Levonorgestrel Implants Enhanced the Suppression of Spermatogenesis by Testosterone Implants: Comparison between Chinese and Non-Chinese Men

Context: Previous male contraceptive studies showed that progestins enhance spermatogenesis suppression by androgens in men.

Objective: We compared the efficacy of spermatogenesis suppression by the combination of levonorgestrel (LNG) with testosterone (T) implants to that by T implants alone in two different ethnic groups.

Design: This was a randomized trial performed in two centers with two treatment groups.

Settings: The study was performed at the Academic Medical Center in the United States and the Research Institute in China.

Participants: Forty non-Chinese and 40 Chinese healthy male volunteers were studied.

Interventions: Subjects were randomized to receive four LNG implants together with four T implants (inserted on d 1 and wk 15–18) vs. T implants alone for 30 wk.

Main Outcome Measures: The primary end point compared the efficiency of suppression to severe oligozoospermia (1 × 10⁶/ml) by LNG plus T implants vs. that by T implants alone. The secondary end point examined differences in spermatogenesis suppression between Chinese and non-Chinese subjects.

Results: LNG plus T implants caused more suppression of spermatogenesis to severe oligozoospermia during the treatment period than T implants alone at both sites (P < 0.02). In Chinese men, severe oligozoospermia was achieved in more than 90% of the men in both treatment groups. Suppression to severe oligozoospermia was less in the non-Chinese men (59%) after T alone (P < 0.020); this difference disappeared with combined treatment (89%). T implant extrusion occurred in six men. Acne and increased hemoglobin were the most common adverse events.

Conclusion: T implants resulted in more pronounced spermatogenesis suppression in Chinese men. Addition of LNG implants to T implants enhanced the suppression of spermatogenesis in the treatment period in both Chinese and non-Chinese men.

ALTHOUGH MANY METHODS of female hormonal contraception are available, including pills, injectables, implants, intrauterine systems, vaginal rings, and transdermal patches, a male hormonal method is still in the development phase. Studies in fertile couples showed that once spermatogenesis suppression to severe oligozoospermia (defined as <1 million/ml semen) is achieved by testosterone (T) ester injections alone (1–3) or T implants plus depo-medroxyprogesterone acetate (4), contraceptive efficacy is similar to that with female contraceptive pills or injections. Studies are ongoing in China with a large phase III study using monthly T undecanoate injections alone for male contraception (Gu, Y. Q., unpublished observations). Because of the potential long-term effects of androgens on the growth of the prostate and the questionable concern about cardiovascular effects, the dose of androgens should be kept at the lowest possible level. In contrast to Chinese subjects, previous studies showed that suppression of spermatogenesis with T alone appeared to be less effective in non-Chinese men (1, 2, 5, 6). Thus, for non-Chinese men, addition of progestins enhanced the effect of androgens on the suppression of gonadotropins and, consequently, spermatogenesis (5–8).

Levonorgestrel (LNG), one of the most commonly used progestins in female oral contraceptive pills, has been shown to enhance the effects of weekly T enanthate injections (9, 10). We (11) and others (12) have reported previously that LNG implants provided a steady-state delivery of LNG in contrast to the bursts of LNG after oral administration, and LNG enhanced the effect of T on suppression of spermatogenesis. LNG with T enanthate or T undecanoate injections were efficacious in inducing severe oligozoospermia in the majority of healthy men, but not when LNG was combined with transdermal T patches. Our data suggested that a critical level of androgens and a balance between progestins and androgens are both required for adequate suppression of spermatogenesis (11).

In the current study the primary aim was to compare the degree of suppression of spermatogenesis and to assess potential adverse effects of steady-state delivery of T and LNG using LNG and T implants vs. T implants alone in normal men. As a secondary aim, we compared the efficacy of suppression of spermatogenesis of non-Chinese subjects recruited in Los Angeles, California, with that of Chinese men in Nanjing, Jiangsu, China, using the same study protocol. The data from this study will provide useful information for the development of implants of modified androgens and progestins for a long-acting male contraceptive method.

Subjects and Methods

Subjects

Forty healthy male volunteers between the ages of 24 and 50 yr (mean ± sem, 37.6 ± 1.0) were recruited through radio, newspaper, and bulletin board advertisements at the Los Angeles Center, and 41 volunteers between the ages of 25 and 45 yr (35.9 ± 0.8) were recruited from a glass/lens factory at the Nanjing Center. The subjects in the Los Angeles Center consisted of 28 white, three African-American, seven Hispanic, and two of other ethnicities. None of the subjects was of East Asian origin. All subjects from the Nanjing Center were Chinese (Table 1). The subjects had no significant medical history, and all had a normal physical examination during recruitment. None of the subjects in either center was undernourished, as determined by history and physical examination. They had normal baseline hematology, blood biochemistry, urinalysis, fasting lipid profile, and three consecutive normal semen analysis at 2-wk intervals (sperm count, >20 million/ml; motility, >50%; oval forms, >10%) according to the World Health Organization Manual for the Examination of Human Semen (13).

Medications

T implants (pellets) consisted of 200 mg crystalline T/implant. These implants each released 1.5 mg/implant·d. They were gifts from N. V. Organon (Oss, The Netherlands). The T implants were anticipated to maintain serum T levels within the adult male physiological range for 4–6 months (4, 14–16).

The LNG implants (Jadelle, Leiras Pharmaceuticals, Turku, Finland) were obtained as a gift from Leiras through Dr. Harold Nash (The Population Council, New York, NY). Each Jadelle system consisted of two 43-mm capsules that release 36–49 μg/implant·d LNG from a cured homogenous mixture of the drug and a polydiniethyl siloxane elastomer covered by thin-walled silicone rubber tubing (11).

Study design

After a pretreatment observation phase of 4 wk, the eligible subjects were randomized into either the T alone or the T plus LNG group within each center. Randomization was achieved using a computer-generated random number generator, and separate randomization schedules were developed for each center. The study was unblinded. On the first day of the treatment phase, four T implants were inserted into the sc fat of the abdominal wall of all subjects under local anesthesia. Two Jadelle systems (four LNG implants) were also inserted on d 1 in the arm of each subject assigned to the T plus LNG group as previously described (11). A second set of four T implants was inserted in the contralateral abdominal side in both groups halfway through the study. Initially, the study was designed such that another four T implants would be inserted at wk 18. This occurred in the first 15 subjects. The team decided to measure serum T levels at wk 15 and 18 for the first 10 subjects who were randomized into the study. This decision was made because data from Turner et al. (4) suggested that four 200-mg T implants might not sustain normal T levels or continued suppression of spermatogenesis for more than 16 wk. The concern was that the original protocol for the insertion of the second sets of T implants at wk 18 might not be adequate to maintain serum T levels. Preliminary data from this study showed serum T levels to be normal in nine of the first 10 subjects at wk 15. However, six of the 10 subjects had serum T levels below the normal range at wk 18. This confirmed that the T implants were insufficient to maintain serum T concentrations within the adult male range, and the subjects might be hypogonadal by wk 18. To prevent this from occurring, we amended the protocol to change the insertion of the second sets of T implants at wk 15 instead of wk 18. This change was introduced after 15 subjects had followed the original protocol at the Los Angeles Center. Because of the change in the date of the second T implantation, the total treatment period was shortened from 36 to 30 wk. The Nanjing Center started about 6 months later than the Los Angeles Center; all subjects in Nanjing followed the amended protocol, with the second set of T implants placed at wk 15.

At the end of the 30-wk treatment phase, the Jadelle implants were removed under local anesthesia. All subjects underwent treatment for 30–36 wk, then completed a recovery period for at least 12 wk or until full recovery, defined as a sperm concentration equivalent to or above the geometric mean of the three pretreatment sperm concentrations, or more than 20 × 10⁶/ml. The insertion of the LNG and T implants and the removal of the LNG implants were performed by a single physician in each center. The physicians were very experienced with Jadelle implant insertion and were trained for T implant insertion by an experienced Australian physician.

All subjects had complete physical examinations and interviews performed by a physician at screening; at wk 6, 15, 30, and 42; and at the end of the study. Digital examination of the prostate was performed at screening, at wk 6 and 30, and at the end of the study. Testicular volume was measured at both centers by different observers using the Prader orchidometer with ellipsoids up to 35 ml. A transrectal ultrasound of the prostate was performed before treatment, at the end of treatment, and at wk 12 after removal of the implants. Semen analysis was performed every 3 wk on samples obtained by masturbation after at least 48 h of abstinence during the 30- to 36-wk treatment period and then every 4 wk during the recovery period until sperm concentrations returned to pretreatment levels or more than 20 × 10⁶/ml. Blood samples for the measurement of clinical chemistry (including glucose, blood urea nitrogen, creatinine, electrolytes, albumin, alanine aminotransferase, aspartate aminotransferase, alkaline phosphatase, bilirubin, and fasting lipid profile), hematology, prostate-specific antigen (PSA), and serum hormones (LH, FSH, T, free T, SHBG, and LNG) were collected at baseline. All patients returned at 3- to 6-wk intervals for blood draws to measure serum hormones until wk 30 and 12 wk after removal of the implants. To investigate whether serum FSH and LH could be early markers to distinguish complete responders (whose sperm concentration was suppressed to less than 1 × 10⁶/ml) vs. incomplete responders, blood samples (3 ml) were obtained on d 4, 7, 14, and 21 for serum FSH and LH measurements in all subjects. Safety laboratory blood samples were drawn at regular intervals during treatment and recovery periods. Sexual function and mood were assessed by questionnaires that the subjects answered daily for 7 consecutive days before every clinic visit. The validated questionnaire covered four domains: sexual desire, sexual enjoyment, sexual activity, and mood (17). This study was approved by the institutional review board/ethical committee of each center. All subjects gave informed written consent. Subject compensation for time lost from work and transportation costs were based on the established guidelines and standards of each institution as appropriate for the research participants and approved by the respective institutional review board/ethical committee.

Hormone assay and semen analyses

All serum samples from the subjects were measured in the same validated assays (11, 18) in the Los Angeles center. Serum LNG was measured by RIA after extraction with hexane and ethyl acetate using reagents obtained through Dr. Saulat Sufi (World Health Organization Collaborating Center, London, UK). The lower limit of quantitation (LOQ) of LNG in serum measured by this assay is 125 pmol/liter. The intra- and interassay coefficients were 6.2% and 8.1%, respectively. Serum T levels were measured after extraction with ethyl acetate and hexane by a specific RIA using reagents from ICN Biomedicals, Inc. (Costa Mesa, CA). The LOQ of serum T measured by this assay was 0.87 nmol/liter. All results below this value were reported as 0.87 nmol/liter. The mean accuracy (recovery) of the T assay, determined by spiking steroid-free serum with varying amounts of T (0.9–52 nmol/liter), was 104% (range, 92–117%). The intra- and interassay coefficients of the T assay were 7.3% and 11.1%, respectively, in the normal adult male range. The adult male serum T reference range in our laboratory was 10.33–36.17 nmol/liter (298–1043 ng/dl), respectively. The serum percent free T was measured by equilibrium dialysis after incubation of the sera with tritium-labeled T and overnight dialysis at 37 C. The intra- and interassay coefficients were 6.2% and 16.4%, respectively. Serum free T was calculate by multiplying the percent free by the total serum T (19). The normal range was 0.13–0.58 nmol/liter. Serum FSH and LH were measured by the highly sensitive and specific fluoroimmunometric assays with reagents provided by Delfia (Wallac, Gaithersburg, MD). The intraassay coefficients of variations for LH and FSH were 3.9% and 5.4%, respectively, and the interassay variations for LH and FSH were 6.8% and 9.8%, respectively (adult normal male range: LH, 1.3–8.1 IU/liter; FSH, 1.4–9.5 IU/liter). For both LH and FSH assays, the LOQ was determined to be 0.1 IU/liter. SHBG was measured by fluoroimmunometric assay with reagents provided by Delfia. The LOQ was determined to be 0.5 nmol/liter. The intra- and interassay coefficients were 2.3% and 6.1%, respectively. The normal range for this assay was 10.8–46.6 nmol/liter. All assays were run within the same series to minimize interassay variability. When results were below the LOQ, the LOQ value was reported and used in data analysis.

Semen samples were collected after 48–72 h of abstinence by masturbation into sterile plastic containers and were analyzed after liquefaction. Semen analyses were performed according to the recommended methods described in the World Health Organization Laboratory Manual for the Examination of Human Semen and Sperm-Cervical Mucus Interaction. Sperm concentration and motility were assessed using the hemocytometer method and visual assessment of motility, respectively.

Statistical analyses

The primary goal of this study was to compare the effectiveness of T implants alone with that LNG implants plus T implants on the suppression of spermatogenesis to severe oligozoospermia (defined as sperm concentration <1 × 10⁶/ml) in normal men at the end of the treatment period. Type I error was set at 0.05. To account for the change in the protocol, we first compared the groups for sperm suppression during the second treatment period using logistic regression, with time to second insertion of T implants as a covariate. Because these analyses showed no effect of a prolonged first period, we combined the groups by dropping the extra 3 wk in the first treatment period and aligning the second treatment period and recovery period with those of the other subjects. Therefore, all subjects were used to compare suppression in the second treatment period. Fisher’s exact test was used to compare groups for the rate of suppression during each week of the two treatment periods. The results were not corrected for multiple testing. We also compared the log sperm concentration in the T plus LNG group to that in the T only group at each treatment week using ANOVA. The time to recovery after the second treatment period was analyzed using survival analysis.

Descriptive statistics, either means and sds or frequency distributions, were calculated for each variable at each time period. Continuous variables were analyzed to determine whether they satisfied the requirements for the ANOVA, either as measured or with suitable transformations, such as the log transformation. Sperm concentration was log transformed, and for azoospermic samples, the nonzero offset of 0.05 was selected to be sufficiently small to have minimal effect on the integrity of the analysis. We used a mixed model, repeated measures ANOVA to assess the efficacy and side effects of LNG implants plus T implant vs. those of T implant alone on other outcome variables, such as semen parameters and FSH, LH, T, and SHBG levels. Differential effects of different treatments and centers were assessed by testing for significant interactions between time and treatment or between time and center. Baseline comparisons were performed using two-way ANOVA, with center and treatment as factors. We also compared the effect of T implants between the Chinese and the non-Chinese men on the primary outcome using logistic regression.

A previous study (10) showed that TE alone (100 mg/wk) suppressed sperm production in 33% of the men and achieved severe oligozoospermia in 61% of the men, whereas TE plus oral LNG (500 μg/d) achieved azoospermia in 67% and severe oligozoospermia in 94% of the men. The percentage of subjects suppressed to severe oligozoospermia between the treatment groups reached statistical significance (P < 0.05) with 18 subjects/group. With 20 subjects/group and at least 18–19 subjects completing the study from each group, we expected that we would observe at least a similar difference in suppression to severe oligozoospermia. We anticipated that both Chinese and non-Chinese subjects would suppress to azoospermia in more than 90% of men with LNG plus T implants, and comparisons between these groups would not be clinically appropriate.

Results

Clinical outcome of the subjects

The baseline characteristics of the subjects are shown in Table 1. The Chinese subjects as a group in Nanjing had significantly lower body weight (P = 0.0001), height (P = 0.0001), and body mass index (BMI; P = 0.0001) than the non-Chinese subjects in Los Angeles. The right and left testicular volumes were also lower in the Chinese men (P = 0.0001) even after covarying for body size, whereas the baseline sperm concentration and serum levels of FSH and LH were not different between the groups. Mean serum total T (P = 0.021) and free T (P = 0.022) were higher in the Nanjing subjects. All subjects in Nanjing completed the study. Six subjects (five in the T implant alone group and one in the T plus LNG implant group) did not complete the study in Los Angeles: three discontinued because of personal reasons, and the other three because of side effects including hair loss, severe acne, and nervousness, and chest pain. Five subjects in the Nanjing Center extruded one or two T implants between d 9 and 26 after the first insertion (two in the T alone group and three in the T plus LNG group). Of these five subjects, four extruded the T implants after the second insertion. In the Los Angeles Center, one subject (T alone group) extruded one implant 84 d after the second insertion and another residual pellet 6 months after he had completed the study. Insertion of LNG implants into the arm was not associated with any adverse events, except transient mild bruising. There was no significant weight gain in either group. Other adverse events associated with T pellet insertion included signs of inflammation in one and pain at the site of insertion in 26 subjects. Other adverse events that might be related to T or LNG included acne in eight subjects, nervousness in four subjects, and hair loss in one subject. Throughout the study the parameters assessed by the psychosexual questionnaire did not show any significant change in either group of subjects. Both left and right testicular volumes decreased during treatment with T implants and T plus LNG implants, increased during the recovery period in all groups, and returned to baseline levels in all except the T plus LNG group in Los Angeles at 42 wk.

Sperm concentration

As a group, both T implants alone and T plus LNG implants suppressed sperm concentrations significantly (P < 0.0001), as shown in Fig. 1. Sperm concentrations were significantly more suppressed in the T plus LNG group compared with the T alone group from wk 15–21 in non-Chinese subjects in Los Angeles (P < 0.03); similar superior suppression was seen at wk 12, 15, and 21 in Chinese subjects in Nanjing (P < 0.02). When subjects from both sites were combined, more suppression in subjects receiving T plus LNG treatment was observed from wk 9–24 (P < 0.02). Because T implants were not removed, the recovery period started on the day when the LNG implants were removed. Sperm concentration recovered to pretreatment levels in 73% of the subjects by 12 wk after removal of the LNG implants and in 93% of the subjects by 16 wk after removal. Of the remaining five subjects, four recovered by 26 wk, and the remaining subject recovered after 38 wk (he did not attend follow-up for a 6-month period during the recovery phase before his last visit); thus, no subjects were censored. Three subjects who were not suppressed at the end of second treatment period were not included in the survival analysis. The median length of time from removal of the LNG implants to recovery was 12 wk in both T alone and T plus LNG groups, respectively. The confidence interval for the median was 10, 12 wk for the T alone group and 12, 12.3 wk for T plus LNG group.

Figure 2 shows the percentage of subjects who reached azoospermia or severe oligozoospermia after T or T plus LNG implant insertion. Comparing the two treatment groups when data from the two sites were combined, the percentage of subjects who reached a sperm concentration less than 1 × 10⁶/ml was significantly higher in the T plus LNG implants group than in the T alone group at wk 12 (P = 0.018), wk 15 (P < 0.001), wk 18 (P = 0.007), and wk 21 (P = 0.015). The number of subjects reaching azoospermia was also significantly higher in the combination group at wk 18 (P = 0.033) and 21 (P = 0.046). In the Chinese subjects in Nanjing, T plus LNG implants resulted in more suppression of spermatogenesis to azoospermia and severe oligozoospermia than T implants at wk 12 and 15 only, whereas in non-Chinese subjects in Los Angeles, the combination treatment was significantly more effective at wk 15 and 18. With time, as more subjects became severely oligozoospermic or azoospermic with T implants alone, the difference between the treatment groups gradually disappeared.

From Fig. 2 it can be noted that T implants alone caused more spermatogenesis suppression in Chinese subjects in Nanjing than in the non-Chinese subjects in Los Angeles. At wk 18, for example, 20%, 26.7%, and 40% of the subjects in Los Angeles compared with 65% (P = 0.016), 65% (P = 0.041), and 80% (P = 0.032) of the subjects in Nanjing achieved suppression of sperm concentration to reach the threshold of 0, less than 1, and less than 3 × 10⁶/ml, respectively. This ethnic difference in responsiveness to T implants alone persisted, but became less prominent with continued treatment. At wk 30, T implants induced azoospermia or less than 1 × 10⁶/ml in 58.8% of the Los Angeles subjects and in 95% of the subjects in Nanjing, respectively (P = 0.014).

Sperm motility and morphology followed the same pattern of suppression by the treatment regimen. The data are not shown.

Serum LNG levels

In all subjects, mean serum LNG levels rose after insertion of the implants to reach relatively steady-state levels of 920-1390 pmol/liter during the treatment period and then returned to baseline levels after implant removal. The mean serum LNG levels achieved in the Chinese subjects were between 1300 and 1700 pmol/liter. The LNG levels were significantly higher at all treatment time periods (P < 0.05) in the Chinese subjects when compared with the LNG levels in the Los Angeles subjects, which ranged from 780-1070 pmol/liter. There was an overall correlation between LNG levels and body size at baseline (r = −0.40; P < 0.02); however, this did not affect the differences between the centers.

Serum T and free T levels

Overall, mean serum total T levels were significantly elevated compared with baseline levels (P < 0.0001) during the treatment period in both treatment groups at both sites (Fig. 3). Serum total T levels rose to peak levels 3 wk after each set of implants was inserted and gradually returned to approximately the pretreatment levels by 15 wk after the first insertion in the T alone group, but became lower than baseline in the T plus LNG group. After the second set of implants, mean serum total T followed the same pattern as after the first insertion. Mean serum total T levels were higher in the T alone compared with the T plus LNG group at all time points after implant insertion (P < 0.0001).

Comparing the subjects from the two sites, mean baseline serum total T was slightly higher in the Chinese subjects (P = 0.021). After T implants in both treatment groups, serum T levels in the Chinese subjects remained significantly higher than those in Los Angeles subjects (P < 0.05). Moreover, as shown in Fig. 3, the peak serum T levels appeared to be significantly higher in the Chinese subjects after the second T implant (wk 3, 28.6 ± 1.9; wk 18, 35.0 ± 2.7 nmol/liter; P = 0.0001). However, the trough serum T levels after the first and second insertions remained the same.

Mean serum free T levels also rose after insertion of T implants in both groups of subjects at both sites (P < 0.0001; Fig. 4). In contrast to serum total T, serum free T levels were similar in both T implant alone and T plus LNG implant groups. The serum free T levels in the Chinese subjects were significantly higher than those in the Los Angeles subjects at baseline (P = 0.022) and throughout the treatment period (P < 0.05). Similar to serum total T levels, there was a time effect on serum free T in the Chinese subjects, where peak serum free T levels were higher after the second insertion of the T implants (peak serum T at wk 3, 0.40 ± 0.035; at wk 18, 0.56 ± 0.06 nmol/liter; P = 0.0001).

Serum SHBG concentrations

Baseline serum SHBG levels were higher in the Chinese men in Nanjing than in the non-Chinese subjects in Los Angeles (P = 0.012; Table 1). LNG implants caused a significant decrease in serum SHBG levels to 54.8% of baseline throughout the treatment period in the T plus LNG group (P < 0.0001; Fig. 5). The difference in serum SHBG between the groups receiving T plus LNG vs. T alone persisted for both sites throughout the treatment period.

Serum gonadotropins

Baseline serum LH and FSH levels were not significantly different between the two treatment groups or between the two sites (Table 1). T implants or T plus LNG implants markedly suppressed both serum LH and FSH levels (Fig. 6). There were no significant differences in the suppression of gonadotropins between the two sites. Longitudinal analyses showed that the suppression of both serum LH and FSH concentrations was significantly greater in the combination treatment with T and LNG than with T alone (P < 0.0001). The suppression of gonadotropins also varied with the time of treatment (P < 0.0001). In both treatment groups, serum FSH and LH levels were maximally suppressed 9 wk after each insertion of T implants. Thereafter, both serum LH and FSH rose to mean levels above 1.0 U/liter at wk 15 after each insertion in the T implant alone group (e.g. at wk 15, 10 of 18 subjects in Los Angeles and 12 of 20 subjects in Nanjing had both LH and FSH levels >1.0 IU/liter), whereas the mean gonadotropin levels remained suppressed to less than 0.5 U/liter in the T plus LNG group (e.g. at wk 15, only two of 19 subjects in Los Angeles and one of 21 subjects in Nanjing had both LH and FSH levels >1.0 IU/liter).

The study design included a period after the insertion of the implants when the subjects returned for blood sampling on d 4, 7, 14, and 21 to determine whether complete responders could be distinguished from the incomplete responders during this early period of spermatogenesis suppression (Fig. 7). Comparing the treatment groups, serum LH levels (P < 0.0001) were significantly more suppressed on d 7, 14, and 21, and serum FSH levels (P < 0.03) were significantly more suppressed on d 14 and 21 in the T plus LNG vs. T alone group (Fig. 7, upper panel). There was no difference in the early suppression of gonadotropins between the Chinese and the non-Chinese (Fig. 7, middle panel) men. If complete responders were classified as those subjects whose sperm concentration decreased to less than 1 × 10⁶/ml, then serum LH (P < 0.007 on d 4 and P < 0.005 on d 7, 14, and 21) and FSH levels (P = 0.059, 0.015, and 0.001 on d 7, 14, and 21, respectively) were significantly more suppressed in these men (Fig. 7, lower panel).

Safety monitoring parameters

Hemoglobin levels were significantly higher in the non-Chinese subjects in Los Angeles compared with the Chinese subjects at baseline and at all time points after treatment (P < 0.0001). Both hemoglobin (P < 0.01) and hematocrit (P < 0.05) increased significantly, without any differences between the two treatment groups (Fig. 8, upper panel). The maximum increases in hemoglobin and hematocrit occurred at 12 wk after the first set of T implants and 9 wk after the second set of implants. Three subjects had elevated hemoglobin above 168 g/liter (normal range, 138–168 g/liter) at wk 12, which became lower at wk 15. The hemoglobin levels in these three men and one additional subject rose again above 168 g/liter at wk 24, which resolved at the end of the recovery period. Despite the elevation of hemoglobin, the hematocrit in all these subjects was less than 52%. Serum total, high-density lipoprotein (HDL), and low-density lipoprotein (LDL) cholesterol and triglyceride levels showed no significant change with treatment (Fig. 8, middle panel). None of the other blood counts or serum clinical chemistry results showed any significant change with treatment. Serum PSA remained within the normal range in all subjects and showed a small, but nonsignificant, increase in the T implants only group. The baseline prostate volume was lower in the Chinese subjects (18.0 ± 0.6 ml; P < 0.0001) than in the non-Chinese subjects (21.7 ± 1.0 ml). The mean prostate volume did not significantly change with treatment (Fig. 8, lower panel).

Discussion

In this study, using implants of T and LNG, we confirmed the findings of previous studies that progestins such as LNG enhance the suppression of spermatogenesis by androgens alone. Our study also confirmed the usefulness of long-acting androgen and progestin implants to deliver steroids for male hormonal contraception. The results from our current study corroborated those of the study by Anderson et al. (20), which showed that a combination of four T implants (200 mg each, every 12 wk) and two etonogestrel implants (68 mg each) were effective in suppressing spermatogenesis to azoospermia in 75% of subjects and near azoospermia in 92% (<0.1 × 10⁶/ml). The modestly better suppression of spermatogenesis in this latter study compared with the present study was probably related to the use of the more potent progestin, etonogestrel. T implants in combination with depot medroxyprogesterone acetate also demonstrated similar successful suppression of spermatogenesis and contraceptive efficacy (4, 16). Despite the implants delivering serum T levels that were above baseline and stably elevated serum LNG levels, the suppression of spermatogenesis was relatively slow. Maximum suppression was reached between wk 21 and 24 of treatment. A similar slow suppression of spermatogenesis was observed in the T plus etonogestrel implants study, where maximum suppression occurred between wk 16 and 24 of treatment (20). These observations were in contrast with the more rapid suppression of sperm output by 8 wk when T implants were administered with depot medroxyprogesterone acetate (4, 16). The rapid suppression observed with injectable or oral progestins might be related to the high serum peaks of the steroids, which were not observed with LNG implants reported in previous studies (11). LNG implants were used with T patches or T enanthate injections (100 mg weekly) in studies reported by our group (11) and with T undecanoate (500 and 1000 mg every 8 wk) in Chinese subjects by Gui et al. (12). In both studies, optimal suppression of spermatogenesis to severe oligozoospermia was attained with injections of T resulting in serum T levels above baseline, showing the important role of androgens in spermatogenesis suppression when used in combination with a less potent progestin, such as LNG. When serum T levels declined in the study, serum gonadotropin levels increased, suggesting the possibility of escape from suppression even in the presence of a steady serum LNG concentration.

In Chinese men, the addition of LNG implants resulted in more effective suppression of spermatogenesis evident only during the early period of treatment. After this period, the proportion of Chinese men who became azoospermic or severely oligozoospermic increased progressively with time, such that by the end of the treatment period, most Chinese men (>90%) became azoospermic in both treatment groups. Unlike the Chinese subjects, the non-Chinese subjects in Los Angeles responded to T implants alone with a maximum suppression of spermatogenesis, resulting in 59% of the men reaching severe oligozoospermia at the end of the treatment period. Addition of LNG improved suppression to severe oligozoospermia to almost 90% at the end of treatment. This failed to reach statistical significance, probably because only 15 of 20 subjects completed the T only arm. Because we were interested in comparing the suppression rate at each time point, and some subjects may lose suppression during treatment, the survival method of analysis was not used for suppression rates. It should be noted that the statistical results were not corrected for multiple comparison; therefore, the chance of false positives may be increased. Our finding of differences in responsiveness to T alone administration between Chinese and non-Chinese men had been previously reported in larger studies (1, 2). The reason for this differential responsiveness to exogenous T administration is not clear and may be ascribed to differences in the metabolism of T (21–24), the responsiveness of the gonadotropins (25), or the apoptotic rate of the germ cells (26).

Similar to previous studies with T implants in healthy men, serum T and free T were elevated to levels above the baseline in treatment groups (4, 15, 16) at both centers. The apparent higher serum T levels in the T alone group vs. the T plus LNG group were the result of suppression of circulating SHBG to about 50% of baseline levels when LNG implants were added to T, as previously reported by our group (11). This finding was confirmed by the serum free T levels, which were not different between the treatment groups. Both the baseline serum T and free T levels were significantly higher in the Chinese men. The Chinese subjects had lower body weight and BMI, but they were healthy and not undernourished, whereas the subjects in Los Angeles were generally heavier. It is well known that higher body weight and BMI are inversely related to total serum T and free T (27–30). In a recent epidemiological study, dietary nutrients were not related to serum T or free T levels (31). Serum SHBG levels were also higher in the Chinese men compared with the subjects in Los Angeles, which concurred with a previous report (32). The elevated SHBG levels may be related to dietary fiber that the Chinese subjects might have consumed and might have partially accounted for the apparent higher baseline serum T in the Chinese. Consistent with previous studies, prostate volumes were lower in Chinese compared with non-Chinese subjects. Serum PSA levels may reflect prostate volume, but because serum PSA was measured by different assay methods at the two centers, the results could not be compared. Serum dihydrotestosterone (DHT) levels were not measured in this study. There is no consistent relationship between serum DHT and prostate volume, nor is the prostate the major source of serum DHT (33). Similar to our previous report as well as others (25), testicular volume appeared to be larger in the non-Chinese men. Serum gonadotropins and sperm concentration were not different between the subjects from the two centers. It is also not clear why serum LNG levels were higher in the Chinese, but the differences could not be explained by the lower body mass alone, suggesting that other factors, such as differences in the clearance rates of progestin in the Chinese subjects, may play a role.

As with other progestin plus androgen combinations, gonadotropin levels were markedly suppressed during treatment, whereas in the androgen only groups, gonadotropin levels appeared to rebound when serum T levels decreased. Nonsuppressed gonadotropins may lead to escape of spermatogenesis suppression. We also examined the suppression of serum gonadotropins during the first 3 wk of treatment to determine whether early measurements of gonadotropins could distinguish complete responders vs. incomplete responders to exogenous hormone treatment. We noted that there was no difference in the suppression of both gonadotropins between the Chinese and the non-Chinese subjects. Comparing subjects who were treated with LNG plus T implants and those with T implants alone, the subjects in the combination group had significantly more suppression of both gonadotropins from d 4–21. When we defined complete responders as those subjects whose sperm concentration was suppressed to below 1 × 10⁶/ml, both serum LH and FSH levels were more suppressed in the complete responders than in those who did not suppress to severe oligozoospermia. This observation suggests that measurements of gonadotropin levels may predict suppression of sperm production with continued treatment. The relation between the suppression of gonadotropin levels and sperm concentration had been previously reported (20, 34–36). McLachlan et al. (36), using a highly sensitive immunofluorometric assay, demonstrated that suppression of serum LH to very low levels was a good predictor of sperm concentration and the likelihood of suppression of spermatogenesis to below 1 × 10⁶/ml, whereas serum FSH was not an independent predictor. We did not perform pharmacogenetic assessment (phenotype/genotype correlations) among these ethnically different groups, although previous studies showed that the CAG repeat polymorphism of the androgen receptor could not clearly distinguish those that would suppress to near azoospermia after administration of hormonal agents in contraceptive clinical trials (37, 38). It would be of interest to discover the basis for enhanced suppression of spermatogenesis in the Chinese subjects, the higher levels of progestin in the Chinese, and the enhanced erythropoiesis in the U.S. group.

In hypogonadal men the estimated release rate of 200-mg implants was 1.3 mg/d (39), and four 200-mg implants should sustain serum T levels within the eugonadal range for more than 20 wk. In our study as well as that reported by Turner et al. in eugonadal men (4), T implants were unable to maintain serum T levels within the adult male range for more than 16 wk. The shorter duration of action of the T implants may be due to changes in T clearance in the eugonadal men and interaction between T and progestin. The insertion of T implants was associated with pain and signs of inflammation in about 35% of the subjects. Extrusion of one or two of the four implants occurred in 7.4% of subjects, which was similar to previous reports (40). Of note, the extrusions recurred with the second implant in four of the six subjects. Acne occurred in 10% of subjects, requiring medication discontinuation in one subject. Hematocrit increased to clinically significant levels in 5% of the subjects. Thus, T implants requiring a minor surgical procedure and lasting for 15–16 wk are not ideal for male contraception. Implants containing more potent androgens and lasting for longer periods of time are required to allow further development of androgen implants as a practical long-acting method of delivery. The LNG implants do not appear to be associated with these insertion problems. In contrast with oral LNG administration, we confirmed our previous findings (11) that implants of LNG did not result in significant weight gain or decreases in HDL cholesterol levels. This difference from oral therapy could be explained by the lower serum LNG levels achieved with the implants and the lack of the hepatic first-pass effect on serum lipids associated with oral administration.

In conclusion, our study showed that T alone or in combination with LNG delivered by implants resulted in marked suppression of spermatogenesis. Addition of LNG to T enhanced the suppression of gonadotropins and spermatogenesis early in the treatment phase. The additive effect of LNG to T was more pronounced in non-Chinese subjects than in Chinese subjects, in whom, given time, T alone effectively suppressed spermatogenesis to azoospermia in most subjects. The adverse events of the combination were mostly related to the use of T implants. Implants of androgen and progestogen combination may be developed into a practical method of male contraception provided that fewer implants can be used, serum hormone levels can be maintained at a stable physiological range, and insertions can be simplified with minimal implant extrusions. To achieve this, more potent androgens, such as 7α-methyl-19-nortestosterone, or selective androgen receptor modulators have to be developed to be used in combination with a more potent progestin, such as etonogestrel, or other selective progesterone receptor modulators.

Acknowledgments

We thank Robert I. McLachlan, M.D., Ph.D. (Monash University Medical Center, Melbourne Australia), for his help with the training of the investigators to insert the T implants; Dr. Y. Q. Gu (National Family Planning Research Institute, Beijing, China) for his advice; the staff of the Endocrine Research Laboratory and the General Clinical Research Center Core Laboratory at Harbor-University of California-Los Angeles Medical Center Biomedical Research Institute for the technical assistance with semen analyses and hormone measurements; the nurses of the General Clinical Research Center at Harbor-University of California-Los Angeles Medical Center for the performance of the study; and Sally Avancena, M.A., for her assistance with manuscript preparation.

This work was supported by the CONRAD Program (CSA-01-289), the Mellon-CONRAD Twinning Grants (MFG-00-53), and the Harbor-University of California-Los Angeles General Clinical Research Center (MO1-RR-00425).

This study was presented in part at the 85th Annual Meeting of The Endocrine Society, Philadelphia, PA, 2003.

Abbreviations:

 
• BMI,

  Body mass index;

 
• DHT,

  dihydrotestosterone;

 
• HDL,

  high-density lipoprotein;

 
• LDL,

  low-density lipoprotein;

 
• LNG,

  levonorgestrel;

 
• LOQ,

  limit of quantitation;

 
• PSA,

  prostate-specific antigen;

 
• T,

  testosterone.

References