- Email: [email protected]
Effect of gonadotropin-releasing hormone agonist on the bone mineral density of patients with endometriosis*†
~"
"8
~OC
'
.:}/rJ;·F~JL!~.i"COlogy~.ndocrinology·
:~Jl;zjiil;H\~;i\(
.··b·;' ... . . . . ... . .
. ....
Vol. 62, No.2, August 1994
FERTILITY AND STERILITY Copyright
©
Printed on acid-free paper in U. S. A.
1994 The American Fertility Society
Effect of gonadotropin-releasing hormone agonist on the bone mineral density of patients with endometriosis*t Tsuguo Uemura, M.D.:j:§ Jun Mohri, M.D. II Hisabumi Osada, M.D.~
Naoyuki Suzuki, M.D.II Nobuyuki Katagiri, M.D.** Hiroshi Minaguchi, M.D.:j:
Yokohama City University, Saiseikai- Yokohama Nanbu Hospital, Yokohama Municipal Citizens' Hospital, and Yokohama-Minamikyosai Hospital, Yokohama, Japan
Objective: To examine changes in bone mineral density (BMD) of patients treated with GnRH agonist (GnRH-a) by dual-energy X-ray absorptiometry and to understand factors related to bone loss. Design: Prospective controlled trial examining BMD during and after GnRH -a therapy every 24 weeks for 18 months in patients with endometriosis compared with nontreated controls. Setting: Outpatients clinic at a university hospital and its affiliated outpatient clinic. Patients: Twenty-two patients with endometriosis as GnRH -a-treated group, 12 healthy women with normal menstrual cycle, and 7 patients with mild endometriosis as control group. Interventions: Patients were treated with a GnRH-a (buserelin acetate) at 900 tLgjd by nasal spray for 24 weeks. Results: The significance of differences in change-rates at all measured points in both groups was assessed by analysis of variance. The interaction between treatment and period was significant, and only week 24 of the GnRH-a-treated group was significantly lower compared with baseline. The reduction rate of BMD was high in patients 33 years of age or younger compared with those who were 34 years of age or older. According to a multiple-regression model, the most important factor related to bone loss was the post-treatment serum levels of E 2 • Conclusion: At the end of treatment, BMD was significantly lower than that of the control group, and the reduction rate was 3.4%. A factor related to bone loss was degree of ovarian suppression. Fertil Steril 1994;62:246-50 Key Words: Endometriosis, bone mineral density, GnRH agonist
Recently, a GnRH agonist (GnRH-a), which inhibits the synthesis of gonadotropin and estrogen
Received May 4,1993; revised and accepted March 31,1994.
* Supported by Hoechst Japan Ltd., Tokyo, Japan.
t Presented at the 3rd International Symposium on GnRH Analogues in Cancer and Human Reproduction, Geneva, Switzerland, February 25 to 28, 1993. :j: Department of Obstetrics and Gynecology, School of Medicine, Yokohama City University. § Reprint requests: Tsuguo Uemura, M.D., Department ofObstetrics and Gynecology, School of Medicine, Yokohama City University, 3-9 Fukuura, Kanazawa-ku, Yokohama 236, Japan (FAX: 81-45-701-3536). II Department of Obstetrics and Gynecology, Saiseikai-Yokohama Nanbu Hospital. 11 Department of Obstetrics and Gynecology, Yokohama Municipal Citizens' Hospital. 246
Uemura et al.
GnRH-a on bone density
by chronic administration, has been used mainly as conservative therapy for endometriosis. Because GnRH-a inhibits ovarian function and induces a hypoestrogenic state, a possibility of a reduction of bone mineral density (BMD) was suggested. Although the mechanism of action of estrogen on bone has not yet been thoroughly clarified, estrogen acts on osteoblast cells directly and indirectly to inhibit bone metabolism (1). Because estrogen also inhibits bone absorption by parathyroid hormone (2), the low estrogen condition is considered to promote bone absorption. In this study, BMD was measured by dual energy X-ray absorptiometry, and changes in BMD during ** Department of Obstetrics and Gynecology, YokohamaMinamikyosai Hospital. Fertility and Sterility
a 24-week treatment with GnRH -a were examined in Japanese women and compared with those of a control group with normal menstrual cycle of the same age as the GnRH-a-treated group, because BMD may change without GnRH -a treatment over 24 weeks in women aged 20 to 40 years.
Table 1
The Changes of BMD: ANOVA
Factor
Sum of squares
d.f.
Mean square F value Pvalue
Treatment (a) Subject Period (b) (a) X (b) Residual Total
39.57607534 308.79192753 56.66868767 31.1 7229324 216.08500478 647.75837512
1 24 3 3 72 103
39.57607534 12.86633031 18.88956253 10.39076441 3.00118062
3.08 4.29 6.29 3.46
0.0922 0.0001 0.0007 0.0260
MATERIALS AND METHODS
Twenty-two patients aged 24 to 43 years who suffered from mild and sever endometriosis were examined at the Department of Obstetrics and Gynecology, Yokohama City University and its affiliated outpatient clinic. They were treated with a GnRHa (buserelin acetate, Spracure; Hoechst Japan Co., Ltd., Tokyo, Japan) at 900 ILg/d by nasal spray for 24 weeks (buserelin group). Twelve healthy women with normal menstrual cycle and seven patients with mild endometriosis were regarded as nontreated controls (control group). Bone mineral density of lumbar spine (L2-L4) was measured by dual-energy x-ray absorptiometry (DX-26; Norland Co., Ltd., Fort Atkinson, WI). The coefficient of variation was found to be 0.32% for spine phantom and 2.18% for healthy volunteers. The BMD was measured four times, namely, before treatment (week 0), at the end of treatment (week 24), at 24 weeks after cessation of treatment (week 48), and at 48 weeks after cessation of treatment (week 72). At the early follicular phase of pretreatment or treatment cycle and thereafter, every 12 weeks during and after treatment regardless of the day of menstrual cycle, E 2 , estrone (E 1 ), and LH in serum were determined by RIA as hormones related to bone metabolism. Alkaline phosphatase calcium in serum, and calcium and creatinine i~ fasting urine were also measured as parameters of bone metabolism. For statistical analysis, the balance of the distribution of subject characteristics between the two groups was examined using the Wilcoxon rank sum test. The change ofBMD was evaluated using analysis of variance (ANOVA), Scheffe's multiple comparison. To examine the factor related to bone loss multiple regression was used. Change of hormone~ during the treatment period was evaluated by using Signed rank test. RESULTS Patients' Background
Bone mineral density in women with endometriosis was not lower than that of women without Vol. 62, No.2, August 1994
endometriosis, and the patients of endometriosis without treatment and women with normal menstrual cycle were combined into one and regarded as a control group. The 22 patients of the buserelin group ranged from 24 to 43 years of age (average 31.6 ± 1.1 years, mean ± SE), whereas the women of the control group ranged from 22 to 49 years of age (average 30.3 ± 1.6 years). There was no significant difference between the buserelin group and the control group in age. The body mass index in the buserelin group (21.39 ± 0.61) was not different from the control group (20.43 ± 0.51). The body weight and the body height were not different between the two groups. Bone Mineral Density of Lumbar Spine (L2-L4)
In the buserelin group, the BMD of the lumbar spine was 1.062 ± 0.016 g/cm2 (n = 22; the control group 1.085 ± 0.025 g/cm 2 , n = 21) at week 0, 1.034 ± 0.019 g/cm2 (n = 20; control group 1.089 ± 0.029 g/cm2 , n = 19) at week 24, 1.037 ± 0.015 g/cm 2 (n = 19; control group 1.084 ± 0.033 g/cm2 , n = 15) at week 48, and 1.059 ± 0.019 g/cm2 (n = 12; control group 1.085 ± 0.032 g/cm2 , n = 17) at week 72. There is some variability due to intersubject variability and different numbers of subjects because some subjects missed one or two of the four BMD measurements. Therefore, to evaluate the changes of BMD, subjects for whom there was complete BMD data at four points (week 0, week 24, week 48, week 72) were selected. Individual changes ofBMD were described as a percentage of baseline. According to the results of ANOVA, neither treatment nor period had a significant effect. But the interaction between treatment and period was significant (P = 0.029) as shown in Table 1. For further analysis using Scheffe's multiple comparison, only week 24 of the buserelin group was significantly lower (96.605 ± 0.812, P < 0.05) compared with baseline (Fig. 1). The correlation of the change-rates of BMD at week 24 to age were investigated in both groups as Uemura et al.
GnRH-a on bone density
247
(%)
o
• Buserelin group(n=12) o Control group (n=14)
105
Mean±SE *0<0.05
:2
m OJ
.!: OJ
(standardized estimate -3.410, t-test, P = 0.0018). No other factors were related to bone loss.
100
Ul
DISCUSSION
jg 95
'0
T 4B
72
Weeks
Figure 1 ment.
Changes in BMD before and after GnRH-a treat-
shown in Figure 2. The change-rate ofBMD did not differ by age in the control group, but the reduction-rate of BMD in the GnRH-a-treated group was high in patients younger than 33 years of age compared with those who were 34 years of age or older. Changes in Hormone and Bone Metabolism Parameters
In the GnRH-a-treated group, serum concentrations of LH, El> and E2 at week 24 showed significant decreases (P < 0.01) from those at week 0 and returned to the pretreatment levels at week 48, though these hormones did not change significantly in the control group. In the GnRH-a-treated group, urinary calcium/creatinine increased significantly (P < 0.05) at week 24. Serum alkaline phosphatase increased and serum calcium slightly decreased at week 24, but these changes were not significant. No significant change could be seen in the control group. Multiple-Regression Analysis on Bone Loss
To know the predicted factor for bone loss, using a multiple-regression model, we chose the following factors: smoking, drinking alcohol, drinking milk, body mass index, age, presence of endometriosis, and the serum level of E2 at week 24. As the habit of drinking milk is not common in Japan, it is an important factor in dietary calcium intake for Japanese people. The presence of endometriosis was also factored into the regression equation because there was an association between the change-rate of BMD and the presence of endometriosis. The most important factor related to bone loss was the post-treatment serum levels of E 2, which correlated negatively with the degree of bone loss 248
Uemura et al.
GnRH-a on bone density
The effects of menopause and the hypoestrogenic condition induced by ovariectomy on bone metabolism have been studied recently. Medicines for endometriosis, such as danazol and GnRH -a, induce a low estrogen condition, and attention was paid to the effects of these medicines on bone metabolism. Bone mineral density during treatment with a GnRH -a has been studied and reported in Europe and America (3-5). But BMD differs among races in that BMD of the spine and proximal femur measured by the same machine for dual-photon absorptiometry was lower in Japanese women (6) than American white women (7). In this study BMD during GnRH-a therapy was studied in Japanese women. To measure BMD, methods such as single-photon absorptiometry, quantitative computed tomography, and dual photon absorptiometry are used. A highly accurate method is dual-energy x-ray absorptiometry, which was developed recently. In most reports, changes in BMD during GnRH-a treatment were examined by quantitative computed tomography and dual-photon absorptiometry, but we applied the highly accurate dual-energy x-ray absorptiometry· to observe the change of BMD. The present study is a prospective controlled trial to examine bone loss and its recovery over 1.5 years in GnRH-a-treated patients with endometriosis compared with controls belonging to the same age group as the GnRH-a-treated group to avoid changes in BMD associated with aging. There is a report that BMD in women with endo-
Buserelin group
Control group
1%)
1%)
110
110
Y=91l.53+0.193X r=[].367
0
55 105
55105
OJ
"
OJ
.!:
3im 100
3im 100
0 0000 0
.c
95
0
0
?Ii. 90
"
.!:
.c
0
Y=103.1-0.1003X r=O.252
0
20
25
30
0
0
00 0
35 40 Age
""
95
?Ii. 45
50
90 20
25
30
35 40 Age
45
50
Figure 2 ~elationships between percentage of baseline BMD and age in the GnRH-a-treated group (n = 20) and the control group (n = 19). Fertility and Sterility
metriosis is significantly lower than in women without endometriosis (3), but we were not able to obtain any significant difference between them (1.072 ± 0.015 gjcm2 vs. 1.095 ± 0.015 gjcm 2 ) as reported by Dawood et al. (4). Therefore, we combined the patients of endometriosis without treatment and women with normal menstrual cycle into one and regarded it as a control group. In this study, BMD decreased significantly to -3.4% for lumbar spine, which is different from the report by Damewood et al. (8), which found no significant change of BMD measured by dual-photon absorptiometry, but similar to the report by Devogelaer et al. (5), which noted a -2.1 % decrease of lumbar spine BMD measured by dual-photon absorptiometry after the intranasal administration of buserelin acetate 900 Jlgjd. Buserelin acetate at a higher dosage than ours (1,200 Jlgjd) was reported to decrease bone mineral content (BMC) even more: -7.4% for thracolumbar vertebrae (4) and -5.9 % for lumbar vertebrae (9) measured by quantitative computed tomography. In the case of nafarelin at 400 Jlgjd, a 2% to 5.9% decrease was observed in BMD of the lumbar spine measured by quantitative computed tomography or dual-photon absorptiometry (10-12). In addition to this, significant increases in the calcium and creatinine ratio in urine were observed in our study, and these increases give biochemical support to our finding that GnRH-a results in a decrease in BMD. With regard to the prognosis after treatment, Matta et al. (9) reported recovery of BMD to the pretreatment level by 6 months after treatment according to restoration of ovarian function. Furthermore, in the reports (10-12) on the administration of nafarelin, the restoration of BMD by 6 months after treatment was observed unanimously. On the other hand, Dawood et al. (4) reported that BMC was still significantly lower (-4.2%) than the pretreatment level at 6 months after treatment, and Nencioni et al. (13) reported that no bone mass restoration took place and a less significant but discernible trend toward further bone loss was apparent in the BMD values measured 6 months after the end of therapy. Bone mineral density in the control group varies during an extended period of observation over 1.5 years, and we should assess changes in BMD precisely by carrying out a comparative study with a control group of the same age group. In this study, BMD gradually increased for 6 months after GnRH -a treatment and was restored to a level of insignificant difference from that of the control Vol. 62, No.2, August 1994
group, whereas the GnRH-a-treated group still had a significant decrease of -2.0 % as compared with the pretreatment level. At the 48th week after treatment, a decrease of -2.2% was observed, but no significant difference from the control group was detected. We investigated the relationship between age and decrease in BMD and found that bone loss was high in the group younger than 33 years of age; however, recovery may be rapid, and we assume that the possibility of their influences remaining for a long period is unlikely. It is important to know whether the rate of bone loss is predicted before and during GnRH -a treatment. Scharla et al. (14) demonstrated that women with elevated 1,25-dehydroxyvitamin D3(OH)2D3 concentrations are at a high risk of developing accelerated bone loss after the onset of estrogen deficiency. In this study we used multiple-regression analysis to observe the relationship between a decrease in BMD and factors such as body mass index, smoking, drinking alcohol, drinking milk, and serum concentrations of E 2. As a result, the value of E2 at week 24 showed the most significant negative correlation to bone loss, and no significant correlation was seen in the other factors.
REFERENCES 1. Takano-Yamamoto T, Rodan GA. Direct effects of 17(3-
2.
3.
4.
5. 6.
7.
8.
estradiol on trabecular bone in ovariectomized rat. Proc Nat! Acad Sci USA 1990;87:2172-6. Orimo H, Fujita T, Yoshikawa M. Increased sensitivity of bone to parathyroid hormone in ovariectomized rats. Endocrinology 1972;90:760-3. Comite F, Delman M, Hutchinson-Williams K, DeCherney AH, Jensen P. Reduced bone mass in reproductive-aged women with endometriosis. J Clin Endocrinol Metab 1989; 69:837-42. Dawood MY, Lewis V, Ramos J. Cortical and trabecular bone mineral content in women with endometriosis: effect of gonadotropin-releasing hormone agonist and danazol. Fertil Steril1989;52:21-6. Devogelaer J-P, De Deuxchaisnes CN, Donnes J, Thomas K. LHRH analogues and bone loss. Lancet 1987; 1:1498, Norimatsu H, Mori S, Uesato T, Yoshikawa T, Katsuyama N. Bone mineral density of the spine and proximal femur in normal and osteoporotic subjects in Japan. Bone Miner 1989;5:213-22. Mazess RB, Barden HS, Ettinger M, Johnston C, DawsonHughes B, Powell M, et al. Spine and femur density using dual photon absorptiometry in US white women. Bone Miner 1987;2:211-9. Damewood MD, Schlaff WD, Hesla JS, Rock JA. Interval bone mineral density with long-term gonadotropin-releasing
Uemura et aI.
GnRH-a on bone density
249
9.
10. 11.
12.
hormone agonist suppression. Fertil Steril 1989;52: 596-9. Matta WH, Shaw RW, Hesp R, Evans R. Reversible trabecular bone density loss following induced hypo-estrogenism with the GnRH analogue buserelin in premenopausal women. Clin Endocrinol (Oxf) 1988;29:45-51. Jacobson JB. Effects of nafarelin on bone density. Am J Obstet Gynecol 1990; 162:591-2. Johansen JS, Riis BJ, Hassager C, Moen M, Jacobson J, Christiansen C. The effect of a gonadotropin releasing hormone agonist analog (nafarelin) on bone metabolism. J Clin Endocrinol Metab 1988;67:701-6. Whitehouse RW, Adams JE, Bancroft K, Vaughan-Williams
250
Uemura et al.
GnRH-a on bone density
CA, Elstein M. The effects of nafarelin and danazol on vertebral trabecular bone mass in patients with endometriosis. Clin Endocrinol (Oxf) 1990; 33:365-73. 13. Nencioni T, Barbieri-Carones PM, Ortolani S, Trevisan C, Polvani F. Gonadotropin releasing hormone agonist therapy and its effect on bone mass. Gynecol Endocrinol 1991;5:4956. 14. Scharla SH, Minne HW, Waibel-Treber S, Schaible A, Lempert UG, Wuester C, et al. Bone mass reduction after estrogen deprivation by long-acting gonadotropin-releasing hormone agonists and its relation to pretreatment serum concentrations of 1,25-dehydroxyvitamin Da. J Clin Endocrinol Metab 1990; 70:1055-61.
Fertility and Sterility
"8
~OC
'
.:}/rJ;·F~JL!~.i"COlogy~.ndocrinology·
:~Jl;zjiil;H\~;i\(
.··b·;' ... . . . . ... . .
. ....
Vol. 62, No.2, August 1994
FERTILITY AND STERILITY Copyright
©
Printed on acid-free paper in U. S. A.
1994 The American Fertility Society
Effect of gonadotropin-releasing hormone agonist on the bone mineral density of patients with endometriosis*t Tsuguo Uemura, M.D.:j:§ Jun Mohri, M.D. II Hisabumi Osada, M.D.~
Naoyuki Suzuki, M.D.II Nobuyuki Katagiri, M.D.** Hiroshi Minaguchi, M.D.:j:
Yokohama City University, Saiseikai- Yokohama Nanbu Hospital, Yokohama Municipal Citizens' Hospital, and Yokohama-Minamikyosai Hospital, Yokohama, Japan
Objective: To examine changes in bone mineral density (BMD) of patients treated with GnRH agonist (GnRH-a) by dual-energy X-ray absorptiometry and to understand factors related to bone loss. Design: Prospective controlled trial examining BMD during and after GnRH -a therapy every 24 weeks for 18 months in patients with endometriosis compared with nontreated controls. Setting: Outpatients clinic at a university hospital and its affiliated outpatient clinic. Patients: Twenty-two patients with endometriosis as GnRH -a-treated group, 12 healthy women with normal menstrual cycle, and 7 patients with mild endometriosis as control group. Interventions: Patients were treated with a GnRH-a (buserelin acetate) at 900 tLgjd by nasal spray for 24 weeks. Results: The significance of differences in change-rates at all measured points in both groups was assessed by analysis of variance. The interaction between treatment and period was significant, and only week 24 of the GnRH-a-treated group was significantly lower compared with baseline. The reduction rate of BMD was high in patients 33 years of age or younger compared with those who were 34 years of age or older. According to a multiple-regression model, the most important factor related to bone loss was the post-treatment serum levels of E 2 • Conclusion: At the end of treatment, BMD was significantly lower than that of the control group, and the reduction rate was 3.4%. A factor related to bone loss was degree of ovarian suppression. Fertil Steril 1994;62:246-50 Key Words: Endometriosis, bone mineral density, GnRH agonist
Recently, a GnRH agonist (GnRH-a), which inhibits the synthesis of gonadotropin and estrogen
Received May 4,1993; revised and accepted March 31,1994.
* Supported by Hoechst Japan Ltd., Tokyo, Japan.
t Presented at the 3rd International Symposium on GnRH Analogues in Cancer and Human Reproduction, Geneva, Switzerland, February 25 to 28, 1993. :j: Department of Obstetrics and Gynecology, School of Medicine, Yokohama City University. § Reprint requests: Tsuguo Uemura, M.D., Department ofObstetrics and Gynecology, School of Medicine, Yokohama City University, 3-9 Fukuura, Kanazawa-ku, Yokohama 236, Japan (FAX: 81-45-701-3536). II Department of Obstetrics and Gynecology, Saiseikai-Yokohama Nanbu Hospital. 11 Department of Obstetrics and Gynecology, Yokohama Municipal Citizens' Hospital. 246
Uemura et al.
GnRH-a on bone density
by chronic administration, has been used mainly as conservative therapy for endometriosis. Because GnRH-a inhibits ovarian function and induces a hypoestrogenic state, a possibility of a reduction of bone mineral density (BMD) was suggested. Although the mechanism of action of estrogen on bone has not yet been thoroughly clarified, estrogen acts on osteoblast cells directly and indirectly to inhibit bone metabolism (1). Because estrogen also inhibits bone absorption by parathyroid hormone (2), the low estrogen condition is considered to promote bone absorption. In this study, BMD was measured by dual energy X-ray absorptiometry, and changes in BMD during ** Department of Obstetrics and Gynecology, YokohamaMinamikyosai Hospital. Fertility and Sterility
a 24-week treatment with GnRH -a were examined in Japanese women and compared with those of a control group with normal menstrual cycle of the same age as the GnRH-a-treated group, because BMD may change without GnRH -a treatment over 24 weeks in women aged 20 to 40 years.
Table 1
The Changes of BMD: ANOVA
Factor
Sum of squares
d.f.
Mean square F value Pvalue
Treatment (a) Subject Period (b) (a) X (b) Residual Total
39.57607534 308.79192753 56.66868767 31.1 7229324 216.08500478 647.75837512
1 24 3 3 72 103
39.57607534 12.86633031 18.88956253 10.39076441 3.00118062
3.08 4.29 6.29 3.46
0.0922 0.0001 0.0007 0.0260
MATERIALS AND METHODS
Twenty-two patients aged 24 to 43 years who suffered from mild and sever endometriosis were examined at the Department of Obstetrics and Gynecology, Yokohama City University and its affiliated outpatient clinic. They were treated with a GnRHa (buserelin acetate, Spracure; Hoechst Japan Co., Ltd., Tokyo, Japan) at 900 ILg/d by nasal spray for 24 weeks (buserelin group). Twelve healthy women with normal menstrual cycle and seven patients with mild endometriosis were regarded as nontreated controls (control group). Bone mineral density of lumbar spine (L2-L4) was measured by dual-energy x-ray absorptiometry (DX-26; Norland Co., Ltd., Fort Atkinson, WI). The coefficient of variation was found to be 0.32% for spine phantom and 2.18% for healthy volunteers. The BMD was measured four times, namely, before treatment (week 0), at the end of treatment (week 24), at 24 weeks after cessation of treatment (week 48), and at 48 weeks after cessation of treatment (week 72). At the early follicular phase of pretreatment or treatment cycle and thereafter, every 12 weeks during and after treatment regardless of the day of menstrual cycle, E 2 , estrone (E 1 ), and LH in serum were determined by RIA as hormones related to bone metabolism. Alkaline phosphatase calcium in serum, and calcium and creatinine i~ fasting urine were also measured as parameters of bone metabolism. For statistical analysis, the balance of the distribution of subject characteristics between the two groups was examined using the Wilcoxon rank sum test. The change ofBMD was evaluated using analysis of variance (ANOVA), Scheffe's multiple comparison. To examine the factor related to bone loss multiple regression was used. Change of hormone~ during the treatment period was evaluated by using Signed rank test. RESULTS Patients' Background
Bone mineral density in women with endometriosis was not lower than that of women without Vol. 62, No.2, August 1994
endometriosis, and the patients of endometriosis without treatment and women with normal menstrual cycle were combined into one and regarded as a control group. The 22 patients of the buserelin group ranged from 24 to 43 years of age (average 31.6 ± 1.1 years, mean ± SE), whereas the women of the control group ranged from 22 to 49 years of age (average 30.3 ± 1.6 years). There was no significant difference between the buserelin group and the control group in age. The body mass index in the buserelin group (21.39 ± 0.61) was not different from the control group (20.43 ± 0.51). The body weight and the body height were not different between the two groups. Bone Mineral Density of Lumbar Spine (L2-L4)
In the buserelin group, the BMD of the lumbar spine was 1.062 ± 0.016 g/cm2 (n = 22; the control group 1.085 ± 0.025 g/cm 2 , n = 21) at week 0, 1.034 ± 0.019 g/cm2 (n = 20; control group 1.089 ± 0.029 g/cm2 , n = 19) at week 24, 1.037 ± 0.015 g/cm 2 (n = 19; control group 1.084 ± 0.033 g/cm2 , n = 15) at week 48, and 1.059 ± 0.019 g/cm2 (n = 12; control group 1.085 ± 0.032 g/cm2 , n = 17) at week 72. There is some variability due to intersubject variability and different numbers of subjects because some subjects missed one or two of the four BMD measurements. Therefore, to evaluate the changes of BMD, subjects for whom there was complete BMD data at four points (week 0, week 24, week 48, week 72) were selected. Individual changes ofBMD were described as a percentage of baseline. According to the results of ANOVA, neither treatment nor period had a significant effect. But the interaction between treatment and period was significant (P = 0.029) as shown in Table 1. For further analysis using Scheffe's multiple comparison, only week 24 of the buserelin group was significantly lower (96.605 ± 0.812, P < 0.05) compared with baseline (Fig. 1). The correlation of the change-rates of BMD at week 24 to age were investigated in both groups as Uemura et al.
GnRH-a on bone density
247
(%)
o
• Buserelin group(n=12) o Control group (n=14)
105
Mean±SE *0<0.05
:2
m OJ
.!: OJ
(standardized estimate -3.410, t-test, P = 0.0018). No other factors were related to bone loss.
100
Ul
DISCUSSION
jg 95
'0
T 4B
72
Weeks
Figure 1 ment.
Changes in BMD before and after GnRH-a treat-
shown in Figure 2. The change-rate ofBMD did not differ by age in the control group, but the reduction-rate of BMD in the GnRH-a-treated group was high in patients younger than 33 years of age compared with those who were 34 years of age or older. Changes in Hormone and Bone Metabolism Parameters
In the GnRH-a-treated group, serum concentrations of LH, El> and E2 at week 24 showed significant decreases (P < 0.01) from those at week 0 and returned to the pretreatment levels at week 48, though these hormones did not change significantly in the control group. In the GnRH-a-treated group, urinary calcium/creatinine increased significantly (P < 0.05) at week 24. Serum alkaline phosphatase increased and serum calcium slightly decreased at week 24, but these changes were not significant. No significant change could be seen in the control group. Multiple-Regression Analysis on Bone Loss
To know the predicted factor for bone loss, using a multiple-regression model, we chose the following factors: smoking, drinking alcohol, drinking milk, body mass index, age, presence of endometriosis, and the serum level of E2 at week 24. As the habit of drinking milk is not common in Japan, it is an important factor in dietary calcium intake for Japanese people. The presence of endometriosis was also factored into the regression equation because there was an association between the change-rate of BMD and the presence of endometriosis. The most important factor related to bone loss was the post-treatment serum levels of E 2, which correlated negatively with the degree of bone loss 248
Uemura et al.
GnRH-a on bone density
The effects of menopause and the hypoestrogenic condition induced by ovariectomy on bone metabolism have been studied recently. Medicines for endometriosis, such as danazol and GnRH -a, induce a low estrogen condition, and attention was paid to the effects of these medicines on bone metabolism. Bone mineral density during treatment with a GnRH -a has been studied and reported in Europe and America (3-5). But BMD differs among races in that BMD of the spine and proximal femur measured by the same machine for dual-photon absorptiometry was lower in Japanese women (6) than American white women (7). In this study BMD during GnRH-a therapy was studied in Japanese women. To measure BMD, methods such as single-photon absorptiometry, quantitative computed tomography, and dual photon absorptiometry are used. A highly accurate method is dual-energy x-ray absorptiometry, which was developed recently. In most reports, changes in BMD during GnRH-a treatment were examined by quantitative computed tomography and dual-photon absorptiometry, but we applied the highly accurate dual-energy x-ray absorptiometry· to observe the change of BMD. The present study is a prospective controlled trial to examine bone loss and its recovery over 1.5 years in GnRH-a-treated patients with endometriosis compared with controls belonging to the same age group as the GnRH-a-treated group to avoid changes in BMD associated with aging. There is a report that BMD in women with endo-
Buserelin group
Control group
1%)
1%)
110
110
Y=91l.53+0.193X r=[].367
0
55 105
55105
OJ
"
OJ
.!:
3im 100
3im 100
0 0000 0
.c
95
0
0
?Ii. 90
"
.!:
.c
0
Y=103.1-0.1003X r=O.252
0
20
25
30
0
0
00 0
35 40 Age
""
95
?Ii. 45
50
90 20
25
30
35 40 Age
45
50
Figure 2 ~elationships between percentage of baseline BMD and age in the GnRH-a-treated group (n = 20) and the control group (n = 19). Fertility and Sterility
metriosis is significantly lower than in women without endometriosis (3), but we were not able to obtain any significant difference between them (1.072 ± 0.015 gjcm2 vs. 1.095 ± 0.015 gjcm 2 ) as reported by Dawood et al. (4). Therefore, we combined the patients of endometriosis without treatment and women with normal menstrual cycle into one and regarded it as a control group. In this study, BMD decreased significantly to -3.4% for lumbar spine, which is different from the report by Damewood et al. (8), which found no significant change of BMD measured by dual-photon absorptiometry, but similar to the report by Devogelaer et al. (5), which noted a -2.1 % decrease of lumbar spine BMD measured by dual-photon absorptiometry after the intranasal administration of buserelin acetate 900 Jlgjd. Buserelin acetate at a higher dosage than ours (1,200 Jlgjd) was reported to decrease bone mineral content (BMC) even more: -7.4% for thracolumbar vertebrae (4) and -5.9 % for lumbar vertebrae (9) measured by quantitative computed tomography. In the case of nafarelin at 400 Jlgjd, a 2% to 5.9% decrease was observed in BMD of the lumbar spine measured by quantitative computed tomography or dual-photon absorptiometry (10-12). In addition to this, significant increases in the calcium and creatinine ratio in urine were observed in our study, and these increases give biochemical support to our finding that GnRH-a results in a decrease in BMD. With regard to the prognosis after treatment, Matta et al. (9) reported recovery of BMD to the pretreatment level by 6 months after treatment according to restoration of ovarian function. Furthermore, in the reports (10-12) on the administration of nafarelin, the restoration of BMD by 6 months after treatment was observed unanimously. On the other hand, Dawood et al. (4) reported that BMC was still significantly lower (-4.2%) than the pretreatment level at 6 months after treatment, and Nencioni et al. (13) reported that no bone mass restoration took place and a less significant but discernible trend toward further bone loss was apparent in the BMD values measured 6 months after the end of therapy. Bone mineral density in the control group varies during an extended period of observation over 1.5 years, and we should assess changes in BMD precisely by carrying out a comparative study with a control group of the same age group. In this study, BMD gradually increased for 6 months after GnRH -a treatment and was restored to a level of insignificant difference from that of the control Vol. 62, No.2, August 1994
group, whereas the GnRH-a-treated group still had a significant decrease of -2.0 % as compared with the pretreatment level. At the 48th week after treatment, a decrease of -2.2% was observed, but no significant difference from the control group was detected. We investigated the relationship between age and decrease in BMD and found that bone loss was high in the group younger than 33 years of age; however, recovery may be rapid, and we assume that the possibility of their influences remaining for a long period is unlikely. It is important to know whether the rate of bone loss is predicted before and during GnRH -a treatment. Scharla et al. (14) demonstrated that women with elevated 1,25-dehydroxyvitamin D3(OH)2D3 concentrations are at a high risk of developing accelerated bone loss after the onset of estrogen deficiency. In this study we used multiple-regression analysis to observe the relationship between a decrease in BMD and factors such as body mass index, smoking, drinking alcohol, drinking milk, and serum concentrations of E 2. As a result, the value of E2 at week 24 showed the most significant negative correlation to bone loss, and no significant correlation was seen in the other factors.
REFERENCES 1. Takano-Yamamoto T, Rodan GA. Direct effects of 17(3-
2.
3.
4.
5. 6.
7.
8.
estradiol on trabecular bone in ovariectomized rat. Proc Nat! Acad Sci USA 1990;87:2172-6. Orimo H, Fujita T, Yoshikawa M. Increased sensitivity of bone to parathyroid hormone in ovariectomized rats. Endocrinology 1972;90:760-3. Comite F, Delman M, Hutchinson-Williams K, DeCherney AH, Jensen P. Reduced bone mass in reproductive-aged women with endometriosis. J Clin Endocrinol Metab 1989; 69:837-42. Dawood MY, Lewis V, Ramos J. Cortical and trabecular bone mineral content in women with endometriosis: effect of gonadotropin-releasing hormone agonist and danazol. Fertil Steril1989;52:21-6. Devogelaer J-P, De Deuxchaisnes CN, Donnes J, Thomas K. LHRH analogues and bone loss. Lancet 1987; 1:1498, Norimatsu H, Mori S, Uesato T, Yoshikawa T, Katsuyama N. Bone mineral density of the spine and proximal femur in normal and osteoporotic subjects in Japan. Bone Miner 1989;5:213-22. Mazess RB, Barden HS, Ettinger M, Johnston C, DawsonHughes B, Powell M, et al. Spine and femur density using dual photon absorptiometry in US white women. Bone Miner 1987;2:211-9. Damewood MD, Schlaff WD, Hesla JS, Rock JA. Interval bone mineral density with long-term gonadotropin-releasing
Uemura et aI.
GnRH-a on bone density
249
9.
10. 11.
12.
hormone agonist suppression. Fertil Steril 1989;52: 596-9. Matta WH, Shaw RW, Hesp R, Evans R. Reversible trabecular bone density loss following induced hypo-estrogenism with the GnRH analogue buserelin in premenopausal women. Clin Endocrinol (Oxf) 1988;29:45-51. Jacobson JB. Effects of nafarelin on bone density. Am J Obstet Gynecol 1990; 162:591-2. Johansen JS, Riis BJ, Hassager C, Moen M, Jacobson J, Christiansen C. The effect of a gonadotropin releasing hormone agonist analog (nafarelin) on bone metabolism. J Clin Endocrinol Metab 1988;67:701-6. Whitehouse RW, Adams JE, Bancroft K, Vaughan-Williams
250
Uemura et al.
GnRH-a on bone density
CA, Elstein M. The effects of nafarelin and danazol on vertebral trabecular bone mass in patients with endometriosis. Clin Endocrinol (Oxf) 1990; 33:365-73. 13. Nencioni T, Barbieri-Carones PM, Ortolani S, Trevisan C, Polvani F. Gonadotropin releasing hormone agonist therapy and its effect on bone mass. Gynecol Endocrinol 1991;5:4956. 14. Scharla SH, Minne HW, Waibel-Treber S, Schaible A, Lempert UG, Wuester C, et al. Bone mass reduction after estrogen deprivation by long-acting gonadotropin-releasing hormone agonists and its relation to pretreatment serum concentrations of 1,25-dehydroxyvitamin Da. J Clin Endocrinol Metab 1990; 70:1055-61.
Fertility and Sterility