- Email: [email protected]
Evaluation of the Azoospermic Patient
0022-534 7/89/1421-0062$02.00/0 Vol. 142, July Printed in U.S.A.
THE JOURNAL OF UROLOGY
Copyright© 1989 by Williams & Wilkins
EVALUATION OF THE AZOOSPERMIC PATIENT JONATHAN P. JAROW, MARK A. ESPELAND AND LARRY I. LIPSHULTZ From the Department of Surgery, Section on Urology, Bowman Gray School of Medicine, Winston-Salem, North Carolina, and Baylor College of Medicine, Houston, Texas
ABSTRACT
Azoospermia is found in up to 10 to 20 per cent of the men who present to an infertility clinic. The main causes are testicular failure and ductal obstruction. Testicular biopsy remains the definitive test used to differentiate these 2 disorders. A retrospective study of 133 azoospermic men was performed to determine the accuracy and limitations of noninvasive variables in predicting testicular failure in an effort to limit the need for diagnostic testicular biopsy. Of 49 patients (37 per cent) with ductal obstruction a third had bilateral vasal agenesis. The remaining 84 azoospermic patients (63 per cent) had testicular failure. The results of the complete evaluation of these patients are described. Among the 101 patients with a testicular biopsy confirmed diagnosis there was a significant difference in testicular size (p less than 0.001), ejaculate volume (p less than 0.001) and serum follicle-stimulating hormone (p less than 0.001) between patients with testicular failure and those with ductal obstruction. The sensitivity and specificity of various parameters were determined. The best criteria to predict ductal obstruction preoperatively are a serum follicle-stimulating hormone level of less than 2 times greater than normal and the absence of bilateral testicular atrophy (100 per cent sensitivity and 71 per cent specificity). An algorithm for evaluation of the azoospermic patient is described such that all men with ductal obstruction and a minimal number with testicular failure undergo testicular biopsy. (J. Urol., 142: 62-65, 1989) The prevalence of azoospermia in the general population has been estimated to be 2 per cent. 1 The incidence at a male infertility clinic has been reported to be as high as 10 to 20 per cent. 2• 3 The over-all prognosis of fertility in men with azoospermia is poor. The majority of these patients have untreatable disorders and, unfortunately, the success rates for the treatable disorders are low. For this reason many patients and physicians have been discouraged from pursuing the evaluation and treatment of this problem. The main differential diagnosis in the azoospermic patient in terms of treatment and prognosis is between testicular failure and obstruction of the excurrent ducts of the testis. Several diagnostic tests to determine the presence and level of obstruction have been described, including measurement of ejaculate volume, biochemical assays of the ejaculate (fructose, carnitine,4 a-glucosidase 5 and glycerylphosphorylcholine6 ) and serum hormone determinations. 7 • 8 All have varying degrees of accuracy that have not been substantiated fully. Despite the availability of these diagnostic tests, testicular biopsy still is necessary to establish the presence of normal spermatogenesis in patients with suspected ductal obstruction. This purpose remains 1 of the few indications for testicular biopsy in the subfertile patient. In addition, scrotal exploration with epididymal inspection and/or vasography is mandatory before any reconstruction is performed. We developed a rational approach to the evaluation of the azoospermic patient through an evaluation of the causes of this disorder, and the measurement of the sensitivity and specificity of various noninvasive diagnostic tests. Ideally, testicular biopsy should be avoided in as many patients as possible without the exclusion of any potentially treatable patient with ductal obstruction.
upon examination of a centrifuged semen sample were reviewed. Patients with either aspermia or a history of elective vasectomy were excluded. Mean age of these 133 patients was 31 years (range 22 to 56 years). The patient population was skewed in that these patients were seen at a tertiary fertility center. Referral sources were a urologist in 65 patients (49 per cent), gynecologist in 45 (34 per cent) and other sources (endocrinologist, internist and self-referred) in 23 (17 per cent). Each patient underwent a complete history, physical examination, semen analysis with post-ejaculatory urinalysis and endocrine profile (testosterone, follicle-stimulating hormone, luteinizing hormone and prolactin). Of the patients 101 underwent testicular biopsy with or without vasography. Vasography was performed in every patient with spermatozoa present on an intraoperative testicular touch-preparation9 and a normalappearing vas deferens. The reason for not performing testicular biopsy in the remaining 32 patients included bilateral vasal agenesis, known genetic or chromosomal abnormalities, or severe bilateral testicular atrophy. Transrectal ultrasound and ultrasonographic evaluation of the upper urinary tract were performed in patients with bilateral vasal agenesis. Testicular biopsies were fixed in Bouin's solution, embedded in paraffin, stained with hematoxylin and eosin, and evaluated subsequently by a pathologist. Intraoperative testicular touchpreparations were stained with hematoxylin and eosin, and examined by a pathologist and the operative surgeon. Testicular biopsy was the standard against which all other diagnostic variables were measured, since the presence of normal spermatogenesis on permanent section and numerous spermatozoa present on a testicular touch-preparation accurately establishes the diagnosis of excurrent ductal obstruction as the cause of azoospermia. The diagnostic variables studied singly and in combination were a history of fertility, testicular size, serum follicle-stimulating hormone and ejaculate volume. These variables were chosen because they are easily obtained by any urologist. From these variables positive predictors of ductal obstruction were defined as a history of fertility, normal testicular size (greater than 3 x 2 cm., at least unilaterally),
MATERIALS AND METHODS
The charts of all men attending a male infertility clinic from January 1984 through December 1986 who had azoospermia Accepted for publication January 6, 1989. 62
63
EVALUATION OF AZOOSPERl.V!IC PATrENT
decreased ejaculate volume (less than 1.5 ml.) and a normal serum follicle-stimulating hormone (50 to 300 ng./dl.). With the 101 patients whose diagnosis was confirmed by a testicular biopsy, statistical analysis was performed to determine whether the aforementioned variables differed significantly between patients with ductal obstruction and those with testicular failure. Student's 2-tailed t tests and/or chi-square tests were used for these comparisons. In addition, the sensitivity and specificity of the noninvasive tests were calculated based on the premise that a normal testicular biopsy represents a true positive result for ductal obstruction. RESULTS
Risk factors. History and physical examination revealed a variety of risk factors in this patient population. A total of 16 patients (12 per cent) had a history of fertility, whereas 117 (88 per cent) had primary infertility. Thirteen patients (10 per cent) had a history of cryptorchidism (8 bilateral and 5 unilateral), 2 had a history of testicular torsion and 19 had a history of exposure to a gonadotoxin (radiation therapy and chemotherapy included). There was a history of mumps orchitis in 4 patients and of epididymitis in 16. One patient had a history of chronic sinusitis associated with bilateral epididymal obstruction, suggesting the diagnosis of Young's syndrome. 10 Physical examination revealed bilateral vasal agenesis in 16 patients. Testicular atrophy (measurement of 3 X 2 cm. or less) was found bilaterally in 26 patients and unilaterally in 10. Laboratory. None of our patients had sperm in the postejaculatory urinalysis. Examination of semen for fructose was positive in all but 10 patients. Of the 10 patients with a negative semen fructose test 7 had bilateral vasal agenesis and only 3 had ejaculatory duct obstruction. Ejaculate volume ranged from 0.1 to 3.5 ml. Endocrine evaluation revealed a normal serum luteinizing hormone except in 1 patient with androgen insensitivity and 1 with Kallmann's syndrome. Of the remaining 131 patients 68 had an isolated elevation of serum follicle-stimulating hormone, which was greater than 2 times the normal value in 49. None of the patients had an elevated serum prolactin level. Three patients with primary gonadal failure had Klinefelter's syndrome, documented by a 47XXY karyotype. Radiological findings. By transrectal ultrasonography 13 of the 16 patients with bilateral vasal agenesis (81 per cent) had an abnormality of the seminal vesicle (either agenesis or hypoplasia). Ultrasonographic evaluation of the upper urinary tract revealed an abnormality in 2 of the 15 patients studied (13 per cent). V asography revealed obstruction at the level of the inguinal canal bilaterally in 1 patient who had undergone bilateral inguinal hernia repair as a child, and at the level of the ejaculatory ducts in 3 other patients. Testicular biopsy. Open testicular biopsy was performed in 101 patients. Ductal obstruction was diagnosed in 43 patients, while 41 had normal histological findings and 2 had evidence of focal scarring in conjunction with normal spermatogenesis. Intrinsic testicular failure was diagnosed in 58 patients. A total of 38 patients had a Sertoli-cell-only pattern and 20 had maturation arrest. Statistical analysis. Only the 101 patients with a biopsy proved diagnosis were used for statistical analysis. Statistical evaluation revealed a significant difference in the distribution of testicular atrophy (p <0.001), mean ejaculate volume (p <0.001) and mean serum follicle-stimulating hormone (p <0.001) between the ductal obstruction and testicular failure groups (table 1). With the chi-square test there was no difference in the frequency of prior fertility between these 2 groups (p = 0.39). Congestion of the epididymal duct observed on scrotal exploration and normal spermatogenesis on testicular biopsy were considered pathognomonic for ductal obstruction in the azoospermic patient, thus, having 100 per cent sensitivity and 100
1. Comparison of noninvasive variables among the 101 patients who had a biopsy confirmed diagnosis of either ductal obstruction or testicular failure
TABLE
Variable Fertility status:* Primary Secondary Testicular atrophy:* None Unilat. Bilat. Ejaculate vol. (ml.)t Follicle-stimulating hormone (ng./ml.)t
Obstruction (43 pts.)
Testicular Failure (58 pts.)
37 (86) 6 (14)
53 (91) 5 (9)
40 (93) 3 (7) 0 (0) 1.5±0.17 156 ± 19
25 (43) 7 (12) 26 (45) 2.8 ± 0.18 818 ± 67
P Value
0.39
<0.001 <0.001 <0.001
* Number
of patients (per cent). t Mean ± standard error; normal 50 to 300 ng./dl.
Sensitivity and specificity (plus or minus standard error) of noninvasive variables in predicting obstruction as the cause of azoospermia in 101 patients who had undergone testicular biopsy
TABLE 2.
Sensitivity
Variable
(%)
Fertility status Ejaculate vol.* Testicular sizet Normal follicle-stimulating hormone 2 x normal follicle-stimulating hormone 3 X normal follicle-stimulating hormone 2 x normal follicle-stimulating hormone and testicular sizet
14 ± 7 74 ± 7 100:j: 88 ± 5 100:j: 100:j: 100:j:
Specificity (%)
91 67 45 81 62 40 71
± 4 ± 6 ± 7
±5 ±6 ±6 ± 6
·•· Ejaculate volume of less than 1.5 ml. indicates obstruction. t Bilateral atrophy indicates testicular failure. :j: Standard error cannot be estimated when the observed sensitivity or specificity is 100 per cent.
per cent specificity rates. The sensitivity and specificity of various noninvasive tests used singly and in combination are listed in table 2. Unfortunately, none of the noninvasive tests was as accurate as testicular biopsy in predicting ductal obstruction. However, with the combined criteria of a serum folliclestimulating hormone level elevated less than 2 times normal and the absence of bilateral testicular atrophy as an indication for testicular biopsy, all of the patients with ductal obstruction were identified (100 per cent sensitivity) and the majority of patients with testicular failure were excluded (71 per cent specificity). DISCUSSION
The various causes of azoospermia identified in this series are listed in table 3, and can be subdivided into treatable, potentially treatable and untreatable disorders. Segmental obstructions of the excurrent ducts (either congenital or acquired) can be treated surgically with varying degrees of success. Hypogonadotropic hypogonadism can be treated successfully with hormonal therapy in approximately 40 per cent of the patients.11 Although vasal agenesis is not treatable successfully to date it is a potentially treatable disorder as further advances are made with micropuncture techniques and surgical placement of alloplastic spermatoceles. 12 The treatment of maturation arrest is limited. If varicoceles are present ligation may result in improvement, 1 "· 1 4 although the over-all response rate in this setting is poor. Potentially, Bertoli cell growth factors" may be used in the future. As further advances in evaluation and treatment are made patients from the idiopathic group may have specific entities that may be treatable. Chromosomal disorders (Klinefelter's syndrome), genetic disorders (androgen insensitivity) and either acquired or congenital germinal cell aplasia are untreatable to date. Several interesting observations were made upon reviewing this patient series. Retrograde ejaculation and hyperprolactinemia are rare, if not nonexistent, causes of azoospermia.
64
JAROW, ESPELAND AND LIPSHULTZ TABLE
3. Final diagnosis of the 133 patients evaluated for azoospermia No.(%)
Obstruction (49 pts., 37% over-all): Vasa! agenesis Epididymal atresia Epididymitis Ejaculatory duct obstruction Young's syndrome Iatrogenic Total Testicular failure (84 pts., 63% over-all): Endocrine Chromosomal Gonadotoxin Mumps orchitis Idiopathic Total
16 (33) 12 (24)
16 (33) 3 1 1
(6) (2) (2)
49 (100) 2 (2) 3 (4) 19 (22) 4 (5) 56 (67)
84 (100)
Patient bias through established referral patterns may have caused the low incidence of an endocrinopathic condition in this series but it probably did not have an effect on the incidence of retrograde ejaculation. In our experience the majority of patients with retrograde ejaculation have sperm in the antegrade specimen or no antegrade specimen at all (aspermia). Additionally, prior fertility status was not helpful to identify patients with acquired obstruction as one might have expected. Fructose, produced and secreted by the seminal vesicles, can be assayed qualitatively within the seminal fluid. This generally is considered an indicator of the patency of the ejaculatory ducts or the presence of seminal vesicles. Although a negative fructose test is strongly indicative of an excurrent ductal abnormality, a positive test does not completely rule out obstruction as demonstrated by this study. Ejaculate volume was significantly lower in patients with obstruction (table 1). However, the mean was within normal limits (1.5 ml.) and significant overlap between the 2 groups adversely affected its clinical use in the differentiation of testicular failure from obstruction (table 2). In general, there is an indirect relationship between the level of serum f~llicle-stimulating hormone and the activity of spermatogenesis, as has been reported by Franchimont16 and de Kretser 17 and their associates. This relationship is believed to be established through the feedback inhibition of the pituitary gland by a factor from the Sertoli cells of the testis (inhibin). Ho~ever, this relationship is understood incompletely, since a patient may have a significant abnormality in spermatogenesis without havii:~ an elevation in serum follicle-stimulating hormone. In addition, the effect of unilateral testicular damage on serum follicle-stimulating hormone levels in the presence of active spermatogenesis contralaterally is not known. Previous studies on the role of follicle-stimulating hormone in the evaluation of the azoospermic man have produced conflicting recommendations. Bablok and associates recommended testicular biopsy in all patients with abnormal follicle-stimulating hormone values and reconstruction in those with a normal folliclestimulating hormone level. 7 Hargreave and J equier recommended biopsy in patients with a serum follicle-stimulating hor~~n~ of less than 3 times the normal value, with a reported sensitivity of 100 per cent and a specificity of 50 per cent in predicting ductal obstruction. 8 Pryor and associates used a cutoff value of 2 times normal follicle-stimulating hormone levels and included testicular size in their criteria. 18 To determine the accuracy of any diagnostic test one must first decide the definition of normal. This entails establishing a threshold level above or below which a value will be declared abnormal. The sensitivity or specificity of a test can be adjusted by raising or lowering that threshold level. For a test to be most useful the cutoff value must be at a point that achieves the d~sired r~sults. For instance, if it is important that all patients with a disease (such as ductal obstruction) be identified, no matter how many patients without the disease are included,
then a high sensitivity is desired. In our example the threshold level of serum follicle-stimulating hormone is moved from normal to either 2 or 3 times higher than normal. On the other hand, if it is important not to include any patients without the disease (in our example those with testicular failure) then a higher _specificity is desired. In our series the specificity could not be improved beyond 71 per cent without moving the threshold level to a below normal value of follicle-stimulating hormone (see figure). This resulted in a great sacrifice in sensitivity (table 2). However, the combined sensitivity and specificity achieved when including testicular size in the equation was much better than using a cutoff follicle-stimulating hormone value alone (table 2 or according to the study of Hargreave and Jeq?ier8)._ We believe that it is more important to include every ratient with ductal obstruction than to avoid a testicular biopsy m a patient with testicular failure. Therefore, a rational approach to the evaluation of the azoosper~ic patient (documented by at least 2 centrifuged semen specimens) would begin with physical examination to rule out either bilateral vasal agenesis or bilateral testicular atrophy. A serum ~o~~one panel (testosterone, and follicle-stimulating and luteimzmg hormones) then should be obtained to rule out a potentially treatable endocrine disorder. Patients with an ele~ation in serum follicle-stimulating hormone of greater than 2 time~ normal have testicular failure. A karyotype, serum prolactm level and post-ejaculatory urinalysis can be performed in select patients. Patients with a serum follicle-stimulating hormone level of less than 2 times normal and at least 1 normal sized testicle and vas deferens potentiaily have a correctable duct~l obs~ructio1: and shoul~ undergo testicular biopsy. With this algorithm 39 patients in our series could have been spared a testicular biopsy on the basis of the physical examination (bilateral vasal agenesis in 13 and bilateral testicular atrophy in 26) and 15 more could have been excluded on the basis of a serum follicle-stimulating hormone of greater than 2 times normal (shaded areas in figure). This would leave 47 patients undergoing testicular biopsy, 17 (36 per cent) of whom would have testicular failure. Of the 30 patients actually having ductal obstruction 4 had a mildly elevated serum follicle-stimulating hormone level and, therefore, a guarded prognosis. It is not possible to identify and exclude all patients with testicular failure.with t~ese no1:invasive variables without simultaneously excludmg patients with ductal obstruction (see figure). I1: c?nclusion, to ensure the greatest diagnostic accuracy and to hm~t an ~.mnecessary operation we believe that only azoospermic patients who have a serum follicle-stimulating hormone level of less than 2 times normal, and at least 1 normal sized testicle and vas deferens on physical examination should DIAGNOSIS
Testis Size"' Normal
FSH
No Vasal Agenesis
NL
TESTICULAR FAILURE
TOTAL
26
10
47
NL-2X
2
6
9
Unilateral
NL
0
0
Atrophy
NL-2X
2
Total
13
30
3
58
101
Dis~ributio~ of 101 azoospermic patients who had undergone testicular b1
EVALUATION OF AZOOSPERMIC PATIENT
undergo scrotal exploration and testicular biopsy. Patients who do not satisfy these criteria have either vasal agenesis or testicular failure. In these cases testicular biopsy should be performed only to obtain a pathological diagnosis. REFERENCES 1. Willott, G. M.: Frequency of azoospermia. Forensic Sci. Int., 20: 9, 1982. 2. Stanwell-Smith, R. E. and Hendry, W. F.: The prognosis of male subfertility: a survey of 1025 men referred to a fertility clinic. Brit. J. Urol., 56: 422, 1984. 3. Jequier, A. M. and Holmes, S. C.: Aetiological factors in the production of obstructive azoospermia. Brit. J. Urol., 56: 540, 1984. 4. Tomamichel, G. R. and Bandhauer, K.: Seminal carnitine content in obstructive azoospermia. Correlation with the anatomic level of obstruction. J. Androl., 7: 328, 1986. 5. Guerin, J. F., Rollet, J., Perrin, P., Menezo, Y., Orgiazzi, A. and Czyba, J.-C.: Enzymes in the seminal plasma from azoospermic men: correlation with the origin of their azoospermia. Fertil. Steril., 36: 368, 1981. 6. Sade, M., Saverymuttu, I., Dinsdale, 0. and Gow, J. G.: The management of azoospermia. Brit. J. Urol., 50: 595, 1978. 7. Bablok, L., Janczewski, Z., Kwiatkowska, Z. and Fracki, S.: The relationship between plasma FSH, testosterone levels and testicular histology in males with azoospermia. Andrologia, 10: 502, 1978. 8. Hargreave, T. B. and Jequier, A. M.: Can follicle stimulating hormone estimation replace testicular biopsy in the diagnosis of
65
obstructive azoospermia? Brit. J. Urol., 50: 415, 1978. 9. Coburn, M., Wheeler, T. and Lipshultz, L. I.: Testicular biopsy: its use and limitations. Urol. Clin. N. Amer., 14: 551, 1987. 10. Hughes, T. M., III, Skolnick, J. L. and Belker, A. M.: Young's syndrome: an often unrecognized correctable cause of obstructive azoospermia. J. Urol., 137: 1238, 1987. 11. Burger, H. G. and Baker, H. W. G.: Therapeutic considerations and results of gonadotropin treatment in male hypogonadotropic hypogonadism. Ann. N. Y. Acad. Sci., 438: 447, 1984. 12. Wagenknecht, L. V., Weitz, K. H., Hoppe, L. P., Krause, D., Becker, H. and Scherrin, C.: Microsurgery in andrologic urology. IL Alloplastic spermatocele. J. Microsurg., 1: 428, 1980. 13. Czaplicki, M., Bablok, L. and Janczewski, Z.: Varicocelectomy in patients with azoospermia. Arch. Androl., 3: 51, 1979. 14. Tulloch, W. S.: Varicocele in subfertility: results of treatment. Brit. Med. J., 2: 356, 1955. 15. Bush, J.P., Lamb, D. J., Lipshultz, L. I. and Smith, R. G.: Partial characterization of a unique growth factor secreted by human Sertoli cells. Fertil. Steril., 49: 658, 1988. 16. Franchimont, P., Millet, D., Vendrely, E., Letawe, J., Legros, J. J. and Netter, A.: Relationship between spermatogenesis and serum gonadotropin levels in azoospermia and oligospermia. J. Clin. Endocr. Metab., 34: 1003, 1972. 17. de Kretser, D. M., Burger, H. G. and Hudson, B.: The relationship between germinal cells and serum FSH levels in males with infertility. J. Clin. Endocr. Metab., 38: 787, 1974. 18. Pryor, J.P., Cameron, K. M., Collins, W. P., Hirsh, A. V., Mahony, J. D. H., Pugh, R. C. B. and Fitzpatrick, J. M.: Indications for testicular biopsy or exploration in azoospermia. Brit. J. Urol., 50: 591, 1978.
THE JOURNAL OF UROLOGY
Copyright© 1989 by Williams & Wilkins
EVALUATION OF THE AZOOSPERMIC PATIENT JONATHAN P. JAROW, MARK A. ESPELAND AND LARRY I. LIPSHULTZ From the Department of Surgery, Section on Urology, Bowman Gray School of Medicine, Winston-Salem, North Carolina, and Baylor College of Medicine, Houston, Texas
ABSTRACT
Azoospermia is found in up to 10 to 20 per cent of the men who present to an infertility clinic. The main causes are testicular failure and ductal obstruction. Testicular biopsy remains the definitive test used to differentiate these 2 disorders. A retrospective study of 133 azoospermic men was performed to determine the accuracy and limitations of noninvasive variables in predicting testicular failure in an effort to limit the need for diagnostic testicular biopsy. Of 49 patients (37 per cent) with ductal obstruction a third had bilateral vasal agenesis. The remaining 84 azoospermic patients (63 per cent) had testicular failure. The results of the complete evaluation of these patients are described. Among the 101 patients with a testicular biopsy confirmed diagnosis there was a significant difference in testicular size (p less than 0.001), ejaculate volume (p less than 0.001) and serum follicle-stimulating hormone (p less than 0.001) between patients with testicular failure and those with ductal obstruction. The sensitivity and specificity of various parameters were determined. The best criteria to predict ductal obstruction preoperatively are a serum follicle-stimulating hormone level of less than 2 times greater than normal and the absence of bilateral testicular atrophy (100 per cent sensitivity and 71 per cent specificity). An algorithm for evaluation of the azoospermic patient is described such that all men with ductal obstruction and a minimal number with testicular failure undergo testicular biopsy. (J. Urol., 142: 62-65, 1989) The prevalence of azoospermia in the general population has been estimated to be 2 per cent. 1 The incidence at a male infertility clinic has been reported to be as high as 10 to 20 per cent. 2• 3 The over-all prognosis of fertility in men with azoospermia is poor. The majority of these patients have untreatable disorders and, unfortunately, the success rates for the treatable disorders are low. For this reason many patients and physicians have been discouraged from pursuing the evaluation and treatment of this problem. The main differential diagnosis in the azoospermic patient in terms of treatment and prognosis is between testicular failure and obstruction of the excurrent ducts of the testis. Several diagnostic tests to determine the presence and level of obstruction have been described, including measurement of ejaculate volume, biochemical assays of the ejaculate (fructose, carnitine,4 a-glucosidase 5 and glycerylphosphorylcholine6 ) and serum hormone determinations. 7 • 8 All have varying degrees of accuracy that have not been substantiated fully. Despite the availability of these diagnostic tests, testicular biopsy still is necessary to establish the presence of normal spermatogenesis in patients with suspected ductal obstruction. This purpose remains 1 of the few indications for testicular biopsy in the subfertile patient. In addition, scrotal exploration with epididymal inspection and/or vasography is mandatory before any reconstruction is performed. We developed a rational approach to the evaluation of the azoospermic patient through an evaluation of the causes of this disorder, and the measurement of the sensitivity and specificity of various noninvasive diagnostic tests. Ideally, testicular biopsy should be avoided in as many patients as possible without the exclusion of any potentially treatable patient with ductal obstruction.
upon examination of a centrifuged semen sample were reviewed. Patients with either aspermia or a history of elective vasectomy were excluded. Mean age of these 133 patients was 31 years (range 22 to 56 years). The patient population was skewed in that these patients were seen at a tertiary fertility center. Referral sources were a urologist in 65 patients (49 per cent), gynecologist in 45 (34 per cent) and other sources (endocrinologist, internist and self-referred) in 23 (17 per cent). Each patient underwent a complete history, physical examination, semen analysis with post-ejaculatory urinalysis and endocrine profile (testosterone, follicle-stimulating hormone, luteinizing hormone and prolactin). Of the patients 101 underwent testicular biopsy with or without vasography. Vasography was performed in every patient with spermatozoa present on an intraoperative testicular touch-preparation9 and a normalappearing vas deferens. The reason for not performing testicular biopsy in the remaining 32 patients included bilateral vasal agenesis, known genetic or chromosomal abnormalities, or severe bilateral testicular atrophy. Transrectal ultrasound and ultrasonographic evaluation of the upper urinary tract were performed in patients with bilateral vasal agenesis. Testicular biopsies were fixed in Bouin's solution, embedded in paraffin, stained with hematoxylin and eosin, and evaluated subsequently by a pathologist. Intraoperative testicular touchpreparations were stained with hematoxylin and eosin, and examined by a pathologist and the operative surgeon. Testicular biopsy was the standard against which all other diagnostic variables were measured, since the presence of normal spermatogenesis on permanent section and numerous spermatozoa present on a testicular touch-preparation accurately establishes the diagnosis of excurrent ductal obstruction as the cause of azoospermia. The diagnostic variables studied singly and in combination were a history of fertility, testicular size, serum follicle-stimulating hormone and ejaculate volume. These variables were chosen because they are easily obtained by any urologist. From these variables positive predictors of ductal obstruction were defined as a history of fertility, normal testicular size (greater than 3 x 2 cm., at least unilaterally),
MATERIALS AND METHODS
The charts of all men attending a male infertility clinic from January 1984 through December 1986 who had azoospermia Accepted for publication January 6, 1989. 62
63
EVALUATION OF AZOOSPERl.V!IC PATrENT
decreased ejaculate volume (less than 1.5 ml.) and a normal serum follicle-stimulating hormone (50 to 300 ng./dl.). With the 101 patients whose diagnosis was confirmed by a testicular biopsy, statistical analysis was performed to determine whether the aforementioned variables differed significantly between patients with ductal obstruction and those with testicular failure. Student's 2-tailed t tests and/or chi-square tests were used for these comparisons. In addition, the sensitivity and specificity of the noninvasive tests were calculated based on the premise that a normal testicular biopsy represents a true positive result for ductal obstruction. RESULTS
Risk factors. History and physical examination revealed a variety of risk factors in this patient population. A total of 16 patients (12 per cent) had a history of fertility, whereas 117 (88 per cent) had primary infertility. Thirteen patients (10 per cent) had a history of cryptorchidism (8 bilateral and 5 unilateral), 2 had a history of testicular torsion and 19 had a history of exposure to a gonadotoxin (radiation therapy and chemotherapy included). There was a history of mumps orchitis in 4 patients and of epididymitis in 16. One patient had a history of chronic sinusitis associated with bilateral epididymal obstruction, suggesting the diagnosis of Young's syndrome. 10 Physical examination revealed bilateral vasal agenesis in 16 patients. Testicular atrophy (measurement of 3 X 2 cm. or less) was found bilaterally in 26 patients and unilaterally in 10. Laboratory. None of our patients had sperm in the postejaculatory urinalysis. Examination of semen for fructose was positive in all but 10 patients. Of the 10 patients with a negative semen fructose test 7 had bilateral vasal agenesis and only 3 had ejaculatory duct obstruction. Ejaculate volume ranged from 0.1 to 3.5 ml. Endocrine evaluation revealed a normal serum luteinizing hormone except in 1 patient with androgen insensitivity and 1 with Kallmann's syndrome. Of the remaining 131 patients 68 had an isolated elevation of serum follicle-stimulating hormone, which was greater than 2 times the normal value in 49. None of the patients had an elevated serum prolactin level. Three patients with primary gonadal failure had Klinefelter's syndrome, documented by a 47XXY karyotype. Radiological findings. By transrectal ultrasonography 13 of the 16 patients with bilateral vasal agenesis (81 per cent) had an abnormality of the seminal vesicle (either agenesis or hypoplasia). Ultrasonographic evaluation of the upper urinary tract revealed an abnormality in 2 of the 15 patients studied (13 per cent). V asography revealed obstruction at the level of the inguinal canal bilaterally in 1 patient who had undergone bilateral inguinal hernia repair as a child, and at the level of the ejaculatory ducts in 3 other patients. Testicular biopsy. Open testicular biopsy was performed in 101 patients. Ductal obstruction was diagnosed in 43 patients, while 41 had normal histological findings and 2 had evidence of focal scarring in conjunction with normal spermatogenesis. Intrinsic testicular failure was diagnosed in 58 patients. A total of 38 patients had a Sertoli-cell-only pattern and 20 had maturation arrest. Statistical analysis. Only the 101 patients with a biopsy proved diagnosis were used for statistical analysis. Statistical evaluation revealed a significant difference in the distribution of testicular atrophy (p <0.001), mean ejaculate volume (p <0.001) and mean serum follicle-stimulating hormone (p <0.001) between the ductal obstruction and testicular failure groups (table 1). With the chi-square test there was no difference in the frequency of prior fertility between these 2 groups (p = 0.39). Congestion of the epididymal duct observed on scrotal exploration and normal spermatogenesis on testicular biopsy were considered pathognomonic for ductal obstruction in the azoospermic patient, thus, having 100 per cent sensitivity and 100
1. Comparison of noninvasive variables among the 101 patients who had a biopsy confirmed diagnosis of either ductal obstruction or testicular failure
TABLE
Variable Fertility status:* Primary Secondary Testicular atrophy:* None Unilat. Bilat. Ejaculate vol. (ml.)t Follicle-stimulating hormone (ng./ml.)t
Obstruction (43 pts.)
Testicular Failure (58 pts.)
37 (86) 6 (14)
53 (91) 5 (9)
40 (93) 3 (7) 0 (0) 1.5±0.17 156 ± 19
25 (43) 7 (12) 26 (45) 2.8 ± 0.18 818 ± 67
P Value
0.39
<0.001 <0.001 <0.001
* Number
of patients (per cent). t Mean ± standard error; normal 50 to 300 ng./dl.
Sensitivity and specificity (plus or minus standard error) of noninvasive variables in predicting obstruction as the cause of azoospermia in 101 patients who had undergone testicular biopsy
TABLE 2.
Sensitivity
Variable
(%)
Fertility status Ejaculate vol.* Testicular sizet Normal follicle-stimulating hormone 2 x normal follicle-stimulating hormone 3 X normal follicle-stimulating hormone 2 x normal follicle-stimulating hormone and testicular sizet
14 ± 7 74 ± 7 100:j: 88 ± 5 100:j: 100:j: 100:j:
Specificity (%)
91 67 45 81 62 40 71
± 4 ± 6 ± 7
±5 ±6 ±6 ± 6
·•· Ejaculate volume of less than 1.5 ml. indicates obstruction. t Bilateral atrophy indicates testicular failure. :j: Standard error cannot be estimated when the observed sensitivity or specificity is 100 per cent.
per cent specificity rates. The sensitivity and specificity of various noninvasive tests used singly and in combination are listed in table 2. Unfortunately, none of the noninvasive tests was as accurate as testicular biopsy in predicting ductal obstruction. However, with the combined criteria of a serum folliclestimulating hormone level elevated less than 2 times normal and the absence of bilateral testicular atrophy as an indication for testicular biopsy, all of the patients with ductal obstruction were identified (100 per cent sensitivity) and the majority of patients with testicular failure were excluded (71 per cent specificity). DISCUSSION
The various causes of azoospermia identified in this series are listed in table 3, and can be subdivided into treatable, potentially treatable and untreatable disorders. Segmental obstructions of the excurrent ducts (either congenital or acquired) can be treated surgically with varying degrees of success. Hypogonadotropic hypogonadism can be treated successfully with hormonal therapy in approximately 40 per cent of the patients.11 Although vasal agenesis is not treatable successfully to date it is a potentially treatable disorder as further advances are made with micropuncture techniques and surgical placement of alloplastic spermatoceles. 12 The treatment of maturation arrest is limited. If varicoceles are present ligation may result in improvement, 1 "· 1 4 although the over-all response rate in this setting is poor. Potentially, Bertoli cell growth factors" may be used in the future. As further advances in evaluation and treatment are made patients from the idiopathic group may have specific entities that may be treatable. Chromosomal disorders (Klinefelter's syndrome), genetic disorders (androgen insensitivity) and either acquired or congenital germinal cell aplasia are untreatable to date. Several interesting observations were made upon reviewing this patient series. Retrograde ejaculation and hyperprolactinemia are rare, if not nonexistent, causes of azoospermia.
64
JAROW, ESPELAND AND LIPSHULTZ TABLE
3. Final diagnosis of the 133 patients evaluated for azoospermia No.(%)
Obstruction (49 pts., 37% over-all): Vasa! agenesis Epididymal atresia Epididymitis Ejaculatory duct obstruction Young's syndrome Iatrogenic Total Testicular failure (84 pts., 63% over-all): Endocrine Chromosomal Gonadotoxin Mumps orchitis Idiopathic Total
16 (33) 12 (24)
16 (33) 3 1 1
(6) (2) (2)
49 (100) 2 (2) 3 (4) 19 (22) 4 (5) 56 (67)
84 (100)
Patient bias through established referral patterns may have caused the low incidence of an endocrinopathic condition in this series but it probably did not have an effect on the incidence of retrograde ejaculation. In our experience the majority of patients with retrograde ejaculation have sperm in the antegrade specimen or no antegrade specimen at all (aspermia). Additionally, prior fertility status was not helpful to identify patients with acquired obstruction as one might have expected. Fructose, produced and secreted by the seminal vesicles, can be assayed qualitatively within the seminal fluid. This generally is considered an indicator of the patency of the ejaculatory ducts or the presence of seminal vesicles. Although a negative fructose test is strongly indicative of an excurrent ductal abnormality, a positive test does not completely rule out obstruction as demonstrated by this study. Ejaculate volume was significantly lower in patients with obstruction (table 1). However, the mean was within normal limits (1.5 ml.) and significant overlap between the 2 groups adversely affected its clinical use in the differentiation of testicular failure from obstruction (table 2). In general, there is an indirect relationship between the level of serum f~llicle-stimulating hormone and the activity of spermatogenesis, as has been reported by Franchimont16 and de Kretser 17 and their associates. This relationship is believed to be established through the feedback inhibition of the pituitary gland by a factor from the Sertoli cells of the testis (inhibin). Ho~ever, this relationship is understood incompletely, since a patient may have a significant abnormality in spermatogenesis without havii:~ an elevation in serum follicle-stimulating hormone. In addition, the effect of unilateral testicular damage on serum follicle-stimulating hormone levels in the presence of active spermatogenesis contralaterally is not known. Previous studies on the role of follicle-stimulating hormone in the evaluation of the azoospermic man have produced conflicting recommendations. Bablok and associates recommended testicular biopsy in all patients with abnormal follicle-stimulating hormone values and reconstruction in those with a normal folliclestimulating hormone level. 7 Hargreave and J equier recommended biopsy in patients with a serum follicle-stimulating hor~~n~ of less than 3 times the normal value, with a reported sensitivity of 100 per cent and a specificity of 50 per cent in predicting ductal obstruction. 8 Pryor and associates used a cutoff value of 2 times normal follicle-stimulating hormone levels and included testicular size in their criteria. 18 To determine the accuracy of any diagnostic test one must first decide the definition of normal. This entails establishing a threshold level above or below which a value will be declared abnormal. The sensitivity or specificity of a test can be adjusted by raising or lowering that threshold level. For a test to be most useful the cutoff value must be at a point that achieves the d~sired r~sults. For instance, if it is important that all patients with a disease (such as ductal obstruction) be identified, no matter how many patients without the disease are included,
then a high sensitivity is desired. In our example the threshold level of serum follicle-stimulating hormone is moved from normal to either 2 or 3 times higher than normal. On the other hand, if it is important not to include any patients without the disease (in our example those with testicular failure) then a higher _specificity is desired. In our series the specificity could not be improved beyond 71 per cent without moving the threshold level to a below normal value of follicle-stimulating hormone (see figure). This resulted in a great sacrifice in sensitivity (table 2). However, the combined sensitivity and specificity achieved when including testicular size in the equation was much better than using a cutoff follicle-stimulating hormone value alone (table 2 or according to the study of Hargreave and Jeq?ier8)._ We believe that it is more important to include every ratient with ductal obstruction than to avoid a testicular biopsy m a patient with testicular failure. Therefore, a rational approach to the evaluation of the azoosper~ic patient (documented by at least 2 centrifuged semen specimens) would begin with physical examination to rule out either bilateral vasal agenesis or bilateral testicular atrophy. A serum ~o~~one panel (testosterone, and follicle-stimulating and luteimzmg hormones) then should be obtained to rule out a potentially treatable endocrine disorder. Patients with an ele~ation in serum follicle-stimulating hormone of greater than 2 time~ normal have testicular failure. A karyotype, serum prolactm level and post-ejaculatory urinalysis can be performed in select patients. Patients with a serum follicle-stimulating hormone level of less than 2 times normal and at least 1 normal sized testicle and vas deferens potentiaily have a correctable duct~l obs~ructio1: and shoul~ undergo testicular biopsy. With this algorithm 39 patients in our series could have been spared a testicular biopsy on the basis of the physical examination (bilateral vasal agenesis in 13 and bilateral testicular atrophy in 26) and 15 more could have been excluded on the basis of a serum follicle-stimulating hormone of greater than 2 times normal (shaded areas in figure). This would leave 47 patients undergoing testicular biopsy, 17 (36 per cent) of whom would have testicular failure. Of the 30 patients actually having ductal obstruction 4 had a mildly elevated serum follicle-stimulating hormone level and, therefore, a guarded prognosis. It is not possible to identify and exclude all patients with testicular failure.with t~ese no1:invasive variables without simultaneously excludmg patients with ductal obstruction (see figure). I1: c?nclusion, to ensure the greatest diagnostic accuracy and to hm~t an ~.mnecessary operation we believe that only azoospermic patients who have a serum follicle-stimulating hormone level of less than 2 times normal, and at least 1 normal sized testicle and vas deferens on physical examination should DIAGNOSIS
Testis Size"' Normal
FSH
No Vasal Agenesis
NL
TESTICULAR FAILURE
TOTAL
26
10
47
NL-2X
2
6
9
Unilateral
NL
0
0
Atrophy
NL-2X
2
Total
13
30
3
58
101
Dis~ributio~ of 101 azoospermic patients who had undergone testicular b1
EVALUATION OF AZOOSPERMIC PATIENT
undergo scrotal exploration and testicular biopsy. Patients who do not satisfy these criteria have either vasal agenesis or testicular failure. In these cases testicular biopsy should be performed only to obtain a pathological diagnosis. REFERENCES 1. Willott, G. M.: Frequency of azoospermia. Forensic Sci. Int., 20: 9, 1982. 2. Stanwell-Smith, R. E. and Hendry, W. F.: The prognosis of male subfertility: a survey of 1025 men referred to a fertility clinic. Brit. J. Urol., 56: 422, 1984. 3. Jequier, A. M. and Holmes, S. C.: Aetiological factors in the production of obstructive azoospermia. Brit. J. Urol., 56: 540, 1984. 4. Tomamichel, G. R. and Bandhauer, K.: Seminal carnitine content in obstructive azoospermia. Correlation with the anatomic level of obstruction. J. Androl., 7: 328, 1986. 5. Guerin, J. F., Rollet, J., Perrin, P., Menezo, Y., Orgiazzi, A. and Czyba, J.-C.: Enzymes in the seminal plasma from azoospermic men: correlation with the origin of their azoospermia. Fertil. Steril., 36: 368, 1981. 6. Sade, M., Saverymuttu, I., Dinsdale, 0. and Gow, J. G.: The management of azoospermia. Brit. J. Urol., 50: 595, 1978. 7. Bablok, L., Janczewski, Z., Kwiatkowska, Z. and Fracki, S.: The relationship between plasma FSH, testosterone levels and testicular histology in males with azoospermia. Andrologia, 10: 502, 1978. 8. Hargreave, T. B. and Jequier, A. M.: Can follicle stimulating hormone estimation replace testicular biopsy in the diagnosis of
65
obstructive azoospermia? Brit. J. Urol., 50: 415, 1978. 9. Coburn, M., Wheeler, T. and Lipshultz, L. I.: Testicular biopsy: its use and limitations. Urol. Clin. N. Amer., 14: 551, 1987. 10. Hughes, T. M., III, Skolnick, J. L. and Belker, A. M.: Young's syndrome: an often unrecognized correctable cause of obstructive azoospermia. J. Urol., 137: 1238, 1987. 11. Burger, H. G. and Baker, H. W. G.: Therapeutic considerations and results of gonadotropin treatment in male hypogonadotropic hypogonadism. Ann. N. Y. Acad. Sci., 438: 447, 1984. 12. Wagenknecht, L. V., Weitz, K. H., Hoppe, L. P., Krause, D., Becker, H. and Scherrin, C.: Microsurgery in andrologic urology. IL Alloplastic spermatocele. J. Microsurg., 1: 428, 1980. 13. Czaplicki, M., Bablok, L. and Janczewski, Z.: Varicocelectomy in patients with azoospermia. Arch. Androl., 3: 51, 1979. 14. Tulloch, W. S.: Varicocele in subfertility: results of treatment. Brit. Med. J., 2: 356, 1955. 15. Bush, J.P., Lamb, D. J., Lipshultz, L. I. and Smith, R. G.: Partial characterization of a unique growth factor secreted by human Sertoli cells. Fertil. Steril., 49: 658, 1988. 16. Franchimont, P., Millet, D., Vendrely, E., Letawe, J., Legros, J. J. and Netter, A.: Relationship between spermatogenesis and serum gonadotropin levels in azoospermia and oligospermia. J. Clin. Endocr. Metab., 34: 1003, 1972. 17. de Kretser, D. M., Burger, H. G. and Hudson, B.: The relationship between germinal cells and serum FSH levels in males with infertility. J. Clin. Endocr. Metab., 38: 787, 1974. 18. Pryor, J.P., Cameron, K. M., Collins, W. P., Hirsh, A. V., Mahony, J. D. H., Pugh, R. C. B. and Fitzpatrick, J. M.: Indications for testicular biopsy or exploration in azoospermia. Brit. J. Urol., 50: 591, 1978.