Highlight
• In a Swedish register-based cohort of 2,095,130 women aged 13–49 followed 2006–2019, ever-use of any hormonal contraceptive was associated with a modestly increased breast cancer risk (HR 1.24).
• Risk differed by progestin content: desogestrel-only and desogestrel-combined oral formulations and etonogestrel-containing implants showed higher relative risks than levonorgestrel-containing combined pills and the levonorgestrel 52 mg intrauterine system.
• Absolute excess risk was small (≈1 extra breast cancer per 7,752 users), underscoring a need to balance small individualized risk differences against contraceptive benefits.
Background
Hormonal contraceptives are widely used for pregnancy prevention and non-contraceptive indications (menstrual regulation, dysmenorrhea, acne). Longstanding epidemiologic interest centers on whether exogenous sex hormones alter breast cancer risk. Earlier pooled analyses and large cohort studies showed a small increase in breast cancer risk associated with current or recent combined oral contraceptive (COC) use, with risk attenuating after cessation. However, modern contraceptives include diverse progestins and delivery systems (oral pills, implants, intrauterine systems, injections, vaginal ring), and whether breast cancer risk varies across formulations has been incompletely defined. Clinicians need formulation-level data to inform contraceptive counseling, particularly for adolescents and premenopausal women in whom baseline absolute risk is low but cumulative exposure may be substantial.
Study design
Hadizadeh et al. conducted a Swedish nationwide population-based cohort study using linked registers to examine breast cancer incidence by hormonal contraceptive formulation. The cohort included all females age 13–49 resident in Sweden on January 1, 2006, without prior relevant cancers, bilateral oophorectomy, or fertility treatment. Participants were followed through 2019 and censored at age 50, diagnosis of exclusion conditions, emigration, death, or end of follow-up. Exposure ascertainment relied on prescription records and categorized hormonal contraceptive use by ever-use and duration, hormone formulation (combined vs progestin-only) and specific progestins (eg, desogestrel, levonorgestrel, etonogestrel), and route (oral, implant, intrauterine system [IUS], injection, vaginal ring). Time-dependent Cox regression models adjusted for sociodemographic and clinical confounders estimated hazard ratios (HRs) for incident in situ and invasive breast cancer.
Key findings
Overall association
Among 2,095,130 women contributing 21,020,846 person-years, 16,385 breast cancer cases occurred. Ever-use of any hormonal contraceptive was associated with an increased breast cancer risk: HR 1.24 (95% CI, 1.20–1.28). The authors translate this to an absolute excess of approximately 1 additional breast cancer case per 7,752 users (95% CI, 5,350–14,070), reflecting the low baseline incidence in this age group.
Formulation-specific differences
Risk varied by hormonal content and delivery system. Key estimates include:
- Combined formulations overall: HR 1.12 (95% CI, 1.07–1.17).
- Progestin-only formulations overall: HR 1.21 (95% CI, 1.17–1.25).
- Oral desogestrel-only: HR 1.18 (95% CI, 1.13–1.23).
- Oral desogestrel-combined products: HR 1.19 (95% CI, 1.08–1.31).
- Etonogestrel implants (etonogestrel is the active metabolite of desogestrel): HR 1.22 (95% CI, 1.11–1.35).
- Levonorgestrel-containing combined pills: HR 1.09 (95% CI, 1.03–1.15).
- Levonorgestrel 52 mg intrauterine system (IUS): HR 1.13 (95% CI, 1.09–1.18).
- No statistically significant increased risk observed for medroxyprogesterone acetate injections, etonogestrel vaginal ring, or combined oral drospirenone preparations despite large user numbers.
Thus, desogestrel/etonogestrel-containing products and some progestin-only methods showed relatively higher HRs than levonorgestrel-containing formulations.
Duration and age patterns
The report analyzed time-dependent exposure and duration, consistent with prior literature showing higher risk during current or recent use and attenuation over time after stopping. Median age at diagnosis was 45 years (IQR 41–48), emphasizing that findings predominantly reflect premenopausal breast cancer risk.
Clinical significance
Although relative risks reached the low-to-modest range (HRs ~1.09–1.24), the absolute excess risk among adolescents and premenopausal women was small. The authors’ absolute-risk framing (≈1 additional case per 7,752 users) is important for counseling: for an individual woman with a low baseline risk of breast cancer, the incremental risk from using most hormonal contraceptives is small compared with the contraceptive and non-contraceptive benefits.
Expert commentary and interpretation
This large, register-based cohort offers several strengths: comprehensive, national-level prescription and cancer registry linkage enables high statistical power and the capacity to disaggregate risk by specific progestin and route. Time-dependent exposure modeling and adjustment for sociodemographic covariates strengthen causal inference in an observational context.
Biologic plausibility
Progestins differ in receptor affinities and metabolic properties—some have androgenic activity (levonorgestrel), others are less androgenic or more progestogenic (desogestrel/etonogestrel), and synthetic progestins vary in impact on breast epithelial proliferation. Differential effects on mammary gland proliferation and estrogen receptor signaling provide biologic plausibility for progestin-specific differences in breast cancer risk, but definitive mechanistic data remain limited.
Potential biases and limitations
- Residual confounding and indication bias: prescription choice may correlate with unmeasured factors (eg, body mass index, lifestyle, breastfeeding history, family history) despite covariate adjustment. If certain formulations are preferentially prescribed to subgroups with different baseline risks, observed associations could be influenced by channeling bias.
- Exposure misclassification: register records reflect dispensed prescriptions, not confirmed ingestion or adherence; implant/IUS insertion data are more directly observed but uptake patterns and removal timing could be variably captured.
- Outcome granularity: registry data do not always provide tumor subtype (hormone-receptor status) in sufficient detail to determine whether associations differ by ER/PR status.
- Follow-up and latency: although 2006–2019 follow-up is substantial, latency for hormonally influenced tumor development can be long; findings primarily apply to relatively short- to medium-term effects in women <50.
- Generalizability: results are from a Swedish population with its own prescribing patterns and demographic profile; effect sizes might differ in other settings.
Comparison with prior literature
The results align with prior large cohort data showing small increased breast cancer risk with recent hormonal contraceptive use (eg, Mørch et al., NEJM 2017), and extend understanding by highlighting formulation-specific differences. Unlike some prior work that grouped all COCs together, this study’s granular approach offers actionable detail for clinicians choosing between progestin types and devices.
Clinical implications
For routine clinical counseling, key messages are:
- The absolute risk increase of breast cancer attributable to hormonal contraceptives in adolescents and premenopausal women is small.
- Formulation matters: clinicians may consider progestin type when individualizing contraceptive choice, particularly for women with additional breast cancer risk factors (strong family history, known genetic predisposition) where even marginal relative differences may be meaningful.
- Non-hormonal options and levonorgestrel-releasing IUS (which still shows a small HR increase in this study) remain highly effective alternatives; contraceptive decisions should weigh pregnancy prevention efficacy, bleeding profile, side effects, patient preferences, and risk tolerance.
- Screening and risk-reduction strategies for women at higher baseline risk should not be altered primarily because of contraceptive use, but contraceptive selection can be integrated into broader risk management discussions.
Research gaps and future directions
Important next steps include pooled analyses across countries to validate these findings in diverse populations; mechanistic studies to elucidate progestin-specific effects on breast epithelium and tumor subtypes; evaluation of long-term risk beyond age 50; and stratified analyses in high-risk groups (BRCA carriers). Improved capture of tumor receptor subtype and inclusion of key confounders such as body mass index, parity, lactation history, and detailed family history will strengthen causal interpretation.
Conclusion
Hadizadeh et al. provide compelling, population-level evidence that breast cancer risk associated with hormonal contraception is not uniform across formulations: desogestrel/etonogestrel-containing products showed relatively higher risk estimates than levonorgestrel-containing formulations. Nevertheless, absolute excess risk in adolescents and premenopausal women is small. These data support more nuanced, formulation-aware contraceptive counseling that frames both relative and absolute risks alongside contraceptive benefits and individual risk profiles.
Funding and clinicaltrials.gov
Funding details are reported by the study authors; consult the original JAMA Oncology publication for the full funding statement. This was an observational register-based study and not registered on ClinicalTrials.gov.
References
1. Hadizadeh F, Koteci A, Karlsson T, Ek WE, Johansson Å. Hormonal Contraceptive Formulations and Breast Cancer Risk in Adolescents and Premenopausal Women. JAMA Oncol. 2025 Oct 30:e254480. doi:10.1001/jamaoncol.2025.4480. PMID: 41165687; PMCID: PMC12576617.
2. Mørch LS, Skovlund CW, Hannaford PC, et al. Contemporary hormonal contraception and the risk of breast cancer. N Engl J Med. 2017;377(23):2228-2239. doi:10.1056/NEJMoa1700732.