Cardiometabolic, Glycaemic, and Metabolic Safety of Hormonal Contraceptives in Women with Medical Conditions: A Systematic Review

Hormonal contraceptives remain a cornerstone of reproductive health care worldwide, with an estimated 151 million women of reproductive age currently using hormonal methods. 1 These include combined estrogen–progestin formulations and progestin-only preparations, which are widely used for both contraceptive and non-contraceptive indications such as menstrual regulation, dysmenorrhea, endometriosis, and polycystic ovary syndrome. 2 Over the past two decades, increased contraceptive use has contributed to reductions in unintended pregnancies, maternal morbidity, and unsafe abortions. 3 However, the safety and effectiveness of hormonal contraceptives in women with underlying medical conditions remain areas of ongoing clinical concern. 4 Non-communicable diseases such as hypertension, diabetes mellitus, and thromboembolic disorders are increasingly prevalent among women of reproductive age. 5 Globally, approximately 22–25% of women aged 18–49 years have hypertension, while over 60 million women live with diabetes, the majority with type 2 disease. 6 In addition, venous thromboembolism affects an estimated 1–2 per 1,000 adults annually, with risk amplified by hormonal exposure, obesity, smoking, and advancing age. 7 These conditions are particularly relevant to contraceptive decision-making because estrogen-containing contraceptives have been associated with elevations in blood pressure, alterations in glucose and lipid metabolism, and increased risk of venous and arterial thrombosis. 8 Furthermore, the population described as “women with medical conditions” extends beyond specific cardiometabolic disorders to include a broader spectrum of medically complex populations, such as women with endocrine, autoimmune, and metabolic conditions. This diversity introduces heterogeneity in baseline risk profiles, contraceptive exposure, and outcome assessment. 9 Variations in formulation, estrogen dose, route of delivery, duration of use, and study design further complicate interpretation, making it difficult to derive uniform conclusions regarding safety. In clinical practice, contraceptive decision-making requires careful balancing of potential cardiometabolic risks against the well-established risks of unintended pregnancy in women with underlying disease, including adverse maternal and fetal outcomes. A clearer synthesis of the available evidence, with attention to these sources of variability, is therefore essential. This systematic review aims to evaluate the cardiometabolic, glycemic, and metabolic safety of hormonal contraceptive use among women with underlying medical conditions.

The database search identified a total of 1,063 records, comprising 264 records from Web of Science (WoS), 154 from Scopus, and 645 from PubMed. After applying initial database-specific filters, 182 WoS records met document type criteria, 169 were in English, and 72 were published between 2015 and 2025. From Scopus, 74 records were retained after year filtering, 44 met document type criteria, and 40 were published in English. PubMed yielded 269 records within the specified publication years, of which 178 had free full-text availability. Following the removal of duplicates, 50 duplicate records were identified, 24 were resolved and 26 were deleted, leaving 264 unique articles for title and abstract screening. After this screening stage, 151 articles were excluded based on irrelevance to the study objectives, leaving 113 articles eligible for full-text assessment. Full-text screening resulted in the exclusion of 69 articles, primarily due to inappropriate study populations, insufficient outcome data, or non-relevance to hormonal contraceptive efficacy or safety in women with specific medical conditions. Ultimately, 44 studies met all inclusion criteria and were included in the qualitative synthesis. These 44 eligible articles formed the final evidence base for the review ( Figure 1 ). Figure 1 Prisma Flowchart. A flowchart showing the identification and screening process of studies through databases. The flowchart illustrates the process of identifying studies through databases. It begins with the identification phase, where articles are retrieved from Web of Science (N=72), Scopus (N=40) and PubMed (N=178), totaling 290 articles. Articles removed before screening include 50 duplicates recorded, 26 duplicates deleted and 24 duplicates resolved. The screening phase involves 264 articles for title and abstract screening, with 151 deemed ineligible. Next, 113 articles proceed to full-text screening, with 69 found ineligible. Finally, 44 articles are considered eligible for inclusion. Prisma Flowchart. The 44 included studies spanned different geographical locations, study designs, populations, and follow-up durations ( Table 1 ). Europe and North America, each contributing 34.1% (15/44) of the total evidence base, followed by Asia (18.2%, 8/44), Oceania including Australia, New Zealand, and multinational studies (6.8%, 3/44), Africa (4.5%, 2/44), and South America (2.3%, 1/44) ( Figure 2 ). Prospective cohort and registry-based studies were most common study designs, particularly large population-based datasets such as the UK Biobank and national health registers. These were complemented by retrospective cohorts, nested case–control studies, cross-sectional analyses, and a smaller number of randomized and interventional studies. Sample sizes varied widely, from fewer than 50 participants to over two million individuals. Study populations ranged from adolescents and young adults to middle-aged and older women. Several studies focused on contraceptive initiation and metabolic outcomes in younger populations, while others examined long-term cardiovascular and metabolic outcomes in older groups. Selected studies also included postmenopausal populations and special clinical groups, such as women living with HIV and individuals with autoimmune or inherited conditions. Exposure duration varied considerably. Long-term follow-up, often exceeding 10 years, was reported in several large cohorts, while other studies assessed medium-term exposure (6 months to 5 years) or short-term use in cross-sectional designs. Some studies evaluated cumulative or lifetime exposure, whereas others used defined exposure intervals. 10–13 The temporal distribution shows that publication output was 2.3% in 2015 and 2018, 9.1% in 2016 and 2019, 4.5% in 2017 and 2020, 13.6% in 2021, 18.2% in 2022, 15.9% in 2023, 9.1% in 2024, and 11.4% in 2025 ( Figure 3 ). Table 1 Characteristics of Included Studies ID First Author Country Study Design Sample Size Age Distribution Duration of Contraceptive Use Follow-up Duration (Outcome) 1 [ 10 ] UK Prospective cohort 156,787 40–70 (Mean 60.1) NR 12.5 years 2 [ 11 ] Denmark Prospective cohort 2,208,172 24–38 NR 10 years (4–17) 3 [ 14 ] USA Cross-sectional 4,765 Mean 15.2 Variable (since initiation) N/A 4 [ 15 ] New Zealand Secondary analysis 16 NR Chronic use (NR) N/A 5 [ 13 ] Iran Prospective cohort 3,043 30–70 15.6 years NR 6 [ 16 ] USA Observational 987 18–40 NR NR 7 [ 17 ] USA Case–control 1,884 / 7,536 15–49 ≤90 days Event-based 8 [ 18 ] Brazil Retrospective 330 Mean 37.6 18 months NR 9 [ 19 ] Australia Prospective cohort 15,244 18–28 NR 5 years 10 [ 20 ] USA Cross-sectional 29 15–20 6–24 months N/A 11 [ 21 ] Germany Cross-sectional 851 Mean 52 NR N/A 12 [ 22 ] Japan Cross-sectional 151 NR 28–32.5 months N/A 13 [ 17 ] USA Prospective cohort 706 16–45 NR 19–23 months 14 [ 23 ] USA Survey ~1,500 NR N/A N/A 15 [ 24 ] Sweden Cohort 23,029 15–52 NR 6.2 years 16 [ 25 ] Poland Interventional 120 Mean 26.9 6 months 6 months 17 [ 26 ] Spain Cross-sectional 4,043 15–30 2 weeks N/A 18 [ 27 ] USA Retrospective cohort 696,193 Mean 48–62 NR 5.5–6.5 years 19 [ 28 ] Italy Cross-sectional 290 Mean 23.2 ≥3 months N/A 20 [ 12 ] USA Cohort 3,594 Mean 62.1 NR 12.45 years 21 [ 29 ] Saudi Arabia Survey 158 Mean 27.5 NR N/A 22 [ 30 ] Iran Retrospective cohort 100 15–35 0–36 months 3–6 months 23 [ 31 ] Europe RCT 1,571 18–45 9–13 cycles Same as treatment 24 [ 32 ] France Case–control 3,144 / 12,158 51–62 3 months Event-based 25 [ 33 ] USA Retrospective cohort 1,464 Mean 32 ≤8 weeks NR 26 [ 34 ] USA Retrospective cohort 710 18–45 1.25 years 3.6 years 27 [ 35 ] Spain Case–control 221 36–43 Discontinued ~8 years prior Event-based 28 [ 36 ] India Cross-sectional 50 Mean 24.8 ≥6 months N/A 29 [ 37 ] USA Retrospective cohort NR NR NR 2000–2013 30 [ 38 ] Poland Observational 168 18–31 6 months NR 31 [ 39 ] Multinational Prospective cohort 129 25–31 NR NR 32 [ 40 ] Netherlands RCT 180 ≥18 12 weeks 12 weeks 33 [ 41 ] Sweden Cohort 38,327 Mean 26 NR 2.6 years 34 [ 42 ] Finland RCT 24 18–40 9 weeks 9 weeks 35 [ 43 ] South Africa Observational 49 18–35 Long-term (NR) NR 36 [ 44 ] Indonesia Cross-sectional 10,279 15–49 0–24+ months N/A 37 [ 45 ] USA Feasibility RCT 0 Planned Planned 6 months Planned 38 [ 46 ] Canada Prospective 154 Mean 34.6 NR 1 year 39 [ 47 ] Germany Cross-sectional 233 20–54 2–35 years N/A 40 [ 48 ] Kenya Prospective 171 Median 31 3 months 3 months 41 [ 49 ] Saudi Arabia Cross-sectional 324 25–34 NR N/A 42 [ 50 ] South Korea Cross-sectional 3,386 Postmenopausal <30 / ≥30 months N/A 43 [ 51 ] Canada Cross-sectional 2,705 Mean 26 1 month N/A 44 [ 52 ] USA Registry cohort 1,659 Up to 40 years NR Up to age 40 Notes : Duration of contraceptive use refers to the period of hormonal contraceptive exposure, while follow-up duration refers to the time over which outcomes were assessed. NR = Not reported; (eg., cross-sectional studies). Abbreviation : N/A, Not applicable. Figure 2 Geographic Distribution of Included Studies by Continent. The results show that the majority of studies were conducted in Europe and North America, each contributing 34.1% (15/44) of the total evidence base, followed by Asia (18.2%, 8/44), Oceania including Australia, New Zealand, and multinational studies (6.8%, 3/44), Africa (4.5%, 2/44), and South America (2.3%, 1/44). A pie chart showing pooled percentage by continent from 2 percent to 34 percent. POOLED PERCENTAGE (percent) A pie chart showing pooled percentage by continent. Slice labels and values: Europe 34 percent; North America 34 percent; Asia 18 percent; Oceania 7 percent; Africa 5 percent; South America 2 percent. Figure 3 Temporal distribution of included studies by year of publication. The temporal distribution shows that publication output was 2.3% in 2015 and 2018, 9.1% in 2016 and 2019, 4.5% in 2017 and 2020, 13.6% in 2021, 18.2% in 2022, 15.9% in 2023, 9.1% in 2024, and 11.4% in 2025. A line graph showing pooled percentage by year of publication. A line graph titled Pooled Percentage (percent). The horizontal axis represents year, ranging from 2014 to 2026. The vertical axis represents percent, ranging from 0.00 percent to 20.00 percent. The plotted line with markers shows labeled values at each year: 2015, 2.30 percent; 2016, 9.10 percent; 2017, 4.50 percent; 2018, 2.30 percent; 2019, 9.10 percent; 2020, 4.50 percent; 2021, 13.60 percent; 2022, 18.20 percent; 2023, 15.90 percent; 2024, 9.10 percent; 2025, 11.40 percent. The highest labeled point is 18.20 percent at 2022 and the lowest labeled points are 2.30 percent at 2015 and 2018. Characteristics of Included Studies Notes : Duration of contraceptive use refers to the period of hormonal contraceptive exposure, while follow-up duration refers to the time over which outcomes were assessed. NR = Not reported; (eg., cross-sectional studies). Abbreviation : N/A, Not applicable. Geographic Distribution of Included Studies by Continent. The results show that the majority of studies were conducted in Europe and North America, each contributing 34.1% (15/44) of the total evidence base, followed by Asia (18.2%, 8/44), Oceania including Australia, New Zealand, and multinational studies (6.8%, 3/44), Africa (4.5%, 2/44), and South America (2.3%, 1/44). Temporal distribution of included studies by year of publication. The temporal distribution shows that publication output was 2.3% in 2015 and 2018, 9.1% in 2016 and 2019, 4.5% in 2017 and 2020, 13.6% in 2021, 18.2% in 2022, 15.9% in 2023, 9.1% in 2024, and 11.4% in 2025. Cardiometabolic conditions were the most frequently reported comorbidities, including hypertension, diabetes mellitus, obesity, and dyslipidemia across the included studies ( Table 2 ). Hypertension was the most common, identified in 13 studies (16.3%), often occurring alongside other metabolic risk factors and more frequently in large cohort and registry-based studies. Diabetes mellitus and obesity were each reported in 8 studies (10.0%), commonly coexisting with hypertension and dyslipidemia. Dyslipidemia was reported in 6 studies (7.5%). Smoking or tobacco exposure was reported in 6 studies (7.5%), particularly in studies assessing cardiovascular or thrombotic outcomes. Established cardiovascular disease was reported in 5 studies (6.3%), primarily among older or higher-risk populations. Polycystic ovary syndrome was reported in 5 studies (6.3%). Autoimmune and inflammatory conditions, including rheumatoid arthritis and multiple sclerosis, appeared in 4 studies (5.0%), while psychiatric conditions were reported in 3 studies (3.8%). Less frequently reported conditions included asthma, cancer, liver disease, migraine, thyroid disorders, HIV infection, and rare genetic conditions such as long QT syndrome, each identified in ≤3.8% of studies. Table 2 Distribution and Prevalence of Comorbidities Across Included Studies Comorbidity Category Frequency (N) Percentage (%) Reference Hypertension / HTN 13 16.3 [ 10 , 11 , 16 ] Diabetes / DM 8 10.0 [ 12 , 26 ] Obesity / High BMI 8 10.0 [ 14 , 20 , 34 ] Dyslipidemia / Hypercholesterolemia 6 7.5 [ 11 , 32 , 44 ] Smoking / Tobacco 6 7.5 [ 13 , 17 ] Cardiovascular disease 5 6.3 [ 17 , 19 , 33 ] PCOS 5 6.3 [ 20 , 25 , 42 ] Autoimmune (RA/MS/CTD) 4 5.0 [ 29 , 40 , 43 ] Asthma / Respiratory 3 3.8 [ 10 , 19 ] Cancer 3 3.8 [ 10 , 12 ] Hepatic / NAFLD / Hepatitis 3 3.8 [ 33 , 34 ] Psychiatric conditions 3 3.8 [ 24 , 41 ] Thyroid disease 3 3.8 [ 19 , 49 ] GI disorders 2 2.5 [ 19 , 39 ] Migraine 2 2.5 [ 16 , 21 ] Other adjusted covariates 2 2.5 [ 37 , 43 ] Reproductive / gynecologic 2 2.5 [ 38 , 41 ] Trauma / SES 2 2.5 [ 14 , 47 ] Valvular / AF / Stroke 2 2.5 [ 17 , 48 ] Athlete / endurance training 1 1.3 [ 15 ] Cystic fibrosis 1 1.3 [ 45 ] HIV 1 1.3 [ 17 ] Long QT syndrome 1 1.3 [ 52 ] Renal disease 1 1.3 [ 33 ] Seizure disorder 1 1.3 [ 33 ] Notes : Hypertension was most reported (16.3%), followed by diabetes and obesity (10.0% each); dyslipidemia and smoking (7.5% each); cardiovascular disease and PCOS (6.3% each); autoimmune conditions (5.0%); respiratory, cancer, hepatic, psychiatric, and thyroid conditions (3.8% each); gastrointestinal, migraine, reproductive, trauma/SES, and valvular/stroke conditions (2.5% each); and rare conditions including HIV, renal disease, and seizure disorders (1.3% each). Distribution and Prevalence of Comorbidities Across Included Studies Notes : Hypertension was most reported (16.3%), followed by diabetes and obesity (10.0% each); dyslipidemia and smoking (7.5% each); cardiovascular disease and PCOS (6.3% each); autoimmune conditions (5.0%); respiratory, cancer, hepatic, psychiatric, and thyroid conditions (3.8% each); gastrointestinal, migraine, reproductive, trauma/SES, and valvular/stroke conditions (2.5% each); and rare conditions including HIV, renal disease, and seizure disorders (1.3% each). Combined estrogen-containing contraceptives were the most frequently evaluated methods, reported in 79.5% of studies ( Table 3 ). These included combined oral contraceptives and other estrogen–progestin formulations assessed across diverse study populations and designs. Progestin-only methods, including oral formulations, depot medroxyprogesterone acetate injections, and other injectable progestins, were reported in 18.2% of studies, often in populations with contraindications to estrogen exposure or specific medical conditions. Long-acting reversible contraceptives, such as levonorgestrel-releasing intrauterine systems, hormonal intrauterine devices, and subdermal implants, were also reported in 18.2% of studies and were commonly examined in the context of long-term use and specialized clinical populations. Mixed hormonal exposure, involving both combined and progestin-only methods within the same study, was identified in 13.6% of studies, enabling comparative evaluation across contraceptive types. In 6.8% of studies, hormonal contraceptive type was not clearly specified, typically where exposure was assessed as part of broader reproductive health outcomes. Less frequently, postmenopausal hormone therapy and planned estradiol exposure in feasibility studies were each reported in 2.3% of studies. Table 3 Distribution of Hormonal Contraceptive Types Across Included Studies Hormonal Contraceptive Category Frequency (N) Percentage (%) Reference Combined oral / estrogen-containing contraceptives (COC, CHC, OC) 35 79.5% [ 10 , 25 , 36 ] Progestin-only methods (POP, DNG, DMPA, injections) 8 18.2% [ 27 , 31 , 37 ] Long-acting reversible contraception (LARC: IUD, IUS, implant) 8 18.2% [ 17 , 48 ] Mixed hormonal methods (combined + progestin-only reported) 6 13.6% [ 11 , 24 , 49 ] Hormonal contraception not specified (NR/unspecified use) 3 6.8% [ 13 , 33 ] Postmenopausal hormone therapy 1 2.3% [ 32 ] Planned / feasibility intervention (estradiol) 1 2.3% [ 45 ] Notes : Percentages are calculated using the total number of included studies (n = 44) as the denominator. Because some studies reported multiple hormonal contraceptive types, the cumulative percentage exceeds 100%. Distribution of Hormonal Contraceptive Types Across Included Studies Notes : Percentages are calculated using the total number of included studies (n = 44) as the denominator. Because some studies reported multiple hormonal contraceptive types, the cumulative percentage exceeds 100%. Oral hormonal contraceptives were the most frequently reported delivery method, identified in 86.4% of studies ( Table 4 ). These were evaluated across diverse study populations and clinical settings, including large population-based cohorts and disease-specific investigations. Intrauterine systems and devices were reported in 20.5% of studies and were commonly examined in the context of long-term hormonal exposure and gynecological and metabolic outcomes. Injectable contraceptives, primarily depot medroxyprogesterone acetate, were identified in 15.9% of studies, mainly within cohort and population-based investigations involving women with chronic medical conditions. Vaginal ring formulations were also reported in 15.9% of studies, often in comparative analyses of hormonal exposure and metabolic effects. Implantable contraceptives were identified in 13.6% of studies, frequently within long-acting contraceptive categories assessing safety and effectiveness. Transdermal patch methods similarly accounted for 13.6% of studies and were evaluated alongside oral and vaginal formulations in comparative studies. A smaller proportion of studies (11.4%) reported mixed, unspecified, or control-only contraceptive categories. Table 4 Distribution of Hormonal Contraceptive Delivery Methods Across Included Studies Contraceptive Delivery Method Frequency (N) Percentage (%) Reference Oral hormonal methods 38 86.4% [ 10 , 25 , 51 ] Injectable methods (eg., DMPA) 7 15.9% [ 48 , 49 , 53 ] Intrauterine systems/devices (IUS/IUD) 9 20.5% [ 18 , 19 , 35 ] Implants 6 13.6% [ 19 , 33 , 39 ] Vaginal ring 7 15.9% [ 16 , 17 , 42 ] Transdermal patch 6 13.6% [ 11 , 16 , 46 ] Route not clearly reported / mixed or controls 5 11.4% [ 13 , 27 , 37 ] Notes : Percentages are calculated using the total number of included studies (n = 44) as the denominator. Because some studies reported multiple contraceptive delivery methods, the cumulative percentage exceeds 100%. Distribution of Hormonal Contraceptive Delivery Methods Across Included Studies Notes : Percentages are calculated using the total number of included studies (n = 44) as the denominator. Because some studies reported multiple contraceptive delivery methods, the cumulative percentage exceeds 100%. Reporting of hormonal contraceptive dose across the included studies was heterogeneous and incomplete ( Table 5 ). The majority of studies did not specify exact hormonal doses and reported contraceptive exposure categorically or omitted dose details altogether. Where dose information was provided, combined oral contraceptives most commonly contained low-to-moderate doses of ethinylestradiol, typically ranging from 20 µg to 35 µg. Reported formulations included ethinylestradiol 30 µg combined with desogestrel 150 µg, drospirenone 3 mg, or levonorgestrel 150 µg. Lower estrogen exposure was also reported in vaginal ring users, delivering approximately 15 µg ethinylestradiol per day combined with etonogestrel 120 µg per day. Progestin-only dose reporting was less frequent. One study evaluated a drospirenone-only oral contraceptive at 4 mg, while another reported estradiol at 2 mg/day in a planned intervention. Injectable progestin exposure was generally described qualitatively, with standard-dose depot medroxyprogesterone acetate reported without specific quantification. Some large observational studies classified estrogen exposure into categories (low, intermediate, or high dose) for oral formulations and grouped transdermal exposure by dose ranges (eg., 50 µg), allowing relative comparison without precise quantification. Long-acting reversible contraceptive dosing was rarely specified, with only one study reporting a levonorgestrel intrauterine system containing 13.5 mg. Table 5 Distribution of Hormonal Contraceptive Dose Reporting Across Included Studies Dose Reporting Category Description Reference Dose not reported No quantitative hormonal dose provided [ 14 ] Low–moderate estrogen dose (20–35 µg EE) Most commonly reported combined oral contraceptive estrogen range [ 11 , 30 , 51 ] Specific combined oral contraceptive formulations Explicit estrogen–progestin combinations reported [ 25 , 38 ] Progestin-only formulations Oral or injectable progestin doses specified [ 31 , 53 ] Long-acting reversible contraception (dose specified) Levonorgestrel-releasing intrauterine system content [ 41 ] Dose category classification Hormone exposure reported as low/intermediate/high [ 32 ] Notes : The table summarizes categories of dose reporting, showing that many studies did not provide quantitative dose information, while reported data most commonly described low-to-moderate ethinylestradiol doses (20–35 µg) in combined oral contraceptives. Other categories include specific formulation reporting, progestin-only dosing, limited reporting for long-acting methods, and dose classification approaches. Distribution of Hormonal Contraceptive Dose Reporting Across Included Studies Notes : The table summarizes categories of dose reporting, showing that many studies did not provide quantitative dose information, while reported data most commonly described low-to-moderate ethinylestradiol doses (20–35 µg) in combined oral contraceptives. Other categories include specific formulation reporting, progestin-only dosing, limited reporting for long-acting methods, and dose classification approaches. The most common comparator was non-use of hormonal contraception, reported in 40.9% of studies, defined as women who had never used or were not currently using hormonal methods ( Table 6 ). In studies involving women with preexisting cardiometabolic conditions, including hypertension, this comparator typically consisted of affected individuals not exposed to hormonal contraception, allowing assessment of the independent effect of hormonal exposure within the same clinical population. Non-hormonal or natural reference groups were reported in 13.6% of studies, including women using non-hormonal methods such as copper intrauterine devices or condoms, as well as naturally menstruating women. In cardiometabolic populations, these groups served as clinically relevant comparators by minimizing hormonal influence while maintaining similar baseline risk profiles. Disease-specific or untreated condition comparators were also reported in 13.6% of studies, particularly in endocrine and gynecologic populations. These included comparisons between treated and untreated individuals with conditions such as polycystic ovary syndrome and endometriosis, and in some cases extended to cardiometabolic conditions where treatment status or exposure to hormonal therapy was evaluated. Head-to-head hormonal comparisons (13.6%) assessed differences between contraceptive formulations, including combined oral contraceptives and progestin-only methods, and were particularly relevant in populations with contraindications to estrogen, such as women with hypertension or elevated cardiovascular risk. Healthy reference populations (9.1%) were used less frequently and typically consisted of women without known chronic disease or contraindications, limiting their comparability in studies focused on cardiometabolic risk. Alternative experimental comparators (6.8%) included placebo, vitamin supplementation, or pre-intervention baseline states, while a small proportion of studies (2.3%) did not include an explicit comparator group. Table 6 Comparator Categories Used Across Included Studies Comparator Category Description Frequency (N) Percentage (%) Reference Non-users of hormonal contraception Comparison between women using hormonal contraception and those who had never used or were not currently using any hormonal method 18 40.9 [ 10 , 17 , 32 ] Non-hormonal or natural controls Women using non-hormonal methods (eg., copper IUD, condoms) or naturally menstruating women 6 13.6 [ 15 , 52 , 53 ] Healthy reference populations Participants without known contraindications, chronic disease, or reproductive pathology used as baseline comparators 4 9.1 [ 16 , 38 ] Disease-specific or untreated condition controls Women with the same medical condition who were untreated or not receiving hormonal therapy 6 13.6 [ 20 , 22 , 36 ] Head-to-head hormonal comparisons Direct comparisons between different hormonal contraceptive formulations or regimens 6 13.6 [ 25 , 37 , 42 ] Alternative intervention or experimental controls Comparisons involving placebo, vitamin supplementation, or pre-intervention baseline states 3 6.8 [ 40 , 45 , 48 ] No explicit comparator Descriptive or utilization studies without a defined comparison group 1 2.3 [ 21 ] Notes : The results show that the most common comparator was non-use of hormonal contraception (40.9%). Disease-specific controls, head-to-head hormonal comparisons, and non-hormonal or natural controls were each used in 13.6% of studies. Healthy reference populations were reported in 9.1%, while alternative experimental comparators accounted for 6.8%. Only one study (2.3%) had no explicit comparator. Comparator Categories Used Across Included Studies Notes : The results show that the most common comparator was non-use of hormonal contraception (40.9%). Disease-specific controls, head-to-head hormonal comparisons, and non-hormonal or natural controls were each used in 13.6% of studies. Healthy reference populations were reported in 9.1%, while alternative experimental comparators accounted for 6.8%. Only one study (2.3%) had no explicit comparator. Medium-term exposure (3–12 months) was the most frequently reported duration, identified in 11 studies (25.0%), followed by cross-sectional or survey-based designs without defined exposure in 8 studies (18.2%), long-term exposure (≥1 year) in 7 studies (15.9%), short-term exposure (<3 months) in 6 studies (13.6%), and prescription- or record-based exposure in 4 studies (9.1%) ( Table 7 ). Exposure duration was not reported or remained unclear in 7 studies (15.9%), and one study (2.3%) reported no exposure due to non-initiation of the planned intervention. Long-term exposure was primarily reported in large population-based cohorts and registry studies using person-years, multi-year datasets, or lifetime exposure, with follow-up extending up to a decade. Medium-term exposure was commonly evaluated in prospective cohort and interventional or comparative studies. Short-term exposure included recall-based designs, cycle-based assessments, and short experimental or crossover studies. Cross-sectional and survey-based studies assessed contraceptive use at a single time point without defined longitudinal exposure, limiting causal interpretation. Prescription- or record-based exposure reported variable duration based on individual use patterns, typically derived from administrative or healthcare databases. Table 7 Exposure Time Characteristics of Included Studies Exposure Time Category Definition / Examples Frequency (N) Percentage (%) Reference Long-term follow-up (≥1 year or person-years reported) Person-years, multi-year cohorts, lifetime or registry-based exposure 7 15.9 [ 10 , 11 , 46 ] Medium-term exposure (3–12 months) Fixed follow-up periods such as 3, 6, or 12 months 11 25.0 [ 30 , 32 , 48 ] Short-term exposure (<3 months) Weeks, cycles, recall windows, testing sessions 6 13.6 [ 31 , 40 , 51 ] Cross-sectional / survey-based (no defined exposure duration) Survey snapshots or non-longitudinal designs 8 18.2 [ 14 , 24 , 35 ] Prescription- or record-based exposure (duration variable) Exposure inferred from prescriptions or administrative data 4 9.1 [ 12 , 19 , 34 ] Exposure not reported / unclear Duration not specified or stated as NR 7 15.9 [ 13 , 16 , 18 ] No exposure (trial not initiated / NA) Planned or inactive exposure 1 2.3 [ 45 ] Notes : The results show that medium-term exposure (3–12 months) was reported in 25.0% of studies, followed by cross-sectional designs without defined duration (18.2%) and long-term exposure (≥1 year) in 15.9%. Short-term exposure (<3 months) accounted for 13.6%, while prescription- or record-based estimates were used in 9.1%. Exposure duration was not reported in 15.9% of studies. Exposure Time Characteristics of Included Studies Notes : The results show that medium-term exposure (3–12 months) was reported in 25.0% of studies, followed by cross-sectional designs without defined duration (18.2%) and long-term exposure (≥1 year) in 15.9%. Short-term exposure (<3 months) accounted for 13.6%, while prescription- or record-based estimates were used in 9.1%. Exposure duration was not reported in 15.9% of studies. Cardiometabolic outcomes were evaluated across three domains, including blood pressure, glycaemic control, and lipid/metabolic profiles ( Table 8 ). Blood pressure outcomes were assessed in 11 studies (25.0%). 5 studies reported no significant change, 4 studies reported increases in systolic blood pressure, and 2 studies reported reductions or lower baseline values. These outcomes varied across studies due to differences in measurement approaches (eg., single vs repeated measurements, clinic-based vs self-reported readings), definitions of hypertension, and reporting of systolic and diastolic parameters. Glycaemic outcomes were reported in 7 studies (15.9%). Among these, 4 studies reported no significant changes in fasting blood glucose or HbA1c, while 3 studies reported adverse metabolic effects, including impaired insulin sensitivity and increased glucose responses during oral glucose tolerance testing, particularly among women with polycystic ovary syndrome. No studies reported consistent improvements in glycaemic parameters. Lipid and metabolic outcomes were assessed in 14 studies (31.8%). 6 studies reported adverse changes, most commonly increases in triglyceride levels and other metabolic markers, 5 studies reported no significant changes, and 3 studies reported neutral or favorable lipid profiles, including stable or improved HDL levels. Effects on total cholesterol, LDL, and HDL varied across studies. Additional findings included increased homocysteine levels, oxidative stress markers, inflammatory biomarkers, and weight gain in selected cohorts. Table 8 Distribution of Blood Pressure, Glycemic, and Metabolic Outcomes Across Included Outcome Domain Studies Assessing Outcome Direction of Effect Reported Reference Blood pressure 11 (25.0%) Mixed (no change, slight increase, or decrease) [ 22 , 30 , 42 ] Glycemic outcomes (FBG/HbA1c/OGTT) 7 (15.9%) Mostly neutral; some worsening insulin resistance [ 25 , 36 , 42 ] Lipid/metabolic outcomes 14 (31.8%) Mild lipid changes; triglycerides most affected [ 10 , 12 , 42 ] Notes : The results show that blood pressure outcomes (25.0%) had no consistent change, with occasional increases in some cohorts. Glycaemic outcomes (15.9%) were unchanged, with limited reports of impaired insulin sensitivity. Lipid and metabolic outcomes (31.8%) most commonly showed increased triglycerides, while total cholesterol and LDL findings were inconsistent. Distribution of Blood Pressure, Glycemic, and Metabolic Outcomes Across Included Notes : The results show that blood pressure outcomes (25.0%) had no consistent change, with occasional increases in some cohorts. Glycaemic outcomes (15.9%) were unchanged, with limited reports of impaired insulin sensitivity. Lipid and metabolic outcomes (31.8%) most commonly showed increased triglycerides, while total cholesterol and LDL findings were inconsistent. Assessment of risk of bias using the ROBINS-I tool demonstrated variability across domains among the included studies ( Figure 4 ). For bias due to confounding (D1), most studies were judged to be at low risk, with a smaller proportion assessed as having moderate risk. Bias due to selection of participants (D2) showed greater variability, with studies distributed across low, moderate, and serious risk categories. Bias in classification of interventions (D3), bias due to deviations from intended interventions (D4), and bias due to missing data (D5) were predominantly assessed as low risk across studies. For bias in measurement of outcomes (D6), most studies were judged to be at low risk, although a subset demonstrated moderate and serious risk. Bias in selection of the reported result (D7) was largely assessed as low risk, with some studies judged to be at moderate risk. Overall risk of bias judgments reflected the highest level of risk identified across individual domains, with the majority of studies classified as low risk and a smaller proportion categorized as having moderate or serious overall risk of bias ( Figure 5 ). Figure 4 ROBINS-I risks of bias assessment across included studies. Traffic-light plot showing domain-specific and overall risk of bias judgments for each study according to the ROBINS-I tool. Domains assessed include bias due to confounding (D1), selection of participants (D2), classification of interventions (D3), deviations from intended interventions (D4), missing data (D5), measurement of outcomes (D6), and selection of the reported result (D7). Green indicates low risk, yellow indicates moderate risk, and red indicates serious risk. Table showing ROBINS-I risk of bias assessments across studies for seven domains and overall judgment. The table lists 44 individuals with assessments across eight columns: D1 to D7 and Overall. Most entries have 'low' ratings across all columns. Exceptions include Tayachew (D1 and Overall serious), Carrasco-Garrido (D7 and Overall serious), Paarth Jain (D4 and Overall serious), Samson (D4 and Overall serious), Wu (D1 and Overall serious) and Bernard (D7 moderate, Overall moderate). Figure 5 Summary of ROBINS-I risk of bias judgments by domain. Bar chart presenting the proportion of studies assessed as low, moderate, or serious risk of bias for each ROBINS-I domain and for overall risk of bias. A stacked horizontal bar graph showing ROBINS I risk of bias judgments by domain and overall risk of bias. A stacked horizontal bar graph titled 'Summary of ROBINS I risk of bias judgments by domain' displays bias percentages across domains. The x-axis shows percent from 0 to 100, while the y-axis lists domains: Bias due to confounding, participant selection, intervention classification, deviations from intended interventions, missing data, outcome measurement, reported result selection and overall bias risk. Legend categories are Low, Moderate and Serious risk. Confounding bias is 80% low risk, 20% moderate. Participant selection is 55% low, 33% moderate, 12% serious. Intervention classification, deviations and missing data are 100% low risk. Outcome measurement is 70% low, 24% moderate, 6% serious. Reported result selection is 80% low, 20% moderate. Overall bias risk is 70% low, 25% moderate, 5% serious. ROBINS-I risks of bias assessment across included studies. Traffic-light plot showing domain-specific and overall risk of bias judgments for each study according to the ROBINS-I tool. Domains assessed include bias due to confounding (D1), selection of participants (D2), classification of interventions (D3), deviations from intended interventions (D4), missing data (D5), measurement of outcomes (D6), and selection of the reported result (D7). Green indicates low risk, yellow indicates moderate risk, and red indicates serious risk. Summary of ROBINS-I risk of bias judgments by domain. Bar chart presenting the proportion of studies assessed as low, moderate, or serious risk of bias for each ROBINS-I domain and for overall risk of bias.

This systematic review demonstrates that the cardiometabolic, glycemic, and metabolic safety of hormonal contraceptive use varies across contraceptive type, population risk profile, exposure duration, and outcome measures. While most studies report no clinically significant adverse effects in the general population, important variations emerge in women with underlying cardiometabolic conditions. Hypertension, diabetes mellitus, and obesity were the most frequently reported comorbidities, highlighting the need for careful risk stratification in contraceptive decision-making. The evidence indicates that combined estrogen-containing contraceptives remain the most studied, despite a clinical shift toward progestin-only and long-acting reversible methods, particularly among high-risk populations. This mismatch between research focus and clinical practice limits the strength of current evidence supporting guideline recommendations. In addition, inconsistencies in dose reporting, exposure duration, and outcome measurement further constrain interpretation of risk and limit the ability to establish clear dose–response relationships. Although blood pressure and glycemic outcomes were largely neutral, subgroup-specific effects particularly in women with pre-existing metabolic vulnerability suggest the importance of individualized contraceptive counseling and monitoring. The consistent observation of triglyceride elevation further emphasizes the metabolic implications of estrogen-containing formulations. Overall, the findings support a tailored, risk-based approach to contraceptive selection, with increased attention to progestin-only and non-oral methods in women with cardiometabolic risk factors. Strengthening methodological rigor through standardized reporting, inclusion of high-risk populations, and well-designed longitudinal studies will be essential to improve evidence-based guidance and optimize contraceptive safety. Hormonal contraceptives showed mostly neutral cardiometabolic and glycaemic effects across studies, with some evidence of modest increases in blood pressure, reduced insulin sensitivity in specific subgroups, and consistent triglyceride elevation, particularly with estrogen-containing methods; however, variability in study design and reporting limits precise estimation of risk.

This systematic review synthesizes evidence on the cardiometabolic, glycemic, and metabolic safety of hormonal contraceptive use among women with underlying medical conditions. While no uniform pattern of harm is observed, the findings vary by contraceptive type, population risk profile, exposure duration, and outcome domain. The current finding revealed that hypertension was the most frequently reported comorbidity, followed by diabetes mellitus and obesity. This finding indicates the growing global prevalence of cardiometabolic risk factors among women of reproductive age. 7 This pattern is consistent with global epidemiological data reporting increase in the rates of hypertension and metabolic disease in younger populations, particularly in urban and transitioning economies. 54 , 55 The co-occurrence of these conditions in these studies revealed the clinical reality that contraceptive decision-making often occurs in the context of multiple overlapping risk factors. 56 The prevalence of cardiometabolic comorbidities has important implications for contraceptive safety evaluation. Many global guidelines classify estrogen-containing contraceptives as higher risk in women with uncontrolled hypertension, diabetes with vascular complications, or established cardiovascular disease. 57 However, the heterogeneity observed in outcome reporting across the included studies limits the ability to draw uniform conclusions about risk magnitude. The relatively lower frequency of reported thrombotic disorders and genetic conditions, such as long QT syndrome, reflects both their lower population prevalence and the tendency for such conditions to be excluded from contraceptive trials. Thus, creating a research gap in the populations where safety data are most clinically relevant. Future research should prioritize well-designed longitudinal studies in high-risk and underrepresented populations, with standardized reporting of cardiometabolic outcomes to enable clearer, evidence-based contraceptive risk stratification. Furthermore, combined estrogen-containing contraceptives were the most frequently evaluated methods, consistent with global use and long-standing role as the reference standard for hormonal contraception. 58 The continued prevalence of combined formulations in the literature contrasts with global shifts toward increased use of progestin-only and long-acting reversible contraceptives. 59 This is particularly among women with medical contraindications to estrogen. This discrepancy suggests that research has not fully kept pace with changes in clinical practice. Progestin-only methods and long-acting reversible contraceptives were less frequently studied, despite their favorable safety profiles in women with cardiometabolic disease as suggested by international guidelines. 60 This imbalance limits the strength of evidence supporting guideline recommendations and highlights the need for more comparative studies focused on non-estrogen methods. Delivery route analysis further demonstrated a prevalence of oral formulations, with fewer studies examining injectable, intrauterine, implantable, or transdermal routes. Given that pharmacokinetics and systemic hormone exposure vary by delivery method, the underrepresentation of non-oral routes limit risk assessment. 61 These findings highlight the need to prioritize progestin-only and long-acting methods in high-risk women and to improve comparative evidence and reporting of delivery routes to support accurate, risk-based contraceptive decisions. Moreover, dose reporting across the included studies was inconsistent and frequently incomplete, a limitation that has been widely acknowledged in global contraceptive safety literature. Many large observational studies categorized exposure simply as use versus non-use, without specifying hormonal dose, formulation, or regimen. This approach facilitates large-scale analyses, it obscures potential dose–response relationships and limits mechanistic interpretation. 62 Where dose data were reported, most combined oral contraceptives contained low-to-moderate doses of ethinylestradiol, reflecting contemporary prescribing practices. However, even within this narrow range, emerging evidence suggests that small differences in estrogen dose may influence blood pressure, lipid metabolism, and thrombotic risk, particularly in susceptible populations. 63 The lack of standardized dose reporting therefore represents methodological gap, with implications for both clinical decision-making and guideline development. Exposure time varied across studies, ranging from short-term experimental interventions to multi-decade registry follow-up. Long-term exposure assessments, though relatively uncommon, provided the strongest evidence for cumulative risk evaluation, particularly for cardiovascular and metabolic outcomes. These findings align with global evidence indicating that duration of hormonal contraceptive use is a key modifier of risk, especially for hypertension and thrombotic events. 63 Blood pressure outcomes across the included studies were mixed, with most investigations reporting no clinically meaningful changes, while a subset identified modest increases associated with longer duration of oral contraceptive use. This pattern is consistent with global meta-analyses showing small average increases in systolic blood pressure among combined oral contraceptive users, with greater effects observed in women with pre-existing hypertension or prolonged exposure. 64 The implication is that while short-term use may be relatively safe for many women, long-term use warrants closer monitoring, particularly in those with additional risk factors. 4 Glycemic outcomes were infrequently assessed and often secondary, reflecting a broader trend in contraceptive research. The neutral findings observed in most studies align with global evidence suggesting minimal impact of hormonal contraceptives on fasting glucose and HbA1c in healthy women. 65 However, the observed worsening of insulin sensitivity in specific subgroups, such as women with polycystic ovary syndrome, is consistent with mechanistic studies demonstrating progestin-related effects on insulin action. 13 , 66 These findings imply the importance of individualized contraceptive counseling in metabolically vulnerable populations. Metabolic and lipid outcomes were the most commonly reported cardiometabolic parameters, with triglyceride elevation emerging as the most consistent finding. This observation mirrors global literature linking estrogen-containing contraceptives to increased triglyceride levels via hepatic lipid metabolism. 67

Significant heterogeneity in study design and outcomes : The included studies differed widely in design (cross-sectional, cohort, case–control, and interventional), outcome definitions, and analytical approaches, which limited direct comparability and precluded quantitative meta-analysis. Inconsistent reporting of hormonal dose and formulation : Many studies classified exposure simply as hormonal contraceptive use versus non-use without specifying estrogen dose, progestin type, formulation, or regimen, limiting evaluation of dose–response relationships and mechanistic interpretation. Variable and frequently unclear exposure duration : Exposure time ranged from short recall periods to decades of follow-up, and several studies did not report duration explicitly, reducing the ability to assess cumulative or long-term risks. Underrepresentation of progestin-only and long-acting reversible methods : Despite their clinical importance for women with contraindications to estrogen, progestin-only and long-acting reversible contraceptives were examined in relatively few studies. Limited and inconsistent assessment of cardiometabolic outcomes : Blood pressure, glycaemic control, and metabolic parameters were not uniformly evaluated, and approximately half of the included studies did not report cardiometabolic outcomes, restricting comprehensive safety assessment. Geographic concentration in high-income settings : Most studies originated from Europe and North America, with limited representation from low- and middle-income regions, which may restrict the global applicability of the findings. Residual confounding and observational bias : A large proportion of studies were observational and relied on administrative or self-reported data, increasing vulnerability to residual confounding, exposure misclassification, and selection bias. Significant heterogeneity in study design and outcomes : The included studies differed widely in design (cross-sectional, cohort, case–control, and interventional), outcome definitions, and analytical approaches, which limited direct comparability and precluded quantitative meta-analysis. Inconsistent reporting of hormonal dose and formulation : Many studies classified exposure simply as hormonal contraceptive use versus non-use without specifying estrogen dose, progestin type, formulation, or regimen, limiting evaluation of dose–response relationships and mechanistic interpretation. Variable and frequently unclear exposure duration : Exposure time ranged from short recall periods to decades of follow-up, and several studies did not report duration explicitly, reducing the ability to assess cumulative or long-term risks. Underrepresentation of progestin-only and long-acting reversible methods : Despite their clinical importance for women with contraindications to estrogen, progestin-only and long-acting reversible contraceptives were examined in relatively few studies. Limited and inconsistent assessment of cardiometabolic outcomes : Blood pressure, glycaemic control, and metabolic parameters were not uniformly evaluated, and approximately half of the included studies did not report cardiometabolic outcomes, restricting comprehensive safety assessment. Geographic concentration in high-income settings : Most studies originated from Europe and North America, with limited representation from low- and middle-income regions, which may restrict the global applicability of the findings. Residual confounding and observational bias : A large proportion of studies were observational and relied on administrative or self-reported data, increasing vulnerability to residual confounding, exposure misclassification, and selection bias.

This systematic review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA 2020) guidelines. The review aimed to evaluate the efficacy and 2safety of hormonal contraceptive use among women with specific medical conditions, including hypertension, diabetes mellitus, and thrombotic disorders. The review focused exclusively on human studies to capture real-world clinical effectiveness and safety outcomes across diverse healthcare settings. A comprehensive literature search was conducted across three electronic databases, PubMed (MEDLINE), Web of Science (Core Collection), and Scopus. The search covered publications from January 2015 to the date of the final search in 2025 to capture contemporary evidence aligned with current clinical guidelines and evolving patterns in hormonal contraceptive use. The search strategy combined Medical Subject Headings and free-text terms using Boolean operators to identify studies evaluating hormonal contraceptive use among women with hypertension, diabetes, or thrombotic disease and reporting contraceptive efficacy and/or safety outcomes. The PubMed search syntax was structured as follows and adapted appropriately for Web of Science and Scopus: (“Contraceptive Agents, Hormonal” OR “oral contraceptives” OR “combined oral contraceptive*” OR “progestin-only” OR “injectable contraceptive*” OR “contraceptive implant*” OR “hormonal contraceptive*”) AND (“Hypertension” OR “high blood pressure” OR “Diabetes Mellitus” OR “type 1 diabetes” OR “type 2 diabetes” OR “thrombosis” OR “venous thromboembolism” OR “deep vein thrombosis” OR “pulmonary embolism” OR “thrombophilia”) AND (“Treatment Outcome” OR efficacy OR effectiveness OR “pregnancy rate*” OR “Pearl Index” OR safety OR “adverse event*” OR “cardiovascular event*” OR stroke OR “myocardial infarction” OR “blood pressure” OR “glycemic control” OR “HbA1c”). Equivalent search strings were constructed using TITLE-ABS-KEY fields for Scopus and TS fields for Web of Science. Search results from each database were exported as comma-separated value files for record management. All records retrieved from PubMed, Web of Science, and Scopus were imported into Rayyan, a web-based systematic review management platform. Duplicate records were identified using Rayyan’s automated deduplication function and manually verified. A total of 50 duplicate records were detected, of which 24 were resolved and 26 were deleted. After deduplication, 264 unique records were retained for title and abstract screening. Eligibility criteria were defined a priori using the PICOS framework. The population included women of reproductive age with hypertension, diabetes mellitus (type 1 or type 2), or current or previous venous thromboembolism, thrombophilia, or clearly elevated thrombotic risk. Studies with mixed populations were included if subgroup data could be extracted. The intervention included any hormonal contraceptive method, including combined estrogen–progestin contraceptives and progestin-only pills, injectables, implants, or levonorgestrel-releasing intrauterine systems. Continuous and cyclic regimens were eligible. Comparators included non-hormonal methods, no contraception, placebo, or alternative hormonal methods. Primary outcomes included contraceptive efficacy and safety outcomes such as major cardiovascular events, venous thromboembolism, blood pressure changes, and glycaemic outcomes. Secondary outcomes included lipid profiles, body weight or BMI, bleeding patterns, discontinuation due to adverse effects, and other serious adverse events. Eligible study designs included randomized controlled trials, non-randomized interventional studies, cohort studies, and case–control studies. Case reports, case series, editorials, reviews, letters, and conference abstracts were excluded. Only full-text English-language articles were included, with no geographical restrictions. Although the review focused on women of reproductive age with cardiometabolic and thrombotic conditions, studies involving broader or related populations were included to capture the full spectrum of hormonal exposure and associated outcomes. Title and abstract screening were conducted in Rayyan by two independent reviewers. Records were classified as include, exclude, or uncertain according to the eligibility criteria. Rayyan’s blinded screening mode was used to minimize selection bias. Following title and abstract screening, 106 articles were retained for full-text review. Full-text screening resulted in the exclusion of 62 articles due to ineligible populations, lack of hormonal contraceptive exposure, or absence of relevant efficacy or safety outcomes. A total of 44 studies met the eligibility criteria and were included in the final synthesis. The study selection process was documented using a PRISMA 2020 flow diagram. Data extraction was performed independently by two reviewers using a standardized Microsoft Excel form ( supplementary file 1 ). Extracted variables included author, year of publication, country, study design, study setting, sample size, age distribution, medical condition, type and route of hormonal contraceptive, dose, regimen, duration of use or follow-up, comparator characteristics, efficacy outcomes, safety outcomes, metabolic outcomes, and adjustment for confounders. Discrepancies were resolved through discussion and consensus. Due to heterogeneity in study designs, populations, contraceptive formulations, and outcome definitions, data were synthesized descriptively. Studies were grouped by underlying medical condition and by hormonal contraceptive type. Efficacy and safety outcomes were summarized narratively, and patterns across study designs and populations were described. No quantitative meta-analysis was conducted. Study selection and data extraction were conducted independently by at least two reviewers to ensure accuracy and consistency. Rayyan facilitated blinded screening, conflict resolution, and transparent documentation of inclusion and exclusion decisions. Reproducibility was ensured through complete documentation of search strategies, exported records, Rayyan logs, and PRISMA flow records. Risk of bias was assessed using the ROBINS-I tool for non-randomized studies, evaluating domains including confounding, selection bias, intervention classification, deviations from intended interventions, missing data, outcome measurement, and selective reporting.

Standardize exposure reporting : Future studies should consistently report hormonal contraceptive type, dose, formulation, delivery route, and regimen to support meaningful comparison and synthesis. Prioritize longitudinal designs with defined follow-up : Prospective cohorts and randomized studies with clearly specified exposure duration are needed to evaluate cumulative and long-term cardiometabolic risks. Expand research on progestin-only and long-acting reversible methods : Comparative safety studies focusing on non-estrogen contraceptives should be prioritized, particularly among women with hypertension, diabetes, or elevated cardiovascular risk. Harmonize cardiometabolic outcome assessment : Standardized measurement of blood pressure, glycaemic indices, lipid profiles, and clinically relevant endpoints would improve comparability across studies. Increase representation from low- and middle-income regions : Research efforts should be extended to settings with high contraceptive use and rising cardiometabolic disease burden to enhance global relevance. Incorporate risk stratification and subgroup analyses : Studies should assess outcomes according to baseline risk factors such as age, obesity, disease severity, and duration of contraceptive use to support individualized clinical decision-making. Integrate clinical, epidemiological, and mechanistic approaches : Combining population-based data with mechanistic and pharmacokinetic studies would strengthen causal inference and better inform clinical guidelines for women with complex medical conditions. Standardize exposure reporting : Future studies should consistently report hormonal contraceptive type, dose, formulation, delivery route, and regimen to support meaningful comparison and synthesis. Prioritize longitudinal designs with defined follow-up : Prospective cohorts and randomized studies with clearly specified exposure duration are needed to evaluate cumulative and long-term cardiometabolic risks. Expand research on progestin-only and long-acting reversible methods : Comparative safety studies focusing on non-estrogen contraceptives should be prioritized, particularly among women with hypertension, diabetes, or elevated cardiovascular risk. Harmonize cardiometabolic outcome assessment : Standardized measurement of blood pressure, glycaemic indices, lipid profiles, and clinically relevant endpoints would improve comparability across studies. Increase representation from low- and middle-income regions : Research efforts should be extended to settings with high contraceptive use and rising cardiometabolic disease burden to enhance global relevance. Incorporate risk stratification and subgroup analyses : Studies should assess outcomes according to baseline risk factors such as age, obesity, disease severity, and duration of contraceptive use to support individualized clinical decision-making. Integrate clinical, epidemiological, and mechanistic approaches : Combining population-based data with mechanistic and pharmacokinetic studies would strengthen causal inference and better inform clinical guidelines for women with complex medical conditions.