Abstract： Objective To evaluate effects and safety of gonadotrophin-releasing hormone agonist (GnRH-a) combined with transdermal estradiol and medroxyprogesterone acetate in the treatment of endometriosis. Methods From January I st, 2007 to July 31 st, 2007, 28 endometriosis patients underwent laparnscopic or transabdominal surgery in Obstetrics and Gynecology Hospital affiliated to Fudan University were randomly divided into group A and group B. 14 patients in group A received 3.6 mg goserelin once every 4 weeks, 12 weeks in all 14 patients in group B received goserelin and added 1/2 piece of half-hydrate estradiol every week and 6 mg oral medroxyprogesterone acetate per day, 12 weeks in all. Serum estradiol (E2 ), follicle stimulating hormone(FSH), bone gla protein levels, visual analogue scale (VAS) of pain, bone mineral density of lumbar spine, vaginal exfoliate cell spurs and the form of Kupperman were compared in patients before and after treatment. Results (1 ) After treatment, the level of FSH and E2levels were (5.0 ± 2. 6 ) U/L and (29 ± 17 ) pmol/L in group A and (3.0 ± 1.5 ) U/L, and (87 ± 53 ) pmol/L in group B, which were significantly lower than those before treatment [FSH (17. 0 ± 12. 2) U/L, and E2 (184 ± 194) pmol/L in group A and FSH :(15.3±13.6)U/L and E2: (281±242) pmol/L in group B, P ＜ 0. 01]. On the seventh day after three-month GnRH-a treatment, it was observed that the level of E2 was higher and FSH was lower in group B than the level of E2 and FSH of group A (P ＜ 0. 01 ). (2 ) After treatment, the basal vaginal exfoliate cell proportion in group A [(66. 2 ± 29. 0) %] was significantly lower than that in group B [(11.8 ± 28. 0) %, P ＜ 0. 01] ; while patients in group A owned a lower proportion of the middle [(29. 1 ± 23.1 ) %], superficial layers [(4. 0 ± 5.5 ) %] and esinophilic cells [(2. 3 ± 2. 6)%]than patients group B [middle layer: (73. 0 ± 25.2)% ; superficial layer: (15. 2 ± 10. 9)% ; esinophilic cells: (10. 8 ± 7.9 ) % ; P ＜ 0. 01]. (3) Before the treatment, patients' VAS scores of total, pelvic pain, dysmenorrheal and dyspareunia were 7.43±3. 20,2. 35 ± 1.82, 4. 93 ± 1.98 and 0. 14±0. 53 in group A and were 7.71±2. 02, 2. 57 ± 1.60, 4. 86 ± 1.56 and 0. 29 ± 1.07 in group B; after treatment, the scores above were changed to 0. 14±0. 36,0. 07±0. 27,0. 07±0. 27and 0 in group A and 0. 36±0. 50, 0. 29±0. 47, 0. 07±0. 27 and 0 in group B, which were all significantly lower than those before treatment separately (P ＜0. 01 ). When menstruation recovered, the scores were 0. 21±0. 43, 0. 07±0. 27, 0. 14 ± 0. 36, and 0 in group A and 0. 50±0. 65, 0. 29±0. 47, 0. 21±0. 43 and 0 in group B, which were also significantly lower than those before treatment (P ＜ 0.01 ), however, no statistical difference was found between groups at any time spot(P ＞ 0. 05). (4) In group A, the bone density after treatment [(0. 96 ± 0. 06 ) g/cm2] was lower than that before treatment [(0. 99 ± 0. 06 ) g/cm2, P ＜ 0.01 )]. In group B, the index was (0. 98 ± 0. 09) g/cm2, which was lower than that before treatment [(0. 99 ± 0. 10 ) g/cm2, P = 0. 201]. No statistical difference was found between groups(P ＞ 0. 05 ). The bone loss rate were (- 2. 77 ± 1.97 ) % in group A and (- 0. 93 ± 2. 86 ) % in group B (P = 0. 058 ). Before treatment, the bone gla protein was (13±3) μg/L in group A and (13±6) μg/L in group B. After treatment, the bone gla protein levels was (17±6)μg/L in group A, which was higher than that before treatment (P ＜ 0. 01 ), the level was (16±6)μg/L in group B, which was higher than that before treatment, however showed no statistical difference(P =0. 053). No difference was found in bone gla protein before and after treatment between two groups (P＞0. 05). (5) The form of Kupperman after treatment were 15±7 in group A and 11±6 in group B, which did not show significant difference (P ＞ 0. 05 ). The incidence of flash and sweat were 93% (13/14)in group A, which was significantly higher than that 57% (8/14) in group B(P ＜0.01 ). Conclusion The add-back therapy that consists of an estradiol patch and oral medroxyprogesterone acetate is effective and safe treatment for endometriosis.