Vasectomy and Risk of Prostate Cancer: A Systematic Review and Meta-analysis

Take Home Message We found a significant association between vasectomy and any-grade, localized, and advanced prostate cancer. However, the effect estimates were increasingly close to null when examining studies of robust design and high quality. Future well-designed studies with attention to potential residual confounders, which were not taken into account in large cohort studies, are needed.

1.11; p < 0.00001; I 2 = 31%), or advanced PCa (OR 1.07; 95% CI, 1.02-1.13; p = 0.006; I 2 = 0%). The association with PCa remained significant when the analyses were restricted to studies with a low risk of bias (OR 1.06; 95% CI, 1.02-1.10; p = 0.02; I 2 = 48%) or cohort studies (OR 1.09; 95% CI, 1.04-1.13; p < 0.0001; I 2 = 64%). Among studies adjusted for PSA screening, the association with localized PCa (OR 1.06; 95% CI, 1.03-1.09; p < 0.001; I 2 = 0%) remained significant. Conversely, vasectomy was no longer associated with localized high-grade (p = 0.19), advanced (p = 0. 22), and lethal (p = 0.42) PCa. Conclusions: Our meta-analysis found an association between vasectomy and any, mainly localized, PCa. However, the effect estimates of the association were increasingly close to null when examining studies of robust design and high quality. On exploratory analyses including studies, which adjusted for PSA screening, the association for aggressive and/or advanced PCa diminished. Patient summary: In this study, we found an association between vasectomy and the risk of developing localized prostate cancer without being able to determine whether the procedure leads to a higher prostate cancer incidence.

Introduction
Vasectomy is the fourth most common method of contraception with increasing interest among men over the past decade. Worldwide, approximately 6-8% of couples choose this method of contraception [1]. Vasectomy is the most effective permanent male contraceptive option with failure rates <1% [2]. The high level of effectiveness and low complication rates made vasectomy the foremost utilized nondiagnostic operation performed by urologists in highly developed countries [3].
Since the first report of a positive relationship between vasectomy and prostate cancer (PCa) [4], there has been an endless debate about possible associations with conflicting results. These discrepancies are due to the paucity of documented causal associations, and possible detection biases related to PCa screening and closer follow-up among vasectomy patients, and modest clinical significance with a relative risk very often close to 1. Recently, several large, high-quality reports demonstrated conflicting results [5][6][7][8][9]. A recent meta-analysis that included the most recent reports found that vasectomy was associated with localized and advanced PCa [10]. However, outcomes by disease stage were not adjusted by prostate-specific antigen (PSA) screening. Given the potential confounding effect of follow-up PSA screening, to best inform our patients on the oncological risks associated with vasectomy, the PCa Oncology Committee of the French Association of Urology conducted a systematic review of the literature and performed a meta-analysis, with a particular focus on whether there is an association between vasectomy and PCa, in both unadjusted and PSA screening-adjusted studies.

Protocol and registration
We conducted a systematic review in line with the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) guidelines [11]. A protocol was registered in PROSPERO (registration number: CRD42022303026).

Search strategy
A literature search was conducted until January 2022 in PubMed/Medline, Scopus, and Web of Science databases. Studies were selected if they included men of any age (patient) who underwent vasectomy (intervention) compared with those who did not undergo vasectomy (comparator). We analyzed any subsequent diagnosis of PCa (outcome) in prospective and retrospective studies (study design). The search strategy used the combination of the following terms grouped according to the Boolean operators (AND, OR, and NOT): vasectomy, deferentectomy, vasoligation, vasoligature, prostate, prostatic, neoplasm, tumor, and cancer. Initial screening was performed independently by two investigators based on the titles and abstracts of the article to identify ineligible reports (M.B. and P.R.). Reasons for exclusion were noted. Potentially relevant reports were subjected to a full-text review, and the relevance of the reports was confirmed after the data extraction process. Disagreements were resolved by consultation with a third coauthor (G.P.).

Inclusion and exclusion criteria
We included prospective and retrospective studies, which analyzed over 1000 patients, that compared the risk of developing PCa in vasectomized and nonvasectomized patients. No patient had a personal history of PCa at baseline. In case of duplicate publications, either the higherquality or the most recent publication was selected. Reviews, meta-analyses, commentaries, meeting abstracts, authors' replies, thesis, and case reports were excluded, but the reference section was checked not to omit relevant articles. Case series lacking comparator groups were also excluded. No restriction on the publication date was applied. Only English-language articles were assessed for eligibility.

Data Extraction
Two authors (M.B. and P.R.) performed an independent initial screening based on the titles and abstracts, and noted the cause of exclusion of ineligible reports. Studies were considered eligible if these reported an effect estimate for an association between vasectomy and any PCa incidence (detection). We independently extracted the following variables from the included studies: first author's name, publication year, country of research, study design, period of patient recruitment, number of patients included, PSA screening, duration of follow-up, PCa detection, tumor characteristics, and potential confounders. We extracted odds ratios (ORs) with 95% confidence intervals (CIs) for the risk of developing PCa in vasectomized versus nonvasectomized patients. The primary outcome was a diagnosis of any PCa. Secondary outcomes included the diagnosis of PCa stratified by disease stages: localized PCa, localized high-grade PCa, advanced PCa, and fatal PCa. No single consensus criterion was used to define high-grade and advanced PCa, and we used the definitions reported in each included study. All discrepancies regarding data extraction were resolved by consensus with a senior author (G.P.).

Quality assessment and risk of bias
We used the Newcastle-Ottawa Scale (NOS) to assess studies' quality and the risk of bias (RoB). This scale assesses RoB in three areas: study group selection, group comparability, and exposure and outcome assessment. Studies that scored 7 were considered of high quality, and those with scores 4-6 were of moderate quality and scores <4 of poor quality. We considered the follow-up adequate if the median or mean follow-up was >5 yr. We assessed publication biases using funnel plots.

Statistical analysis
We used the inverse variance technique to calculate the pooled ORs for PCa risk and corresponding 95% CIs. We assessed heterogeneity using the Q test and quantified it using I 2 values [12]. We used either a fixed-or a randomeffect model for calculations of ORs according to the heterogeneity of the pooled studies. We assessed heterogeneity using the Cochrane's Q test and quantified it using I 2 values. In the case of heterogeneity (Cochrane's Q test p < 0.05 and I 2 > 50%), we used a random-effect model (DerSimonian method) and attempted to investigate and explain the heterogeneity; otherwise, the fixed-effect model (Mantel-Haenszel method) was used. Meta-analyses and graph figures were generated using the Cochrane Review Manager 5.4 (RevMan 5.4; The Cochrane Centre, Copenhagen, Denmark). Statistical significance was set at p < 0.05.

Subgroup analysis
Subgroup analyses were planned a priori. First, we evaluated the outcomes in subgroups of patients who were submitted to PSA screening versus those in patients in whom PSA screening was not performed or was very uncommon. Second, we examined studies according to publication year (1990-2000, 2001-2010, and 2011-2021). Third, we limited our analyses to studies identified as having a low RoB. Fourth, we evaluated the series according to study design (cohort vs cross sectional vs case control). Finally, we compared cohort studies with a follow-up of <10 versus >10 yr.

Study selection
The study selection process is outlined in the PRISMA flow diagram (Fig. 1). A total of 466 unique records were identified. Of these, 70 full-text articles were assessed for eligibility and 37 met the inclusion criteria for qualitative and quantitative analysis [4,[6][7][8][9]. The reasons for exclusion are summarized in Figure 1.

Study characteristics
The baseline characteristics of the included studies are presented in Table 1 A sensitivity analysis was performed to assess the influence of individual studies on the overall risk of PCa. After excluding any study that did not substantially influence the direction and magnitude of the cumulative estimates, we obtained similar results.

Vasectomy and PCa stratified by disease stage
The association between vasectomy and localized, localized high-grade, advanced, and fatal PCa was investigated through the analyses of ten, eight, 13, and nine studies, respectively (Fig. 3) Table 2). The Cochrane's Q and I 2 tests did not show any heterogeneity in all pooled analyses. Indeed, the increase in any PCa risk decreased from 23% when all studies were considered to 9% when the analysis was limited to cohort studies, and to 6% for studies with a low RoB.

Subgroup analyses
Restricting the analyses to studies published <10 yr ago, there was a significant association between vasectomy and any type of PCa (   3.8.

Discussion
In this systematic review and meta-analysis, we found that vasectomy was significantly associated with a low risk of developing PCa. This association remained after restriction of our analyses to high-quality and cohort studies. However, the effect estimates of the association between vasectomy and PCa were increasingly closer to the null when analyzing studies with robust study design and study quality. Indeed, the increase in PCa risk fell from 23% when all studies were considered to 9% when the analysis was limited to cohort studies and to 6% for high-quality studies. It is questionable whether such low statistical significance may have a true clinical impact and whether it should influence vasectomy decision-making. It has been suggested that an individual cancer risk assessment could be considered before vasectomy, depending on other risk factors such as Afro-Caribbean origin or a family history of PCa [45]. Neverthe- Fig. 2 -Forest plots for meta-analyses of the adjusted estimates for the association between vasectomy and prostate cancer by study design. Data were pooled separately by study design. As significant heterogeneity (I 2 > 50%) was found, a pooled estimate was calculated with a random-effect model (DerSimonian and Laird method). CI = confidence interval; df = degrees of freedom; IV = inverse variance; SE = standard error.
less, a statistically significant association is different from causation. Some preclinical studies have tried to explain the association [46,47]. Possible explanations for the increased risk of PCa in vasectomized individuals include a decrease in prostatic secretory volume resulting in prolonged exposure to certain carcinogens, an increase in circulating androgens or in the binding capacity of androgenbinding proteins, development of antisemen antibodies that can affect immunological processes, and reduced levels of certain molecules in seminal plasma, such as IGF-1 and IGFBP3, known to be involved in prostate carcinogenesis. However, these molecular mechanisms underlying the link between vasectomy and PCa remain speculative. Therefore, we cannot argue with certainty that a causal association exists due to potential residual confounders. Indeed, it has been suggested that men undergoing vasectomy likely have multiple factors that bias PCa detection, such as the intensity for follow-up PSA screening. Given the strong confounding effect of PSA screening, we assessed the risk of any PCa in studies adjusted for PSA screening. Vasectomy and PCa remained significantly associated but with an excess risk of only 6%, with no significant heterogeneity between the included studies. Compared with the last meta-analysis published by Xu et al [10], in a subgroup analysis including studies adjusted with PSA screening, we found no association between vasectomy and high-grade, advanced, or fatal localized PCa. Thus, our conclusions are more moderate and cautious than those of Xu et al [10]. At present, it is unknown whether residual confounding factors could be responsible for the modest excess of PCa incidence or whether this association should be considered definitive.
If assuming that PSA screening is a potential bias, it is expected to influence outcomes by disease stage. Indeed, PSA screening has been associated with increased detection of localized disease and decreased advanced PCa [48,49]. Similar to two previous meta-analyses [10,50], we found a positive association between vasectomy and advanced PCa. Nevertheless, as expected, when we restricted our analysis to studies adjusted for PSA screening, this association was no longer significant. Finally, PCa mortality was not influenced by vasectomy, which was consistent with previous reports [10,[50][51][52].
This review has several limitations that should be acknowledged. First, several confounding factors were not taken into account in the individual studies, making it impossible to establish definitively a causality between vasectomy and PCa. Second, substantial heterogeneity was observed across the included studies. Third, our results are based primarily on North American studies where vasectomy is more common (22% in Canada, 12% in USA, 11% in Oceania and Northern Europe, 3-5% in South America, and <1% in Africa [53]) and where PCa screening practices varied considerably over the study periods. Nevertheless, our study also has several strengths, including the number of patients included, the a priori definition of subgroup analyses, and the consideration of a detection bias by PSA screening for each PCa stage to refine the evaluation of the association.

3.9.
Implications for practice and future research The results of this study need to be interpreted with caution. Translating our results into clinical practice is likely to dissuade patients from undergoing vasectomy, whereas the absolute risk may be close to zero. Clear, fair, and understandable information should be provided about a possible association between vasectomy and PCa, without being able to determine whether there is any causality. To definitively address this question, future collaborative, well-designed, international studies are needed to prospectively assess this risk of PCa among vasectomized patients with particular attention to potential confounders such as wellestablished risk factors for developing PCa.

Conclusions
Our meta-analysis found a significant association between vasectomy and the risk of any, mainly localized, PCa. However, the effect estimates for the association between vasectomy and PCa were increasingly close to null when examining studies of robust design and quality. When we limited our analysis to studies adjusted for PSA screening, the association remained significant only for localized disease, but not for aggressive and/or advanced PCa. Future studies are needed to prospectively assess the possible  causality between vasectomy and PCa, with attention to potential residual confounders that were not taken into account in large cohort studies.
Author contributions: Michael Baboudjian had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Administrative, technical, or material support: None.
Financial disclosures: Michael Baboudjian certifies that all conflicts of interest, including specific financial interests and relationships and affiliations relevant to the subject matter or materials discussed in the manuscript (eg, employment/affiliation, grants or funding, consultancies, honoraria, stock ownership or options, expert testimony, royalties, or patents filed, received, or pending), are the following: None.
Funding/Support and role of the sponsor: None.