A predictive score for testicular sperm extraction quality and surgical ICSI outcome in non-obstructive azoospermia: a retrospective study

Human Reproduction, Vol.26, No.12 pp. 3215 3221, 2011 Advanced Access publication on September 18, 2011 doi:10.1093/humrep/der314 ORIGINAL ARTICLE Andrology A predictive score for testicular sperm extraction quality and surgical ICSI outcome in non-obstructive azoospermia: a retrospective study F. Boitrelle 1,2, *, G. Robin 3,4, F. Marcelli 3, M. Albert 2, B. Leroy-Martin 1, D. Dewailly 4, J.-M. Rigot 3, and V. Mitchell 1,5 1 Department of Reproductive Biology and CECOS, Lille University Medical Center, Lille 59000, France 2 Department of Reproductive Biology and Cytogenetics, Hospital of Poissy, Poissy 78303, France 3 Department of Andrology and CECOS, Lille University Medical Center, Lille 59000, France 4 Department of Gynecology and Reproductive Medicine, Lille University Medical Center, Lille 59000, France 5 University of Medicine and Science, EA 4308, Lille 59000, France *Correspondence address. E-mail: florenceboitrelle@yahoo.fr Submitted on March 20, 2011; resubmitted on June 17, 2011; accepted on August 1, 2011 background: In non-obstructive azoospermia (NOA), testicular sperm extraction (TESE) is successful in 50% of cases. A parameter for predicting TESE quality and pregnancy rates after ICSI of testicular spermatozoa is still lacking. methods: We retrospectively evaluated the total testicular volume (TTV), hormone levels and TESE quality in 280 patients with NOA. After successful TESE, the characteristics of the ICSI cycles and the take-home baby rates were evaluated. results: TESE was successful in 149 patients (53.2%). In a multivariate logistic regression analysis, only TTV, FSH and inhibin B were correlated with the TESE outcome. A score including these three parameters was the best predictor of successful TESE (positive likelihood ratio: +3.01). When the score was,18.5, TESE was successful in 77.4% of cases and sperm rich (i.e. yielding.100 spermatozoa) in 91.1% of cases; 42.8% of couples took a baby home. The take-home baby rate did not depend on whether the score was,18.5 or between 18.5 and 3700. If the score was.3700, TESE was successful in 37.8% of cases and sperm rich in 14.3% of cases; only one couple took a baby home (a rate of 7.7 versus 42.8% when the score was,18.5; P, 0.001). Owing to low sperm retrieval on the day of oocyte retrieval, fewer oocytes were injected when the score was.3700 than when,3700. conclusions: Thanks to better knowledge of TESE quality, the present score could help to improve care and pre-icsi counseling for patients with NOA. Key words: non-obstructive azoospermia / prediction / testicular sperm extraction outcome / testicular sperm / ICSI Introduction Although the lack of ejaculated spermatozoa makes azoospermic men naturally infertile, these men can become fathers thanks to ICSI (Palermo et al., 1992) recovered by testicular sperm extraction (TESE) (Schoysman et al., 1993). However, in non-obstructive azoospermia (NOA), testicular sperm recovery is only successful in 50% of cases (Tournaye et al., 1995; Chan and Schlegel, 2000; Ald et al., 2004). Testicular biopsy is invasive and may impair future spermatogenesis or induce testicular atrophy (Schlegel and Su, 1997; Fenichel et al., 1999). Even though the pregnancy rate after ICSI is encouragingly high (Devroey et al., 1996; Devroey and Van Steirteghem, 2004; Verheyen et al., 2004; Andersen et al., 2008), it is known to be (i) lower in NOA than in obstructive azoospermia (OA) (Nicopoullos et al., 2004 for review) and (ii) lower in NOA than in men with normal spermatogenesis when ejaculated spermatozoa were used for ICSI (for a review, see Tournaye, 2010). Hence, patients should be informed about their chances of successful TESE and their likelihood of becoming parents. Some studies have suggested that testicular volume (Bohring et al., 2002; Goulis et al., 2008), serum FSH level (von Eckardstein et al., 1999; Bohring et al., 2002; Vernaeve et al., 2002; Zitzmann et al., 2006; Goulis et al., 2008), serum inhibin B level (von Eckardstein et al., 1999; Bohring et al., 2002; Nagata et al., 2005; Goulis et al., 2008; Chen et al., 2010) or combinations of two hormones (von Eckardstein et al., 1999; Bohring et al., 2002) may be valuable parameters in the prediction of TESE outcomes. & The Author 2011. Published by Oxford University Press on behalf of the European Society of Human Reproduction and Embryology. All rights reserved. For Permissions, please email: journals.permissions@oup.com

3216 Boitrelle et al. However, neither clinical nor hormonal parameters can predict TESE outcomes with sufficient accuracy at present (Tunc et al., 2006; Carpi et al., 2009). The present study focused on in a large population of patients with NOA (n ¼ 280). For the first time, rates of sperm extraction were analyzed qualitatively and quantitatively. A multivariate logistic regression model was used to determine whether any of the measured parameters were predictive of the TESE outcome. Next, the characteristics of the subsequent ICSI cycles and the resulting pregnancy rates were analyzed. We sought to answer the following questions. Which men will yield sperm in a TESE? How can we improve the counseling and management of patients with NOA and their partners in order to increase pregnancy rates? Materials and Methods Patient selection and evaluation This retrospective case series included patients with NOA having undergone TESE in our andrology clinic between 2002 and 2009. All patients suffered from azoospermia (i.e. no sperm after centrifugation of total ejaculate in at least two semen analyses with a 3-month interval) (WHO, 2009). Before surgical sperm extraction, each patient underwent a complete andrologic evaluation to determine the etiology of azoospermia (clinical history, physical examination, seminal markers, ultrasound imaging of the urogenital tract and testes, serum FSH, inhibin B, total testosterone and LH levels, karyotype analysis and Yq microdeletion analysis), as previously recommended (Huyghe et al., 2008; Robin et al., 2010). Patients with OA, a vasectomy, anejaculation or hypogonadotrophic hypogonadism was excluded. As a result, 280 patients with NOA were included in the study. Testicular volumes Testicular volumes were calculated following an ultrasound examination. Total testicular volume (TTV) was defined as the sum of the two testis volumes. The normal minimum value for single testis volume is 16 ml (i.e. a TTV of 32 ml; Schiff et al., 2004). Hormone assays Serum FSH and LH levels were measured by an immunoassay with a limit of detection (LOD) of 0.05 IU/l. The intra- and inter-assay coefficients of variation were,3 and,5%, respectively. The normal concentration ranges were 2 10 IU/l for FSH and 1 12 IU/l for LH. Testosterone levels in serum samples collected in the morning were determined using a radioimmunoassay. The normal concentration range for testosterone was 2.3 6.7 ng/ml. Serum inhibin B levels were measured using a commercially available, dimer-specific, solid-phase, sandwich enzyme-linked immunosorbent assay (Oxford Bio-Innovation, Oxford, UK) with an LOD of 15 pg/ml. Patients with inhibin B concentrations below the detection limit were assigned with a value of 1 pg/ml. The normal concentration for serum inhibin B was 238 + 32 pg/ml (range: 102 459 for fertile patients). Testicular sperm extraction Testicular biopsies were performed under general anesthesia, as described previously (Tournaye et al., 1997). For each patient, recovered testicular sperm were cryopreserved prior to subsequent oocyte retrieval. A portion of the testicular biopsy was sent to the reproductive biology laboratory. After each biopsy had been weighed, the testicular tissue was prepared as described previously (Robin et al., 2010) by mincing and mechanical shedding in culture medium, in order to dilacerate the seminiferous tubules. An aliquot was examined under an inverted microscope at 400 magnification. When spermatozoa were found, their number and vitality were rated. If at least one motile or immotile but living spermatozoon (in a hypo-osmotic swelling test) was observed in a volume of 10 ml, the testicular cell suspension was frozen for later use in ICSI. If very few spermatozoa were observed (,100 spermatozoa in the entire sample prior to cryopreservation), the sample was frozen but labeled as coming from a sperm-poor extraction. Extractions with.100 spermatozoa in the sample were considered to be sperm rich. If no spermatozoa were found after centrifugation of the preparation (300 g for 10 min) and re-examination of the pellet, samples were not frozen. ICSI cycles Eighty-nine couples (in which the woman was under the age of 35 years and had no known etiologies for infertility) underwent one or more ICSI cycles in our IVF center between December 2002 and September 2010 and were included in the study. Controlled ovarian hyperstimulation was performed with recombinant FSH (rfsh, Puregon, Schering-Plough, The Netherlands, or Gonal-F, Merck-Serono, Switzerland) after pituitary desensitization with a GnRH agonist: daily injections of triptorelin 0.1 mg (Decapeptyl, Ipsen, France) were initiated in the mid-luteal phase of the preceding cycle (for women with regular menstruations) or on the first day of bleeding (for women with oligomenorrhea or amenorrhea). Desensitization was checked 12 15 days after the first administration of GnRH agonist. Daily injections of rfsh were initiated only if (i) the estradiol (E 2 ) level was,50 pg/ml, (ii) the endometrium was,6 mm thick and (iii) there were no functional ovarian cysts. Hormone assays and ultrasound monitoring were then performed regularly. Five thousand IU of purified urinary hcg (Schering-Plough) was administered as soon as three follicles (or more) reached a mean diameter of 17 mm, with a consistent rise in serum E 2 concentration. Oocytes were retrieved 35 36 h after the hcg injection by transvaginal, ultrasound-guided needle aspiration. After removing cumulus corona cells, mature oocytes (with the first polar body) were injected with motile spermatozoa, when available. For samples with non-motile sperm, sperm viability was assessed with the hypo-osmotic swelling test (WHO, 2009). All embryo transfers were performed 2 or 3 days after oocyte retrieval. Luteal phase support consisted of 600 mg vaginal micronized progesterone and 6 mg oral E 2 per day (both initiated on the day of retrieval). If a pregnancy occurred, progesterone and E 2 administration was maintained until ultrasound evidence of fetal heart activity was detected. Supernumerary embryos were cryopreserved and some were transferred in further cycles. Clinical pregnancy was defined as a gestational sac with fetal heartbeats at least 5 weeks after embryo transfer. Clinical pregnancy and take-home baby rates were calculated per couple. Statistical analysis Statistical analyses were carried out with SAS software, version 9.1 (SAS Institute, Inc., USA). Data are expressed as the mean + SEM. Two groups of patients (successful versus unsuccessful sperm recovery) were compared with a Student s t-test. The ability of TTV and hormonal parameters to discriminate between successful and unsuccessful TESE was evaluated in terms of the positive predictive value (PPV), negative predictive value (NPV) and positive and negative likelihood ratios (LR+ and LR2, respectively) at different cut-offs determined by receiver operating characteristic (ROC) curve analysis. To calculate the score, a multivariate logistic regression model was used to determine which parameters were independently correlated with a successful TESE. The characteristics of the TESE and subsequent ICSI cycles were analyzed and compared as a

Predicting sperm extraction quality and pregnancy rate 3217 Figure 1 ROC curves for TTV, FSH, inhibin B and our score as a guide to the presence of spermatozoa in 280 men with NOA. function of the score. Student s t-test was used to compare numerical parameters (the woman s age, the number of ICSI cycles, the number of oocytes retrieved and the number of mature oocytes injected with spermatozoa). If data were not normally distributed, a Wilcoxon sum of ranks test was applied. The chi-squared test was used to compare the groups in term of the clinical pregnancy and take-home baby rates. The threshold for statistical significance was set to P, 0.05 for all tests. Results TESE outcomes and predictive parameters The average age (range) of the 280 patients with NOA was 33.2 + 0.4 years (range: 22 44). The average biopsy weight was 186 + 6 (30 540) mg. No etiology was found for 130 (36.8%) of the patients with NOA (i.e. idiopathic NOA), 95 had a history of cryptorchidism (33.9%) and 23 had Klinefelter syndrome (8.2%). Other patients presented with Y microdeletions (n ¼ 6), Y inversion (n ¼ 4) or a history of inguinoscrotal surgery (n ¼ 11), testicular tumor (n ¼ 8) or chemo/radiotherapy (n ¼ 7). Twenty-three patients had a combination of several etiologies. TESE was successful in 149 of the 280 patients (53.2%). There was no significant difference between the groups of patients with successful or unsuccessful TESE in terms of the man s age, total testosterone levels, LH levels or biopsy weight. Patients with successful sperm retrieval had a greater mean TTV than patients with unsuccessful sperm retrieval (18 + 0.7 versus 15 + 0.6 ml, respectively; P ¼ 0.0013), along with a lower mean FSH concentration (18.7 + 1.5 versus 24.9 + 1.5 IU/l, respectively; P ¼ 0.003) and a higher mean serum inhibin B concentration (74.8 + 6.8 versus 27.2 + 3.5 pg/ml, respectively; P, 0.001). The predictive values of TTV, FSH and inhibin B for discriminating between successful or unsuccessful TESE were assessed by calculating the corresponding area under the ROC curve (AUC) (Fig. 1). On a stand-alone basis, inhibin B (cut-off at 27.5 pg/ml; AUC: 0.683) was the best predictor of TESE outcome, with a PPV and NPV of 69.2 and 60.7%, respectively and with an LR+ and LR2 of 1.98 and 0.57, respectively (Table I). Inhibin B was not detected in 40.4% of the patients (n ¼ 113 out of 280) and TESE was unsuccessful for 69 patients with undetectable inhibin B (n ¼ 69 out of 113; 61.1%). After multivariate logistic regression analysis, TTV, FSH and inhibin B were retained as being independently correlated with the TESE outcome. The PPV, NPV, LR+ and LR2 for various ratios [FSH/inhibin B; FSH/TTV (FSH TTV)/inhibin B; FSH/(TTV inhibin B)] were also evaluated. The latter ratio had the best predictive values for TESE outcome (AUC: 0.663). To obtain information on hormone levels in NOA (relative to OA and unimpaired spermatogenesis), the score was multiplied by the reciprocal of the mean score obtained for 50 men with OA and unimpaired spermatogenesis [i.e. (inhibin B TTV)/FSH) ¼ 915)]. Furthermore, this coefficient yielded a unitless score: 915 FSH Score = TTV inhibin B

3218 Boitrelle et al. Table I Data for the parameters investigated as possible predictors of TESE quality and pregnancy rates after ICSI. Cut-off Se Sp PPV NPV LR1 LR2... TTV (ml) 16 59.7 56.5 61 55.2 1.37 0.7 FSH (IU/l) 20.5 68.5 55.7 63.8 60.8 1.54 0.56 Inhibin B (pg/ml) 27.5 60.4 69.5 69.2 60.7 1.98 0.57 Score 18.5 53 82.4 77.4 60.7 3.01 0.57 Se, sensitivity; SP: specificity. Multiplication by this coefficient did not change the statistical results (the AUC for a score computed in this way was 0.663 (Fig. 1). The mean (range) score for the 280 patients with NOA was 1791 + 17.8 (0.2 37656). The most discriminative cut-off was 18.5, yielding the highest PPV (77.4%) and the highest LR+ (3.01) (Table I). The ROC analysis also revealed that a score cut-off of 3700 had the best NPV (62.3%) (Fig. 1). Table II shows the TESE outcome as a function of the score. When the score was 18.5, sperm recovery was more frequently successful than when the score was.3700; the higher the score, the less successful the TESE. In 12 of the 23 patients with a score of 18.5 but no sperm retrieval, a histological biopsy showed maturation arrest. Thirty-seven successful TESEs (24.8%) yielded fewer than 100 spermatozoa per extraction. These patients with a sperm-poor TESE did not differ from patients with a sperm-rich TESE in terms of biopsy weight (190 + 12 versus 199 + 6 mg, respectively; P ¼ 0.55). With a score of 18.5, 91.1% of the TESEs were sperm rich. With a score of.3700, 14.3% of the TESEs were sperm rich (Table II); the higher the score, the less sperm rich the extraction. Characteristics of ICSI cycles, pregnancy rates and take-home baby rates Of the 89 couples having undergone ICSI, 33 (37%) took a baby home. Thirty-eight couples (42.7%) had at least one pregnancy, with an average of 1.07 pregnancies per couple. Of the 38 pregnant women, three suffered a miscarriage before 12 weeks of pregnancy and two had an extra-uterine pregnancy (Table III). The woman s age, the number of ICSI cycles and the number of mature oocytes retrieved did not differ as a function of the man s score. Owing to a lack of retrieved sperm on the day of oocyte retrieval, significantly fewer oocytes (5.3 + 0.5) were injected when the score was.3700 than when the score was,18.5 (7.6 + 0.6; P ¼ 0.05) or between 18.5 and 3700 (7.2 + 0.6; P ¼ 0.03). For men scoring,3700, the clinical pregnancy rate and the take-home baby rate did not depend on whether the score was,18.5 or between 18.5 and 3700 [48.6 versus 48.8%, respectively; P ¼ 0.87 and 42.8 versus 41.5%, respectively; P ¼ 0.2]. If the score was.3700, the clinical pregnancy rate and the take-home baby rate were lower than those obtained for couples in which the man obtained a score of,18.5 (7.7 versus 48.6%; P, 0.001 and 7.7 versus 42.8%; P, 0.001) or between 18.5 and 3700 (7.7 versus 48.8%; P, 0.001 and 7.7 versus 41.5%; P, 0.001). Table II Presence and quantity of spermatozoa retrieved (TESE quality) as a function of the score. TESE (n 5 280) Successful Percentage of TESE rate (%) sperm rich TESEs (i.e. more than 100 spermatozoa cryopreserved)... Score 18.5 (n ¼ 102) 77.4 (n ¼ 79/102) 91.1 (n ¼ 72/79) 18.5, score, 3700 39.7 (n ¼ 56/141) 67.9 (n ¼ 38/56) (n ¼ 141) Score 3700 (n ¼ 37) 37.8 (n ¼ 14/37) 14.3 (n ¼ 2/14) Discussion To the best of our knowledge, this is the first time that TESE outcomes have been evaluated quantitatively and qualitatively in a large population of patients with NOA. We evaluated the pregnancy and take-home baby rates per couple as a function of the quality of sperm extraction in a large sample of patients with NOA. In line with the literature data (von Eckardstein et al., 1999; Chan and Schlegel, 2000; Bohring et al., 2002; Vernaeve et al., 2002; Ald et al., 2004; Zitzmann et al., 2006; Goulis et al., 2008), TESE was successful in 53.2% of cases. To date, it has not been possible to predict the TESE outcome in NOA on the basis of non-invasive parameters (Carpi et al., 2009; Adamopoulos and Koukkou, 2010). When used alone, neither TTV (Bohring and Krause, 1999; Kolb et al., 2000; Hipler et al., 2001; Bohring et al., 2002; Carpi et al., 2009) nor FSH levels (Carpi et al., 2009) were robust enough. A few authors have stated than inhibin B may be a better predictive factor than FSH (von Eckardstein et al., 1999; Ballesca et al., 2000; Brugo-Olmedo et al., 2001) and various cut-off have been reported [e.g. 94 pg/ml (von Eckardstein et al., 1999), 40 pg/ml (Ballesca et al., 2000) and 13.7 pg/ml (Vernaeve et al., 2002)]. A recent meta-analysis found that the inhibin B concentration was most predictive of the presence of spermatozoa but could not serve as a stand-alone marker of spermatogenesis in men with NOA (Toulis et al., 2010). A multivariate regression model was used in the present study. The linearity of the relationship between dependent and independent variables could not be proved. Hence, we applied a log-transformation to determine the predictive parameters of TESE outcome; but to

Predicting sperm extraction quality and pregnancy rate 3219 Table III Characteristics of the ICSI cycles and clinical pregnancy and take-home baby rates as a function of the score. ICSI (n 5 89) Score 18.5 (n 5 35) 18.5 < score < 3700 Score 3700 (n 5 13) P (Wilcoxon s test) (n 5 41)... Woman s age (years) 30.9 + 0.7 ab 30.9 + 0.6 ac 33.3 + 1.5 bc a ¼ 0.95 b ¼ 0.18 c ¼ 0.14 Mean number of ICSI cycles 2.1 + 0.2 de 2 + 0.2 df 1.7 + 0.1 ef d ¼ 0.59 e ¼ 0.20 f ¼ 0.53 Mean number of mature oocytes by cycle 8 + 0.6 gh 7.6 + 0.5 gi 7.5 + 0.6 hi g ¼ 0.69 h ¼ 0.86 i ¼ 0.95 Mean number of injected oocytes 7.6 + 0.6 jk 7.2 + 0.6 jl 5.3 + 0.5 kl j ¼ 0.92 k 5 0.05 l 5 0.03 Clinical pregnancy rate 48.6% mn (n ¼ 17) 48.8% mo (n ¼ 20) 7.7% no (n ¼ 1) m ¼ 0.87 n 5 2.4 3 10 24 o 5 5.1 3 10 25 Take-home baby rate 42.8% pq (n ¼ 15) 41.5% pr (n ¼ 17) 7.7% qr (n ¼ 1) p ¼ 0.2 q 5 8.6 3 10 25 Bold entries highlight the statistically significant P. r 5 1.1 3 10 25 compute a score which was easy to calculate, we decided to use initial values (instead of log values). Although this constituted a limitation, regression analysis remained a powerful method for determining which main parameters were independently correlated with the TESE outcome. Hence, only TTV, FSH and inhibin B values were selected for calculation of the score. According to the ROC curves, inhibin B was more predictive of TESE outcome than FSH or TTV. However, when the inhibin B level was,15 pg/ml (in 40.4% of men), it had low specificity (as shown by the linear, final part of the ROC curve; Fig. 1). In this situation, the score (which combines two other parameters) was able to provide valuable information about the TESE outcome. Furthermore, the score had the highest LR+ (3.01) and was therefore considered to be the best predictor of the TESE outcome. It should be noted that patients with inhibin B concentrations below the detection limit (15 pg/ml) were arbitrarily assigned with a value of 1 pg/ml for convenience; this assignment did not influence the score s predictive values and diagnostic performance when compared with a value of 14 pg/ml. Calculation of our score prior to surgery and the use of two thresholds (18.5 with the best PPV and 3700 with the best NPV) could therefore improve the counseling and management of individual cases: If the score is,18.5, the patient has a rather good chance of a successful TESE (77.4%; LR+: 3.01). The extraction was sperm rich in 91.1% of cases and 42.8% of these couples took a baby home. Our results suggest that the asynchronous TESE procedure (i.e. not carried out at the same time as oocyte retrieval) is efficient and does not have an overly negative impact on ICSI outcome (i.e. it gives high pregnancy rates). The conclusions of a Cochrane Collaboration meta-analysis (Van Peperstraten et al., 2008) have validated this procedure. If the score is between 18.5 and 3700, sperm retrieval is less likely to be successful (39.7%) and so patients should be informed of this likelihood. However, when TESE is successful, 41.5% of couples take a baby home. Again, the asynchronous TESE procedure is efficient. Our data suggest that testicular sperm quality and clinical pregnancy rates are not influenced by male FSH or inhibin B levels because the pregnancy rates in this group were just high as those obtained for men with a score of,18.5. If the score is.3700, the extraction was sperm rich in only 14.3% of cases. Only one couple took a baby home (7.7%); this is very much lower than for patients with a score of,3700 (41.5 42.8%), and is probably related to a lack of retrieved sperm on the day of oocyte retrieval because significantly fewer oocytes were injected when the score was.3700 than when the score was,18.5 (P ¼ 0.05) or between 18.5 and 3700 (P ¼ 0.03). Hence, improved counseling for couples with a high score could prompt them to postpone or refuse surgery, or choose another way of becoming parents. For example, they may decide to use donor sperm if the score predicts a very low likelihood of intraconjugal pregnancy. Furthermore, calculation of the score prior to surgery would also enable andrologists to improve the management of patients with a score of.3700 and who are therefore at risk of a poor-quality TESE result. In such cases, synchronized TESE with ICSI could be suggested; our data suggest that asynchronous ICSI after a spermpoor TESE could give a disappointingly low clinical pregnancy rate because of the lack of spermatozoa retrieved on the day of oocyte injection. These results agree with a recent report (Cavallini et al., 2011). In fact, as has been demonstrated for spermatozoa (McLaughlin

3220 Boitrelle et al. et al., 1992; Henry et al., 1993; Park et al., 2003; Griveau et al., 2006) and testicular sperm (Verheyen et al., 1997), a significant decrease in motility and viability occurs with freezing and thawing. Our study suggests that the negative impact of sperm cryopreservation could be even worse when very few spermatozoa are initially retrieved. For most of the men with a score.3700, the low number of sperm retrieved before and after cryopreservation prevented all the mature oocytes from being injected in an asynchronous TESE programme (on average, 5.3 of the 7.5 available oocytes were injected). In our view, synchronized ICSI cycles may increase clinical pregnancy rates for couples at risk of a poor TESE result (i.e. with a score of.3700). Although there would still be a risk of oocyte retrieval with no sperm to inject, the score may provide an efficient method of preselecting patients at risk of poor TESE outcomes. Furthermore, repeat TESE accounts for,1% of cases in our center. Our present data suggest that clinical pregnancy rates can be improved by increasing the number of retrieved spermatozoa. Hence, repeat TESE could be suggested more frequently to patients with positive but low-quality TESEs (score of.3700). Indeed, Vernaeve s group continued to retrieve spermatozoa in men in whom a first TESE was positive. This same team shows that when no spermatozoa were obtained after thawing cryopreserved testicular sperm for ICSI in patients with NOA, a repeat TESE procedure can be scheduled (Vernaeve et al., 2006). Finally, this score could be used by reproductive biologists as an internal quality control. In patients with a score of.3700 and who are at risk of a poor TESE outcome, biologists could adapt the extraction procedure and increase the number of spermatozoa retrieved by using, for example, chemical techniques with collagenase, as previously described (Vernaeve et al., 2002). In conclusion, the score presented here is a tool which could help andrologists, gynecologists and reproductive biologists to improve the quality of care for men with NOA. In our center, the score is currently being used to improve pre-surgery counseling of patients with NOA by predicting their TESE quality and ICSI outcome. Thanks to better knowledge of TESE quality prior to surgery and better counseling of the couples involved, it should be possible to increase pregnancy rates. In future work, we hope to be able to determine the value of synchronized ICSI cycles or repeat TESE for patients whose score is.3700. In the future, only a few spermatozoa may suffice for successful IVF. Authors roles F.B. conceived the study, performed experiments and wrote the manuscript. G.R. participated in the study design and revised the manuscript critically. F.M. recruited patients and performed testicular biopsies. M.A. revised the manuscript critically. B.L.M. participated in TESE. D.D. revised the manuscript critically. J.M. recruited patients, performed testicular biopsies and revised the manuscript critically. V.M. supervised the experiments, interpretation of the data and the drafting of the manuscript. All authors approved the final manuscript. Acknowledgements The authors thank N. Sebbane at Lille 2 University for statistical analysis of the results. The study was funded by grants from the Interregional G4 Collaboration on Reproduction (Amiens, Caen, Rouen and Lille University Hospitals), Lille University Hospital (PHRC 2004 R-1908), the French Language Society for Andrology (SALF), the GRAL Andrology Research Group, the French Agency for Biomedicine (ABM) and the Fonds d Aide à la Recherche Organon (FARO, Laboratoires Organon, France). References Adamopoulos DA, Koukkou EG. Value of FSH and inhibin-b measurements in the diagnosis of azoospermia a clinician s overview. Int J Androl 2010;33:109 113. Ald M, Niederberger CS, Ross LS. Surgical sperm retrieval for assisted reproduction. Minerva Ginecol 2004;56:217 222. Andersen AN, Goossens V, Ferraretti AP, Bhattacharya S, Felberbaum R, de Mouzon J, Nygren KG. Assisted reproductive technology in Europe, 2004: results generated from European registers by ESHRE. Hum Reprod 2008;23:756 771. Ballesca JL, Balasch J, Calafell JM, Alvarez R, Fabregues F, de Osaba MJ, Ascaso C, Vanrell JA. Serum inhibin B determination is predictive of successful testicular sperm extraction in men with non-obstructive azoospermia. Hum Reprod 2000;15:1734 1738. Bohring C, Krause W. Serum levels of inhibin B in men with different causes of spermatogenic failure. Andrologia 1999;31:137 141. Bohring C, Schroeder-Printzen I, Weidner W, Krause W. Serum levels of inhibin B and follicle-stimulating hormone may predict successful sperm retrieval in men with azoospermia who are undergoing testicular sperm extraction. Fertil Steril 2002;78:1195 1198. Brugo-Olmedo S, De Vincentiis S, Calamera JC, Urrutia F, Nodar F, Acosta AA. Serum inhibin B may be a reliable marker of the presence of testicular spermatozoa in patients with nonobstructive azoospermia. Fertil Steril 2001;76:1124 1129. Carpi A, Sabanegh E, Mechanick J. Controversies in the management of nonobstructive azoospermia. Fertil Steril 2009;91:963 970. Cavallini G, Cristina Magli M, Crippa A, Resta S, Vitali G, Pia Ferraretti A, Gianaroli L. The number of spermatozoa collected with testicular sperm extraction is a novel predictor of intracytoplasmic sperm injection outcome in non-obstructive azoospermic patients. Asian J Androl 2011; 13:312 316. Chan PT, Schlegel PN. Nonobstructive azoospermia. Curr Opin Urol 2000; 10:617 624. Chen SC, Hsieh JT, Yu HJ, Chang HC. Appropriate cut-off value for follicle-stimulating hormone in azoospermia to predict spermatogenesis. Reprod Biol Endocrinol 2010;8:108. Devroey P, Van Steirteghem A. A review of ten years experience of ICSI. Hum Reprod Update 2004;10:19 28. Devroey P, Nagy P, Tournaye H, Liu J, Silber S, Van Steirteghem A. Outcome of intracytoplasmic sperm injection with testicular spermatozoa in obstructive and non-obstructive azoospermia. Hum Reprod 1996;11:1015 1018. Fenichel P, Rey R, Poggioli S, Donzeau M, Chevallier D, Pointis G. Anti-Mullerian hormone as a seminal marker for spermatogenesis in non-obstructive azoospermia. Hum Reprod 1999;14:2020 2024. Goulis DG, Polychronou P, Mikos T, Grimbizis G, Gerou S, Pavlidou V, Papanikolaou A, Tarlatzis BC, Bontis IN, Papadimas I. Serum inhibin-b and follicle stimulating hormone as predictors of the presence of sperm in testicular fine needle aspirate in men with azoospermia. Hormones (Athens) 2008;7:140 147. Griveau JF, Lobel B, Laurent MC, Michardiere L, Le Lannou D. Interest of pentoxifylline in ICSI with frozen-thawed testicular spermatozoa from

Predicting sperm extraction quality and pregnancy rate 3221 patients with non-obstructive azoospermia. Reprod Biomed Online 2006; 12:14 18. Henry MA, Noiles EE, Gao D, Mazur P, Critser JK. Cryopreservation of human spermatozoa. IV. The effects of cooling rate and warming rate on the maintenance of motility, plasma membrane integrity, and mitochondrial function. Fertil Steril 1993;60:911 918. Hipler UC, Hochheim B, Knoll B, Tittelbach J, Schreiber G. Serum inhibin B as a marker for spermatogenesis. Arch Androl 2001;46:217 222. Huyghe E, Izard V, Rigot JM, Pariente JL, Tostain J. Optimal evaluation of the infertile male. French urological association guidelines. Prog Urol 2008;18:95 101. Kolb BA, Stanczyk FZ, Sokol RZ. Serum inhibin B levels in males with gonadal dysfunction. Fertil Steril 2000;74:234 238. McLaughlin EA, Ford WC, Hull MG. Motility characteristics and membrane integrity of cryopreserved human spermatozoa. J Reprod Fertil 1992; 95:527 534. Nagata Y, Fujita K, Banzai J, Kojima Y, Kasima K, Suzuki M, Tanaka K. Seminal plasma inhibin-b level is a useful predictor of the success of conventional testicular sperm extraction in patients with non-obstructive azoospermia. J Obstet Gynaecol Res 2005; 31:384 388. Nicopoullos JD, Gilling-Smith C, Almeida PA, Norman-Taylor J, Grace I, Ramsay JW. Use of surgical sperm retrieval in azoospermic men: a meta-analysis. Fertil Steril 2004;82:691 701. Palermo G, Joris H, Devroey P, Van Steirteghem AC. Pregnancies after intracytoplasmic injection of single spermatozoon into an oocyte. Lancet 1992;340:17 18. Park YS, Lee SH, Song SJ, Jun JH, Koong MK, Seo JT. Influence of motility on the outcome of in vitro fertilization/intracytoplasmic sperm injection with fresh vs. frozen testicular sperm from men with obstructive azoospermia. Fertil Steril 2003;80:526 530. Robin G, Boitrelle F, Leroy X, Peers MC, Marcelli F, Rigot JM, Mitchell V. Assessment of azoospermia and histological evaluation of spermatogenesis. Ann Pathol 2010;30:182 195. Schiff JD, Li PS, Goldstein M. Correlation of ultrasonographic and orchidometer measurements of testis volume in adults. BJU Int 2004; 93:1015 1017. Schlegel PN, Su LM. Physiological consequences of testicular sperm extraction. Hum Reprod 1997;12:1688 1692. Schoysman R, Vanderzwalmen P, Nijs M, Segal L, Segal-Bertin G, Geerts L, van Roosendaal E, Schoysman D. Pregnancy after fertilisation with human testicular spermatozoa. Lancet 1993;342:1237. Toulis KA, Iliadou PK, Venetis CA, Tsametis C, Tarlatzis BC, Papadimas I, Goulis DG. Inhibin B and anti-mullerian hormone as markers of persistent spermatogenesis in men with non-obstructive azoospermia: a meta-analysis of diagnostic accuracy studies. Hum Reprod Update 2010;16:713 724. Tournaye H. Update on surgical sperm recovery the European view. Human Fertil(Camb) 2010;13:242 246. Tournaye H, Camus M, Goossens A, Liu J, Nagy P, Silber S, Van Steirteghem AC, Devroey P. Recent concepts in the management of infertility because of non-obstructive azoospermia. Hum Reprod 1995;10(Suppl. 1):115 119. Tournaye H, Camus M, Vandervorst M, Nagy Z, Joris H, Van Steirteghem A, Devroey P. Surgical sperm retrieval for intracytoplasmic sperm injection. Int J Androl 1997;20(Suppl. 3):69 73. Tunc L, Kirac M, Gurocak S, Yucel A, Kupeli B, Alkibay T, Bozkirli I. Can serum Inhibin B and FSH levels, testicular histology and volume predict the outcome of testicular sperm extraction in patients with non-obstructive azoospermia? Int Urol Nephrol 2006;38:629 635. Van Peperstraten A, Proctor ML, Johnson NP, Philipson G. Techniques for surgical retrieval of sperm prior to intra-cytoplasmic sperm injection (ICSI) for azoospermia. Cochrane Database Syst Rev 2008;CD002807. Verheyen G, Nagy Z, Joris H, De Croo I, Tournaye H, Van Steirteghem A. Quality of frozen thawed testicular sperm and its preclinical use for intracytoplasmic sperm injection into in vitro-matured germinal-vesicle stage oocytes. Fertil Steril 1997;67:74 80. Verheyen G, Vernaeve V, Van Landuyt L, Tournaye H, Devroey P, Van Steirteghem A. Should diagnostic testicular sperm retrieval followed by cryopreservation for later ICSI be the procedure of choice for all patients with non-obstructive azoospermia? Hum Reprod 2004;19:2822 2830. Vernaeve V, Tournaye H, Schiettecatte J, Verheyen G, Van Steirteghem A, Devroey P. Serum inhibin B cannot predict testicular sperm retrieval in patients with non-obstructive azoospermia. Hum Reprod 2002; 17:971 976. Vernaeve V, Verheyen G, Goossens A, Van Steirteghem A, Devroey P, Tournaye H. How successful is repeat testicular sperm extraction in patients with azoospermia? Hum Reprod 2006;21:1551 1554. von Eckardstein S, Simoni M, Bergmann M, Weinbauer GF, Gassner P, Schepers AG, Nieschlag E. Serum inhibin B in combination with serum follicle-stimulating hormone (FSH) is a more sensitive marker than serum FSH alone for impaired spermatogenesis in men, but cannot predict the presence of sperm in testicular tissue samples. J Clin Endocrinol Metab 1999;84:2496 2501. World Health Organization. WHO Laboratory Manual for the Examination of Human Semen and Sperm Cervical Mucus Interaction, 5th edn. Cambridge, UK: Cambridge Press, 2009. Zitzmann M, Nordhoff V, von Schonfeld V, Nordsiek-Mengede A, Kliesch S, Schuring AN, Luetjens CM, Kamischke A, Cooper T, Simoni M et al. Elevated follicle-stimulating hormone levels and the chances for azoospermic men to become fathers after retrieval of elongated spermatids from cryopreserved testicular tissue. Fertil Steril 2006;86:339 347.