Sexual Function/Infertility

Successful Repeat Microdissection Testicular Sperm Extraction

in Men With Nonobstructive Azoospermia

Ranjith Ramasamy, Joseph A. Ricci, Robert A. Leung and Peter N. Schlegel*From the Department of Urology, New York-Presbyterian Hospital, Weill Cornell Medical College, New York, New York

Purpose: We studied factors that can predict successful repeat microdissectiontesticular sperm extraction in men with nonobstructive azoospermia.Materials and Methods: We retrospectively analyzed the records of 126 menwith nonobstructive azoospermia who underwent 1 successful microdissectiontesticular sperm extraction attempt. Clinical factors identifiable at the secondprocedure, including age, testicular volume, endocrinological data and histology,were analyzed.Results: Overall testicular spermatozoa were successfully retrieved at 103 of 126repeat attempts (82%). Men with a successful repeat attempt had lower follicle-stimulating hormone (mean � SD 23.1 � 12.4 vs 29.2 � 12.8, p � 0.04) and largertesticular volume (mean 10 � 5 vs 7 � 4, p � 0.0001) at the repeat procedurecompared to men with a failed repeat attempt. Adjusted associations from amultiple logistic regression model showed that no factors predicted sperm re-trieval during repeat microdissection testicular sperm extraction. An ROC curveshowed a fair prediction model (AUC � 0.71).Conclusions: The follicle-stimulating hormone level and testicular volume at therepeat attempt appear to have predictive value to determine the success of asecond attempt. These observations are interesting since testicular volume andfollicle-stimulating hormone in men with nonobstructive azoospermia do notpredict sperm retrieval at a primary microdissection testicular sperm extractionattempt.

Key Words: testis; infertility, male; sperm retrieval;

Abbreviations

and Acronyms

FSH � follicle-stimulatinghormone

ICSI � intracytoplasmic sperminjection

micro-TESE � microdissectionTESE

NOA � nonobstructiveazoospermia

TESE � testicular spermextraction

Submitted for publication July 8, 2010.Study received Weill Cornell Medical College

institutional review board approval.* Correspondence: Department of Urology,

525 East 68th St., Starr 900, New York, New York10065 (telephone: 212-746-5491; FAX: 212-746-8425; e-mail: pnschleg@med.cornell.edu).

Klinefelter syndrome; reoperation

MEN with NOA have long been consid-ered infertile due to severely deficientspermatogenesis with inadequate spermproduction to reach the ejaculate.1

However, recent studies suggest thatalmost 60% of men with NOA havesome sperm production in the testes.2

These men are now no longer consid-ered irrevocably infertile since treat-ment options have rapidly progressedin the last several years.2–6 The devel-opment of micro-TESE has allowedphysicians to localize small regions of

sperm production in the testes, serv-

0022-5347/11/1853-1027/0THE JOURNAL OF UROLOGY®

© 2011 by AMERICAN UROLOGICAL ASSOCIATION EDUCATION AND RES

ing as a highly efficacious treatment inmen with NOA.3 After retrieval thesesperm can be combined with ICSI to pro-vide an opportunity for reproduction.4

In a substantial number of patientswith nonobstructive azoospermia, in-cluding those with Klinefelter’s syn-drome, pregnancy and live birth havebeen achieved with the spermatozoaextracted by this procedure.5

Based on the successful fertilitytreatment outcomes in men with NOAreported in the literature, much re-

search in this area has focused on

Vol. 185, 1027-1031, March 2011Printed in U.S.A.

EARCH, INC. DOI:10.1016/j.juro.2010.10.066www.jurology.com 1027

SUCCESSFUL REPEAT MICRODISSECTION TESTICULAR SPERM EXTRACTION1028

identifying factors that predict the success of spermretrieval in patients with NOA who undergo micro-TESE.2,7–14 Although some remain controversial,factors implicated in predicting the success of spermretrieval at a micro-TESE attempt include testicularvolume, hormone levels, testicular histology, chro-mosomal deletions and patient age.2,7–14 With theadvent of micro-TESE men who desire more than 1offspring have been undergoing repeat attempts.However, to our knowledge there has been no studyto date comparing successful and failed repeat mi-cro-TESE attempts. We investigated whether thesuccess or failure of a repeat attempt can be pre-dicted based on clinically identifiable factors mea-sured at the first micro-TESE attempt.

MATERIALS AND METHODS

Patient SelectionWe retrospectively analyzed the records of all 963 consec-utive patients treated between March 1999 and June 2009with nonobstructive azoospermia who underwent micro-TESE. Patients were grouped based on the outcome (suc-cess or failure) of the repeat operation. The 2 groups ofmen were then compared based on clinical factors presentat the repeat operation. Azoospermia in all of these menwas confirmed by analysis of at least 2 centrifuged semenspecimens according to WHO guidelines.15 An additionalsemen sample was confirmed to be azoospermic on the dayof planned micro-TESE.

Karyotype analysis was performed in all patients andthe diagnosis of nonmosaic Klinefelter’s syndrome (47,XXY karyotype) was confirmed by analyzing peripherallymphocytes with at least 50 cells cytogenetically ana-lyzed per patient. Testicular volume was measured byphysical examination using an orchidometer and the av-erage volume of the 2 testes was used for analysis. Also,physical examination was done to detect varicocele. Tes-ticular histology was determined based on the results ofprevious biopsy done elsewhere or on the results of intra-operative testicular exploration during micro-TESE. Hor-monal evaluation included FSH with levels measuredwithin 2 months before the micro-TESE attempt.

Clinical pregnancy in female partners was defined byidentification of a gestational sac with a fetal heartbeat ontransvaginal ultrasound examination 6 weeks after em-bryo transfer. Confirmation of live birth was obtained bytelephone interviews of couples identified with clinicalpregnancy. The study protocol was approved by the WeillCornell Medical College institutional review board.

Micro-TESEMicro-TESE has been described previously.16 Briefly, thistechnique allows an extensive search of multiple areas ofthe testis rather than a limited biopsy sample that mayreflect little of total testicular function. Sperm retrievalsurgery was typically attempted on the day before oocyteretrieval. Briefly, a midline incision was made in the scro-tum and the testis with the spermatic cord was preferen-tially delivered from the side with the larger testis. The

tunica vaginalis was opened and the tunica albuginea was

visualized. Under an operative microscope the tunica al-buginea was widely opened in an equatorial plane aroundapproximately 270 degrees of the testicular circumferencewith preservation of subtunical vessels.

After the tunica albuginea was opened the testicularparenchyma was directly examined at between 12� and18� magnification under an operating microscope. Exam-ination included as much of the testicular parenchyma aspossible until spermatozoa were found. Samples (1 to 15mg) were excised by teasing out larger, more opaque tu-bules from surrounding Leydig cell nodules or hyperplasiain the testicular parenchyma. Excised samples were ex-amined immediately for the presence of the testicularspermatozoa by mechanically disrupting the testicular tis-sue and placing a small droplet of dispersed tissue sus-pension on a glass slide under a phase contrast microscopeat 200� magnification.17 Each sample was examined byan experienced embryologist from the in vitro fertilizationteam. If no spermatozoa were identified in the initialsample, subsequent samples were taken from the sametestis and, if needed, from the contralateral testis. Dissec-tion was done through all regions of testicular tissue ifneeded to find sperm, preserving the centrifugal pattern ofthe testicular blood supply. The procedure was terminatedwhen sperm were documented in testicular tissue, allregions of each testis had been examined with excision ofthe best appearing tubules or further dissection wasthought likely to jeopardize the testicular blood supply.The presence or absence of sperm was documented in theoperating room by examination of wet preparation speci-mens under a phase contrast microscope. After the TESEprocedure the best testicular samples were maintainedovernight in sperm wash medium at 37C. The surgicalprocedure was terminated when spermatozoa were re-trieved. All repeat procedures were done in exactly thesame manner as the initial procedure.

Statistical AnalysisMicrosoft® Excel® 2000, GraphPad® Prism 5® and IBM®SPSS® Statistics 18 were used to perform all statisticalcalculations with p �0.05 considered statistically signifi-cant. The men were classified into 2 groups based on theoutcome of repeat micro-TESE. After the groups wereestablished clinical parameters at the time of the initialprocedure were then evaluated for predictive value. Stu-dent’s unpaired t test was used to compare factors presentat the repeat TESE attempt, such as age, testicular vol-ume, FSH and operative time, between men with success-ful vs failed repeat TESE attempts. Chi-square analysiswas done to compare the presence of Klinefelter’s syn-drome, the presence of varicocele and testicular histologyon biopsy at the first TESE attempt between the 2 groups.Fisher’s exact test rather than chi-square analysis wasused when the value was less than 10 in 1 cell of the 2 �2 contingency table. Multiple logistic regression was usedto assess the association of FSH and testicular volumewith the success of repeat sperm retrieval, adjusting forpotential confounding variables, including the presence ofKlinefelter’s syndrome and patient age. Confounding vari-ables were included if they satisfied the criteria of chang-ing the estimated association between FSH or testicular

volume and sperm retrieval by at least 10%. We performed

SUCCESSFUL REPEAT MICRODISSECTION TESTICULAR SPERM EXTRACTION 1029

ROC curve analysis using a final model with the backwardelimination method that included FSH, testicular volume,Klinefelter’s syndrome and patient age.

RESULTS

Of 963 men who underwent TESE during this period126 had a successful initial micro-TESE attemptand returned for a repeat procedure. Sperm re-trieval was successful in 103 of these repeat at-tempts (82%). Of men who had sperm retrieved dur-ing the repeat attempt clinical pregnancy wasachieved at 54 initial (42%) and 49 repeat (39%)attempts.

Clinically identifiable factors present at repeatmicro-TESE were analyzed for the prediction ofsperm retrieval and stratified by the outcome of therepeat procedure (table 1). Mean FSH was lower inpatients with a successful repeat operation than inpatients in whom no sperm were retrieved at therepeat procedure (mean � SD 23.1 � 12.4 vs 29.2 �12.8, p � 0.04). There was a greater increase in FSHfrom the first to the second micro-TESE attempt inthe group in which no sperm were found but thedifference was not statistically significant (9.1 vs1.9, p � 0.10). Also, mean testicular volume at thetime of initial micro-TESE was larger in the groupwith a successful repeat attempt than in the groupwith a failed repeat attempt (mean 10 � 5 vs 7 � 4,p � 0.0001). There was no difference in the otherpatient characteristics, including age, Klinefelter’ssyndrome, varicocele, interval between the first andthe second operation, and testicular histology pat-tern on biopsy, between the successful and failedgroups. We performed multiple logistic regressionanalysis with serum FSH and testicular volume topredict sperm retrieval during repeat micro-TESE(table 2). Included in the model were Klinefelter’s

Table 1. Successful and failed repeat micro-TESE attempts

Sperm No Sperm p Value

No. attempts 103 23 —Mean � SD age 35 � 7 37 � 10 Not significantMean � SD FSH:

Level (IU/l) 23.1 � 12.4 29.2 � 12.8 0.04Change (IU/l) 1.9 � 15.1 9.1 � 20.2 Not significantBefore 1st/before repeat

micro-TESE1.2 � 0.8 1.3 � 0.7 Not significant

Mean � SD testicular vol (cc) 10 � 5 7 � 4 0.0001No. Klinefelter’s syndrome (%) 12 (12) 4 (17) Not significantNo. varicocele (%) 26 (25) 5 (22) Not significantNo. Sertoli’s cell Only (%) 49 (48) 15 (67) Not significantNo. maturation arrest (%) 19 (18) 5 (22) Not significantNo. hypospermatogenesis (%) 35 (34) 3 (11) Not significantMean � SD time 1st-repeat

attempts (yrs)1.7 � 1.5 2.2 � 1.4 Not significant

Av operative time (mins) 101 127 0.03

syndrome and age. Adjusted associations from themodel showed that the chance of retrieving spermduring repeat micro-TESE cannot be predicted by anyindividual variable. However, a ROC curve based onthe 4 factors included in logistic regression analysisshowed a fair prediction model (AUC � 0.71, seefigure).

DISCUSSION

Micro-TESE combined with ICSI has rapidly be-come a treatment of choice in men with nonobstruc-tive azoospermia who want to father children.2 As aresult of the unprecedented high success of theseprocedures, a dramatic increase in the number ofpatients who desire a repeat attempt has been ob-served. During the first attempt at micro-TESEspermatozoa that are retrieved are primarily usedfor ICSI and any remaining sperm samples are cryo-preserved. For a second attempt at conception se-men analysis is repeated and cryopreserved sperm,if available, are thawed and analyzed for viabilitybefore ICSI. However, only about 30% of samples

Table 2

OR (95% CI) p Value

Age 0.953 (0.922–0.984) 0.115Testicular size 1.123 (1.055–1.191) 0.087FSH 0.980 (0.958–1.002) 0.362Klinefelter’s syndrome 0.734 (�0.108–1.576) 0.713

ROC curve of pertinent preoperative parameters to discriminate

successful and failed repeat micro-TESE (AUC � 0.71).

SUCCESSFUL REPEAT MICRODISSECTION TESTICULAR SPERM EXTRACTION1030

thawed from prior TESE have documented viablesperm for ICSI from a prior cycle.18 Hence, if cryo-preserved sperm are not viable after a thaw cycle,patients undergo repeat micro-TESE.

At this time the prediction of repeat micro-TESEsuccess remains controversial at best since to ourknowledge the predictive factors, if any, have neverbeen studied. Only 1 prior study describes micro-TESE as a technique to salvage failed conventionalTESE.19 While that study demonstrated no signifi-cant difference in sperm retrieval rates between thegroup with salvage micro-TESE after failed conven-tional TESE and the group with only an initial lim-ited set of testis biopsies (2 or less per testis), thestudy did not directly address the issue of predictingthe outcome of repeat micro-TESE procedures.

The results of our univariate analysis indicate asignificant difference in testicular size and meanFSH at the initial procedure between men who hadsuccessful and failed repeat micro-TESE. Severalinvestigators have reported that testicular size isnot a predictor of successful sperm retrieval at aninitial micro-TESE attempt.7,8 However, our analy-sis revealed that larger testicular size is associatedwith increased sperm retrieval at a repeat attempt.This is most likely because men with larger testeshave a larger volume, potentially with more sitesthat can be searched for sperm during the proce-dure, increasing the likelihood of a successful out-come. Furthermore, our results also demonstratethat lower average FSH is associated with men whounderwent a successful repeat operation. This is incontrast to previous results noted by our group inpatients who had nonobstructive azoospermia dur-ing the first attempt at sperm retrieval, when FSHand testicular volume do not affect the chance ofsperm retrieval.11 The chance of sperm being pres-ent in a testis is clearly independent of overall tes-ticular function, as reflected by FSH and testicularvolume. However, prior studies suggested that asubset of men with nonobstructive azoospermia, lowFSH and normal testicular volume, ie those withdiffuse maturation arrest, have a poor prognosis.20

This lower FSH may reflect the larger number ofSertoli’s cells in a larger testis, providing more con-trol feedback to suppress FSH production. Men withlarger testes and lower FSH but sperm found at aprior TESE attempt seemed to have a better chanceof a successful repeat micro-TESE attempt. We be-lieve that this could be due to a greater number oftubules available in which to search for sperm at therepeat attempt.

In the search for factors that can predict success-ful sperm retrieval in patient undergoing repeat mi-

cro-TESE we created a multiple logistic regression

model to investigate factors present at the initialprocedure. We included the variables FSH and tes-ticular volume, given their statistical significance onunivariate analysis, as well as patient age at theinitial procedure and the presence or absence ofKlinefelter’s syndrome, which in prior studies pre-dicted the outcome of initial micro-TESE at-tempts.13,21,22 The apparent inconsistency betweenunivariate and multiple regression analyses proba-bly stems from the issue of collinearity, especiallybetween FSH and testicular volume. Given thehighly significant effect of testicular volume on uni-variate analysis, an interpretation of the results isthat after controlling for this effect FSH no longerhas a statistically significant effect. To achieve themost accurate prediction we used the backwardelimination method to select the best model for mul-tiple logistic regression and ROC curve analysis.Based on ROC curve data we can say that these 4factors fairly predict successful sperm retrieval atrepeat micro-TESE. The other reason that the modelmay not have identified any predictive factors wasthe high success rate of repeat TESE attempts (81%)compared to the number of patients in which theprocedure failed.

The lack of research surrounding repeat micro-TESE procedures limits the information that we canprovide patients regarding the prognosis for success-ful treatment. Given the recent increase in the num-ber of inquiries for repeat procedures based largelyon the success of the initial micro-TESE procedure,this poses a significant challenge for physicians,namely determining which patients are likely tobenefit from a repeat procedure.2 Despite the short-comings of the predictive model described, this studyrepresents initial steps toward defining the chanceof successful sperm retrieval at repeat proceduresand evaluating the whether there exist any factorspredicting successful sperm retrieval at repeat mi-cro-TESE operations. The identification of addi-tional approaches to recognizing which men havesperm preoperatively may help us optimize the abil-ity of these men with severe male factor infertility tofather offspring without resorting to substitutivetreatment, such as donor insemination or adoption.

CONCLUSIONS

Repeat micro-TESE is successful in 81% of attempts.Patients with lower FSH and/or larger testes mayhave a better chance of sperm retrieval, possiblysince they may have more testicular tissue in whichto search for sperm. Hopefully future studies canidentify parameters to predict the characteristics ofpatients who are better candidates for a repeat pro-

cedure.

SUCCESSFUL REPEAT MICRODISSECTION TESTICULAR SPERM EXTRACTION 1031

REFERENCES

1. Irvine DS: Epidemiology and aetiology of maleinfertility. Hum Reprod 1998; 13: 33.

2. Schlegel PN: Non-obstructive azoospermia: a rev-olutionary surgical approach and results. SeminReprod Med 2009; 27: 156.

3. Tournaye H, Staessen C, Liebaers I et al: Testic-ular sperm recovery in nine 47, XXY Klinefelterpatients. Hum Reprod 1996; 11: 1644.

4. Palermo GD, Cohen J and Rosenwaks Z: Intracy-toplasmic sperm injection: a powerful tool toovercome fertilization failure. Fertil Steril 1996;65: 899.

5. Palermo GD, Schlegel PN, Sills ES et al: Birthsafter intracytoplasmic injection of sperm ob-tained by testicular extraction from men withnonmosaic Klinefelter’s syndrome. N Engl J Med1998; 338: 588.

6. Friedler S, Raziel A, Strassburger D et al: Out-come of ICSI using fresh and cryopreserved-thawed testicular spermatozoa in patients withnon-mosaic Klinefelter’s syndrome. Hum Reprod2001; 16: 2616.

7. Tsujimura A: Microdissection testicular spermextraction: prediction, outcome, and complica-tions. Int J Urol 2007; 14: 883.

8. Tournaye H, Verheyen G, Nagy P et al: Are thereany predictive factors for successful testicularsperm recovery in azoospermic patients? Hum

Reprod 1997; 12: 80.

9. Su LM, Palermo GD, Goldstein M et al: Testicularsperm extraction with intracytoplasmic sperm in-jection for nonobstructive azoospermia: testicularhistology can predict success of sperm retrieval.J Urol 1999; 161: 112.

10. Seo JT and Ko WJ: Predictive factors of success-ful testicular sperm recovery in non-obstructiveazoospermia patients. Int J Androl 2001; 24: 306.

11. Ramasamy R, Lin K, Godsen LV et al: High serumFSH levels in men with nonobstructive azoosper-mia does not affect success of microdissectiontesticular sperm extraction. Fertil Steril 2009; 92:590.

12. Jezek D, Knuth UA and Schulze W: Successfultesticular sperm extraction (TESE) in spite of highserum follicle stimulating hormone and azoosper-mia: correlation between testicular morphology,TESE results, semen analysis and serum hormonevalues in 103 infertile men. Hum Reprod 1998;13: 1230.

13. Ramasamy R, Ricci JA, Palermo GD et al: Suc-cessful fertility treatment for Klinefelter’s syn-drome. J Urol 2009; 182: 1108.

14. Madgar I, Dor J, Weissenberg R et al: Prognosticvalue of the clinical and laboratory evaluation inpatients with nonmosaic Klinefelter syndromewho are receiving assisted reproductive therapy.Fertil Steril 2002; 77: 1167.

15. WHO Laboratory Manual for the Examination of

Human Semen and Sperm-Cervical Mucus Inter-

action. Cambridge: Cambridge University Press1999.

16. Ramasamy R, Yagan N and Schlegel PN: Struc-tural and functional changes to the testis afterconventional versus microdissection testicularsperm extraction. Urology 2005; 65: 1190.

17. Ostad M, Liotta D, Ye Z et al: Testicular spermextraction for nonobstructive azoospermia: re-sults of a multibiopsy approach with optimizedtissue dispersion. Urology 1998; 52: 692.

18. Schlegel PN, Liotta D, Hariprashad J et al: Freshtesticular sperm from men with nonobstructiveazoospermia works best for ICSI. Urology 2004;64: 1069.

19. Tsujimura A, Miyagawa Y, Takao T et al: Salvagemicrodissection testicular sperm extraction afterfailed conventional testicular sperm extraction inpatients with nonobstructive azoospermia. J Urol2006; 175: 1446.

20. Hung AJ, King P and Schlegel PN: Uniform tes-ticular maturation arrest: a unique subset of menwith non-obstructive azoospermia. J Urol 2007;178: 608.

21. Emre Bakircioglu M, Erden HF, Kaplancan T et al:Aging may adversely affect testicular sperm re-covery in patients with Klinefelter syndrome.Urology 2006; 68: 1082.

22. Okada H, Goda K, Yamamoto Y et al: Age as alimiting factor for successful sperm retrieval inpatients with nonmosaic Klinefelter’s syndrome.

Fertil Steril 2005; 84: 1662.