Abstract
Purpose
We sought to examine the literature reporting the effect of urinary tract infection (UTI) on non-schistosomiasis-related UBC (UBCNS) through a systematic review and meta-analysis.
Methods
A predefined study protocol was developed according to PRISMA. Medline and Scopus were searched for all studies investigating exposure to UTI with UBCNS as the primary outcome. Potential studies were screened against eligibility criteria. Clinical heterogeneity was assessed and groups with more than two studies were evaluated by random effect meta-analysis. Study-level bias was assessed with the Newcastle–Ottawa Scale (NOS). In cases of substantial between study heterogeneity (I2 > 50%), predefined sensitivity and subgroup analyses were performed.
Results
Of 16 eligible studies, eight case–control studies spanning four decades and five countries were suitable for quantitative analysis. Main analysis favored exposure to UTI increasing risk of subsequent UBCNS (RR 1.33 [95% CI 1.14–1.55]). This effect was no longer statistically significant after excluding studies published prior to year 2000 and at high risk of bias. Between study heterogeneity was considerable for nearly all analyses and not reduced by predefined sensitivity or subgroup analyses.
Conclusion
Exposure to UTI favors increased risk for UBCNS, particularly in men, but these effects were statistically insignificant when pooling data from the most recent and highest quality studies. These data do not support findings of previously published studies, that report on heterogenous populations with poor definitions of UTI and minimal control for important confounders. Results from previous studies should be viewed as hypothesis generating. This review highlights the need for higher quality investigation.
Similar content being viewed by others
Avoid common mistakes on your manuscript.
Introduction
A diverse body of literature has shown associations between urinary tract infection (UTI) and the development of urinary bladder cancer (UBC). The most definitive link established between UTI and UBC consists of the association between Schistosoma haematobium, a parasitic worm that causes urogenital schistosomiasis and squamous cell carcinoma (SCC) of the bladder [1]. Although exact pathogenesis is not well defined, egg deposition in the bladder wall is thought to be the major contributor to increased cancer risk.
A link between bacterial UTI and non-schistosomiasis-related UBC (UBCNS), particularly urothelial carcinoma (UC), is less clear. Numerous case–control studies constructed over the last several decades have fairly consistently shown exposure to bacterial UTI is associated with an increased risk of bladder cancer. The concept that chronic bacterial UTI can lead to bladder cancer has permeated urology textbooks. However, these studies were observational and designed to show association, not causation. Furthermore, these studies include weak definitions of UTI and poorly control for confounding variables and biases, including diagnostic bias (i.e., a patient with irritative voiding symptoms may be incorrectly diagnosed with a UTI when the symptoms are caused by undiagnosed UBC).
We sought to clarify the potential connection between UTI and UBCNS through a systematic review and meta-analysis. Our primary objective was to define if exposure to UTI affects the risk for future UBCNS. Evaluation of the influence of lifestyle factors and comorbidities on the association between UTI exposure and UBCNS risk was a secondary objective. At the time of this writing, a meta-analysis of this scope has not been published. One recent publication did not follow established criteria for a systematic review and findings were reported in a narrative format [2].
Methods
Protocol and registration
This review was performed in accordance with Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) [3]. A full protocol for this systematic review and meta-analysis was developed according to PRISMA for systematic review protocols (PRISMA-P) guidelines [4]. It was registered with the International Prospective Register of Systematic Reviews (PROSPERO) [5]. The protocol was developed after preliminary literature searches and piloting of the study selection process but before formal screening and data extraction.
Study eligibility criteria
A preliminary literature search did not reveal prospective studies relevant to the topic. We considered all observational studies, including cohort and case–control studies, which reported histologically confirmed UBC as the primary outcome and UTI as a primary or separate exposure variable. We excluded studies that did not confirm UBC by histology or reported SCC as the primary UBC variant (i.e., to exclude, as best as possible, cases that may be caused by schistosomiasis). Likewise, we excluded studies with participants and/or comparators with neurological conditions (e.g., spinal cord injury, as bladder catheterization has been linked to SCC of the bladder) or located in areas where schistosomiasis is endemic.
Identification and selection of studies
In January 2017, Medline and Scopus were searched for potentially eligible studies using a predefined strategy with medical subject headings and keywords. The full search strategy for both databases is available through our PROSPERO registration [5]. All identified studies were screened by title and abstract (e.g., “first-level” assessment) for further review. Eligibility criteria were applied to full-text articles (e.g., “second-level” assessment) using a predefined worksheet. In cases of duplicate studies or studies on duplicate populations, the study presenting the most recent data was assessed (unless there was a contraindication to doing so). References of all full-text articles evaluated as part of the second-level assessment were screened to ensure literature saturation. All literature searches and reviews were independently performed by two unblinded reviewers (CB, DF) before developing consensus lists. Disagreements were resolved by a third reviewer (MH).
Data collection and assessing risk
For studies meeting eligibility criteria, data were extracted using a predefined worksheet. For our primary outcome of UBC and exposure history of UTI, we extracted total raw numerator and denominator data for controls and comparators and, when possible, excluded data ≤ 2 years preceding UBC diagnosis for cases and interview for comparators to control for diagnostic bias. Studies were excluded for quantitative analysis if raw data were not presented in the published article. When possible, we extracted secondary outcome data related to UTI exposure subgroups, such as gender, smoking status, and recurrent UTI exposure. We did not collect adjusted outcome statistics given that the methods and variables were inconsistent across studies.
We developed a grading system (1–5) to categorize study definitions of UTI exposure in which higher grades indicate higher quality definitions of UTI (Supplementary Table 1). When defined by individual studies, only data related to exposure to lower tract UTI (e.g., cystitis) were extracted.
We used the Newcastle–Ottawa Scale (NOS), a validated tool for assessing risk of bias in nonrandomized clinical studies, to assess study-level bias [6]. NOS uses a “star system” to grade studies on selection of study groups (up to 4 stars), comparability of the groups (up to 2 stars), and ascertainment of exposure or outcome (up to 3 stars). Higher total star ratings equate to higher study quality and less risk of bias. Two reviewers (CB, DF) independently assessed all eligible studies according to NOS scales. If consensus was not reached, a third reviewer (MH) served as arbiter.
Data synthesis
Clinical heterogeneity was assessed by grouping studies according to design, UTI definition, and UTI subgroups. For groups with two or more studies, data were pooled using Mantel–Haenszel random effect meta-analysis. Risk ratios (RR) with 95% confidence intervals (CI) were calculated as they are more intuitive to understand than odds ratio (OR). CIs that did not cross RR 1.0 were considered statistically significant.
Between study heterogeneity was assessed visually using forest plots and statistically using the I2 statistic. No threshold of heterogeneity excluded pooled analysis, but substantial heterogeneity (considered I2 > 50%) was further investigated through predefined, logical sensitivity analyses excluding those studies investigating hospital-based sample populations (under the assumption hospitalized patients are more likely to have cancer and/or UTI than the general population), published prior to year 2000 (to exclude less sophisticated and/or rigorous methods of investigation), and with highest risk of bias (NOS total ≤ 6). We also performed sensitivity analyses excluding either European or U.S. studies (as these populations may be different in racial makeup as well as social, cultural, and environmental factors). Additional subgroup analyses separated data by gender, smoking status, and multiple UTI status. As all pooled analyses included < 10 studies, we assessed the potential for small study effects only visually using a funnel plot [7, 8].
All data were maintained in and meta-analyses were performed with Review Manager (RevMan) 5 [9].
Results
Study selection
The study selection process is shown in Supplementary Figure 1. Full text for all 21 studies passing the first-level assessment was available for review. Of these, five were excluded based on eligibility criteria. A total of 16 studies met eligibility criteria and were considered for quantitative synthesis [10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25]. All studies were retrospective in nature, and all but one was of case–control design. Of the 16, only 8 studies were suitable for pooled analysis [10,11,12,13,14,15,16,17]. Table 1 identifies details for all 16 full-text articles meeting eligibility criteria.
Eligible studies
The eight studies considered suitable for pooled analysis were published over a 30-year period from a range of countries. All were designed as case–control studies with the majority selecting community-based case and control populations. The studies included a range of sample sizes for case (n = 170–2932) and control populations (n = 282–5698). All studies utilized the lowest grade of our predefined UTI definitions. Vermeulen et al. were the only investigators to collect data for UTI definitions 1 and 2, thus only data pertaining to definition 1 were synthesized in the overall pooled analysis. According to the NOS, only four (50%) studies, all published in the last 11 years, were judged to be at low risk of bias (NOS total > 6).
Main analysis: UTI exposure and risk of UBCNS
The pooled overall effect favored a statistically significant increased risk of UBCNS following exposure to UTI (RR 1.33 [95% CI 1.14–1.55, eight studies]; Fig. 1). Only one study (Jiang et al.) reported exposure to UTI reduced the risk of UBCNS. Between study heterogeneity was considerable (I2 = 87%). A funnel plot (Supplementary Figure 2) did not reveal an asymmetric distribution, making it unlikely the overall pooled effect was strongly impacted by small study biases. The absence of studies along the lower left-side of the graph prevents us from excluding publication bias.
To investigate the substantial between study heterogeneity, predefined sensitivity analyses were performed (Fig. 2a–e). Between study heterogeneity remained considerable in all analyses (I2 ≥ 79%). After excluding studies published prior to year 2000 and those with the highest risk of bias (NOS total ≤ 6), pooled effects no longer met criteria for statistical significance as the CIs crossed 1.0 (RR 1.21 [CI 0.88, 1.68, four studies] and RR 1.15 [95% CI 0.83, 1.59, four studies], respectively; Fig. 2d, e) and with considerable between study heterogeneity (I2 = 91%).
Subanalyses: UTI exposure subgroups and risk of UBCNS
Pooled analysis was performed for raw data available for five subgroups of UTI exposure: males, females, exposure to UTI with smoking history, exposure to UTI without smoking history, and exposure to multiple UTIs.
Five (62.5%) studies reported raw data on UTI exposure by gender (Supplementary Figures 2A–B). Overall effect of UTI exposure on UBCNS risk increased when considering only males (RR 1.67 [95% 1.14, 2.45, five studies]; Supplementary Figure 3A) but was statistically insignificant when pooling only female data (RR 1.27 [95% CI 0.96, 1.69, five studies]; Supplementary Figure 3B). Both analyses demonstrated considerable between study heterogeneity (I2 = 95 and 94%). The effect for only males was no longer significant after excluding studies published prior to year 2000 and at highest risk of bias (RR 2.59 [95% 0.3, 22.3, two studies]; data not shown). Predefined sensitivity analyses did not sizably reduce between study heterogeneity for either subgroup comparison (data not shown).
Two (25%) studies provided data permitting subgrouping by smoking status, though Jiang et al. only provided data regarding female subjects (Supplementary Figure 4A–B). Pooling data from subjects with (RR 1.03 [95% CI 0.34, 3.06, two studies]; Supplementary Figure 4A) and without smoking history (RR 1.05 [95% CI 0.35, 3.11, two studies]; Supplementary Figure 4B) revealed no effect given the broad CIs of both analyses. Between study heterogeneity was considerable (I2 = 98 and 92%) for both analyses.
Five (62.5%) studies reported data on multiple UTI exposures (Supplementary Figures 5A–B). Each study categorized frequency of UTI differently. For purposes of comparison, data from Kantor et al. and Vermeulen et al. were extracted as ≥ 3 episodes of UTI compared to no exposure. Data from the remaining three studies were extracted as ≥ 4 episodes of UTI compared to no exposure. Notably, in collecting data regarding the number of UTI episodes, Vermeulen et al. changed their definition of UTI, eliciting the number to times subjects were treated with an antibiotic (corresponding to UTI definition 2). Jhamb et al. excluded data within 3 years of cancer diagnosis. Jiang et al. reported episodic UTI data for females only and excluded data within 5 years of cancer diagnosis. Pooled analysis showed no association between recurrent UTI exposure and risk of UBCNS (RR 1.03 [95% CI 0.47, 2.28, five studies]; Supplementary Figure 5A) with considerable between study heterogeneity (I2 = 98%). Predefined sensitivity analyses did not sizably alter these results until studies with published prior to year 2000 and highest risk of bias were eliminated (for this analysis, Kantor et al. and La Vecchia et al. met both criteria). Under these criteria, exposure to multiple UTIs favored a statistically significant reduction in risk of UBCNS (RR 0.56 [95% CI 0.38, 0.81, three studies]; Supplementary Figure 5B) but between study heterogeneity remained substantial (I2 = 63%).
Discussion
We performed a systematic review and meta-analysis to evaluate the effect of UTI on the risk of developing UBCNS. Eight case–control studies suitable for pooled analysis, according to predefined criteria, revealed an overall effect favoring exposure to UTI increasing the risk of UBCNS (RR 1.33 [95% CI 1.14–1.55, eight studies]). The statistical significance of this effect was lost after excluding European studies, studies published prior to year 2000, and studies at the highest risk of bias. There may be a gender-specific increased risk for males (RR 1.67 [95% 1.14, 2.45, five studies]), yet this effect was statistically insignificant after excluding studies published prior to year 2000 and at highest risk for bias. Female-only data showed no statistically significant effect. Pooling data from subjects with and without smoking history showed no effect. Exposure to multiple UTIs was not associated with UBCNS risk until studies published prior to year 2000 and those with a high risk of bias were excluded. Under these conditions, analysis favored a decreased risk of UBCNS (RR 0.56 [95% CI 0.38, 0.81, three studies]) with substantial between study heterogeneity (I2 = 63%). For all analyses, between study heterogeneity was considerable and was not explained by logical exclusion of studies with different sample populations (e.g., hospital versus community, European versus American), early publication dates, or high risk of bias.
The main effect from pooled case–control studies corroborates data from the only cohort study we identified in our systematic search [25]. Sun et al. reviewed 9 years of reimbursement data in Taiwan and reported those diagnosed with UTI experienced a significantly increased risk of urinary tract cancer during follow-up compared to a non-UTI exposed group (hazard ratio [HR] 4.66 [95% CI 3.55–6.1]). Results grouped by UTI and cancer location reinforce an association between lower UTIs and future bladder cancer (HR 5.68 [95% CI 3.91, 8.25]). A notable limitation of this study is the Taiwanese database did not identify cancer histology.
The finding that increased frequency of UTI exposure may decrease the risk of UBCNS is interesting and warrants discussion. First, it argues against the common criticism of diagnostic bias in research on UTI exposure and UBC risk. Second, it does not necessarily follow the intuitive extrapolation of our main pooled analysis. If any exposure to UTI is associated with an increased risk for UBCNS, why would repeated exposure decrease oncogenic potential? Some [15, 17] have theorized that repeated antibiotic exposure may exhibit a dose-dependent anti-cancer effect as fluoroquinolones, trimethoprim/sulfamethoxazole, and nitrofurantoin have been found to inhibit cell proliferation of bladder cancer cell lines [26, 27]. However, closer inspection of the Nijmegen data suggests a threshold above which the “protective” effect of multiple UTIs is lost: for men and women, the adjusted OR of UBC rose above 1.0 after 6 UTIs and increased sharply for men reporting ≥ 11 UTIs [17].
Finally, in our analysis, the decreased cancer risk associated with multiple UTIs was strengthened and between study heterogeneity slightly decreased following sensitivity analyses selecting for the most recently published and highest quality data. Regarding this result, it is important to consider that while Vermeulen et al. present the highest quality and overall most recent study on this topic, in eliciting information on UTI frequency, the authors changed their definition of UTI from self-reported symptoms (definition 1) to a definition based on the prescription of antibiotics (definition 2). As these definitions do not necessarily extract the same information, their data may be incomplete when trying to decipher whether multiple exposures to UTI affect cancer risk. The authors even highlighted small overlap between patients who answered “no” to the lesser UTI definition but later reported having been prescribed antibiotics for a UTI.
Most of the pooled analyses in this review identified considerable between study heterogeneity, evidence of the diverse body of research published on this topic. Obvious sources of heterogeneity include the several decades and countries from which these studies were individually performed. Differences in sample populations regarding age and smoking status cannot be understated given the impact these variables have on overall bladder cancer risk. Individual studies attempted to control these confounders with sophisticated methods of adjusted OR reporting. While it would be convenient to compare these ORs at face value, we made an a priori decision to extract and compare raw data given the dissimilar methods authors used to categorize and control for different confounders.
Potential criticism of our work includes meta-analysis in the face of persistent between study heterogeneity despite predefined sensitivity and subgroup analyses. We stress the heterogeneity reflects the quality of the existing literature, which we deem to be overall mediocre, and should not bar pooled synthesis as long as the heterogeneity is cited as a clear limitation. In this analysis, the persistently elevated I2 values underscore the caution with which readers should view the original studies. In other words, our pooled results should not be perceived to substantiate those studies that previously reported positive associations between bacterial UTI and risk of UBCNS, especially increasingly positive associations after multiple infections. We again highlight that statistical significance of our overall effect was lost following exclusion of European studies, studies published prior to year 2000, and at a high risk of bias.
Readers should consider theories of bacterial UTIs contributing or initiating UBCNS carcinogenesis have developed from the consecutive publication of case–control studies reviewed here as a well as a linkage to other models of UTI and UBC, such as schistosomiasis and patients with spinal cord injury. However, urogenital schistosomiasis is pathophysiologically distinct from bacterial UTI, and many patients with spinal-related neuropathic bladder conditions require intermittent or chronic catheterization, which have been separately implicated as risk factors for UBC [28]. Increased inflammation and cell turnover have been proposed as mechanisms by which schistosomiasis and chronic catheterization increase risk of bladder cancer. The association between UTI and UBCNS is unlikely to be explained by a similar mechanism. Urogenital schistosomiasis and chronic catheterization most commonly cause non-keratinizing forms of SCC. Conversely, UC is the most common form of bladder cancer in the general population of industrialized countries.
There are study-, outcome- and review-level limitations that warrant discussion. At the study level, we decided to exclude (when possible) data at least 1 year prior to the diagnosis of UBC for cases and data acquisition (e.g., subject interview) for controls. The purpose of this decision was to minimize diagnostic bias. Studies in the pooled analysis inconsistently excluded data: some did not categorize and present data this way, and others excluded 2 years of data. Also, as mentioned earlier, all studies were observational in nature and cannot identify causality. At the outcome level, our primary interest, UBCNS, was rarely reported in a subgroup of cases and controls without previous tobacco exposure. Therefore, our pooled results cannot account for the effect of tobacco use. At the review level, access to unpublished data may have clarified or allowed more meaningful analyses. For example, with complete access to raw data, we may have been able to control tobacco exposure or consistently exclude years prior to UBC diagnosis. Additionally, Kantor et al. and Hartge et al. may have used similar control populations. Contacting the authors may have clarified these points. However, given the wide publication timeframe, we decided it was unlikely we would be able to uniformly access all unpublished raw data, so we did not pursue this option. Finally, though the NOS is widely used and has been validated by expert opinion, it has substantial limitations, as noted by Stang [29]. Despite this, we felt the NOS was more appropriate for this investigation than other tools as it was specifically designed for nonrandomized, nonintervention observational studies.
Conclusion
While the results we present are provocative and in some cases statistically significant, we encourage readers to “turn back the dial” and view any association between UTI and UBCNS as hypothesis generating. The heterogeneity and overall mediocre quality of the literature on bacterial UTI and future risk of bladder cancer limit its applicability to clinical medicine. Until future investigation can better control important confounders, such as smoking status, and consistently use a more stringent definition of UTI, the data we present here cannot substantiate earlier reports that bacterial infection of the urinary tract alone modifies risk for bladder cancer.
Abbreviations
- CI:
-
Confidence interval
- HR:
-
Hazard ratio
- NOS:
-
Newcastle–Ottawa Scale
- OR:
-
Odds ratio
- PRISMA:
-
Preferred Reporting Items for Systematic Reviews and Meta-Analyses
- PROSPERO:
-
International Prospective Register of Systematic Reviews
- RR:
-
Risk ratio
- SCC:
-
Squamous cell carcinoma
- UBC:
-
Urinary bladder cancer
- UBCNS :
-
Non-schistosomiasis-related UBC
- UC:
-
Urothelial carcinoma
- UTI:
-
Urinary tract infection
References
International Agency for Research on Cancer Agents classified by the IARC monographs, vol. 1–118. http://monographs.iarc.fr/ENG/Classification/index.php. Accessed 21 Dec 2016
Anderson-Otunu O, Akhtar S (2016) Chronic infections of the urinary tract and bladder cancer risk: a systematic review. Asian Pac J Cancer Prev 17:3805–3807
Moher D, Liberati A, Tetzlaff J, Altman DG, The PRISMA Group (2009) Preferred Reporting Items for Systematic Reviews and Meta-Analyses: the PRISMA statement. Ann Intern Med 2009(151):264–269
Moher D, Shamseer L, Clarke M, Ghersi D, Liberati A, Petticrew M, Shekelle P, Stewart LA (2015) Preferred reporting items for systematic review and meta-analysis protocols (PRISMA-P) 2015 statement. Syst Rev 4:1
Farah D, Hsieh M, Bayne C (2016) Role of urinary tract infection in non-schistosomiasis bladder cancer: protocol for a systematic review and planned meta-analysis. PROSPERO CRD42016053888. http://www.crd.york.ac.uk/PROSPERO/display_record.asp?ID=CRD42016053888
Wells GA, Shea B, O’Connell D, et al. The Newcastle-Ottawa Scale (NOS) for assessing the quality of nonrandomized studies in meta-analyses. http://www.ohri.ca/programs/clinical_epidemiology/oxford.asp. Assessed December 12, 2016
The Cochrane Collaboration (2011) 10.4.3.1 Recommendations on testing for funnel plot asymmetry. In: Higgins JPT, Green S (eds) Cochrane handbook for systematic reviews of interventions version 5.1.0 [updated March 2011]. http://www.handbook.cochrane.org. Accessed 31 May 2016
Sterne JA, Sutton AJ, Ioannidis JP et al (2011) Recommendations for examining and interpreting funnel plot asymmetry in meta-analyses of randomized controlled trials. BMJ 343:d4002
Review Manager (RevMan) (2014) [Computer program]. Version 5.3. The Nordic Cochrane Centre, The Cochrane Collaboration, Copenhagen
Kantor AF, Hartge P, Hoover RN, Narayana AS, Sullivan JW, Fraumeni JF Jr (1984) Urinary tract infection and risk of bladder cancer. Am J Epidemiol 119:510–515
Kjaer SK, Knudsen JB, Sørensen BL, Møller Jensen O (1989) The Copenhagen case–control study of bladder cancer. V. Review of the role of urinary-tract infection. Acta Oncol 28:631–636
Hartge P, Harvey EB, Linehan WM et al (1990) Unexplained excess risk of bladder cancer in men. J Natl Cancer Inst 82:1636–1640
La Vecchia C, Negri E, D’Avanzo B, Savoldelli R, Franceschi S (1991) Genital and urinary tract diseases and bladder cancer. Cancer Res 51:629–631
Jhamb M, Lin J, Ballow R, Kamat AM, Grossman HB, Wu X (2007) Urinary tract diseases and bladder cancer risk: a case–control study. Cancer Causes Control 18:839–845
Jiang X, Castelao JE, Groshen S et al (2009) Urinary tract infections and reduced risk of bladder cancer in Los Angeles. Br J Cancer 100:834–839
Erdurak K, Dundar PE, Ozyurt BC, Negri E, La Vecchia C, Tay Z (2014) Smoking, occupation, history of selected diseases and bladder cancer risk in Manisa, Turkey. Eur J Cancer Prev 23:58–61
Vermeulen SH, Hanum N, Grotenhuis AJ et al (2015) Recurrent urinary tract infection and risk of bladder cancer in the Nijmegen bladder cancer study. Br J Cancer 112:594–600
Howe GR, Burch JD, Miller AB et al (1980) Tobacco use, occupation, coffee, various nutrients, and bladder cancer. J Natl Cancer Inst 64:701–713
Sullivan JW (1982) Epidemiologic survey of bladder cancer in greater New Orleans. J Urol 128:281–283
Claude J, Kunze E, Frentzel-Beyme R, Paczkowski K, Schneider J, Schubert H (1986) Life-style and occupational risk factors in cancer of the lower urinary tract. Am J Epidemiol 124:578–589
Piper JM, Matanoski GM, Tonascia J (1986) Bladder cancer in young women. Am J Epidemiol 123(6):1033–1042
González CA, Errezola M, Izarzugaza I et al (1991) Urinary infection, renal lithiasis and bladder cancer in Spain. Eur J Cancer 27:498–500
Kunze E, Chang-Claude J, Frentzel-Beyme R (1992) Life style and occupational risk factors for bladder cancer in Germany. A case–control study. Cancer 69:1776–1790
D’Avanzo B, La Vecchia C, Negri E, Decarli A, Benichou J (1995) Attributable risks for bladder cancer in northern Italy. Ann Epidemiol 5:427–431
Sun LM, Lin CL, Liang JA et al (2013) Urinary tract infection increases subsequent urinary tract cancer risk: a population-based cohort study. Cancer Sci 104:619–623
Seay TM, Peretsman SJ, Dixon PS (1996) Inhibition of human transitional cell carcinoma in vitro proliferation by fluoroquinolone antibiotics. J Urol 155:757–762
Kamat AM, Lamm DL (2004) Antitumor activity of common antibiotics against superficial bladder cancer. Urology 63:457–460
Gui-Zhong L, Li-Bo M (2017) Bladder cancer in individuals with spinal cord injuries: a meta-analysis. Spinal Cord 55:341–345
Stang A (2010) Critical evaluation of the Newcastle-Ottawa scale for the assessment of the quality of nonrandomized studies in meta-analyses. Eur J Epidemiol 25:603–605
Acknowledgements
We thank the anonymous reviewers for their role in improving our manuscript.
Funding
CEB is supported by the Joseph E. Robert, Jr. Fellowship at the Sheikh Zayed Institute for Pediatric Surgical Innovation, part of the Children’s National Health System. MHH is supported by a National Institutes of Health R56 (AI119168) grant.
Author information
Authors and Affiliations
Contributions
CEB: Protocol/project development, data collection, data analysis, manuscript writing. DF: Protocol/project development, data collection. KWH: Data analysis. MHH: Protocol/project development.
Corresponding author
Ethics declarations
Conflict of interest
The authors declare that they have no conflict of interest.
Informed consent
For this type of study formal consent is not required.
Research involving human participants and/or animals
All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. This article does not contain any studies with human participants or animals performed by any of the authors.
Protocol registration
This study protocol (2016:CRD42016053888) was developed prior to data collection according to PRISMA and PRISMA-P guidelines and registered with the International Prospective Register of Systematic Reviews (PROSPERO): http://www.crd.york.ac.uk/PROSPERO/display_record.asp?ID=CRD42016053888.
Electronic supplementary material
Below is the link to the electronic supplementary material.
345_2018_2257_MOESM2_ESM.png
Supplementary material 2 (PNG 4 kb) Funnel plot illustrating risk ratio according to standard errorLegend: RR, risk ratio; SE, standard error; dashed line = overall effect; RR of 1 equates to no intervention effect
345_2018_2257_MOESM3_ESM.tif
Supplementary material 3 (TIFF 83 kb) Pooled analysis for exposure to UTI on risk of UBC according to sex. A) Males only. B) Females only
345_2018_2257_MOESM4_ESM.tif
Supplementary material 4 (TIFF 55 kb) Pooled analysis exposure to UTI on UBC risk according to smoking history*. A) With smoking history. B) Without smoking history. *Jiang et al data reported smoking data for women only and excluded data within 5 years of cancer diagnosis
345_2018_2257_MOESM5_ESM.tif
Supplementary material 5 (TIFF 63 kb) Pooled analysis for exposure to multiple UTI (versus no exposure) on risk of UBC*. A) Overall analysis. B) Analysis after excluding studies prior to year 2000 and with high risk of bias. * Jhamb et al excluded data within 3 years of cancer diagnosis; Jiang et al reported episodes of bladder infection for women only and excluded data within 5 years of cancer diagnosis
Rights and permissions
About this article
Cite this article
Bayne, C.E., Farah, D., Herbst, K.W. et al. Role of urinary tract infection in bladder cancer: a systematic review and meta-analysis. World J Urol 36, 1181–1190 (2018). https://doi.org/10.1007/s00345-018-2257-z
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00345-018-2257-z