Clinical predictors of successful outcomes for couples with nonobstructive azoospermic male partners undergoing micro-TESE

Abstract

Nonobstructive azoospermia (NOA) is the most challenging and complex clinical scenario for infertile men. Besides circumstances such as hypogonadotropic hypogonadism, surgical sperm retrieval is typically necessary, and microdissection testicular sperm extraction (micro-TESE) is the procedure of choice for men with NOA desiring to father children with their own gametes. Micro-TESE results in the highest numbers of sperm cells retrieved for use with in vitro fertilization/intracytoplasmic sperm injection (ICSI) in comparison to all other techniques for surgical sperm retrieval in men with NOA. Several factors may affect sperm retrieval rate and ICSI outcomes, including the patient’s age, testicular volume, histopathological and genetic profile, and serum hormone levels. This article aims to review the medical literature describing predictors of successful micro-TESE and the outcomes of ICSI in men with NOA.

INTRODUCTION

The majority of men with nonobstructive azoospermia (NOA) do not have an identifiable underlying etiology. NOA may be classified in terms of pretesticular etiology or testicular failure. Pretesticular azoospermia is the failure of sperm production due to a hypothalamic or pituitary abnormality, presenting as hypogonadotropic hypogonadism (HH). This could be due to a congenital abnormality (normosmic idiopathic HH or anosmic HH like Kallmann syndrome) or acquired abnormalities such as a pituitary tumor, pituitary or hypothalamic trauma, and hormone therapy impact on the hypothalamic–pituitary function (e.g., with exogenous testosterone). Azoospermia due to primary testicular failure is the impairment of sperm production due to testicular cellular dysfunction, presenting as HH. This may be due to congenital abnormalities (chromosomal inversion/translocation, Klinefelter syndrome, and Y-chromosome microdeletions of azoospermia factor a [AZFa]/AZFb/AZFc), cryptorchidism, or may be acquired (chemotherapy, radiation therapy, infection, and testicle traumatic injury).1,2 The histopathological classification could be used to divide testicular failure into different patterns: hypospermatogenesis, maturation arrest, and Sertoli cell-only pattern (Table 1).

Table 1.

Etiological factors of nonobstructive azoospermia

HH	Pretesticular NOA	Testicular NOA
Congenital/genetic	-Normosmic idiopathic HH	-Klinefelter syndrome
	-Anosmic HH (Kallmann syndrome)	-Y-chromosome microdeletions
		-Cryptorchidism
		-Idiopathic
Acquired	-Pituitary tumor	-Chemotherapy
	-Pituitary or hypothalamic trauma	-Radiation therapy
	-Through exogenous testosterone or hormone therapies	-Infection
		-Trauma

HH: hypogonadotropic hypogonadism; NOA: nonobstructive azoospermia

One percent of the general population and 10%–15% of males presenting for fertility assessment are found to be azoospermic.3,4,5 Approximately 60% of men with azoospermia have NOA.1,6 Primary testicular failure affects 49%–93% of males with azoospermia, while pretesticular azoospermia affects 2% of azoospermic cases.1,6,7,8 Multifactorial control governs spermatogenesis, including age, testicular volume, histopathological genetic profile, and serum hormone levels.9 These factors could be considered potential predictors of sperm retrieval, and several studies have evaluated their impact and usefulness to avoid unnecessary procedures.10

Isolating sperm through microdissection testicular sperm extraction (micro-TESE) is considered the gold-standard technique to yield the highest odds of surgical sperm retrieval (SSR) rate. This is critical to ultimately assist the couple in achieving the goal of having their own biological offspring. However, a number of obstacles may preclude the ability to reach the final desired outcome, including the chance of sperm retrieval. It is helpful to counsel the infertile couples on the odds of successful outcome based on available data on potential predictors of success in each step. The current medical literature assessed the reliability of potential predictors of success of several outcomes, including sperm retrieval, surplus sperm retrieved beyond a single in vitro fertilization (IVF)/intracytoplasmic sperm injection (ICSI) cycle, embryology progression to embryo transfer (ET), clinical pregnancy (CP) rate, miscarriage rate, and finally live birth (LB) rate. The aim of this review is to discuss the predictors of successful micro-TESE and ICSI outcomes in men with NOA.

PREDICTORS OF SPERM RETRIEVAL IN MEN WITH NOA UNDERGOING MICRO-TESE

Age

Deng et al.11 published a retrospective study of 200 men with NOA who underwent micro-TESE with an overall SSR rate of 33.5% and demonstrated there was no difference in age in men who had a successful SSR versus a failed SSR. A narrative review of 77 studies suggested that neither age nor body mass index predicted SSR.9 Other retrospective studies supported the finding that age did not impact the odds of SSR with micro-TESE. One of these studies which showed no difference based on age included 72 men with a 70.8% SSR rate,12 and another included 1067 men with an SSR rate of 56.6%,12,13

Testicular volume

The retrospective study by Deng et al.11 revealed that testicular volume was not a predictor of successful SSR with micro-TESE. This study included 67 patients with an overall SSR rate of 33.5%.11 A previous meta-analysis, by Li et al.14, including 4354 patients reported no correlation between testicular volume and SSR with micro-TESE. Furthermore, severely atrophic testicles were not correlated with negative SSR with micro-TESE in a review including 1127 men with an overall SSR rate of 56%.15 However, one retrospective study reported that testicular volumes ≥15 ml were associated with poor sperm retrieval rates on reviewing the outcomes in 640 patients, with an overall SSR rate of 44.5% in men with Sertoli cell-only histopathological patterns.16 A narrative review suggested that testicular volume may predict successful SSR in men with NOA, although there is controversy, and no threshold testicular volume has been established to predict successful SSR.9

Reproductive hormones

Deng et al.11 reported a significantly lower median follicle-stimulating hormone (FSH) level in men with a successful SSR with micro-TESE (17.2 mIU ml⁻¹) in comparison with those who did not have sperm retrieved (23.2 mIU ml⁻¹).11 There was no difference in luteinizing hormone (LH), prolactin, estradiol (E2), or testosterone levels in the successful group in comparison to the unsuccessful group. There were significantly lower anti-Müllerian hormone (AMH) levels and higher inhibin B levels in the group with successful sperm retrievals in comparison to the unsuccessful group.11 A systematic review including 20 studies with 4760 patients demonstrated that for every 1.19 mIU ml⁻¹ increase in FSH level, a 1% decrease in SSR would be demonstrated, indicating that FSH levels can predict SSR rates in men undergoing micro-TESE. The overall SSR rate for men undergoing micro-TESE in this review was 51.9%.17 Serum AMH was independently assessed and was found to be a moderate predictor of SSR with an AMH cut-off of 0.133 ng ml⁻¹ with a sensitivity of 0.91 and a specificity of 0.29 for SSR in 46 men who had an overall SSR rate of 61.1%.18 A narrative review indicated that testosterone, LH, and prolactin do not predict SSR, but FSH and E2 have controversial data on their predictive ability.9 A study of 294 men with NOA suggested that a higher FSH level was a negative prognostic predictor of SSR.19 A retrospective review of 100 men who underwent micro-TESE for NOA reported that serum FSH, testosterone, and inhibin B levels were the most robust predictive factors for successful SSR; they reported an overall SSR rate of 40%.20 One study defined FSH levels between 10 mIU ml⁻¹ and 15 mIU ml⁻¹ as predictive for poorer SSR rates.16

Conversely, another study found that FSH did not predict SSR odds.12 A similar observation was reported by Ramasamy et al.,21 whose study included 792 men with an overall SSR rate of 60%.14 However, the former study reported that higher FSH levels had a negative relationship with surplus sperm being retrieved to be adequate for use for additional IVF/ICSI cycles beyond the initial cycle.12

Etiology of NOA

Etiologies of NOA have been well established, such as Klinefelter syndrome, Y-chromosome microdeletions, cryptorchidism, mumps orchitis, exposure to gonadotoxins such as chemotherapy or radiation therapy, and cases in which an etiology is not identifiable are categorized as idiopathic (Table 1). A retrospective review by Gao et al.22 of 335 NOA men who underwent micro-TESE studied the odds of SSR correlating them with the etiology of NOA. In that study, the overall SSR rate was 40.9% with the poorest SSR rates in men with idiopathic NOA (31.22%) compared to Klinefelter syndrome (48.6%), Y-chromosome microdeletion (60.9%), cryptorchidism (80.9%), and mumps orchitis (75%).22 A study suggested the finding of Klinefelter syndrome and Y-chromosome microdeletion of the AZFc region as positive predictors of SSR, in contrast with the finding of AFZa and AZFb microdeletions which have poor SSR outcomes.9 A systematic review and meta-analysis by Corona et al.23 including 37 trials with 1248 patients reported that SSR in men with Klinefelter syndrome offers a nearly 50% chance of retrieving sperm with subsequent pregnancy rates and live birth rates of approximately 50%.

A retrospective study of 509 men with NOA suggested that the etiology of NOA impacted the SSR rates, with an overall SSR rate of 40.3% and significantly higher SSR rates in men with a history of mumps orchitis or cryptorchidism as etiologies of NOA in comparison to men with idiopathic NOA, Klinefelter syndrome, AZFc microdeletions, or men who received chemotherapy.24 However, Deng et al.11 demonstrated that the etiology of NOA did not predict the SSR. A clear exception that has been well established is the poor SSR outcomes in men with AZFa and AZFb microdeletions in contrast to men with AZFc microdeletions who demonstrate a higher SSR rate.25

A recent review exploring the potential predictors of SSR in men with NOA undergoing micro-TESE suggested that there is no clinical predictor that should be considered valuable enough to discourage the attempt for SSR based on the current level of evidence in the medical literature.26 This review accounted for etiologic factors, including Klinefelter syndrome, Y-chromosome microdeletions, cryptorchidism, and varicocele, as well as clinical factors such as testicular volume, serum hormonal parameters, testicular histology, and seminal plasma molecular markers. The authors of this review suggested that men with NOA with complete AZFc Y-chromosome microdeletions and a history of cryptorchidism may have the highest odds of SSR.26

Testicular histology

The severity of the pattern of testicular histology has been correlated with the odds of successfully retrieving sperm with micro-TESE in a study of 143 men.27 A report suggested that the severity of testicular histology abnormality was a powerful negative predictor of SSR in men with NOA.9 A recent study agreed that men with less severe histology patterns such as hypospermatogenesis had higher SSR rates, as well as having higher odds of having surplus sperm retrieved for subsequent IVF/ICSI cycle.12 A retrospective study reported that men with a Sertoli cell-only pattern had the poorest SSR rates.16 Although most data correlate more severe testicular histology with poorer SSRs, there is still some controversy.14 A study including 86 men with an overall SSR rate of 71% suggested an association between more severe forms of abnormal testicular histological patterns in men with NOA, higher FSH levels, and smaller testicular volumes.28

Seminal plasma biomarkers

Multiple seminal plasma biomarkers have been studied for their ability to predict SSR through micro-TESE in men with NOA. These include seminal plasma RNAs, metabolites, AMH, inhibin B, leptin, survivin, clusterin, lectin galactoside-binding soluble 3 binding protein (LGALS3BP), extra-embryonic tissue-spermatogenesis-homeobox gene 1 (ESX1), testis-expressed gene 101 (TEX101), transition nuclear protein 1 (TNP1), deleted in azoospermia (DAZ), protamine-1 (PRM1), and PRM2. A narrative review of 21 studies suggested that seminal AMH and inhibin B do not predict SSR in these men, and leptin, survivin, clusterin, LGALS3BP, ESX1, TEX101, TNP1, DAZ, PRM1, and PRM2 may have potential for prediction of SSR, but the data reviewed in these studies did not show a definitive correlation, and further large prospective studies are needed to elucidate these as potential predictors.29 A study assessing seminal phosphoglycerate kinase 2 (PGK2) and acrosin (ACR) protein expression revealed that men with NOA who had a successful SSR by micro-TESE demonstrated higher concentrations of PGK2 and ACR proteins in the seminal plasma.30 Another study including 41 men with a 44% SSR rate found a positive correlation between seminal plasma extracellular vesicles tRF-Val-AAC-010 and successful SSR by micro-TESE in men with NOA.31 Another study evaluated seminal microRNAs as a potential predictor of SSR in men with NOA, but it was limited by a very small sample size of 6 in each group.32

Seminal plasma extracellular vesicles and PIWI-interacting RNAs (piRNAs) demonstrated favorable SSR in men with NOA undergoing micro-TESE in a small series of eight men.33 A larger study of 96 men with NOA created a seminal plasma biomarker panel of nine testis-specific extracellular vesicle long noncoding RNAs (lncRNAs) with a calculated score higher than 0.532 correlating with a positive SSR.34

Newly emerging predictors

Several new predictors of SSR in men with NOA undergoing micro-TESE have been evolving in recent years. An example is testicular tissue touch print smear (TPS) which includes the process of obtaining a testicular sample intraoperatively, smearing it on a sterile slide, staining it with thionine, and microscopically examining it for the presence of sperm.35 One hundred and forty-eight men had TPS performed and had the results compared to bulk retrieved tissue specimens that were processed and evaluated by the IVF laboratory for the presence of sperm for use with IVF/ICSI. The TPS test was found to have a 100% positive predictive value and a 95.5% negative predictive value for predicting the presence or absence of sperm in the IVF laboratory, respectively. In the same study, needle testicular biopsies were performed in a separate group of 360 azoospermic men, and histology was compared to TPS results in these men. A 93.6% correlation with the histologic diagnosis was found.35

Another predictor studied was circular RNA monoglyceride lipase (circ_MGLL) found in testicular tissue. One hundred and fourteen men with NOA who underwent micro-TESE had circ_MGLL expression from the testicular tissue assessed.36 In that study, higher levels of circ_MGLL expression were inversely related to the odds of SSR. A nomogram was designed to predict SSR based on circ_MGLL in conjunction with serum FSH, LH, and testosterone levels, and histopathology.36 Another study aimed to assess seminal circular RNAs as a predictor for SSR in men undergoing micro-TESE. The investigators assessed circular RNAs hsa_circ_0060394, hsa_circ_0000277, and hsa­_circ_0007773, but the outcomes were significantly limited by the very small sample size of 25.37 Seminal proliferating cell nuclear antigen (PCNA) and LIM15 gene expression levels have been studied for their ability to predict successful SSR rates in men with NOA. The PCNA and LIM15 gene expression levels in testicular tissue and the LIM15 gene expression in the ejaculate, determined by real-time polymerase chain reaction (PCR), significantly correlated with the odds of SSR with micro-TESE. Although PCNA gene expression levels measured in the ejaculate were higher in men with successful SSR, the difference was not statistically significant.27

Bastug et al.38 published a study assessing hematologic inflammatory parameters in the serum to predict the outcomes in micro-TESE cases. They reported that the neutrophil–lymphocyte ratio, platelet–lymphocyte ratio, and calculated systemic immune inflammation index were predictors of successful SSR, and the neutrophil–lymphocyte ratio was the most robust predictor of outcomes.38

A small series reported that 12 men with NOA who had successful SSR had higher signal intensities of choline metabolite signal and creatine signal compared to those of 12 fertile controls, which were measured by magnetic resonance spectroscopy. Eighteen men who did not have sperm retrieved at the time of micro-TESE had significantly lower signal intensities of choline metabolite signal and creatine signal.39 In another small case–control study, plasma exosomal transfer RNA-derived fragments (tRFs) were assessed as potential predictors of SSR outcomes in men with NOA undergoing micro-TESE. Expression of plasma exosomal tRF-Gly-GCC-002 and tRF-Glu-CTC-005 was reported as positive predictors of SSR.40

The presence of Beclin-1, an autophagy gene suggestive of testicular damage, was assessed for 62 men with NOA undergoing micro-TESE by real-time PCR in testicular tissue as a potential predictor of SSR. Beclin-1 expression was significantly upregulated in men with histology revealing a Sertoli cell-only pattern. A logistic regression model suggested that the level of Beclin-1 expression is associated with failed SSR.41 A case–control series of testicular tissue samples from 32 men who had successful SSR with micro-TESE was compared to 25 who failed SSR. There were nine controls who were obstructive azoospermics. All groups assessed for testis-specific genes zona pellucida binding protein 2 (ZPBP2), PGK2, and acrosomal vesicle protein 1 (ACRV1) by DNA flow cytometry in testis biopsy samples. The haploid cell percentage was significantly higher in the NOA group with positive SSR in comparison to the NOA group with negative SSR. Expression of ZPBP2 and PGK2 was higher in the samples with higher haploid cell percentages. Sperm quality when retrieved was positively associated with PGK2 expression. The results suggest that the expression of ZPBP2 and PGK2 may be useful predictors of successful SSR.42

A narrative review suggested that contrast-enhanced scrotal ultrasound may predict SSR in men with NOA.9 A retrospective study including 806 men with NOA suggested that gray-scale testicular ultrasound determination of seminiferous tubule size may be used to predict positive SSR with micro-TESE. However, the overall SSR was fairly low in these 806 men at 29.8%.43

PREDICTORS OF SPERM RETRIEVAL IN SALVAGE MICRO-TESE FOLLOWING INITIAL TESE

A recent meta-analysis and systemic review by Zhang et al.44 included four retrospective studies with 332 men. These were men with NOA who underwent a salvage micro-TESE following a previously failed initial micro-TESE. It also included three retrospective studies of 177 men with NOA who underwent a salvage micro-TESE following a failed initial conventional TESE. In those who failed an initial micro-TESE, predictors of successful SSR at the time of salvage micro-TESE included younger age, smaller testicular volumes, lower FSH levels, lower LH levels, and histology consistent with hypospermatogenesis. In men who underwent salvage micro-TESE following a failed conventional TESE, testicular histology consistent with hypospermatogenesis was found to be a predictor of SSR success.44

A study assessing multiple repeat micro-TESEs performed following an initial successful micro-TESE evaluated success rates in the subsequent micro-TESEs and factors correlated with successful SSR in the repeat case.45 In that study, 88, 61, and 40 men with NOA who had previously undergone micro-TESE underwent a 2^nd, 3^rd, and 4^th micro-TESE, respectively. Successful SSR was reported in 73.9%, 86.9%, and 92.5% of men undergoing 2^nd, 3^rd, and 4^th micro-TESE procedures. These data suggest that repeat micro-TESE offers high sperm retrieval rates when the sperm retrieved from the initial micro-TESE for use with IVF/ICSI is exhausted.45

It is well established that micro-TESE is the SSR technique of choice in men with NOA; however, it is not uncommon to see men presenting who have previously undergone failed conventional TESE. A multicenter cross-sectional study revealed that SSR was successful in 49.2% of men who underwent micro-TESE following failed cTESE.26

PREDICTORS OF ART OUTCOME IN MEN WITH NOA UNDERGOING MICRO-TESE

The majority of the medical literature evaluates SSR predictors, which is of course the necessary first step to be able to proceed further. However, there is a paucity of data on predictors of artificial reproductive technology (ART) outcomes in men with NOA who underwent micro-TESE for use with IVF/ICSI. Predictive data on ART outcomes are important for counseling these couples on outcomes beyond simply SSR, such as CP rate and LB rate.

Age

The male partner’s age at the time of micro-TESE did not impact CP rates or LB rates. However, a study reported an unexpected finding that older men had higher odds of their embryos reaching the Day 5 blastocyst stage for embryo transfer.12 The median age was 35 of the patient cohort, with only 15% of men in their 40s, 2.8% in their 50s, and 2.8% in their 60s. Although age was found to be a statistically significant predictor of embryos reaching embryo transfer, most men in this cohort were below the ages considered to be advanced paternal age.12

Testicular volume

Testicular volume did not impact CP rates or LB rates when sperm was retrieved through micro-TESE in men with NOA for IVF/ICSI.12

Reproductive hormone profile

FSH levels were not predictive of CP or LB rates when micro-TESE sperm was used in couples undergoing ICSI.12

Etiology of NOA

The retrospective study assessing not only SSR rates, but IVF/ICSI outcomes based on the etiology of NOA by Gao et al.22 demonstrated no difference in CP or LB rate with IVF/ICSI using micro-TESE retrieved sperm, between men with different etiologies including idiopathic azoospermia (68.6% and 71.4%), Klinefelter syndrome (65% and 53.8%), Y-chromosome microdeletion (44.4% and 50%), cryptorchidism (66.7% and 75%), and mumps orchitis (85.7% and 66.7%), respectively.

A retrospective case–control study revealed that sperm used from men with a Y-chromosome AZFc microdeletion resulted in poorer clinical outcomes when used with ICSI (345 cycles) in comparison to idiopathic patients with no Y-chromosome microdeletions (363 cycles). Both groups included men who were NOA and had sperm retrieved with micro-TESE as well as severely oligozoospermic men necessitating ICSI. Female factor was controlled for by excluding female partners with 35 years of age or older or with serum testing consistent with diminished ovarian reserve. Although previous studies have indicated that an AZFc microdeletion may be a favorable predictor of SSR,26 this study suggests it as a poor prognostic predictor of reaching an ET, CP rate, and LB rate.46

Testicular histology

Men with less severe patterns of testicular histology such as hypospermatogenesis pattern were reported to have better CP rates and LB rates with the use of micro-TESE sperm with IVF/ICSI in comparison to men with more severe histological patterns such as Sertoli cell-only.12

CONCLUSIONS

The areas of near consensus of predictors of SSR and ART outcomes include age not being a predictor of SSR, most data indicating that idiopathic NOA has poorer SSR rates than known etiologies, and most data agreeing that poorer testicular histological patterns are associated with poorer SSR rates. To date, the data on seminal plasma biomarkers and the majority of other biochemical and imaging potential predictors are limited by small sample-sized studies and a lack of larger well-performed studies to reproduce the reported outcome predictors as well as complex technical techniques that may be beyond the scope of what is feasible in clinical practice. Age and testicular volume have not been shown to impact ART outcomes once sperm has been retrieved with micro-TESE. The severity of testicular histology appears to predict poorer ART outcomes. The remainder of the predictors assessed for either SSR or ART outcomes have conflicting data leaving a lack of clarity on their predictive utility. Currently, there may not be robust data to discourage a man with NOA who is a candidate for micro-TESE from pursuing it once or in repeated trials.

AUTHORS CONTRIBUTIONS

PKK carried out study design, literature search, and manuscript writing. NG carried out the literature search and manuscript writing. RS carried out the critical review and revision. All authors read and approved the final manuscript.

COMPETING INTERESTS

All authors declare no competing interests.