These studies established that bone density decreases with the initiation of DMPA, but none of them addressed the key issue of whether BMD remains at lower levels long term (ie, decades) and thereby increases future fracture risk.
Studies of adolescents. Fewer studies have examined the relationship between DMPA use and BMD in adolescents ( TABLE 2 ). Most available studies have small sample sizes and methodological limitations (high dropout rates, different age criteria, and significant differences in the comparison groups). In this population, DMPA seems to cause a mild decrease in BMD. There are not enough data to evaluate BMD recovery after DMPA discontinuation. Therefore, it is hard to extrapolate the information about BMD in an adolescent to future fracture risk.
One study examined serum estradiol levels and BMD in 22 adolescents ages 15 to 19 years who were new users of DMPA.16 Only 6 participants were still using DMPA at 1 year, and 4 used it throughout the 2 years of the study. The trend over 2 years was toward decreasing BMD. Serum estradiol levels were low, but were not correlated with BMD.
Another related study measured bone biochemical markers in 3 groups: 53 adolescents ages 12 to 18 starting DMPA; 165 adolescents starting oral contraceptive pills; and 152 adolescent women not using hormonal contraception.17 There was no relationship between bone biochemical markers and BMD at either the LS spine or the femoral neck.
TABLE 1
DMPA’s effect on BMD in adult women: What the studies reveal
AUTHOR (TYPE OF STUDY) | # OF PARTICIPANTS/ POPULATION DESCRIPTION | OUTCOME MEASURE | RESULTS | COMMENTS |
---|---|---|---|---|
Gbolade, 199824 (cross-sectional) | N=185 Ages 17-52 (mean 33) Using DMPA for 1-16 years | DEXA of LS spine and femoral neck | Z-score lower at LS spine (P<.001) but not at the femoral neck (P=.25) | No significant association between duration of DMPA use and Z-score |
Ryan, 200225 (cross-sectional) | N=32 Ages 19-53 Using DMPA >2 years Low serum estradiol level or menopausal symptoms | DEXA of LS spine and femoral neck | Z-scores were lower at both femoral neck (-0.84; 95% confidence interval [CI], -1.17 to -0.52) and LS spine (-0.32; 95% CI, -0.62 to -0.02) | 18 women had osteopenia at LS spine 3 women had osteoporosis at LS spine |
Petitti, 200026 (cross-sectional) | n=350 (DMPA) n=695 (control) Ages 30-34 Using DMPA ≥2 years Control group: women who never used hormonal contraception | SXA of wrist | BMD was lower for DMPA current users vs nonusers 0.465 vs 0.471 g/cm2 in midshaft ulna (P<.001) 0.369 vs 0.382 g/cm2 in distal radius (P<.001) | Large WHO-sponsored, multinational study Past users of DMPA had bone densities not significantly different from nonusers Large variations in BMD among sites |
Wanichsetakul, 200227 (cross-sectional) | n=34 (DMPA) n=62 (comparison) Ages 30-34 Using DMPA ≥2 years Comparison groups of women on no steroid contraception in prior 6 months | DEXA of LS spine, distal radius, and femoral neck | BMD at femoral neck and distal radius was not different between DMPA users and controls (P=.335 and P=.398) DMPA users had lower BMD at LS spine (P=.007) | Study conducted in Thailand |
Beksinska, 200528 (cross-sectional) | n=127 (DMPA) n=161 (comparison) Ages 40-49 Using DMPA ≥1 year | DEXA of radius and ulna | No significant difference in BMD at distal radius (P=.26) or ulna (P=.21) | Higher BMD was associated with higher BMI Higher FSH levels were associated with lower BMD |
Tang, 200029 (cohort) | N=59 Ages 37-49 Using DMPA for a mean of 10 years | DEXA of LS spine and femoral neck Annual measurements for 3 years | Small annual decreases in BMD at LS spine (-0.44%), femoral neck (-0.4%), and Ward’s triangle (-1.05%) | Duration of DMPA use not related to BMD Decreases in BMD less than projected for age Study conducted in China |
Scholes, 200213 (cohort) | n=183 (DMPA) n=258 (comparison) Ages 18-39 Comparison group not exposed to DMPA | DEXA of LS spine and proximal femur Measurements every 6 months for 4 years | Total hip and LS spine BMD were lower for DMPA users (P=.002 at LS spine; P<.005 for proximal femur) | New users lost bone faster than longer-term users Women who discontinued DMPA showed increasing BMD levels, which reached levels of nonusers after 30 months 33% dropout rate among both groups at 3 years, 44% of DMPA users discontinued use within first 6 months of the study |
Cundy, 199411 (cohort) | n=36 (DMPA) n=18 (comparison) Ages 25-51 (mean 41-45) 14 women had used DMPA for ≥3 years and discontinued 22 women were long-term DMPA users Individuals in comparison group were never users of DMPA | DEXA of LS spine and femoral neck Measured twice in each woman | Group I (discontinuers) BMD change at LS spine 3.4% per year (1.6% to 5.2%) and at femoral neck 0.8% per year (-1.8% to 3.4%) Group II (long-term users) BMD change at LS spine -0.2% per year (-2.0% to 1.6%) and at femoral neck -1.1% per year (-2.6% to 0.4%) Group III (nonusers) BMD change at LS spine 0.3% per year (-2.2% to 2.8%) and at femoral neck -1.5% per year (-3.2% to 0.2%) | BMD in LS spine in both groups of DMPA users was 9% lower than control group at baseline |
Berenson, 200130 (cohort) | n=33 (DMPA) n=59 (comparison) Ages 18-33 New users of DMPA Comparison group not using any hormonal contraception | DEXA at LS spine 2 measurements for each participant 12 months apart | Adjusted percent change in BMD for DMPA users was -2.7% (-4.44% to -1.05%) and in nonusers was -0.37% (-1.98% to 1.25%), P=.01 | 39% dropout rate among both groups |
Merki-Feld, 200031 (cohort) | N=36 Ages 30-45 Using DMPA ≥6 months | Quantitative CT of radius Measured twice over 12 months | Trabecular bone mass increased 1.6% (P=.8) Cortical bone mass decreased 0.6% (P<.04) | Duration of DMPA use was not associated with BMD change |
Clark, 200414 (cohort) | n=178 (DMPA) n=145 (comparison) Ages 18-35 New users of DMPA Comparison group not using hormonal contraception | DEXA of LS spine and total hip Measured every 3 months for 2 years | At 24 months, change in BMD in DMPA users was -5.8% (SE=0.096) at hip and -5.7% (SE=0.034) at LS spine Significant difference between DMPA group and comparison group (P=.001) | Dropout rate 22% in both groups over 2 years Duration of use predicted decrease in BMD Among DMPA users, increasing BMI was protective against BMD loss at hip |
Kaunitz, 200615 (cohort) | n=248 (DMPA) n=360 (comparison) Ages 25-35 New users of DMPA Comparison group not using hormonal contraception | DEXA LS spine, total hip, femoral neck, and trochanter Measured at baseline and every 48 weeks for up to 5 years | Mean decrease in BMD in DMPA users was 5.16% (±3.6) at hip and 5.38% (±3.57) at LS spine At 96 weeks after discontinuation, change was -0.20% at hip and -1.19% at LS spine | Decreases in BMD were linearly associated with duration of use up to 5 years 17% of DMPA group and 33% of comparison group completed entire 5 years of study |
Clark, 200612 (cohort) | n=178 (DMPA) n=145 (comparison) Ages 18-35 New DMPA users Comparison group not using hormonal contraception | DEXA total hip and LS spine Measured every 3 months for up to 4 years | Mean change in BMD in DMPA users was -7.7% (±0.11) at hip and -6.4% (±0.36) at LS spine DMPA users of 24-36 months had BMD of -4.7% (hip) and -2.9% (spine) compared with baseline 18 months after discontinuation | Most loss was noted first 2 years after initiation of DMPA Most users of DMPA up to 2 years returned to baseline BMD by 3 years 36% dropout rate in both groups after second year of study Only 45% of DMPA group completed 4 years of study |
BMD, bone mineral density; BMI, body mass index; CT, computed tomography; DEXA, dual-energy x-ray absorptiometry; DMPA, depot- medroxyprogesterone acetate; FSH, follicle-stimulating hormone; LS, lumbosacral; SE, standard error; SXA, single-energy x-ray absorptiometry; WHO, World Health organization. |