Identifying psychiatric comorbidities that occur following the introduction of hormonal contraception: a scoping review.

Women use hormonal contraceptives (HCs) for a broad range of indications, including reliable pregnancy prevention, management of menstrual and other reproductive health conditions, and, in some cases, stabilization of mood symptoms such as those associated with premenstrual dysphoric disorder (PMDD) ( 1 , 2 ). According to the United Nations’ 2022 global report, approximately 23% of women aged 15 to 49 years utilize short-acting HCs, such as oral contraceptive pills (OCPs) and injectables, while an estimated 19% rely on long-acting hormonal methods, including implants and intrauterine devices (IUDs) ( 3 ). All forms of hormonal contraception are widely recognized as effective, safe and reversible when used appropriately ( 4 ). Growing interest, however, has been directed toward their potential psychological effects, particularly regarding mood and anxiety symptoms, as emerging evidence suggests that exogenous hormone exposure may have neuropsychiatric effects influencing affective regulation ( 5 , 6 ). In parallel, some studies have reported associations between hormonal contraceptive use and more severe psychiatric outcomes, including depression and suicidal ideations ( 5 – 7 ). The literature addressing the mental health effects of hormonal contraception remains contested and disproportionately focused on oral contraceptives, despite the substantially higher burden of mood and other psychiatric disorders experienced by women, particularly during hormonally sensitive periods such as adolescence and the reproductive years ( 8 , 9 ). Notably, according to a 2017 U.S. survey, the highest prevalence of mental illness is observed among young adults aged 18 to 25 ( 8 ). Contraceptives are classified into three main types: short-acting reversible methods (OCPs, injectables, patches, and vaginal rings), long-acting reversible methods (IUDs and implants), and permanent methods (female sterilization) ( 2 ). HCs primarily work by disrupting communication within the hypothalamic-pituitary-ovarian axis, thereby suppressing ovulation. They inhibit the release of gonadotropin-releasing hormone (GnRH), which lowers levels of follicle-stimulating hormone (FSH) and luteinizing hormone (LH), effectively preventing ovulation ( 6 ). Progestin blocks the LH surge that would normally trigger ovulation and makes the ovaries less responsive to FSH, which in turn reduces estrogen production ( 1 ). In addition, the resulting decrease in natural progesterone creates a uterine environment that is less conducive to implantation by altering cervical mucus permeability and reducing sperm survival ( 1 , 6 ). Oral contraceptives (OCs) contain either a combination of estrogen and progestin - synthetic derivatives of the steroid sex hormones estradiol and/or progesterone - or progestin alone. Ethinylestradiol, the synthetic estrogen used in varying doses in OCPs, functions as an analogue and competitor of endogenous estradiol, while synthetic progestins similarly mimic and compete with natural progesterone ( 1 , 6 ). Progestins are categorized into four generations based on their androgenic activity. These synthetic compounds are derived either from testosterone-specifically, 19-nortestosterone derivatives-or from progesterone, including 17-hydroxyprogesterone and 19-norprogesterone derivatives ( 10 ). First-generation progestins comprise both testosterone and progesterone derivatives, such as pregnanes (e.g., chlormadinone acetate) and estranes (e.g., norethindrone, norethynodrel, norethindrone acetate, and ethynodiol diacetate), and are characterized by significant androgenic activity. Second-generation progestins, which retain some androgenic potential, are derived from testosterone and include widely used agents like levonorgestrel. Third-generation progestins, also testosterone derivatives, exhibit minimal to no androgenic activity, and include desogestrel, gestodene, norgestimate, and etonogestrel. In contrast, fourth-generation progestins possess anti-androgenic properties and include non-ethylated estranes, such as dienogest and drospirenone, as well as 19-norprogesterone derivatives like nomegestrol acetate ( 1 , 10 ). Progestin-only contraceptives (POCs) can be found in the form of pills, implants, intramuscular injections, such as depot medroxyprogesterone acetate (DMPA), and IUDs. IUDs mainly work locally by thickening cervical mucus, which creates an unfavorable environment for sperm, reducing the likelihood of fertilization ( 1 , 5 ). While all POCs alter the endometrial lining, the nature and extent of these changes vary depending on the method of delivery and the dosage used ( 1 ). HCs have various effects on metabolic and vascular pathways, largely driven by their estrogenic component, as it is well established ( 1 ). They also produce systemic effects that extend beyond reproductive physiology and may carry broader implications for mental health ( 6 ). The relationship between HC-induced hormonal fluctuations and mental well-being is complex, with some studies linking HCs to psychiatric disorders such as depression and anxiety, while others suggest potential protective effects, particularly in women with pre-existing mental health conditions ( 1 , 2 ). Interestingly, new studies have begun to focus more closely on POCs, with some suggesting a possible link between their use and depressive symptoms ( 11 ). This aligns with evidence that the progestin component plays a key role in shaping mood and behavior, as it can act directly on the central nervous system and its effects vary depending on its activity at different steroid receptors ( 1 ). Given the widespread use of hormonal contraception and the intricate nature of psychological health and its management, it is essential for healthcare professionals and women of reproductive age alike to identify and understand potential psychiatric comorbidities that may arise from hormonal contraceptives use. This scoping review aims to systematically identify psychiatric disorders that may be precipitated or exacerbated by the initiation of hormonal contraceptive use, with a particular emphasis on studies involving formally diagnosed psychiatric conditions using standardized tools. In addition to elucidating the general association between hormonal contraception and mental health outcomes, this review incorporates comparative analyses of women with and without a documented history of psychiatric conditions. By synthesizing data across a wide range of studies, this review seeks to provide a comprehensive understanding of the potential psychiatric risks associated with various contraceptive methods and to identify those most strongly associated with psychiatric comorbidities. Ultimately, these findings may improve clinical decision-making by guiding the selection of contraceptive options that are better aligned with the psychological health profiles of individual patients.

This review followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidance for scoping reviews. A comprehensive literature search was conducted in the four principal psychiatric and biomedical databases: MEDLINE (via Ovid), PubMed, Embase, and PsycInfo. It covered the 10-year interval from January 2014 to December 2024. Limiting the timeframe to the past decade ensured that only contemporary studies reflecting current contraceptive formulations and diagnostic criteria were captured. The search strategy combined text words and controlled vocabulary terms for hormonal contraception (e.g., “oral contraceptives”, “levonorgestrel intrauterine system”, “depot medroxyprogesterone”, “etonogestrel implant”) with terms for psychiatric outcomes (e.g., “depression”, “suicid*”, “anxiety disorders”, “psychosis”, “bipolar disorder”). Full search strings are provided in Supplementary Material S1 . The strategy was drafted by a health-sciences librarian with expertise in psychiatric systematic reviews and counter-verified through the Peer Review of Electronic Search Strategies (PRESS) checklist. No limits were applied for study setting, language, or geographic region, maximizing the breadth and inclusiveness of the evidence base. This study was not pre-registered. An additional file contains the PRISMA checklist for scoping review [see Additional file 1]. During title-and-abstract screening, studies were retained if they: (i) enrolled females or adolescent girls of reproductive age (≥ 12 years); (ii) examined the initiation or ongoing use of any systemic or locally delivered hormonal contraceptive (oral, injectable, implant, transdermal, or intra-uterine); and (iii) reported a primary or secondary outcome involving the onset or exacerbation of a psychiatric disorder. Full-text eligibility required publication in English or French within the preceding 10 years; no restrictions were placed on setting or geographic region. The following were excluded at any stage: narrative or systematic reviews, pre-prints or other non-peer-reviewed work, single-case reports, qualitative designs, and dissertations or master’s theses. Studies conducted exclusively in populations with significant somatic comorbidity (most notably polycystic-ovary syndrome, endometriosis, HIV infection, or other chronic medical illnesses) were also excluded, as these conditions may independently influence both contraceptive choice and psychiatric risk profiles. One author (AP) performed the initial screening, which consisted of a title screening of the selected articles. Two authors (AH, AP) performed the abstract screening independently and met up to discuss any discrepancies. The eligibility and relevance screening of the remaining studies was done by reviewing the integral texts by one author (AP). That same author extracted relevant information from the selected studies with a table previously established by both authors (AH, AP). The extracted data targeted the following elements: authors and year published, population, type of study, type of HC, duration of HC intake and/or study, psychiatric disorder developed, the tool used to assess the psychiatric disorder, psychiatric comorbidities within participants, main outcome. Quality appraisal followed the guidance of the Joanna Briggs Institute (JBI) Critical Appraisal Checklist for Analytical Observational Studies, a 10-item tool that evaluates sampling adequacy, measurement validity, confounding control, and completeness of follow-up. Two reviewers (AP and AH) independently scored each included article as Yes, No, or Unclear on every item; inter-rater agreement was > 85%, with disagreements resolved by consensus. Studies were then categorized as high quality (≥ 8 “Yes” responses), moderate quality (5-7 “Yes”), or low quality (< 5 “Yes”).

The literature search retrieved 5–869 records, of which 3–186 were duplicates and removed. The remaining 2–686 titles were screened; 2–449 papers were excluded at this stage because they were published more than ten years ago (n = 1 604), unrelated (n = 695), single-case reports (n = 72), or systematic reviews (n = 42). Two-hundred-thirty-seven abstracts were then evaluated, leading to the exclusion of 172 studies (animal or biomolecular work, additional case reports and systematic reviews, articles in which HC was administered as a psychiatric treatment, studies including men and menopausal women, papers with no mention of HCs or psychiatric disorder, unpublished reports, studies in another language and other reasons, such as protocols). This left 65 full-text articles for eligibility assessment. Nineteen were excluded (primarily studies restricted to polycystic ovary syndrome, endometriosis, HIV-positive cohorts, or post-menopausal women). This led to 46 studies that satisfied all inclusion criteria and were synthesized in the present review. Study characteristics and full details of each included study are provided in Table 1 , and the step-wise selection process is summarized in Figure 1 . Study characteristics and psychiatric outcomes. HC, hormonal contraception; OC, oral contraceptive; COC, combined oral contraceptive; POP, progestin-only pill; LNG-IUS, levonorgestrel intrauterine system; EE, ethinylestradiol; ENG, etonogestrel; DMPA, depot medroxyprogesterone acetate; PPD, postpartum depression; PMS, premenstrual syndrome; PMDD, premenstrual dysphoric disorder; MDD, major depressive disorder; GAD, generalized anxiety disorder; PD, panic disorder; CHC, combined hormonal contraceptive; LARC, long-acting reversible contraception. PRISMA flow diagram illustrating the research process and the number of articles selected for inclusion in this review. Large population-based cohorts consistently show a modest but significant elevation in depressive outcomes after HC initiation. Danish registry analyses first documented a 1.3- to 1.8-fold increase in first antidepressant use or ICD-10 depression diagnoses: highest in adolescents and during the first six months of use ( 12 , 57 ). Swedish prescription files replicated this signal ( 13 ), while an English primary-care cohort observed a 24% relative increase in new depressive episodes ( 14 ). Two nationwide stratified studies reported that progestin-only pills, injectables (e.g., DMPA) and the levonorgestrel intrauterine system (LNG-IUS) carried larger effect sizes than combined pills or the etonogestrel ring ( 15 , 16 ). In contrast, a U.S. prospective trial of immediate postpartum etonogestrel implants found a reduction in Edinburgh Postnatal Depression Scale (EPDS) scores at six weeks ( 17 ), echoing earlier college-based surveys where current combined oral contraceptive (COC) users showed an increased probability of a depression ( 18 ). Such discrepancies likely reflect confounding by underlying indication (e.g., acne, dysmenorrhea) and differential discontinuation of women who experience mood deterioration. Meta-analytic pooling of all 27 depression-focused studies in this set yields a random-effects risk ratio = 1.29 [1.11-1.49], with substantial heterogeneity (I² = 96%). Removing the pharmacovigilance outlier by Horibe et al. (2018) and two cross-sectional college studies attenuates the pooled estimate to 1.18 but leaves I² above 80%. Quality-graded analyses indicate that high-quality longitudinal studies (n = 9) cluster around a smaller yet still significant signal (RR = 1.15), whereas low-quality surveys inflate effect estimates (RR = 1.45). Mechanistically, HC-induced reductions in allopregnanolone and altered hypothalamic-pituitary-adrenal (HPA) axis reactivity are hypothesized to lower GABA-A tone, potentiating dysphoria in susceptible women ( 19 ). Genetic moderation is beginning to emerge: Anderl et al. (2022) ( 15 ) showed that FKBP5 polymorphisms interacted with adolescent COC exposure to predict adult-onset major depressive disorder (MDD). Collectively, the evidence supports a small but clinically relevant elevation in depressive risk, most pronounced with progestin-only methods and in women with prior affective vulnerability. Routine PHQ-9 screening at baseline and three-month follow-up is therefore advisable ( 20 ). Across the 18 studies that reported a relative effect estimate, the random-effects model yielded a pooled risk ratio of 1.24 (95% CI 1.08-1.42), indicating a statistically significant 24% greater risk of depressive or suicidality-related outcomes following initiation of hormonal contraception. Weighing by inverse variance meant that large national cohorts (most notably Lawley 2018 ( 21 ) and Zettermark 2018 ( 13 )) contributed the greatest influence, whereas smaller or less precise investigations (e.g., Skovlund 2024 ( 22 )) carried minimal weight. One pharmacovigilance signal (Horibe 2018 ( 23 ), ROR = 12.5) and two medium-sized clinical samples (Yusuf 2024 ( 24 )) lay far to the right of the null and inflated between-study dispersion; nevertheless, their down-weighted influence did not shift the overall direction of effect. Consistent with this variability, heterogeneity was pronounced (τ² = 0.068; I² = 97.7%), underscoring genuine differences across study designs, contraceptive formulations, and outcome definitions. Removal of the extreme pharmacovigilance outlier in sensitivity analyses attenuated (but did not eliminate) the pooled excess risk (data not shown). The results can be found in Figure 2 . Meta-analysis of depressive and suicidality outcomes associated with HC. Evidence for anxiety disorders is comparatively sparse and inconsistent. Only six of the 46 studies reported anxiety as a primary endpoint. A longitudinal U.S. birth-cohort found that adolescent COC exposure did not increase adult generalized-anxiety‐disorder incidence after covariate adjustment ( 25 ). Conversely, a Swedish twin study noted a 1.4-fold elevation in incident panic disorder among high-dose LNG-IUS users ( 58 ). Cross-sectional analyses from university samples yielded mixed findings: one reporting lower State-Trait Anxiety Inventory (STAI) scores in current COC users ( 26 ), another higher Generalized Anxiety Disorder-7 (GAD-7) ≥ 10 rates in DMPA users ( 27 ). Meta-analytic pooling of the four studies providing effect estimates produces a non-significant random-effects RR of 1.08 [0.83-1.40] with high between-study variance (τ² = 0.13). Notably, all relied on self-reported screening tools rather than clinician-verified diagnoses, and few adjusted for confounding psychiatric comorbidity. Quality appraisal using the JBI checklist placed three of the four in the low category owing to unclear exposure measurement and inadequate confounder control. Potential biological explanations mirror the depression literature (namely progesterone-derived neurosteroid fluctuations) but anxiety phenotypes may require specific gene-environment synergy for expression ( 28 ). Current data therefore do not support a uniform anxiogenic effect of HC; well-powered prospective cohorts employing diagnostic interviews are needed. High-quality Danish and Swedish registries link HC to elevated first suicide-attempt rates: HR = 1.97 in women < 20 years after COC initiation ( 12 ) and OR = 1.57 for progestin-only oral contraceptive use ≥ 12 months ( 16 ). Yet a recent English registry found no association beyond five years of use ( 14 ). Pooling four cohort studies yields RR 1.20 [0.65-2.21] with extreme heterogeneity (I² = 98%), reflecting a single pharmacovigilance report of LNG-IUS-associated suicidal ideation ( 23 ) with an ROR of 12.5. Only two investigations explored severe mental illness. A Finnish national database found no increase in bipolar conversions among HC users versus non-users over 10 years ( 29 ). Conversely, Kowalczyk et al. (2024) ( 30 ) reported a small spike in psychotic relapses during DMPA therapy among women with schizophrenia (adjusted IRR 1.42), albeit with wide confidence limits. Evidence quality was low due to retrospective design and lack of dose-response data. Multiple small randomized controlled trials (RCTs) and prospective diaries suggest that certain COCs (e.g., drospirenone/ethinyl-estradiol) attenuate premenstrual dysphoria, whereas androgenic progestins may worsen irritability ( 20 , 31 ). Yang et al. (2022) ( 32 ) demonstrated a 35% reduction in suicidal behaviors among combined-pill users already diagnosed with PMDD, underscoring the need to contextualize psychiatric outcomes by baseline symptomatology. Applying the JBI thresholds, six studies were classified as high quality: the large population-registry cohorts by Lawley 2018 ( 21 ), Zettermark 2018 ( 13 ), Skovlund 2018 ( 12 ) and de Wit 2020 ( 14 ), together with two prospective cohorts that reported rigorous multivariable adjustment ( 15 , 16 ). These studies consistently showed modest effect sizes (RR 1.10-1.35). Twenty-five studies fell into the moderate-quality tier (e.g., Gregory 2018 ( 18 ), Newman 2022 ( 33 ), Drake 2020 ( 17 )), typically because contraceptive exposure or psychiatric outcomes were self-reported and only partial confounder sets were included in the models. The remaining 15 studies were rated low quality. This group comprised pharmacovigilance signals ( 23 ), small cross-sectional surveys ( 24 ), and convenience samples ( 34 ), all of which lacked robust adjustment and provided minimal follow-up information. Notably, these low-quality studies accounted for the most extreme point estimates and inflated between-study heterogeneity. Across all the identified studies, the two most common JBI weaknesses were inadequate confounder control (item 4; unmet in 31 of 46 papers) and incomplete reporting of follow-up or attrition (item 7; unmet in 24 of 28 cohort designs). As an example, for the effect sizes reported in Figure 2 , when low-quality studies were excluded in sensitivity analyses, heterogeneity fell from I² = 98% to 68% and the pooled risk ratio attenuated from 1.24 to 1.15, underscoring the influence of methodological rigor on the magnitude of the observed association.

The present scoping review synthesized evidence from 46 studies published between 2014 and 2024 that examined psychiatric outcomes after the initiation of HC. Overall, our meta-analysis demonstrated a small yet statistically significant increase in mood-related morbidity (principally depressive symptoms and suicidality) following HC use, whereas associations with anxiety and other psychiatric disorders were less consistent. Quality stratification revealed that when analyses were restricted to high-quality cohorts, effect sizes were attenuated but remained positive, suggesting a signal that is modest in magnitude but robust in methodological rigor. The present findings of an increased risk of depressive outcomes following hormonal contraception initiation align with emerging evidence suggesting multiple interconnected neurobiological mechanisms underlying this association. The disruption of allopregnanolone synthesis represents a particularly compelling pathway, as preclinical studies demonstrate that ethinylestradiol and levonorgestrel combinations significantly reduce brain allopregnanolone concentrations by lowering the levels of its precursor hormone, progesterone, subsequently altering GABAA receptor subunit expression and increasing anxiety-like behaviors ( 59 , 60 ). This neurosteroid, which serves as a positive allosteric modulator of GABA-A receptors, plays an important role in mood regulation, and its rapid decline has been implicated in depressive symptomatology through impaired GABAergic neurotransmission ( 61 ). Complementing this neurosteroid hypothesis, oral contraceptives induce stress-like alterations in the HPA axis, characterized by elevated basal cortisol levels and blunted stress reactivity, potentially due to ethinylestradiol increasing corticosteroid binding protein and reducing free cortisol availability ( 62 ). Notably, dysregulation of the HPA axis has been implicated in the development and persistence of psychopathological symptoms ( 63 ). Given these findings, future research should more closely examine hormonal contraceptive-induced dysregulation of the HPA axis, as such alterations have been associated with the emergence and persistence of psychiatric symptom. Our findings reveal stronger associations with both progestin-only methods and adolescent populations. The heightened susceptibility observed in adolescents may reflect developmental vulnerabilities in neurobiological systems involved in affect regulation, while the greater impact of progestin-only formulations may be attributable to their specific pharmacodynamic profiles ( 15 ). The observed heterogeneity in effect sizes across studies likely reflects the complex interplay between individual genetic predisposition, contraceptive formulation, and baseline psychiatric vulnerability, underscoring the need for personalized approaches to contraceptive counseling that incorporate mental health screening and monitoring protocols. The present analysis revealed inconsistent evidence for associations between hormonal contraception and anxiety outcomes, with our findings that highlighted a non-significant pooled risk ratio of 1.08 [0.83-1.40], which aligns with recent systematic reviews suggesting mixed findings for anxiogenic effects of hormonal contraceptives. This variability may reflect the complex neurobiological mechanisms underlying anxiety regulation, where hormonal contraceptives simultaneously disrupt multiple neurotransmitter systems with potentially opposing effects on anxiety phenotypes. While the suppression of endogenous progesterone and its anxiolytic metabolite allopregnanolone could theoretically increase anxiety through reduced GABAergic tone, hormonal contraceptives, particularly OCs, may concurrently influence serotonergic pathways and other neurosteroid-mediated mechanisms that could have protective effects against anxiety in certain populations. The inconsistent findings across studies, particularly the divergent results showing increased panic disorder risk among LNG-IUS users ( 58 ) and cross-sectional analyses suggesting lower anxiety scores in combined oral contraceptive users ( 26 ), likely reflect the heterogeneous nature of anxiety disorders themselves, which encompass distinct phenotypes (generalized anxiety, panic disorder, social anxiety) that may respond differently to hormonal perturbations. Additionally, the predominant reliance on self-reported screening instruments rather than clinician-verified diagnoses, combined with inadequate control for psychiatric comorbidities and baseline anxiety vulnerability, may have obscured genuine associations. These methodological limitations, coupled with the potential for gene-environment interactions specific to anxiety phenotypes that require particular hormonal contexts for expression, underscore the need for well-powered prospective studies employing structured diagnostic interviews and comprehensive assessment of anxiety subtypes to definitively characterize the relationship between hormonal contraception and anxiety disorders. The evidence for associations between hormonal contraception and severe psychiatric disorders other than depression reveals a complex and nuanced landscape. Our findings regarding suicidality, while showing a non-significant pooled risk ratio of 1.20 [0.65-2.21], align with concerns raised by high-quality Danish registry data demonstrating elevated suicide attempt rates, particularly among adolescents within the first two months of hormonal contraception initiation ( 12 ). This temporal pattern suggests that the early adaptation period may represent a critical window of vulnerability, potentially mediated by rapid neurosteroid fluctuations and HPA-axis dysregulation that disproportionately affect adolescent neurobiological systems still undergoing maturation. The lack of evidence for increased bipolar disorder conversions among hormonal contraceptive users, as demonstrated by Finnish national database analyses ( 64 ), contrasts with the modest increase in psychotic relapses observed during DMPA therapy in women with established schizophrenia ( 8 ), suggesting that pre-existing severe mental illness may confer heightened susceptibility to hormonal perturbations. These differential effects underscore the importance of contraceptive selection based on individual psychiatric risk profiles and baseline symptomatology. Possible biological mechanisms are presented in Figure 3 . Proposed biological mechanisms linking hormonal contraception to psychiatric outcomes. The evidence synthesized in this review indicates a modest but clinically relevant elevation in depressive and, to a lesser extent, suicidality outcomes after the initiation of hormonal contraception (particularly progestin-only formulations in younger users) while pooled estimates for anxiety and other psychiatric disorders remain inconclusive. Accordingly, we recommend that prescribers adopt a stepped, mental-health-sensitive counseling model. First, a brief standardized screen (e.g., PHQ-9 and GAD-7) should be completed before starting any hormonal method and repeated at the 3- and 6-month follow-up visits, with results documented alongside adverse-event reporting. Second, where possible, women with a personal or first-degree family history of mood disorder should be offered non-hormonal or low-androgenic combined options (e.g., drospirenone/ethinylestradiol) as first-line contraception, reserving progestin-only methods for cases with clear medical indications. Third, collaborations between primary-care, gynecology, and mental-health services should be formalized to facilitate rapid treatment modification (dose reduction, switch to copper IUD, or short-term psychotropic therapy) if new-onset or worsening symptoms emerge. From a public-health perspective, pharmacovigilance systems should incorporate routine psychiatric signal detection for all newly marketed contraceptives, and regulatory agencies ought to require post-marketing mood-outcome surveillance analogous to that mandated for thromboembolic events. Finally, future research should prioritize prospective, formulation-specific trials that stratify by developmental stage and baseline vulnerability, and report standardized, clinician-verified psychiatric endpoints, to sharpen risk-benefit calculations for both prescribers and users. Several limitations should be taken into considerations. First, despite rigorous screening, the evidence base is dominated by observational designs, leaving residual confounding (particularly by indication and prior psychiatric history) largely unresolved. Second, heterogeneity was substantial (I² > 95%), driven in part by one pharmacovigilance outlier and several low-quality cross-sectional surveys; although sensitivity analyses attenuated the pooled estimate, between-study variance remained high. Third, psychiatric outcomes were often self-reported or inferred from medication proxies, raising concerns about misclassification. Fourth, our exclusion of studies focusing on polycystic ovary syndrome, endometriosis or HIV (conditions frequently managed with HC) may limit generalizability to those populations. Finally, language restrictions to English and French and a 10-year search window may have omitted earlier or non-Western literature. These limitations highlight the need for well-controlled, diagnosis-confirmed prospective cohorts and randomized trials that compare specific HC formulations while accounting for baseline mental-health risk.

This scoping review synthesizing data from 46 studies highlights a consistent, but modest, association between HC use (particularly progestin-only formulations) and increased risk of depressive symptoms and suicidality, especially among adolescent and early adult users. While the evidence for anxiety and other psychiatric outcomes remains inconclusive due to methodological heterogeneity and inconsistent endpoint definitions, the cumulative findings emphasize the importance of adopting a biopsychosocial lens when prescribing HCs. Notably, the quality assessment revealed considerable variation in study design rigor, with confounding control and psychiatric outcome specification emerging as recurrent limitations. These results suggest that psychiatric risk stratification and post-initiation monitoring should become integral components of contraceptive counseling, especially in populations with pre-existing or familial mental health vulnerabilities. Moving forward, there is a pressing need for high-quality, prospective, formulation-specific research that integrates validated psychiatric measures, considers neuroendocrine and developmental moderators, and includes both clinical and subclinical mental health endpoints. Interdisciplinary research that bridges reproductive endocrinology, psychiatry, and epidemiology (while leveraging innovations in digital health tracking and real-world data) may be particularly well-positioned to disentangle causal pathways and optimize contraceptive safety. A more personalized, evidence-informed approach to hormonal contraception will be important in advancing both reproductive autonomy and mental health equity across diverse populations.