1 July 1992

Volume 117

Number 1

Annals of Internal Medicine Treatment of Postmenopausal Osteoporosis with Transdermal Estrogen Edward G. Lufkin, MD; Heinz W. Wahner, MD; William M. O'Fallon, PhD; Stephen F. Hodgson, MD; Mark A. Kotowicz, MBBS, Ann W. Lane, MPH; Howard L. Judd, MD; Robert H. Caplan, MD; and B. Lawrence Riggs, MD

• Objective: To evaluate the tolerance and effectiveness of transdermal estrogen for women with established postmenopausal osteoporosis and vertebral fractures. • Design: Double-blind, randomized, placebo-controlled clinical trial lasting 1 year. • Setting: Referral-based outpatient clinic. • Patients: Seventy-five postmenopausal women, 47 to 75 years of age, with one or more vertebral fractures due to osteoporosis. • Interventions: Thirty-nine women received dermal patches delivering 0.1 mg of 17p-estradiol for days 1 to 21 and oral medroxyprogesterone acetate for days 11 to 21 of a 28-day cycle. Another 39 women received placebo. • Measurements: Bone turnover assessed by biochemical markers and iliac bone histomorphometry; bone loss assessed by serial measurement of bone density; and vertebral fracture rate. • Results: Compared with the placebo group, the median annual percentage change in bone mineral density in the estrogen group reflected increased or steady-state bone mineral density at the lumbar spine (5.3 compared with 0.2; P = 0.007), femoral trochanter (7.6 compared with 2.1; P = 0.03), and midradius (1.0 compared with - 2.6, P < 0.001) but showed no significant difference at the femoral neck (2.6 compared with 1.4; p = 0.17). Estrogen treatment uniformly decreased bone turnover as assessed by several methods including serum osteocalcin concentration (median change, - 0.35 compared with 0.02 nmol/L; P < 0.001). Histomorphometric evaluation of iliac biopsy samples confirmed the effect of estrogen on bone formation rate per bone volume (median change, - 12.9 compared with - 6.2% per year; P = 0.004). Also, 8 new fractures occurred in 7 women in the estrogen group, whereas 20 occurred in 12 women in the placebo group, yielding a lower vertebral fracture rate in the estrogen group (relative risk, 0.39; 95% CI, 0.16 to 0.95). • Conclusions: Transdermal estradiol treatment is effective in postmenopausal women with established osteoporosis.

Annals of Internal Medicine. 1992;117:1-9. For current author affiliations and addresses, see end of text.

JC/Strogen deficiency is an important contributing cause of osteoporosis. When begun at menopause, estrogen replacement therapy decreases the occurrence of fractures due to osteoporosis by at least half (1-3). Some believe, however, that estrogen therapy is not as effective after substantial bone loss and fractures have occurred (4), and interest has focused on the use of new nonestrogenic drugs to treat women with established postmenopausal osteoporosis (5). In the past, most estrogen therapy has been administered orally. The high concentration of estrogen in the portal circulation after oral administration, however, leads to increased hepatic production of estrogendependent proteins and bile changes that reportedly account for several of the complications of estrogen therapy including cholelithiasis, thrombophlebitis, and pulmonary embolism (6). Transdermal administration presumably can prevent these complications and has the theoretic advantage of continuous rather than bolus absorption into the circulation. Oral estrogen treatment does, however, have beneficial lipid effects because of altered liver metabolism. The increased levels of high-density lipoprotein (HDL) cholesterol and the decreased levels of total and low-density lipoprotein (LDL) cholesterol associated with oral administration are not seen with transdermal estrogen. Transdermal administration of estrogen to perimenopausal women has been shown to prevent postmenopausal bone loss (7-9), but no data exist regarding its effectiveness in the treatment of established osteoporosis. We therefore conducted a 1-year prospective, randomized, doubleblind, placebo-controlled trial of transdermal estradiol therapy in postmenopausal women with vertebral fractures due to established osteoporosis. In making this comprehensive evaluation, we assessed all available indicators of efficacy including bone density, bone biochemical markers, bone histomorphometric values, and fracture occurrence. We also examined the possibility that several baseline covariates modulate the therapeutic response. © 1992 American College of Physicians

Downloaded From: http://annals.org/ by a University of California San Diego User on 12/29/2016

1

Table 1. Clinical Characteristics of the Women in the Estrogen and Placebo Groups at Baseline Characteristic

Age, v Height, m Relative weight, actual weight/ ideal weight Time since menopause, v Pregnancies, n Previous hysterectomy, n Previous estrogen replacement therapy, n Nonsmokers, n Nondrinkers, n Caffeinated beverages consumed, cups/day Decaffeinated coffee consumed, cups/day Previous vertebral fractures, n Bone mineral density Lumbar spine, g/cm2 Femoral neck, g/cm2 Femoral trochanter, g/cm2 Midradius, g/cm

Estrogen (n = 36) Median (Range)f Patients (Percent)

Placebo (n = 39) Patients Median (Range)t (Percent)

P Value*

65.5 (54.6 to 72.1) 1.60(1.51 to 1.68)

64.1 (55.1 to 70.4) 1.60(1.53 to 1.67)

> 0.2 > 0.2$

1.27(1.02 to 1.71) 16.6 (5.7 to 27.6) 3 (1.7 to 9)

1.19 (1.05 to 1.45) 14.0 (5.0 to 25.0) 3 (1 to 6) 17 (47)

14 (36)

0.11$ > 0.2$ > 0.2 > 0.2

10 (28) 28 (82) 17 (47)

13 (33) 28 (76)§ 24 (62)

> 0.2 > 0.2 > 0.2

3 (0 to 8)

2 (0 to 5)

0 (0 to 3) 4 (1 to 9.3)

1 (0 to 3) 4 (2 to 9)

0.79 0.64 0.54 0.72

(0.65 (0.50 (0.39 (0.61

to to to to

0.77 0.65 0.52 0.75

0.91) 0.79) 0.64) 0.91)

(0.65 (0.50 (0.36 (0.58

0.14 > 0.2 > 0.2 to to to to

1.03) 0.78) 0.68) 0.89)

> > > >

0.2 0.2$ 0.2$ 0.2$

* For difference between groups, by rank-sum test for continuous variables or chi-square for categoric variables unless otherwise specified. t Shown as 10th and 90th percentiles. X From /-test. § Information not available for two women.

Methods

E x p e r i m e n t a l Design

Patients

Of the 75 women enrolled, 36 were randomly assigned to treatment with estrogen and 39 to placebo. Each woman's calcium intake was determined by a registered dietitian based on interviews and a review of a 7-day diet diary. Women whose calcium intake was less than 800 mg/d were instructed to maintain a diet providing this amount. All women received adequate calories to maintain their weight. The women in the estrogen group were treated with 20-cm2 transdermal patches (Estraderm, CIBA Pharmaceuticals, Edison, New Jersey) delivering 0.1 mg of estradiol for days 1 to 21 and with medroxyprogesterone acetate (10 mg/d orally) for days 11 to 21 of a 28-day cycle, regardless of previous hysterectomy. This estrogen dose was chosen for its known ability to produce midfollicular phase levels of serum estradiol and estrone (26). This dose of medroxyprogesterone acetate was chosen for its ability to prevent endometrial hyperplasia and thus to prevent endometrial carcinoma, which can occur with unopposed estrogen (6). The placebo group received a matching regimen of placebo, identical in appearance to the study drugs. Each patient was evaluated four times (at baseline, 1 month, 6 months, and 12 months) during the year of study.

All patients were studied at the Mayo Clinic, including 14 who were referred from the Gunderson Clinic, La Crosse, Wisconsin. The 75 women with established type I (postmenopausal) osteoporosis who were enrolled in our study were fully ambulatory, postmenopausal, white women 47 to 75 years of age who had documented osteoporosis but no evidence of an associated disease or a history of use of any drug known to cause osteoporosis or to affect calcium levels. The criteria for diagnosis of osteoporosis were a bone mineral density at the lumbar spine and proximal femur below the 10th percentile of normal premenopausal women and one or more vertebral fractures defined as a decrease in vertebral height of more than 15% (10). At the time of recruitment, 38 women were receiving estrogen, calcium supplements (> 500 mg/d), or vitamin D (> 400 IU/d). None of the women had ever received sodium fluoride or bisphosphonate. These treatments were discontinued for 3 months before the start of the study in women receiving only calcium supplements and for 6 months in women receiving estrogen, vitamin D, or both. Of 23 women reporting any previous estrogen replacement therapy (Table 1), 17 had discontinued use 4 or more years before the study, and only 3 were still receiving estrogen 6 months before the study. Of the 75 women in the study, 3 (1 in the estrogen group and 2 in the placebo group) refused therapy and discontinued treatment before information was collected at 6 months. An additional five women dropped out of the study before completing the examination at 1 year (two in the estrogen group and one in the placebo group because of skin reactions to the patches, one in the placebo group because of lung cancer, and one in the placebo group because of failure to keep appointments). Because of an error, one woman was randomized twice, first to the estrogen group and then, after 1 month of treatment, to the placebo group. In statistical analyses, she was considered to be a member of the estrogen group. All women gave written informed consent. The study was reviewed and approved by the Mayo Institutional Review Board. 2

L a b o r a t o r y Studies At baseline and at 1 year, participants had a general medical examination that included physical examination, radiographic evaluation of the thorax and of the thoracic and lumbar spine, complete blood cell count, chemistry group analysis, urinalysis, mammographic examination, and endometrial (Vabra) biopsy. Bone mineral density of the lumbar spine was measured on two occasions (approximately 1 week apart) at baseline, at 6 months, and at 1 year. Bone mineral density of the midradius, femoral neck, and femoral trochanter was measured only at baseline and at 1 year, using two measurements (approximately 1 week apart) at each visit for the proximal femur and one measurement at each visit for the midradius. Bone densitometry of the lumbar spine and proximal femur was done using dual-photon absorptiometry (11); that of the midradius was done using single-photon absorptiometry (12). All studies

1 July 1992 • Annals of Internal Medicine • Volume 117 • Number 1

Downloaded From: http://annals.org/ by a University of California San Diego User on 12/29/2016

were done at Mayo Clinic except for serum estrogen measurements, which were done at the University of California at Los Angeles. Serum and urine biochemical measurements were done at baseline, at 6 months, and at 1 year; serum hormone levels were measured at baseline and at 1 year. The serum and urine samples were frozen and stored in batches for periodic analysis. Serum and urine calcium levels were measured by atomic absorption spectrophotometry. Serum phosphate and creatinine levels and urinary creatinine levels were measured by routine methods using an autoanalyzer. Serum osteocalcin (bone Gla-protein) levels were measured by radioimmunoassay (13), and serum bone isoenzyme of alkaline phosphatase was measured by enzyme-linked immunosorbent assay (ELISA) (14). Serum estrone, estradiol, and follicle-stimulating hormone levels were measured by radioimmunoassay (15). Urine hydroxyproline levels were measured colorimetrically after fractionation by high-pressure liquid chromatography. The glomerular filtration rate was estimated by measuring creatinine clearance. Transiliac bone biopsy specimens were obtained after double-labeling with tetracycline at baseline and after 1 year of treatment using a 7.5-mm trephine as previously described (16). Bone histomorphometry was done using established techniques. A Bioquant IV computer system (R & M Biometrics, Nashville, Tennessee) was used to quantitate surface measurements on fixed, stained sections. Calculations were made only on cancellous bone (17). Lateral radiographs of the thoracic and lumbar spine were obtained at a standard target-to-film distance of 105 cm. We used a transparent plastic ruler to evaluate and confirm each fracture by measuring the vertical height of the anterior (ha), middle (hm), and posterior (hp) regions of the vertebra (10). From these measurements, fractures were classified as anterior wedge (ha/hp < 85%), central biconcave (hm/hp < 85%), or compression (hp < 85% of hp of the nearest uncompressed vertebra). A new fracture was considered to have occurred if at least one of the three measurements was decreased by more than 15% compared with baseline radiographic measurements (regardless of whether the vertebra was fractured at baseline). Statistical Analysis Baseline characteristics of women in the estrogen and placebo groups and of the changes occurring between the baseline and 1-year visits were compared using Mests, rank-sum tests, or ordinary least-squares extensions of these tests (18), as appropriate, for continuous variables, and using chi-square or exact tests for categoric variables. The primary indicator of response at the lumbar spine was the rate of change in bone mineral density over time as calculated for each woman from an ordinary least-squares linear regression using successive measurements of bone mineral

density. The absolute rate of change (g/cm2 per year) for a woman was equal to the slope of the regression line, whereas the percentage rate of change per year was calculated as 100 times the slope divided by the intercept of the line at study entry. For the other scanning sites, the absolute rate of change in bone mineral density for each woman was calculated as the difference between the mean measurements at baseline and at 1 year divided by time. The percentage rate of change was 100 times the absolute rate divided by the mean measurement at baseline. These rates can be calculated only for women who completed the 1-year trial and who had measurements at both visits. Relations between the observed rates of change in lumbarspine bone mineral density and characteristics of the women at baseline were investigated using multiple-regression models, both with and without weighting according to estimated variances for the rates of change (19). Rates of change in bone mineral density (expressed in mg/cm2 per year or in percent per year) were the dependent variables in these models, and baseline characteristics, an indicator for treatment, and interactions were the independent variables. Correlates of treatment-related changes in lumbar-spine bone mineral density were studied using the changes in or final values of biochemical, histologic, and other densitometric measures as independent variables after inclusion of an indicator for treatment. The group fracture rate was defined as the total number of new fractures in a group per 100 person-years of follow-up for group members. Confidence limits for these rates and for the fracture rate ratio (that is, relative risk in the estrogen group compared with that in the placebo group) were calculated as described by Cochran (20). Relations between the occurrence of fractures and baseline characteristics, treatment, or concomitant changes in measurements related to bone turnover were investigated by ordered categoric data analysis (proportional odds model) using the logistic regression model (21). The dependent variable in this analysis was the number of fractures (zero, one, or more than one) that a patient had experienced during the trial. Both predictive and concurrent relation were considered (as was done for the rates of change in bone mineral density). In all analyses, the results reflect all available information for the 75 women studied. Results Clinical Findings D e m o g r a p h i c , b e h a v i o r a l , medical history, clinical, biochemical, and b o n e histologic characteristics of the w o m e n in the estrogen and p l a c e b o g r o u p s at baseline are s u m m a r i z e d in T a b l e s 1 and 2. T h e g r o u p s did not

Table 2. Major Biochemical, Hormonal, and Histologic Variables at Baseline in the Estrogen and Placebo Groups* Variable Serum Estrone, pmol/L Estradiol, pmol/L Osteocalcin, nmol/L PTH, pmol/L Bone alkaline phosphatase, fikat/L Urine Calcium, ymol/dL. GF/d Hydroxyproline, /imol/dL GF/d Iliac histomorphometry§ Activation frequency, n/d *: 10~4 Bone formation rate per bone volume, %/y

Estrogen (n =36) Median (Range)$

Placebo (n = 39) Median (Range)$

P Valuet

96 (64 to 143) 26 (17 to 54) 1.6 (0.9 to 2.6) 3.6 (1.9 to 6.5) 0.42 (0.25 to 0.69)

92 (55 to 155) 29 (15 to 48) 1.7 (0.4 to 2.5) 4.0 (2.6 to 6.0) 0.43 (0.30 to 0.83)

> 0.2 > 0.2 0.16 > 0.2 > 0.2

4.2 (2.0 to 7.8) 0.14 (0.09 to 0.26)

4.6 (2.3 to 7.9) 0.15 (0.08 to 0.24)

> 0.2 > 0.2

9 (4 to 20) 19 (7 to 46)

9 (2 to 26) 21 (5 to 42)

> 0.2 > 0.2

* GF = glomerular filtrate; PTH = parathyroid hormone. t For difference between groups, by rank-sum test unless otherwise specified. $ Shown as 10th and 90th percentiles. § Information not available for two women. 1 July 1992 • Annals of Internal Medicine

Downloaded From: http://annals.org/ by a University of California San Diego User on 12/29/2016

• Volume 117 • Number 1

3

Figure 1. Mean (± SE) bone mineral density at lumbar spine, femur, and radius in the estrogen and placebo groups. Evaluations included 30 women in the estrogen group and 34 in the placebo group for the lumbar spine, 32 in the estrogen group and 34 in the placebo group for the radius, and 32 in the estrogen group and 33 in the placebo group for the femur. P values are from the rank-sum test comparing differences from baseline in the two groups.

differ significantly in any of these characteristics, nor were any significant differences found between the subgroups of the estrogen and placebo groups used in different sections of the analysis. Adverse experiences included skin irritation in 8 (11%) women (4 [11%] in the estrogen group and 4 [10%] in the placebo group), breast tenderness in 22 (29%) women (20 [56%] in the estrogen group and 2 [5%] in the placebo group), and endometrial hyperplasia in 3 (8%) women in the estrogen group. Breast cancer was diagnosed shortly after the conclusion of the study in two women, one from each group. In the estrogen group, all women with intact uteri experienced the expected menstrual bleeding associated with estrogen plus progestin. None in the placebo group had vaginal bleeding. No clinically important abnormalities were found with regard to physical findings; weight; blood pressure;

Table 3. Rates of Change in Bone Mineral Density in Estrogen and Placebo Groups Site

Estrogen Median (Change ± SE)

Lumbar spine Number Rate, mglcm2ly Rate, %/y Femoral neck Number Rate, mg/cm2ly Rate, %/y Femoral trochanter Number Rate, mg/cm2/y Rate, %/y Midradius Number Rate, mg/cm/y Rate, %/y

Placebo Median (Change ± SE)

P

Value*

33 45 (46 ± 10) 5.3 (6.2 ± 1.4)

37 1 (8 ± 7) 0.2(1.2 ± 0.9)

0.003 0.007t

32 16 (23 ± 8) 2.6(4.1 ± 1.3)

33 9 (7 ± 7) 1.4(1.6 ± 1.4)

0.13 0.17t

32 39 (40 ± 8) 7.6 (8.6 ± 1.7)

33 10 (16 ± 7) 2.1 (3.8 ± 1.6)

0.02 0.03f

32 9 (5 ± 6) 1.0 (0.8 ± 0.8)

34 -20 (-20 ± 3) -2.6 (-2.7 ± 0.4)

< O.OOlt < O.OOlt

* For difference between groups, by Mest, unless otherwise specified. t From the rank-sum test. 4

1 July 1992 • Annals of Internal Medicine

or vaginal or cervical cytologic, hematologic, or clinical chemistry results. Laboratory Findings Values for bone mineral density at the femur, lumbar spine, and radius are shown in Figure 1. For women receiving transdermal estrogen treatment, the median annual rates of change in lumber spine bone mineral density were 5.3% and 45 mg/cm 2 (Table 3). Both of these values were significantly greater than zero (P < 0.001), indicating an increase in bone mineral density. For women in the placebo group, the median annual rates were 0.2% and 1 mg/cm 2 , indicating no change in bone mineral density over the duration of the study. The rates in the estrogen group differed significantly from those in the placebo group (P < 0.007). At midradius, the estrogen group had median annual rates of 1.0% and 9 mg/cm; neither differed significantly from zero, indicating that women receiving estrogen maintained a constant bone mineral content at this site. In contrast, the placebo group had median annual rates of - 2.6% and - 20 mg/cm, both of which were significantly less than zero (P < 0.001). The rates in the two groups also differed significantly from each other (P < 0.001). In the estrogen group, bone mineral density appeared to increase in the hips; rates at both scanning sites were significantly greater than zero (P < 0.01). Only at the femoral trochanter, however, was the increase significant compared with changes in the placebo group (P < 0.03). Serum levels of both estradiol and estrone increased markedly among women in the estrogen group (Figure 2) and reached mean levels similar to those of premenopausal women at midcycle (mean levels at 1 year, 348 pmol/L for estradiol and 260 pmol/L for estrone). The increases in estradiol and estrone levels in the estrogen group were significantly greater than the small changes in the placebo group (P < 0.001).

• Volume 117 • Number 1

Downloaded From: http://annals.org/ by a University of California San Diego User on 12/29/2016

Both serum markers of bone turnover (osteocalcin and bone alkaline phosphatase) showed large and consistent decreases in the estrogen group but remained unchanged in the placebo group (P < 0.001 for the comparison between groups) (Figure 3). Urinary hydroxyproline excretion corrected for glomerular filtration rate and histomorphometric values related to turnover (bone formation rate per bone volume and activation frequency) also decreased greatly in the estrogen group (compared with no change, P < 0.001 in all cases), and the differences from the corresponding changes in the placebo group were all significant (P < 0.02). Several biochemical characteristics did not change during the year of the trial. The two groups did not differ significantly regarding changes in serum intact parathyroid hormone concentration, urinary calcium concentration (see Figure 3), or serum lipid levels (Table 4). Vertebral Fractures Repeat radiographs done after 1 year of treatment were available for 34 women in each group. Eight new vertebral fractures occurred in 7 women in the estrogen group, whereas 20 new fractures occurred in 12 women in the placebo group. Thus, the group fracture rate for the estrogen group (23 fractures per 100 person years) was less than half the rate for the placebo group (58 fractures per 100 person-years). Also, the relative risk was significantly different from 1 (relative risk = 0.39; P = 0.04), indicating a protective effect of transdermal estrogen. Effect of Covariates on Response to Treatment The variables in Tables 1 and founders and as gression models

evaluated at baseline and summarized 2 were all assessed as potential coneffect modifiers (22) using multiple rewith rates of change in lumbar-spine

bone mineral density as the dependent variable. Individually and in combination, several of the variables were significantly related to changes in bone mineral density and thus had the potential to confound the relation between treatment and response. Of these variables, markers of bone turnover, initial bone density, age or time since menopause, factors related to reproductive history, initial serum estrone levels, previous estrogen replacement therapy, and consumption of alcohol and caffeinated beverages did not affect the size or direction of the predicted treatment response. However, three variables—decaffeinated coffee consumption, relative weight, and serum estradiol levels (below or above the median at baseline)—were significant modifiers of the effect of treatment on lumbar-spine bone mineral density. The difference between the estrogen and placebo groups was not the same for patients with different values of these variables. Our analyses indicate that the difference between the estrogen and placebo groups increased with increased decaffeinated coffee consumption, decreased with increased relative weight, and was less for those with an initial serum estradiol concentration greater than 25.7 pmol/L. Despite the influence of these factors, our data indicated that estrogen treatment increases the bone mineral density of the lumbar spine. Similar analyses using the proportional-odds model reinforce the conclusion that the estrogen group has a significantly lower vertebral fracture rate than does the placebo group. In analyses of subgroups above and below the median age of 65 years, the differences between the estrogen and placebo groups among older women were at least as large as those among younger women for all main end points of the trial. The median rate of change in lumbar-spine bone mineral density for women in the estrogen group who were 65 years of age or older was 5.3% per year (0.046 g/cm2 per year) compared with 4.4% per year (0.038 g/cm2 per year) for those under age 65, showing that gains in bone mass with estrogen

Figure 2. Serum estrogen levels and histologic indices of bone turnover in the estrogen and placebo groups. Mean ± SE for women with measurements at baseline and at 1 year. Thirty-three women in the estrogen group and 34 in the placebo group were evaluated for serum estrogens, and 27 in the estrogen group and 31 in the placebo group were evaluated for histologic indices. P values are as in Figure 1. TBV = total bone volume. 1 July 1992 • Annals of Internal Medicine • Volume 117 • Number 1

Downloaded From: http://annals.org/ by a University of California San Diego User on 12/29/2016

5

Figure 3. Serum and urine biochemical markers of bone turnover and serum parathyroid hormone in the estrogen and placebo groups. Mean ± SE for women with measurements at baseline and at 1 year. Thirty-three women in the estrogen group and 34 in the placebo group were evaluated for serum measurements. Thirty-two women in the estrogen group and 34 in the placebo group were evaluated for urine measurements. P values are as in Figure 1. AP = alkaline phosphatase; GF = glomerular filtrate; PTH = parathyroid hormone.

did not decrease among older women. Analyses of subgroups above or below the median time since menopause (15 years) yielded similar results. Previous estrogen replacement therapy was found to have no effect on the differences between the estrogen and placebo groups with respect to the main end points of our study. Discussion We found that transdermal estrogen treatment had important beneficial effects on bone metabolism. The increased indices of bone turnover present at baseline were decreased among women receiving estrogen to levels found in normal premenopausal women. In the lumbar spine, a site containing mostly cancellous bone, bone mineral density increased after 1 year of treatment. Bone mineral density increased to a somewhat lesser extent in the proximal femur, a site containing both cancellous and compact bone. Treatment stabilized bone mineral density at the midradius, a site containing almost exclusively cortical bone. These changes represent a reversal of the changes induced by estrogen deficiency. After menopause, bone turnover increases, and the rate of cancellous bone loss is much greater than that of cortical bone (23). Our results are consistent with those of Lindsay and colleagues (24) who treated 40 postmenopausal, osteoporotic women with calcium supplementation, 1500 mg/d, combined with 1.25 mg/d of conjugated estrogen or placebo, and with those of Civitelli and associates (25) who treated a group of 21 postmenopausal, osteoporotic women for 1 year with either conjugated estrogen or placebo. Because transdermal estrogen and medroxyprogesterone acetate were given together, we cannot exclude the 6

possibility that the latter hormone may have had an additive effect to that of estrogen. Among women in the placebo group, bone mineral density of the lumbar spine and proximal femur (as assessed by dual-photon absorptiometry) did not decrease significantly, whereas significant decreases in bone mineral density were detected at the midradius (as assessed by single-photon absorptiometry). We attribute these differences in results to the greater precision and accuracy of single-photon absorptiometry (11). The adjustment of calcium intake and institution of general supportive measures that were made in all patients could, however, have contributed to the decrease in the rate of bone loss in the placebo group. Although we cannot exclude the possibility of a systematic error in the dual-photon absorptiometry measurements, the randomized study design prevents this factor from affecting the validity of comparisons between groups. Our results provide no indication that transdermal estrogen therapy is less effective in older osteoporotic women or in those with longer times since menopause; estrogen effects were, in fact, greater in these subgroups. No serious side effects attributable to estrogen occurred during the trial. The patch irritation that occurred in 11% of all patients, but equally in both groups, was presumably caused by the alcohol vehicle. In the estrogen group, 56% of patients had mastodynia and 8% had endometrial hyperplasia. These side effects probably resulted from the reestablishment of physiologic concentrations of serum estrogens during treatment. The dosage of 0.1 mg 17/3-estradiol per day resulted in serum concentrations of both estradiol and estrone that were similar to those in the midfollicular phase of the menstrual cycle of premenopausal women

1 July 1992 • Annals of Internal Medicine • Volume 117 • Number 1

Downloaded From: http://annals.org/ by a University of California San Diego User on 12/29/2016

(26). The achievement of these normal premenopausal serum concentrations of both estrogens, despite administration of only 17/3-estradiol, undoubtedly resulted from peripheral conversion of 17/3-estradiol to estrone. This pattern of serum estrogen differs from that obtained by oral treatment of postmenopausal women with conjugated estrogens, which results mainly in increases in estrone and equine compounds not present in humans (26). Others have noted that transdermal estrogen treatment eliminates hepatic overproduction of estrogendependent proteins (27). Results from previous studies have shown that the gain in bone mass after administration of antiresorptive agents such as estrogen is transient (28). The increase continues until a new steady state is reached, at which time the decrease in bone resorption is matched by a coupled decrease in bone formation. Subsequently, bone mass is maintained, but no further increases occur. Theoretic estimates based on tetracycline-based histomorphometry (28) suggest that this "leveling off" in bone mass should occur within 3 to 4 months in normal women, although it may take longer in patients with osteoporosis. In our study, bone turnover as assessed by biochemical indices showed a continued decrease, and bone mineral density of the lumbar spine as assessed by bone densitometry showed a continued increase between 6 and 12 months. In a 2-year study, Lindsay and colleagues (24) found that increases in bone density of the lumbar spine leveled off by 18 months but that a gain at the femoral neck continued for 24 months. We found a slight decrease in serum total cholesterol levels in the estrogen group (which was not significant when compared with changes in the placebo group) and no changes in LDL cholesterol, HDL cholesterol, or triglyceride levels. Others have reported no change (29)

or decreases (30, 31) in total cholesterol and LDL cholesterol levels with transdermal estrogen treatment. It is unclear whether the lipid effects of transdermal estradiol are influenced by coadministered progestins, which differ in type and dose in these studies. Oral administration of progestins may have adverse effects on lipids (32) and thus could explain why we did not find the same decrease in cholesterol levels (30, 31) reported by others. Observational studies have suggested that oral estrogen treatment may decrease the incidence of myocardial infarction by 40% to 50% (33, 34), but this result remains to be confirmed by prospective, randomized trials. This decrease in the rate of myocardial infarction has been attributed to an associated decrease in LDL cholesterol and an increase in HDL cholesterol levels (35, 36). Results from studies in primate models have suggested, however, that the cardioprotective effect of estrogen may result from a direct action of estrogen on the arterial wall (37-40). Ours is the first demonstration that estrogen treatment decreases vertebral fracture rates in women with established osteoporosis. A relatively small increase in bone mass was associated with a greater than 50% decrease in the vertebral fracture rate, despite the relatively short duration and small sample size of this trial. Although the reason for this apparent paradox is unclear, we suggest that it is related to the suppression of bone turnover by estrogen therapy. Studies using stereologic analysis of cancellous-bone biopsy samples have indicated that bone loss in women with postmenopausal osteoporosis is caused not only by thinning of trabeculae but also by perforation of trabecular plates associated with increased osteoclastic activity (41). The perforative resorption leads to a loss of structural elements within cancellous bone and predisposes the bone to fracture. Transdermal estrogen treatment would

Table 4. Changes in Serum Lipids in Estrogen and Placebo Groups Variable Values Median (Range)t Total cholesterol, mmollL Baseline

Estrogen (n = 34) Difference Median (Mean ± SE)

5.8 (5.0 to 7.6)

Values Median (Range)t

Placebo (n = 34) Difference Median (Mean ± SE)

6.2 (4.9 to 7.2) - 0.10 ( - 0.34 ± 0.14)

1 Year High-density lipoprotein cholesterol:}: mmollL Baseline

5.5 (4.8 to 7.4)

0.08 ( - 0.09 ± 0.18)

1.5 (1.1 to 2.1)

1.4(1.0 to 1.8)

1.4(1.1 to 1.9) 3.9 (3.0 to 5.5)

* t t §

> 0.2

0.00 ( - 0.13 ± 0.15)

> 0.2

- 0.12 ( - 0.05 ± 0.09)

> 0.2

4.2 (3.0 to 5.2)

3.9 (3.0 to 5.6)

4.0 (2.9 to 4.9)

0.9 (0.6 to 1.8)

1.1 (0.7 to 2.1) - 0.02 ( - 0.12 ± 0.07)

1 Year

0.04 (0.03 ± 0.05) 1.5 (1.0 to 1.9)

- 0.10 ( - 0.18 ± 0.14) 1 Year Triglycerides, mmollL Baseline

0.16

5.8 (4.7 to 6.8)

- 0.08 ( - 0.02 ± 0.06) 1 Year Low-density lipoprotein cholesterol§ Baseline

P Value*

0.9 (0.5 to 1.6)

1.1 (0.6 to 2.0)

For differences between groups, by rank-sum test. Shown as 10th and 90th percentiles Information not available for two women in each group. Information not available for three women in each group.

1 July 1992 • Annals of Internal Medicine • Volume 117 • Number 1

Downloaded From: http://annals.org/ by a University of California San Diego User on 12/29/2016

7

probably also decrease hip fracture rates, as it does vertebral fracture rates, because bone mineral density in the estrogen group increased at the hip sites studied. Many patient-years of follow-up are, however, necessary to show such a decrease. Despite the obvious physiologic benefits of estrogen therapy and the causal or contributing role of estrogen deficiency in pathogenesis, estrogen is not widely used in the treatment of established osteoporosis, perhaps because of the mistaken belief that the disease has progressed too far for such treatment to be useful. The results that we obtained with transdermally administered estrogen are, however, as good as or better than those reported with other regimens. Although treatment with sodium fluoride results in large increases in cancellous bone mass, two recent prospective, placebocontrolled clinical trials (42) showed that it did not decrease the vertebral fracture rate, and, in one trial (10), it increased the rate of nonvertebral fractures threefold. This result suggests, at least for the doses used, that fluoride therapy increases bone fragility. Based on the favorable results of two recently published clinical trials (43, 44), cyclic therapy of osteoporosis with etidronate has attracted wide interest. The increase in bone density of the lumbar spine and the decrease in vertebral fracture rate achieved after 1 year of treatment with transdermal estrogen were, however, similar to results achieved with 2 to 3 years of cyclic etidronate treatment. Our definition of a new fracture was, however, somewhat less stringent than that used in their study. Although quantitative comparison of two distinct therapeutic programs would require a randomized comparison within the same clinical trial, our data indicate that estrogen can substantially increase bone mass over a short term in patients with established osteoporosis. Our results are also similar to those obtained using a combination of estrogen and norethisterone, an analog of 19-nortestosterone (45), or using calcitonin (46) in patients with established postmenopausal osteoporosis. We conclude that transdermal estrogen treatment, at least for the conditions and 1-year duration of this study, is safe and effective for women with established postmenopausal osteoporosis. Although the relative efficacies of the various available regimens can be assessed only by randomized studies using side-by-side comparisons, we found no evidence that estrogen therapy is less effective than any currently available regimen for treatment of women with established postmenopausal osteoporosis. Acknowledgments: The authors thank Ms. Ginny Wong for her management of the study as research nurse coordinator and to the staff of the Endocrine Testing Unit for obtaining the bone biopsy specimens. We also thank Drs. A. M. Pecora and K. Wolter of Ciba-Geigy Corporation for their assistance. Requests for Reprints: Edward G. Lufkin, MD, Mayo Clinic, 200 First Street S.W., Rochester, MN 55905. Grant Support: By a grant from the Ciba-Geigy Corporation, Summit, New Jersey. Current Author Addresses: Drs. Lufkin, Hodgson, and Riggs: Division of Endocrinology and Metabolism, Mayo Clinic, Rochester, MN 55905. Dr. Wahner: Section of Diagnostic Nuclear Medicine, Mayo Clinic, Rochester, MN 55905. Dr. O'Fallon and Ms. Lane: Department of Health Sciences Research, Mayo Clinic, Rochester, MN 55905.

8

Dr. Kotowicz: University Department of Medicine, Geelong Hospital, Geelong, 3220 Victoria, Australia. Dr. Judd: Department of Obstetrics & Gynecology, Division of Reproductive Endocrinology, UCLA School of Medicine, Center for the Health Sciences, Los Angeles, CA 90024. Dr. Caplan: Endocrinology Section, Department of Internal Medicine, Gundersen Clinic, Ltd., La Crosse, WI 54601. References 1. Kiel DP, Felson DT, Anderson JJ, Wilson PWF, Moskowitz MA. Hip fracture and the use of estrogens in postmenopausal women. The Framingham study. N Engl J Med. 1987;317:1169-74. 2. Paganini-Hill A, Ross RK, Gerkins VR, Henderson BE, Arthur M, Mack TM. Menopausal estrogen therapy and hip fractures. Ann Intern Med. 1981;95:28-31. 3. Ettinger B, Genant HK, Conn CE. Long-term estrogen replacement therapy prevents bone loss and fractures. Ann Intern Med. 1985; 102:319-24. 4. Smith R. Osteoporosis after 60. Br Med J. 1990;301:452-3. 5. Riggs BL. A new option for treating osteoporosis. N Engl J Med. 1990;323:124-5. 6. Judd HL, Meldrum DR, Deftos LJ, Henderson BE. Estrogen replacement therapy: indications and complications. Ann Intern Med. 1983;98:195-205. 7. Ribot C, Tremollieres F, Pouilles JM, Louvet JP, Peynon R. Preventive effects of transdermal administration of 17-/3-estradiol on postmenopausal bone loss: A 2-year prospective study. Obstet Gynecol. 1990;75(Suppl):42S-6S. 8. Stevenson JC, Anst MP, Gangar KF, Hillard TC, Lees B, Whitehead MI. Effect of transdermal versus oral hormone replacement therapy on bone density in spine and proximal femur in postmenopausal women. Lancet. 1990;336:265-8. 9. Adami S, Suppi R, Bertoldo F, Rossini M, Residori M, Maresca V, et al. Transdermal estradiol in the treatment of postmenopausal bone loss. Bone Miner. 1989;7:79-86. 10. Riggs BL, Hodgson SF, O'Fallon WM, Chao EY, Wahner HW, Muhs JM, et al. Effect of fluoride treatment on the fracture rate in postmenopausal women with osteoporosis. N Engl J Med. 1990;322: 802-9. 11. Riggs BL, Wahner HW, Seeman E, Offord KP, Dunn WL, Mazess RB. Changes in bone mineral density of the proximal femur and spine with aging: differences between the postmenopausal and senile osteoporosis syndromes. J Clin Invest. 1982;70:716-23. 12. Riggs BL, Wahner HW, Melton LJ III, Richelson LS, Judd HL, Offord KP. Rates of bone loss in the appendicular and axial skeletons of women: Evidence of substantial vertebral bone loss before menopause. J Clin Invest. 77:1487-91, 1986. 13. Delmas PD, Wahner HW, Mann KG, Riggs BL. Assessment of bone turnover in postmenopausal osteoporosis by measurement of serum bone Gla-protein. J Lab Clin Med. 1983;102:470-6. 14. Duda RJ Jr, O'Brien JF, Katzman JA, Peterson JM, Mann KG, Riggs BL. Concurrent assays of circulating bone Gla-protein and bone alkaline phosphatase: Effects of sex, age, and metabolic bone disease. J Clin Endocrinol Metab. 1988;66:951-7. 15. Judd HL, Judd GE, Lucas WE, Yen SS. Endocrine function of the postmenopausal ovary: concentrations of androgens and estrogens in ovarian and peripheral vein blood. J Clin Endocrinol Metab. 1974;39:1020-4. 16. Hodgson SF, Johnson KA, Muhs JM, Lufkin EG, McCarthy JT. Out patient percutaneous biopsy of the iliac crest: Methods, morbidity, and patient acceptance. Mayo Clin Proc. 1986;61:28-33. 17. Eriksen EF, Hodgson SF, Eastell R, Cedel SL, O'Fallon WM, Riggs BL. Cancellous bone remodeling in type I (postmenopausal) osteoporosis: Quantitative assessment of rates of formation, resorption, and bone loss at tissue and cellular levels. J Bone Miner Res. 1990;5:311-9. 18. O'Brien PC. Comparing two samples: Extensions of the t, ranksum, and log rank tests. JASA 1988;83:52-61. 19. Draper NR, Smith H. Applied Regression Analysis. 2d edition. New York: John Wiley; 1981:108-16. 20. Cochran WG. Sampling Techniques. 3d edition. New York: John Wiley; 1977:153-6. 21. McCullagh P. Regression models for ordinal data. J R Stat Soc [B]. 1980;42:109-42. 22. Kleinbaum DG, Kupper LL, Morgenstern H. Epidemiologic Research: Principles and Quantitative Methods. New York: Van Nostrand Reinhold; 1982. 23. Heaney RP, Recker RR, Saville PD. Menopausal changes in bone remodeling. J Lab Clin Med. 1978;92:964-70. 24. Lindsay R, Tohme JF. Estrogen treatment of patients with established osteoporosis. Obstet Gynecol. 1990;76:290-5. 25. Civitelli R, Agnusdei D, Nardi P, Zacchei F, Avioli LV, Gennari C. Effects of one-year treatment with estrogens on bone mass, intestinal calcium absorption, and 25-hydroxyvitamin D-la-hydroxylase reserve in postmenopausal osteoporosis. Calcif Tissue Int. 1988;42: 77-86.

1 July 1992 • Annals of Internal Medicine • Volume 117 • Number 1

Downloaded From: http://annals.org/ by a University of California San Diego User on 12/29/2016

26. Powers MS, Schenkel L, Darley BS, Good WR, Balestra JC, Place VA. Pharmacokinetics and pharmacodynamics of transdermal dosage forms of 17/3-estradiol: comparison with conventional oral estrogens used for hormone replacement. Am J Obstet Gynecol. 1985; 152:1099-1106. 27. Chetkowski RJ, Meldrum DR, Steingold KA, Randle D, Lu JK, Eggena P, et al. Biologic effects of transdermal estradiol. N Engl J Med. 1986;314:1615-20. 28. Parfitt AM. Morphologic basis for bone mineral measurements: transient and steady state effects of treatment in osteoporosis. Miner Electrolyte Metab. 1980;4:273-87. 29. Faguer de Moustier B, Conard J, Guyene TT, Sitt Y, Denys I, Arnoux-Rouveyre M, et al. Comparative metabolic study of percutaneous versus oral micronized 17-0-oestradiol in replacement therapy. Maturitas. 1989;11:275-86. 30. Keller PJ. Percutaneous estrogen therapy in the postmenopausal period. Schweizerische Medizinische Wochenschrift. J Suisse Med. 1989;119:999-1004. 31. Stanczyk FZ, Shourpe D, Nunez V, Macrias-Gonzales P, Vijod MA, Lobo RA. A randomized comparison of normal estradiol delivery in postmenopausal women. Am J Obstet Gynecol. 1988;159:1540-6. 32. Hirvonen E, Malkonen M, Manninen V. Effects of different progestogens on lipoproteins during postmenopausal replacement therapy. N Engl J Med. 1981;304:560-3. 33. Barrett-Connor E, Wingard DL, Ariguis MH. Postmenopausal estrogen use and heart disease risk factors in the 1980s. JAMA. 1989; 261:2095-100. 34. Stampfer MJ, Colditz GA, Willett WC, Manson JE, Rosner B, Speizer FE, et al. Postmenopausal estrogen therapy and cardiovascular disease: Ten-year follow-up from the nurses' health study. N Engl J Med. 1991;325:756-62. 35. Walsh BW, Schief I, Rosner B, Greenberg L, Ravnikar V, Sacks FM. Effects of postmenopausal estrogen replacement on the concentrations and metabolism of plasma lipoproteins. N Engl J Med. 1991; 325:1196-204.

36. Barrett-Connor E, Bush TL. Estrogen and coronary heart disease in women. Clin Cardiol. 1991;265:1861-7. 37. Lin AL, Gonzalez R, Carsey KD, Shain SA. Estradiol-17 j3 affects estrogen receptor distribution and elevates progesterone receptor content in baboon aorta. Arteriosclerosis. 1986;6:495-503. 38. Williams JK, Adams MR, Klopfenstein HS. Estrogen modulates responses of atherosclerotic coronary arteries. Circulation. 1990;81: 1680-7. 39. Harder DR, Coulson PB. Estrogen receptors and effects of estrogen on membrane electrical properties of coronary vascular smooth muscle. J Cell Physiol. 1979;100:375-82. 40. Adams MR, Kaplan JR, Manuck SB, Koritnik DR, Parks JS, Wolfe MS, et al. Inhibition of coronary artery atherosclerosis by 17-beta estradiol in ovariectomized monkeys. Arteriosclerosis. 1990; 10: 1051-7. 41. Parfitt AM, Mathews CH, Villaneuva AR, Kleerekoper M, Frame B, Rao DS. Relationship between surface, volume, and thickness of iliac trabecular bone in aging and in osteoporosis. J Clin Invest. 1983;72:1396-1409. 42. Kleerekoper M, Peterson EL, Nelson DA, Phillips E, Schork MA, Tilley BC, et al. A randomized trial of sodium fluoride as a treatment for postmenopausal osteoporosis. Osteoporosis Int 1991;1:155161. 43. Watts NB, Harris SJ, Genant HK, Wasnich RD, Miller PD, Jackson RD, et al. Intermittent cyclical etidronate treatment of postmenopausal osteoporosis. N Engl J Med. 1990;323:73-9. 44. Storm J, Thamsberg G, Steiniche T, Genant HK, Sorenson OH. Effect of intermittent cyclical etidronate therapy on bone mass and fracture rate in women with postmenopausal osteoporosis. N Engl J Med. 1990;322:1265-171. 45. Christiansen C, Riis BJ. 17j3-Estradiol and continuous norethisterone: A unique treatment for established osteoporosis in elderly women. J Clin Endocrinol Metab. 1990;71:836-41. 46. Civitelli R, Gonnelli S, Zacchei F, Bigazzi S, Vattimo A, Avioli LV, et al. Bone turnover in postmenopausal osteoporosis: Effect of calcitonin treatment. J Clin Invest. 1988;82:1268-74.

1 July 1992 • Annals of Internal Medicine • Volume 117 • Number 1

Downloaded From: http://annals.org/ by a University of California San Diego User on 12/29/2016

9