Contraception 86 (2012) 332–336

Original research article Effect of two kinds of different combined oral contraceptives use on bone mineral density in adolescent women☆ ⁎ Ling Gaia, , Yifang Jiab, Meihua Zhanga, Ping Gaib, Sumei Wanga, Hong Shia, Xiaojie Yua, Yonghong Liuc aKey Laboratory for Improving Birth Outcome Technique, Shandong Provincial Institute of Science and Technology for Family Planning, Jinan, Shandong 250002, China bProvincial Hospital Affiliated to Shandong University, Jinan, Shandong 250021, China cShandong Medical Imaging Research Institute, Jinan, Shandong 250021, China Received 26 December 2011; revised 16 January 2012; accepted 18 January 2012

Abstract

Background: hormonal contraceptives are highly effective and widely used. Most studies have shown a negative effect of combined oral contraceptives (COCs) on the bone mineral density (BMD) of adolescents. The study was conducted to compare BMD among users of /, users of ethinylestradiol/ and nonhormonal control subjects in women aged 16–18 years. Study Design: The study included 450 women 16–18 years of age. One hundred fifty women were using ethinylestradiol/desogestrel, 150 women were using ethinylestradiol/, and 150 women were using nonhormonal contraception as control subjects. BMD of the lumbar spine and femoral neck was obtained using dual-energy X-ray absorptiometry, and mean BMD changes in COCs users and nonusers were compared. Results: At 24 months of treatment, lumbar spine and femoral neck mean BMD values in women (n=127) who used ethinylestradiol/ desogestrel were slightly lower compared with baseline, but these effects did not reach statistical significance (p=.837 and p=.630, respectively). The mean lumbar spine and femoral neck BMD values in women (n=134) who used ethinylestradiol/cyproterone acetate were slightly higher compared with baseline, but there was no statistical significance (p=.789 and p=.756, respectively). The increases in mean percent change in lumbar spine and femoral neck BMD in the ethinylestradiol/cyproterone acetate group were less than those in the control group (1.88% vs. 0.30% and 0.98% vs. 0.49%, respectively). There were no significant differences in mean BMD of the lumbar spine and femoral neck between the users of ethinylestradiol/desogestrel or ethinylestradiol/cyproterone acetate and nonusers (pN.05). Conclusion: Our study indicates that 2 years of COCs therapy had no significant effect on bone density in adolescents, but it remains unknown whether therapy longer than 2 years has a significant adverse effect on the attainment of peak bone mass. © 2012 Elsevier Inc. All rights reserved.

Keywords: Combined oral contraceptive (COCs); Ethinylestradiol/desogestrel; Ethinylestradiol/cyproterone acetate; Bone mineral density (BMD); Contraception

1. Introduction

Steroid hormonal contraceptives, including oral and in- jectable contraceptives, are highly effective and widely used. These contraceptives have important health benefits, includ- ☆ This study was supported by a grant from the Research Foundation of ing contraceptive and noncontraceptive benefits, but also the Shandong Provincial Committee of Family Planning Research, Jinan, have some health risks, such as bone health. Some studies Shandong, China. have demonstrated that use of progestin-only contraceptive ⁎ Corresponding author. Key Laboratory for Improving Birth Outcome depot acetate (DMPA) may cause Technique, Shandong Provincial Institute of Science and Technology for Family Planning, Jinan, Shandong 250002, China. Tel.: +86 531 82597813; bone loss [1,2]. Combined oral contraceptives (COCs) may fax: +86 531 82597807. also adversely affect bone health, especially when used in E-mail address: [email protected] (L. Gai). adolescents [3] and young women [4]. Use of COCs may

0010-7824/$ – see front matter © 2012 Elsevier Inc. All rights reserved. doi:10.1016/j.contraception.2012.01.009 L. Gai et al. / Contraception 86 (2012) 332–336 333 impede attainment of peak bone mass [4]. Most COCs users baseline. All subjects had BMD determined approxi- are under 30 years of age [5]. Previous investigations have mately every 12 months for 24 months. The data were sta- examined the effect of COCs on bone mineral density tistically analyzed. (BMD) in adolescents [3–9], but the results are inconsistent. The objective of the present study was to evaluate 2.3. Statistical analysis whether use of ethinylestradiol (EE)/desogestrel and EE/ All analyses were performed using SPSS version 13.0. cyproterone acetate affects bone health in adolescent women The one-way analysis of variance was used to test for dif- by comparing their BMDs with those of controls. ferences of BMD values and selected characteristics among the three groups, and two-group comparisons for these vari- ables were conducted using t tests. A statistical test with a 2. Materials and methods p valueb.05 was considered statistically significant. Data 2.1. Subjects are presented as mean±SD. We originally enrolled 450 women from 16 to 18 years old attending family planning clinics and requesting birth 3. Results control. All of these women had never used hormonal con- 3.1. Demographic and anthropometric characteristics traception prior to recruitment to this study. Three hundred women requesting COCs contraception were randomized In total, 450 women aged 16–18 years were recruited. A into two treatment groups by drawing lots, receiving either total of 138 (92.0%) EE/desogestrel users, 139 (92.7%) EE/ an EE/desogestrel oral contraceptive with EE 30 mcg and cyproterone acetate users and 136 (90.7%) nonhormonal desogestrel 0.15 mg (N.V. Organon, the ) (group users completed the first 12 months of observation, and a A, n=150) or EE/cyproterone acetate oral contraceptives total of 127 (84.7%) women in the EE/desogestrel group, with EE 35 mcg and cyproterone acetate 2 mg (Schering 134 (89.3%) women in EE/cyproterone acetate group and GmbH and Co., Germany) (group B, n=150). Women of 115 (76.7%) women in the nonhormonal group completed groups A and B were instructed to start pill intake from the the entire 24 months of observation. The reasons for fifth day after the beginning of the next spontaneous menses treatment-phase withdrawal (n=23) in group A were as for 21 days with a 7-day pill-free interval. Women who did follows: nine (39.1%) women terminated for side effects of not wish to use hormonal were recruited as the method, such as irregular bleeding, , etc.; three nonhormonal controls (group C, n=150). They used intra- (13.0%) became pregnant; five (21.7%) moved; four (17.4%) uterine device or condom for contraception. were noncompliant; and two (8.7%) were lost to follow-up. Inclusion and exclusion criteria include the following: All The reasons for treatment-phase withdrawal (n=16) in group subjects had regular menses, should not have used any kind B were as follows: two (12.5%) had irregular bleeding, two of , and had no or (12.5%) became pregnant, three (18.8%) moved, six (37.5%) delivery for at least 6 months. They were also required not to were noncompliant, and three (18.8%) were lost to follow- become pregnant. They did not take any , vitamin D up. The reasons for treatment-phase withdrawal (n=35) in and bone-affecting medication. Women who had chronic group C were as follows: 22 (62.9%) became pregnant, 6 disease, such as diabetes mellitus, renal dysfunction, thyroid (17.1%) moved, 3 (8.6%) were noncompliant, and 4 (11.4%) and parathyroid diseases, hepatitis or pituitary diseases, were were lost to follow-up. excluded from this study. Written informed consent was Table 1 shows baseline information about these women. obtained from subjects and their parents or legal guardians There were no significant differences between the three when the subjects were under the age of 18 years. Study contraceptive user groups regarding mean age, BMI, age at protocols were approved by the Institutional Review Board of menarche, menstrual cycle or number of . Shandong Provincial Institute of Family Planning Research. 3.2. BMD measurements 2.2. Variables There were no significant differences in lumbar spine and The subjects were asked about their age, number of femoral neck mean BMD among groups A, B and C at pregnancies and deliveries, menstrual status, etc. Physical baseline. At 24 months of treatment, in group A, as compared measurements were carried out for their height and weight. to baseline, the mean BMD in lumbar spine and femoral neck Height was measured using a stadiometer. Body weight was revealed a slight decrease. The mean percentage change from measured with calibrated electronic scales. Body mass index baseline in lumbar spine and femoral neck had decreased by (BMI) was calculated as body mass/(height)2. 0.30% and 0.61%, respectively. The mean lumbar spine and BMD at the lumbar spine (L2-4) and femoral neck was femoral neck BMD values at 24 months were not signi- measured by dual-energy X-ray absorptiometry (QDR- ficantly different compared to baseline and subjects in the 4500W, Hologic, Bedford, MA, USA). Results were recorded nonuser group (pN.05). While in groups B and C, there was a as g/cm2. Before using COCs, BMD was determined at trend toward increasing BMD. In group B, the mean 334 L. Gai et al. / Contraception 86 (2012) 332–336 Table 1 adolescents (14–18 years), mean BMD did not differ by Demographic and anthropometric characteristics of the participants COCs duration of use or EE dose. However, in women aged Group A Group B Group C p value between 19 and 30 years, the mean BMD was lower with N 150 150 150 longer COCs use for the spine and whole body (p=.004 and Chronological age 17.09±0.79 17.05±0.78 17.13±0.78 .718 .02, respectively) and lowest for N12 months of low-dose (years) COCs for the hip, spine and whole body (p=.02, .003 and Gynecologic age (years) 3.72±1.22 3.65±1.20 3.72±1.32 .867 .002, respectively). Puberty is very important for the Weight (kg) 49.51±3.49 49.53±4.01 49.46±3.66 .987 Height (cm) 158.11±4.29 158.92±4.27 158.17±4.36 .194 achievement of peak bone mass in adolescent girls. Estro- BMI (kg/m2) 19.79±0.95 19.58±0.92 19.75±0.84 .107 gen is required for normal pubertal skeletal growth and Age at menarche (years) 13.38±1.02 13.35±0.98 13.41±1.08 .879 maturity in adolescent women. COCs inhibit the hypotha- Menstrual cycle (days) 9.63±2.56 9.71±2.64 9.71±2.57 .945 lamic–pituitary axis and maintain a constant serum No. of pregnancies .21±0.47 .22±0.53 .17±0.47 .618 concentration comparable to the level present in the early No. of deliveries .027±0.16 .020±0.14 0.013±0.12 .713 follicular phase of the menstrual cycle [10]. Hypoestrogen- fi Data are expressed as mean±SD. There were no signi cant differences emia is one of the most important causes of bone loss in among the three groups. women. Therefore, the suppression of endogenous sex steroid production by COCs may interfere with the increase percentage change at 24 months from baseline in the lumbar in bone mass during the adolescence. The high levels of spine and femoral neck increased by 0.30% and 0.49%, ethinyl estradiol may overcome this interference. respectively. There were no significant differences compared Many longitudinal studies have examined the effect of to baseline and nonusers. In group C, the mean percentage COCs on BMD changes in the adolescent population. change from baseline in lumbar spine and femoral neck Although COCs may have beneficial effect on BMD in increased by 1.88% and 0.98%, respectively. There were perimenopausal or postmenopausal women, studies of ado- also no significant differences compared to baseline. There lescent and young women generally found lower mean BMD were no significant differences in lumbar spine and femoral among COCs users than nonusers [11]. Pikkarainen et al. [3] neck BMD among the three groups after 24 months of followed BMC in 122 adolescent women aged 12–19 years. treatment (pN.05) (Table 2). After 4 years of follow-up, there was a significant trend showing less increase in the mean BMC of lumbar spine in the group of adolescent women who had used COCs for 4. Discussion more than 2 years compared with the groups with 1–2 years of use and the nonusers. In the mean BMC of the femoral Osteoporosis is characterized by low BMD, compromised neck, there was a significant trend of a smaller increase in architectural stability and increased risk of fracture. Attain- COCs users of more than 2 years compared with those with ment of peak bone mass is essential for the prevention of 1–2 years of use. Beksinska et al. [9] investigated BMD in osteoporosis. Adolescence is the critical period for bone 15- to 19-year-old new users of DMPA, mass accrual. Maximizing peak bone mass in the adolescent enanthate and COCs. They found that, in nonusers, mean period may reduce the risk of osteoporotic fractures in later BMD increased by 1.49% per annum, while in the COCs life. There is no agreement on whether the use of COCs during the adolescent ages has any effect on BMD. Table 2 Our study indicated that, at 24 months of treatment, there Changes in lumbar spine and femoral neck BMD after 24 months of were no significant differences in mean BMD of the lumbar contraceptive use spine and femoral neck between the users of EE/desogestrel Group A Group B Group C p value N or EE/cyproterone acetate and nonusers (p .05). Age, BMI, Value of BMD at baseline age at menarche, menstrual and status of the three N 150 150 150 groups were statistically matched to eliminate possible con- Lumber spine 1.010±0.107 1.009±0.107 1.008±0.109 .998 founding factors. Some cross-sectional studies were con- Femoral neck 0.818±0.089 0.818±0.087 0.816±0.087 .970 ducted to assess the effects of COCs on BMD in adolescent Value of BMD during treatment N 138 139 136 women. Hartard et al. [4] reported that women aged between Lumber spine 1.008±0.106 1.011±0.105 1.018±0.106 .751 18 and 24 years who had ever used COCs had significantly (12 months) lower mean bone mass at the femoral neck and at the tibial Femoral neck 0.815±0.089 0.819±0.087 0.819±0.088 .931 shaft relative to control subjects who had never used COCs, (12 months) and long duration and early start of COCs use were asso- N 127 134 115 Lumber spine 1.007±0.108 1.012±0.107 1.027±0.106 .340 ciated with lower area mean BMD of the femoral neck and (24 months) lower total bone mineral content (BMC) at the distal tibia Femoral neck 0.813±0.090 0.822±0.088 0.824±0.089 .562 and the tibial shaft. Scholes et al. [5] enrolled 606 women (24 months) aged 14 to 30 years to evaluate the effect of COCs dura- Data are expressed as mean±SD. There were no significant differences tion of use and dose on BMD. They found that, in among the three groups. L. Gai et al. / Contraception 86 (2012) 332–336 335 users, the increases were less, only 0.84%. There was evi- different from controls among users of norethindrone- dence for lower BMD increases per annum in COCs users containing COCs (p=.01), but not among users of deso- compared to nonusers (p=.010). Cromer et al. [7] compared gestrel-containing COCs (p=.99). It is not clear whether the BMD of 12- to 18-year-old women who were current use of a desogestrel-containing COCs affects BMD. Our users of COCs and DMPA for 2 years with that of same-age results showed a trend toward a decrease in BMD with EE/ nonusers. They found that, over the first 12 months, the desogestrel use, which did not reach significance. Further mean percent changes in BMD at the spine were as follows: studies are needed to determine whether there is any effect of COCs +2.3% [95% confidence interval (CI) +1.49, +3.18]; desogestrel on bone heath. untreated +3.8% (95% CI +3.11, +4.57). At the femoral Our study suggests that, at 24 months of treatment, mean neck, the mean percent changes were as follows: COCs lumbar spine and femoral neck BMD values in women who +0.3% (95% CI −0.87, +1.41); untreated +2.3%. Mean used EE/desogestrel were slightly lower compared with percent change in BMD at the lumbar spine over 24 baseline values, while the mean lumbar spine and femoral months was+4.2% in the COCs group and +6.3% in the neck BMD values in women who used EE/cyproterone untreated group. At the femoral neck, mean percent change acetate were slightly higher compared with baseline values. in BMD over 24 months was +3.0% in the COCs group There were no significant differences in BMD of the lumbar and +3.8% in the untreated group. One comparison spine and femoral neck between the users of EE/desogestrel between the COCs group and untreated group reached or EE/cyproterone acetate and nonusers. But it remains marginal statistical significance: BMD at the lumbar spine unknown whether therapy longer than 2 years has a was lower in mean absolute value among the COCs users significant adverse effect on the attainment of peak bone than among the untreated group (p=.03). Berenson et al. [8] mass. Further studies are needed to determine whether there found that COCs users 16–33 years old had a slight mean is any effect of desogestrel on bone heath, BMD increase at the spine during the first 12 months, followed by a slow and gradual decrease in the second and Acknowledgment third years. However, COCs users 16–24 years old lost fi signi cantly less mean bone density at the spine (0.4% vs. We thank Xiaoruo Gai, University of Michigan, USA, for – 0.8%, p=.013) than women 25 33 years of age. In contrast, data analysis and assistance in manuscript revision. nonhormonal users 16–24 years old gained significantly more BMD at the spine (3.3% vs. 1.3%, p=.001) than those 25–33 years of age. Reed et al. [12] reported that COCs References users showed a nonsignificant trend for smaller changes in mean BMD at all sites after 24 and 36 months when [1] Harel Z, Johnson CC, Gold MA, et al. Recovery of bone mineral compared with similarly aged nonusers. Their results were density in adolescents following the use of depot medroxyprogesterone – similar with ours. Our results showed that, at 24 months acetate contraceptive injections. Contraception 2010;81:281 91. [2] Gai L, Zhang JL, Zhang HZ, Gai P, Zhou L, Liu YH. The effect of of treatment, in the EE/desogestrel group, the mean per- depot medroxyprogesterone acetate (DMPA) on bone mineral density centage change from baseline in lumbar spine and femoral (BMD) and evaluating changes in BMD after discontinuation of neck had decreased by 0.30% and 0.61%, respectively, DMPA in Chinese women of reproductive age. Contraception 2011; while in the EE/cyproterone acetate group, the mean per- 83:218–22. centage change from baseline in lumbar spine and femoral [3] Pikkarainen E, Lehtonen-Veromaa M, Mottonen T, Kautiainen H, Viikari J. Estrogen-progestin contraceptive use during adolescence neck had increased by 0.30% and 0.49%, respectively. But prevents bone mass acquisition: a 4-year follow-up study. Contracep- there were no significant differences in mean BMD of the tion 2008;78:226–31. lumbar spine and femoral neck between the COCs users [4] Hartard M, Kleimond C, Wiseman M, Weissenbacher ER, Felsenberg and nonusers. D, Erben RG. Detrimental effect of oral contraceptives on parameters Our study compares the effects of two different types of bone mass and geometry in a cohort of 248 young women. Bone 2007;40:444–50. of COCs on BMD. We found that, at 24 months of treat- [5] Scholes D, Ichikawa L, LaCroix AZ, et al. Oral contraceptive use and ment, in the users of EE/desogestrel, the mean BMD in bone density in adolescent and young adult women. Contraception lumbar spine and femoral neck had slightly decreased as 2010;81:35–40. compared to baseline, but did not reach significance, whereas [6] Nappi C, Di Spiezio Sardo A, Greco E, Tommaselli GA, Giordano E, in the users of EE/cyproterone acetate, the mean BMD in Guida M. Effects of an oral contraceptive containing on bone turnover and bone mineral density. Obstet Gynecol 2005;105: lumbar spine and femoral neck slightly increased as com- 53–60. pared to baseline. In the study of adult women (25–40 years), [7] Cromer BA, Bonny AE, Stager M, et al. Bone mineral density in we observed a similar scenario [13]. A study of women with adolescent females using injectable or oral contraceptives: a 24-month a mean age of 25 years by Berenson et al. [14] found that prospective study. Fertil Steril 2008;90:2060–7. women who used desogestrel- and norethindrone-containing [8] Berenson AB, Rahman M, Breitkopf CR, Bi LX. Effects of depot medroxyprogesterone acetate and 20-microgram oral contraceptives on COCs experienced a mean BMD gain of approximately bone mineral density. Obstet Gynecol 2008;112:788–99. 0.33% and 2.33%, respectively, compared with a mean [9] Beksinska ME, Kleinschmidt I, Smit JA, Farley TM. Bone mineral 0.37% loss in nonhormonal users, which was significantly density in a cohort of adolescents during use of norethisterone 336 L. Gai et al. / Contraception 86 (2012) 332–336

enanthate, depot-medroxyprogesterone acetate or combined oral con- [12] Reed SD, Scholes D, LaCroix AZ, Ichikawa LE, Barlow WE, Ott SM. traceptives and after discontinuation of . Longitudinal changes in bone density in relation to oral contraceptive Contraception 2009;79:345–9. use. Contraception 2003;68:177–82. [10] Agostino H, Di Meglio G. Low-dose oral contraceptives in adoles- [13] Liu RJ, Gai L, Jia YF, Gai P, Yu XJ. Effect of two kinds of different cents: how low can you go? J Pediatr Adolesc Gynecol 2010;23: combined oral contraceptives use on bone mineral density in women of 195–201. reproductive age. J Reprod Contracept 2011;22:145–51. [11] Martins SL, Curtis KM, Glasier AF. Combined hormonal contracep- [14] Berenson AB, Radecki CM, Grady JJ, Rickert VI, Thomas A. A tion and bone health: a systematic review. Contraception 2006;73: prospective, controlled study of the effects of hormonal contraception 445–69. on bone mineral density. Obstet Gynecol 2001;98:576–82.