Extended sperm preparation: an alternative to testicular sperm extraction in non-obstructive azoospermia

Transcription

1 Human Reproduction vol.12 no.6 pp , 1997 Extended sperm preparation: an alternative to testicular sperm extraction in non-obstructive azoospermia R.Ron-El 1, D.Strassburger, S.Friedler, D.Komarovski, O.Bern, Y.Soffer and A.Raziel IVF and Infertility Unit, Assaf Harofeh Medical Center, Zerifin 70300, Israel 1 To whom correspondence should be addressed Testicular sperm retrieval for the treatment of nonobstructive azoospermia requires the execution of an invas- ive procedure, with all its possible attending complications and subsequent long-term effects. This study suggests a new non-invasive approach for collection of spermatozoa in these patients: the extended sperm preparation (ESP). ESP consists of conducting a thorough microscopic search through many droplets of ejaculate sediment. ESP was performed for 49 patients; in 17 patients (35%), spermatozoa were found and subsequently used in intracytoplasmic sperm injection (ICSI). Of these preparations, five yielded fewer motile spermatozoa than the number of corresponding oocytes available, and in one patient only non-motile spermatozoa were recovered. The remaining 32 ESP-negative patients underwent testicular sperm extraction (TESE) from testicular biopsy. Spermatozoa were found in 16 of 32 biopsies (50%) and subsequently used in ICSI. Fertilization and cleavage rates were comparable in both ESP and TESE groups, yielding four clinical pregnancies in each group (27 and 29% respectively). Embryo morphology was defined as excellent in significantly more cases in the ESP group than the TESE group, and implantation rate appeared somewhat higher in the ESP group (16%) than the TESE group (13%). The ESP technique yields results similar to TESE, and can be applied in cases of nonobstructive azoospermia as a prerequisite modality enabling us to avoid testicular biopsy in 35% of cases. Key words: azoospermia/intracytoplasmic sperm injection/testicular spermatozoa Introduction The introduction of intracytoplasmic sperm injection (ICSI) (Palermo et al., 1992; Van Steirteghem et al., 1993a,b) was originally considered for couples with male factor infertility when spermatozoa could be found in the ejaculate. Epididymal spermatozoa (Tournaye et al., 1994) and testicular spermatozoa (Craft et al., 1993; Schoysman et al., 1993) are used for ICSI in patients with obstructive azoospermia. Men with impaired sperm quality and those with obstructive azoospermia have relatively good chances of becoming fathers with the help of the recent achievements in assisted reproductive technologies. However, patients with non-obstructive azoospermia due to primary testicular failure are still the most difficult group to treat and have a relatively poor prognosis. Recent advances with the technique of testicular sperm extraction (TESE) in these individuals and oocyte injection of the spermatozoa retrieved from the testes have resulted in pregnancies (Devroey et al., 1995; Tournaye et al., 1996). Occasional examination of the ejaculate of the patients with non-obstructive azoospermia who were programmed for TESE by testicular biopsy surprised us when occasional sperm cells were found after meticulous microscopic investigation, some of the sperm cells being motile. The same examination in patients with obstructive azoospermia revealed absolutely no spermatozoa. Thereafter, an extended sperm preparation (ESP) was performed on the ejaculate in all cases with non-obstructive azoospermia, consisting of a meticulous search for sperm cells as a standard procedure prior to any surgical intervention. In this study we investigated the frequency of the presence of spermatozoa in ESP preparations and patient characteristics among our patients with non-obstructive azoospermia. We also compared the outcome of the ICSI procedure performed in these patients with the outcome of the ICSI performed with spermatozoa retrieved from the testes in those men when no spermatozoa could be seen in their ESP. Materials and methods From November 1995 until June 1996, 49 patients with non- obstructive azoospermia entered our in-vitro fertilization (IVF) pro- gramme. Non-obstructive azoospermia was diagnosed according to the following clinical criteria: no spermatozoa were observed in several ejaculates or their centrifuged specimens; the vas deferens could be palpated on both sides; the existence of seminal vesicles and prostate without engorgement was demonstrated by transrectal sonography; and the absence of any history of genital infection, surgery or vasectomy was confirmed. The follicle stimulating hormone (FSH) concentration was not incorporated as a diagnostic criterion in view of the study by Hauser et al. (1995). All patients underwent chromosomal analysis which, was normal, and sonography of the testes to rule out any macroscopic pathology. The mean age of the patients was 32.7 years (range 22 45). On the morning of the programmed oocyte retrieval in his partner, the patient was asked to produce semen. The liquefied ejaculate was prepared as described below and carefully examined. In the cases where spermatozoa were found the programmed testicular biopsies were not performed. In those patients whose ESP revealed no spermatozoa, testicular biopsies were taken and elaborated for sperm searching. In these patients a small specimen of the removed tissue was sent to histology. Classification of the histological findings was done according to Levin (1979) European Society for Human Reproduction and Embryology

2 ESP versus TESE in non-azoospermia Extended sperm preparation 3.75 mg microcapsules (Ferring, Malmö, Sweden) and human menopausal The patient was asked to abstain from intercourse for at least 1 week gonadotrophin (HMG, Pergonal; Teva, Israel) as described before the morning of oocyte retrieval, when he produced semen. previously (Ron-El et al., 1991). Human chorionic gonadotrophin After liquefaction a droplet was put in a Makler chamber whereby (HCG, Chorigon; Teva, Israel) was administered when the leading azoospermia was confirmed. The semen was then washed twice by follicle reached mm diameter. Oocytes were retrieved h centrifugation at 350 g for 20 min. The supernatant was discarded, later. The cumulus corona cells were removed after exposure to the pellet gently resuspended and again a droplet was aspirated and hyaluronidase (Hyaluronidase, type IV-S; Sigma) 80 IU/ml in human placed in the Makler chamber to confirm that no spermatozoa were tubal fluid medium (HTF; Irvine Scientific) for s. The oocytes visible. Aliquots of 4 6 µl were aspirated from the pellet and spread were observed for nuclear maturation and only metaphase II eggs in droplets, each of 10 µl flushing medium (Earle s balanced were injected. salt solution; Gibco BRL Life Technologies, Paisley, UK). A collecting pipette with a 10 µm opening with no spike (Perry Medical Instru- ICSI procedure mental Development Ltd, Raanana, Israel) was mounted. Once motile ICSI was carried out according to the methodology described by Van or immotile spermatozoa were visible under an inverted microscope Steirteghem et al. (1993a,b) on an inverted microscope (Diaphot, (Diaphot 300; Nikon Corp., Tokyo, Japan) at a 200 or 400 Nikon Corp.) at 400 magnification using the Hoffman modulation magnification, the cell was aspirated and put in the central droplet contrast system (Modulation Optics Inc., Greenvale, NY, USA) with polyvinylpyrrolidone (PVP; Irvine Scientific, Santa Ana, CA, equipped with two coarse positioning manipulators and with two USA). The motile and immotile spermatozoa were separately located three-dimensional hydraulic micromanipulators (Narishige, Tokyo, in the PVP droplet. The oocytes were inserted into the droplets with Japan). Further culture of the injected oocytes took place in 25 µl the flushing medium, the ICSI procedure could be initiated, and droplets of Irvine HTF medium (Irvine Scientific #9962) under testicular biopsy in the patient was cancelled. lightweight paraffin oil (embryo-tested) (Sigma, ). In cases where no spermatozoa were visible, the patient was taken Fertilization was confirmed after h if two distinct pronuclei to the operating theatre for TESE. In the cases where sporadically were observed under the inverted microscope. Cleavage was assessed immotile spermatozoa could be seen after examining many droplets 24 h later and the embryos were classified according to their and the overall screening of the droplets yielded fewer than five morphological appearance: grade I: symmetrical blastomeres with no immotile spermatozoa, the patient was also designated to undergo anucleated fragments; grade II: asymmetrical blastomeres with or testicular biopsy. Moreover, in such cases the spermatozoa had an without 20% of the volume filled with fragments; grade III: with indistinct outline and the manipulation of that spermatozoon by the 20 50% of the volume covered with fragments; and grade IV: pipette often caused detachment of the head from the tail. The average extremely asymmetrical blastomeres and 50% of the volume filled time needed to complete an ESP procedure was ~45 min. with anucleated fragments. Embryos classed I III were considered for transfer and were introduced into the uterine cavity h after Testicular sperm extraction the ICSI procedure. TESE was performed under general anaesthesia. The scrotal skin and Luteal supplementation consisted of either HCG administration, tunica vaginalis were opened, a 0.5 cm incision was made in the 2500 IU every third day up to four times, or progesterone in oil tunica albuginea and one or two pieces of the extruding testicular (Gestone, Paines & Byrne, Surrey, UK) 50 mg per day i.m. tissue were excised using a pair of curved scissors. The specimen Pregnancy rate was calculated considering clinical pregnancies was then transferred into a Petri dish (Falcon; Becton-Dickinson, only, determined by visualization of at least one gestational sac by Aalst, Belgium) filled with ~2 ml modified HEPES-buffered Earle s transvaginal ultrasound 4 weeks after embryo transfer. medium and heparin 0.4% (H 3149; Sigma, St Louis, MO, USA). Statistical analysis for fertilization and cleavage rates was performed The biopsy specimen was shredded into small pieces with two sterile on the percentage values of the variables within each cycle. Mean glass microscope slides under a stereomicroscope to release the number of blastomeres, mean degree of the embryological morphology spermatozoa from the seminiferous tubules into the medium. The and number of embryos replaced were also expressed as the mean of presence of spermatozoa was assessed using the inverted microscope. the values of the variables within each cycle. Global significance When spermatozoa could be identified among debris, immature cells, was tested by paired comparison of t-test. The comparison of the red blood cells or Sertoli cells no further testicular incision was made. pregnancies in the different groups was compared by means of the In the case where no spermatozoa were visible, up to three biopsies χ 2 -test. When applicable Fisher s exact test was used. P 5% was were performed on different areas of each testis. The shredded biopsy considered statistically significant for all tests. tissue was gently homogenized by repeated aspiration and spillingout manoeuvre into a tuberculin syringe. The effluent and the minced tissue were transferred into a Falcon tube (Becton-Dickinson) and Results incubated for 2 3 h in 5% CO 2 in air at 37 C. The fluid was then ESP yielded available spermatozoa (ESP ), either motile or aspirated, put into another tube and centrifuged at 300 g for 5 min. non-motile, in 17 (35%) of the 49 patients with non-obstructive The supernatant was removed, the pellet gently resuspended. Droplets azoospermia enabling progress to the ICSI procedure and the were aspirated from the suspension as described in the ESP for cancellation of TESE by testicular open biopsy. The age of searching of spermatozoa. this group of patients (ESP ) was years (mean Female patients and oocyte preparation SD) ranging from 22 to 42. In the ESP of the remaining 32 male patients no motile The mean age of the female partners was 28.3 years (range 21 41). The mean age of the females in whose male partners ESP spermatozoa spermatozoa were observed, or occasionally fewer than five could be seen was comparable to the mean age of the those whose non-motile spermatozoa could be seen in all the examined partners underwent TESE (26.6 years, range 22 39; 28.5, range 22 droplets. Subsequently, they underwent testicular biopsies for 41 respectively). All were hyperstimulated using the combination of TESE. Their mean age was comparable to the age of the gonadotrophin-releasing agonist (GnRHa) suppression, Decapeptyl ESP patients, namely, years (range 23 45). 1223

3 R.Ron-El et al. Table I. Follicle stimulating hormone (FSH) and testosterone concentrations in the patients of the extended sperm preparation (ESP) and testicular sperm extraction (TESE) groups Table III. Outcome of intracytoplasmic sperm injection procedure in the extended sperm preparation (ESP) and the testicular sperm extraction (TESE) groups a Patients FSH (IU/l) Testosterone ESP TESE P (nmol/l) Injected oocytes ESP 2 PN (%) 100 (41) 61 (49) NS Spermatozoa found PN (%) 19 (8) 15 (13) NS No spermatozoa PN (%) 9 (4) 2 (1.6) NS TESE Embryos (%) b 88 (88) 57 (93) Spermatozoa found No. of replaced embryos No spermatozoa No. of excellent embryos (I) No. of good embryos NS There were no statistical differences between the four groups. (II; II III) No. of fair embryos (III) 8 12 NS Number of blastomeres NS Table II. The oocytes and the injected oocytes in the extended sperm Cycles with replaced embryos 15/17 14/16 NS preparation (ESP) and the testicular sperm extraction (TESE) groups No. embryos replaced NS Cycles ESP TESE P Implantations (%) 9/57 (16) 6/45 (13) NS Clinical pregnancies per 4/15 (27) 4/14 (29) NS No. of cases replacement (%) No. of metaphase II oocytes Clinical pregnancies per cycle 4/17 (23) 4/32 (12.5) NS Oocytes per case NS a Values in parentheses are percentages. For details of embryo morphological Injected oocytes b grading see Materials and methods. % of metaphase II oocytes a The ESP and TESE groups are not strictly comparable since (i) more eggs injected c were injected in the ESP group and (ii) fertilization rates from ejaculated Injected oocytes per cycle a spermatozoa are known to be higher than those from testicular spermatozoa. b Percentage out of 2 PN (pronuclear) oocytes. a t-test; NS not significant. 62% of the eggs in the TESE group. This was due to fewer χ 2 -Test. The remaining oocytes were not injected since no more spermatozoa were available spermatozoa for ICSI in the TESE group as compared available. with the ESP group. The fertilization rate was similar in both groups as was the Spermatozoa were recovered from the biopsy (TESE ) in number of oocytes with single pronucleus (Table III). The 16 (50%) of these 32 patients. No spermatozoa could be found number of oocytes with multiple pronuclei was more than (TESE ) in the remaining 16 patients either in their immediate twice as high in the ESP group (4%) as compared with the sampling during the operation (wet preparation) or after TESE patients (1.6%). However, this difference was not centrifugation, preparation and incubation. statistically significant, probably due to the small numbers. Five (29%) of the 17 ESP males had a testicular volume The number of the replaced embryos and their mean number of 15 ml on each side compared with eight (50%) of the 16 of blastomeres was comparable in both groups. The embryos TESE patients from whom spermatozoa could be recovered classified as having excellent morphology (grade I) were more (not significant). Three of the 17 ESP and four of the 16 frequently seen in the ESP than in the TESE group: 18 out of TESE patients were diagnosed with cryptorchidism (not 57 compared with six out of 45 (P 0.05). significant) in childhood. The number of cycles with embryo replacements, percent Three patients in the ESP group had documented testicular of implantation (visible sacs/replaced embryos) and clinical histology from previous occasions in other centres. Severe pregnancies were comparable in both groups (ESP and TESE) hypospermatogenesis was described in two of them, and in (Table III). The pregnancy rate per cycle in the ESP group the third patient the histology of the biopsy was defined as was 23% compared with 12.5% in the TESE group. This maturation arrest. difference was not statistically significant. The ESP group Table I shows the basal concentrations of FSH and testosterone contained one singleton gestation, two twins and one quadruence in the different groups. No statistically significant differ- plet. The quadruplets were reduced to triplets. The patient was found, either in the patients with ESP compared refused a further reduction to twins. All four are ongoing with those who were ESP, or in those with TESE compared pregnancies of more than 20 weeks gestation. The TESE group with TESE. consisted of two singletons and two twins. One delivered a Table II shows that the cohort of mature oocytes retrieved healthy boy, and another singleton is now at the gestational from the women in the ESP group was larger (299) compared age of 23 weeks. The third pregnancy started as a twin to the corresponding number of oocytes (201) collected from pregnancy with one of the fetuses not developing beyond the women in the TESE group. There were significantly more the ninth week and is now at 14 weeks gestation. Another injected oocytes per cycle in the ESP group ( ) than twin pregnancy terminated as a missed abortion at 9 weeks in the TESE group ( ; P 0.005). Thus 79% of the gestation. oocytes in the ESP group were injected compared with only The number of available spermatozoa for injection in the 1224

4 ESP versus TESE in non-azoospermia Devroey et al., 1994; Silber et al., 1995a,b) and non-obstructive Table IV. Outcome of the intracytoplasmic injection of motile and non- azoospermia (Devroey et al., 1995, 1996; Tournaye et al., motile spermatozoa in the cases with extended sperm preparation (ESP) 1995, 1996). This novel approach has shattered the previous Spermatozoa definition of the term azoospermia as the lack of spermatozoa Motile Non motile P in seminal fluid. Tournaye et al., (1995) divided azoospermia into two sub- No. of injected oocytes (A) groups, absolute and virtual. According to their definition Degenerated oocytes 21 (10) a 4 (10) a NS* absolute azoospermia means lack of spermatozoa in all the 2PN (B) 78 (39) a 11 (26) a NS* 1PN 11(5) a 7 (17) a 0.02* ejaculates studied, including the sediment after centrifugation. 3PN 14(7) a ** In virtual azoospermia the ejaculate occasionally contains a Embryos (C) 77 (38) a 9 (21) a NS* few spermatozoa during examinations prior to the oocyte (99) b (82) b Excellent embryos (I) 16 (21) c 1(11) c NS** retrieval day, but no spermatozoa are visible in the ejaculate Good embryos (II; II III) 20 (26) c 4 (44) c obtained on the ovum retrieval day. In four (27%) of 15 cases Fair embryos (III) 22 (29) c 4 (44) c with absolute azoospermia spermatozoa could not be seen in Values in parentheses are percentages. PN pronuclear. For details of the testicular biopsy (Tournaye et al., 1995). In a more recent embryo morphological grading see Materials and methods. study by the same group (Tournaye et al., 1996), in eight a Percentage from A. (28%) of 29 patients with absolute azoospermia no spermatob Percentage from B. Percentage from C. zoa could be retrieved from the testicular biopsy. However, *χ 2 -Test. testicular spermatozoa were attained in the remaining 25 **Fisher s exact test. patients with virtual azoospermia mentioned in the same study. ESP procedure, especially the motile ones, was limited. There- In the present study we included in the group of nonfore, when oocytes were left over and no more originally obstructive azoospermia only patients in whom both the motile spermatozoa could be found, we injected non-motile ejaculate and its sedimentation after centrifugation revealed spermatozoa selected according to their morphological appearno spermatozoa. Our inclusion criteria are concomitant with ance. In 11 of the 17 patients in the ESP group, all oocytes the definition of absolute azoospermia mentioned above. The were injected with originally motile spermatozoa. In another results in the current study show that in 35% (17 out of 49) five of the ESP patients, 37 (58%) of the oocytes were injected of the patients diagnosed as having absolute azoospermia, with motile and the remaining 27 oocytes (42%) with nonspermatozoa could be obtained by the ESP technique. In 11 motile spermatozoa. In one patient all the 14 eggs were injected (22%) of these patients we could collect motile spermatozoa with non-motile spermatozoa, since no motile spermatozoa for the injection of all the oocytes retrieved from their partners; were seen in any of the droplets. All four pregnancies achieved in five (10%) of them enough motile spermatozoa were were among the 11 women whose oocytes were all injected with available to inject 58% of the oocytes, while the remaining originally motile spermatozoa. Table IV shows a comparison of 42% were injected with immotile spermatozoa. This means the outcome of ICSI with motile and non-motile spermatozoa that by using the ESP, in 33% (16 out of 49) of the nonin the ESP cases. Although such a comparison is not strictly obstructive azoospermic patients, motile spermatozoa could be orthodox since it was not known what proportion of the nonfound for the injection of part or all of the collected oocytes. motile spermatozoa were actually dead, the fertilization rate Since the majority of the patients with azoospermia will not appeared higher (39%), in the oocytes injected with motile achieve pregnancy in their first treatment cycle and the yield spermatozoa compared with the eggs injected with originally of frozen testicular spermatozoa is relatively low (Gil-Salom non-motile spermatozoa (26%); however, this difference was et al., 1996), most of them will be referred for repeated not statistically significant. testicular biopsies for TESE. This fact is a crucial shortcoming Oocytes with a single pronucleus were considerably more in the treatment of patients with non-obstructive azoospermia. frequent among the eggs injected with non-motile spermatozoa ESP contributes to avoiding unnecessary multiple testicular (17%) than eggs injected with originally motile ones (5%; biopsies for TESE which can lead to transient adverse physio- P 0.02). There were no oocytes with multiple pronuclei in logical effects on the testis (Schlegel, 1996). Although initially the small group of the 42 oocytes injected with non-motile the meticulous search for spermatozoa with the ESP method spermatozoa. Fourteen (7%) multipronucleated embryos is time consuming (an average of 45 min), with experience resulted from injection with motile spermatozoa. The cleavage the method is performed with greater confidence and less time rate and the number of embryos with class I morphology was is required per case. lower, but with no significant difference, in the group of The introduction of ESP as part of the procedure in the oocytes injected with non-motile spermatozoa compared with treatment of non-obstructive azoospermia preselects the cases the numbers in the cohort of oocytes undergoing ICSI with into those with spermatozoa present in the ESP (ESP ) and originally motile spermatozoa. those undergoing testicular biopsy, since they were defined as ESP. One could presume that the patients in the ESP group Discussion might have a better chance to achieve pregnancy. However, The introduction of ICSI has made it possible to use testicular the differences in the fertilization and cleavage rates were spermatozoa in cases of obstructive (Schoysman et al., 1993a,b; not statistically significant. Nevertheless, there were some 1225

5 R.Ron-El et al. differences: the embryos with excellent morphology were Levin, H.S. (1979) Testicular biopsy in the study of male infertility. Its current usefulness, histologic techniques, and prospects for the future. Hum. Pathol., significantly more frequent in the ESP than in the ESP 10, groups. This does not reflect the pregnancy rate in both groups, Palermo, G., Joris, H., Devroey, P. et al. (1992) Pregnancies after but probably influences the implantation rate (16 versus 13%, intracytoplasmic sperm injection of sperm injection. Lancet, 340, Ron-El, R., Herman, A., Golan, A. et al. (1991) Gonadotropins and combined not significant) and the later well-being of the fetus. While all gonadotropins releasing-hormone agonist gonadotropin protocols in a pregnancies are ongoing in the ESP group, there were two randomized prospective study. Fertil. Steril., 55, miscarriages out of four pregnancies in the TESE group. Schlegel, P.N. (1996) Physiological consequences of testicular sperm extraction. Hum. Reprod., 11 (Abstr. Book 1), 74, Abstr During the year 1995 the fertilization rate in our ICSI Schoysman, R., Van der Zwalmen, P., Nijs, M. et al. (1993a) Successful programme was 62%. The clinical pregnancy rate per cycle fertilization by testicular spermatozoa in an in-vitro fertilization programme. was 33% (153 gestations per 459 cycles). These rates are not Hum. Reprod., 8, significantly different than the corresponding rates in the ESP Schoysman, R., Van der Zwalmen, P., Nijs, P. et al. (1993b) Pregnancy after fertilization with human testicular sperm. Lancet, 342, group (four pregnancies out of 15 cases, 27%; P 0.2, Silber, S.J., Van Steirteghem, A., Liu, J. et al. (1995a) High fertilization and Fisher s exact test). This means that in cases with extremely pregnancy rate after intracytoplasmic sperm injection with spermatozoa low counts of spermatozoa, the results with ESP will resemble obtained from testicle biopsy. Hum. Reprod., 10, Silber, S.J., Nagy, Z., Liu, J. et al. (1995b) The use of epididymal and those with testicular spermatozoa using TESE, and both will testicular spermatozoa for intracytoplasmic sperm injection: the genetic be somewhat lower than, yet not significantly, with ejaculated implications for male infertility. Hum. Reprod., 10, spermatozoa in the regular ICSI programme. Tournaye, H., Devroey, P., Liu, J. et al. (1994) Microsurgical epididymal Normally we replace two embryos with excellent morphoapproach to infertility as a result of congenital bilateral absence of the vas sperm aspiration and intracytoplasmic sperm injection: a new effective logy when the patient is 25 years, three embryos with good deference. Fertil. Steril., 61, morphology at age and four embryos when the patient Tournaye, H., Camus, M., Goossens, A. et al. (1995) Recent concepts in the is 36 years old. We also replaced four embryos in women management of infertility because of non-obstructive azoospermia. Hum. Reprod., 10 (Suppl. 1), between 26 and 35 years old if some of the embryos were of Tournaye, H., Liu, H., Nagy, P.Z. et al. (1996) Correlation between testicular a lesser morphological quality (class II or III). Two patients histology and outcome after intracytoplasmic sperm injection using testicular aged 24 and 27 years insisted on the replacement of four spermatozoa. Hum. Reprod., 11, Van Steirteghem, A., Liu, J., Joris, H. et al. (1993a) Higher success rate of embryos although the possible risks were explained to them. intracytoplasmic sperm injection than by subzonal insemination. Report of Both conceived, one a quadruplet pregnancy and the other a the second series of 300 consecutive treatment cycles. Hum. Reprod., 8, twin pregnancy Van Steirteghem, A., Nagy, Z., Joris, H. et al. (1993b) High fertilization rates When analysing the outcome of the TESE group in the after intracytoplasmic sperm injection. Hum. Reprod., 8, present study the pregnancy rate is lower than in the other studies: 29% compared with 41%, the average percentage in Received on November 26, 1996; accepted on April 2, 1997 the categories of germ-cell aplasia, hypoplasia and maturation arrest (Tournaye et al., 1996). We counted only clinical pregnancies and our TESE cases are a selection of the more severe cases since all were ESP, which may explain the differences in the results. In our opinion ESP should be a prerequisite procedure when treating patients with non-obstructive azoospermia before performing testicular biopsy. The fact that ESP yields results as good as TESE, while avoiding an invasive procedure, justifies this approach. Hopefully, further randomized studies will confirm these results. References Craft, I., Bennett, V. and Nicholson, N. (1993) Fertilising ability of testicular spermatozoa. (Letter.) Lancet, 342, 864. Devroey, P., Liu, J., Nagy, Z. et al. (1994) Normal fertilization of human oocytes after testicular sperm extraction and intracytoplasmic sperm injection. Fertil. Steril., 62, Devroey, P., Liu, J. and Nagy, P. (1995) Pregnancies after testicular sperm extraction (TESE) and intracytoplasmic sperm injection (ICSI) in nonobstructive azoospermia. Hum. Reprod., 10, Devroey, P., Nagy, P., Tournaye, H. et al. (1996) Outcome of intracytoplasmic sperm injection with testicular spermatozoa in obstructive and nonobstructive azoospermia. Hum. Reprod., 5, Hauser, R., Temple-Smith, P.D., Southwick, G.J. et al. (1995) Fertility in cases of hypergonadotropic azoospermia. Fertil. Steril., 63, Gil-Salom, M., Romero, J., Minquez, Y. et al. (1996) Pregnancies after intracytoplasmic sperm injection with cryopreserved testicular spermatozoa. Hum. Reprod., 6,