Effect of childhood cancer treatment on fertility markers in adult male long-term survivors

Abstract

Background

Although it is accepted that pediatric cancer treatment harbors a risk of gonadal damage, large cohort studies using up-to-date fertility markers are lacking.

Procedure

The aim of our study was to evaluate the gonadal toxicity of childhood cancer treatment using fertility markers. We included 248 adult male long-term survivors of childhood cancer. Median age at diagnosis: 5 years, median age at follow-up: 24 years, median follow-up time 18 years. We evaluated patient characteristics, treatment modalities, testicular size, and endocrinological parameters including Inhibin B, luteinizing hormone (LH), follicle-stimulating hormone (FSH), and testosterone.

Results

The median value of Inhibin B in the cancer survivor group was 126 ng/L versus 177 ng/L in the control group (P < 0.001). In the survivors, 67% had Inhibin B levels below the normal reference value of 150 ng/L compared with 26% in the control group (P < 0.05). Inhibin B was the most sensitive discriminator between survivors and controls. Significantly decreased Inhibin B levels and increased FSH levels were found in men treated for Hodgkin and non-Hodgkin lymphoma, acute-myeloid leukemia, neuroblastoma, and sarcoma as compared to other malignancies. Cumulative dosages of procarbazine and cyclophosphamide were the only independent chemotherapy-related predictors for decrease of Inhibin B levels and increase of FSH. Age at time of treatment did not influence post-treatment Inhibin B or FSH levels.

Conclusions

Severe gonadal impairment is a risk in a considerable subgroup of childhood cancer survivors based on current fertility markers like Inhibin B. Males receiving gonadotoxic treatment before puberty are not protected from post treatment gonadal dysfunction. Pediatr Blood Cancer 2009;52:108–112. © 2008 Wiley-Liss, Inc.

INTRODUCTION

In the past decades, survival of childhood cancer improved significantly following optimized treatment modalities 1. Therefore, the number of survivors of childhood cancer in the general population is estimated to increase 2, 3. Several studies have shown an impaired gonadal function in adult male survivors of childhood cancer 4, 5. Post-treatment gonadal damage showed to be depending on the type of used agents and cumulative dosages administered during childhood 6–8. Usually, studies were based on one type of malignancy and treatment protocol and had limited number of survivors included 6–8. Age at time of treatment was suggested to give some protection against chemotherapy induced gonadal damage due to quiescent stage of the testis in the pre-pubertal age 9. At present, semen analysis serves as the gold-standard for predicting the reproductive capacity of men in their reproductive age. However, young adults treated for childhood cancer are reluctant to provide a semen sample for analysis. Therefore, reliable serum markers that reflect gonadal function in men may serve as a more convenient screening tool in determining reproductive function in male cancer survivors at least to start with. In recent studies, we and others have shown that Inhibin B is the best serum marker for spermatogenesis 10–15. Inhibin B values after puberty depend on FSH secretion and are associated with sperm concentration and testicular volume not only in healthy individuals but also after chemotherapy and/or radiotherapy 16, 17. However, studies in large cohorts of adults treated for childhood cancer describing the role of Inhibin B as a male fertility marker are not available as yet. The aim of this study was to evaluate gonadal damage in a large single center cohort of male childhood cancer survivors using modern fertility markers.

PATIENTS AND METHODS

Between January 2003 and September 2006, 291 male survivors of childhood cancer were recruited to our late effects outpatient clinic. All had been treated in our center. Of them, 30 did not respond and 13 survivors were not included in the analysis because they did not receive chemotherapy or radiotherapy during their childhood treatment. Written informed consent was obtained from all participants, according to protocols approved by the ethical review board of the Erasmus MC. The distribution by diagnosis is given in Table I. Of the Hodgkin lymphoma (HL) group, 22 included survivors have been described in a previous larger series 15. In order to describe the complete cohort of survivors these patients were included in the study.

Table I. Patient Characteristics per Diagnosis Group

Diagnosis	N	Age at diagnosis	Age at visit	Inhibin B	FSH	LH	Testosterone	Testis volume
ALL, acute lymphoblastic leukemia; AML, acute myeloid leukemia; HL, Hodgkin lymphoma; NHL, non-Hodgkin lymphoma; MMT, malignant mesenchymal tumor. Others consist out of 1 embryonal adenocarcinoma, 1 langerhans histiocytosis, 2 hepatoblastomas, 1 malignant schwannoma, 1 nasopharynx carcinoma, 2 yolk sac tumors, 1 Grawitz tumor. Values are given as median and range.
ALL	88	5.0 (0.3–15.0)	23.0 (18.0–40.0)	140 (0–393)	3.6 (0.5–58.7)	3.5 (0.1–14.9)	15.5 (3.7–27.3)	20.0 (2.0–25.0)
AML	11	3.0 (0.5–14.0)	21.0 (18.0–33.0)	68 (10–195)	3.9 (0.3–15.0)	3.2 (0.1–8.9)	12.8 (8.5–45.1)	15.0 (10.0–20.0)
HL	28	11.0 (2.0–15.0)	24.0 (18.0–41.0)	54 (0–274)	9.2 (1.7–51.0)	3.6 (1.2–10.9)	13.2 (8.8–32.7)	15.0 (8.0–22.5)
NHL	42	8.0 (2.0–14.0)	23.0 (18.0–36.0)	88 (5–232)	5.5 (0.9–31.7)	4.0 (0.9–11.5)	15.5 (7.6–27.9)	15.0 (9.0–22.5)
Renal tumors	27	3.0 (0.7–11.0)	23.0 (18.0–39.0)	146 (13–244)	4.0 (0.5–47.4)	3.3 (0.1–16.3)	14.5 (4.5–28.9)	17.5 (10.0–25.0)
MMT	18	5.5 (1.0–14.0)	25.0 (18.0–36.0)	55 (0–226)	14.4 (0.5–65.8)	4.7 (0.1–12.9)	14.6 (2.8–28.9)	15.0 (4.0–16.5)
Osteosarcoma	9	12.0 (1.0–15.0)	23.0 (18.0–37.0)	146 (38–339)	6.0 (0.7–17.3)	3.6 (2.0–8.5)	14.8 (9.7–23.2)	20.0 (11.0–25.0)
Ewing sarcoma	6	8.0 (5.0–15.0)	25.5 (18.0–36.0)	36 (15–299)	8.6 (1.7–15.8)	3.6 (1.1–8.1)	16.1 (9.7–20.2)	16.3 (15.0–20.0)
Neuroblastoma + ganglioneuroblastoma	11	0.1 (0.1–1.0)	24.0 (18.0–30.0)	79 (21–212)	6.3 (3.3–20.6)	3.9 (1.3–7.5)	15.0 (7.6–20.4)	15.0 (12.0–20.0)
Others	8	1.5 (0.3–9.0)	20.0 (18.0–30.0)	114 (10–295)	3.6 (1.4–21.4)	3.7 (1.4–7.2)	14.4 (12.5–24.7)	22.5 (15–25.0)
Controls	74	—	33.4 (23.3–50.6)	176.5 (60–556)	3.6 (0.6–18.0)	2.7 (0.3–19.2)	16.8 (4.7–27.7)	16.5 (11.0–30.0)

At the time of the study, all survivors were at least 18 years of age and at least 5 years after cessation of treatment. Median age at initial diagnosis was 5 years (range 0–15 years). Median age at follow-up was 23 years (range 18–41 years) with a median follow-up time of 18 years (range 5–39 years, Table I). Complete physical examination including measurement of testicular volume and serum hormone analysis consisting of LH, FSH, testosterone, and Inhibin B was performed. Testicular volume was measured by Prader Orchidometer. Sperm analysis performed in a subset of cancer survivors was performed according to the 1999 WHO laboratory manual for the examination of human semen and sperm–cervical mucus interaction 18.

The control group consisted of 74 normospermic men, with a sperm concentration >20 × 10⁶/ml and progressive motility >50%, aged median 33.4 years (range 23–51 years). Peripheral blood samples in survivors and controls were taken at time of their visit. The normospermic men were recruited from the fertility clinic. During andrological work-up, they were found to be normospermic. The reference values of LH, FSH, Inhibin B, and testosterone values for male adults in our institute are 1.5–8.0 U/L, 2.0–7.0 U/L, 150–400 ng/L, and 10.0–30.0 nmol/L, respectively. Within-assay and between-assay coefficients of variation (CV) for Inhibin B were 9% and 15%, LH 6% and 9%, FSH 5% and 11%, testosterone 3% and 4.5%, respectively 19. The cut-off value of 150 ng/L for Inhibin B is taken as it has been shown that this provided the highest sensitivity and specificity in identifying low semen quality 14.

Statistical Analyses

The computer program Statistical Package for Social Science (Version 11.1, SPSS Inc., Chicago, Illinois) was used for data analysis. Non-parametric analyses were done as a non-normal distribution if the distribution was not normal. P-values <0.05 were considered significant. Survivor characteristics are expressed as median with range.

RESULTS

The range of serum hormone values of the complete survivor group and the control group are depicted in Table I. In general, the survivors had higher FSH and LH levels and lower testosterone and Inhibin B levels than the controls (Tables I and II). Inhibin B was the strongest discriminator between the survivors (median of 123.0 ng/L, range 0–393 ng/L) and the controls (median 176.5 ng/L, range 60–556 ng/L; P < 0.0001, Fig. 1).

Table II. Comparison of Serum Hormone Levels Between Survivors and Controls

	Survivors (n = 248)	Percentage of abnormal levels	Controls (n = 74)	Percentage of abnormal levels	P-value
Values are indicated as median and range. Percentages of abnormal levels for each serum marker and sperm concentration are given. a n = 221; b n = 244; c n = 21; d n = 233.
Inhibin B (ng/L)	123.0 (0–393)a	65%	176.5 (60–556)	25%	<0.0001
Testosterone (nmol/L)	14.7 (2.8–45.1)b	8%	16.8 (4.7–27.7)	10%	0.05
FSH (U/L)	4.6 (0.3–65.8)b	32%	3.6 (0.6–18.0)	17%	0.01
LH (U/L)	3.6 (0.1–16.3)b	10%	2.7 (0.3–19.2)	2%	<0.0001
Sperm concentration (1 × 10^6/ml)	0.01 (0.0–158.0)c	76%	52.5 (21.1–234.0)	0%	<0.0001
Testicular volume (ml)	17.5 (2–35)d	17%	16.5 (11–30)	14%	0.391

Figure 1. Inhibin B values of controls and childhood cancer survivors. Each dot represents one person. The horizontal lines represent the median values of Inhibin B in the normospermic controls (176.5 ng/L) and in the survivors (123.0 ng/L, P < 0.0001).

One hundred forty-five of the 221 survivors (65%) in whom Inhibin B was measured had values below 150 ng/L in contrast to 19 (26%) of the 74 controls (Fig. 1). When analyzed by type of cancer, survivors of HL, non-HL, malignant mesenchymal tumors (MMT), acute myeloid leukemias, sarcoma's, and neuroblastomas showed median Inhibin B levels below 100 ng/L suggestive for severe gonadal dysfunction (Table I, Fig. 2). Survivors of acute lymphoblastic leukemia and renal tumors had relative normal median Inhibin B levels of respectively, 140 ng/L and 146 ng/L (Table I, Fig. 2).

Figure 2. Median Inhibin B values and quartile ranges per diagnosis. The horizontal dotted line represents the lower threshold of normal (150 ng/L).

We compared sperm analysis with endocrinological parameters in a subgroup of 21 patients; nine had an azoospermia with a median Inhibin B level of median 39 ng/L (range 0–79 ng/L). Two patients were diagnosed with an oligospermia (sperm concentration between 5 and 20 million/ml), had sperm concentrations of 13 × 10⁶ and 15 × 10⁶ spermatozoa/ml, and had Inhibin B levels of 67 ng/L and 72 ng/L. Five patients who had severe oligo-asthenospermia had sperm concentrations below 5 × 10⁶/ml and Inhibin B levels of median 25 ng/L (range 2–39 ng/L). Two of them had severe oligospermia and with only a few immotile spermatozoa in their semen, with Inhibin B levels of 2 ng/L and 19 ng/L. The five survivors with a normospermia had median Inhibin B levels of 145 ng/L (range 125–234 ng/L). Inhibin B levels showed a significant correlation with sperm concentration in both survivors (r = 0.671, P = 0.01) and controls (r = 0.345, P = 0.03).

Using a linear regression model, treatment determinants for low Inhibin B levels (Inhibin B < 150 ng/L) were calculated. Only regimens containing cyclophosphamide or procarbazine but not any other drugs, were independent determinants for low Inhibin B levels in the patients treated with chemotherapy (P < 0.01). None of the other chemotherapeutic agents was independently associated with altered Inhibin B levels. Cyclophosphamide was part of the treatment protocol in 131 of the 248 survivors with a median dosage of 4.8 g/m² (range 0.25–32 g/m²). Inhibin B levels showed a significant negative correlation with total cumulative dosage (TCD) of cyclophosphamide (r = −0.531, P < 0.001) (Fig. 3). Inhibin B levels were abnormal in all but one survivor receiving a TCD of more than 10 g/m² (Fig. 3).

Figure 3. Dose-effect of cyclophosphamide (g/m²) on Inhibin B levels. The horizontal line shows the lower threshold of normal for Inhibin B (150 ng/L) the other line shows the linear regression (Rsq = 0.201, correlation coefficient R = −0.531, P > 0.001).

HL survivors treated with MOPP had lower Inhibin B levels (median 15 ng/L vs. 143 ng/L, P < 0.001), smaller testis volumes, (i.e., 12.7 ml vs. 17.8 ml, P < 0.01), higher FSH levels (median 22.6 U/L vs. 4.7 U/L, P < 0.001), and higher LH levels (5.9 U/L vs. 2.4 U/L, P < 0.001) than the HL survivors who did not receive procarbazine. These values indicate a severe gonadal dysfunction in the 15 HL patients treated with procarbazine.

Age at which cyclophosphamide and procarbazine were administered showed no correlation with Inhibin B levels (r = −0.06, P = 0.526 and r = 0.182, P = 0.516, respectively). Young age at time of cyclophosphamide administration did not show a protective effect on gonadal function whether they received low or high TCD of cyclophosphamide (Supplemental Fig. 1A and B). This was even more obvious when we evaluated patients younger and older than 10 years with a relatively low (TCD < 10 g/m²) dosage of cyclophosphamide. Median Inhibin B level in the patients younger than 10 years (n = 58) was 138 ng/L compared to 127 ng/L in the older group (n = 29) (P = 0.402). Although the groups were smaller, no age-effect was seen in the patients receiving cyclophosphamide in higher concentrations (TCD > 10 g/m²). The survivors treated before the age of 10 (n = 28, Inhibin B available in 22) had median Inhibin B level of 21 ng/L (range 0–218 ng/L) compared to 10 ng/L (range 5–78 ng/L) in the older group (n = 5, P = 0.845).

The 10 survivors (1 HL, 4 AML, 4 ALL, and 1 NHL), who received total body irradiation (TBI 7.5 or 12 Gy) as part of myelo-ablative therapy, had extremely low Inhibin B levels (median 10.0, range 0.0–62.0 ng/L). Of those, only two survivors had low testosterone levels (7.6 and 8.6 nmol/L) whereas the others had normal testosterone levels (≥10 nmol/L). No impairment in gonadotrophins was found in the subgroups. Additionally, four survivors (3 ALL, 1 AML) had received testicular irradiation (24 Gy) on both testes. One survivor with a MMT was treated with pelvic radiotherapy (45 Gy), which included the scrotum. Testicular volume was evaluated in three as one had testicular implants (Inhibin B = 0 ng/L) and in one patient, it was unfortunately not documented. These three survivors showed severe impairment of testicular growth (testis volume median 2.0 ml, range 1–3 ml). Median Inhibin B level available in the four survivors, in whom the testicles were not removed, was 10 ng/L, (range 0–13 ng/L).

Testicular volume did not differ between the controls and the survivors in the total group nor in one of the subgroups by diagnosis (P = 0.391). This is reflected by the fact that only 18 males (7.3%) of the cancer survivors had subnormal levels of testosterone (<10 nmol/L) compared to eight men (10.8%) in the control group. All but three men out the patient group fully completed pubertal development compared to all in the control group.

DISCUSSION

Although several studies have shown the value of modern fertility markers for prediction of spermatogenesis, studies on gonadal function after chemotherapy and irradiation in large series of adult male cancer survivors, using Inhibin B as a novel serum markers are lacking 5, 15, 20–22. Therefore, we performed a study in a full single center cohort of male childhood cancer survivors (n = 248).

We found a significant difference in reproductive hormone levels between survivors and controls, of which Inhibin B proved to be a better discriminator than FSH for the assessment of spermatogenesis between the two groups. This was already known in healthy individuals but this study shows that this is also the case in cohorts of male long-term survivors of childhood cancer. We used a cut-off value of 150 ng/L. We acknowledge that this cut-off value is arbitrary, and therefore performed the same analyses with a cut-off value of 80 ng/L and found that cyclophosphamide and procarbazine were the only independent treatment determinants for impaired gonadal function.

Patients with HL, non-HL, MMTs, acute myeloid leukemia, Ewing sarcoma, and neuroblastoma have a relative high risk of subfertility after treatment. This is in line with the review article by Brougham and Wallace 23, who divided childhood cancer types into three subgroups according to risk of subfertility. Our study confirms that diagnostic subgroups are indicative for post-treatment gonadal function, but mainly explained by the different treatment regimes accompanying these diagnoses.

In our study, cyclophosphamide and procarbazine containing regimens, both compounds with alkylating properties were the most deleterious drugs for the gonadal function. This finding supports earlier studies in which these drugs were associated with severe gonadal dysfunction 24, 25. Cyclophosphamide showed a negative dose correlation with Inhibin B levels. Almost all survivors who received a cyclophosphamide TCD of 10 g/m² or more were at risk for severe spermatogenic dysfunction. These findings support the earlier observations in which a dose-dependent effect of cyclophosphamide was found 20, 24. This dose effect was not found for procarbazine containing regimens, which indicates that this agent might already be gonadotoxic in low concentrations in males. Our study confirms that not only the type of agent used but also the cumulative dosage is important in predicting post-treatment gonadal function 26, 27.

Total body- and testicular irradiation causes severe gonadal dysfunction 28–30, which is illustrated by the fact that eight out of 10 survivors treated with TBI had almost undetectable Inhibin B levels (≤26 ng/L). In addition, local radiotherapy to the testis was administered in five men who all showed extremely low Inhibin B levels. Although semen analyses were not available of these men, the previous published correlations between Inhibin B levels and sperm count indicates that these survivors probably have a severely impaired spermatogenesis 10, 15.

This study illustrates that young age at time of childhood cancer treatment does not prevent post-treatment gonadal damage. Although such protection at very young age has been suggested in the past, our findings confirm those of Thomson et al. 22 who observed severe gonadal dysfunction in male survivors even if treated in pre-pubertal life. Pre-pubertal age has been classically characterized as a quiescent period potentially providing some kind of protection against chemotherapy 9. However, during this period, there is active proliferation of sertoli and leydig cells, which might explain the damage after cytotoxic treatment 31–34. Childhood cancer patients and/or their parents should therefore be informed about the risk of infertility after chemotherapy independent of age and pubertal stage. Therefore, sperm preservation should be considered for risk groups as young as possible, and if necessary assisted by artificial procedures like electro-ejaculation methods 19, 35.

The question whether Inhibin B can replace semen analysis as a gold-standard remains unanswered. Although Inhibin B is highly specific, it still is an indirect marker and controversies exist. Sperm analysis remains the gold-standard especially as with modern assisted reproduction techniques, like intra-cytoplasmic sperm injection, only a few normal spermatozoa are needed for conception. Therefore, we propose to encourage survivors to have Inhibin B levels measured after chemotherapy as a first step to assess gonadal damage in order to identify survivors at risk for impaired spermatogenesis.

In the total group treated with chemotherapy, testosterone levels were slightly lower than the controls; however, only 7% of the survivors had low testosterone values. This might indicate that the leydig cells are more resistant to gonadotoxic treatment than the germinal cells. The minimal dysfunction of leydig cells is supported by the observation that all except three survivors, of which two received testicular radiotherapy, fully completed pubertal development, which is mainly dependent on testosterone production. It is conceivable that most boys will complete their pubertal development possibly by an increase in LH levels, which occurred in our cancer survivors, that is, significant higher levels of LH were found in the survivors compared to in the normospermic men. This might indicate that in some survivors (10% in this study) a leydig cell dysfunction exists, which can be compensated by elevated LH levels. This underscores that the assessment of reproductive capacity in childhood cancer survivors by solely evaluating secondary sexual characteristics or testosterone is insufficient.

This study demonstrates the importance of fertility counseling and pre-treatment semen cryopreservation in male survivors of childhood cancer. Moreover, the results of this study show the ongoing need to consider the risk of gonadal toxicity of certain agents in the design of new protocols for newly diagnosed children with cancer. Inhibin B levels may be helpful to assess the gonadal function. Determination of Inhibin B levels, being the best marker for assessing gonadal function and recovery, is a relative easy and quick non-invasive method to assess gonadal function. Although it can be used to identify patients at risk with impaired reproductive capacity, combination with sperm analysis remains mandatory to evaluate reproductive options when there is an active child wish. Our results emphasize the need for gonadal function assessment and fertility counseling in long-term survivors of childhood cancer.

Acknowledgements

We acknowledge M. van Baalen for her support in this study.