An open randomized comparative study of the effect of goserelin depot and danazol in the treatment of endometriosis

Transcription

1 FERTILITY AND STERILITY Copyright ~ 1992 The American Fertility Society Printed on acid-free paper in U.S.A. An open randomized comparative study of the effect of goserelin depot and danazol in the treatment of endometriosis Robert W. Shaw, M.D.t Zoladex Endometriosis Study Team:!: Academic Department of Obstetrics and Gynaecology, Royal Free Hospital School of Medicine, Royal Free Hospital, London, United Kingdom Objective: To compare the efficacy and safety of goserelin depot and danazol for endometriosis. Design: Open, randomized comparative trial. Setting: Multicenter European academic clinical institutions. Patients: A total of 307 patients with laparoscopically diagnosed endometriosis were randomized to goserelin (n = 204) or danazol (n = 103); 249 patients underwent second look laparoscopy (175 received goserelin and 74 danazol) and were analyzed for efficacy. Interventions: A 3.6-mg depot of goserelin monthly subcutaneously or oral danazol200 mg three times a day administered for 24 weeks. Main Outcome Measures: Efficacy assessments were based on changes in visible deposits at laparoscopy before and after treatment and subjective symptom scores at 4-week intervals during treatment and 8-week intervals after treatment for up to 24 weeks. Safety was assessed by adverse event reporting and clinical laboratory measures. Results: There were similar proportions of symptomatic (73 %) and asymptomatic (but infertile) (27%) and comparable distribution of different severity of endometriosis randomized to each treatment. Significantly fewer patients randomized to goserelin (6.4%) withdrew during treatment compared with 20.4% randomized to danazol (P < 0.05). There were significantly reduced visible deposits of endometriosis found post-treatment (P < ) within each group but no differences between the treatments. The mean total subjective symptoms scores remained significantly less than entry at 24 weeks post-treatment (P < 0.05). Hypoestrogenic side effects were more common in those receiving goserelin, particularly hot flushes, but anabolic/androgenic side effects of weight gain and muscle cramps were more common in those receiving danazol. Conclusions: The monthly administered 3.6-mg depot preparation of goserelin was highly effective at inducing resolution of endometriotic implants and relieving the symptoms of endometriosis with prevention of their return during 24 weeks follow-up in the majority of patients. However, results were not significantly different from those achieved with danazol600 mg/d. Fertil Steril 1992;58: Key Words: Endometriosis, gonadotropin-releasing hormone analogues, goserelin depot, danazol Endometriosis is a common condition in the reproductive years with a peak incidence between 30 and 45 years of age and associated with symptoms of cyclical pelvic pain and infertility. The symptoms can be severe enough in many patients to interfere significantly with normal life and justify treatment. The dependence of endometriotic tissue for its continued growth on ovarian steroid hormones, particularly estrogen (E) (1), has resulted in medical Received November 7, 1991; revised and accepted April 8, * Supported by ICI Pharmaceuticals pic, Macclesfield, United Kingdom. t Reprint requests and present address: Robert W. Shaw, M.D., Department of Obstetrics and Gynaecology, University of Wales College of Medicine, Heath Park, Cardiff CF4 4XN, United Kingdom. :j: The following clinical investigators and centers participated in the conduct and analysis ofthis study: Robert W. Shaw, M.D., (principal investigator), Rosemary L. Gardner, M.R.C.O.G., Royal Free Hospital, and Gillian L. Rose, M.D., Panos Maouris, M.D., Queen Charlotte's and Chelsea Hospital, London United Kingdom; Richard R. Fleming, B.Sc., Margaret J. Bonduelle, M.D., Shaw et ai. Goserelin versus danazol in endometriosis 265

2 treatments aimed at inducing ovarian suppression and inhibition of cyclic menstruation. To date, danazol (an isoxazol derivative of 17 a-ethinyl-testosterone [T]) with androgenic and anabolic properties has been the first choice by many gynecologists because of its proven efficacy but because of its induced side effects may not be tolerated well by patients. This has led to a search for therapeutic alternatives. Continued exposure of the pituitary gonadotropes to agonistic analogues of gonadotropin-releasing hormone (GnRH-a) results in desensitization or down regulation (2, 3) with resultant reduction in circulating serum gonadotropin levels and inhibition of ovarian steroidogenesis. These effects thus offered a potential alternative in the medical therapy of endometriosis. Initial pilot studies with GnRH-a administered as daily subcutaneous (SC) injections or multiple daily intranasal (IN) applications confirmed that they were effective at relieving symptomatology related to endometriosis and in inducing resolution of endometriotic implants (4-8). Sustained slow release depot preparations of GnRH -a are known to achieve more consistent and a greater degree of suppression of serum estradiol- 17{3 (E 2 ) than IN preparations (9). This study therefore aimed to assess the efficacy of the biodegradeable depot GnRH-a goserelin (D-Ser-[Bu t ]6- Aza GlylOLH-RH, Zoladex; ICI Pharmaceuticals pic, Macclesfield, United Kingdom) administered monthly SC in comparison with danazol (Danocrine, Sterling Winthrop, Albany, NY) orally. Ethical considerations did not allow the study to be conducted double-blind with placebo implants nor the Royal Infirmary, Glasgow, United Kingdom; Rune Rolland, M.D., Arne T. Franssens, M.D., Radboudziekenhuis, Nijmegan, Holland; Lars Hamberger, M.D., Bjorn Kallfelt, M.D., Sahlgrens Hospital, Goteborg, Sweden; Manjit Obhrai, M.D., Jag S. Samra, M.B., B.S., Birmingham Maternity Hospital, Birmingham, United Kingdom; Paul Devroey, M.D., Academic Hospital University of Brussels, Belgium; Colm O'Herlihy, M.D., Michael J. Brassil, M.D., National Maternity Hospital, Dublin, Ireland; Arne Skryten, M.D., Eli Smedvig, M.D., Sentralsjukehuset 1. Rogaland, Stavanger, Norway; Paul Lebech, M.D., Henning Djursing, M.D., Frederiksberg Hospital, Copenhagen, Denmark; Aksel Lange, M.D., Mogens Norgaard, M.D., KAS Glostrup, Denmark; Arne Berget, M.D., Jens B. Hertz, M.D., KAS Gentofte, Hellerup, Denmark; Nicole Madenelat, M.D., Henri Foulet, M.D., Hopital Bichet, Paris, France; Jean Dubuisson, M.D., Hopital Cochin, Paris, France; Andre Peters, M.D., Inge Smidt, M.D., Academic Hospital Leiden, Leiden, Holland; Frederick Petersson, M.D., Kvinnokliniken Lanssjukhuset, Halmstad, Sweden; Max Elstein, M.D., Catherine Vaughn-Williams, M.D., University Hospital, Manchester, United Kingdom; Michael C. White, M.D., and Elaine Turner, M.B., Ch.B., Royal Victoria Infirmary, Newcastle upon Tyne, United Kingdom. use of a placebo nontreatment arm in symptomatic patients. Patients MATERIALS AND METHODS Three hundred seven women with laparoscopically confirmed endometriosis gave informed consent to be studied. Ethical committee approval was obtained from all 18 centers involved. Patients were randomly allocated to 24 weeks of treatment with either goserelin depot (3.6 mg goserelin acetate in a lactide-glycolide copolymer rod, inserted into the SC tissue of the anterior abdominal wall every 28 days using a prefilled syringe applicator with a 16-gauge needle or danazol capsules 200 mg by mouth three times a day. Randomization was in the ratio two goserelin:one danazol with each center having its randomization list. After 8 weeks of treatment, the dose of goserelin could be increased to two depots (7.2 mg) every 28 days and the dose of danazol to 200 mg four times a day if the patient failed to achieve amenorrhea (vaginal spotting was allowed). If an adverse event occurred on the higher dose, the dose of goserelin could be reduced to one depot (3.6 mg) every 28 days, and danazol could also be reduced to 400 mg/ dafter 8 weeks of treatment if side effects occurred provided that the patient was amenorrheic. Treatment was begun between days 1 to 7 of a menstrual cycle. No hormonal agents (including oral contraceptives) were allowed during the 8 weeks before starting study therapy; neither were GnRH-a or danazol allowed during the 24 weeks before starting study therapy. The patients were otherwise in good health, not taking anticoagulants, and had been advised to use barrier contraception for the duration of the study until resumption of menses post-treatment. The age range of the subjects was from 18 to 40 years, with normal menstrual cycle length of between 21 to 42 days. The degree of endometriosis was assessed according to the Revised American Fertility Society (AFS) classification (10) and the Additive Diameter of Implants (11) in which a point score of 1 is given for every square millimeter of endometriotic implant visualized regardless of depth of invasion. No more than 12 weeks were allowed to elapse between the diagnostic laparoscopy and entry of the patient into the study. At the initiallaparoscopy, no procedures other than biopsy of a deposit were undertaken, and video recordings were not routinely performed in all centers. The follow-up laparoscopy had to be per- 266 Shaw et al. Goserelin versus danazol in endometriosis Fertility and Sterility

3 formed within 5 weeks of the end of treatment and usually by the same clinician who had performed the original. Patients were clinically examined before treatment and at 4-week intervals during the 24-week treatment period. During the 24 weeks after treatment follow-up, patients were examined at 8- week intervals. Patients were questioned at each visit for pelvic symptoms (pelvic pain, dyspareunia, dysmenorrhea) and examined for physical findings (pelvic tenderness and induration). Each was graded as: none (0); mild (1); moderate (2); or severe (3). Scores from pelvic pain, dyspareunia, and dysmen- 0rrhea were combined to give a total pelvic symptoms score, and scores for all the symptoms and physical findings were combined to give a total subjective score. The presence and severity of certain symptoms considered to be recognized pharmacological effects of GnRH -a or danazol were elicited by direct questioning. Direct questions were asked about vaginal dryness, depression, acne, oily hair or skin, hirsutism' ankle oedema, mood swings, headache, and voice changes. Each was graded: none (0); mild (1); moderate (2); or severe (3). Changes in libido and breast size were graded as compared with entry: increase (1); no change (2); decrease (3). Hot flushes, if present, were assessed by their number and severity, whereas sweating was recorded as present or absent. Serum luteinizing hormone (LH), follicle-stimulating hormone (FSH), and E2 were measured at each visit, and hematologic and biochemical profiles were measured before treatment and at 4, 12, 24, 32, and 48 weeks. Those patients desiring pregnancy were monitored for a further 24 weeks after the standard 24 weeks after treatment symptomatic follow-up. Statistical Consideration Two hundred fifty patients, randomized in a ratio of two goserelin:one danazol, were calculated to be a large enough number to demonstrate equivalence between treatments at the 5% level of statistical significance with a 90% power, assuming that the objective response rate of each treatment is 70% and equivalence between treatments is defined as a difference of <20% in response rate. Goserelin and danazol were compared irrespective of dose. All randomized patients were included in the overall analysis of objective response. The proportion of responders was analyzed using logistic regression with the pretreatment total revised AFS (10) score as covariate. Of the 307 who were randomized to receive treatment, 286 satisfied the selection criteria precisely. Patients are included in efficacy analyses ac- cording to treatment as randomized and in the safety analysis according to treatment received. Demographic Data RESULTS There was equal distribution of symptomatic and asymptomatic (but infertile) subjects between the two treatment groups and no differences in mean age or mean pretreatment weight. Two hundred four women were randomized to receive goserelin and 103 to receive danazol. More patients randomized to goserelin had recurrent disease with 30.4% having received previous hormone therapy or surgery compared with 16.5% of those randomized to danazol. Patients were not stratified at randomization in relation to initial revised AFS (10), but there was approximately equal percentage distribution of stages I to IV between the two treatment groups. Two hundred one of the 204 patients randomized to goserelin received the drug for a mean of 23.8 weeks, 194 (96.5%) remained on 3.6 mg, and only 7 (3.5%) had the increased dosage of 7.2 mg. Of the 103 patients who received danazol for a mean of 22.5 weeks, 72 (69.9%) remained on 600 mg, 20 (19.4%) had the dose increased to 800 mg, but 6 reduced back again to 600 mg, and in 11 (10.7%) the dose was reduced to 400 mg as per the protocol. Withdrawals A total of 81 (39.7%) patients randomized to goserelin were withdrawn from the study, 13 (6.4%) before completion of the treatment period. This compared with 54 (52.4%) of patients randomized to danazol, of whom 21 (20.4%) withdrew before completing the treatment period. Reasons for withdrawal were classified into four categories: lack of effect, adverse findings, pregnancy, and administrative (patient unwilling to continue, additional therapy given, leaving area). During treatment a greater proportion of patients withdrew on danazol (9.7%) than on goserelin therapy (1.0%) because of adverse findings or adverse events. On goserelin these were mood changes (1) and voice change (1), and on danazol hirsutes and voice change (3), jaundice (1), rash (1), nausea and weight gain (2), allergic reaction (1), muscle cramps (1). After treatment there was an equivalent number of withdrawals between treatment groups and no differences noted between incidence of pregnancy or other causes of withdrawa~. Objective Response-Changes in Laparoscopic Findings Two hundred forty nine of the 307 patients randomized were eligible for analysis of change in total Shaw et ai. Goserelin versus danazol in endometriosis 267

4 o Implants Score Adhesions Score Stage I Stage II Stage III Stage IV GOSERELIN.} Pre-treatment tie }post-treatment DANAZOL ~ Figure 1 Changes in total revised American Fertility Society (R-AFS) implant and adhesion component scores separately and differing revised AFS stages before and after treatment in patients randomized to goserelin (n = 175) or danazol (n = 74). revised AFS (10) score, (total adhesions revised AFS score, total implants revised AFS score), and total Additive Diameter of Implants (11) scores. Only those patients who satisfied the selection criteria and had both before and after treatment laparoscopies were eligible. Figures 1 and 2 compare the mean total revised AFS and total Additive Diameter of Implants scores before and at the end of treatment with either goserelin or danazol and within each of the stages I to IV. The reduction in mean (±SD) total revised AFS (10) score in the goserelin treated patients was from 16.4 ± 20.1 to 6.7 ± 12.4 (-59.1 %) and in the danazol patients from 14.5 ± 13.7 to 7.0 ± 11.0 (-51.7%). Corresponding changes in mean (±SD) total Additive Diameter of Implants scores were from 37.0 ± 35.2 to 6.9 ± 12.7 (-81.4%) for goserelin patients and from 38.8 ± 34.5% to 10.3 ± 14.7 (-73.5%) in the danazol treated group and were comparable in all stages of endometriosis. The differences were statistically significant within each group (P = <0.0001), but there was no difference between the treatments. 268 Shaw et a1. Goserelin versus danazol in endometriosis U sing the definition of a response as a decrease of 50% or more in total revised AFS (10) score, 127 of 174 patients (73%) treated with goserelin and 52 of 74 (70.3%) treated with danazol responded. (Comparison between treatments P = 0.42, 95% confidence interval for odds ratio 0.69, 2.43). Eightyone patients (46.3%) receiving goserelin and 25 (33.8%) treated with danazol showed a clinical "cure" with a total Additive Diameter of Implants score of zero at post-treatment laparoscopy. Treatment failures, classified as a <50% decrease from pretreatment revised AFS score, no change, or an increase of <25% occurred in 43 of 175 (24.5%) on goserelin and 17 of 74 (21.6%) on danazol. Objective progression (increase in revised AFS score > 25% baseline) was found in 2.0% of goserelintreated and 4.9% danazol-treated patients. Subjective Scores Patients were considered symptomatic at entry to the study if they had a pelvic symptoms score of 3 or more. Two hundred twenty-three patients were symptomatic at entry with equal percentage distri- Fertility and Sterility

5 o TotalADI Stage I ADI Stage II ADI Stage III ADI Stage IV ADI GOSERELIN DANAZOL : }Pn>t~atm~t Figure 2 Changes in Additive Diameter of Implant (AD!) scores, total, and ADI scores in revised AFS stages I to IV. (mean ± SD) before and after treatment in patients randomized to goserelin (n = 175) or danazol (n = 74). bution between treatments. Both treatments were associated with a significant reduction in mean total subjective scores (pelvic symptoms + physical findings) but with no difference between the treatments. A significant reduction in subjective scores had been achieved by 8 weeks on treatment, and the score continued to fall until the 16 to 20-week assessment. Although there was a slight increase in total subjective scores during the follow-up post-treatment, primarily because of the contribution from dysmen- 0rrhea that returned to some degree with the onset of menstruation, the mean scores remained significantly below those pretreatment even at 24 weeks off treatment (P < 0.05) (Fig. 3). Endocrine Changes, Menstruation, and Pregnancies Goserelin therapy suppressed mean serum LH and FSH values (P < 0.05), whereas no significant change was seen in danazol-treated patients. A fall in mean serum E2 to values below the early follicular range «110 pmolfl) was induced by goserelin and was significantly lower than changes induced by danazol (P < 0.05). One hundred fifty-two of the 189 goserelin patients assessed (80.4%) became amenorrheic by the 8th week on treatment compared with 40 of the 88 danazol patients assessed (45.4%). Menstruation returned 47.4 days (range 23 to 142) after last effective goserelin day (calculated as 28 days after last goserelin depot) and 34.0 days (range 2 to 64) after the last danazol capsule. Of the goserelin-treated patients, 113 were noted to be infertile at entry into the study, and 33 (29.2%) successfully conceived during the 12 months after cessation of treatment. The comparable figures for danazol-treated patients were 54 infertile with 13 (24.1%) achieving pregnancy. In addition, 14 goserelin patients and 2 danazol patients were noted to have been fertile at entry to the study and at the end of treatment desired pregnancy. Ten goserelin patients (71.4%) and 1 danazol patient (50%) conceived; 4 goserelin- and 4 danazol-treated patients who did not desire pregnancy on entry also con- Shaw et al. Goserelin versus danazol in endometriosis 269

6 9-0 DANAZOL GOSERELIN Pre WEEKS End of treatment Figure 3 Changes in mean (±SD) of total subjective scores (pelvic symptoms and physical findings) during treatment and follow-up in symptomatic patients treated with goserelin (n = 158) or danazol (n = 65). ceived. Of the 47 pregnant patients who had received goserelin, 34 (72.3%) had live births; ofthe 18 danazol-treated patients, 14 (77.8%) had live births. There is no statistically significant difference between the groups in terms of percentage of pregnancies nor mean time to conception. Elicited Pharmacological Effects and Adverse Events The most common adverse side effects reported in the study are summarized in Table 1. Hot flushes were the most common side effect reported in 98% Table 1 The Most Common Adverse Side Effects Reported or Elicited During the Study Goserelin Danazol % % Pain 7 12 Headache 5 3 Nausea 5 12 Breast pain 5 1 Weight gain 1 27 Muscle cramps 1 12 Hot flushes and sweats Acne Oily hair and skin Reduction in libido of goserelin-treated and 58% of danazol-treated patients. Although in some patients these were severe and frequent, these alone were not the reason to cease treatment. Androgenic/anabolic side effects were more common in patients receiving danazol, whereas E deficiency symptoms were more common in patients treated with goserelin. DISCUSSION This study reports the results of a multicenter open randomized trial of a monthly depot preparation of the GnRH agonist goserelin compared with oral danazol. Although not stratified for revised AFS score (10) at entry, there were comparable percentage distributions of stages I to IV, symptomatic (77%) and asymptomatic (but infertile 23%) ofpatients between the two treatment groups, and no other differences in baseline demographic data were found. Both treatments were shown to be equally effective at inducing resolution of endometriotic deposits and reducing subjective symptom scores. Differences between the treatments resulted from side effects and adverse effects experienced by the two essentially different endocrine environments induced by the treatments. Although the revised AFS scoring system has been used in other comparative trials (12), the scoring 270 Shaw et al. Goserelin versus danazol in endometriosis Fertility and Sterility

7 system is arbitrary and subjective (13). The use of the Additive Diameter of Implants to assess response to therapies in this study adds an additional dimension and demonstrated a comparable reduction in total Additive Diameter of Implants scores in all revised AFS stages treated. The percentage of goserelin-treated patients with zero total scores at second-look laparoscopy was 46.3% compared with 33.8% of danazol treated subjects (P = not significant). The induction of a sustained hypoestrogenic environment has been thought to be a critical factor in the treatment of endometriosis. The mean circulating levels of E2 were significantly lower in the goserelin than in the danazol-treated patients. Sex hormone-binding globulin (SHBG) levels on goserelin are not significantly altered, whereas they are decreased while on danazol. Estradiol is more loosely bound to SHBG than T; thus free E2 is expected to be altered less than free T by reduction in SHBG (14), and indeed percent free E2 in danazol-treated patients has been shown to increase from 1.01% to 1.64% (15). The greater reduction in total E2 was not reflected in any significantly increased resolution of endometriotic deposits on goserelin. This may indicate that although a relative degree of hypoestrogenism is essential to allow an endocrine suppressive effect on endometrial deposits, further reduction below this threshold level has no further additional benefits. Because the adverse effects of goserelin treatment are a consequence of the hypoestrogenemia, it may be possible to counteract these with some E supplementation without detriment to the effect on the response of endometriotic tissue. Such an approach would also help prevent the changes in calcium metabolism and decrease in trabecular bone mineral content known to occur after 24 weeks of therapy with goserelin (16) and other GnRH-a (17, 18). Although these are unlikely to be of any significance from a single 6-month course of treatment, they might be relevant if prolonged or repeat courses of therapy are administered. Danazol has both androgenic and anabolic properties and can induce associated side effects of weight gain, edema, myalgia, acne, hirsutism, and hepatocellular damage (19) that are dose related. In a previous comparative trial of danazol and the GnRH-a nafarelin, a dose of 800 mg/d danazol was used (12). In this trial, the standard dose was 600 mg/ d, but the reduction in dosage does not appear to have resulted in any deleterious significant difference in symptomatic response or altered resolution of endometriotic deposits com- pared with the higher 800 mg dose. Studies with even lower doses of danazol are justified because the incidence of androgenic/anabolic side effects might also be reduced because these were the reason for discontinuation of therapy in 9.5% during therapy in this study. The data from this comparative trial have demonstrated that the depot GnRH agonist goserelin administered once a month is at least as effective as danazol in the treatment of endometriosis and is well tolerated by patients. The consequences of the hypoestrogenemia induced by goserelin account for its side effects and although 24 weeks of treatment is unlikely to have any long lasting or irreversible effects, longer term therapy or repeat treatments of a disorder that is commonly recurrent may necessitate the use of combined analogue plus bone protective therapies. REFERENCES 1. Schweppe KW, Wynn RM, Beller FK. Ultrastructural comparison of endometriotic implants and ectopic endometrium. Am J Obstet Gynecol 1984;148: Sandow J. Clinical applications of LHRH and its analogues. Clin Endocrinol (Oxf) 1983;18: Bergqvist C, Nillius SJ, Wide L. Intranasal LHRH agonist treatment for inhibition of ovulation in women: clinical aspects. Clin Endocrinol (Oxf) 1982;17: Meldrum DR, Chang RJ, Lu J, Vale W, Rivier J, Judd HL. 'Medical oophorectomy' using a long-acting GnRH agonist: a possible new approach to the treatment of endometriosis. J Clin Endocrinol Metab 1982;54: Shaw RW, Fraser HM, Boyle H. Intranasal treatment with luteinising hormone-releasing hormone agonist in women with endometriosis. Br Med J 1983;287: Lemay A, Maheux R, Faure N, Jean C, Fazekas ATA. Reversible hypogonadism induced by a luteinising hormonereleasing hormone (LH-RH) agonist (buserelin) as a new therapeutic approach for endometriosis. Fertil Steril 1984;41: Schriock E, Monroe SE, Henzl M, Jaffe RB. Treatment of endometriosis with a potent agonist of gonadotrophin releasing hormone (nafarelin). Fertil Steril 1985;44: Steingold KA, Cedars M, Lu JKH, Randle RN, Judd HL, Meldrum DR. Treatment of endometriosis with a long acting gonadotropin-releasing hormone agonist. Obstet Gynecol 1987;69: Shaw RW. LHRH analogues in the treatment of endometriosis-comparative results with other treatments. Clin Obstet Gynaecol 1988;2: American Fertility Society. Revised American Fertility Society classification of endometriosis: Fertil Steril 1985;43: Doberl A, Bergqvist A, Jeppson S, Koskimies AI, Ronnberg L, Segerbrand E, et al. Regression of endometriosis following shorter treatment with, or lower dose of danazol. Acta Obstet Gynaecol Scand SuppI1984;123: Henzl MR, Cordon SL, Moghissi K, Buttram VC Jr, Bergqvist C, Jacobson J. Administration of nasal nafarelin as compared Shaw et al. Goserelin versus danazol in endometriosis 271

8 with oral danazol for endometriosis. N Engl J Med 1988;318: Dmowski WP. Visual assessment of peritoneal implants for staging endometriosis: do number and cumulative size of lesions reflect the severity of a systemic disease? Fertil Steril 1987;47: Dunn JF, Nisula BC, Rodbard D. Transport of steroid hormones: binding of 21 endogenous steroids to both testosterone binding globulin and corticosteroid-binding globulin in human plasma. J Clin Endocrinol Metab 1981;53: Dowsett M, Mansfield MD, Griggs DJ, Jeffcoate SL. Validation and use of centrifugal ultrafiltration dialysis in the measurement of percent free oestradiol in serum. J Steroid Biochem 1984;21: Stevenson JC, Lees B, Gardner R, Shaw RW. A comparison of the skeletal effects of goserelin and danazol in premenopausal women with endometriosis. Horm Res 1989;32(Suppl): Cann CE, Henzl MR, Burry K, Andreyko J, Hanson F, Adamson GD, et al. Reversible bone loss is produced by the GnRH agonist nafarelin. In: Cohn DV, Martin TJ, Meunier PJ, editors. Calcium regulation and bone metabolism: basic and clinical aspects. Amsterdam: Elsevier 1987; Matta WH, Shaw RW, Hesp RR, Katz D. Hypogonadism induced by luteinizing hormone-releasing hormone agonist analogues: effects on bone density in premenopausal women. Br Med J 1987;294: Barbieri RL, Evans S, Kistner RW. Danazol in the treatment of endometriosis: analysis of 100 cases with a 4-year followup. Fertil Steril1982;37: Shaw et al. Goserelin versus danazol in endometriosis Fertility and Sterility