EUROPEAN UROLOGY 60 (2011) 796–808

available at www.sciencedirect.com journal homepage: www.europeanurology.com

Collaborative Review – Andrology

Varicocele and Male Factor Infertility Treatment: A New Meta-analysis and Review of the Role of Varicocele Repair Abdulaziz Baazeem a,*, Eric Belzile b, Antonio Ciampi b, Gert Dohle c, Keith Jarvi d, Andrea Salonia e, Wolfgang Weidner f, Armand Zini g a

Department of Surgery, Umm Al-Qura University, Makkah, Saudi Arabia; b Department of Clinical Epidemiology and Community Studies, St. Mary’s Hospital

Center, Montreal, Quebec, Canada; c Department of Urology, Erasmus MC, University Medical Center, Rotterdam, The Netherlands;

d

Division of Urology,

Samuel Lunenfeld Research Institute and the Institute of Medical Sciences, University of Toronto, Toronto, Ontario, Canada; e Department of Urology, VitaSalute University, San Raffaele Hospital, Milan, Italy; f Department of Urology, Pediatric Urology and Andrology, Justus-Liebig University Giessen, Giessen, Germany; g Division of Urology, McGill University, Montre´al, Quebec, Canada

Article info

Abstract

Article history: Accepted June 9, 2011 Published online ahead of print on June 20, 2011

Context: Varicocele is a common condition, found in many men who present for infertility evaluation. Objective: To assess the effect of varicocelectomy on male infertility. Evidence acquisition: A literature search was performed using Embase and Medline. Literature reviewed included meta-analyses and randomized and nonrandomized prospective (controlled and noncontrolled) studies. In addition, a new meta-analysis was performed. Evidence synthesis: Four randomized controlled trials reporting on pregnancy outcome after repair of clinical varicoceles in oligozoospermic men were identified. Using the random effect model, the combined odds ratio was 2.23 (95% confidence interval [CI], 0.86–5.78; p = 0.091), indicating that varicocelectomy is moderately superior to observation, but the effect is not statistically significant. We identified 22, 17, and 5 prospective studies reporting on sperm concentration, total motility, and progressive motility, respectively, before and after repair of clinical varicocele. The random effect model combined improvement in sperm concentration was 12.32 million sperm per milliliter (95% CI, 9.45– 15.19; p < 0.0001). The random effect model combined improvement in sperm total and progressive motility were 10.86% (95% CI, 7.07–14.65; p < 0.0001) and 9.69% (95% CI, 4.86–14.52; p = 0.003), respectively. These results indicate that varicocelectomy is associated with a significant increase in sperm concentration as well as total and progressive motility. Prospective studies also show that varicocelectomy reduces seminal oxidative stress and sperm DNA damage as well as improving sperm ultramorphology. Studies indicate that a microsurgical approach to a varicocele repair results in less recurrence and fewer complications than other techniques. Conclusions: Although there is no conclusive evidence that a varicocele repair improves spontaneous pregnancy rates, varicocelectomy improves sperm parameters (count and total and progressive motility), reduces sperm DNA damage and seminal oxidative stress, and improves sperm ultramorphology. The various methods of repair are all viable options, but microsurgical repair seems to be associated with better outcomes. # 2011 European Association of Urology. Published by Elsevier B.V. All rights reserved.

Keywords: Varicocele Varicocelectomy Infertility male Sperm count Spermatozoa Oligospermia Sperm motility Pregnancy outcome Semen and sperm

* Corresponding author. PO Box 11777, Jeddah 21463, Saudi Arabia. Tel. +966 2 650 9000x4812; Fax: +966 2 650 9001. E-mail address: [email protected] (A. Baazeem).

0302-2838/$ – see back matter # 2011 European Association of Urology. Published by Elsevier B.V. All rights reserved.

doi:10.1016/j.eururo.2011.06.018

EUROPEAN UROLOGY 60 (2011) 796–808

1.

Introduction

Varicocele is a collection of abnormally dilated, tortuous spermatic veins. A clinical varicocele is found in about 15% of all adult males, up to 35% of men who present for infertility evaluation [1,2], and as many as 81% of men with secondary infertility [3]. When classified according to semen analysis results, 11.7% of infertile men with normal semen analyses and 25.4% of those with abnormal analyses were found to have clinical varicocele [4]. Few topics in urology have been as controversial as the effect of varicocelectomy on male factor infertility. Several nonrandomized studies have suggested that repairing a clinically palpable varicocele in the presence of an abnormal semen analysis results in improvements in semen parameters and pregnancy rates [5,6]. Reports on varicocele and infertility issued by the Practice Committee of the American Society for Reproductive Medicine and the American Urological Association’s Male Infertility Best Practice Policy Committee suggest that for infertile couples in which the female partner has no proven (or a potentially treatable) cause of infertility, repair of the male partner’s varicocele should be considered if it is clinically palpable and he has abnormal semen parameters [7,8]. However, the European Association of Urology’s guidelines on male infertility cautioned that the issue of varicocele treatment with the aim of achieving pregnancy is controversial [9]. Meanwhile, the National Institute for Health and Clinical Excellence’s clinical guideline on fertility states, ‘‘men should not be offered surgery for varicocele as a form of fertility treatment because it does not improve pregnancy rates’’ [10]. Part of the controversy lies in the fact that, despite considerable research, the exact pathophysiologic mechanism by which varicocele can induce male factor infertility is not known with certainty, although several potential causes have been postulated [11–17]. This is further complicated by the fact that most of the evidence suggesting a positive effect for varicocele repair on male fertility potential is based, to a great part, on retrospective or poorly controlled studies. Furthermore, out of just four published randomized controlled trials (RCTs) that assess the effect of repairing a clinical varicocele in men with abnormal semen analyses on the couple’s pregnancy rate, only two showed a benefit [18–21]. Some of these trials might have failed to demonstrate a significant advantage in favor of therapy to a certain extent because of issues with patient selection, debatable classification, and/or questionable semen analysis data before starting therapy [22]. Although the percentage of couples with male factor infertility has remained relatively stable, intracytoplasmic sperm injection utilization has dramatically increased [23]. This information suggests that there might be a trend toward a reduced role for therapy directed toward treating the male partner. If true, doing so might carry the risks of potentially unnecessary and substantially higher costs and transferring the treatment burden to the female partner, with potentially serious complications such as multiple pregnancies and the ovarian hyperstimulation syndrome. Also, in most developing countries, artificial reproductive

797

techniques (ART) are not widely available. There continues to be a need for simple and effective treatments for male factor infertility. We aim to shed light on this topic by reviewing the available high-level studies in the literature in an attempt to answer some of the questions regarding this issue. We also present a new meta-analysis of these studies. 2.

Evidence acquisition

A literature search was performed using Embase (from 1974 to May 2010) and Medline (from 1949 to May 2010) with the aid of an expert external methodologist. The keywords used for the search were varicocele, varicocelectomy, infertilitymale, sperm-count, spermatozoa, oligospermia, spermatogenesis, sperm-motility, fertility, pregnancy-outcome, semen, and sperm. Another search using Medline and the keywords varicocele and surgery was performed by one of the authors. Search limits were set for the English language and human species. The retrieved articles were gathered and examined. Reference lists of retrieved articles as well as relevant review articles were also studied. Meta-analyses as well as randomized and nonrandomized prospective controlled studies (level 1 and level 2a evidence, Oxford Centre for Evidence Based Medicine levels of evidence) [24] were included in the review. Also included were prospective noncontrolled trials. Studies that evaluated the effect of varicocele repair on pregnancy rates, semen parameters, seminal oxidative stress, and sperm DNA damage as well as those that compared different varicocele repair techniques were chosen and included by consensus. After the studies were reviewed, it was noted that none of the previously performed meta-analyses included the most recent RCT that reported on pregnancy rates [18], some of them did not include sclerotherapy or embolization, and some focused primarily on pregnancy rates. Therefore, we performed a new meta-analysis for the purpose of this review. The results of the RCTs evaluating pregnancy rates were used as odds ratios (ORs) and were pooled by metaanalysis [25]. A meta-analysis using weighted mean difference [25] was conducted on repeated measures to estimate the improvement in sperm concentration and motility (total and progressive) after varicocelectomy (only prospective studies that reported the mean plus or minus standard deviation [SD] preoperative and postoperative sperm parameters were included). The Q statistic [25] was used to test between study homogeneity: Homogeneity was rejected when the Q statistic p value was < 0.10. Depending on whether homogeneity was accepted or rejected, we used the fixed or the random effect model to compute the combined OR (for pregnancy data) or mean improvement (for sperm parameters) and the 95% confidence interval (CI). Given the high variability in sperm morphology criteria used in the selected studies, we did not perform a meta-analysis on the change in sperm morphology following varicocelectomy. Forest plots were presented to show (1) the effect of varicocelectomy (vs observation) on pregnancy rates and (2) the mean improvement in sperm parameters after

798

EUROPEAN UROLOGY 60 (2011) 796–808

varicocelectomy. A forest plot contains (1) a vertical solid line representing the null effect; (2) the pooled estimate of the effect (OR or weighted mean difference) as a dashed vertical line with a diamond at the bottom to represent its 95% CI; and (3) individual studies represented as squares with their confidence intervals, the surface of the square being proportional to the weight of the study in the pooled estimate. For each of the last three outcomes, sensitivity analysis was performed to account for the correlation between before and after outcomes measurement [26]. The meta-analysis was conducted using the Stata v.10.0 statistical software (StataCorp, College Station, TX, USA). 3.

Evidence synthesis

3.1.

Does varicocele repair result in an improvement of male

factor infertility?

In 2001, a Cochrane review of the effect of varicocele treatment by surgery or embolization in subfertile men concluded that there was insufficient evidence to support the idea that repairing varicocele in subfertile men improves couples’ spontaneous pregnancy rates [27]. This metaanalysis was criticized because it included studies of men with normal semen parameters and subclinical varicocele. Furthermore, in its updates, it did not include the effect of varicocele repair on semen parameters, a factor that might affect the outcomes of couples’ future attempts at ART use [28–30]. This information prompted the Italian Varicocele Study Group to perform a new meta-analysis based on the same eight studies that were included in the Cochrane review [22]. Five of the eight studies were excluded because they included

men with normal semen parameters and/or subclinical varicocele. An ‘‘as-treated’’ analysis of the three remaining studies revealed a significantly higher pregnancy rate in the treatment group than in controls. As others have, the authors noted the heterogeneity of data in available RCTs as well as the high dropout rates. They concluded that data from these studies cannot result in a good-quality meta-analysis. Several other meta-analyses have also been published, at least partly in response to the Cochrane review. One such study found the odds of spontaneous pregnancy in infertile men with clinical varicocele who underwent surgical repair were significantly higher than in those who received no or medical treatment [31]. It included two RCTs and three observational studies. The OR (2.87) was in favor of therapy. This first integration of observational studies into a metaanalysis was based on the Potsdam criteria for meta-analysis performance and a scoring system aimed at reducing bias [32]. Another meta-analysis from largely the same author group addresses the effect of surgical correction of clinical varicocele on the semen parameters of men with abnormal preoperative analyses [33]. It also included results from RCTs and observational studies and suggested that surgical varicocele repair may result in a significant improvement in sperm concentration, motility, and morphology. As mentioned earlier, we performed a new meta-analysis as part of this review. We identified four RCTs reporting on pregnancy outcome after repair of clinical varicocele in oligospermic men (Fig. 1). These RCTs involved 380 couples (192 randomized to treatment and 188 to observation). The OR ranged from 0.83 to 13.5; in two out of four estimates, these ORs were statistically different from unity (Fig. 1). The fixed-effect model combined OR was 2.10 (95% CI, 1.31–3.38; p = 0.002), suggesting that the effect of varicocelectomy is

[(Fig._1)TD$IG] Study

Odds ratio (95% CI)

Treatment n/N

Control n/N

Abdel-Meguid et al (2011)

24/73

10/72

3.04 (1.33,6.95)

29.7

Krause et al (2002)

5/32

6/33

0.83 (0.23,3.06)

22.3

Nieschlag et al (1998)

18/62

16/63

1.20 (0.55,2.65)

30.3

Madgar et al (1995)

15/25

2/20

13.50 (2.55,71.40)

17.6

Overall (95% CI)

% Weight

2.23 (0.86,5.78)

.014005

1

71.4029 Odds ratio

Fig. 1 – Intention-to-treat analysis: Forest plot depicting odds ratio and 95% confidence interval for pregnancy outcome in the four randomized controlled trials on varicocele repair or observation in men with abnormal semen parameters and clinical varicocele [18–21]. The random effect model was used (heterogeneity x2 = 14.60 [df = 3]; p = 0.024). OR = odds ratio; CI = confidence interval.

799

EUROPEAN UROLOGY 60 (2011) 796–808

[(Fig._2)TD$IG] Study

Odds ratio (95% CI)

Treatment n/N

Control n/N

24/73

10/72

3.04 (1.33,6.95)

30.8

6/20

6/37

2.21 (0.61,8.09)

21.5

Nieschlag et al (1998)

18/62

16/63

1.20 (0.55,2.65)

31.6

Madgar et al (1995)

15/25

2/20

13.50 (2.55,71.40)

16.1

Abdel-Meguid et al (2011) Krause et al (2002)

Overall (95% CI)

% Weight

2.69 (1.16,6.24)

.014005

1

71.4029 Odds ratio

Fig. 2 – ‘‘As-treated’’ analysis: Forest plot depicting odds ratio and 95% confidence interval for pregnancy outcome in the four randomized controlled trials on varicocele repair or observation in men with abnormal semen parameters and clinical varicocele [18–21]. The random effect model was used (heterogeneity x2 = 7.41 [df = 3]; p = 0.060). OR = odds ratio; CI = confidence interval.

statistically superior to observation. However, the Q statistic p value was 0.024, indicating the nonhomogeneity of the studies (heterogeneity x2 = 14.60 [df = 3]; p = 0.024). Using the random effect model, the combined OR was 2.23 (95% CI, 0.86–5.78; p = 0.091), indicating that difference in effect of varicocelectomy compared to observation is not statistically significant (Fig. 1). On reviewing these studies, their heterogeneity and some of their limitations become apparent. For example, the study by Krause et al [21] suffered low recruitment, a high dropout rate, and considerable loss to follow-up. Furthermore, only 125 of 226 patients who fulfilled the study’s criteria completed the Nieschlag study [20]. The Madgar study [19] shows a strikingly high treatment arm pregnancy rate (60%), while most randomized studies show pregnancy rates in the treated group of 25–35% after 1 yr of follow-up. The Abdel-Meguid study [18] has a surprisingly low number of pregnancies in the control group (14%), considering the short duration of the infertility (1.5 yr), the young age of the female partners, and the relatively good sperm parameters. It also had an unusually high recruitment rate and extraordinarily low dropout rate. We also performed an ‘‘as-treated’’ analysis (Fig. 2); like the Ficarra study [22], a difference in favor of repair was found. Intention-to-treat analysis (the results of which were described earlier), however, is generally believed to aid in minimizing bias [34]. We identified 22 prospective studies reporting on sperm concentration before and after clinical varicocele repair (Fig. 3). The mean improvement in sperm concentration in these studies ranged from  4.0 to 60.0  106 sperm/ml; in 15 of 22 studies, the improvement was statistically significant (Fig. 3). The Q statistic p value was