Depot medroxyprogesterone acetate (DMPA), approved by the U.S. Food and Drug Administration since 1992, functions as an effective, long-acting, and reversible contraceptive for more than 2 million U.S. women annually.1,2 In fact, it has been stated that the decline in unintended pregnancies in the United States between 1987 and 1994 was partly due to increased use of this contraceptive.3,4 However, its adverse effect on bone mineral density (BMD) is a concern among clinicians, researchers, and policy makers. Studies demonstrated that use of DMPA among adolescents and adults for 2–4 years results in a 1.5–6.8% loss of BMD at the spine and a 3.6–7.7% loss at the hip or femoral neck.5–12 One limitation of these studies on this topic, however, is that they report mean bone loss. Thus, it has not been determined whether certain women lose more bone mass whereas others do not. Furthermore, it has not been determined whether it can be predicted who is at greatest risk of losing higher BMD while using DMPA.
Several published studies have attempted to predict who is at increased risk of bone loss while using DMPA. For example, Clark et al10 found that a greater percentage increase in fat mass was protective against bone loss at the spine and hip, whereas women with higher baseline fat mass lost less hip bone mass over time. Berenson et al11 found that DMPA users aged 16–24 years lost more bone at the spine and femoral neck compared with those aged 25–33 years. This same study reported that current smokers were more likely to lose bone mass at the spine, whereas lean mass was protective at the femoral neck. However, information is lacking on clinically important predictors of higher bone loss which could be used in clinical practice. The purpose of this study was to identify possible predictive factors of higher bone loss, defined as at least 5%, at the spine or femoral neck over time in DMPA users.
MATERIALS AND METHODS
As part of a larger study, 805 non-Hispanic African-American, non-Hispanic white, and Hispanic women aged between 16 and 33 years were recruited between October 9, 2001, and September 14, 2004. The methods for the larger study are reported in detail elsewhere.11 Briefly, recruitment was conducted to achieve a sample that was balanced by age group (16–24 years and 25–33 years), race (African American, white, and Hispanic), and contraceptive method (nonhormonal, oral contraceptives, and injections [DMPA]). Of the 805 women, five withdrew before completing the first visit, 92 had abnormal laboratory results, and five had T scores less than −2.5 on their bone scans. Thus, baseline data for the larger study were analyzed for 703 women. Those excluded (n=102) did not differ from women included in the analyses (n=703) on age, marital status, parity, or education (all P>.05). A total of 245 women chose oral contraceptives, 240 chose DMPA, and 218 chose nonhormonal contraceptives. Comprehensive follow-up visits (eg, phlebotomy, urine pregnancy testing, written questionnaires, calcium checklist,13 bone densitometry) were conducted every 6 months after the initial interview whereas contraceptives were dispensed every 3 months and minor follow-up visits (ie, anthropometry) were conducted every 3 months up to 36 months after the baseline visit. A registered dietician conducted a 24-hour recall of dietary intake annually.
Our analyses focused only on the DMPA users (N=240). Of the 240 initial users, 145 discontinued DMPA use sometime during the next 2 years, with 95 completing the study. There were no differences in baseline characteristics between DMPA users who completed the 24-month follow-up (n=95) and those who did not (n=145) with regard to age, race, and BMI (P>.05). For the analysis purpose, an arbitrary value of at least 5% BMD loss from the baseline value after 24 months of DMPA use at any one of the two sites was considered as the higher BMD loss. The larger study had 58 DMPA users after 36 months of follow-up. However, a decision was made to use 24-month follow-up data (n=95) rather than 36-month data (n=58) to provide adequate sample size for the multivariable analyses. Bone densitometry, total body fat, percent body fat, and total lean mass were measured using dual-energy x-ray absorptiometry (QDR 4500W Elite fan-beam densitometer; Hologic Inc., Bedford, MA). All procedures were approved by the Institutional Review Board of the University of Texas Medical Branch.
Univariable comparisons were performed to compare the two groups (at least 5% compared with less than 5% bone loss at the lumbar spine and femoral neck after 24 months of DMPA use) at baseline using the χ2, Fisher exact test, or Student t test as appropriate. Multivariable logistic regression was used to identify correlates of higher BMD loss after 24 months. Variables were screened for inclusion in an initial multivariable model. Candidate variables with P≤.20 were included in the initial multivariable model, whereas variables with P>.20 were excluded from the final model. The Hosmer-Lemeshow test14 and area under the receiver operating characteristic curve were used to assess the fit and predictive ability of the final model. All analyses were performed using Stata 10 (Stata Corporation LP, College Station, TX).
At baseline, the total sample (N=240) had a mean age of 23.8 years with 72 African-American, 82 white, and 86 Hispanic women. Of the initial 240 DMPA users, 95 completed 24 months of follow-up. Forty-five of the 95 DMPA users (47.4%) had at least 5% BMD loss at the lumbar spine or femoral neck. A total of 50 women had less than 5% bone loss at both sites during the same period. The two groups (at least 5% BMD loss at spine or femoral neck compared with less than 5% at the two sites) were similar in terms of race, height, weight, BMI, lean mass, fat mass, age at menarche, smoking, alcohol use, weight-bearing exercise, daily calcium intake, use of calcium supplements, and baseline spine and femoral neck BMD (Table 1). However, the former group of DMPA users were younger, were less likely to have used DMPA in the past, and had lower mean parity than the latter group.
Variables that met the screening criteria for inclusion in the multivariable model were age, BMI, parity, previous use of DMPA, total calcium intake per day, and current smoking while race/ethnicity was forced into the model. Variables with P>.20 (age and BMI) were excluded from the final model. The final logistic model yielded a P value for the Hosmer-Lemeshow test of .25, and the estimated area under the receiver operating characteristic curve was 0.79 (95% confidence interval [CI] 0.70–0.88). Adjusted odds ratios (ORs) for the final logistic model are presented in Table 2. Higher BMD loss after 24 months of DMPA use was associated with current smoking (adjusted OR 3.88, 95% CI 1.26–11.96), calcium intake (in 100 mg) (OR 0.81, 95% CI 0.65–0.99), and parity (OR 0.49, 95% CI 0.29–0.82).
Figure 1A–F shows the mean percentage BMD loss at the spine and femoral neck by smoking status, calcium intake, and history of parity over 24 months with 6-month increments based on the available observations at each follow-up period. By and large, BMD loss was higher in women who were current smokers, had never been pregnant, and had a daily calcium intake of 600 mg or less (mean daily calcium intake) compared with their counterparts, although not all comparisons achieved statistical significant difference. Moreover, the difference in BMD loss substantially increased between the women with all three risk factors and without any risk factors (Fig. 1G and H).
Thirty-six–month data were available for 27 of the 45 women who had at least 5% BMD loss at 24 months at either site. Data on these 27 women showed that the BMD loss increased at the spine from 5.3% to 5.8% and at the femoral neck from 5.7% to 7.7%, on average, between 24 and 36 months of use. In contrast, those who lost less than 5% BMD by 24 months and had a scan at 36 months (n=29) experienced an average loss of 1.7% at the lumbar spine and 2.5% at the femoral neck by the 36 month visit, as compared with 1.2% and 0.7% at 24 months, respectively.
This study examined the correlates of BMD loss associated with DMPA use based on a distinct cutoff point (at least 5% compared with less than 5%). We observed that almost half of DMPA users experience a bone loss of this magnitude within 2 years of initiating DMPA. Furthermore, these women continued to experience significant losses in BMD during their third year of use, especially at the femoral neck. These losses are of significant concern because they are likely to take an extended period of time to recover, especially at the hip. In a previous publication that included data on BMD changes after DMPA discontinuation, we reported that women regained an average of 2.4% BMD annually at the lumbar spine after they stop receiving injections.11 Bone lost at the femoral neck, however, recovered more slowly (1.6% per year). Clark et al10 observed that hip BMD was 4.7% below baseline 18 months after DMPA discontinuation among women who used this contraceptive for 24–36 months, demonstrating that this site was slow to recover in their study as well. The same study also reported spine BMD 2.9% below baseline 18 months after DMPA discontinuation. In contrast, Kaunitz et al15 observed more rapid recovery at the hip and femoral trochanter. These values were close to their baseline levels after 96 weeks of DMPA discontinuation among women who used this contraceptive up to 5 years. However, spine and femoral neck BMD (1.2% and 3.1% lower than baseline level after 96 weeks, respectively) showed only partial recovery in that study, similar to the studies of Berenson et al11 and Clark et al.10 Because the femoral neck is the most common fracture site after menopause, DMPA use for multiple years may put women at risk of future fractures if these detrimental changes are not completely reversible. Unfortunately, we were unable to examine reversibility among women who experienced at least 5% loss in their BMD in the current investigation, because too few women were followed after discontinuation with this large of a loss. Thus, the reversibility of higher bone loss among DMPA users and the amount of time that it takes to return to baseline values needs to be explored in future studies.
This study reports that BMD loss is not a significant concern for all women who choose DMPA for contraception, because it is associated with certain risk factors. Those who had delivered a child, did not smoke, and consumed at least 600 mg/d of calcium did not lose more than 2% of their BMD at the spine or hip over 24 months. Thus, concerns about their bone health are minimal. Furthermore, clinicians can counsel women who have a modifiable risk factor and wish to use DMPA that they can protect their bone health by quitting smoking and consuming a diet adequate in calcium. Given that the average calcium intake of 600 mg/d among women in this study is still far below recommended amounts,16 this counseling would probably benefit users of all types of contraceptives. Brochures that depict calcium-rich foods could be disseminated, and information on the role of calcium supplementation could be provided for those with inadequate dietary intake. Physicians may also wish to disseminate information about types of supplements and which are most efficiently absorbed. We observed that only approximately 8.9% of the women using DMPA took calcium supplements at least 4 d/wk, demonstrating that few young women of reproductive age understand the importance of adequate calcium consumption. Finally, family planning visits could provide an opportunity to discuss smoking cessation programs in the local area as well as medications that can assist with their efforts to quit.
Another important finding of this study was that increased parity resulted in lower BMD loss. However, results of prior published studies on the relationship between parity and BMD are not consistent.17–19 Some studies showed higher BMD with increased parity in later life, whereas others showed lower BMD or did not find any association. Pregnancy-related changes in estrogen concentration, and increase in periosteal expansion and bone size due to weight gain during pregnancy could act as protective factors for BMD loss in later life.
Prevention of bone loss while using DMPA is not well understood and remains an important topic for future research. One option that has been suggested is to give estrogen supplementation. A double-blind, placebo-controlled trial in women with a mean age of 37 years demonstrated that this can result in a 1% increase in spinal BMD per year among users of DMPA as compared with a 2.6% loss annually among those who did not take estrogen.20 Similar results were also found for the femoral and total body BMD, although the statistical significance was less. Another study of adolescent DMPA users (mean age of 16 years) who had monthly estradiol cypionate injections showed a 2.8% spine BMD increase over 24 months compared with a 1.8% decrease in DMPA users who did not have estradiol injections.21 This study also showed a similar pattern at the femoral neck. Future studies that consider fracture as an outcome are needed to examine the practicality of using this supplement along with DMPA.
This study has several limitations. First, the sample size was small, so there may be associations that we were unable to detect in our analyses. Second, data such as calcium intake and amount of exercise are subject to recall bias. Third, the discontinuation rate among DMPA users was high. However, this is a problem inherent to many contraceptive studies, because there are many reasons that women may choose to change or discontinue their method.22–24 Fourth, the larger study did not randomly assign women to one of the three contraceptive groups, because these methods have different efficacies and randomization could have led to unintended pregnancies. Finally, we did not have enough women in the at least 5% bone loss group who were followed after DMPA discontinuation to examine the reversibility of these findings among this subset.
In conclusion, women who have not delivered a child, smoke, and do not consume much calcium in their diet are at risk for higher bone loss while using DMPA. These women require additional counseling on how to decrease their risk of BMD loss to avoid putting their bone health at risk.
1. Kaunitz AM, Arias R, McClung M. Bone density recovery after depot medroxyprogesterone acetate injectable contraception use. Contraception 2008;77:67–76.
2. Mosher WD, Martinez GM, Chandra A, Abma JC, Willson SJ. Use of contraception and use of family planning services in the United States: 1982–2002. Adv Data 2004;10(350):1–36.
3. Henshaw SK. Unintended pregnancy in the United States. Fam Plann Perspect 1998;30:24–9, 46.
4. Achievements in public health, 1900–1999: family planning. MMWR Morb Mortal Wkly Rep 1999;48(47):1073–80.
5. Scholes D, Lacroix AZ, Ichikawa LE, Barlow WE, Ott SM. Injectable hormone contraception and bone density: results from a prospective study [published erratum appears in Epidemiology 2002:6:749]. Epidemiology 2002;13:581–7.
6. Berenson AB, Breitkopf CR, Grady JJ, Rickert VI, Thomas A. Effects of hormonal contraception on bone mineral density after 24 months of use. Obstet Gynecol 2004;103:899–906.
7. Clark MK, Sowers MR, Nichols S, Levy B. Bone mineral density changes over two years in first-time users of depot medroxyprogesterone acetate. Fertil Steril 2004;82:1580–6.
8. 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.
9. Scholes D, LaCroix AZ, Ichikawa LE, Barlow WE, Ott SM. Change in bone mineral density among adolescent women using and discontinuing depot medroxyprogesterone acetate contraception. Arch Pediatr Adolesc Med 2005;159:139–44.
10. Clark MK, Sowers M, Levy B, Nichols S. Bone mineral density loss and recovery during 48 months in first-time users of depot medroxyprogesterone acetate. Fertil Steril 2006;86:1466–74.
11. Berenson AB, Rahman M, Breitkopf CR, Bi LX. Effects of depot medroxyprogesterone acetate and 20-microgram oral contraceptives on bone mineral density. Obstet Gynecol 2008;112:788–99.
12. Cromer BA, Bonny AE, Stager M, Lazebnik R, Rome E, Ziegler J, et al. Bone mineral density in adolescent females using injectable or oral contraceptives: a 24-month prospective study. Fertil Steril 2008;90:2060–7.
13. Hertzler AA, Frary RB. A dietary calcium rapid assessment method (RAM). Top Clin Nutr 1994;9:76–85.
14. Hosmer DW Jr, Lemeshow S. Applied logistic regression. 2nd ed. New York (NY): John Wiley & Sons; 2000.
15. Kaunitz AM, Miller PD, Rice VM, Ross D, McClung MR. Bone mineral density in women aged 25–35 years receiving depot medroxyprogesterone acetate: recovery following discontinuation. Contraception 2006;74:90–9.
17. Henderson PH III, Sowers M, Kutzko KE, Jannausch ML. Bone mineral density in grand multiparous women with extended lactation. Am J Obstet Gynecol 2000;182:1371–7.
18. Specker B, Binkley T. High parity is associated with increased bone size and strength. Osteoporos Int 2005;16:1969–74.
19. Karlsson MK, Ahlborg HG, Karlsson C. Maternity and bone mineral density. Acta Orthop 2005;76:2–13.
20. 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. J Clin Endocrinol Metab 2003;88:78–81.
21. Cromer BA, Lazebnik R, Rome E, Stager M, Bonny A, Ziegler J, et al. Double-blinded randomized controlled trial of estrogen supplementation in adolescent girls who receive depot medroxyprogesterone acetate for contraception. Am J Obstet Gynecol 2005;192:42–7.
22. Berenson AB, Odom SD, Breitkopf CR, Rahman M. Physiologic and psychologic symptoms associated with use of injectable contraception and 20 μg oral contraceptive pills. Am J Obstet Gynecol 2008;199:351.e1–12.
23. Murphy PA, Brixner D. Hormonal contraceptive discontinuation patterns according to formulation: investigation of associations in an administrative claims database. Contraception 2008;77:257–63.
24. Rosenberg MJ, Waugh MS. Oral contraceptive discontinuation: a prospective evaluation of frequency and reasons. Am J Obstet Gynecol 1998;179(pt 1):577–82.