis‎based‎on‎the‎ treatment‎duration‎showed‎that‎antibiotics‎were‎not‎ben- eficial‎in‎the‎management‎of‎category‎III‎prostatitis‎when‎ the‎ treatment‎ duration‎ was‎ more‎ than‎ 12‎ weeks,‎ even‎ where‎ beneficial‎ when‎ less‎ than‎ 12‎ weeks.‎The‎ short- term‎ curative‎ effect‎ is‎ actually‎ the‎ analgesic‎ effect‎ of‎ antibiotics,‎which‎caused‎by‎patients’‎subjective‎feeling.‎ In‎fact,‎the‎long-term‎curative‎effect‎showed‎that‎the‎anti- inflammatory‎effects‎of‎antibiotics‎were‎not‎obvious‎in‎ prostate category III, which was the reason that antibiot- ics‎could‎not‎improve‎category‎III‎prostatitis.‎ In‎our‎data,‎several‎issues‎warrant‎further‎discussion.‎The‎ etiology‎for‎category‎III‎prostatitis‎has‎not‎been‎fully‎elu- cidated‎and‎the‎criteria‎used‎for‎classifying‎the‎treatment‎ response‎were‎varied,‎so‎we‎did‎not‎analyze‎the‎treatment‎ responsiveness.‎Compared‎with‎the‎systematic‎reviews‎ published‎by‎Thakkinstian‎and‎colleagues‎and‎Cohen‎and‎ colleagues,(5,12)‎we‎included‎the‎latest‎seven‎randomized‎ Figure 1. Study selection strategy. 1383Vol. 11 | No. 02 | March- April 2014 |U R O LO G Y J O U R N A L Antibiotics and Category III Prostatitis | Zhu et al controlled‎trials,‎which‎have‎the‎highest‎quality‎in‎pub- lished literature. Although the search strategy was with- out‎restriction‎on‎language,‎some‎literature‎may‎be‎omit- ted‎because‎of‎the‎limitation‎of‎the‎internet.‎Meanwhile,‎ as‎we‎know,‎the‎grey‎literature‎was‎difficult‎to‎obtain. In addition, we included the trials representing the op- posite results, so a high heterogeneity was detected in the studies‎and‎forced‎us‎to‎apply‎the‎random‎effect‎model‎ reducing‎the‎credibility‎and‎increasing‎the‎imprecision‎of‎ the‎results.‎As‎we‎know,‎it‎was‎questionable‎that‎pooled‎ the‎data‎by‎a‎meta-analysis‎when‎the‎heterogeneity‎was‎ too‎high‎(I2‎>‎75%),‎and‎its‎effect‎would‎not‎be‎overcome‎ by‎the‎random‎effect‎model.‎However,‎according‎to‎some‎ authoritative‎literatures,(22,23) we‎also‎used‎this‎method‎to‎ Figure 2. Forest plot of change in (A) total score, (B) pain score, (C) voiding score and (D) Qol score of National Institutes of Health Chronic Prostatitis Symptom Index. 1384 | evaluate‎the‎results,‎even‎though‎it‎could‎not‎avoid‎the‎ huge‎heterogeneity.‎Considering‎a‎high‎heterogeneity‎and‎ a‎publication‎bias‎existed,‎we‎found‎several‎sources‎in‎ these‎RCTs.‎First,‎the‎patients‎chosen‎to‎study‎were‎dif- ferent.‎The‎men‎with‎refractory‎long-standing‎symptoms‎ represented‎a‎small‎subpopulation‎of‎the‎overall‎group‎ with‎CP/CPPS.‎Second,‎the‎agents‎might‎be‎more‎effec- tive‎ in‎men‎who‎had‎received‎less‎previous‎ treatment.‎ The‎ideal‎study‎should‎involve‎patients‎who‎were‎naïve‎ to‎the‎antimicrobial‎therapy.‎Third,‎the‎duration‎of‎treat- ment‎was‎different.‎Some‎studies‎did‎not‎test‎the‎use‎of‎ drugs‎for‎longer‎than‎6‎weeks‎but‎longer‎treatment‎may‎ be‎warranted.‎Fourth,‎the‎combination‎therapy‎was‎dif- ferent‎from‎the‎mono-therapy.‎Fifth,‎the‎antibiotic‎ther- apy‎for‎category‎IIIA‎was‎justified,‎but‎not‎for‎category‎ IIIB‎in‎some‎trials;‎others‎found‎no‎significant‎differenc- es‎between‎categories‎II,‎IIIA,‎or‎IIIB‎to‎the‎antibiotic‎ treatment.‎Sixth,‎the‎dose‎of‎antibiotics‎was‎change‎from‎ 100‎mg‎twice‎daily‎to‎500‎mg‎daily‎in‎different‎countries.‎ Seventh,‎the‎revisiting‎time‎for‎treatment‎ranged‎from‎3‎ months‎to‎1‎year‎follow-up‎period.‎Lastly,‎some‎patients‎ were‎wrongly‎diagnosed.‎Although‎the‎combining‎esti- mates‎were‎greatly‎heterogeneous,‎the‎mixed‎model‎with‎ random‎intercept‎gave‎consideration‎to‎variations‎at‎the‎ study‎level.‎What’s‎more,‎the‎measurement‎that‎we‎used‎ was‎the‎score‎reduced‎difference‎instead‎of‎the‎score‎of‎ CPSI‎directly,‎which‎may‎reflect‎the‎role‎of‎antibiotics‎ better‎than‎the‎direct‎CPSI‎score.‎ There‎were‎some‎limitations‎that‎needed‎to‎be‎taken‎into‎ account.‎The‎number‎of‎patients‎enrolled‎was‎small‎and‎ the‎total‎sample‎sizes‎were‎relatively‎small,‎so‎the‎cred- ibility‎of‎the‎conclusion‎was‎not‎strong‎enough‎and‎the‎ representative‎required‎considering.‎The‎sample‎sizes‎of‎ the‎studies‎were‎so‎different‎that‎the‎weight‎was‎not‎the‎ same,‎which‎led‎to‎a‎high‎heterogeneity‎after‎ the‎data‎ combination.‎The‎incorporative‎results‎were‎often‎heter- ogeneous‎and‎the‎origin‎of‎this‎difference‎was‎not‎obvi- ous.‎Category‎III‎prostatitis‎remains‎a‎disputed‎condition‎ with‎little‎consensus‎regarding‎the‎best‎treatment‎option. (1)‎The‎treatment‎benefits‎were‎modest‎for‎some‎therapies‎ and‎nonexistent‎for‎others,‎which‎probably‎reflected‎the‎ individual‎differences.‎ CONCLUSION Although‎it‎is‎commonly‎known‎that‎there‎is‎a‎great‎ben- efit‎from‎antibiotics‎for‎category‎III‎prostatitis,‎we‎found‎ no‎significant‎associations‎between‎them‎when‎analyz- ing‎ the‎ published‎ studies‎ by‎ meta-analysis.‎ Our‎ meta- analysis‎reveals‎that‎antibiotics‎are‎not‎beneficial‎in‎the‎ management‎of‎category‎III‎prostatitis.‎Future‎research‎to‎ confirm‎these‎findings‎is‎warranted,‎and‎we‎may‎reduce‎ the‎usage‎of‎antibiotics‎in‎such‎a‎population. ACKNOWLEDGMENT Yongtong‎Zhu,‎Chunyan‎Wang‎and‎Xiang‎Pang‎contrib- uted‎equally‎to‎this‎work. CONFLICT OF INTEREST None declared. REFERENCES 1. Lee KS, Choi JD. Chronic prostatitis: approaches for best manage- ment. Korean J Urol. 2012;53:69-77. 2. Guo YL, Li HJ. Prostatitis. Beijing, China, People's Military Medical Press; 2007. p. 63-8. 3. Schaeffer AJ. Classification (traditional and National Institutes of Health) and demographics of prostatitis. Urology. 2002;60:5-6. 4. Kim SH, Ha US, Sohn DW, et al. Preventive effect of ginsenoid on chronic bacterial prostatitis. J Infect Chemother. 2012;18:709-14. Review Article Figure 3. Funnel plot of change in total score of National Insti- tutes of Health Chronic Prostatitis Symptom Index. 1385Vol. 11 | No. 02 | March- April 2014 |U R O LO G Y J O U R N A L 5. Thakkinstian A, Attia J, Anothaisintawee T, Nickel JC. alpha-block- ers, antibiotics and anti-inflammatories have a role in the manage- ment of chronic prostatitis/chronic pelvic pain syndrome. BJU Int. 2012;110:1014-22. 6. Anothaisintawee T, Attia J, Nickel JC, et al. Management of chronic prostatitis/chronic pelvic pain syndrome: a systematic review and network meta-analysis. JAMA. 2011;305:78-86. 7. Lombardo F, Fiducia M, Lunghi R, et al. Effects of a dietary supple- ment on chronic pelvic pain syndrome (Category IIIA), leucocyto- spermia and semen parameters. Andro. 2012;44 Suppl 1:672-8. 8. Nickel JC, Downey J, Clark J, et al. Levofloxacin for chronic prosta- titis/chronic pelvic pain syndrome in men: a randomized placebo- controlled multicenter trial. Urology. 2003;62:614-7. 9. Kim TH, Lee KS, Kim JH, et al. Tamsulosin Monotherapy versus Com- bination Therapy with Antibiotics or Anti-Inflammatory Agents in the Treatment of Chronic Pelvic Pain Syndrome. Int Neurourol J. 2011;15:92-6. 10. De Rose AF, Gallo F, Giglio M, Carmignani G. Role of mepartricin in category III chronic nonbacterial prostatitis/chronic pelvic pain syndrome: a randomized prospective placebo-controlled trial. Urol- ogy. 2004;63:13-6. 11. Wang WF, Dong DX, Cen S, et al. Preliminary Study of Incurable Causes of Chronic Prostatitis. Nat J Andro. 2001;7:233-6. 12. Cohen JM, Fagin AP, Hariton E, et al. Therapeutic Intervention for Chronic Prostatitis/Chronic Pelvic Pain Syndrome (CP/CPPS): A Sys- tematic Review and Meta-Analysis. PLoS One. 2012;7:e41941. 13. Dong J, Su SY, Wang MY, Zhan Z. Shenqi fuzheng, an injection con- cocted from Chinese medicinal herbs, combined with platinum- based chemotherapy for advanced non-small cell lung cancer: a systematic review. J Exp Clin Cancer Res. 2010;29:137. 14. Hozo SP, Djulbegovic B, Hozo I. Estimating the mean and variance from the median, range, and the size of a sample. BMC Med Res Methodol. 2005;5:13. 15. Ye ZQ, Lan RZ, Yang WM, Yao LF, Yu X. Tamsulosin treatment of chronic non-bacterial prostatitis. J Int Med Res. 2008;36:244-52. 16. Jeong CW, Lim DJ, Son H, Lee SE, Jeong H. Treatment for chronic prostatitis/chronic pelvic pain syndrome: levofloxacin, doxazosin and their combination. Urol Int. 2008;80:157-61. 17. Alexander RB, Propert KJ, Schaeffer AJ, et al. Ciprofloxacin or tamsu- losin in men with chronic prostatitis/chronic pelvic pain syndrome: a randomized, double-blind trial. Ann Intern Med. 2004;141:581-9. 18. Zhou Z, Hong L, Shen X, et al. Detection of nanobacteria infection in type III prostatitis. Urology. 2008;71:1091-5. 19. Nickel JC, Downey J, Johnston B, Clark J. Predictors of patient re- sponse to antibiotic therapy for the chronic prostatitis/chronic pelvic pain syndrome: a prospective multicenter clinical trial. J Urol. 2001;165:1539-44. 20. Yoshimura T, Kurita C, Usami E, et al. Immunomodulatory action of levofloxacin on cytokine production by human peripheral blood mononuclear cells. Chemotherapy. 1996;42:459-64. 21. Roblot P, Becq-Giraudon B. [Non-antibiotic effects of antibiot- ics: from side effects to therapeutic uses]. Pathol Biol (Paris). 1997;45:751-7. Antibiotics and Category III Prostatitis | Zhu et al 22. Aboumarzouk OM, Stein RJ, Eyraud R, et al. Robotic versus laparo- scopic partial nephrectomy: a systematic review and meta-analysis. Eur Urol. 2012;62:1023-33. 23. Wu LM, Hu J, Gu HY, Hua J, Xu JR. Can diffusion-weighted magnetic resonance imaging (DW-MRI) alone be used as a reliable sequence for the preoperative detection and characterisation of hepatic me- tastases? A meta-analysis. Eur J Cancer. 2013;49:572-84.