International Journal of Urology (2012) 19, 149154 doi: 10.1111/j.1442-2042.2011.02890.

Original Article: Clinical Investigation iju_2890 149..154

Effect of surgical repair on testosterone production in

infertile men with varicocele: A meta-analysis
Fuping Li,1,2 Huanxun Yue,2 Kohei Yamaguchi,1 Keisuke Okada,1 Kei Matsushita,1 Makoto Ando,1
Koji Chiba1 and Masato Fujisawa1
1
Division of Urology, Department of Surgery Related , Faculty of Medicine, Kobe University Graduate School of Medicine, Kobe,
Japan; and 2Human Sperm Bank, West China Second University Hospital of Sichuan University, Chengdu, China

Abbreviations & Acronyms
CI = confidence interval
RCT = randomized
controlled trial
RIA = radioimmunoassay
ROS = reactive oxygen
species
Correspondence: Huanxun Yue
M.D., Human Sperm Bank, West
China Second University
Hospital of Sichuan University,
20 The People of South Road ,
Chengdu 610041, China. Email:
yuehuanx@hotmail.com
Received 1 August 2011;
accepted 6 October 2011.
Online publication 8 November
2011
Objectives: To determine the effect of surgical varicocele repair in improving testicu-
lar Leydig cell function as shown by increased testosterone production.
Methods: Eligible studies were searched in Medline and the Pubmed database, and
cross-referenced as of 31 May 2011 using the terms varicocele, testosterone and
surgery. The database search, quality assessment and data extraction were indepen-
dently carried out by two reviewers. Only studies including patients with testosterone
evaluation before and after surgery were considered for the analysis. A systematic
review and meta-analysis was carried out for continues variables using random effect
models.
Results: Out of 125 studies, a total of nine were selected , including 814 patients. The
combined analysis showed that mean serum testosterone levels after surgical treatment
increased by 97.48 ng/dL (95% condence interval 43.73151.22, P = 0.0004) compared
with preoperative levels.
Conclusions: Surgical treatment of varicocele signicantly increases testosterone
production and improves testicular Leydig cell function.
Key words: Leydig cell, meta-analysis, testosterone, varicocele, varicocelectomy.

Introduction
Varicocele is now accepted as an important etiological factor in male infertility and has
been determined to be the most common cause.1 It can adversely affect testicular function
in a variety of ways.2,3 Treatment of varicocele by surgical repair can improve the poten-
tial of fertility.4 However, the mechanisms of impairment by varicocele and treatment by
surgical repair are not well understood , and various theories have been suggested , but
none of them are conclusive.5 In addition to effects on spermatogenesis, growing evidence
documents that varicocele also impairs testicular Leydig cell function with a consequen-
tial decrease in testosterone production.68 How varicocele and its repair can affect tes-
ticular Leydig cell function and result in changes in testosterone production are not
clearly identified. It is a question that has been the intensely debated for decades. Many
studies report that varicocelectomy promotes Leydig cells with increased testosterone
after surgery, but other studies have suggested no benefit. Recently, Tanrikut C et al.
reported that data are still conflicting, although the evidence shows an adverse effect of
varicocele on testosterone production and the possible benefit of varicocelectomy.9 To
resolve this issue, we carried out a meta-analysis to assess the effect of surgical repair of
varicocele on testosterone production. Meta-analysis is a useful statistical method that
statistically combines results from several studies to produce a single estimate of a treat-
ment effect or size of an association, considering variations in characteristics that can
influence the overall estimate of the outcome of interest.10 Our aim was to identify the
potential mechanism of varicocele and its repair on Leydig cell function and improvement
of testosterone production.

2011 The Japanese Urological Association 149

F LI ET AL.
Methods
Eligible articles had to meet the following criteria: studies
should describe patients who underwent varicocelectomy
for unilateral or bilateral palpable varicocele. There were at
least two evaluations of serum testosterone before and after
surgery, respectively. The period of follow up should be
longer than 1 month. The value of testosterone should be
determined by the same methods, such as radioimmunoas-
say (RIA), for the possible combining of data. The patients
served as their own controls, with the effects compared over
time. The outcome measure was the effect of surgical vari-
cocelectomy on serum testosterone after surgery. Patients
age, varicocele grade and surgical management were not
considered as factors for exclusion. We did not include the
studies evaluating treatment in adolescents with varicocele.
Studies included were original patient series and studies
containing duplicate material were excluded. Studies
included not only randomized controlled trials (RCT), but
also observational studies including controlled clinical trials
and comparative studies, in which risk factors in research
generally cannot be randomized because they relate to inher-
ent human characteristics or practices, and exposing sub-
jects to harmful risk factors is unethical.11 These criteria
complied with the recommendation of broad inclusiveness
for meta-analysis of observational studies. To adjust and
quantify various potential biases, a special checklist was
made according to a new scoring system.12
A literature search was carried out for all publications up
until 31 May 2011 in Medline and the Pubmed database. A
cross-reference search of the eligible articles was carried out
to identify additional studies not found in the computerized
search. The following key words were used to search the
databases: varicocele, testosterone and surgery. The limit
was humans. No exclusions were made on the basis of
language. The strategy for study identification and data
extraction is outlined in Figure 1.
After identifying the potential relevant studies through
searching the literature, the abstracts of these studies were
read to decide whether the studies were eligible for inclusion
in the meta-analysis. The full article was retrieved when the
information in the abstract appeared to meet the inclusion
criteria. The abstract and the full text of all potentially rel-
evant studies were independently evaluated and selected by
two reviewers (LF and YH), and the final decision was
determined by a discussion between the two reviewers.
The two reviewers independently assessed the selected
studies, and tabulated and extracted data on predefined
sheets. Studies were evaluated methodologically for hetero-
geneity. The results of continuous variables were reported as
means and standard deviations used to calculate a weighted
mean difference. Statistical significance was set at a P-value
less than 0.05. Statistical analysis was carried out using the
Review Manager Software (Version 5.1.1, the Cochrane Col-
150
Potential relevant studies
through literature search
(n = 125)
Articles considered not
relevant based on title
and abstract(n = 39)
Studies retrieved for more
detailed information
(n = 86)
Articles excluded on
study design and
outcome measures
Potentially appropriate (n = 69)
studies to be included in
the review (n = 17)
6 subgroup data, 2
special patient population
(n = 8)
Studies ultimately
included in the review
(n = 9)
Fig. 1 Flow diagram of selection of studies.
laboration, Oxford , UK). Statistical heterogeneity of studies
was investigated using I-squared statistics. The fixed and the
random effect will be used respectively when the I-squared is
less than and more than 30% in the meta-analysis.13
Results
A total of 125 potentially relevant studies were found
through combined research. After screening all titles and
abstracts, 39 articles were rejected because the focus was on
a different topic or the article did not report original
research, but a commentary. Then 86 studies were selected
for further review. According to the inclusion and exclusion
criteria, another 69 articles had to be excluded. After an
independent review, 17 articles fulfilled the inclusion crite-
ria. Six more studies had to be excluded from review,
because the testosterone data only showed a different sub-
group of patients with varicocele (age, surgical treatment).
Statistics bias can be produced by combining them directly.
Another two articles were subsequently excluded as a result
of the populations in the studies being special patients, such
as patients with persistent and recurrent varicocele. At the
conclusion of the evaluation period , just nine studies were
included in the meta-analysis. Among them was one pro-
spective study,14 and the other eight were observational
clinical studies. There is a paucity of studies looking spe-
cifically at these related topics. Testosterone data were
extracted from these studies in which the data were
mentioned as a secondary data point with other primary
objectives.
The summary of the methodology and intervention char-
acteristics in the studies is shown in Table 1. The nine
studies included in this systematic review involved 814
2011 The Japanese Urological Association
 Varicocelectomy and testosterone production

Table 1 Summary of characteristics of included studies

Author, year Study type Age (years) No. Varicocele Follow up Surgical approach
patients grade I, II, III (months)
Su et al. (1995)19 Retrospective 35 (2257) 53 13/53 (I) Up to 3 Inguinal or subinguinal
22/53 (II) varicocelectomy
18/53 (III)
Cayan et al. (1999)7 Retrospective 29.5 (2140) 78 NA Up to 12 Subinguinal
varicocelectomy
Pierik et al. (2001)23 Retrospective 35.5 (2941) 30 NA Up to 3 Palomo
Fujisawa et al. (2001)22 Retrospective 32.5 (2442) 52 13/52 (I) Up to 6 Subinguinal
19/52 (II) varicocelectomy
20/52 (III)
Lee et al. (2007)21 Retrospective 39.6 (NA) 12 NA Up to 3 Subinguinal
varicocelectomy
Di Bisceglie et al. Retrospective 28 (1643) 38 NA Up to 6 Sclerotherapy
(2007)24
Ozden et al. (2008)14 Prospective 24 (NA) 30 10/30 (II) Up to 6 Subinguinal
20/30 (III) varicocelectomy
Rodriguez et al. (2009)20 Retrospective 23.5 (1535) 202 142/202 (II) Up to 6 Inguinal varicocelectomy
60/202 (III)
Tanrikut et al. (2011)16 Retrospective 35 (1870) 325 NA 312 Subinguinal
varicocelectomy

NA, not available.

patients who underwent varicocelectomy and follow up for
at least 3 months. The number of patients included in each of
the nine studies was relatively small, varying between 12
and 325 patients. Patients (mean age 31.4 years) were diag-
nosed clinically by palpable varicocele and underwent sur-
gical correction. All men had serum testosterone levels
measured both before and after surgery. The unit of
testosterone is expressed as ng/dL and different units of
measurement were converted to ng/dL. After surgical vari-
cocelectomy, mean serum testosterone was significantly
increased by 97.48 ng/dL (95% CI 43.73151.22,
P = 0.0004) compared with preoperative levels. The random
effect was applied in the statistic, as I-squared was more
than 30% (Fig. 2).
Discussion
Growing evidence has shown that the pathophysiology of
varicocele might partially be involved in dysfunction of the
hypothalamicpituitarygonadal axis that occurs at the level
of Leydig cells and consequently leads to hormonal dys-
function. In patients with idiopathic varicocele, testicular
biopsy showed a decreased tubular diameter, hyperplasia
in the number of Leydig cells (in which many showed
cytoplasmic vacuolization) and atrophy. The number of
testosterone-positive Leydig cells also decreased in testicu-
lar tissue.15 The finding of reduced serum testosterone con-
centrations in infertile men with varicoceles has also led to
the hypothesis that varicoceles result in Leydig cell dysfunc-
tion and subsequently to decreased testosterone produc-
tion.8,16 Tanrikut et al. noted that patients with varicoceles
had significantly lower testosterone levels than a control
group without varicoceles.17 Hurtado et al. found that test-
osterone production decreased significantly in varicocele
patients compared with age-paired fertile men.18 Taken
together, altered testosterone might be the result of a defect
in the testes of men with varicocele. However, it is unclear if
altered testosterone is the cause or the effect of decreased
Leydig cell function.
The relationship between varicocele and impaired test-
osterone production is less well established. Furthermore,
the effect of surgical repair of varicocele to promote Leydig
cell function and testosterone production remains controver-
sial. Many studies reported that serum testosterone levels
were improved after surgical varicocelectomy.7,15,19 Cayan
et al. retrospectively evaluated 78 infertile patients who
underwent microsurgical inguinal varicocelectomy and
found that the mean serum testosterone levels increased
significantly.7 Su et al. found a significant change in serum
testosterone levels, especially in patients with low preopera-
tive values.19 Recently, Tanrikut et al. showed that 325 men
with varicocele had lower testosterone levels compared
with a control group without varicocele across age catego-
ries and this increased significantly after microsurgical

2011 The Japanese Urological Association 151

F LI ET AL.

Pre-operation Post-operation Mean difference Mean difference

Study or subgroup Mean SD Total Mean SD Total Weight IV, Random, 95% CI IV, Random, 95% CI
Su et al. 199519 319 12 53 409 23 53 13.0% -90.00 [-96.98, -83.02]
Cayan et al. 19997 563 140 78 837 220 78 11.3% -274.00 [-331.87, -216.13]
22
Fujisawa et al. 2001 460 160 52 470 190 52 10.8% -10.00 [-77.51, 57.51]
Pierik et al. 200123 542 31 30 571 25 30 12.9% -29.00 [-43.25, -14.75]
Di Bisceglie et al. 200724 650 50 38 660 50 38 12.7% -10.00 [-32.48, 12.48]
Lee et al. 200721 348 202 12 416 358 12 3.8% -68.00 [-300.57, 164.57]
Ozden et al. 200814 660 130 24 720 130 24 10.5% -60.00 [-133.55, 13.55]
Rodriguez et al. 200920 648 156 202 709 232 202 12.2% -61.00 [-99.55, -22.45]
Tanrikut et al. 201116 200 126 325 454 168 325 12.7% -254.00 [-276.83, -231.17]

Total (95% CI) 814 814 100.0% -97.48 [-151.22, -43.73]

Heterogeneity: Tau = 5770.38; Chi = 362.13, df = 8 (P < 0.00001); I = 98%
-200 0 100 200
Test for overall effect: Z = 3.55 (P = 0.0004)
Pre-operation Post-operation

Fig. 2 Meta-analysis of testosterone pre- and postoperation. CI, condence interval.

varicocelectomy.16 They also proposed that the mean
improvement in testosterone >150 ng/dL would be consid-
ered as clinically significant in those patients who showed a
postoperative testosterone increase. Contrary to the findings
of those studies, other studies have shown no significant
difference in testosterone production in varicocele patients
after surgery.14,2024 In two recent studies, the indicators
documented conflicting outcomes. Hsiao et al. found micro-
surgical varicocelectomy resulted in a significant increase in
testosterone in all age groups including a total of 272 men
divided into three groups based on age at surgery (less than
30 years, 3039 years and 40 years or greater).25 However,
Resorlu et al. noted that no significant change in testoster-
one levels observed between each group, which were cat-
egorized into three age groups (1825 years, 2635 years
and older than 36 years).26 These two studies were not
included in our meta-analysis, because combining subgroup
data might lead to statistics bias. Taken together, these data
assessed in the present meta-analysis showed postoperative
mean serum testosterone was significantly higher than pre-
operative levels. The result of this meta-analysis showed the
benefit of surgical varicocelectomy as a means to improve
serum testosterone levels and to partially restore Leydig cell
function.
Because the basic function of Leydig cells is testosterone
biosynthesis and secretion, testosterone levels depend on the
number and function of Leydig cells. As we have shown,
serum testosterone levels decrease in patients with varico-
cele and can be improved after surgical treatment. However,
the exact mechanism of effects of varicocele on Leydig cells
and testosterone production are poorly understood so far.
Several hypotheses have been suggested to explain the del-
eterious effect of varicocele on testis function. The most
accepted hypothesis is related to alterations in the testicular
thermal environment. The spermatic veins leaving the tes-
ticles form a communicating meshwork of veins that pro-
duces a counter-current heat-exchange mechanism to cool
the arterial blood.27 This mechanism is eliminated in patients
with varicocele, causing elevated scrotal temperatures.
Furthermore, spermatic enzyme activity, such as 17
a-hydroxyprogesteronealdolase, is inhibited at higher tem-
peratures, resulting in decreased testosterone production.28
This can be reversed after surgery, improving the Leydig cell
environment for enzyme activity, thus leading to a postop-
erative increase in serum testosterone levels as shown in the
present meta-analysis. Cayan et al. also reported that in the
testosterone increase group, sperm count and motility
improved significantly after the removal of the hyperthermic
effects of the varicocele by surgical repair. This finding
suggested that recovery of testosterone synthesis had the
advantage of epididymal function and spermatogenesis in
males with varicocele.7
Another recent possible theory has greatly emphasized
the role of oxidative stress (OS) on the pathogenesis of
infertility patients with varicocele. A previous work in our
laboratory showed that the expression of 8-hydroxy-2-
deoxyguanosine (8-OHdG) in the seminiferous tubules was
more prevalent in patients with varicocele.29 The release of
reactive oxygen species (ROS) adversely affected testicular
function, including Leydig cells. ROS-mediated perturba-
tion of Leydig cell mitochondria resulting in inhibition of
StAR protein expression and function might well be an
important contributing factor to the decrease in testosterone
associated with varicocele.30
Recently, a direct impact on Leydig cell function as
reflected by changes in intratesticular testosterone in an
experimental rat model of surgically-induced varicocele was
observed in two studies.6,31 This impairment appeared to be

152 2011 The Japanese Urological Association


fully corrected by surgical repair. Luo et al. also showed that
impairment of testis associated with varicocele was bilateral
and gradual, and the level of intratesticular testosterone
declined earlier than the level of serum testosterone.6 The
time difference of testosterone changes between intra-
testicular and periphery condition might be one of the
reasons for the contradictory outcomes reported by many
investigators.
In the current review, most articles included are observa-
tional studies, which could be controversial. Meta-analysis
has been applied with increasing frequency to RCT, which
are considered to provide the strongest evidence regarding
an intervention. In fact, because studies of risk factors gen-
erally relate to inherent human characteristics or medical
practices and cannot be randomized ,11 and RCT design are
not feasible in many clinical trials, only data from clinical
observational studies are available. It is also recommended
by Postdam Consultation that observational studies are
accepted if the methodological designs of these studies are
critically evaluation.32 The strategy in the current review was
to compare the effect of change over time among patients
undergoing surgery, not to compare the patients against the
no treatment controls.
Conclusions
This meta-analysis shows that varicocele causes a distur-
bance of Leydig cell function resulting in decreased test-
osterone biosynthesis. Surgical repair significantly increases
testosterone levels in men with varicocele. As a result of the
limitation of clinical observation and potential publishing
bias, further controlled studies are warranted to more clearly
define this clinical problem and animal studies are required
to investigate the mechanism of varicocele affecting test-
osterone production.
Conict of interest
None declared.
References
1 World Health Organization. The influence of varicocele on
parameters of fertility in a large group of men presenting in
infertility clinics. Fertil. Steril. 1992; 57: 128993.
2 Gorelick JI, Goldstein M. Loss of fertility in men with
varicocele. Fertil. Steril. 1993; 59: 61316.
3 Jarow JP, Coburn M, Sigman M. Incidence of varicoceles
in men with primary and secondary infertility. Urology
1996; 47: 736.
4 Marmar JL, Agarwal A, Prabakaran S et al. Reassessing the
value of varicocelectomy as a treatment for male
subfertility with a new meta-analysis. Fertil. Steril. 2007;
88: 63948.
2011 The Japanese Urological Association
Varicocelectomy and testosterone production
5 Jarow JP. Effects of varicocele on male fertility.
Hum. Reprod. Update 2001; 7: 5964.
6 Luo DY, Yang G, Liu JJ, Yang YR, Dong Q. Effects of
varicocele on testosterone, apoptosis and expression of
StAR mRNA in rat Leydig cells. Asian J. Androl. 2011; 13:
28791.
7 Cayan S, Kadioglu A, Orhan I, Kandirali E, Tefekli A,
Tellaloglu S. The effect of microsurgical varicocelectomy
on serum follicle stimulating hormone, testosterone and
free testosterone levels in infertile men with varicocele.
BJU Int. 1999; 84: 10469.
8 And S, Giacchetto C, Colpi G et al. Physiopathologic
aspects of Leydig cell function in varicocele patients.
J. Androl. 1984; 5: 16370.
9 Tanrikut C, Goldstein M. Varicocele repair for treatment of
androgen deficiency. Curr. Opin. Urol. 2010; 20: 5002.
10 Zwahlen M, Renehan A, Egger M. Meta-analysis in
medical research: potentials and limitations. Urol. Oncol.
2008; 26: 3209.
11 Stroup DF, Berlin JA, Morton SC et al. Meta-analysis of
observational studies in epidemiology: a proposal for
reporting. Meta-analysis Of Observational Studies in
Epidemiology (MOOSE) group. JAMA 2000; 283:
200812.
12 Agarwal A, Deepinder F, Cocuzza M et al. Efficacy of
varicocelectomy in improving semen parameters: new
meta-analytical approach. Urology 2007; 70: 5328.
13 Huser W, Bernardy K, Ueyler N. Treatment of
fibromyalgia syndrome with antidepressants: a
meta-analysis. JAMA 2009; 301: 198209.
14 Ozden C, Ozdal OL, Bulut S, Guzel O, Koyuncu HH,
Memis A. Effect of varicocelectomy on serum inhibin B
levels in infertile patients with varicocele. Scand. J. Urol.
Nephrol. 2008; 42: 4413.
15 Sirvent JJ, Bernat R, Navarro MA, Rodriguez Tolra J,
Guspi R, Bosch R. Leydig cell in idiopathic varicocele.
Eur. Urol. 1990; 17: 25761.
16 Tanrikut C, Goldstein M, Rosoff JS, Lee RK, Nelson CJ,
Mulhall JP. Varicocele as a risk factor for androgen
deficiency and effect of repair. BJU Int. 2011; 108:
14804.
17 Tanrikut C, Choi JM, Lee RK. Varicocele is a risk factor
for androgen deficiency. Fertil. Steril. 2007; 88 (Supp 1):
S386.
18 Hurtado de Catalfo GE, Ranieri-Casilla A, Marra FA, de
Alaniz MJ, Marra CA. Oxidative stress biomarkers and
hormonal profile in human patients undergoing
varicocelectomy. Int. J. Androl. 2007; 30: 51930.
19 Su LM, Goldstein M, Schlegel PN. The effect of
varicocelectomy on serum testosterone levels in infertile
men with varicoceles. J. Urol. 1995; 154: 17525.
20 Rodriguez Pea M, Alescio L, Russell A, Lourenco da
Cunha J, Alzu G, Bardoneschi E. Predictors of improved
seminal parameters and fertility after varicocele repair in
young adults. Andrologia 2009; 41: 27781.
21 Lee RK, Li PS, Goldstein M. Simultaneous vasectomy and
varicocelectomy: indications and technique. Urology 2007;
70: 3625.
153
F LI ET AL.
22 Fujisawa M, Dobashi M, Yamasaki T et al. Significance of
serum inhibin B concentration for evaluating improvement
in spermatogenesis after varicocelectomy. Hum. Reprod.
2001; 16: 19459.
23 Pierik FH, Abdesselam SA, Vreeburg JT, Dohle GR, De
Jong FH, Weber RF. Increased serum inhibin B levels after
varicocele treatment. Clin. Endocrinol. (Oxf.) 2001; 54:
77580.
24 Di Bisceglie C, Bertagna A, Baldi M et al. Varicocele
sclerotherapy improves serum inhibin B levels and seminal
parameters. Int. J. Androl. 2007; 30: 5316.
25 Hsiao W, Rosoff JS, Pale JR, Greenwood EA, Goldstein M.
Older age is associated with similar improvements in
semen parameters and testosterone after subinguinal
microsurgical varicocelectomy. J. Urol. 2011; 185: 6205.
26 Resorlu B, Kara C, Sahin E, Unsal A. The significance of
age on success of surgery for patients with varicocele. Int.
Urol. Nephrol. 2010; 42: 3516.
27 Goldstein M, Eid JF. Elevation of intratesticular and scrotal
skin surface temperature in men with varicocele. J. Urol.
1989; 142: 7435.
154
28 Ando S, Giacchetto C, Colpi G. Plasma levels of
17-OH-progesterone and testosterone in patients with
varicoceles. Acta Endocrinol. (Copenh.) 1983; 102:
4639.
29 Ishikawa T, Fujioka H, Ishimura T, Takenaka A, Fujisawa
M. Increased testicular 8-hydroxy-2-deoxyguanosine in
patients with varicocele. BJU Int. 2007; 100: 8636.
30 Diemer T, Allen JA, Hales KH, Hales DB. Reactive
oxygen disrupts mitochondria in MA-10 tumor Leydig
cells and inhibits steroidogenic acute regulatory (StAR)
protein and steroidogenesis. Endocrinology 2003; 144:
288291.
31 Zheng YQ, Zhang XB, Zhou JQ, Cheng F, Rao T, Yao Y.
The effects of artery-ligating and artery-preserving
varicocelectomy on the ipsilateral testes in rats. Urology
2008; 72: 117984.
32 Cook DJ, Sackett DL, Spitzer WO. Methodologic
guidelines for systematic reviews of randomized control
trials in health care from the Potsdam Consultation on
Meta-Analysis. J. Clin. Epidemiol. 1995; 48: 16771.
2011 The Japanese Urological Association