Hormonal Contraception and Initiation of Semaglutide Therapy.

We conducted a nationwide nested case-control study using Danish national registries. Since 1968, all Danish residents have been assigned a unique civil personal register number at birth or at immigration, enabling unambiguous individual-level linkage across multiple national registries. 9 For this study, data were obtained and linked from several nationwide registries to extract relevant information. First-time users of semaglutide were defined as females aged 12 to 49 years who filled their first prescription for semaglutide (Anatomical Therapeutic Chemical [ATC] code A10BJ06) of any dose, in the Danish National Prescription Registry, without any prior fill of any drug to lower glucose levels (ATC code A10 except for A10BJ06). 10 Each participant in the semaglutide user group was incidence matched on birth year to 10 nonusers (control group) from the nationwide cohort of all Danish females aged 12 to 49 years from January 1, 1996 to December 31, 2023, without a previous prescription fill of any drug to lower glucose levels (ATC code A10) and with no history of hysterectomy, oophorectomy, sterilization, infertility treatment, use of hormone therapy, polycystic ovary syndrome, endometriosis, thrombosis, thrombophilia, liver disease, kidney disease, or pancreatitis. eTable 1 in Supplement 1 provides details on all data sources and definitions. Matching was performed to improve statistical precision and maximize use of the nationwide registry data while maintaining comparability between groups. 11 In accordance with the General Data Protection Regulation and the Danish Data Protection Act, this study was registered and approved by the data-responsible institute in the Capital Region of Denmark. Informed consent or ethical approval are not mandatory for registry-based studies of significant societal importance that are conducted for the sole purpose of statistics and scientific research. This study complied with the Strengthening the Reporting of Observational Studies in Epidemiology ( STROBE ) reporting guideline for case-control studies. Hormonal contraception use was determined through filled prescriptions in The Danish National Prescription Registry. 10 Exposure time was calculated from purchased daily doses using the medicinMacro package in R, version 4.4.1 (R Project for Statistical Computing), which has been described previously and is available on Github. 12 For each female, hormonal contraceptive exposure was assessed by reviewing all filled hormonal contraceptive prescriptions from the index date back to 12 years of age (or study entry). Each female’s history of hormonal contraception use was summarized based on all types of hormonal contraceptives for which she had filled prescriptions and the chronological order in which they were used throughout this period. We included all types of hormonal contraception available on the Danish market, consisting of combined oral contraceptives, progestin-only oral contraceptives, progestin-only intrauterine devices (IUDs), subcutaneous implants, patches, vaginal rings, and injectable depot medroxyprogesterone acetate. However, only the most common utilization patterns were presented in the analyses. Accumulated duration of hormonal contraception use was calculated and stratified into 1-year intervals. eTable 2 in Supplement 1 lists the ATC codes and categories of hormonal contraception used in this study. The indication for semaglutide therapy 1 in Denmark is treatment of type 2 diabetes, whereas semaglutide therapy 2 is recommended for individuals with obesity (body mass index [BMI; calculated as weight in kilograms divided by height in meters squared] ≥30.0) or with overweight and obesity-related comorbidities (BMI ≥27.0 and dysglycemia [prediabetes or type 2 diabetes], hypertension, dyslipidemia, obstructive sleep apnea, or cardiovascular disease). In Denmark, semaglutide is officially dispensed by prescription only, and dispensing occurs exclusively through pharmacies, resulting in complete capture in the Danish National Prescription Registry. We classified the 2 semaglutide types into therapy 1 and therapy 2 using their item number for subgroup analysis. eTable 3 in Supplement 1 lists the codes used to identify semaglutide types. BMI was ascertained using the Danish Medical Birth Register, which contains information on 99.8% of all births in Denmark since 1973. 13 In this register, all pregnancies managed within the public Danish health care system are recorded, including maternal prepregnancy height and weight. BMI information was only available for females who had given birth and could not be ascertained for those without registered pregnancies. For each woman, prepregnancy height and weight were obtained from the most recent registration prior to semaglutide initiation. BMI was categorized into 4 groups: (1) underweight with BMI less than 18.5, (2) normal weight with BMI of 18.5 to 24.9, (3) overweight with BMI of 25.0 to 29.9, and (4) obesity with BMI greater than 30.0. Educational status was obtained using the Danish education registers, 14 which hold information on the highest completed level of education for each individual on an annual basis. Educational level was categorized based on the International Standard Classification of Education (ISCED) into 4 groups: (1) ISCED levels 0 to 2, equivalent to early childhood, primary, and lower secondary education; (2) ISCED level 3, equivalent to general upper secondary education and vocational level; (3) ISCED levels 5 to 6, equivalent to short-cycle tertiary, medium-length tertiary, and bachelors level education; and (4) ISCED levels 7 to 8, equivalent to second-cycle, masters, or doctorate level education. ISCED level 4 was not used, as no educational attainment has this classification in Denmark. Females with missing data on education level were assigned the lowest educational level (ISCED 1), as an informed imputation strategy, based on the high completeness and quality of the Danish educational registries, indicating that missing educational information most often reflects very limited education. Income was obtained from the Income Statistics Register as equivalized disposable income, which accounts for household income and composition. 15 Income level was categorized as quartiles from 1 to 4 (first indicating lowest income and second, third, and fourth indicating higher income) based on the entire Danish population according to age group and calendar year. The most recently available income level at 1 to 5 years prior to the index date was used. Females with missing data on income were assigned the lowest (first) income quartile as an informed imputation strategy, because missing income information most often reflects limited or unstable labor market attachment and lower income. Parity was obtained using the Danish Medical Birth Register. 13 Each woman’s parity is recorded for each pregnancy, and the most recent registration prior to semaglutide initiation was used. For females without recorded parity, this information was derived from the number of live births documented in the register. Parity was subsequently categorized as 0, 1, or 2 or greater. Immigrant status was obtained from the Danish Civil Registration System 16 ; individuals were classified as persons of Danish origin, immigrants, or descendants (ie, persons born in Denmark to parents who were both born outside Denmark and are not Danish citizens). Information on race and ethnicity was not available in the Danish nationwide registries and was therefore not included in this study. Immigrant status was included as a covariate due to patterns of hormonal contraception use as well as the prevalence of obesity and the uptake of pharmacologic weight management, including semaglutide, that differ across groups and may therefore confound the association. Data were analyzed from November 25, 2025, to May 18, 2026. Categorical variables are expressed in counts and percentages, while continuous variables are expressed as medians and IQRs. Adjusted hazard ratios (AHRs) with 95% CIs and 2-sided P values were estimated using conditional logistic regression models adjusted for age, educational level, income, immigrant status, parity, and accumulated duration of hormonal contraception use in 1-year intervals. Statistical significance was defined as a 95% CI that did not include 1.00. Covariates were selected a priori based on clinical knowledge on established factors occurring before the exposure, associated with both hormonal contraceptive use and semaglutide initiation, that were available through the Danish nationwide registers. The primary analysis was repeated within subgroups of BMI categories (<25.0 or ≥25.0), age, parity (0, 1, or ≥2), ISCED educational level (0-2, 3, 5-6, or 7-8), income level (first, second, third, or fourth quantile), immigrant status (Danish origin, immigrant, or descendant), and semaglutide type (therapy 1 or 2). In an additional analysis, the main model was further adjusted for BMI category (underweight [<18.5], normal weight [18.5-24.9], overweight [25.0-29.9], or obesity [≥30.0]). Furthermore, we conducted a separate analysis in which exposure was limited to the 2-year period prior to the index date. All statistical analyses were performed using R, version 4.4.1 (R Project for Statistical Computing).

We included 22 694 semaglutide users and 229 640 matched controls (median age, 37 [IQR, 30-44] years) ( Figure 1 ), with similar baseline characteristics, apart from BMI category, where a greater proportion of semaglutide users than controls were categorized as having overweight or obesity (11 157 of 14 949 [74.6%] vs 42 634 of 134 108 [31.8%]) ( Table ). A total of 2246 females (0.9%) with missing data on educational level and 1903 females (0.8%) with missing data on income were imputed. a Participants could be excluded for more than 1 reason. Abbreviations: BMI, body mass index (calculated as weight in kilograms divided by height in meters squared); ISCED, International Standard Classification of Education; NA, not applicable. Refers to individuals who were born in Denmark to parents who were both born outside Denmark and are not Danish citizens. All hormonal contraception utilization patterns were associated with semaglutide initiation compared with never users of hormonal contraception ( Figure 2 ). Among utilization patterns involving a single contraceptive type, AHRs ranged from 1.42 (95% CI, 1.34-1.51) for combined oral tablets to 1.63 (95% CI, 1.46-1.82) for progestin-only IUDs. For utilization patterns involving 2 or more contraceptive types, AHRs ranged from 1.64 (95% CI, 1.47-1.82) for combined oral tablets followed by progestin-only oral tablets to 2.11 (95% CI, 1.97-2.25) for other utilization patterns. AHRs were adjusted for age, educational level, income, immigrant status, parity, and accumulated duration of hormonal contraception use in 1-year intervals. IUD indicates intrauterine device; NA, not applicable. When adjusting for BMI category in the subgroup of the study population with available BMI (n = 149 057), all associations persisted ( Figure 3 ). We also stratified the analysis by BMI less than 25.0 (underweight and normal weight) and 25.0 or greater (overweight and obesity) ( Figure 4 ). For the subgroup of females with normal weight and underweight, hormonal contraception utilization patterns were associated with semaglutide initiation compared with never users of hormonal contraception, with the exception of only having used the progestin-only oral tablet (AHR, 1.08 [95% CI, 0.56-2.11]). Among females with overweight and obesity, all hormonal contraception utilization patterns remained associated with semaglutide initiation. Based on a subgroup of the population with available body mass index measurements from the Danish Medical Birth Register. AHRs were adjusted for body mass index, age, educational level, income, immigrant status, parity, and accumulated duration of hormonal contraception use in 1-year intervals. IUD indicates intrauterine device; NA, not applicable. Based on a subgroup of the population with available BMI measurements from the Danish Medical Birth Register. BMI (calculated as weight in kilograms divided by height in meters squared) stratified between those with normal weight or underweight (BMI <25.0) and those with overweight or obesity (BMI ≥25.0). AHRs were adjusted for age and accumulated duration of hormonal contraception use in 1-year intervals. IUD indicates intrauterine device; NA, not applicable. In age-stratified analyses (<30 vs ≥30 years), AHRs for semaglutide initiation were statistically significant for nearly all hormonal contraception utilization patterns, but in females younger than 30 years, use of progestin-only IUDs was not associated with semaglutide initiation (AHR, 1.17 [95% CI, 0.94-1.46]) (eFigure 1 in Supplement 1 ). In income-stratified analyses, hormonal contraception utilization patterns were associated with semaglutide initiation across all income quartiles, except for progestin-only oral contraception in the third income quartile (AHR, 1.22 [95% CI, 0.92-1.62]) (eFigure 2 in Supplement 1 ). Across educational levels, all hormonal contraception utilization patterns were associated with semaglutide initiation, with overlapping effect estimates across ISCED categories (eFigure 3 in Supplement 1 ). In parity-stratified analyses, most hormonal contraception utilization patterns were associated with semaglutide initiation. Among females with parity of 2 or more, progestin-only oral contraception was not associated with semaglutide initiation (AHR, 1.28 [95% CI, 1.00-1.64]) (eFigure 4 in Supplement 1 ). Across immigrant status groups, all hormonal contraception utilization patterns were associated with semaglutide initiation (eFigure 5 in Supplement 1 ). When initiation of the semaglutide therapy types were analyzed separately, associations were observed across all hormonal contraception utilization patterns in both analyses (eFigure 6 in Supplement 1 ). In a separate analysis restricting exposure to the 2-year period prior to the index date, the estimated associations were attenuated compared with the main analysis. All hormonal contraception exposure categories remained associated with the outcome (eFigure 7 in Supplement 1 ).

In this nationwide nested case-control study, we found that all utilization patterns of hormonal contraception were associated with semaglutide initiation compared with never using hormonal contraception. These findings persisted after adjustment for BMI and were largely consistent across sociodemographic and clinical subgroups. Adjustment for BMI attenuated the observed associations between hormonal contraception use and semaglutide initiation. However, the associations remained across all utilization patterns, suggesting that BMI alone does not fully account for them. This implies that other factors—such as perceived weight changes, health-seeking behavior, contraceptive dissatisfaction or switching patterns, increased contact with health care services, lifestyle factors, or metabolic concerns—may also contribute to the increased likelihood of initiating pharmacologic weight management among females with a history of hormonal contraception use. In the subgroup analyses, a few analyses involving females who exclusively used 1 type of progestin-only contraceptive throughout the observation period were not associated with semaglutide initiation, including among females with normal or underweight BMI, females younger than 30 years, females in the third income quartile, and females with parity of 2 or greater. This observation aligns with existing evidence suggesting that potential weight changes associated with hormonal contraception are primarily estrogen driven rather than caused by progestin alone. 1 Furthermore, associations remained when exposure was restricted to the 2 years preceding the index date, suggesting that more proximate hormonal contraception use was also associated with semaglutide initiation. This finding strengthens the relevance of the observed associations, suggesting that recent hormonal contraceptive use reflect females’ health perceptions and health care-seeking behavior at the time of semaglutide initiation. Current evidence on the association between hormonal contraception and weight changes is limited and inconsistent. Two Cochrane systematic reviews on combined hormonal contraception 1 and progestin-only contraceptives 2 found insufficient evidence to draw conclusions on hormonal contraception and weight gain, noting the overall quality of evidence to be low. More recent cohort studies suggest that certain types of hormonal contraception may be associated with modest BMI increases over several years. 3 , 4 Overall, the literature is limited by small sample sizes and a lack of adequate control groups without hormonal contraception. In the present study, we examined the 2 principal semaglutide formulations available on the market in Denmark. The use of semaglutide has increased markedly in recent years, and although Ozempic is not approved for weight management in individuals without type 2 diabetes, approximately one-third of new users in 2022 did not have type 2 diabetes, indicating substantial off-label use. 6 Current evidence shows that this increase is not evenly distributed across the population, with a disproportionate uptake among females, particularly those of higher socioeconomic status. 7 , 8 A major strength of this study is the large sample sourced from the entire Danish population, with comprehensive prescription histories derived from the Danish registries making it possible to assess utilization patterns of hormonal contraception over decades. Another key strength of this study is the availability of BMI measurements for a large proportion of the females included, allowing us to account for BMI in the analyses, which is typically not available in register-based studies. However, BMI measurements were obtained at varying time points before semaglutide initiation, which may have introduced some degree of misclassification. Our study also has limitations. With an observational design, causality cannot be established, and residual confounding by unmeasured factors such as lifestyle or psychosocial motivations cannot be excluded. Moreover, confounding by indication may have influenced the observed associations, as females using hormonal contraception may differ systematically from nonusers in many terms. Additionally, we lack direct measures of perceived weight change or contraceptive-related adverse effects, which may influence both contraceptive use patterns and treatment decisions. Finally, information on nonhormonal contraceptive methods, such as copper IUDs, is not available in the Danish national registers. Consequently, the reference group likely consisted of a heterogeneous population including both females not using contraception and females using nonhormonal contraceptive methods. As contraceptive-seeking behavior itself may be associated with health characteristics and health care utilization, this may have introduced residual confounding and should be considered when interpreting the observed associations.

In this nationwide case-control study of females of reproductive age, use of hormonal contraception was associated with initiation of semaglutide across all contraceptive utilization patterns compared with never use. Associations persisted after adjustment for BMI, and subgroup analyses showed consistent associations across most utilization patterns. These findings highlight the need for further research into factors associated with health care utilization, body weight concerns, and initiation of pharmacologic weight-management treatment among women of reproductive age.

Hormonal contraception is frequently discussed in relation to weight changes, metabolic symptoms, and other hormonally mediated effects that may influence females’ health perceptions and health care-seeking behavior. Although existing evidence remains inconsistent regarding an association between hormonal contraceptive use and objectively measured weight gain, 1 , 2 , 3 , 4 many females report perceived weight changes during use. 5 These perceptions may be clinically relevant, as they may prompt health care contact or contraceptive discontinuation and often increased contact with health care services and consideration of interventions aimed at weight management. Semaglutide, a glucagonlike peptide-1 receptor agonist, was initially approved for the treatment of type 2 diabetes (Ozempic, licensed in Denmark in 2018 and hereinafter referred to as semaglutide therapy 1) and has more recently been introduced as a highly effective pharmacologic treatment for weight management (Wegovy, licensed in Denmark in 2022 and hereinafter referred to as semaglutide therapy 2). 6 In Denmark, the use of semaglutide for weight loss has increased substantially, particularly among females, with a higher prevalence of use among females with higher income levels. 7 , 8 Initiation of semaglutide treatment is typically preceded by concerns related to weight, metabolic health, or obesity-associated comorbidities—factors that may overlap with concerns reported by users of hormonal contraception. Understanding whether hormonal contraception use is associated with semaglutide initiation may therefore provide insight into health care–seeking behavior and patterns of pharmacologic weight management among women of reproductive age, including factors such as perceived weight changes, contraceptive dissatisfaction, switching patterns, increased contact with health care services, and metabolic concerns. We hypothesized that females with a history of hormonal contraception use have a higher likelihood of initiating semaglutide treatment compared with those who never used hormonal contraception. Using nationwide Danish health registers, we investigated this possible association among females of reproductive age. By leveraging comprehensive prescription and health registry data, we sought to improve understanding of patterns of pharmacologic weight-loss treatment initiation and to explore whether hormonal contraceptive use may serve as an early indicator that patients may have a greater likelihood of seeking treatment with semaglutide.