Vasectomy and the risk of prostate cancer: a meta-analysis of cohort studies

Abstract

Background: The relationship of vasectomy to prostate cancer has great public health significance. However, the results of observational studies were conflicting. To determine whether vasectomy is associated with the risk of prostate cancer, we performed a meta-analysis of cohort studies. Methods: A literature search was carried out using Pubmed, Embase, Cochrane Libraryl, and China National Knowledge Infrastructure (CNKI) between January 1966 and July 2013. Before meta-analysis, between-study heterogeneity and publication bias were assessed using adequate statistical tests. Fixed-effect and random-effect models were used to estimate summary relative risks (RR) and the corresponding 95% confidence intervals (CIs). Potential sources of heterogeneity were detected by meta-regression. Subgroup analyses and sensitivity analysis were also performed. Results: A total of nine cohort studies contributed to the analysis. There was heterogeneity among the studies but no publication bias. Pooled results indicated that vasectomy was not associated with a significant increase of total prostate cancer risk (RR = 1.07, 95% CI [0.79, 1.46]). When stratified the various studies by geographic location, we found a significant association between vasectomy and increased PCa risk among studies conducted in the USA (RR = 1.54, 95% CI [1.23, 1.93]), however, there was no significant association between vasectomy and PCa risk among studies conducted in non-USA countries (RR = 0.74, 95% CI [0.50, 1.09]). Furthermore, sensitivity analysis confirmed the stability of the results. Conclusions: In conclusion, the present meta-analysis of cohort studies suggested that vasectomy was not associated with increased risk of prostate cancer. More in-depth studies are warranted to report more detailed results, including stratified results by age at vasectomy, tumor grade, and tumor stage.

Introduction

Prostate cancer (PCa) is the second-most frequently diagnosed cancer and the sixth-leading cause of cancer death in males worldwide [1,2]. The cause of PCa is not well known, but multiple risk factors have been identified, including age, race, and family history of PCa [3-5]. Many putative risk factors, including androgens, diet, physical activity, sexual factors, inflammation, and obesity, have been implicated, but their roles in PCa etiology remain unclear.

Vasectomy is an important method of birth control, with approximately 500,000 vasectomies performed annually in the United States alone [6]. About 12% of married men have had a vasectomy and are generally under the age of 40 years when the procedure is performed [7]. The relationship of vasectomy to PCa has great public health significance [8-10]. Vasectomy has been hypothesized to increase the risk of PCa by diminishing the secretion of prostatic fluid or by altering immune response to sperm antigens [11-13]. Several epidemiological studies have investigated the association of vasectomy with PCa, however, their results were conflicting. The previous meta-analysis by LK Dennis et al [14] in 2002 suggested that men with a prior vasectomy may be at an increased risk of PCa (RR = 1.37, 95% CI, 1.15-1.62). However, the majority of observational studies included in their meta-analysis were case-control studies, and there were only five cohort studies. Since 2002, more cohort studies are published. To further evaluate the effect of vasectomy on the risk of developing PCa, we now performed a meta-analysis of cohort studies.

Materials and methods

Systematic search strategy

The present meta-analysis was conducted following the Preferred Reporting Items for Systematic reviews and Meta-Analyses guidelines (PRISMA) [15], and the meta-analysis of observational studies in epidemiology (MOOSE) guidelines [16]. A literature search was carried out using Pubmed, Embase, Cochrane Library, and China National Knowledge Infrastructure (CNKI) between January 1966 and July 2013. Search terms included: “vasectomy” OR “deferentectomy” OR “gonangiectomy” OR “vas ligation” OR “vasoligation” OR “vas ligature” OR “vasoligature” OR “vas occlusion” AND “prostate cancer”. The reference lists of each comparative study included in this meta-analysis and previous reviews were manually examined to identify additional relevant studies.

Study selection

Two reviewers independently selected eligible cohort studies that investigated vasectomy and the risk of PCa. Disagreement between the two reviewers was settled by discussing with the third reviewer. Inclusion criteria were: (i) used a cohort study design; (ii) evaluated the association between vasectomy and PCa risk; (iii) presented odds ratio (OR), relative risk (RR), or hazard ratio (HR) estimates with its 95% confidence interval (CI). When there were multiple publications from the same population, only data from the most recent report was included in the meta-analysis and the remaining was excluded. Studies reporting different measures of RR like risk ratio, rate ratio, hazard ratio, and odds ratio were included in the meta-analysis. In practice, these measures of effect yield a similar estimate of RR, since the absolute risk of PCa is low.

Assessment of study quality, data extraction, and analysis

Newcastle-Ottawa scale (NOS) was used to assess the methodological quality of cohort studies [17]. The NOS contains eight items that are categorized three categories: selection (four items, one star each), comparability (one item, up to two stars), and outcome (three items, one star each). A “star” presents a “high-quality” choice of individual study. Two reviewers assessed the methodological quality independently. Disagreement between the two reviewers was settled by discussing with the third reviewer.

The following data was collected by two reviewers independently using a purpose-designed form: name of first author, publishing time, country of the population studied, study design, study period, follow-up time, number of PCa cases and subjects, the study-specific adjusted ORs, RRs, or HRs with their 95% CIs, confounding factors for matching or adjustments.

Heterogeneity was assessed using the Cochran Q and I² statistics. For the Q statistic, a P value < 0.10 was considered statistically significant for heterogeneity; for the I² statistic, heterogeneity was interpreted as absent (I²: 0%-25%), low (I²: 25.1%-50%), moderate (I²: 50.1%-75%), or high (I²: 75.1%-100%) [18]. The overall analysis including all eligible studies was performed first, and subgroup analyses were performed according to (i) study design (prospective cohort study and retrospective cohort study), (ii) Study location (USA and non-USA), (iii) publication year (before 2000 and after 2000), and (iv) control for confounding factors (n ≥ 4, and n ≤ 3) to examine the impact of these factors on the association. When substantial heterogeneity was detected, the summary estimate based on the random-effect model (DerSimonian-Laird method) [19] was reported, which assumes that the studies included in the meta-analysis had varying effect sizes. Otherwise, the summary estimate based on the fixed-effect model (the inverse variance method) was reported, which assumes that the studies included in the meta-analysis had the same effect size. To derive the relationship between time since vasectomy and risk for PCa, we carried out dose-response analysis. For the dose-response meta-analysis, methods proposed by Greenland [20] and Orsini [21] were used to estimate study-specific slopes. To test the robustness of association and characterize possible sources of statistical heterogeneity, sensitivity analysis were carried out by excluding studies one-by-one and analyzing the homogeneity and effect size for all of rest studies. To better investigate the possible sources of between-study heterogeneity, a meta-regression analysis was performed [22]. Publication bias was assessed using Begg and Mazumdar adjusted rank correlation test and the Egger regression asymmetry test [23,24]. All analyses were performed using Stata version 11.0 (Stata Corp, College Station, TX).

Results

Search results and reporting quality

Figure 1 illustrates the search process and the final selection of relevant studies. 590 records were identified through database searching. On the basis of the titles and abstracts, we identified nine full-text articles. After further evaluation, one study was excluded because it was from the same population. One study was identified from reference lists. At last, a total of nine eligible cohort studies published between 1991 and 2012 were identified [10,25-32] (Baseline data and other details of included studies are shown in Table 1). A total of 331,436 male subjects, including 1,245 PCa cases were involved. Of the nine included studies, three studies were conducted in Europe [10,28,29], four studies in the USA [25-27,30], and remaining two studies in other countries [31,32]. Only two studies [26,32] were prospective cohort studies, and the others were retrospective cohort studies. The NOS scores for the included cohort studies ranged from 5 to 8, with a median 6; about 60% (6/9) of included studies were deemed to be of a high quality (≥ 6) (shown in Table S1).

Figure 1.

Flow diagram of screened, excluded, and analysed publications.

Table 1.

Characteristics of cohort studies included in the present meta-analysis

Author	Year	Country	Study design	All male subjects	PCa cases	Age (y)	Study period	Follow-up (y)	Confounders for adjustment	Adjusted risk estimate, RR (95% CI)
Sidney S	1991	USA	Retrospective cohort study	20,466	135	45.8	1977-1987	6.8 (mean)	age, race, marital status, date and location of multiphasic health exam	Total 1.1 (0.7-1.6)
										Years Since Vasectomy
										< 10 1.2 (0.5-2.6)
										10-20 1.2 (0.6-2.6)
										> 20 1.2 (0.6-2.2)
										Age at vasectomy
										< 40 1.0 (0.5-1.9)
										≥ 40 1.4 (0.8-2.3)
Nienhuis H	1992	Britain	Retrospective cohort study	35,442	6	25-49	1970-1986	7.5 (mean)	age	Total 0.44 (0.1-4.0)
Giovannucci E	1993	USA	Retrospective Cohort Study	29,214	96	42	1976-1989	NA	age	Total 1.56 (1.03-2.37)
										Years Since Vasectomy
										1-9 1.11 (0.46-2.70)
										10-19 1.26 (0.75-2.10)
										≥ 20 1.89 (1.14-3.14)
Giovannucci E	1993	USA	Prospective Cohort Study	47,855	300	40-75	1986-89	NA	age, marital status, race, and geographicalregion	Total 1.66 (1.25-2.21)
										Years Since Vasectomy
										< 13 1.24 (0.61-2.55)
										13-21 1.39 (0.83-2.31)
										≥ 22 1.77 (1.18-2.64)
Møller H	1994	Denmark	Retrospective Cohort Study	73,917	165	NA	1977-1989	NA	age	Total 0.98 (0.84-1.14)
Lynge E	2002	Denmark	Retrospective Cohort Study	115,862	93	NA	1977-1995	12.7 (mean)	none	Total 0.98 (0.73 -1.31)
										Years Since Vasectomy
										0-4 0.95 (0.31-2.21)
										5-9 1.24 (0.71-2.01)
										10-14 1.12 (0.69-1.72)
										≥ 15 0.40 (0.11-1.02)
										Age at vasectomy
										≤ 30 14.26 (1.73-51.57)
										30-39 0.84 (0.31-1.82)
										40-49 0.80 (0.48-1.25)
										50-59 1.06 (0.56-1.81)
										≥ 60 1.65 (0.61-3.60)
Rohrmann S	2005	USA	Retrospective Cohort Study	3,373	78	54.8	1996-2004	8.0 (mean)	age	Total 2.03 (1.24-3.32)
										Low-stage 1.47 (0.55-3.90)
										High-stage 1.52 (0.46-5.06)
										Low-grade 2.87 (1.49-5.54)
										High-grade 0.99 (0.36-2.76)
Yong N	2008	China	Retrospective Cohort Study	3,186	314	NA	1996-2005	NA	none	Total 0.50 (0.37-0.67)
										Years Since Vasectomy
										20-29 0.39 (0.14-1.09)
										30-39 0.31 (0.20-0.47)
										≥ 40 1.12 (0.75-1.68)
Romero FR	2012	Brazil	Prospective Cohort Study	2,121	58	≥ 40	2006-2011	21.5 (mean)	none	Total 0.23 (0.03-1.70)

NA, not available; RR, relative risk; CI, confidence interval.

Meta-analysis results

Because of significant heterogeneity was observed (I² = 83.4%, P < 0.001), a random-effects model was chosen over a fixed-effects model and we found that vasectomy did not significantly affect the risk of PCa (RR = 1.07, 95% CI [0.79, 1.46]). Both multivariable adjusted RR estimates with 95% CIs of each study and combined RR is shown in Figure 2. In the present meta-analysis, no publication bias was observed among studies using Begg’s P value (P = 0.677); Egger’s (P = 0.966) test, which suggested there was no evidence of publication bias (Figure 3).

Figure 2.

Forest plot: overall meta-analysis of vasectomy and prostate cancer risk. Squares indicated study-specific risk estimates (size of square reflects the study-statistical weight, i.e. inverse of variance); horizontal lines indicate 95% confidence intervals; diamond indicates summary relative risk estimate with its corresponding 95% confidence interval.

Figure 3.

Funnel plot for publication bias in the studies investigating risk for prostate cancer associated with vasectomy.

Subgroup analyses, and sensitivity analysis

No significant association was detected between vasectomy and PCa risk among prospective cohort studies (RR = 0.80, 95% CI [0.12, 5.20]), as well as retrospective cohort studies (RR = 1.03, 95% CI [0.75, 1.42]), presented in Table 2. When stratified the various studies by geographic location, we found a significant association between vasectomy and increased PCa risk among studies conducted in the USA (RR = 1.54, 95% CI [1.23, 1.93]), however, there was no significant association between vasectomy and PCa risk among studies conducted in non-USA countries (RR = 0.74, 95% CI [0.50, 1.09]). When we examined whether the associations differed by publication year, no significant association was detected between vasectomy and PCa risk among studies published before 2000 (RR = 1.24, 95% CI [0.92, 1.67]), as well as studies published after 2000 (RR = 0.86, 95% CI [0.44, 1.70]). Further, we found that the association was not affected by the number of adjustment factors (see in Table 2).

Table 2.

Summary of pooled relative risks of prostate cancer in subgroups

	No. of studies	RR	95% CI	P for heterogeneity	I2 (%)
All	9	1.07	0.79-1.46	< 0.001	83.4
Region
USA	4	1.54	1.23-1.93	0.259	25.4
Non-USA	5	0.74	0.50-1.09	0.001	78.4
Study type
Prospective cohort study	2	0.80	0.12-5.20	0.057	72.3
Retrospective cohort study	7	1.03	0.75-1.42	< 0.001	82.3
Publication year
Before 2000	5	1.24	0.92-1.67	0.008	70.7
After 2000	4	0.86	0.44-1.70	< 0.001	88.7
Number of adjustment factors
n ≥ 4 confounders	2	1.39	0.93-2.07	0.108	61.3
n ≤ 3 confounders	7	0.98	0.68-1.41	< 0.001	83.0

To test the robustness of association and characterize possible sources of statistical heterogeneity, sensitivity analyses were carried out by excluding studies one-by-one and analyzing the homogeneity and effect size for all of the rest studies. Sensitivity analysis indicated that no significant variation in combined RR by excluding any of the study, confirming the stability of present results.

Dose-response analysis and meta-regression analysis

We evaluated evidence for a dose-response relationship among the five studies [25-27,29,31] which reported information on time since vasectomy and PCa risk. However, there was no evidence of a linear association between time since vasectomy and PCa risk (P for linearity = 0.565; Figure 4). To better investigate the possible sources of between-study heterogeneity, a meta-regression analysis was performed. Study design, geographic area, publication year, and major confounders adjusted, which may be potential sources of heterogeneity, were tested by a meta-regression method. We found that only geographic area (P < 0.05) had statistical significance in a multivariate model.

Figure 4.

The dose-response analysis between time since vasectomy and prostate cancer risk obtained by the restricted cubic spline regression model. P for linearity = 0.565. The solid line and the short dash line represent the estimated RR and its 95% CI. Long dash line represents the linear relationship.

Discussion

Vasectomy is a common birth control method, and prostate cancer is the most frequently diagnosed solid tumor in men, so the relationship of vasectomy to PCa has great public health significance [8,33]. The present meta-analysis included nine cohort studies currently available (two prospective cohort studies and seven retrospective cohort studies), involving a total of 331,436 male subjects and 1,245 PCa cases. There was statistically significant heterogeneity among the nine included cohort studies, so a random-effects model was chosen over a fixed-effects model. Finally, we found that vasectomy did not significantly affect the risk of PCa. Further, there was no evidence of a linear association between time since vasectomy and PCa risk. Meta-regression analysis revealed that geographic area may be the source of heterogeneity. Sensitivity analysis indicated that an omission of any studies did not alter the magnitude of observed effect, suggesting a stability of our findings. Moreover, the results of Begg’s test and Egger’s test did not support the existence of major publication bias. In our subgroup analyses, the results were not substantially affected by study design, publication year, and confounder adjustment. Prospective cohort and retrospective cohort studies alone showed no significant association between vasectomy and the risk of PCa. However, we should notice that there were only two prospective cohort studies investigating the association between vasectomy and PCa risk. That number was rather low to draw firm conclusions. As we know, compared with retrospective cohort studies, prospective cohort studies are less susceptible to bias (e.g. recall bias, selection bias) due to their nature [34-36]. So, more prospective cohort studies are needed to confirm the associations found in the present meta-analysis. Although the association was not affected by the number of adjustment factors, we should notice that the number of adjustment factors was rather low among the included studies (shown in Table 1). There are a lot of factors which may affect the risk of PCa, such as age, race, family history of PCa, androgens, diet, physical activity, inflammation, and obesity. Further, sexual activity may also affect the association between vasectomy and PCa risk. Since an increased number of sexual partners and history of a sexually transmitted infection appear to be related to PCa but likely inversely related to having a vasectomy [37,38], this would cause negative confounding. The future studies should adjust as more confounders as possible [39]. We found a significant association between vasectomy and increased PCa risk among studies conducted in the USA but not non-USA countries. Explanations for the inconsistent findings between study location are not known. There are many possible reasons which will lead to the difference in association between different areas. The differences in genetic susceptibility, culture, and lifestyles may explain part of the inconsistency of the results [40].

Among the nine included studies, only two studies stratified the association by age at vasectomy. Sidney S et al [25] found that the RR of PCa associated with vasectomy increased with age at vasectomy (1.4 in men 40 or more years old and l.0 in men less than 40 years old), but the CIs around these RRs were wide and included one. In the study by Lynge E et al [29], no difference was seen in PCa risk by age at vasectomy. So whether age at vasectomy will affect PCa risk is unclear, and this topic should be further investigated in the future. Rohrmann S et al found that the risk of low-grade disease (HR = 2.87; 95% CI 1.49-5.54), but not high-grade disease (HR = 0.99; 95% CI 0.36-2.76), was higher in men who had had a vasectomy. No statistically significant associations were observed for low-stage or high-stage disease. Their findings should be confirmed by more cohort studies in the future.

The strength of the present meta-analysis lies in a large sample size (331,436 male subjects and 1,245 PCa cases) and no significant evidence of publication bias. Two investigators independently performed the article identification, data extraction, and verification and resolved all discrepancies. Furthermore, our findings were stable and robust in sensitivity analyses. However, several limitations to this meta-analysis should be noted. Firstly, as a meta-analysis of observational data, the possibility of recall and selection biases cannot be ruled out. Compared with retrospective cohort studies, prospective cohort studies are less susceptible to bias due to their nature. However, the present meta-analysis included only two prospective cohort studies, so more prospective cohort studies are need to confirm the associations in the future. Secondly, we did not search for unpublished studies, so only published studies were included in our meta-analysis. Therefore, publication bias may have occurred although no publication bias was indicated from both visualization of the funnel plot and Egger’s test. Thirdly, the number of adjustment factors was rather low among the included studies, so the future studies should adjust as more confounders as possible.

In conclusion, the present meta-analysis of cohort studies suggests that vasectomy is not associated with increased risk of PCa. More in-depth studies are warranted to report more detailed results, including stratified results by age at vasectomy, tumor grade, and tumor stage.

Disclosure of conflict of interest

None.