- Email: [email protected]
Double-blinded randomized controlled trial of estrogen supplementation in adolescent girls who receive depot medroxyprogesterone acetate for contraception
American Journal of Obstetrics and Gynecology (2005) 192, 42e7
www.ajog.org
Double-blinded randomized controlled trial of estrogen supplementation in adolescent girls who receive depot medroxyprogesterone acetate for contraception Barbara A. Cromer, MD,a,b,* Rina Lazebnik, MD,a,c Ellen Rome, MD, MPH,a,d Margaret Stager, MD,a,b Andrea Bonny, MD,a,b Julie Ziegler, MA,a,b Sara M. Debanne, PhDa Case Western Reserve University School of Medicine,a MetroHealth Medical Center,b University Hospitals of Cleveland,c and The Cleveland Clinic,d Cleveland, Ohio Received for publication April 24, 2004; revised July 20, 2004; accepted July 20, 2004
KEY WORDS Depot medroxyprogesterone acetate Adolescent Bone mineral density Estrogen
Objective: The purpose of this clinical trial was to evaluate the effect of estrogen supplementation on bone mineral density in adolescent girls who received depot medroxyprogesterone acetate for contraception. Study design: One hundred twenty-three adolescents who began receiving depot medroxyprogesterone acetate injections every 12 weeks were assigned randomly to receive monthly injections of estradiol cypionate or placebo. The main outcome was bone mineral density that was measured by dual energy x-ray absorptiometry for 12 (n = 69) to 24 (n = 36) months. Participants, technicians, and physicians were blinded to estrogen treatment. Results: Over the 24-month period, the percentage of change from baseline bone mineral density at the lumbar spine was 2.8% in the estradiol cypionate group versus 1.8% in the placebo group (P !.001). At the femoral neck, the percentage of change from baseline bone mineral density was 4.7% in the estradiol cypionate group versus 5.1% in the placebo group (P !.001). Conclusion: Our results suggest that estrogen supplementation is protective of bone in adolescent girls who receive depot medroxyprogesterone acetate injections. Ó 2005 Elsevier Inc. All rights reserved.
Depot medroxyprogesterone acetate (DMPA) fulfills several criteria for an ideal birth control method. As an intramuscular injection every 3 months, it is very
Supported by National Institutes of Health grant R01HD39009 and General Clinical Research Center grant M1RR00080I2. * Reprint requests: Barbara Cromer, MD, Department of Pediatrics, MetroHealth Medical Center, 2500 MetroHealth Dr, Cleveland, OH 44109. E-mail: [email protected] 0002-9378/$ - see front matter Ó 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.ajog.2004.07.041
effective, convenient, and reversible.1 For these reasons, DMPA has become a popular contraceptive choice for adolescents in the United States since its approval by the Food and Drug Administration in 1992.2 It is estimated that more than 1 million American teens currently use DMPA for contraception.3,4 The mechanism of contraceptive action for DMPA is ovulation suppression; in addition, DMPA suppresses ovarian production of estrogen, which could affect bone mass. In support of this, cross-sectional studies in adult
Cromer et al women who use DMPA have shown that its long-term use is associated with a significant reduction in bone density.5,6 Prospective studies in adult women have also demonstrated that bone loss occurs after the initiation of DMPA and that bone mass can be regained when women stop using DMPA and become estrogen replete.7,8 Relatively little research has been directed to the examination of bone mass in adolescents who receive DMPA. A small number of prospective studies have demonstrated a loss of bone mass in adolescents on DMPA at a time when non-DMPA users were still gaining bone mass.9-11 This is a very significant concern because adverse effects on the skeleton in this age period could affect the acquisition of peak bone mass. Total bone mass normally increases by at least 50% during adolescence.12 Bone fragility and increased risk for fracture later in life may stem from decreased bone accrual during this critical period of development. If estrogen deficiency is the cause of bone loss that is associated with DMPA, then logically, treatment with estrogen concurrent with DMPA should prevent bone loss. The purpose of this double-blinded, randomized controlled clinical trial was to compare bone mineral density (BMD) in adolescent girls who receive DMPA plus monthly intramuscular injections of either estradiol cypionate or placebo over a 24-month period.
Material and methods The study population was drawn from 4 general adolescent health clinics that are located in a large, metropolitan setting between May 2000 and December 2002. Adolescent girls, who ranged in age from 12 to 18 years, who were seeking contraception, and who selected DMPA for their method of contraception were eligible for enrollment. Exclusion criteria included the use of DMPA, pregnancy or abortion over the past 6 months, the use of oral contraceptives over the past 3 months, a chronic medical condition or treatment that may have an effect on bone, or a need for confidentiality in contraception management. Written, informed consent and written assent were obtained from each custodial parent and study enrollee, respectively. This study received approval from the Institutional Review Boards at each recruitment site. The double-blinded, randomized, controlled clinical trial was conducted as follows: Patients were stratified by recruitment site (4 health clinics), race (black vs nonblack), and gynecologic age (ie, years since menarche [%3 and O3 years]). The rationale for this stratification is the well-known difference in BMD according to race and sexual maturity. Within each stratum, each subject entered the estrogen supplement arm or the placebo arm with the use of blocked randomization techniques. The treatments comprised monthly intra-
43 muscular injections of either 5 mg estradiol cypionate (supplement) or 5 mL normal saline solution (placebo) that was administered by a nurse who was not blinded. Those blinded to the treatment included the study subjects, the technicians who were conducting the dual energy x-ray absorptiometry scans, and the clinicians who were providing health care to the participants. DMPA was administered as a 150-mg deep intramuscular (gluteus or triceps) injection every 12 weeks. The level of compliance was calculated by chart review for DMPA and for estradiol cypionate and placebo as the number of injections divided by number of prescribed injections ! 100. Tobacco use was elicited as number of cigarettes per day over the past 30 days from a questionnaire that was derived from the Youth Risk Behavior Survey.13 Calcium intake was elicited with a focused 24-hour dietary recall combined with the Calcium Rapid Assessment Method.14 Each subject was also asked to choose with an overall physical activity level from the following levels: very inactive, inactive, active, and very active. Menstrual bleeding pattern information was collected by direct interview. Height and weight were measured with the same stadiometer (Easy Glide Bearing stature board; Portage, Mich) and a Mettler-Toledo scale (Worthington, Ohio), respectively. Serum 17b-estradiol was measured at the end of the DMPA injection interval with a double antibody radioimmunoassay kit (Diagnostics Products Corporation, Los Angeles, Calif). At baseline and 12 and 24 months, BMD measurements were obtained that included L1 to L4 lumbar vertebrae, total hip (left), femoral neck, trochanter, and Ward’s triangle. The measurement technique that was used was dual energy x-ray absorptiometry with a fan-beam densitometer (model QDR 4500W; Hologic Inc, Bedford, Mass). Technicians received training from Hologic Inc and were certified for machine operation by the Ohio Department of Health. Calibration protocol for the dual energy x-ray absorptiometry machine included a daily phantom reading and a weekly check for drift by a radiologist. In vivo intraindividual coefficients of variation were 1.2% at the spine and 1.4% at the femoral neck; interindividual 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 12-month intervals. Because longitudinal growth typically does not cease altogether until late adolescence, skeletal size 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 (ie, increased bone thickness may falsely appear as increased bone density). BMD was calculated by the amount of scanned bone mineral content (BMC) within a projected area (Ap), termed areal density, and is expressed as grams per
44
Cromer et al
Table I
Demographic characteristics at baseline
Demographic Chronologic age (y)* Gynecologic age (y)* Body weight (kg)* Mean body mass index (kg/m2)* Regular menses (n) Racial background: Black (n) Cigarettes per moy Physical activity (n) Inactive Active Very active Baseline BMD (g/cm2)* Spine Femoral neck
DMPA-E (n = 65)
DMPA-P (n = 58)
15.8 G 1.4 3.9 G 2.0 65.1 G 16.0 24.5 G 5.0
15.8 G 1.6 3.9 G 1.8 60.5 G 12.8 23.0 G 4.1
56 (86%) 40 (61%)
45 (76%) 37 (64%)
0 (0, 105)
0 (0, 4)
5 (8%) 32 (49%) 28 (43%)
9 (16%) 29 (50%) 20 (34%)
1.003 G 0.114 0.937 G 0.130
0.980 G 0.090 0.925 G 0.137
* Data are given as mean G SD. y Data are given as median (interquartile range).
square centimeter. Therefore, to correct for volumetric variations in bone, we performed an additional calculation to achieve bone mineral apparent density (BMAD; g/cm3), with the following formula15: BMAD ¼ BMC ðL1 -L4 OAp3=2 BMAD ¼ BMCðfemoral neckÞ OAp2ðfemoral neckÞ The association of changes over time in BMD or BMAD with group membership, controlling for possible effects because of potentially confounding variables (weight and gynecologic age at recruitment and race) was assessed in 2 ways: First, analysis of covariance was used, where the posttreatment value of BMD or BMAD was used as the end point, and the corresponding baseline value and demographics were used as covariates in the model. Second, analysis of covariance of the percentage of change in BMD or BMAD was examined, where the time factor was incorporated into the end point by the calculation of a percentage change from baseline. Final models retained only the significant demographic covariates. Before modeling, the distributional properties of continuous variables were examined. BMD and BMAD were normally distributed; estrogen was not. Its natural log was used in the analyses, because this transformation normalized the data.
Results At baseline, the study population comprised 123 adolescent girls who selected DMPA as their contraceptive
method; 65 girls were assigned randomly to the DMPAestrogen (DMPA-E) group, and 58 girls were assigned randomly to the DMPA-placebo (DMPA-P) group. In December 2003, the Data Safety Monitoring Board supervised interim analysis of the trial with the recommendation that, if our spine BMD findings at 12 months were significantly different between the 2 treatment groups at a probability value of !.001, the trial would be stopped. The results did reach that level of significance, and, with the guidance of the MetroHealth Medical Center Institutional Review Board, the trial was stopped. The protocol was instituted to (1)inform all subjects/guardians of the results of the trial, (2) offer estrogen supplementation to subjects who had received placebo, and (3) provide exercise and dietary counseling to all active participants. As a result of the early discontinuation of the trial, there were subjects who had not yet completed their period of observation at 12 (n = 24) and 24 months (n = 33). Also, there were subjects who withdrew from the study by 12 months (n = 30) and 24 months (n = 53). At 12 months, the numbers of the study participants were as follows: DMPA-E, 35 girls; DMPA-P, 34 girls. At 24 months, the numbers were as follows: DMPA-E, 19 girls; DMPA-P, 17 girls. No significant differences were found between the 2 groups at baseline on demographic, anthropometric, and behavioral characteristics (Table I). Compliance with DMPA injections was 93% G 17% for the DMPA-E group and 91% G 17% for the DMPA-P group. The DMPA-E compliance with monthly estradiol injections was 86% G 19%. The DMPA-P group compliance with monthly placebo injections was 91% G 13%. Table II shows the adjusted group means for BMD and BMAD at 12 months, respectively. Adjusted for body weight and baseline bone density, the absolute mean BMD value at the lumbar spine in the DMPA-E group, 1.088 g/cm2, was significantly higher than that seen in the DMPA-P group, 0.978 g/cm2 (P !.001). At the lumbar spine, the percent change from baseline was as follows: 1.3% the DMPA-E group versus 1.6% in the DMPA-P group (P !.001). At the femoral neck, the absolute mean value at 12 months was also significantly higher in the DMPA-E group, 0.934 g/cm2, than in the DMPA-P group, 0.904 g/cm2 (P =.001). The percent change from baseline was also significantly different between the 2 treatment groups: 1.0% increase for the DMPA-E group and 2.2% decrease for the DMPA-P group (P =.001). Trends in BMAD were similar to those observed in BMD; however, the comparisons of absolute values and percent change from baseline BMAD between the 2 treatment groups was statistically significant only at the spine (P !.001). Table III shows the adjusted group means for BMD and BMAD at 24 months, respectively. Adjusted for body weight and baseline bone density, the absolute
Cromer et al Table II
45
BMD and BMAD of the lumbar spine and femoral neck at 12 months for DMPA-E and DMPA-P
Anatomic site BMD Spine Absolute value (g/cm2) Change from baseline (%) Femoral neck Absolute value (g/cm2) Change from baseline (%) BMAD Spine Absolute value (g/cm2) Change from baseline (%) Femoral neck Absolute value (g/cm2) Change from baseline (%)
DMPA-E (n = 35)
DMPA-P (n = 34)
P value
1.008 G 0.0038 C1.3% G 0.39%
0.978 G 0.0039 1.6% G 0.40%
! .001 ! .001
0.934 G 0.0060 C1.0 G 0.66
0.904 G 0.0061 2.2 G 0.67
.001 .001
0.158 G 0.0008 C1.2 G 0.39
0.153 G 0.0008 1.5 G 0.40
! .001 ! .001
0.196 G 0.0022 0.3 G 1.15
0.193 G 0.0022 1.5 G 1.17
.38 .46
* Data are given as adjusted group means G SE: Adjusted for body weight and baseline bone density.
Table III
BMD and BMAD of the lumbar spine and femoral neck at 24 months for DMPA-E and DMPA-P
Anatomic site BMD Spine Absolute value (g/cm2) Change from baseline (%) Femoral neck Absolute value (g/cm2) Change from baseline (%) BMAD Spine Absolute value (g/cm2) Change from baseline (%) Femoral neck Absolute value (g/cm2) Change from baseline (%)
DMPA-E (n = 18)
DMPA-P (n = 16)
P value
1.019 G 0.0059 C2.8 G 0.61
0.970 G 0.0062 1.8 G 0.64
! .001 ! .001
0.964 G 0.0135 C4.7 G 1.69
0.878 G 0.0143 5.1 G 1.80
! .001 ! .001
0.159 G 0.0010 C3.0 G 0.63
0.152 G 0.0010 1.9 G 0.66
! .001 ! .001
0.202 G 0.0043 C4.0 G 2.31
0.189 G 0.0047 3.1 G 2.45
.04 .04
* Data are given as adjusted group means G SE: Adjusted for body weight and baseline bone density.
mean value at the lumbar spine BMD was significantly higher in the DMPA-E group, 1.019 g/cm2, than in the DMPA-P group, 0.970 g/cm2 (P !.001). The comparison in percent change from baseline BMD between the 2 groups was as follows: 2.8% in the DMPA-E group versus 1.8% in the DMPA-P group (P !.001). At the femoral neck, the mean absolute BMD value was significantly higher in the DMPA-E group, 0.964 g/ cm2, than in the DMPA-P group, 0.878 g/cm2 (P !.001). The comparison in percent change between the 2 groups was as follows: 4.7% in the DMPA-E group versus 5.1% in the DMPA-P group (P !.001). The pattern of findings was similar after adjusting for changes in bone size; however, the results at the femoral neck BMAD were marginally significant (P =.04). At 12 months, the mean estradiol levels were significantly higher in the DMPA-P group, 36.2 G 1.09 pg/ mL (132.0 G 4.00 pmol/L) than in the DMPA-E group, 26.6 G 1.09 pg/mL (97.6 G 4.00 pmol/L; P =.01).
Results were similar at 24 months, with mean estradiol levels in the DMPA-P group at 36.6 G 1.09 pg/mL (134.3 G 4.00 pmol/L) and in the DMPA-E group at 26.8 G 1.08 pg/mL (98.4 G 4.00 pmol/L; P =.01). No significant relationship was found between estradiol levels and BMD or BMAD in either treatment group.
Comment In this randomized clinical trial, we found that adolescents who were receiving DMPA-E had significantly higher BMD at both the lumbar spine and femoral neck at 12 and 24 months compared with that seen in girls who were receiving DMPA-P. When the data were adjusted for volumetric changes in bone size, the findings followed a similar pattern at the lumbar spine at 12 and 24 months and the femoral neck at 24 months. The only other randomized controlled clinical trial of
46 estrogen supplementation in DMPA users was conducted by Cundy et al16 in adult premenopausal women, with a mean age of 37 years. Thirty-eight women were assigned randomly to receive orally either conjugated estrogens (0.625 mg) or placebo daily over 24 months. At the lumbar spine, BMD increased by 1.0% over 24 months in the estrogen group, whereas, in the placebo group, BMD decreased 2.6% (P !.01). BMD trends at the femoral neck and total body were in the same direction but had less statistical significance. Thus, estrogen appeared to prevent the bone loss that was seen in the DMPA placebo group. This study provides strong corroborating evidence for our primary hypothesis (ie, that estrogen deficiency is the cause of bone loss in premenopausal women who receive DMPA). One important feature that distinguishes adolescents from adult women is that adolescents are normally in a dynamic period of skeletal development; the vast majority of bone mass is accrued before the age of 20 years. From the onset of puberty, around age 11 years, until its completion, usually age 15 years (about gynecologic age 3), total BMC increases by 50%.12 Therefore, bone loss in an adolescent who receives DMPA suggests not only the loss but also a concomitant failure of bone acquisition. The clinical question that emerges from our study is whether estrogen supplementation should be recommended routinely in adolescents who are receiving DMPA. We suggest that this change in clinical practice would be premature for multiple reasons. First, there is the issue of potential increases in BMD after the discontinuation of DMPA. It may be that the bone loss seen in users of DMPA is a temporary phenomenon and recovery of bone mass occurs quickly after the cessation of treatment. One analogy would be breastfeeding, a period marked by dramatic bone loss; however, as shown by Sowers et al17 and Karlsson et al,18 there is total return of bone mass within several months after breastfeeding has been stopped. Two studies to date have examined the response of bone after the discontinuation of DMPA, and both studies found at least partial recovery within 12 to 30 months after the cessation of the method.7,8 Of note, in the 1 study that included young women,7 the increase in BMD was largest in the youngest age group (18-21 year) after discontinuation of DMPA; however, BMD values did not reach that of control subjects. Thus, we cannot conclude that observed bone loss is irreversible until further study is conducted and the relevant data have been obtained. Our study indicated that bone loss in the girls who received DMPA-P is consistent with other prospective studies of bone and DMPA in adolescents.9-11 However, this apparent adverse effect must be weighed against the considerable benefits that are associated with DMPA as a contraceptive.19 DMPA has features
Cromer et al that are particularly appealing to teens: it is highly effective; it obviates the need for daily compliance with a pill, and it can be used privately.20 The use of DMPA has been credited in part with the decrease in teen pregnancy in this country over the past decade. There also may be important environmental factors (such as nutrition and exercise) that ultimately contribute more substantively to bone health after menopause than the early use of DMPA, particularly given the short duration of DMPA use that is seen typically in this population.21 The challenge then that faces the clinician who is providing contraceptive care is to balance the risk versus the benefit ratio of the use DMPA in a very young woman. The benefit from the very high contraceptive efficacy of DMPA is considerable and would appear to outweigh the risk that is associated with bone loss in adolescents. We failed to find a significant relationship between serum estradiol levels and BMD after 12 months of treatment with either DMPA-P or DMPA-E. Other previous investigations have reported similar findings.5 The reason for this pattern of findings is unclear. However, given the demonstrated variability in serum estrogen levels during treatment with DMPA,22 a single serum estradiol level may not be an accurate reflection of tissue levels of estrogen throughout the treatment cycle. A significant limitation of this study was the low numbers of girls who were available for follow-up. Thus, there is no guarantee that the group that was measured at either 12 or 24 months was representative of the group that was measured at baseline. A source of bias may have been introduced by a coincidence of differing behavioral attributes between those girls who withdrew from the study, those girls who were still active (and did not complete the study protocol due to early discontinuation of trial), and those girls who completed the study. However, we assessed the major attributes that were relevant to BMD and found no differences among these groups. In conclusion, we found that BMD was significantly higher over a 24-month period in adolescent girls who received DMPA plus 5-mg monthly injections of estradiol cypionate compared with girls who received DMPA plus received monthly injections of placebo. Research efforts must address the degree of bone recovery after DMPA use in adolescents and the effects of nutrition and exercise on BMD, to help define the best clinical practice regarding estrogen supplementation in adolescents who use DMPA for contraception.
Acknowledgments We thank the following team members for their exceptional efforts: Kelly Camlin-Shingler, MSSA, Ray Harvey, BS, Mary Jo Day, LPN, Darlene Lewis, RN, Rachel Whitsel, BA.
Cromer et al
References 1. Kaunitz AM. Current concepts regarding use of DMPA. J Reprod Med 2002;47(suppl):785-9. 2. Creinin MD, Cullins VE, Emans SJ, Adams Hillard PJ, Nelson AL, Poindexter NA, et al. Implants and injectables: recent developments. The Contraception Report 2000;10:26-30. 3. Piccinino LJ, Mosher WD. Trends in contraceptive use in the United States: 1982-1995. Fam Plann Perspect 1998;30:4-10. 4. US Census Bureau. Census 2000 summary file 1. Washington (DC): US Department of Commerce; Economics and Statistics Bureau; 2000. 5. Cundy T, Evans M, Roberts. Bone mineral density in women receiving depot medroxyprogesterone acetate for contraception. BMJ 1991;303:13-9. 6. Scholes D, Lacroix AZ, Ott SM, Ichikawa IE, Barlow WE. Bone mineral density in women using depot medroxyprogesterone acetate for contraception. Obstet Gynecol 1999;93:253-8. 7. Scholes D, Lacroix AZ, Ichikawa LE, et al. Injectable hormone contraception and bone density: results from a prospective study. Epidemiology 2002;13:581-7. 8. Cundy T, Cornish J, Evans MC, Roberts H, Reid IR. Recovery of bone mineral density in women who stop using depot medroxyprogesterone acetate. BMJ 1993;308:247-8. 9. Cromer BA, McArdle BJ, Mahan JD, Zibners L, Naumovski Z. A 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. 10. Busen NH, Britt NB, Reanin 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. 11. 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.
47 12. Sabatier JP, Guaydier-Souquieres G, Benmalek A, Marcelli C. Evolution of bone mineral content during adolescence and adulthood: a longitudinal study in 295 healthy females 10-24 years of age and 206 premenopausal women. Osteoporos Int 1999;9:476-82. 13. Grunbaum JA, Kann L, Kinchen SA, Williams B, Ross JG, Lowry R, et al. Youth risk behavior surveillance: United States 2001. J School Health 2002;72:313-28. 14. Hertzler AA, Trary RB. A dietary calcium rapid assessment method (RAM): innovations in practice: a dietary calcium RAM. Top Clin Nutr 1994;9:76-85. 15. 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. 16. Cundy T, Ames R, Horne A, Clearwater J, Roberts H, Gamble G, et al. A randomized controlled trial of estrogen replacement therapy in long-term users of depot medroxyprogesterone acetate. 2003;88:78-81. 17. Sowers M, Corton G, Shapiro B, Jannausch ML, Crutchfield M, Smith ML, et al. Changes in bone density with lactation. JAMA 1993;269:3130-5. 18. Karlsson C, Obrant KJ, Karlsson M. Pregnancy and lactation confer reversible bone loss in humans. Osteoporos Int 2001;12: 828-34. 19. Westhoff C. Bone mineral density and DMPA. J Reprod Med 2002;61:77-82. 20. Cromer BA, Smith RD, McArdle JM, Dwyer J, Brown RT. 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. 21. Zibners A, Cromer BA, Hayes J. Comparison of continuation rates for hormonal contraception among adolescents. J Pediatr Adolesc Gynecol 1999;12:90-4. 22. Clark MK, Sowers M, Levy BT, Tenhundfeld P. Magnitude and variability of sequential estradiol and progesterone concentrations in women using depot medroxyprogesterone acetate for contraception. Fertil Steril 2001;75:871-7.
www.ajog.org
Double-blinded randomized controlled trial of estrogen supplementation in adolescent girls who receive depot medroxyprogesterone acetate for contraception Barbara A. Cromer, MD,a,b,* Rina Lazebnik, MD,a,c Ellen Rome, MD, MPH,a,d Margaret Stager, MD,a,b Andrea Bonny, MD,a,b Julie Ziegler, MA,a,b Sara M. Debanne, PhDa Case Western Reserve University School of Medicine,a MetroHealth Medical Center,b University Hospitals of Cleveland,c and The Cleveland Clinic,d Cleveland, Ohio Received for publication April 24, 2004; revised July 20, 2004; accepted July 20, 2004
KEY WORDS Depot medroxyprogesterone acetate Adolescent Bone mineral density Estrogen
Objective: The purpose of this clinical trial was to evaluate the effect of estrogen supplementation on bone mineral density in adolescent girls who received depot medroxyprogesterone acetate for contraception. Study design: One hundred twenty-three adolescents who began receiving depot medroxyprogesterone acetate injections every 12 weeks were assigned randomly to receive monthly injections of estradiol cypionate or placebo. The main outcome was bone mineral density that was measured by dual energy x-ray absorptiometry for 12 (n = 69) to 24 (n = 36) months. Participants, technicians, and physicians were blinded to estrogen treatment. Results: Over the 24-month period, the percentage of change from baseline bone mineral density at the lumbar spine was 2.8% in the estradiol cypionate group versus 1.8% in the placebo group (P !.001). At the femoral neck, the percentage of change from baseline bone mineral density was 4.7% in the estradiol cypionate group versus 5.1% in the placebo group (P !.001). Conclusion: Our results suggest that estrogen supplementation is protective of bone in adolescent girls who receive depot medroxyprogesterone acetate injections. Ó 2005 Elsevier Inc. All rights reserved.
Depot medroxyprogesterone acetate (DMPA) fulfills several criteria for an ideal birth control method. As an intramuscular injection every 3 months, it is very
Supported by National Institutes of Health grant R01HD39009 and General Clinical Research Center grant M1RR00080I2. * Reprint requests: Barbara Cromer, MD, Department of Pediatrics, MetroHealth Medical Center, 2500 MetroHealth Dr, Cleveland, OH 44109. E-mail: [email protected] 0002-9378/$ - see front matter Ó 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.ajog.2004.07.041
effective, convenient, and reversible.1 For these reasons, DMPA has become a popular contraceptive choice for adolescents in the United States since its approval by the Food and Drug Administration in 1992.2 It is estimated that more than 1 million American teens currently use DMPA for contraception.3,4 The mechanism of contraceptive action for DMPA is ovulation suppression; in addition, DMPA suppresses ovarian production of estrogen, which could affect bone mass. In support of this, cross-sectional studies in adult
Cromer et al women who use DMPA have shown that its long-term use is associated with a significant reduction in bone density.5,6 Prospective studies in adult women have also demonstrated that bone loss occurs after the initiation of DMPA and that bone mass can be regained when women stop using DMPA and become estrogen replete.7,8 Relatively little research has been directed to the examination of bone mass in adolescents who receive DMPA. A small number of prospective studies have demonstrated a loss of bone mass in adolescents on DMPA at a time when non-DMPA users were still gaining bone mass.9-11 This is a very significant concern because adverse effects on the skeleton in this age period could affect the acquisition of peak bone mass. Total bone mass normally increases by at least 50% during adolescence.12 Bone fragility and increased risk for fracture later in life may stem from decreased bone accrual during this critical period of development. If estrogen deficiency is the cause of bone loss that is associated with DMPA, then logically, treatment with estrogen concurrent with DMPA should prevent bone loss. The purpose of this double-blinded, randomized controlled clinical trial was to compare bone mineral density (BMD) in adolescent girls who receive DMPA plus monthly intramuscular injections of either estradiol cypionate or placebo over a 24-month period.
Material and methods The study population was drawn from 4 general adolescent health clinics that are located in a large, metropolitan setting between May 2000 and December 2002. Adolescent girls, who ranged in age from 12 to 18 years, who were seeking contraception, and who selected DMPA for their method of contraception were eligible for enrollment. Exclusion criteria included the use of DMPA, pregnancy or abortion over the past 6 months, the use of oral contraceptives over the past 3 months, a chronic medical condition or treatment that may have an effect on bone, or a need for confidentiality in contraception management. Written, informed consent and written assent were obtained from each custodial parent and study enrollee, respectively. This study received approval from the Institutional Review Boards at each recruitment site. The double-blinded, randomized, controlled clinical trial was conducted as follows: Patients were stratified by recruitment site (4 health clinics), race (black vs nonblack), and gynecologic age (ie, years since menarche [%3 and O3 years]). The rationale for this stratification is the well-known difference in BMD according to race and sexual maturity. Within each stratum, each subject entered the estrogen supplement arm or the placebo arm with the use of blocked randomization techniques. The treatments comprised monthly intra-
43 muscular injections of either 5 mg estradiol cypionate (supplement) or 5 mL normal saline solution (placebo) that was administered by a nurse who was not blinded. Those blinded to the treatment included the study subjects, the technicians who were conducting the dual energy x-ray absorptiometry scans, and the clinicians who were providing health care to the participants. DMPA was administered as a 150-mg deep intramuscular (gluteus or triceps) injection every 12 weeks. The level of compliance was calculated by chart review for DMPA and for estradiol cypionate and placebo as the number of injections divided by number of prescribed injections ! 100. Tobacco use was elicited as number of cigarettes per day over the past 30 days from a questionnaire that was derived from the Youth Risk Behavior Survey.13 Calcium intake was elicited with a focused 24-hour dietary recall combined with the Calcium Rapid Assessment Method.14 Each subject was also asked to choose with an overall physical activity level from the following levels: very inactive, inactive, active, and very active. Menstrual bleeding pattern information was collected by direct interview. Height and weight were measured with the same stadiometer (Easy Glide Bearing stature board; Portage, Mich) and a Mettler-Toledo scale (Worthington, Ohio), respectively. Serum 17b-estradiol was measured at the end of the DMPA injection interval with a double antibody radioimmunoassay kit (Diagnostics Products Corporation, Los Angeles, Calif). At baseline and 12 and 24 months, BMD measurements were obtained that included L1 to L4 lumbar vertebrae, total hip (left), femoral neck, trochanter, and Ward’s triangle. The measurement technique that was used was dual energy x-ray absorptiometry with a fan-beam densitometer (model QDR 4500W; Hologic Inc, Bedford, Mass). Technicians received training from Hologic Inc and were certified for machine operation by the Ohio Department of Health. Calibration protocol for the dual energy x-ray absorptiometry machine included a daily phantom reading and a weekly check for drift by a radiologist. In vivo intraindividual coefficients of variation were 1.2% at the spine and 1.4% at the femoral neck; interindividual 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 12-month intervals. Because longitudinal growth typically does not cease altogether until late adolescence, skeletal size 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 (ie, increased bone thickness may falsely appear as increased bone density). BMD was calculated by the amount of scanned bone mineral content (BMC) within a projected area (Ap), termed areal density, and is expressed as grams per
44
Cromer et al
Table I
Demographic characteristics at baseline
Demographic Chronologic age (y)* Gynecologic age (y)* Body weight (kg)* Mean body mass index (kg/m2)* Regular menses (n) Racial background: Black (n) Cigarettes per moy Physical activity (n) Inactive Active Very active Baseline BMD (g/cm2)* Spine Femoral neck
DMPA-E (n = 65)
DMPA-P (n = 58)
15.8 G 1.4 3.9 G 2.0 65.1 G 16.0 24.5 G 5.0
15.8 G 1.6 3.9 G 1.8 60.5 G 12.8 23.0 G 4.1
56 (86%) 40 (61%)
45 (76%) 37 (64%)
0 (0, 105)
0 (0, 4)
5 (8%) 32 (49%) 28 (43%)
9 (16%) 29 (50%) 20 (34%)
1.003 G 0.114 0.937 G 0.130
0.980 G 0.090 0.925 G 0.137
* Data are given as mean G SD. y Data are given as median (interquartile range).
square centimeter. Therefore, to correct for volumetric variations in bone, we performed an additional calculation to achieve bone mineral apparent density (BMAD; g/cm3), with the following formula15: BMAD ¼ BMC ðL1 -L4 OAp3=2 BMAD ¼ BMCðfemoral neckÞ OAp2ðfemoral neckÞ The association of changes over time in BMD or BMAD with group membership, controlling for possible effects because of potentially confounding variables (weight and gynecologic age at recruitment and race) was assessed in 2 ways: First, analysis of covariance was used, where the posttreatment value of BMD or BMAD was used as the end point, and the corresponding baseline value and demographics were used as covariates in the model. Second, analysis of covariance of the percentage of change in BMD or BMAD was examined, where the time factor was incorporated into the end point by the calculation of a percentage change from baseline. Final models retained only the significant demographic covariates. Before modeling, the distributional properties of continuous variables were examined. BMD and BMAD were normally distributed; estrogen was not. Its natural log was used in the analyses, because this transformation normalized the data.
Results At baseline, the study population comprised 123 adolescent girls who selected DMPA as their contraceptive
method; 65 girls were assigned randomly to the DMPAestrogen (DMPA-E) group, and 58 girls were assigned randomly to the DMPA-placebo (DMPA-P) group. In December 2003, the Data Safety Monitoring Board supervised interim analysis of the trial with the recommendation that, if our spine BMD findings at 12 months were significantly different between the 2 treatment groups at a probability value of !.001, the trial would be stopped. The results did reach that level of significance, and, with the guidance of the MetroHealth Medical Center Institutional Review Board, the trial was stopped. The protocol was instituted to (1)inform all subjects/guardians of the results of the trial, (2) offer estrogen supplementation to subjects who had received placebo, and (3) provide exercise and dietary counseling to all active participants. As a result of the early discontinuation of the trial, there were subjects who had not yet completed their period of observation at 12 (n = 24) and 24 months (n = 33). Also, there were subjects who withdrew from the study by 12 months (n = 30) and 24 months (n = 53). At 12 months, the numbers of the study participants were as follows: DMPA-E, 35 girls; DMPA-P, 34 girls. At 24 months, the numbers were as follows: DMPA-E, 19 girls; DMPA-P, 17 girls. No significant differences were found between the 2 groups at baseline on demographic, anthropometric, and behavioral characteristics (Table I). Compliance with DMPA injections was 93% G 17% for the DMPA-E group and 91% G 17% for the DMPA-P group. The DMPA-E compliance with monthly estradiol injections was 86% G 19%. The DMPA-P group compliance with monthly placebo injections was 91% G 13%. Table II shows the adjusted group means for BMD and BMAD at 12 months, respectively. Adjusted for body weight and baseline bone density, the absolute mean BMD value at the lumbar spine in the DMPA-E group, 1.088 g/cm2, was significantly higher than that seen in the DMPA-P group, 0.978 g/cm2 (P !.001). At the lumbar spine, the percent change from baseline was as follows: 1.3% the DMPA-E group versus 1.6% in the DMPA-P group (P !.001). At the femoral neck, the absolute mean value at 12 months was also significantly higher in the DMPA-E group, 0.934 g/cm2, than in the DMPA-P group, 0.904 g/cm2 (P =.001). The percent change from baseline was also significantly different between the 2 treatment groups: 1.0% increase for the DMPA-E group and 2.2% decrease for the DMPA-P group (P =.001). Trends in BMAD were similar to those observed in BMD; however, the comparisons of absolute values and percent change from baseline BMAD between the 2 treatment groups was statistically significant only at the spine (P !.001). Table III shows the adjusted group means for BMD and BMAD at 24 months, respectively. Adjusted for body weight and baseline bone density, the absolute
Cromer et al Table II
45
BMD and BMAD of the lumbar spine and femoral neck at 12 months for DMPA-E and DMPA-P
Anatomic site BMD Spine Absolute value (g/cm2) Change from baseline (%) Femoral neck Absolute value (g/cm2) Change from baseline (%) BMAD Spine Absolute value (g/cm2) Change from baseline (%) Femoral neck Absolute value (g/cm2) Change from baseline (%)
DMPA-E (n = 35)
DMPA-P (n = 34)
P value
1.008 G 0.0038 C1.3% G 0.39%
0.978 G 0.0039 1.6% G 0.40%
! .001 ! .001
0.934 G 0.0060 C1.0 G 0.66
0.904 G 0.0061 2.2 G 0.67
.001 .001
0.158 G 0.0008 C1.2 G 0.39
0.153 G 0.0008 1.5 G 0.40
! .001 ! .001
0.196 G 0.0022 0.3 G 1.15
0.193 G 0.0022 1.5 G 1.17
.38 .46
* Data are given as adjusted group means G SE: Adjusted for body weight and baseline bone density.
Table III
BMD and BMAD of the lumbar spine and femoral neck at 24 months for DMPA-E and DMPA-P
Anatomic site BMD Spine Absolute value (g/cm2) Change from baseline (%) Femoral neck Absolute value (g/cm2) Change from baseline (%) BMAD Spine Absolute value (g/cm2) Change from baseline (%) Femoral neck Absolute value (g/cm2) Change from baseline (%)
DMPA-E (n = 18)
DMPA-P (n = 16)
P value
1.019 G 0.0059 C2.8 G 0.61
0.970 G 0.0062 1.8 G 0.64
! .001 ! .001
0.964 G 0.0135 C4.7 G 1.69
0.878 G 0.0143 5.1 G 1.80
! .001 ! .001
0.159 G 0.0010 C3.0 G 0.63
0.152 G 0.0010 1.9 G 0.66
! .001 ! .001
0.202 G 0.0043 C4.0 G 2.31
0.189 G 0.0047 3.1 G 2.45
.04 .04
* Data are given as adjusted group means G SE: Adjusted for body weight and baseline bone density.
mean value at the lumbar spine BMD was significantly higher in the DMPA-E group, 1.019 g/cm2, than in the DMPA-P group, 0.970 g/cm2 (P !.001). The comparison in percent change from baseline BMD between the 2 groups was as follows: 2.8% in the DMPA-E group versus 1.8% in the DMPA-P group (P !.001). At the femoral neck, the mean absolute BMD value was significantly higher in the DMPA-E group, 0.964 g/ cm2, than in the DMPA-P group, 0.878 g/cm2 (P !.001). The comparison in percent change between the 2 groups was as follows: 4.7% in the DMPA-E group versus 5.1% in the DMPA-P group (P !.001). The pattern of findings was similar after adjusting for changes in bone size; however, the results at the femoral neck BMAD were marginally significant (P =.04). At 12 months, the mean estradiol levels were significantly higher in the DMPA-P group, 36.2 G 1.09 pg/ mL (132.0 G 4.00 pmol/L) than in the DMPA-E group, 26.6 G 1.09 pg/mL (97.6 G 4.00 pmol/L; P =.01).
Results were similar at 24 months, with mean estradiol levels in the DMPA-P group at 36.6 G 1.09 pg/mL (134.3 G 4.00 pmol/L) and in the DMPA-E group at 26.8 G 1.08 pg/mL (98.4 G 4.00 pmol/L; P =.01). No significant relationship was found between estradiol levels and BMD or BMAD in either treatment group.
Comment In this randomized clinical trial, we found that adolescents who were receiving DMPA-E had significantly higher BMD at both the lumbar spine and femoral neck at 12 and 24 months compared with that seen in girls who were receiving DMPA-P. When the data were adjusted for volumetric changes in bone size, the findings followed a similar pattern at the lumbar spine at 12 and 24 months and the femoral neck at 24 months. The only other randomized controlled clinical trial of
46 estrogen supplementation in DMPA users was conducted by Cundy et al16 in adult premenopausal women, with a mean age of 37 years. Thirty-eight women were assigned randomly to receive orally either conjugated estrogens (0.625 mg) or placebo daily over 24 months. At the lumbar spine, BMD increased by 1.0% over 24 months in the estrogen group, whereas, in the placebo group, BMD decreased 2.6% (P !.01). BMD trends at the femoral neck and total body were in the same direction but had less statistical significance. Thus, estrogen appeared to prevent the bone loss that was seen in the DMPA placebo group. This study provides strong corroborating evidence for our primary hypothesis (ie, that estrogen deficiency is the cause of bone loss in premenopausal women who receive DMPA). One important feature that distinguishes adolescents from adult women is that adolescents are normally in a dynamic period of skeletal development; the vast majority of bone mass is accrued before the age of 20 years. From the onset of puberty, around age 11 years, until its completion, usually age 15 years (about gynecologic age 3), total BMC increases by 50%.12 Therefore, bone loss in an adolescent who receives DMPA suggests not only the loss but also a concomitant failure of bone acquisition. The clinical question that emerges from our study is whether estrogen supplementation should be recommended routinely in adolescents who are receiving DMPA. We suggest that this change in clinical practice would be premature for multiple reasons. First, there is the issue of potential increases in BMD after the discontinuation of DMPA. It may be that the bone loss seen in users of DMPA is a temporary phenomenon and recovery of bone mass occurs quickly after the cessation of treatment. One analogy would be breastfeeding, a period marked by dramatic bone loss; however, as shown by Sowers et al17 and Karlsson et al,18 there is total return of bone mass within several months after breastfeeding has been stopped. Two studies to date have examined the response of bone after the discontinuation of DMPA, and both studies found at least partial recovery within 12 to 30 months after the cessation of the method.7,8 Of note, in the 1 study that included young women,7 the increase in BMD was largest in the youngest age group (18-21 year) after discontinuation of DMPA; however, BMD values did not reach that of control subjects. Thus, we cannot conclude that observed bone loss is irreversible until further study is conducted and the relevant data have been obtained. Our study indicated that bone loss in the girls who received DMPA-P is consistent with other prospective studies of bone and DMPA in adolescents.9-11 However, this apparent adverse effect must be weighed against the considerable benefits that are associated with DMPA as a contraceptive.19 DMPA has features
Cromer et al that are particularly appealing to teens: it is highly effective; it obviates the need for daily compliance with a pill, and it can be used privately.20 The use of DMPA has been credited in part with the decrease in teen pregnancy in this country over the past decade. There also may be important environmental factors (such as nutrition and exercise) that ultimately contribute more substantively to bone health after menopause than the early use of DMPA, particularly given the short duration of DMPA use that is seen typically in this population.21 The challenge then that faces the clinician who is providing contraceptive care is to balance the risk versus the benefit ratio of the use DMPA in a very young woman. The benefit from the very high contraceptive efficacy of DMPA is considerable and would appear to outweigh the risk that is associated with bone loss in adolescents. We failed to find a significant relationship between serum estradiol levels and BMD after 12 months of treatment with either DMPA-P or DMPA-E. Other previous investigations have reported similar findings.5 The reason for this pattern of findings is unclear. However, given the demonstrated variability in serum estrogen levels during treatment with DMPA,22 a single serum estradiol level may not be an accurate reflection of tissue levels of estrogen throughout the treatment cycle. A significant limitation of this study was the low numbers of girls who were available for follow-up. Thus, there is no guarantee that the group that was measured at either 12 or 24 months was representative of the group that was measured at baseline. A source of bias may have been introduced by a coincidence of differing behavioral attributes between those girls who withdrew from the study, those girls who were still active (and did not complete the study protocol due to early discontinuation of trial), and those girls who completed the study. However, we assessed the major attributes that were relevant to BMD and found no differences among these groups. In conclusion, we found that BMD was significantly higher over a 24-month period in adolescent girls who received DMPA plus 5-mg monthly injections of estradiol cypionate compared with girls who received DMPA plus received monthly injections of placebo. Research efforts must address the degree of bone recovery after DMPA use in adolescents and the effects of nutrition and exercise on BMD, to help define the best clinical practice regarding estrogen supplementation in adolescents who use DMPA for contraception.
Acknowledgments We thank the following team members for their exceptional efforts: Kelly Camlin-Shingler, MSSA, Ray Harvey, BS, Mary Jo Day, LPN, Darlene Lewis, RN, Rachel Whitsel, BA.
Cromer et al
References 1. Kaunitz AM. Current concepts regarding use of DMPA. J Reprod Med 2002;47(suppl):785-9. 2. Creinin MD, Cullins VE, Emans SJ, Adams Hillard PJ, Nelson AL, Poindexter NA, et al. Implants and injectables: recent developments. The Contraception Report 2000;10:26-30. 3. Piccinino LJ, Mosher WD. Trends in contraceptive use in the United States: 1982-1995. Fam Plann Perspect 1998;30:4-10. 4. US Census Bureau. Census 2000 summary file 1. Washington (DC): US Department of Commerce; Economics and Statistics Bureau; 2000. 5. Cundy T, Evans M, Roberts. Bone mineral density in women receiving depot medroxyprogesterone acetate for contraception. BMJ 1991;303:13-9. 6. Scholes D, Lacroix AZ, Ott SM, Ichikawa IE, Barlow WE. Bone mineral density in women using depot medroxyprogesterone acetate for contraception. Obstet Gynecol 1999;93:253-8. 7. Scholes D, Lacroix AZ, Ichikawa LE, et al. Injectable hormone contraception and bone density: results from a prospective study. Epidemiology 2002;13:581-7. 8. Cundy T, Cornish J, Evans MC, Roberts H, Reid IR. Recovery of bone mineral density in women who stop using depot medroxyprogesterone acetate. BMJ 1993;308:247-8. 9. Cromer BA, McArdle BJ, Mahan JD, Zibners L, Naumovski Z. A 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. 10. Busen NH, Britt NB, Reanin 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. 11. 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.
47 12. Sabatier JP, Guaydier-Souquieres G, Benmalek A, Marcelli C. Evolution of bone mineral content during adolescence and adulthood: a longitudinal study in 295 healthy females 10-24 years of age and 206 premenopausal women. Osteoporos Int 1999;9:476-82. 13. Grunbaum JA, Kann L, Kinchen SA, Williams B, Ross JG, Lowry R, et al. Youth risk behavior surveillance: United States 2001. J School Health 2002;72:313-28. 14. Hertzler AA, Trary RB. A dietary calcium rapid assessment method (RAM): innovations in practice: a dietary calcium RAM. Top Clin Nutr 1994;9:76-85. 15. 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. 16. Cundy T, Ames R, Horne A, Clearwater J, Roberts H, Gamble G, et al. A randomized controlled trial of estrogen replacement therapy in long-term users of depot medroxyprogesterone acetate. 2003;88:78-81. 17. Sowers M, Corton G, Shapiro B, Jannausch ML, Crutchfield M, Smith ML, et al. Changes in bone density with lactation. JAMA 1993;269:3130-5. 18. Karlsson C, Obrant KJ, Karlsson M. Pregnancy and lactation confer reversible bone loss in humans. Osteoporos Int 2001;12: 828-34. 19. Westhoff C. Bone mineral density and DMPA. J Reprod Med 2002;61:77-82. 20. Cromer BA, Smith RD, McArdle JM, Dwyer J, Brown RT. 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. 21. Zibners A, Cromer BA, Hayes J. Comparison of continuation rates for hormonal contraception among adolescents. J Pediatr Adolesc Gynecol 1999;12:90-4. 22. Clark MK, Sowers M, Levy BT, Tenhundfeld P. Magnitude and variability of sequential estradiol and progesterone concentrations in women using depot medroxyprogesterone acetate for contraception. Fertil Steril 2001;75:871-7.