The approval of the first hormonal contraceptive (HC) in 1960 was a historical turning point for women, offering unprecedented autonomy over reproductive choices and greater control over the trajectories of their lives. As such, access to contraception is now recognised as an international human right (World Health Organisation, 2014).

An estimated 40% of reproductive-aged women use hormonal contraceptive (HCs) (Kristensen, 2021), but discontinuation is common due to perceived psychological side effects (Martell, 2023). Despite this, epidemiological research assessing the link between HCs and mood changes or depression has yielded mixed results. Some studies report a protective influence of HCs on mood in specific populations, e.g., those with premenopausal dysphoric disorder (PMDD) (Robakis, 2019), while other studies have reported a negative association or null findings for HCs and depressive symptoms (Skovlund, 2016; Worly, 2018). These inconsistencies may be explained by between-study variation; more research is needed to understand this relationship, accounting for nuances regarding the influence of different exogenous hormone formulations and deliveries (e.g., long-acting versus oral methods), and in different populations.

This blog entry reviews findings from a recent Danish study (Larsen et al, 2024) which estimated the risk of depression in new users of three different doses of the levonorgestrel-releasing intrauterine system (LNG-IUS). This work is necessary not only to provide informed counselling to users but also to support equitable access to contraception by addressing barriers related to concern regarding side effects.

Methods

This was a prospective study using data from the Danish national health registry. The study sample included women residing in Denmark who:

• Were born in 1978 or later,
• Started using an LNG-IUS between the ages of 15 and 44, and
• Had no previous mental health disorder.

Incident depression was determined using diagnostic records or prescription of antidepressants within 12 months of LNG-IUS insertion. Risk of depression was calculated using Cox regression, adjusted for other known and measured confounders (i.e., age, education, family history of mental health disorder, postpartum IUS use and medical disorders for which a high-dose IUS is indicated).

Results

The analysis included 149,200 new users of the LNG-IUS; 14.8%, 32% and 53.3% of whom used a low, medium and high dose device, respectively. High-dose IUS users were more likely to be older, to have had children, and to have been diagnosed with a medical indication for an LNG-IUS (e.g., heavy menstrual bleeding, dysmenorrhoea, endometriosis or uterine fibroids).

The total number of depression cases, based on both diagnoses and antidepressant prescriptions, was 2,258. After 12 months, the percentage of low, medium and high dose LNG-IUS users with depression was 1.21%, 1.46%, and 1.84%, respectively.  This translates to 0.26% more of the medium dose group having depression compared to the low dose group, and 0.63% more in the high compared to low dose group; relatively modest differences in risk.

Cases of depression were predominantly identified through antidepressant prescriptions (n=2,110), whereas 366 women had received a formal diagnosis. To overcome potential bias from including women who received antidepressants for conditions other than depression (e.g., anxiety disorders, chronic pain conditions, premenstrual dysphoric disorder [PMDD]), the analysis was rerun with those for which depression was specified as the indication (n=1,198). In this case, the absolute risks were 0.59% (low dose), 0.70% (medium dose), and 1.02% (high dose).

The risk differences between low–medium and low–high doses were small; however, a dose–response pattern (i.e., increasingly higher risk with a higher levonorgestrel dose) was evident regardless of the method used to identify depression cases.

The elevated risk of depression in the high- compared to the low-dose LNG-IUS group persisted following a series of robustness checks. For example, the authors conducted separate analyses with the following parameters: nulliparous women (i.e., with no birth history); women under the age of 30; and those who initiated an LNG-IUS after 2017 (to overcome bias from variable prescription patterns over time). Using propensity score weighting, an analysis technique that aims to mimic randomisation of baseline characteristics within the study sample, the authors again found a dose–response relationship between levonorgestrel dose and depression risk.

Another analysis using only those with a prior mental health disorder diagnosis (48,937 women) reported risks for low, medium and high dose LNG-IUS to be 4.80%, 5.22% and 5.91%, respectively. In this analysis, the difference in the proportion of high compared to low dose LNG-IUS users who developed depression (1.11%) is larger than in the group with no prior mental health diagnosis, suggesting that this group may be slightly more sensitive to the increased LNG dose.

Conclusions

The authors concluded that their study:

provides evidence of a dose-dependent association between LNG exposure and risk of subsequent depression across three dosages, which was consistent after considering potential confounders, such as menstrual bleeding indications for high-dose LNG-IUS use.

A range of additional robustness checks were conducted to overcome sources of bias, and this dose-dependent relationship persisted regardless of the analysis parameters. However, it must be acknowledged that the risk differences (0.27 to 0.66%) were modest between low and medium to high LNG doses, respectively, and that this small increased risk of depression must be considered alongside other potential beneficial effects. Clinicians should be mindful that higher LNG-IUS doses may marginally increase the risk of depression, but the choice of LNG dose should ultimately be holistically informed, accounting for comorbid conditions and patient preferences.

Strengths and limitations

Nordic countries are renowned for their excellent contributions to epidemiological research due to their population-based registries that provide routinely collected, virtually complete prospective data (Laugesen et al. 2021). This Danish study is an example of such, benefitting from a large sample size spanning the extent of the national population, and consequently, has a high level of statistical power leading to precise estimates and a vastly reduced chance of selection bias. Furthermore, such registries have a high positive predictive value for depression (individuals flagged as having depression are more likely to be accurately identified—there are fewer “false positives”).

The study also did well to address confounding by gynaecological conditions that are often treated with high-dose LNG-IUS, postpartum initiation, and history of mental health disorders. Potential sources of bias were addressed through conducting a series of sensitivity analyses, each of which maintained a dose-dependent pattern for depression risk by increasing levonorgestrel dose.

However, some limitations must be noted. As suggested by Soares (2024) in an editorial response, the exclusion of non-LNG-IUS users means that the absolute risks of depression are not given in context of the general population. The annual incidence rate of depression in Danish women is approximately 1.3% (Musliner et al. 2019); therefore, low to medium dose LNG-IUS users may not incur a meaningfully higher risk than would be expected in general. This is an important consideration when interpreting results.

It is possible that this study underestimated the true risk of incident depression, as depression diagnoses were taken from the Psychiatric Central Register but were unavailable for primary care. There may be a subset of patients with milder depressive symptoms who may have received counselling focused on lifestyle changes rather than medical intervention. Furthermore, mild unfavourable reactions to previous hormonal contraceptives were unknown; it is possible that women were switched to LNG-IUS who have an existing sensitivity to exogenous hormones.

Lastly, while differences between the three dosing groups were adjusted for in analyses, there remains a chance that other unknown or unmeasured factors may have influenced results.

Implications for practice

Contraceptive counselling is not always straightforward. Medical professionals must weigh up the risks and benefits of a range of hormonal or non-hormonal options, all while considering the unique needs of each individual in terms of their medical history, lifestyle, previous contraceptive experiences, and their personal preferences. The complex relationship between hormonal contraceptives and mental health is not well understood, with mixed findings for protective effects, increased risks or null findings for different contraceptives and depression.

Larsen et al. (2024) provide novel evidence for an increasing risk of incident depression as the dose of LNG-IUS increases. Considering the dose–response pattern and the complementary preclinical evidence for the effects of levonorgestrel on relevant mechanistic pathways in the central and peripheral nervous systems, it is plausible that levonorgestrel does modestly increase depression risk. However, given the observational nature of this study, causal relationships cannot be established.

It is also noteworthy that the absolute depression risk was only marginally higher than that expected in the general population. For high-dose LNG-IUS, approximately one in every 152 women may develop clinical depression within 12 months following initiation, compared to low dose or the general population. This could be considered a small risk overall.

At the same time, it is important for healthcare providers to be aware of this small risk. It has been thought that the LNG-IUS acts only within the intrauterine environment and that it has little influence on widespread physiological processes, for example, within the central nervous system. The dose–dependent pattern identified in this research suggests that exogenous hormones from the LNG-IUS may indeed influence depression risk, perhaps explained by influence of neuroendocrine pathways, such as potentiation of hypothalamic-pituitary-adrenal (HPA) axis reactivity (Aleknaviciute et al. 2017). Healthcare providers should validate patient’s concerns regarding anticipated effects of the LNG-IUS on mental health outcomes and consider monitoring for symptoms, especially in those with a psychiatric history or prior sensitivity to other hormonal contraceptives.

Furthermore, this study captured depression diagnoses or antidepressant prescriptions, which likely represents women with more severe depressive symptoms; yet it is possible that women may experience milder symptoms that still have an impact on their quality of life. Therefore, the relatively small risk of depression estimated by this study should not be used to dismiss patients’ concerns. Future research is warranted to explore LNG-IUS influences on a wider range of mental health symptoms and at different severities.

Statement of interests

I currently supervise a DPhil student who is studying the influence of postpartum LNG-IUS on breastfeeding rates.

Links

Primary paper

Larsen SV, Mikkelsen AP, Ozenne B, et al. Association Between Intrauterine System Hormone Dosage and Depression Risk. American Journal of Psychiatry 2024; 181(9) 834–841. https://psychiatryonline.org/doi/full/10.1176/appi.ajp.20230909.

Other references

World Health Organisation (WHO). Ensuring human rights in the provision of contraceptive information and services: Guidance and recommendations. 2014. Geneva Switzerland Department of Reproductive Health and Research.
https://iris.who.int/bitstream/handle/10665/102539/9789241506748_eng.pdf;sequence=1

Kristensen SI, Lidegaard Ø. Hormonal contraceptive use in Denmark 2010-2019. Dan Med J. 2021; 68(6).
https://content.ugeskriftet.dk/sites/default/files/scientific_article_files/2021-05/a08200599_web.pdf

Martell S, Marini C, Kondas CA et al. Psychological side effects of hormonal contraception: a disconnect between patients and providers. Contracept Reprod Med. 2023; 8(9). https://doi.org/10.1186/s40834-022-00204-w

Robakis T, Williams KE, Nutkiewicz L. et al. Hormonal Contraceptives and Mood: Review of the Literature and Implications for Future Research. Curr Psychiatry Rep 2019;21(57). https://doi.org/10.1007/s11920-019-1034-z

Skovlund CW, Mørch LS, Kessing LV, Lidegaard Ø. Association of Hormonal Contraception with Depression. JAMA Psychiatry. 2016;73(11):1154–1162. doi:10.1001/jamapsychiatry.2016.2387

Worly BL, Gur TL, Schaffir J.  The relationship between progestin hormonal contraception and depression: a systematic review. Contraception. 2018;97(6):478–489.
https://www.sciencedirect.com/science/article/pii/S0010782418300325?casa_token=wS982ncxPCUAAAAA:y-UQcrwyQHRkj81OYeLNcHr-B8581rp58G5XERqbb1kWrxwB9_1_MorHaUAxt0As_N7kO4byprA#ab0005

Laugesen K, Ludvigsson JF, Schmidt M, et al. Nordic Health Registry-Based Research: A Review of Health Care Systems and Key Registries. Clin Epidemiol. 2021;13:533–554. https://pmc.ncbi.nlm.nih.gov/articles/PMC8302231/#:~:text=The%20Nordic%20registries%20provide%20population,up%20and%20exact%20censoring%20information.

Soares CN. Contraception, Intrauterine Systems, and Depression: Can We Spot the Real Perpetrator? Am J of Psych. 2024; 181(9):789–92. https://psychiatryonline.org/doi/full/10.1176/appi.ajp.20240647.

Musliner KL, Liu X, Gasse C, et al. Incidence of medically treated depression in Denmark among individuals 15–44 years old: a comprehensive overview based on population registers. Acta Psychiatr Scand. 2019; 139:548–557. https://pubmed.ncbi.nlm.nih.gov/30908590/

Aleknaviciute J, Tulen JHM, De Rijke YB, et al; The levonorgestrel-releasing intrauterine device potentiates stress reactivity. Psychoneuroendocrinology 2017; 80:39–45.
https://europepmc.org/article/med/28315609

Photo credits

• Photo by Omar Lopez on Unsplash
• Photo by Becca Tapert on Unsplash