The effects of different types of hormonal contraception on bone mineral density in women of reproductive age: a systematic review and meta-analysis

Approximately 150 million women worldwide use hormonal contraceptives; however, formulation-specific effects on bone mineral density (BMD) during the reproductive years remain poorly characterized. Prior meta-analyses are limited by the absence of untreated control groups and inadequate heterogeneity quantification, leaving clinicians without reliable guidance. To determine the effect of hormonal contraceptives on lumbar spine and femoral neck bone mineral density (BMD) in reproductive-aged women relative to untreated non-user controls stratified by contraceptive category, progestin class, and ethinylestradiol (EE) dose. This was a pre-registered systematic review and meta-analysis (PROSPERO: CRD420250640432; PRISMA, 2020). The MEDLINE, Embase, Web of Science, and Cochrane CENTRAL of Controlled Trials databases were searched from inception to December 2024. Eligible studies required a mandatory untreated control arm and ≥ 12-month DXA follow-up in women of reproductive age. Effect sizes were pooled as weighted mean differences (WMD) using the DerSimonian–Laird random-effects model. Evidence certainly was rated using GRADE. Seven studies (k = 15 comparisons; n = 1,488 completers) met all inclusion criteria. The pooled WMD for lumbar spine BMD at 12 months was − 0.003 g/cm² (95% CI: −0.010 to + 0.005; p = 0.470; I² = 38.7%), below DXA precision limits and representing less than 0.3% of mean baseline BMD. COC users showed no significant BMD difference from controls across all progestin classes (WMD = − 0.0004 g/cm²; p = 0.925). A clinically interpretable EE-dose gradient emerged: low-dose formulations (≤ 20 µg; k = 8) showed a small but significant homogeneous reduction (WMD = − 0.011 g/cm²; 95% CI: −0.016 to − 0.005; p < 0.001; I² = 0%), while standard-dose formulations (30–35 µg) did not (WMD = + 0.009 g/cm²; p = 0.193). DMPA demonstrated a significant 2.0% lumbar spine BMD reduction (WMD = − 0.020 g/cm²; 95% CI: −0.038 to − 0.002; p = 0.030). RCT-only sensitivity analysis (k = 10) yielded WMD = − 0.006 g/cm² ( p = 0.002; I² = 0%). The GRADE certainty was moderate. COCs do not materially alter BMD over 12 months in reproductive-age women, with consistent null effects across all progestin classes. A biologically plausible EE dose gradient was identified; DMPA use warrants individualized counselling given clinically relevant, though reversible, bone loss. Adequately powered long-term randomized trials targeting low-dose formulations, adolescents, and DMPA continuers are urgently needed. Effect of different types of hormonal contraception on bone mineral density in women of reproductive age