Clinical Outcomes and Live Birth Rate (LBR) of Micro-TESE With ICSI-IVF in 968 Non-obstructive Azoospermia Patients From One Reproductive Medicine Center: a Retrospective Cohort Study

Background: Most of data available in the literature reported the sperm retrieval rate (SRR) and ICSI results of microdissection testicular sperm extraction (micro-TESE) in non-obstructive azoospermia (NOA) patients with different etiologies. Unfortunately, there is currently a lack of comprehensive Intent-to-treat (ITT) data to guide clinicians in conducting comprehensive consultations with NOA patients. The aim of current study was to obtain more comprehensive evidence-based data and clinical outcomes for better consultation of NOA patients who opted to undergo micro-TESE combined with ICSI-IVF. Materials and methods: A retrospective study involved 968 NOA patients underwent micro-TESE during the period between January 2015 and December 2019. Those who had successful sperm retrieval and performed ICSI-IVF cycles were included in intent-to-treat (ITT) analysis. The primary outcome measure was the live birth rate (LBR). The cumulative pregnancy or live birth was dened as clinical pregnancy or at least one live-born baby resulting from an ICSI initiated cycle. Two kinds of stratication analyses were performed based on different etiologies of NOA and various amounts of sperm retrieved. The 'Student's t-test was used for comparison of continuous variables. One-way ANOVA was used to assess outcomes among more than two groups. Chi-squared (χ2) or Fisher's exact test was used for proportions. Results: The SRR of all 968 NOA patients undergoing micro-TESE was 44.6% (n=432). ITT analysis was performed in 424 patients, and ICSI-IVF was applied in 362 couples, leading to 171 clinical pregnancies (40.3%) and 161 live-birth deliveries (38.0%) in the rst embryo transfer cycle. No signicant difference was observed for per-protocol analysis between the groups of frozen sperm and fresh sperm in cumulative clinical pregnancy rate (CPR, 51.0% vs. 45.2%) and live-birth rate (LBR, 47.5% vs. 42.9%). NOA patients with Y chromosome azoospermia factor c (AZFc) microdeletions had the lowest rate of the high-score embryo Conclusions: Micro-TESE is an effective sperm retrieval technique for NOA patients. Our data indicated no signicant difference in the LBR between ICSI-IVF cycles using frozen or fresh testicular sperm.

The development of intracytoplasmic sperm injection (ICSI) in 1992 provided a novel opportunity for azoospermia patients to become fathers [7,8]. Sperm obtained through testicular sperm extraction (TESE) was rst performed on NOA patients for ICSI in 1995 [9]. TESE-ICSI then became a routine procedure to treat NOA patients. Studies reported that the sperm retrieval rate (SRR) of TESE in the NOA population was approximately 30-50%. However, these studies' selective bias makes numbers controversial [10][11][12]. Subsequently, various surgical techniques, such as multiple testicular biopsies and ne-needle aspiration, were used to improve SRR and reduce complications in NOA patients, but each has its limitations [13][14][15]. In the past 20 years, microdissection testicular sperm extraction (micro-TESE) has gradually become a popular surgical technique with a high SRR and low tissue loss [16][17][18][19]. However, most of the previous studies were focus on SRR, and few could provide comprehensive follow-up data on ICSI-IVF outcomes of NOA patients, explicitly focusing on the use of fresh or frozen sperm, cumulative pregnancy rate (PR), and live birth rate (LBR), not alone birth defect rate (20)(21)(22)(23)(24).
Our goal was to obtain more comprehensive evidence-based data and clinical outcomes for better preoperative consultation and counseling of NOA patients who opted to undergo micro-TESE combined with ICSI-IVF. This retrospective study aimed to analyze the effects of frozen or fresh sperm, different etiologies, and retrieved sperm quantity on ICSI-IVF outcomes after micro-TESE. All NOA patients were offered the option to receive fresh or frozen testicular sperm for ICSI cycles before surgical sperm retrieval and be informed of their respective advantages and disadvantages. Controlled ovarian stimulation (COS) will be initiated after the sperm is frozen. However, the simultaneous ICSI cycles of fresh sperm and egg collection may result in passive egg freezing due to micro-TESE failure or unimplemented surgery.

Micro-TESE
A single surgeon (GA) performed the micro-TESE surgical procedures as described in the previous literature [25], a few minor and technical modi cations, including the use of saline instead of Ringer's solution as the rinsing uid and suturing the albuginea testis with a 5-0 polypropylene suture instead of 6-0 . Instead of the surgeon grinding the testicular tissue on the operating table, an embryologist grind the  testicular tissue and microscopically search sperm on a bench setting near the operating table [26]. If no sperm was identi ed in one testis, micro-TESE of the contralateral testis was performed immediately.
Evaluation of sperm quantity and quality Sperm count was determined by examining the processed homogenate under high power magni cation (200×), and 20 high power elds (HPF) were microscopically totaled. Motile spermatozoa rate was determined (motile spermatozoa rate = motile sperm/total sperm × 100%). If the total number of sperm was greater than 100, the number of motile sperm was quanti ed in 100 sperm. The sperm deformity rate was assessed in the same manner as the motile spermatozoa rate.

Testicular tissue suspension cryopreservation and thawed
After testicular tissue was retrieved from the testis and transferred to the embryology laboratory, a freezing process was immediately performed. First, testicular tissue was further processed to release spermatozoa. Procedures were performed at room temperature as follows: wash testicular tissue with 1 ml G-MOPS-Plus uid (Vitrolife, Sweden) to remove the red blood cells; Move testicular tissue into a new petri dish and add 1 ml new G-MOPS-Plus uid and grind the tissue into tiny patches by using microscopic forceps to release spermatozoa; Transfer all uid and tissues into a 15 ml centrifuge tube; Add 1 ml fresh G-MOPS-Plus uid to the petri dish to wash additional tissue from the dish and transfer into the 15-ml tube; Incubate at room temperature for ve minutes; Aspirate the supernatant to a fresh new 15-ml centrifuge tube; Add 1ml G-MOPS-Plus uid to the pellet, and transfer the supernatant to the fresh new 15-ml tube. Centrifuge the tube at 400g for 10 minutes, and remove all the supernatant. Next, the freezing procedures were performed as follows: add freezing medium (Test York Buffer with Gentamicin Sulfate) (Irvine Scienti c, USA) and G-MOPS-Plus in a 1:1 ratio into a pellet and resuspend, aliquot resuspended pellet into 2-4 cryovials; and place at t 4℃ for 30 minutes; Resuspend the mixture prior to placing in liquid nitrogen for 1 hour. Transfer the cryovials into a liquid nitrogen container for longterm cryopreservation. The thawing process was performed as follows. Remove the cryovial from the liquid nitrogen container, place it at room temperature for 10 minutes; Transfer the thawed liquid to a fresh 15ml centrifuge tube; Add drops of IM washing buffer (Vitrolife, Sweden) to the tube, and mix gently. Centrifuge at 400g for 10 minutes, and then remove the supernatant. Resuspend the pellet with 1-2 ml IM washing buffer. After the second wash, resuspend the sperm pellet with 50-100 μl G-MOPS-Plus uid. The count, motility, and deformity rate of thawed sperm were reassessed.

Ovarian stimulation and oocytes retrieval
In couples who had testicular sperm was retrieved and frozen, or in couples undergoing a synchronous micro-TESE-ICSI procedure, female partners underwent ovarian stimulation using recombinant FSH or hMG combined GnRH antagonists or GnRH-a [27]. Oocyte-cumulus complexes were recovered 36h after administering 5000 or 10 000 IU of hCG.
ICSI procedure, embryo culture, and transfer For couples who had sperm cryopreserved, sperm were thawed only when females had oocytes retrieved. For couples who underwent synchronous micro-TESE-ICSI treatment, oocytes were vitri ed if no sperm retrieved. Next, ICSI-IVF and fertilization assessment were performed as previously described by Liu [27].
Fertilization rates were expressed as the percentage of oocytes with two distinct pronuclei per injected metaphase II oocytes. Embryos were scored by their morphological appearance according to the Society for Assisted Reproductive Technology scoring system [28]. Normally cleaving embryos with ≥5 cells and ≤20% fragmentation were considered eligible for transfer. Up to two embryos were transferred into the uterine cavity on day 3 (preferred) or day 5 (blastocyst culture was carried out only when no embryos were available on day 3) after injection. The remaining embryos were frozen directly for the next thawed transfer cycles.

Pregnancy follow-up
Pregnancy was diagnosed by elevated serum hCG levels (≥25IU/L) 14 days after embryo transfer. Clinical pregnancy was de ned as a visible gestational sac at transvaginal ultrasound 4-5 weeks after embryo transfer. Pregnancy loss was de ned as the loss of a clinical pregnancy before 28 weeks of gestation. Live birth was de ned as the birth of at least one living child, irrespective of gestation duration. The cumulative pregnancy or live birth was de ned as clinical pregnancy or at least one live-born baby resulting from an ICSI-IVF initiated cycle.

Statistical analysis
Statistical analyses were performed with SPSS statistical software for Windows, version 22.0 (SPSS, Chicago, IL, USA). Continuous variables were expressed as mean ± SD. The 'Student's t-test was used for comparison of continuous variables. One-way ANOVA was used to assess outcomes among more than two groups. Chi-squared (χ2) or Fisher's exact test was used for proportions. Differences were considered statistically signi cant when the p-value was <0.05.
Outcomes of ICSI-IVF ICSI-IVF was performed in 362 couples in the ITT set, leading to 171 clinical pregnancies and 161 livebirth deliveries in the rst embryo transfer cycle (Fig. 1). All patients had completed at least one embryo transfer by the end of this follow-up, and 25 couples have not yet achieved live birth but still have cryopreserved embryos. All couples undergoing the ICSI-IVF cycle were divided into frozen sperm group and fresh sperm group (323 cases with frozen sperm vs. 101 cases with fresh sperm (Fig. 1). In ITT analysis, baseline demographic data were comparable between the two groups and fertilization rate, 2 primary nucleus (PN) rate, numbers of available and high-score embryos on day 3 (Table 1). Though the clinical pregnancy rate (CPR) in the rst embryo transfer cycle was lower in frozen sperm than that in the fresh sperm group (37.5% vs. 49.5%, p < 0.05, Table 1), there was no signi cant difference in live birth rate (LBR) between these two groups (35.9% vs. 44.6%, p > 0.05, Table 1).
In the PP set, 299 couples were undergoing ICSI-IVF treatments (257 cases with frozen sperm vs. 42 cases with fresh sperm, Fig. 1), leading to 136 clinical pregnancies and 130 live-birth deliveries in the rst embryo transfer cycle. There were no signi cant differences in outcomes of embryo culture, including fertilization rate, day 3 utilization rate of MII eggs for ICSI, and high-score embryo rate on day 3, between these two groups (Table 3). Besides, the percentage of patients with no available embryo was low in both frozen and fresh sperm groups and was not signi cantly different between the two groups (11.3% vs. 14.3%, Table 3). Then, accumulative embryo transfer cycles were analyzed in PP set, and there was no signi cant difference in cumulative CPR between the frozen and fresh groups (51.0% vs. 45.2%, p = 0.491). Also, no signi cant difference was observed in the cumulative LBR (PP: 47.5% vs. 42.9%, p = 0.579) between the two groups (Table 3).
There were three cases of birth defects recorded. One case is a cardiovascular malformation, one is a cleft lip and palate, and the other is congenital hypospadias (  (Table 3). Overall, LBR was 17.6% (170/968) in the rst embryo transfer cycle for NOA patients with different etiologies who underwent the micro-TESE procedure for the rst time, regardless of fresh or frozen testicular spermatozoa (Fig. 1). #: 3 cases of birth defects were recorded: one of them is a boy with congenital hypospadias whose father was cryptorchidism and come from fresh sperm. Two were all from frozen sperm: one case is a girl with cardiovascular malformation whose father was Klinefelter syndrome; and the other is also a girl with cleft lip and palate whose father was idiopathic NOA patient.

Discussion
Our study aimed to comprehensively analyze the outcomes of the ICSI-IVF cycle of NOA patients who underwent micro-TESE. This study provides a tremendous clinical basis for reproductive medicine specialists to better understand the clinical Outcome of micro-TESE combined with ICSI-IVF cycle, which allows for more comprehensive preoperative clinical consultation and preparations.
For NOA patients, SRR has been the most studied clinical outcome. Previously, SRR has often been different in studies examining outcomes after micro-TESE, likely and mainly due to the limited amount of data and patient selection bias [19,29]. In this study, we comprehensively analyzed micro-TESE data from 968 men with NOA diagnosis, and found that the SRR was 44.6%. The SRR did differ in NOA patients with different etiology. The SRR was highest in orchitis with 81.2%, followed by Y chromosome AZFc microdeletions (68.6%), cryptorchidism (62.4%), KS (43.6%), and lowest in idiopathic NOA (31.1%).
Not all NOA patients have access to fresh sperm, and couples who want to synchronize the ICSI-IVF cycles with fresh sperm and oocyte have to face the risk of oocyte freezing and sperm donation-IVF. Therefore, sperm cryopreservation before ICSI may be more reasonable and reduce unnecessary risks for females. Cryopreservation of testicular sperm has long been used in assisted reproductive technology [30,31]. Previous studies con rmed signi cant differences between fresh and frozen sperm by ejaculation, regardless of total sperm count, motility, or morphology of sperm [32]; however, the differences contributed little to the outcome of ICSI cycles [33,34]. Due to controversial evaluations with NOA patients, ICSI-IVF outcomes are controversial because of different cryopreservation methods [20,[35][36][37].
Our study reported an improved laboratory technique of testicular tissue suspension cryopreservation in NOA patients after sperm retrieval from micro-TESE. We established a technique and method to assess the quantity and quality of sperm by counting the average number of sperm in high power and calculating the percentage of motile sperm after standardizing the total volume of sperm suspension.
This evaluation method is simple, reliable, and can be easily used as a conventional assessment method, even for too extremely few spermatozoa patients. We found no difference in the average number and the percentage of motile spermatozoa between frozen-thawed sperm and fresh sperm. At the same time, frozen-thawed sperm with ICSI resulted in a similar fertilization rate (70.1% vs. 68.2%, p = 0.403) and day 3 utilization rate of oocytes (28.7% vs. 26.2%, p = 0.286) compared with fresh sperm. The cumulative live birth rate (LBR) of one ICSI cycle in the frozen-thawed sperm group was 47.5%, high up to the same as conventional ICSI-IVF as reported [38].
According to previous researches, the clinical pregnancy rate between fresh sperm and frozen sperm remains controversial. Park reported that patients with frozen spermatozoa had signi cantly higher pregnancy and implantation rates than fresh sperm [39]. Others showed that the fertilization rate and clinical pregnancy rate were higher in fresh sperm from non-mosaic KS patients by TESE [40]. A systematic review and meta-analysis revealed that in men with NOA showed that the ICSI-IVF outcome was not affected by whether the retrieved testicular sperm is fresh or frozen [41]. Nevertheless, cumulative pregnancy or cumulative live birth was not mentioned in any of the above studies, especially the cumulative live birth. After comparing the outcomes of ICSI-IVF cycles between fresh and frozenthawed spermatozoa, we found that there were no signi cant differences in all laboratory parameters, not only fertilization rate and day 3 oocytes utilization rate we mentioned above, but also cleavage rate, rate of the high-score embryo, and the ratio of patients who had no available embryos. The main results we found are consistent with previous studies [20,37,42]. It is worth to mention that about 1/7 to 1/9 patients may not be able to obtain available embryos according to one ICSI cycle. This detailed data will help physicians provide with su cient counseling and avoid patients` over-high expectations of treatment.
Then, we conducted a further analysis of clinical outcomes after embryo transfer, including the rst embryo transfer cycle in ITT and PP analysis and cumulative embryo transfer cycles in PP analysis. Our results showed that CPR in the rst embryo transfer cycle was lower in frozen sperm with ITT analysis; however, not signi cantly different in PP analysis. This difference may result from social factors; for example, a high proportion of patients with frozen sperm in the ITT set who did not receive ICSI treatment after sperm acquired for divorce, singlehood, nally change for sperm donation or loss of communication. In PP analysis, the CPR or cumulative CPR and LBR showed no signi cant differences between the fresh and frozen groups after excluding patients with preimplantation genetic testing treatment (spouse's chromosome showed a Robertson translocation), warmed oocytes cycles, or those men whose spouse had history of recurrent pregnancy loss and uterine abnormality, and those without ICSI treatment. Therefore, more than 40% NOA patients can achieve live birth in one ICSI cycle, no matter with frozen or fresh sperm (47.5% and 42.9%).
Few studies provided a detailed analysis of ICSI-IVF outcomes according to the different pathological types and etiologies of NOA patients [20,21,43]. De Croo et al. reported maturation arrest had a lower fertilization rate and clinical pregnancy rate compared with hypospermatogenesis and Sertoli Cell Only syndrome [23]. Some studies demonstrated no signi cant differences in ICSI-IVF outcomes between different histopathological subsets in NOA [40,44] [45,46]. Our results showed that lower day 3 oocytes utilization rate and high-score embryo rate and lower cumulative CPR and cumulative LBR were observed in patients with Y chromosome AZFc microdeletion. These results are also consistent with those reported by Van et al. [47], and the primary function of the AZFc region in the Y chromosome is involvement in spermatozoa quality or function than in spermatogenesis.
Besides, we found that compared to patients who got more sperm (> 20 approximately), NOA patients with fewer sperm (≤ 20 approximately) were detected with signi cantly higher serum FSH level, lower oocytes utilization rate, lower high-score embryo rate, and a higher ratio of cycles without available embryos. Some studies reported that sperm from NOA patients have aneuploidy, mosaicism, and DNA damage that contribute to decreased clinical outcomes [22,48]. Similarly, in this study, a signi cantly lower clinical pregnancy rate and live birth rate were observed as expected in patients with fewer sperm. These results provide us with a better understanding of treatment outcomes for patients with different laboratory ndings after testicular tissue processing.
Before micro-TESE surgery was conducted, complete clinical consulting was performed regarding the treatment protocol selection and male factors that may in uence the outcome of ICSI-IVF treatment. In addition, the female's ovarian reserve would be assessed in detail before the eventual decision on the conduction of micro-TESE surgery to reduce the risk of cycle cancellation due to the female factors, especially for couples who had di culties accepting the consequences of failure. In this study, advanced female age (≥ 40 years old, 6 cases), diminished ovarian reserve (DOR, AMH ≤ 1ng/ml, 16 cases), or poor ovarian response (POR, ≤ 4 oocytes retrieved, 29 cases) were all de ned as risk factors for adverse outcomes, and patients would be informed of high risk of ICSI-IVF failure. All of these patients treated with ICSI in this study lead to one live-birth delivery in six patients aged ≥ 40 years, three live-birth deliveries in 16 DOR patients, and 15 live-birth deliveries in 29 POR patients (data not included in the results section). The female-factor combined with male-factor assessment re ned our clinical consultation on ICSI treatment for couples with NOA.

Conclusions
This study's main strength is comprehensively analyzing clinical outcomes for NOA patients undergoing micro-TESE and ICSI-IVF cycle. Information and data from both male and female sides were included in our analysis; besides, we included CPR and LBR with intent-to-treat (ITT) analysis and per-protocol (PP) analysis minimally or barely reported in the past literature. In addition, female factors that may affect the ICSI-IVF outcomes were excluded in the PP analysis, including recurrent pregnancy loss, abnormal karyotype, known uterine anomalies, and adenomyosis, improving extrapolation of our results. However, our study still has limitations, including the sample size of the fresh sperm group was not large enough, the need to supplement follow-up data on live births, and patient selection bias. In the future, multicenter data and randomized controlled trials are needed to determine clinical predictors of successful outcomes for NOA couples.
Abbreviations NOA: non-obstructive azoospermia; Micro-TESE: microdissection testicular sperm extraction; SRR: sperm retrieval rate; LBR: live birth rate; ICSI-IVF: intracytoplasmic sperm injection in-vitro fertilization; ITT: intentto-treat; CPR: cumulative clinical pregnancy rate; PP: per-protocol; CI: Con dential interval; OR: Odds ratio. University (reference number 2017-055) and was carried out in accordance with theHelsinki Declaration. Due to the retrospective nature, informed consent was not required, and patients' data were used anonymously.

Consent for publication
Not applicable.
Availability of data and materials The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.