- Email: [email protected]
Depot medroxyprogesterone acetate, oral contraceptives and bone mineral density in a cohort of adolescent girls
JOURNAL OF ADOLESCENT HEALTH 2004;35:434 – 441
ORIGINAL ARTICLE
Depot Medroxyprogesterone Acetate, Oral Contraceptives and Bone Mineral Density in a Cohort of Adolescent Girls BARBARA A. CROMER, M.D., MARGARET STAGER, M.D., ANDREA BONNY, M.D., RINA LAZEBNIK, M.D., ELLEN ROME, M.D., M.P.H., JULIE ZIEGLER, M.A., AND SARA M. DEBANNE, Ph.D.
Purpose: To conduct a longitudinal comparison of bone mineral density (BMD) in 370 adolescent girls, aged 12–18, who self-selected depot medroxyprogesterone acetate (DMPA) or an oral contraceptive (OC) containing 20g ethinyl estradiol/100g levonorgestrel with that in girls who received no hormonal treatment (control group). Methods: Lumbar spine and femoral neck BMD measurements were obtained by dual energy x-ray absorptiometry at baseline and 12 months. Data were analyzed with repeated measures analysis of covariance methods. Results: Over 12 months, lumbar spine BMD decreased in the DMPA group (n ⴝ 29), with a mean percent change of ⴚ1.4% (95% confidence interval [CI] ⴚ2.73, ⴚ0.10), and increased by a mean of 3.8% (95% CI 3.11, 4.57) in the control group [n ⴝ 107 (p < .001)]. The increase in mean percent change in lumbar spine BMD in the OC group (n ⴝ 79), 2.3% (95% CI 1.49, 3.18), was significantly smaller than in the control group (p ⴝ .03). Over 12 months, the mean percent change in femoral neck BMD was ⴚ2.2% (95% CI ⴚ3.95, ⴚ0.39) in the DMPA group, but increased 2.3% (95% CI 1.29, 3.27) in the control group (p < .001). The increase in mean percent change at the femoral neck in the OC group, 0.3% (95% CI ⴚ0.87, 1.41), was significantly lower than in the control group (p ⴝ .03).
From the MetroHealth Medical Center; Rainbow Babies & Children’s Hospital, University Hospitals of Cleveland; The Children’s Hospital at the Cleveland Clinic; and Department of Epidemiology and Biostatistics, Case Western Reserve University School of Medicine, Cleveland, Ohio . Address correspondence to: Dr. Barbara Cromer, Department of Pediatrics, MetroHealth Medical Center, 2500 MetroHealth Dr., Cleveland, OH 44109. E-mail: [email protected] Manuscript accepted July 29, 2004. 1054-139X/04/$–see front matter doi:10.1016/j.jadohealth.2004.07.005
Conclusions: Our study contributes to an increasing body of knowledge indicating a negative impact of DMPA on bone health in young women. Additional findings suggest a potential adverse effect of an OC containing 20 g ethinyl estradiol/100 g levonorgestrel on bone health in adolescents. © Society for Adolescent Medicine, 2004
KEY WORDS:
Bone mineral density Adolescents Depot medroxyprogesterone acetate Oral contraceptives
Adolescence is a crucial period for skeletal growth [1]. Sex hormones, especially estrogen, play a central role in bone development [2]; therefore, alterations in their circulating levels have a significant impact on bone and future peak bone mass. The question that emerges is whether use of exogenous sex hormones, in the form of hormonal contraception, affect bone health in the adolescent. The two most commonly used forms of hormonal contraception by adolescent girls in this country are oral contraceptives (OC) and depot medroxyprogesterone acetate (DMPA) [3]. Since its approval by the FDA as a contraceptive in 1992, DMPA use by young women has been credited, in part, with the decrease in adolescent pregnancy rates observed over the past decade [4,5]. Overall, the method has been shown to be very effective, safe and to have few adverse © Society for Adolescent Medicine, 2004 Published by Elsevier Inc., 360 Park Avenue South, New York, NY 10010
December 2004
clinical effects [6,7]. A previous study, however, demonstrated bone loss in a small group of adolescent DMPA users [8]. The goal of our study was to investigate whether these previous findings were replicable in a larger adolescent population. Over the past 30 years, secular trends in product development in the pharmaceutical industry resulted in a progressive lowering of estrogen dose in OC from over 150 g to 20 g ethinyl estradiol [9], and in the most recent formulation, to 15 g ethinyl estradiol (with 60 g gestodene) [10]. No published data are available regarding the relationship between OC containing less than 30 g ethinyl estradiol and bone in the adolescent population. The second goal of our study was to examine whether bone mineral density (BMD) increased in adolescent girls using an OC with 20 g ethinyl estradiol/100 g levonorgestrel at a rate consonant with that seen in untreated, healthy adolescent girls. In December 2003, before completion of study enrollment, we performed an interim analysis of our concurrent clinic trial [11] at the request of our Data Safety Monitoring Board. Because of the highly significant results in the clinical trial, the Institutional Review Board requested a hold on recruitment of subjects initiating DMPA. Therefore, we froze the data set as of this date and elected to report our findings from the interim analysis. The data set included subjects who completed a 12-month observation, subjects who withdrew before their 12-month observation, and subjects who had not yet reached their 12-month observation.
Materials and Methods This study was designed as a 24-month, prospective examination of BMD and hormonal contraception in adolescents. This report presents results from a 12month interim analysis. Postmenarcheal girls, ranging from 12-18 years of age, were recruited from four general adolescent health clinics located in a large, urban, metropolitan setting. Specifically, adolescent girls requesting contraception, and selecting either DMPA or OC, were eligible for enrollment. Concurrently, girls attending the same clinics, who planned to receive no treatment with sex hormones, were eligible for enrollment. Exclusion criteria included use of DMPA, pregnancy, or abortion over the past 6 months, use of oral contraceptives over the past 3 months, chronic medical condition or treatment that may have an effect on BMD, or need for confidentiality in contraception management. The Institutional
BONE HEALTH IN YOUNG WOMEN
435
Review Boards of all participating institutions approved the study. Written, informed consent and assent were obtained from each custodial parent and study enrollee, respectively. At baseline, 6, and 12 months, clinical and behavioral information was obtained from each study participant. Tobacco use was reported as number of cigarettes per day over the past 30 days, as it is reported on the Youth Risk Behavior Survey [12]. Calcium intake was elicited with a focused 24-hour dietary recall combined with the Calcium Rapid Assessment Method [13]. Girls who consumed ⬍ 1200 mg/day dietary calcium were counseled by a dietician; if the level of intake did not improve after 3 months, the subject was given a sample of Tums® 500 mg to be taken once daily for 3 months. Physical activity was assessed using a survey that asked each subject to classify herself as very inactive, inactive, normal, active, and very active. Clinical information was collected by direct interview and included menstrual bleeding patterns, general medical symptoms, and medication use. Height and weight were measured using the same stadiometer (Easy Glide Bearing stature board) and Mettler-Toledo scale every visit. Gynecologic age was defined as the number of years since menarche. At baseline, 6, and 12 months, all study participants underwent measurement of BMD that included L1–L4 lumbar vertebrae, total hip (left), femoral neck, trochanter, and Ward’s triangle. The lumbar vertebrae were chosen because of the high percentage of trabecular bone and the sensitivity of lumbar spine BMD to hormone-related changes in bone mass. The hip was also chosen because osteoporosis fractures at this site are associated with the greatest morbidity later in life. The measurement technique employed was dual energy x-ray absorptiometry (DEXA) with the model QDR 4500W fanbeam densitometer (Hologic Inc., Bedford, MA). Windows 11.2 was the software, which included a low-density measurement option. Technicians received training from Hologic and were certified for machine operation by the Ohio Department of Health. Calibration protocol for the DEXA machine included a daily phantom reading and a weekly check for drift by a radiologist. In vivo intra-individual coefficients of variation were 1.2% at the spine and 1.4% at the femoral neck; inter-individual coefficients of variation were 1.3% at the spine and 2.2% at the femoral neck. All scans were obtained within 4 weeks of the scheduled 6-month intervals. Because longitudinal growth typically does not cease altogether until late adolescence, skeletal size
CROMER ET AL
436
JOURNAL OF ADOLESCENT HEALTH Vol. 35, No. 6
Table 1. Characteristics of Study Participants at Baselinea Characteristics
DMPA (n ⫽ 53)
OC (n ⫽ 165)
Control (n ⫽ 152)
Age in years Gynecological age in years Race, n (% black) Body Weight (kg) Height in (cm) Body Mass Index Spine BMD Femoral Neck BMD Vitamin D Regular Menses, n (%) Number of Cigarettes per Month n (%) Physical Activityh Active, n (%) Normal, n (%) Inactive, n (%)
15.6 ⫾ 1.5 3.7 ⫾ 1.6 34 (64) 60.5 ⫾ 13.0d 161.7 ⫾ 6.9 23.1 ⫾ 4.2e 0.98 ⫾ 0.09 0.92 ⫾ 0.14 57.1 ⫾ 25.8 46 (87) 39.7 ⫾ 97.7 43 (81)
16.0 ⫾ 1.4 4.0 ⫾ 1.6 103 (62) 69.1 ⫾ 17.3 161.8 ⫾ 6.4 26.4 ⫾ 6.4 1.03 ⫾ 0.11f 0.98 ⫾ 0.16g 54.4 ⫾ 30.0 133 (81) 32.8 ⫾ 92.8 119 (72)
14.8 ⫾ 1.5b 2.6 ⫾ 1.6c 95 (63) 63.2 ⫾ 16.9 160.9 ⫾ 6.6 24.3 ⫾ 5.8 0.98 ⫾ 0.12 0.92 ⫾ 0.14 55.6 ⫾ 29.9 123 (81) 8.7 ⫾ 50.2 117 (77)
14 (29) 27 (55) 8 (16)
67 (43) 78 (50) 10 (7)
69 (48) 62 (43) 13 (9)
All numbers represent mean ⫾ SD except: Race, Regular Menses, and Physical Activity, which are reported as percentages. Chronological Age is significantly lower in control group than DMPA group (p ⬍ .001) and OC group (p ⬍ .001). Gynecological Age is significantly lower in control group than DMPA group (p ⬍ .001) and OC group (p ⬍ .001). d Body Weight is significantly higher in OC group than DMPA group (p ⫽ .001) and control group (p ⬍ .002). e Body Mass Index is significantly higher in OC group than DMPA group (p ⬍ .001) and control group (p ⫽ .002). f Spine BMD Spine is significantly higher in OC group than DMPA group (p ⬍ .001) and control group (p ⬍ .001). g Femoral Neck BMD is significantly higher in OC group than DMPA group (p ⫽ .002) and control group (p ⫽ .002). h Missing data: DMPA group (n ⫽ 4), OC group (n ⫽ 10), control group (n ⫽ 8). a
b c
may have changed in our study subjects over the observation period. As bones grow and increase in width and height, the bone thickness increases, and BMD can be confounded by changes in thickness, i.e., increased bone thickness may falsely appear as increased bone density. BMD was calculated by amount of scanned bone mineral content (BMC) within a projected area (Ap), termed areal density, and is expressed as g/cm2. Therefore, to correct for volumetric variations in bone, we performed an additional calculation to achieve bone mineral apparent density (BMAD), utilizing the following formula: spine BMAD ⫽ BMC (L1–L4) ⫼ Ap3/2 and femoral neck BMAD ⫽ BMC(femoral neck) ⫼ Ap2(femoral neck) [14]. DMPA was administered as a 150-mg deep intramuscular (gluteus or triceps) injection every 12 weeks. All girls in this study received an OC containing 20 g ethinyl estradiol and 100 g levonorgestrel. Compliance with DMPA injections was assessed by chart review and was calculated as number of injections divided by number of prescribed injections (1 every 3 months) ⫻ 100. Compliance rates with OC use was assessed by monthly self-report and calculated as the number of pills taken divided by number of pills prescribed (1 every day) ⫻ 100. Girls who chose to switch their contraceptive method during the 12-month observation period were withdrawn from the study.
The association of changes in both BMD and BMAD over time was assessed in two ways. First, repeated measures analysis of covariance (ANCOVA) was used, where the time of the visit (baseline, 6 months, and 12 months) was used as a factor in the model. The approach allows full utilization of all available follow-up data. Linear modeling is justified because the time intervals are only 6 months in length. ANCOVA was also used to examine the annual percentage change, where the time factor was incorporated into the endpoint by calculating a percentage change in BMD or BMAD from baseline to 12 months. Before modeling, the distributional properties of continuous variables were examined for possible departures from normality. Multi-collinearity of predictors was also examined. The high correlation between chronologic age and gynecologic age (Spearman correlation rho ⫽ 0.69, p ⬍ .001, n ⫽ 370), and between body weight and BMI (rho ⫽ 0.94, p ⬍ .001, n ⫽ 370) led to the decision to use only one from each pair in the multivariate models. Because three baseline characteristics, i.e., chronological age, weight, and race are powerful predictors of BMD [6], they were treated as confounding variables and were included in all models, whether significant or not. Regarding physical activity, tobacco use, and vitamin D status, these were not significantly related to BMD and, therefore, were not included in the models. Second order interaction terms between treatment type and other covariates
December 2004
BONE HEALTH IN YOUNG WOMEN
437
Table 2. Adjusteda Means and Standard Deviations (SD) of Absolute and Percent Change for L1-L4 Spine BMD and Femoral Neck BMD Anatomical Site
BMD
DMPA Months
OC Months
Control Months
Model Baseline 6 12 Baseline 6 12 Baseline 6 12 p (n ⫽ 53) (n ⫽ 48) (n ⫽ 29) (n ⫽ 165) (n ⫽ 129) (n ⫽ 79) (n ⫽ 152) (n ⫽ 138) (n ⫽ 107) Valueb
0.977 L1-L4 Spine (g/cm2) (SD) (0.094)
0.969 (0.089)
1.009 (0.102)
1.019 (0.091)
⫺1.414 (3.524)
% Change L1-L4 spine (SD)
0.930 Femoral (0.130) neck (g/cm2) (SD) % Change femoral neck (SD)
0.960 (0.074)
0.904 (0.123)
0.894 (0.095)
1.024 (0.071)
0.984 (0.098)
1.008 (0.094)
2.338 (3.780)
0.956 (0.128)
0.953 (0.124)
⫺2.172 (4.773)
0.953 (0.097) 0.269 (5.122)
1.021 (0.082)
3.837 (3.809)
0.923 (0.131)
0.928 (0.126)
0.939 (0.114) 2.280 (5.148)
Tukey’s Adjusted p Values Baseline-12 monthsc
⬍.001 Control vs. DMPA (p ⫽ .006) DMPA vs. OC (p ⬍ .003) ⬍.001 Control vs. DMPA (p ⬍ .001) DMPA vs. OC (p ⬍ .001) Control vs. OC (p ⫽ .03) ⬍.001 ⬍.001 Control vs. DMPA (p ⬍ .001) Control vs. OC (p ⫽ .03)
a
Adjusted for race, chronological age, and body weight. Model p values represent significance from the interaction effect of time and method for absolute data and from the method effect for the percent change data. c Only significant comparisons reported at p ⬍ .05. b
were considered for inclusion in the models with 0.25 as entry criterion. Regarding adjustments of p values, the nonindependence of repeated measurements on the same individuals over time was accounted for within the modeling of the repeated measures ANCOVA and within the percentage change of the ANCOVA. Within each of the models, post hoc pair-wise comparisons were done using Tukey’s adjustment for multiple comparisons. Final results were considered statistically significant at the .05 level. All analyses were conducted using SAS version 8.2 (SAS Institute Inc., Cary, NC).
Results At baseline, the study population included 370 adolescent girls who self-selected either DMPA (n ⫽ 53) or OC (n ⫽ 165) as their contraceptive method, or who were abstinent or using a nonhormonal form of contraception (n ⫽ 152). Descriptive baseline data are presented in Table 1. The control group was significantly younger in chronologic age and gynecologic age than the other two treatment groups. The OC group had significantly higher body weight and body mass index than the other two groups. No
statistically significant differences were found among the groups regarding race, percent having regular menses, number of cigarettes smoked per month, or physical activity levels. At the time of this analysis, of the 370 subjects, 215 subjects had completed a 12-month study visit including lumbar spine and femoral neck DEXA scans. There were 29 in the DMPA group, 79 in the OC group, and 107 in the control group. The remaining 155 subjects either withdrew from the study before their 12-month observation (12 in the DMPA group, 55 in the OC group and 20 in the control group) or were still enrolled in the study, but did not yet have their 12-month observation (12 in the DMPA group, 31 in the OC group and 25 in the control group). There were no significant differences in baseline characteristics between these subjects and those who completed their 12-month observation. The DMPA group compliance was 100%; the OC group compliance was 86% (95% CI 82% to 88%). Table 2 presents lumbar spine and femoral neck BMD in both mean absolute values at baseline, 6 months, and 12 months and mean percent change from baseline to 12 months. All BMD values were adjusted for race, chronological age, and body
CROMER ET AL
438
JOURNAL OF ADOLESCENT HEALTH Vol. 35, No. 6
Table 3. Adjusteda Means and Standard Deviations (SD) of Absolute and Percent Change for L1-L4 Spine BMAD and Femoral Neck BMAD Anatomical Site
BMAD
DMPA Months
OC Months
Control Months
Model Baseline 6 12 Baseline 6 12 Baseline 6 12 p (n ⫽ 53) (n ⫽ 48) (n ⫽ 29) (n ⫽ 165) (n ⫽ 129) (n ⫽ 79) (n ⫽ 152) (n ⫽ 138) (n ⫽ 107) Valueb
0.154 L2-L4 Spine (g/cm2) (SD) (0.014)
0.152 (0.014)
0.159 (0.013)
0.160 (0.011)
⫺1.280 (4.175)
% Change L2-L4 spine (SD) 0.199 Femoral (0.036) neck (g/cm2) (SD) % Change femoral neck (SD)
0.151 (0.011)
0.193 (0.034)
0.191 (0.032)
0.160 (0.009)
0.156 (0.012)
0.159 (0.012)
1.419 (4.478)
0.207 (0.038)
0.208 (0.034)
⫺0.870 (8.112)
0.206 (0.027)
0.198 (0.037)
0.632 (8.708)
0.197 (0.035)
0.160 (0.010)
⬍.001
2.954 (4.499)
⬍.001
0.200 (0.031)
⬍.001
1.907 (8.741)
Tukey’s Adjusted p Values Baseline-12 monthsc Control vs. DMPA (p ⬍ .02) DMPA vs. OC (p ⬍ .02) Control vs. DMPA (p ⬍ .001) DMPA vs. OC (p ⫽ .012)
.1940
a
Adjusted for race, chronological age, and body weight. Model p values represent significance from the interaction effect of time and method for absolute data and from the method effect for the percent change data. c Only significant comparisons reported at p ⬍ .05. b
weight. The mean absolute values for lumbar spine in the DMPA group decreased ⫺0.02 g/cm2, increased 0.02 g/cm2 in the OC group, and increased 0.04 g/cm2 in the control group. At 12 months, the mean absolute value for spine BMD was significantly lower in the DMPA group than in the OC group (p ⬍ .003) and the control group (p ⫽ .006). The mean percent change in lumbar spine was ⫺1.4%, 2.3%, and 3.8% for the DMPA, OC and control groups, respectively. In mean percent change, spine BMD was significantly lower in the DMPA group than in both the OC group and the control group (p ⬍ .001). Additionally, the mean percent change in the OC group was significantly lower than in the control group (p ⫽ .03). The mean BMD absolute values at the femoral neck in the DMPA group decreased ⫺0.04 g/cm2, decreased ⫺0.003 g/cm2 in the OC group, and increased 0.02 g/cm2 in the control group. No significant differences were found when comparing mean absolute values among the three groups. The mean percent change from baseline to 12 months was ⫺2.2%, 0.3%, and 2.3% for the DMPA, OC, and control groups, respectively. Mean percent change in femoral neck BMD was significantly lower in the DMPA group than in the control group (p ⬍ .001) and significantly lower in the OC group the in the control group (p ⫽ .03).
BMD results were recalculated correcting for possible volumetric changes in bone and were expressed as BMAD. These results are presented in Table 3. The mean absolute value at the lumbar spine was significantly lower in the DMPA group than in either the OC or the control group (p ⬍ .02). The mean percent change in lumbar spine from baseline to 12 months was significantly lower in the DMPA group than in the OC group (p ⫽ .012) and the control group (p ⬍ .001). There were no statistically significant differences in BMAD at the femoral neck.
Discussion The major question generated from the findings in this study is whether two currently prescribed methods of contraception, DMPA and an OC containing 20 g ethinyl estradiol/100 g levonorgestrel, exert a deleterious effect on the skeleton in the adolescent girl. The preponderance of research examining the relationship of hormonal contraception to bone health has been conducted in adult pre-menopausal women of widely varying ages. The implied assumption from these studies is that the results can be generalized to all pre-menopausal women, including adolescents. However, this study challenges that assumption. There is a vast difference in bone
December 2004
growth and metabolism between a 15- and a 35-yearold woman. For example, a 35-year-old woman typically experiences steady, small amounts of loss in bone mass; in contrast, a 15-year-old girl undergoes large increases in bone mass, between 2% and 10% per year, resulting in a 40% total increase during her second decade of life [1]. Therefore, any losses experienced before 18 years of age have potentially greater impact on bone health when compared with the same amount of loss seen in the older adult woman. Our data indicate a loss in BMD at both the spine (mean percent change ⫺1.4%) and femoral neck (mean percent change ⫺2.2%) in girls using DMPA over 12 months. This finding was in contrast to significant increases in BMD at both anatomic sites (mean percent change 3.8% and 2.3% at the spine and femoral neck, respectively) in the adolescent girls who were abstinent or using nonhormonal methods. Previous data in three longitudinal studies on the effects of DMPA on BMD within the adolescent age group were consistent with ours; however, in one cross-sectional study, nonsignificant differences were found between adolescent DMPA users and nonusers [15]. The results across the longitudinal studies were as follows: BMD changes at the lumbar spine (the only anatomic site measured) in the DMPA group ranged from ⫺1.8% to ⫺3.5% over 12 months [8,16,17] compared with mean increases of 1.2% to 2.9% BMD in the control group [8,16]. None of these studies calculated their data using the BMAD formula. Among 18 –30-year-olds followed over 1 year, Berenson et al found a similar decrease to that in our study in mean lumbar spine BMD (⫺2.7%) in the DMPA group; in contrast, they found a ⫺0.4% mean decrease in mean lumbar spine BMD in their control group [18]. Although the degree of bone loss was similar to that found in the younger age group, the discrepancy between the DMPA and control groups was smaller in this older age group. Equally important is the relationship between OC and bone mass in the adolescent age group. The dose of ethinyl estradiol in OC has been lowered over the past 30 years owing to a clinical concern regarding the dose-response relationship between estrogen and thromboembolic phenomena. However, the risk for such adverse events is extremely low in healthy adolescents. Yet, very little attention has been directed toward establishing the adequate dose of estrogen for optimal bone growth and consolidation before attainment of peak bone mass for women in their teens and 20s.
BONE HEALTH IN YOUNG WOMEN
439
This study reports results of BMD in adolescents using an OC containing 20 g ethinyl estradiol/100 g levonorgestrel. We found a smaller mean gain over 12 months in percent change of lumbar spine BMD in the OC group (2.3%) and femoral neck (0.3%), than in the control group (3.8% and 2.3%, respectively). The balance of evidence to date indicates that OC do not exert an adverse effect on bone. In their meta-analysis, Kuohung et al reported either a positive effect or no effect on BMD in 13 studies of “low-dose” (range 20 – 40 g) OC [19]. However, none of these studies included adolescents and only one examined an OC containing 20 g ethinyl estradiol [20]. The only other published study specifically evaluating an OC containing 20 g ethinyl estradiol was conducted in young adult women (age range 19 –22 years) and found that BMD remained unchanged in the OC users but increased 7.8% in a comparison group of non-OC users [21]. Other studies, focusing on OC containing 30 – 40 g ethinyl estradiol in adolescents [8,22] and young adults [18,23–25], produced results consistent with those summarized by Kuohung, i.e., either a positive effect [18,23] or no effect on bone mineral density [8,22,24,25]. In summary, there is consistent evidence that OC containing at least 30 g ethinyl estradiol are adequate for maintenance of bone health in the young adult woman and there is preliminary evidence that this dose is sufficient for bone accrual in the adolescent. We present, in this article, findings that suggest that an OC containing 20 g ethinyl estradiol may not be enough for optimal bone mass acquisition in the very young woman who is still undergoing skeletal maturation. The crucial issue of “recovery” of bone after discontinuation of hormonal contraception has received limited examination. In the two studies published to date dealing directly with this issue [26,27], there was evidence of at least partial recovery over time. However, Scholes et al, comparing a cohort of women ranging in age from 18 to 39 years, who either discontinued DMPA or never used it, found that bone mineral density among the 18 –21-year-olds, over a mean of 15 months, had not reached that of the untreated group [27]. No study has addressed prospectively the response of bone after discontinuation of oral contraceptives; however, retrospective work suggests that additional factors, such as physical activity and nutrition, contribute to ultimate bone health in the adult woman [28]. These same factors may also be important in mitigating the losses in bone density experienced by women using DMPA [29].
440
CROMER ET AL
The strengths of this study include (a) the large numbers of study participants across the treatment groups, (b) the longitudinal design that provides a comparison of BMD measurements over time, and (c) the inclusion of very young adolescents (i.e., ⬍ gynecologic age 3, typically 12–14-year-olds) in whom bone development is particularly active. The limitations relate to (a) a high degree of attrition, particularly in the OC group, (b) significant differences in BMD, chronological/gynecologic age, and body weight in the OC group, vs. the other two groups, and (c) the results based on an interim analysis. However, three findings mitigate the importance of these limitations: (a) no significant differences were found in background variables between subjects who withdrew from the study and those who completed their 12-month observation, (b) we adjusted all BMD results by age, body weight, and racial background, and (c) no significant differences were found in background variables between subjects who had not yet reached their 12-month observation, who represented a small percentage (18%) of the total sample, and those who completed their 12-month observation. In conclusion, our study contributes to an increasing body of knowledge indicating a negative impact of DMPA on bone health in females. Although somewhat less clear, our findings suggest a potential adverse effect of an OC containing 20 g ethinyl estradiol and 100 g levonorgestrel on bone health in the same population. Attention needs to be focused on designing the optimal formulations of hormonal contraception that provide both effective contraception as well as stimulation for bone accrual toward acquisition of optimal peak bone mass. This work was supported by NIH grant R01HD39009 and General Clinical Research Center Grant M1RR00080I2. We thank Ray Harvey, B.S., Kelly Camlin Shingler, M.S.S.A, Rachel Whitsel, B.A., Mary Jo Day, L.P.N., Darlene Lewis, R.N., and Judy Schwartzbaum, Ph.D., for their help in preparation of this manuscript.
References 1. Sabatier JP, Guadyier-Souquiese G, Benmalek A, Macelli C. Evolution of lumbar bone mineral content during adolescence and adulthood: A longitudinal study in 395 healthy females 10 –24 years of age and 206 premenopausal women. Osteoporos Int 1999;476 – 82. 2. Compston JE. Sex steriods and bone. Physiol Rev 2001;81:419 – 47. 3. Alan Guttmacher Institute. Occasional Report: Why is Teenage Pregnancy Declining? The Roles of Abstinence, Sexual Activity and Contraceptive Use. National Survey of Family Growth 1988 and 1995 Cycles. New York, NY: Alan Guttmacher Institute, 1999.
JOURNAL OF ADOLESCENT HEALTH Vol. 35, No. 6
4. Contraceptives and teens. What are the options? Contracept Technol Update 2000;21:109 –11. 5. Kaufmann RB, Spitz AM, Strauss LT, et al. The decline in US teen pregnancy rates, 1990 –1995. Pediatrics 1998;102:1141–7. 6. Cromer BA, Smith RD, Blair JM, et al. A prospective study of adolescents who choose among levonorgestrel implant (Norplant), medroxyprogesterone acetate (Depo-Provera), or the combined oral contraceptive pill as contraception. Pediatrics 1994;94:687–94. 7. Nelson A. Merits of DMPA relative to other reversible contraceptive methods. J Reprod Med 2002;47:781– 4. 8. Cromer BA, Blair JM, Mahan JD, et al. Prospective comparison of bone density in adolescent girls receiving depot medroxyprogesterone acetate (Depo-Provera), levonorgestrel (Norplant), or oral contraceptives. J Pediatr 1996;129:671– 6. 9. Trends in oral contraceptive development and utilization. looking to the future. Contracept Rep 1997;7:4 –13. 10. Nappi C, Di Spiezio Sardo A, Acunzo G, et al. Effects of a low-dose and ultra-low-dose combine oral contraceptive use on bone turnover and BMD in young fertile women: A prospective randomized study. Contraception 2003;67:355–9. 11. Cromer BA, Lazebnik R, Rome E, et al. Double-blinded randomized controlled trial of estrogen supplementation in adolescent girls receiving depot medroxyprogesterone acetate for contraception. Am J Obstet Gynecol 2004, in press. 12. Grunbaum JA, Kann L, Kinchen SA, et al. Youth risk behavior surveillance—United States, 2001. J Sch Health 2002;72:313–28. 13. Hestzler AA, Trary RB. A dietary calcium rapid assessment method (RAM). Innovations in practice: A dietary calcium RAM. In Top Clin Nutr 1994;9:76 – 85. 14. Katzman DK, Bachrach LK, Carter DR, Marcus R. Clinical and anthropometric correlates of bone mineral acquisition in healthy adolescent girls. J Clin Endocrinol Metab 1991;73: 1332–9. 15. Scholes D, LaCroix AZ, Ichikawa LE, et al. The association between depot medroxyprogesterone acetate contraception and bone mineral density in adolescent women. Contraception 2004;69:99 –104. 16. Lara-Torre E, Edwards CP, Perlman S, Hertweck SP. Bone mineral density in adolescent females using depot medroxyprogesterone acetate. J Pediatr Adolesc Gynecol 2004;17:17–21. 17. Busen NH, Britt RB, Rianon N. Bone mineral density in a cohort of adolescent women using depot medroxyprogesterone acetate for one to two years. J Adolesc Health 2003;32: 257–9. 18. Berenson AB, Radecki CM, Grady JJ, et al. A prospective, controlled study of the effects of hormonal contraception on bone mineral density. Obstet Gynecol 2001;98:576 – 82. 19. Kuohung W, Borgatta L, Stubblefield P. Low-dose oral contraceptives and bone mineral density: an evidence-based analysis. Contraception 2000;61:77– 82. 20. Mais V, Fruzetti F, Ajossa S, et al. Bone metabolism in young women taking a monophasic pill containing 20 g ethinyl estradiol. Contraception 1993;48:445–52. 21. Polatti F, Perotti F, Filippa N, et al. Bone mass and long-term monophasic oral contraceptive treatment in young women. Contraception 1995;51:221– 4. 22. Lloyd T, Taylor DS, Lin HM, et al. Oral contraceptive use by teenage women does not affect peak bone mass: A longitudinal study. Fertil Steril 2000;74:734 – 8. 23. MacDougall J, Davies M, Overton CE, et al. Bone density in a population of long term oral contraceptive pill users does not differ from that in menstruating women. Br J Fam Plann 1999;25:96 –100.
December 2004
24. Recker RR, Davies M, Hinders SM, et al. Bone gain in young adult women. JAMA 1992;268:2403– 8. 25. Reed SD, Scholes D, LaCroix AZ, et al. Longitudinal changes in bone mineral density in relation to oral contraceptive use. Contraception 2003;68:177– 82. 26. Cundy T, Cornish J, Evans MC, et al. Recovery of bone density in women who stop using medroxyprogesterone acetate. BMJ 1993;308:242– 8.
BONE HEALTH IN YOUNG WOMEN
441
27. Scholes D, LaCroix AZ, Ichikawa LE, et al. Injectable hormonal contraception and bone density: Results from a prospective study. Epidemiology 2002;13:581–7. 28. Pasco JA, Katowicz MA, Henry MJ, et al. Oral contraceptives and bone mineral density: A population-based study. Am J Obstet Gynecol 2000;182:265–9. 29. Westhoff C. Bone mineral density and DMPA. J Reprod Med 2002;47:795–9.
ORIGINAL ARTICLE
Depot Medroxyprogesterone Acetate, Oral Contraceptives and Bone Mineral Density in a Cohort of Adolescent Girls BARBARA A. CROMER, M.D., MARGARET STAGER, M.D., ANDREA BONNY, M.D., RINA LAZEBNIK, M.D., ELLEN ROME, M.D., M.P.H., JULIE ZIEGLER, M.A., AND SARA M. DEBANNE, Ph.D.
Purpose: To conduct a longitudinal comparison of bone mineral density (BMD) in 370 adolescent girls, aged 12–18, who self-selected depot medroxyprogesterone acetate (DMPA) or an oral contraceptive (OC) containing 20g ethinyl estradiol/100g levonorgestrel with that in girls who received no hormonal treatment (control group). Methods: Lumbar spine and femoral neck BMD measurements were obtained by dual energy x-ray absorptiometry at baseline and 12 months. Data were analyzed with repeated measures analysis of covariance methods. Results: Over 12 months, lumbar spine BMD decreased in the DMPA group (n ⴝ 29), with a mean percent change of ⴚ1.4% (95% confidence interval [CI] ⴚ2.73, ⴚ0.10), and increased by a mean of 3.8% (95% CI 3.11, 4.57) in the control group [n ⴝ 107 (p < .001)]. The increase in mean percent change in lumbar spine BMD in the OC group (n ⴝ 79), 2.3% (95% CI 1.49, 3.18), was significantly smaller than in the control group (p ⴝ .03). Over 12 months, the mean percent change in femoral neck BMD was ⴚ2.2% (95% CI ⴚ3.95, ⴚ0.39) in the DMPA group, but increased 2.3% (95% CI 1.29, 3.27) in the control group (p < .001). The increase in mean percent change at the femoral neck in the OC group, 0.3% (95% CI ⴚ0.87, 1.41), was significantly lower than in the control group (p ⴝ .03).
From the MetroHealth Medical Center; Rainbow Babies & Children’s Hospital, University Hospitals of Cleveland; The Children’s Hospital at the Cleveland Clinic; and Department of Epidemiology and Biostatistics, Case Western Reserve University School of Medicine, Cleveland, Ohio . Address correspondence to: Dr. Barbara Cromer, Department of Pediatrics, MetroHealth Medical Center, 2500 MetroHealth Dr., Cleveland, OH 44109. E-mail: [email protected] Manuscript accepted July 29, 2004. 1054-139X/04/$–see front matter doi:10.1016/j.jadohealth.2004.07.005
Conclusions: Our study contributes to an increasing body of knowledge indicating a negative impact of DMPA on bone health in young women. Additional findings suggest a potential adverse effect of an OC containing 20 g ethinyl estradiol/100 g levonorgestrel on bone health in adolescents. © Society for Adolescent Medicine, 2004
KEY WORDS:
Bone mineral density Adolescents Depot medroxyprogesterone acetate Oral contraceptives
Adolescence is a crucial period for skeletal growth [1]. Sex hormones, especially estrogen, play a central role in bone development [2]; therefore, alterations in their circulating levels have a significant impact on bone and future peak bone mass. The question that emerges is whether use of exogenous sex hormones, in the form of hormonal contraception, affect bone health in the adolescent. The two most commonly used forms of hormonal contraception by adolescent girls in this country are oral contraceptives (OC) and depot medroxyprogesterone acetate (DMPA) [3]. Since its approval by the FDA as a contraceptive in 1992, DMPA use by young women has been credited, in part, with the decrease in adolescent pregnancy rates observed over the past decade [4,5]. Overall, the method has been shown to be very effective, safe and to have few adverse © Society for Adolescent Medicine, 2004 Published by Elsevier Inc., 360 Park Avenue South, New York, NY 10010
December 2004
clinical effects [6,7]. A previous study, however, demonstrated bone loss in a small group of adolescent DMPA users [8]. The goal of our study was to investigate whether these previous findings were replicable in a larger adolescent population. Over the past 30 years, secular trends in product development in the pharmaceutical industry resulted in a progressive lowering of estrogen dose in OC from over 150 g to 20 g ethinyl estradiol [9], and in the most recent formulation, to 15 g ethinyl estradiol (with 60 g gestodene) [10]. No published data are available regarding the relationship between OC containing less than 30 g ethinyl estradiol and bone in the adolescent population. The second goal of our study was to examine whether bone mineral density (BMD) increased in adolescent girls using an OC with 20 g ethinyl estradiol/100 g levonorgestrel at a rate consonant with that seen in untreated, healthy adolescent girls. In December 2003, before completion of study enrollment, we performed an interim analysis of our concurrent clinic trial [11] at the request of our Data Safety Monitoring Board. Because of the highly significant results in the clinical trial, the Institutional Review Board requested a hold on recruitment of subjects initiating DMPA. Therefore, we froze the data set as of this date and elected to report our findings from the interim analysis. The data set included subjects who completed a 12-month observation, subjects who withdrew before their 12-month observation, and subjects who had not yet reached their 12-month observation.
Materials and Methods This study was designed as a 24-month, prospective examination of BMD and hormonal contraception in adolescents. This report presents results from a 12month interim analysis. Postmenarcheal girls, ranging from 12-18 years of age, were recruited from four general adolescent health clinics located in a large, urban, metropolitan setting. Specifically, adolescent girls requesting contraception, and selecting either DMPA or OC, were eligible for enrollment. Concurrently, girls attending the same clinics, who planned to receive no treatment with sex hormones, were eligible for enrollment. Exclusion criteria included use of DMPA, pregnancy, or abortion over the past 6 months, use of oral contraceptives over the past 3 months, chronic medical condition or treatment that may have an effect on BMD, or need for confidentiality in contraception management. The Institutional
BONE HEALTH IN YOUNG WOMEN
435
Review Boards of all participating institutions approved the study. Written, informed consent and assent were obtained from each custodial parent and study enrollee, respectively. At baseline, 6, and 12 months, clinical and behavioral information was obtained from each study participant. Tobacco use was reported as number of cigarettes per day over the past 30 days, as it is reported on the Youth Risk Behavior Survey [12]. Calcium intake was elicited with a focused 24-hour dietary recall combined with the Calcium Rapid Assessment Method [13]. Girls who consumed ⬍ 1200 mg/day dietary calcium were counseled by a dietician; if the level of intake did not improve after 3 months, the subject was given a sample of Tums® 500 mg to be taken once daily for 3 months. Physical activity was assessed using a survey that asked each subject to classify herself as very inactive, inactive, normal, active, and very active. Clinical information was collected by direct interview and included menstrual bleeding patterns, general medical symptoms, and medication use. Height and weight were measured using the same stadiometer (Easy Glide Bearing stature board) and Mettler-Toledo scale every visit. Gynecologic age was defined as the number of years since menarche. At baseline, 6, and 12 months, all study participants underwent measurement of BMD that included L1–L4 lumbar vertebrae, total hip (left), femoral neck, trochanter, and Ward’s triangle. The lumbar vertebrae were chosen because of the high percentage of trabecular bone and the sensitivity of lumbar spine BMD to hormone-related changes in bone mass. The hip was also chosen because osteoporosis fractures at this site are associated with the greatest morbidity later in life. The measurement technique employed was dual energy x-ray absorptiometry (DEXA) with the model QDR 4500W fanbeam densitometer (Hologic Inc., Bedford, MA). Windows 11.2 was the software, which included a low-density measurement option. Technicians received training from Hologic and were certified for machine operation by the Ohio Department of Health. Calibration protocol for the DEXA machine included a daily phantom reading and a weekly check for drift by a radiologist. In vivo intra-individual coefficients of variation were 1.2% at the spine and 1.4% at the femoral neck; inter-individual coefficients of variation were 1.3% at the spine and 2.2% at the femoral neck. All scans were obtained within 4 weeks of the scheduled 6-month intervals. Because longitudinal growth typically does not cease altogether until late adolescence, skeletal size
CROMER ET AL
436
JOURNAL OF ADOLESCENT HEALTH Vol. 35, No. 6
Table 1. Characteristics of Study Participants at Baselinea Characteristics
DMPA (n ⫽ 53)
OC (n ⫽ 165)
Control (n ⫽ 152)
Age in years Gynecological age in years Race, n (% black) Body Weight (kg) Height in (cm) Body Mass Index Spine BMD Femoral Neck BMD Vitamin D Regular Menses, n (%) Number of Cigarettes per Month n (%) Physical Activityh Active, n (%) Normal, n (%) Inactive, n (%)
15.6 ⫾ 1.5 3.7 ⫾ 1.6 34 (64) 60.5 ⫾ 13.0d 161.7 ⫾ 6.9 23.1 ⫾ 4.2e 0.98 ⫾ 0.09 0.92 ⫾ 0.14 57.1 ⫾ 25.8 46 (87) 39.7 ⫾ 97.7 43 (81)
16.0 ⫾ 1.4 4.0 ⫾ 1.6 103 (62) 69.1 ⫾ 17.3 161.8 ⫾ 6.4 26.4 ⫾ 6.4 1.03 ⫾ 0.11f 0.98 ⫾ 0.16g 54.4 ⫾ 30.0 133 (81) 32.8 ⫾ 92.8 119 (72)
14.8 ⫾ 1.5b 2.6 ⫾ 1.6c 95 (63) 63.2 ⫾ 16.9 160.9 ⫾ 6.6 24.3 ⫾ 5.8 0.98 ⫾ 0.12 0.92 ⫾ 0.14 55.6 ⫾ 29.9 123 (81) 8.7 ⫾ 50.2 117 (77)
14 (29) 27 (55) 8 (16)
67 (43) 78 (50) 10 (7)
69 (48) 62 (43) 13 (9)
All numbers represent mean ⫾ SD except: Race, Regular Menses, and Physical Activity, which are reported as percentages. Chronological Age is significantly lower in control group than DMPA group (p ⬍ .001) and OC group (p ⬍ .001). Gynecological Age is significantly lower in control group than DMPA group (p ⬍ .001) and OC group (p ⬍ .001). d Body Weight is significantly higher in OC group than DMPA group (p ⫽ .001) and control group (p ⬍ .002). e Body Mass Index is significantly higher in OC group than DMPA group (p ⬍ .001) and control group (p ⫽ .002). f Spine BMD Spine is significantly higher in OC group than DMPA group (p ⬍ .001) and control group (p ⬍ .001). g Femoral Neck BMD is significantly higher in OC group than DMPA group (p ⫽ .002) and control group (p ⫽ .002). h Missing data: DMPA group (n ⫽ 4), OC group (n ⫽ 10), control group (n ⫽ 8). a
b c
may have changed in our study subjects over the observation period. As bones grow and increase in width and height, the bone thickness increases, and BMD can be confounded by changes in thickness, i.e., increased bone thickness may falsely appear as increased bone density. BMD was calculated by amount of scanned bone mineral content (BMC) within a projected area (Ap), termed areal density, and is expressed as g/cm2. Therefore, to correct for volumetric variations in bone, we performed an additional calculation to achieve bone mineral apparent density (BMAD), utilizing the following formula: spine BMAD ⫽ BMC (L1–L4) ⫼ Ap3/2 and femoral neck BMAD ⫽ BMC(femoral neck) ⫼ Ap2(femoral neck) [14]. DMPA was administered as a 150-mg deep intramuscular (gluteus or triceps) injection every 12 weeks. All girls in this study received an OC containing 20 g ethinyl estradiol and 100 g levonorgestrel. Compliance with DMPA injections was assessed by chart review and was calculated as number of injections divided by number of prescribed injections (1 every 3 months) ⫻ 100. Compliance rates with OC use was assessed by monthly self-report and calculated as the number of pills taken divided by number of pills prescribed (1 every day) ⫻ 100. Girls who chose to switch their contraceptive method during the 12-month observation period were withdrawn from the study.
The association of changes in both BMD and BMAD over time was assessed in two ways. First, repeated measures analysis of covariance (ANCOVA) was used, where the time of the visit (baseline, 6 months, and 12 months) was used as a factor in the model. The approach allows full utilization of all available follow-up data. Linear modeling is justified because the time intervals are only 6 months in length. ANCOVA was also used to examine the annual percentage change, where the time factor was incorporated into the endpoint by calculating a percentage change in BMD or BMAD from baseline to 12 months. Before modeling, the distributional properties of continuous variables were examined for possible departures from normality. Multi-collinearity of predictors was also examined. The high correlation between chronologic age and gynecologic age (Spearman correlation rho ⫽ 0.69, p ⬍ .001, n ⫽ 370), and between body weight and BMI (rho ⫽ 0.94, p ⬍ .001, n ⫽ 370) led to the decision to use only one from each pair in the multivariate models. Because three baseline characteristics, i.e., chronological age, weight, and race are powerful predictors of BMD [6], they were treated as confounding variables and were included in all models, whether significant or not. Regarding physical activity, tobacco use, and vitamin D status, these were not significantly related to BMD and, therefore, were not included in the models. Second order interaction terms between treatment type and other covariates
December 2004
BONE HEALTH IN YOUNG WOMEN
437
Table 2. Adjusteda Means and Standard Deviations (SD) of Absolute and Percent Change for L1-L4 Spine BMD and Femoral Neck BMD Anatomical Site
BMD
DMPA Months
OC Months
Control Months
Model Baseline 6 12 Baseline 6 12 Baseline 6 12 p (n ⫽ 53) (n ⫽ 48) (n ⫽ 29) (n ⫽ 165) (n ⫽ 129) (n ⫽ 79) (n ⫽ 152) (n ⫽ 138) (n ⫽ 107) Valueb
0.977 L1-L4 Spine (g/cm2) (SD) (0.094)
0.969 (0.089)
1.009 (0.102)
1.019 (0.091)
⫺1.414 (3.524)
% Change L1-L4 spine (SD)
0.930 Femoral (0.130) neck (g/cm2) (SD) % Change femoral neck (SD)
0.960 (0.074)
0.904 (0.123)
0.894 (0.095)
1.024 (0.071)
0.984 (0.098)
1.008 (0.094)
2.338 (3.780)
0.956 (0.128)
0.953 (0.124)
⫺2.172 (4.773)
0.953 (0.097) 0.269 (5.122)
1.021 (0.082)
3.837 (3.809)
0.923 (0.131)
0.928 (0.126)
0.939 (0.114) 2.280 (5.148)
Tukey’s Adjusted p Values Baseline-12 monthsc
⬍.001 Control vs. DMPA (p ⫽ .006) DMPA vs. OC (p ⬍ .003) ⬍.001 Control vs. DMPA (p ⬍ .001) DMPA vs. OC (p ⬍ .001) Control vs. OC (p ⫽ .03) ⬍.001 ⬍.001 Control vs. DMPA (p ⬍ .001) Control vs. OC (p ⫽ .03)
a
Adjusted for race, chronological age, and body weight. Model p values represent significance from the interaction effect of time and method for absolute data and from the method effect for the percent change data. c Only significant comparisons reported at p ⬍ .05. b
were considered for inclusion in the models with 0.25 as entry criterion. Regarding adjustments of p values, the nonindependence of repeated measurements on the same individuals over time was accounted for within the modeling of the repeated measures ANCOVA and within the percentage change of the ANCOVA. Within each of the models, post hoc pair-wise comparisons were done using Tukey’s adjustment for multiple comparisons. Final results were considered statistically significant at the .05 level. All analyses were conducted using SAS version 8.2 (SAS Institute Inc., Cary, NC).
Results At baseline, the study population included 370 adolescent girls who self-selected either DMPA (n ⫽ 53) or OC (n ⫽ 165) as their contraceptive method, or who were abstinent or using a nonhormonal form of contraception (n ⫽ 152). Descriptive baseline data are presented in Table 1. The control group was significantly younger in chronologic age and gynecologic age than the other two treatment groups. The OC group had significantly higher body weight and body mass index than the other two groups. No
statistically significant differences were found among the groups regarding race, percent having regular menses, number of cigarettes smoked per month, or physical activity levels. At the time of this analysis, of the 370 subjects, 215 subjects had completed a 12-month study visit including lumbar spine and femoral neck DEXA scans. There were 29 in the DMPA group, 79 in the OC group, and 107 in the control group. The remaining 155 subjects either withdrew from the study before their 12-month observation (12 in the DMPA group, 55 in the OC group and 20 in the control group) or were still enrolled in the study, but did not yet have their 12-month observation (12 in the DMPA group, 31 in the OC group and 25 in the control group). There were no significant differences in baseline characteristics between these subjects and those who completed their 12-month observation. The DMPA group compliance was 100%; the OC group compliance was 86% (95% CI 82% to 88%). Table 2 presents lumbar spine and femoral neck BMD in both mean absolute values at baseline, 6 months, and 12 months and mean percent change from baseline to 12 months. All BMD values were adjusted for race, chronological age, and body
CROMER ET AL
438
JOURNAL OF ADOLESCENT HEALTH Vol. 35, No. 6
Table 3. Adjusteda Means and Standard Deviations (SD) of Absolute and Percent Change for L1-L4 Spine BMAD and Femoral Neck BMAD Anatomical Site
BMAD
DMPA Months
OC Months
Control Months
Model Baseline 6 12 Baseline 6 12 Baseline 6 12 p (n ⫽ 53) (n ⫽ 48) (n ⫽ 29) (n ⫽ 165) (n ⫽ 129) (n ⫽ 79) (n ⫽ 152) (n ⫽ 138) (n ⫽ 107) Valueb
0.154 L2-L4 Spine (g/cm2) (SD) (0.014)
0.152 (0.014)
0.159 (0.013)
0.160 (0.011)
⫺1.280 (4.175)
% Change L2-L4 spine (SD) 0.199 Femoral (0.036) neck (g/cm2) (SD) % Change femoral neck (SD)
0.151 (0.011)
0.193 (0.034)
0.191 (0.032)
0.160 (0.009)
0.156 (0.012)
0.159 (0.012)
1.419 (4.478)
0.207 (0.038)
0.208 (0.034)
⫺0.870 (8.112)
0.206 (0.027)
0.198 (0.037)
0.632 (8.708)
0.197 (0.035)
0.160 (0.010)
⬍.001
2.954 (4.499)
⬍.001
0.200 (0.031)
⬍.001
1.907 (8.741)
Tukey’s Adjusted p Values Baseline-12 monthsc Control vs. DMPA (p ⬍ .02) DMPA vs. OC (p ⬍ .02) Control vs. DMPA (p ⬍ .001) DMPA vs. OC (p ⫽ .012)
.1940
a
Adjusted for race, chronological age, and body weight. Model p values represent significance from the interaction effect of time and method for absolute data and from the method effect for the percent change data. c Only significant comparisons reported at p ⬍ .05. b
weight. The mean absolute values for lumbar spine in the DMPA group decreased ⫺0.02 g/cm2, increased 0.02 g/cm2 in the OC group, and increased 0.04 g/cm2 in the control group. At 12 months, the mean absolute value for spine BMD was significantly lower in the DMPA group than in the OC group (p ⬍ .003) and the control group (p ⫽ .006). The mean percent change in lumbar spine was ⫺1.4%, 2.3%, and 3.8% for the DMPA, OC and control groups, respectively. In mean percent change, spine BMD was significantly lower in the DMPA group than in both the OC group and the control group (p ⬍ .001). Additionally, the mean percent change in the OC group was significantly lower than in the control group (p ⫽ .03). The mean BMD absolute values at the femoral neck in the DMPA group decreased ⫺0.04 g/cm2, decreased ⫺0.003 g/cm2 in the OC group, and increased 0.02 g/cm2 in the control group. No significant differences were found when comparing mean absolute values among the three groups. The mean percent change from baseline to 12 months was ⫺2.2%, 0.3%, and 2.3% for the DMPA, OC, and control groups, respectively. Mean percent change in femoral neck BMD was significantly lower in the DMPA group than in the control group (p ⬍ .001) and significantly lower in the OC group the in the control group (p ⫽ .03).
BMD results were recalculated correcting for possible volumetric changes in bone and were expressed as BMAD. These results are presented in Table 3. The mean absolute value at the lumbar spine was significantly lower in the DMPA group than in either the OC or the control group (p ⬍ .02). The mean percent change in lumbar spine from baseline to 12 months was significantly lower in the DMPA group than in the OC group (p ⫽ .012) and the control group (p ⬍ .001). There were no statistically significant differences in BMAD at the femoral neck.
Discussion The major question generated from the findings in this study is whether two currently prescribed methods of contraception, DMPA and an OC containing 20 g ethinyl estradiol/100 g levonorgestrel, exert a deleterious effect on the skeleton in the adolescent girl. The preponderance of research examining the relationship of hormonal contraception to bone health has been conducted in adult pre-menopausal women of widely varying ages. The implied assumption from these studies is that the results can be generalized to all pre-menopausal women, including adolescents. However, this study challenges that assumption. There is a vast difference in bone
December 2004
growth and metabolism between a 15- and a 35-yearold woman. For example, a 35-year-old woman typically experiences steady, small amounts of loss in bone mass; in contrast, a 15-year-old girl undergoes large increases in bone mass, between 2% and 10% per year, resulting in a 40% total increase during her second decade of life [1]. Therefore, any losses experienced before 18 years of age have potentially greater impact on bone health when compared with the same amount of loss seen in the older adult woman. Our data indicate a loss in BMD at both the spine (mean percent change ⫺1.4%) and femoral neck (mean percent change ⫺2.2%) in girls using DMPA over 12 months. This finding was in contrast to significant increases in BMD at both anatomic sites (mean percent change 3.8% and 2.3% at the spine and femoral neck, respectively) in the adolescent girls who were abstinent or using nonhormonal methods. Previous data in three longitudinal studies on the effects of DMPA on BMD within the adolescent age group were consistent with ours; however, in one cross-sectional study, nonsignificant differences were found between adolescent DMPA users and nonusers [15]. The results across the longitudinal studies were as follows: BMD changes at the lumbar spine (the only anatomic site measured) in the DMPA group ranged from ⫺1.8% to ⫺3.5% over 12 months [8,16,17] compared with mean increases of 1.2% to 2.9% BMD in the control group [8,16]. None of these studies calculated their data using the BMAD formula. Among 18 –30-year-olds followed over 1 year, Berenson et al found a similar decrease to that in our study in mean lumbar spine BMD (⫺2.7%) in the DMPA group; in contrast, they found a ⫺0.4% mean decrease in mean lumbar spine BMD in their control group [18]. Although the degree of bone loss was similar to that found in the younger age group, the discrepancy between the DMPA and control groups was smaller in this older age group. Equally important is the relationship between OC and bone mass in the adolescent age group. The dose of ethinyl estradiol in OC has been lowered over the past 30 years owing to a clinical concern regarding the dose-response relationship between estrogen and thromboembolic phenomena. However, the risk for such adverse events is extremely low in healthy adolescents. Yet, very little attention has been directed toward establishing the adequate dose of estrogen for optimal bone growth and consolidation before attainment of peak bone mass for women in their teens and 20s.
BONE HEALTH IN YOUNG WOMEN
439
This study reports results of BMD in adolescents using an OC containing 20 g ethinyl estradiol/100 g levonorgestrel. We found a smaller mean gain over 12 months in percent change of lumbar spine BMD in the OC group (2.3%) and femoral neck (0.3%), than in the control group (3.8% and 2.3%, respectively). The balance of evidence to date indicates that OC do not exert an adverse effect on bone. In their meta-analysis, Kuohung et al reported either a positive effect or no effect on BMD in 13 studies of “low-dose” (range 20 – 40 g) OC [19]. However, none of these studies included adolescents and only one examined an OC containing 20 g ethinyl estradiol [20]. The only other published study specifically evaluating an OC containing 20 g ethinyl estradiol was conducted in young adult women (age range 19 –22 years) and found that BMD remained unchanged in the OC users but increased 7.8% in a comparison group of non-OC users [21]. Other studies, focusing on OC containing 30 – 40 g ethinyl estradiol in adolescents [8,22] and young adults [18,23–25], produced results consistent with those summarized by Kuohung, i.e., either a positive effect [18,23] or no effect on bone mineral density [8,22,24,25]. In summary, there is consistent evidence that OC containing at least 30 g ethinyl estradiol are adequate for maintenance of bone health in the young adult woman and there is preliminary evidence that this dose is sufficient for bone accrual in the adolescent. We present, in this article, findings that suggest that an OC containing 20 g ethinyl estradiol may not be enough for optimal bone mass acquisition in the very young woman who is still undergoing skeletal maturation. The crucial issue of “recovery” of bone after discontinuation of hormonal contraception has received limited examination. In the two studies published to date dealing directly with this issue [26,27], there was evidence of at least partial recovery over time. However, Scholes et al, comparing a cohort of women ranging in age from 18 to 39 years, who either discontinued DMPA or never used it, found that bone mineral density among the 18 –21-year-olds, over a mean of 15 months, had not reached that of the untreated group [27]. No study has addressed prospectively the response of bone after discontinuation of oral contraceptives; however, retrospective work suggests that additional factors, such as physical activity and nutrition, contribute to ultimate bone health in the adult woman [28]. These same factors may also be important in mitigating the losses in bone density experienced by women using DMPA [29].
440
CROMER ET AL
The strengths of this study include (a) the large numbers of study participants across the treatment groups, (b) the longitudinal design that provides a comparison of BMD measurements over time, and (c) the inclusion of very young adolescents (i.e., ⬍ gynecologic age 3, typically 12–14-year-olds) in whom bone development is particularly active. The limitations relate to (a) a high degree of attrition, particularly in the OC group, (b) significant differences in BMD, chronological/gynecologic age, and body weight in the OC group, vs. the other two groups, and (c) the results based on an interim analysis. However, three findings mitigate the importance of these limitations: (a) no significant differences were found in background variables between subjects who withdrew from the study and those who completed their 12-month observation, (b) we adjusted all BMD results by age, body weight, and racial background, and (c) no significant differences were found in background variables between subjects who had not yet reached their 12-month observation, who represented a small percentage (18%) of the total sample, and those who completed their 12-month observation. In conclusion, our study contributes to an increasing body of knowledge indicating a negative impact of DMPA on bone health in females. Although somewhat less clear, our findings suggest a potential adverse effect of an OC containing 20 g ethinyl estradiol and 100 g levonorgestrel on bone health in the same population. Attention needs to be focused on designing the optimal formulations of hormonal contraception that provide both effective contraception as well as stimulation for bone accrual toward acquisition of optimal peak bone mass. This work was supported by NIH grant R01HD39009 and General Clinical Research Center Grant M1RR00080I2. We thank Ray Harvey, B.S., Kelly Camlin Shingler, M.S.S.A, Rachel Whitsel, B.A., Mary Jo Day, L.P.N., Darlene Lewis, R.N., and Judy Schwartzbaum, Ph.D., for their help in preparation of this manuscript.
References 1. Sabatier JP, Guadyier-Souquiese G, Benmalek A, Macelli C. Evolution of lumbar bone mineral content during adolescence and adulthood: A longitudinal study in 395 healthy females 10 –24 years of age and 206 premenopausal women. Osteoporos Int 1999;476 – 82. 2. Compston JE. Sex steriods and bone. Physiol Rev 2001;81:419 – 47. 3. Alan Guttmacher Institute. Occasional Report: Why is Teenage Pregnancy Declining? The Roles of Abstinence, Sexual Activity and Contraceptive Use. National Survey of Family Growth 1988 and 1995 Cycles. New York, NY: Alan Guttmacher Institute, 1999.
JOURNAL OF ADOLESCENT HEALTH Vol. 35, No. 6
4. Contraceptives and teens. What are the options? Contracept Technol Update 2000;21:109 –11. 5. Kaufmann RB, Spitz AM, Strauss LT, et al. The decline in US teen pregnancy rates, 1990 –1995. Pediatrics 1998;102:1141–7. 6. Cromer BA, Smith RD, Blair JM, et al. A prospective study of adolescents who choose among levonorgestrel implant (Norplant), medroxyprogesterone acetate (Depo-Provera), or the combined oral contraceptive pill as contraception. Pediatrics 1994;94:687–94. 7. Nelson A. Merits of DMPA relative to other reversible contraceptive methods. J Reprod Med 2002;47:781– 4. 8. Cromer BA, Blair JM, Mahan JD, et al. Prospective comparison of bone density in adolescent girls receiving depot medroxyprogesterone acetate (Depo-Provera), levonorgestrel (Norplant), or oral contraceptives. J Pediatr 1996;129:671– 6. 9. Trends in oral contraceptive development and utilization. looking to the future. Contracept Rep 1997;7:4 –13. 10. Nappi C, Di Spiezio Sardo A, Acunzo G, et al. Effects of a low-dose and ultra-low-dose combine oral contraceptive use on bone turnover and BMD in young fertile women: A prospective randomized study. Contraception 2003;67:355–9. 11. Cromer BA, Lazebnik R, Rome E, et al. Double-blinded randomized controlled trial of estrogen supplementation in adolescent girls receiving depot medroxyprogesterone acetate for contraception. Am J Obstet Gynecol 2004, in press. 12. Grunbaum JA, Kann L, Kinchen SA, et al. Youth risk behavior surveillance—United States, 2001. J Sch Health 2002;72:313–28. 13. Hestzler AA, Trary RB. A dietary calcium rapid assessment method (RAM). Innovations in practice: A dietary calcium RAM. In Top Clin Nutr 1994;9:76 – 85. 14. Katzman DK, Bachrach LK, Carter DR, Marcus R. Clinical and anthropometric correlates of bone mineral acquisition in healthy adolescent girls. J Clin Endocrinol Metab 1991;73: 1332–9. 15. Scholes D, LaCroix AZ, Ichikawa LE, et al. The association between depot medroxyprogesterone acetate contraception and bone mineral density in adolescent women. Contraception 2004;69:99 –104. 16. Lara-Torre E, Edwards CP, Perlman S, Hertweck SP. Bone mineral density in adolescent females using depot medroxyprogesterone acetate. J Pediatr Adolesc Gynecol 2004;17:17–21. 17. Busen NH, Britt RB, Rianon N. Bone mineral density in a cohort of adolescent women using depot medroxyprogesterone acetate for one to two years. J Adolesc Health 2003;32: 257–9. 18. Berenson AB, Radecki CM, Grady JJ, et al. A prospective, controlled study of the effects of hormonal contraception on bone mineral density. Obstet Gynecol 2001;98:576 – 82. 19. Kuohung W, Borgatta L, Stubblefield P. Low-dose oral contraceptives and bone mineral density: an evidence-based analysis. Contraception 2000;61:77– 82. 20. Mais V, Fruzetti F, Ajossa S, et al. Bone metabolism in young women taking a monophasic pill containing 20 g ethinyl estradiol. Contraception 1993;48:445–52. 21. Polatti F, Perotti F, Filippa N, et al. Bone mass and long-term monophasic oral contraceptive treatment in young women. Contraception 1995;51:221– 4. 22. Lloyd T, Taylor DS, Lin HM, et al. Oral contraceptive use by teenage women does not affect peak bone mass: A longitudinal study. Fertil Steril 2000;74:734 – 8. 23. MacDougall J, Davies M, Overton CE, et al. Bone density in a population of long term oral contraceptive pill users does not differ from that in menstruating women. Br J Fam Plann 1999;25:96 –100.
December 2004
24. Recker RR, Davies M, Hinders SM, et al. Bone gain in young adult women. JAMA 1992;268:2403– 8. 25. Reed SD, Scholes D, LaCroix AZ, et al. Longitudinal changes in bone mineral density in relation to oral contraceptive use. Contraception 2003;68:177– 82. 26. Cundy T, Cornish J, Evans MC, et al. Recovery of bone density in women who stop using medroxyprogesterone acetate. BMJ 1993;308:242– 8.
BONE HEALTH IN YOUNG WOMEN
441
27. Scholes D, LaCroix AZ, Ichikawa LE, et al. Injectable hormonal contraception and bone density: Results from a prospective study. Epidemiology 2002;13:581–7. 28. Pasco JA, Katowicz MA, Henry MJ, et al. Oral contraceptives and bone mineral density: A population-based study. Am J Obstet Gynecol 2000;182:265–9. 29. Westhoff C. Bone mineral density and DMPA. J Reprod Med 2002;47:795–9.