|Home | About | Journals | Submit | Contact Us | Français|
Many male survivors of childhood cancer are at risk for azoospermia. Although both the levels of follicle-stimulating hormone (FSH) and inhibin B are correlated with sperm concentration, their ability to predict azoospermia in survivors of childhood cancer remains uncertain.
Semen analysis was performed and serum levels of FSH and inhibin B were measured in 275 adult male survivors of childhood cancer who had received gonadotoxic therapy. Receiver operating characteristic (ROC) analysis was performed to determine the optimal inhibin B and FSH values for identifying patients with azoospermia. The patient sample was divided into a learning set and a validation set. Sensitivity, specificity, and positive and negative predictive value were calculated.
Inhibin B was dichotomized as ≤ 31 ng/L or more than 31 ng/L and FSH was dichotomized as ≤ 11.5 mIU/mL or more than 11.5 mIU/mL based on results of the ROC analysis. Using these values, the specificity of the serum level of inhibin B for identifying azoospermic survivors was 45.0%, and the positive predictive value was 52.1%. The specificity for FSH was 74.1%, and the positive predictive value was 65.1%.
Neither serum inhibin B nor FSH is a suitable surrogate for determination of sperm concentration in a semen sample. Young men and their physicians should be aware of the limitations of these measures for assessment of fertility potential.
The treatment of children and adolescents with cancer has become increasingly successful. Approximately 80% of all patients diagnosed before 15 years of age will survive for five years. The majority are expected to survive for many years after diagnosis.1 In boys and young men, irradiation of the gonads and/or the use of certain classes of chemotherapeutic agents (eg, alkylating agents) may damage spermatogenesis leading to impaired fertility.2
Adult survivors of childhood cancer place high value on fertility and desire information regarding their potential to sire/conceive a pregnancy.3,4 Semen analysis is a noninvasive method for estimation of male fertility potential, but religious, cultural, and personal barriers may deter individuals from completing the procedure.5
Inhibin B is a dimeric protein that consists of alpha and beta subunits that are synthesized in the Sertoli cells of the testis.6 The level of inhibin B is inversely related to the level of follicle-stimulating hormone (FSH) in normal males and those with a variety of reproductive abnormalities.7 Jensen et al8 reported that the serum level of inhibin B was directly related, whereas the serum level of FSH was inversely related, to sperm concentration in healthy Danish males. Others have reported that the serum inhibin B level is highly correlated with sperm concentration among men undergoing evaluation for male factor infertility.9–11 Studies of the relationship between serum levels of inhibin B and FSH and sperm concentration in adult survivors of childhood cancer have been limited in the number of evaluated patients who received gonadotoxic treatment and/or the number of patients from whom semen samples were obtained.12–14
We undertook these analyses to determine the sensitivity, specificity, and positive and negative predictive value of the serum levels of inhibin B and FSH for the identification of azoospermia. This assessment is possible because of the availability of a large population of long-term survivors of childhood cancer who had been treated with alkylating agents, direct gonadal irradiation (any dose), and/or hypothalamic/pituitary irradiation (≥ 40 Gy), all of which are associated with a significant risk of azoospermia.15
A cohort of patients (St. Jude Lifetime Cohort Study [SJLIFE]) was identified that fulfilled the following criteria: diagnosis of childhood malignancy treated at St. Jude Children's Research Hospital (SJCRH), survival ≥ 10 years from diagnosis, and current age ≥ 18 years. The detailed methods used for ascertainment, recruitment, and evaluation of the members of this cohort have been reported previously.16 This investigation was approved by the institutional review board at SJCRH, and all participants and/or their legal guardians provided informed consent.
The cumulative doses for 32 specific chemotherapeutic agents (5-azacytidine, bleomycin, busulfan, carboplatin, carmustine, cisplatinum, cyclophosphamide [intravenously [IV] or orally], cytarabine [IV, intramuscularly, intrathecally, subcutaneously], dacarbazine, dactinomycin, daunorubicin, dexamethasone, doxorubicin, etoposide [IV, orally], fludarabine, fluorouracil, hydroxyurea, idarubicin, ifosfamide, l-asparaginase, lomustine, melphalan, methotrexate [IV, intramuscularly, intrathecally], nitrogen mustard, prednisone, procarbazine, teniposide, thioguanine, thiotepa, tretinoin, vinblastine, vincristine), surgical procedures, and radiation treatment fields, dose, and energy source were abstracted from the medical records according to a protocol similar to that used in the Childhood Cancer Survivor Study (CCSS).17
Participants underwent a risk-based assessment as suggested by the Children's Oncology Group Long-Term Follow-Up Guidelines for Survivors of Childhood, Adolescent and Young Adult Cancer (COG Guidelines).18 Semen analysis was offered to men who had received potentially gonadotoxic treatment, including exposure to an alkylating agent, direct testicular irradiation (any dose), or hypothalamic/pituitary irradiation (≥ 40 Gy). FSH was to be measured in all men who received gonadotoxic therapy, and inhibin B was to be determined in all men who underwent semen analysis. Those who had received hypothalamic/pituitary irradiation ≥ 40 Gy or had a tumor in the hypothalamic/pituitary region (eg, craniopharyngioma) and those who were receiving exogenous androgen treatment were excluded from these analyses. Eleven participants had undergone unilateral orchiectomy. No participant underwent bilateral orchiectomy.
To analyze the serum hormone levels, we obtained morning peripheral blood samples and separated the serum. The serum samples were stored in liquid nitrogen at −180° Celsius until they were shipped on dry ice to Quest Diagnostics (Valencia, CA) for inhibin B analysis by immunoassay. The lower limit of the assay is 30 ng/L.
Serum FSH was assayed by using a two-step sandwich-type electrochemiluminescence immunoassay (Roche Diagnostics, Indianapolis, IN) in the SJCRH clinical laboratory. Within-assay coefficient of variation (CV) was 1.2% (n = 20; mean, 4.72 mIU/mL). Day-to-day imprecision for three control materials (Liquichek Immunoassay Plus Control, Trilevel; Bio-Rad Laboratories, Hercules, CA) was as follows: Level 1, n = 155; mean, 7.33 mIU/mL; CV, 3.9%; Level 2, n = 155; mean, 19.72 mIU/mL; CV, 3.7%; Level 3, n = 150; mean, 46.65 mIU/mL; CV, 3.7%. The normal range for males age 20 to 50 years in the SJCRH laboratory is 2.0 to 9.2 mIU/mL.
The semen samples were collected after a minimum of 2 days and a maximum of 7 days of sexual abstinence and were processed within 30 minutes of collection following the WHO Guidelines, 5th Edition.19 Every semen sample was allowed to liquefy, and the time to liquefaction was recorded. The raw sample was microscopically evaluated. If no sperm were detected, the sample was centrifuged and concentrated. The concentrated sample was again evaluated before it was considered azoospermic. A recent medical history was taken at the time of collection and, if any historical issues were revealed (recent fever above 102°F, certain medication use, recent genitourinary tract infection, or injury), a request for a repeat specimen in 1 month to confirm azoospermia was made.
Receiver operating characteristic (ROC) analysis was undertaken to identify the optimal dichotomization values for FSH, inhibin B, and the inhibin B:FSH ratio for identifying patients with azoospermia. Specifically, the patient sample was divided into a learning set (n = 140) and a validation set (n = 135) by random assignment. The ROC analysis was performed by using PROC LOGISTIC in SAS 9.2 (SAS Institute, Cary, NC).
Two hundred ninety-eight (53%) of 565 SJLIFE males who received gonadotoxic treatment, were not receiving exogenous androgens, and participated in SJLIFE before February 23, 2011, submitted a semen specimen for evaluation (Fig 1). Of these 298, we excluded those who had received ≥ 40 Gy hypothalamic/pituitary irradiation and/or had a tumor in the hypothalamic/pituitary region (n = 23), for a total of 275 eligible cases. The majority were non-Hispanic white and had at least a high school education. The most frequent diagnosis was acute lymphoblastic leukemia (Table 1). The median age at semen analysis was 30.5 years (range, 19.7 to 59.1 years). Inhibin B was measured in 238 and FSH in 275 participants. One hundred five (38.2%) of 275 were found to be azoospermic. Most of those who submitted semen specimens reported that they had no children (Appendix Table A1, online only). Among those who did not submit semen specimens, live births were reported by one participant who did not complete the SJLIFE clinical evaluation, two who declined semen analysis, one who was unable to produce a semen specimen, one who had erectile dysfunction, and two who had undergone vasectomy.
ROC analysis was performed to identify the optimum values of inhibin B, FSH, and the inhibin B:FSH ratio for identification of patients with azoospermia. On the basis of the learning set (Table 2), we selected the point on the ROC curve that was closest to the point (0,1) as the optimal threshold point. We also evaluated two other criteria: the minimum absolute difference between sensitivity and specificity, and the Youden index.20 The cutoffs selected by these criteria were similar. Furthermore, the validation set provided sensitivity, specificity, and positive and negative predictive values similar to those of the learning set (Table 2). As a result, we combined the two sets and calculated sensitivity, specificity, and positive and negative predictive values by using standard methods. The results for the sensitivity, specificity, and positive and negative predictive values for the learning data set, the validation data set, and the combined data set are provided in Table 2. An inhibin B value of ≤ 31 ng/L (Fig 2A), a FSH value of more than 11.5 mIU/mL (Fig 2B), and an inhibin B:FSH ratio of ≤ 2.52 pg/mIU (Fig 2B) were determined to be the optimum levels.
The scatterplot of the serum level of inhibin B and sperm concentration is shown in Figure 3, with Spearman's correlation coefficient rs = 0.70. Eighty-nine of 89 azoospermic individuals (100.0%) had a serum level of inhibin B ≤ 31 ng/L (Table 3). Of 149 individuals with oligospermia or a normal sperm count, 67 (45.0%) had a serum inhibin B level more than 31 ng/L. The specificity of serum inhibin B level ≤ 31 ng/L for identifying azoospermia was 45.0%, and the positive predictive value was 52.1% (Table 3).
A scatter plot of serum level of FSH and sperm concentration is shown in Figure 4 with Spearman's correlation coefficient rs = −0.71. Eighty-two (78.1%) of 105 azoospermic patients had a FSH level more than 11.5 mIU/mL (Table 3). One hundred twenty-six (74.1%) of those with oligospermia or a normal sperm count had a FSH level ≤ 11.5 mIU/mL. The specificity of a FSH level more than 11.5 mIU/mL for identifying azoospermia was 74.1%, and the positive predictive value was 65.1% (Table 2).
The optimal value for the inhibin B:FSH ratio for identifying participants with azoospermia was 2.52 pg/mIU. The ratio of inhibin B:FSH was not more useful for identifying participants with azoospermia than was FSH more than 11.5 mIU/mL. The specificity of the ratio ≤ 2.52 pg/mIU was 74.5%, and the positive predictive value was 63.8% (Table 2). The area under the ROC curves was 0.83 for FSH and 0.83 for the inhibin B:FSH ratio (P = .31; Fig 2B).
We have demonstrated that, although the serum level of inhibin B is directly and that of FSH is inversely correlated with sperm concentration, determination of the serum levels of neither inhibin B nor FSH, nor their ratio, is adequate for distinguishing between azoospermic and nonazoospermic long-term survivors of childhood cancer because of the lack of specificity and positive predictive value of both serum markers.
Previous investigations demonstrated a direct correlation between sperm concentration and inhibin B level in relatively small cohorts of long-term survivors of Hodgkin lymphoma treated before age 16 years (n = 21),12 of Ewing or soft tissue sarcoma before age 22 years (n = 13),21 and a variety of malignancies before age 16 years (n = 21 to 23).13,22 Thomson et al23 reported that basal inhibin B was significantly lower among azoospermic adult male survivors of childhood cancer (n = 7) than nonazoospermic adult male survivors of childhood cancer (n = 20).
Although the preceding publications demonstrated a significant correlation between the serum level of inhibin B and FSH and sperm concentration, only Romerius et al14 previously examined the sensitivity, specificity, and positive and negative predictive values for these markers and azoospermia in survivors of childhood cancer. By using ROC analysis, they found that the optimal value for inhibin B was 50 ng/L, which was also the lower limit of the normal range for the inhibin B assay they used. When using this value, the sensitivity was 0.91 and the specificity was 0.90. The positive predictive value was 66% (95% CI, 47% to 81%), and the negative predictive value was 98% (95% CI, 93% to 100%).14 In addition, they reported that the sensitivity of FSH for azoospermia was 0.96 and the specificity was 0.96 by using a FSH threshold of 10.9 IU/L. The positive predictive value was 50% (95% CI, 35% to 67%), and the negative predictive value was 99% (95% CI, 94% to 100%).14
The study by Romerius et al14 included only 19 participants (14.7%) who had been treated with sterilizing doses and 40 (32.5%) who had been treated with nonsterilizing doses of cisplatin or alkylating agents with or without radiation therapy to cranial, supradiaphragmatic, and/or infradiaphragmatic treatment volumes and/or total-body irradiation. By contrast, all of the individuals evaluated in this study were exposed to potentially gonadotoxic treatment. In the study by Romerius et al,14 only 23 (17.8%) of 129 participants were azoospermic compared with 105 (38.2%) of 275 in this study. The differences in the composition of the study populations, the larger number of events, and the larger size of the cohort may have contributed to the differences in specificity observed between the two studies.
We evaluated the inhibin B:FSH ratio based on the data of Andersson et al24 who reported that an inhibin B:FSH ratio more than 23.5 ng/IU gave a sensitivity of 62% and specificity of 95% for identifying men of proven fertility in their study population. The ratio was more sensitive than either inhibin B or FSH for identifying men of proven fertility.24 Our data suggest that the ratio is not more sensitive or specific for identification of azoospermic cancer survivors than either FSH or inhibin B. However, the limited recruitment of only 60.6% of the potentially eligible patients should be considered in the interpretation of our study findings. Notably, many of those who did not submit semen specimens had proven fertility (89 [49.7%] of 179).
Our results have important implications for counseling male long-term survivors of childhood cancer regarding their potential for fertility. Although the presence of a detectable level of inhibin B excluded azoospermia in the majority of men evaluated in this study, neither inhibin B nor FSH levels, individually or combined, should be used for advising men regarding the adequacy of spermatogenesis. All sexually active males should be instructed that, regardless of the levels of these biomarkers, adequate methods of contraception should be used if paternity is not a desired outcome of sexual activity. Moreover, because the return of spermatogenesis has been reported after prolonged periods of post-treatment azoospermia,25–29 similar precautions should be exercised regardless of the results of a single semen analysis. Additional research is needed to identify a surrogate marker that has greater specificity and positive predictive value for sperm concentration than any of the currently available clinical biomarkers.
|Characteristic||Semen Analysis (n = 275)||No Semen Analysis (n = 179)|
|No. of children|
|Cancelled appointment/no show||0||4|
|Unable to produce specimen||0||7|
Supported in part by Grant No. CA-21765 from the United States Public Health Service and by funding from the American Lebanese Syrian Associated Charities (to St. Jude Children's Research Hospital).
Presented in part at the 43rd Annual Conference of the International Society of Paediatric Oncology, Auckland, New Zealand, October 28-30, 2011.
Authors' disclosures of potential conflicts of interest and author contributions are found at the end of this article.
The author(s) indicated no potential conflicts of interest.
Conception and design: Daniel M. Green, William H. Kutteh, James L. Klosky, Monika L. Metzger, Fariba Navid, Melissa M. Hudson
Administrative support: Leslie L. Robison, Melissa M. Hudson
Collection and assembly of data: Raymond W. Ke, William H. Kutteh, Leslie L. Robison, Melissa M. Hudson
Data analysis and interpretation: Liang Zhu, Nan Zhang, Charles A. Sklar, Sheri L. Spunt, DeoKumar Srivastava, Leslie L. Robison, Melissa M. Hudson
Manuscript writing: All authors
Final approval of manuscript: All authors