The potential association between psychiatric symptoms and the use of levonorgestrel intrauterine devices (LNG-IUDs): A systematic review

Abstract

Background

Levonorgestrel (LNG)-intrauterine devices (IUDs) are an effective method of contraception; however, there is growing evidence regarding potential psychiatric side effects such as depressive symptoms, anxiety, and suicidal thoughts. Therefore, we conducted this systematic review to summarise the psychiatric effects of using LNG-IUDs.

Methods

We searched six databases (MEDLINE, Web of Science, Scopus, Science Direct, Cochrane Library, and PsycInfo), and we included all study designs. The included studies were extracted, quality assessed, and qualitatively summarised.

Results

Out of the screened studies, only 22 were finally included. While ten studies showed increased depressive symptoms, two studies showed reduced symptoms. Moreover, one study showed increased anxiety, another one reported an increased risk of suicide, four studies concluded no association with depressive symptoms, and four other studies showed uncertainty about a potential association but mentioned other psychiatric symptoms.

Conclusion

Despite unreliable data, many studies report psychiatric symptoms associated with LNG-IUDs, predominantly depression. Gynaecologists, general practitioners, and psychiatrists should therefore be aware of these potential risks, especially depressive symptoms and suicidality. Counselling patients about these risks should be mandatory. Further studies should investigate the absolute risk of mental disorders associated with LNG-IUDs and other hormonal contraceptives.

KEY MESSAGES

• Many researchers are reporting adverse psychiatric events associated with levonorgestrel intrauterine devices (LNG-IUDs).

• Despite their effectiveness, a proper psychiatric assessment should be done before inserting LNG-IUDs.

• Proper counselling regarding the depressive symptoms and suicidality should be done by the treating obstetrician.

• Further studies should investigate the absolute risk of mental disorders associated with LNG-IUDs and other hormonal contraceptives.

Keywords:

• Levonorgestrel
• intrauterine device
• depressive disorder
• mental disorders
• psychiatric symptoms

Introduction

Several intrauterine devices (IUDs) have been available on the market for over thirty years (Kaiser Family Foundation 2020) and have been proven to be a popular choice with women in developed countries (World Health Organization 2020). The Levonorgestrel (LNG)-releasing-IUD is considered one of the most effective and safe methods of reversible contraception. These devices are classified as a progestin class of agents, which unfolds their effect in particular, by the release of the LNG in the cavum uteri (Sturridge and Guillebaud 1996; Beatty and Blumenthal 2009; Bednarek and Jensen 2010; Kaneshiro and Aeby 2010). LNG-IUDs exert their mechanism of action by the stimulation of the progesterone receptors carrying out two main functions (Salmi et al. 1998): firstly, they cause thickening of the cervical mucus, thus inhibiting the occurrence of fertilisation (Peck et al. 2016); secondly, LNG has a locally suppressive endometrial effect that results in minimal or indolent menstrual blood flow (Lu and Yang 2018). Moreover, the use of low-dose oestrogen combined with LNG-IUDs may improve common perimenopausal symptoms (Santoro et al. 2015). Solicited adverse reactions include menstrual cycle irregularities and/or abnormal menstrual bleeding (Mansour 2012). Owing to the local release of the drug in the uterine cavity, LNG-IUDs are thought to exert only a few systemic effects (Attia et al. 2013).

At present, there are three LNG-IUDs available: Mirena^TM, Kyleena^TM, and Skyla^TM (Healthline 2022). Mirena^TM releases 52 mcg/day of LNG, Kyleena^TM 17.5 mcg/day, and Skyla^TM 13.5 mcg/day of LNG. Of those, Mirena^TM is the most commonly used LNG-IUD (Medscape 2022). Following the initial insertion of Mirena^TM, 20 mcg/day are released for three months (Datapharm 2022). One year following the implantation of the IUD, approximately 18 mcg/day of LNG are released whereas only 10 and 9 mcg/day are released five years and six years post-insertion, respectively (Beatty and Blumenthal 2009).

The devices are considered safe with less pronounced side effects of LNG-IUDs than those with no hormonal components (Bednarek and Jensen 2010; Reinecke et al. 2018). However, depressive mood is one of the most frequent adverse effects, where between 1–100 and 1–10 users are affected (Datapharm 2022). Recent data from published studies have emerged on psychiatric symptoms or even disorders such as depression, emotional lability, panic attacks, and suicidal ideation in women using LNG-IUDs (Nilsson et al. 1982; Concin et al. 2009; Nelson 2017; Horibe et al. 2018; Skovlund et al. 2018; Langlade et al. 2019; Zeiss et al. 2020). Subsequently, a nationwide register study examined almost half a million women for an average of 8 years and linked the use of hormonal contraception (including LNG-IUDs) to a higher risk of suicide, especially among women in their adolescence (Skovlund et al. 2018). It is worth mentioning that non-pharmacological treatments such as enhancing exercise have been shown to be effective both in reducing mortality and treating symptoms of major depression (Belvederi Murri et al. 2018).

Despite several assumptions, the underlying etiopathological mechanisms of the association between psychiatric symptoms and hormonal contraception, in particular with LNG-IUDs are not fully understood and are still subject to research. One hypothesis comprises that LNG-IUDs potentiate the systemic response to psychosocial stress and induce a centrally mediated sensitisation of the hypothalamic-pituitary-adrenal (HPA) axis responsivity with a subsequent increase in cortisol secretion (Aleknaviciute et al. 2017). In addition, a recent study outlined the effects of progesterone and oestradiol on brain functioning within regions that regulate mood, as well as the effects of these sex hormones on neurotransmission within the serotonergic and dopaminergic system, relevant in the pathophysiology of depression (Nutt 2008; Rio et al. 2018). Motivated by this controversial data and the ongoing debate on this topic, we conducted a systematic review to delineate any potential associations between the use of LNG-IUDs and psychiatric adverse reactions in the current scientific literature.

Methods

Search strategy and database searching

We searched the databases PubMed, Web of Science, Scopus, Science Direct, Cochrane Library, and PsycInfo on the 6th of October 2022 for the association between LNG-IUD and psychiatric symptoms. A broad search strategy was followed to cover most of the available literature. The search strategy for all databases was (‘Norgestrel’ OR ‘Mirena’ OR ‘Norplant’ OR ‘Plan B’ OR ‘NorLevo’ OR ‘Norplant2’ OR ‘Cerazet’ OR ‘Capronor’ OR ‘Microlut’ OR ‘Microval’ OR ‘Levonorgestrel’ OR’ LNG-IUS’ OR ‘Hormonal contraception’) AND (‘Panic attacks’ OR ‘Panic disorder’ OR ‘Depression’ OR ‘Depressive disorder’ OR ‘Suicide’) except for ScienceDirect, which was (‘LNG-IUD’ OR’ LNG-IUS’) AND (‘Panic attacks’ OR ‘Panic disorder’ OR ‘Depression’ OR ‘Depressive disorder’ OR ‘Suicide’). The search was conducted by using the title and abstract filters in all databases except for PubMed Web of Science, which were set to include all fields.

Selection criteria

Inclusion criteria: All kinds of peer-reviewed studies, including case series/reports, examining the association between LNG-IUDs and the occurrence of the above-mentioned psychiatric symptoms or disorders. Exclusion criteria: Studies that did not involve LNG-IUDs, animal studies, conference papers, and non-English studies. We followed the ‘Preferred Reporting Items for Systematic reviews and Meta-Analyses’ (PRISMA) criteria (Page et al. 2021) for study selection (Figure 1) and the PICO (‘Population, Intervention, Comparison, Outcome’) schema for systematic reviews as follows:

Figure 1. PRISMA flow diagram that depicts the search process and the number of retrieved articles for inclusion in this review.

Population: Women of childbearing age

Intervention: Using Levonorgestrel-IUD

Comparison: None

Outcome: Psychiatric disorders (depressive symptoms/anxiety/suicidal attempts).

Screening and data extraction

Three authors (KT, ME, AE) performed the screening independently. The process consisted of two steps, first a title and abstract screening, then studies were screened for eligibility and relevance by reviewing the full texts. Three researchers examined full texts and reviewed them for relevance (ME, AE, CS-L). Five researchers (KT, ME, NK, IP, AE) extracted relevant information from the included studies. References of included studies were screened for additional relevant studies not included in the initial search process. The main outcome was the association between LNG-IUD and the occurrence of depression or other psychiatric disorders (Figure 1 PRISMA flow diagram).

Quality assessment

Two authors (KT and ME) independently assessed the quality of the included studies. Any disagreement was referred to a senior study member (CS-L). Four tools were used for the quality assessment of included studies: (i) NIH tool for Quality Assessment of Observational Cohort and Cross-Sectional Studies; (ii) NIH tool for Quality Assessment Tool for before-after (pre-post) studies with no control group; (iii) Newcastle Ottawa scale for non-randomised trials (NOS) and (iv) Cochrane Risk of Bias 2 (RoB2) tool for randomised clinical trials. The NOS tool consists of three main domains: selection, comparability, and exposure. The total score is given out of 9. Studies were deemed high quality if the score was 7–9, fair quality for 4–6, and low quality for 0–3. The NIH tools are composed of 14 criteria for cross-sectional studies and 12 criteria for prospective studies with no control group. Cross-sectional and observational studies were deemed high quality if they scored 11 or more, fair quality for 7–11, and poor quality if the score was lower than 7. The quality of prospective studies with no control was considered high quality (if 9 or more), fair quality (if 6–9), and poor quality (if scores were lower than 6). The RoB 2 assesses the quality of randomised trials based on several domains: randomisation process, allocation concealment, blinding of personnel and participants, blinding of assessors, incomplete outcome data, selective reporting of outcomes, and other biases. The scoring for each domain is divided into low risk, some concerns of bias, and high risk.

Results

Study characteristics

A total of 2020 studies were primarily found from our search. Half of them (999) were excluded using EndNote 20 software. After the further exclusion of 996 studies for the reasons depicted in the PRISMA flow diagram (Figure 1), twenty-two studies were included in this systematic review. Ten studies showed a positive association between LNG-IUD and depressive symptoms while four showed no association, and two showed a negative association. The summary of the included studies is shown in Table 1, while the characteristics of the included studies are shown in Tables 2 and 3.

Table 1. Summary of the included studies.

Study	Design	Country	Population	Association of LNG with depression
a) Studies associated with increased or reported depressive symptoms:
Langlade et al. 2019	Retrospective	France	3324 Reports of adverse effects related to LNG-IUD	Psychiatric disorders including depression symptoms were reported in the population.
Horibe et al. 2018	Retrospective	Multi country	253 Reports of postpartum depression received from FAERS † database	LNG-IUD was associated with higher risk of postpartum depression
Skovlund et al. 2018	Prospective	Denmark	Half a million women followed through The National Health Register of Denmark	LNG-IUD use was associated with higher risk of suicide and suicide attempts compared to non-users of hormonal contraceptives
Daud and Ewies 2008	Cross-sectional	UK	145 Women with LNG-IUD use between 1998 and 2003	Depression was reported as a side effect of LNG-IUD use
Robinson et al. 2008	Retrospective	UK	1056 Women using LNG-IUD were surveyed	Depression was reported as a side effect in 60 (6%) of women
Elovainio et al. 2007	Cohort study	Finland	119 Women undergoing treatment for menorrhagia	Depression was stated as a reason for discontinuation of LNG-IUS use, however whether the depressive symptoms was caused by the intrauterine device is unclear
Backman et al. 2000	Retrospective	Finland	17914 Women using LNG-IUD were surveyed	Premature removal of LNG-IUS was associated with higher risk of depression
Andersson et al. 1994	Randomised trial	Sweden	1821 Women using LNG-IUD and 937 women using Nova T	Cumulative termination rate due to depression was higher in the LNG-IUD group
Cox et al. 2000	Phase-III multicenter study	UK	40 Women	3 Complaints of depression were reported to be associated with LNG-IUD use
Nilsson et al. 1982	Randomised trial	Finland	Women between 18 and 38 years were enrolled	Increase of depression symptoms was reported in the first year following LNG-IUS use
b) Studies associated with decreased depressive symptoms:
Roberts and Hansen 2017	Retrospective	US	75,528 Records of postpartum women	LNG-IUD was associated with lower risk of depression diagnosis
Keyes et al. 2013	Prospective	USA	Data of 6654 women were assessed from the National Longitudinal Study of Adolescent Health	Hormonal contraceptives were associated with lower risk of depression and suicide attempts.
c) Studies with no association of depressive symptoms:
Tazegül Pekin et al. 2014	Prospective	Turkey	120 Premenopausal women	Depression scores were assessed before insertion of LNG-IUS and 6 months after. No significant change in depression scores were seen after the follow up, indicating no effect of LNG-IUD on depression.
Enzlin et al. 2012	Cross-sectional	Belgium	353 Women using LNG-IUS and 49 women using Cu-IUD ††††	No significant difference regarding depression symptoms was found between both groups.
Toffol et al. 2012	Retrospective	Finland	Data of 8586 women were obtained from the national FINRISK Study Surveys and analysed †††††	No association of depression and LNG-IUD use was found.
Toffol et al. 2011	Retrospective	Finland	Data of 3223 women were obtained from the Finnish-population-based Health 2000 study	No association between current use of LNG-IUD and depression scores were found
d) Studies associated with increased risk of anxiety:
Slattery et al. 2018	Prospective	UK	Reports retrieved from THIN †† database of UK general practice	Positive association of LNG-IUD with depression, anxiety and sleep problems was found in users compared to non-users
e) Studies associated with increased risk of suicide:
Skovlund et al. 2016	Prospective	Denmark	Data from National Prescription Register and the Psychiatric Central Research Register in Denmark	Risk of first-use antidepressant and depression diagnosis was higher in women using the LNG-IUD
f) Studies with unclear association of depressive symptoms:
Aleknaviciute et al. 2017	Cross-sectional	Netherlands	55 Women divided into 3 groups of LNG-UD only, EE30/LNG ††† and no contraception	LNG potentiates the autonomic response of heart rate during psychosocial stress
Eisenberg et al. 2015	Partially randomised trial	USA	1600 Women under 35 years and 151 women between 35-45	Depression occurred in 91 (5.4%) of women in a 3-year follow up period
Heikinheimo et al. 2014	Prospective	Sweden	170 Women currently using a LNG-IUD presenting to replace it with a second device	Discontinuation of LNG-IUD was reported in 2 women due to depression. However whether the depression was related to the LNG-IUD is not clear
Sivin and Stern 1994	Randomised trial	USA	Parous women aged 18 to 38 years randomised to LNG-IUD or copper IUD	First occurrence of depression was described in some women with LNG-IUS however this was not statistically significant compared to copper IUD

†FAERS: Food and Drug Administration Adverse Event Reporting System

††THIN: The Health Improvement Network

†††EE30/LNG: oral ethinylestradiol/levonorgestrel (0.03mg/0.15mg).

††††Cu-IUD: Copper IUD

†††††FINRISK: Finland Risk factor survey.

Table 2. Characteristics of the included study (Part 1).

Study	Age	Controls	follow up	Country	Sample size	Setting	Precise design
Langlade et al. 2019	37 Years, IQR: 32–42	No controls	–	France	3224 Case reports	Case reports of adverse events of Mirena in France	Case reports involving Mirena in the French national pharmacovigilance database. A comparison was made before and after the media coverage of the adverse effects of LNG IUD.
Horibe et al. 2018	–	No controls	–	Japan	7,561,254 Reports	FAERS dataset was downloaded from the FDA website	The dataset was screened for postpartum depression being assessed in women taking different contraceptives
Skovlund et al. 2018	21, SD4	Never users and former users	8.3years, 4.7sd	Denmark	475802 Women, 3,920,818 person–years	The Danish Sex Hormone Register	National cohort study that includes all women living in Denmark starting from age 15 and followed up for adverse effects of contraceptive use
Slattery et al. 2018	36.9 Years LNG and 33.1 years non LNG	Non–medicated IUD users	4.2 LNG and 4.4 years non LNG	UK	10,872 LNG and 6871 non LNG users	Data from UK general practice electronic medical records	Comparison between LNG IUD and non-LNG IUDs over a follow up period of 5 years. The time to first manifestation of the disease following the insertion of IUD was analysed and compared with the control group.
Aleknaviciute et al. 2017	18–45 Years	Copper IUD	–	The Netherlands	210 Women	Medical centres	Analysis of three studies where women with LNG-IUS were compared to oral levonorgestrel and Copper IUD regarding stress reactivity
Roberts and Hansen 2017	28.5 SD 5.2	Women who were not exposed to hormonal contraception during the postpartum period	8.9m (SD 4.0)	USA	75528 Postpartum women	US military medical system	A secondary analysis of insurance records postpartum women enrolled in the US military medical system assessing the association between hormonal contraceptives and depressive symptoms
Skovlund et al. 2016	24.4 Range 15–34	Non users of hormonal contraceptives	Followed from entry till event or January 2013	Denmark	1,061,997 Women	Data from the National Prescription Register and the Psychiatric Central Research Register in Denmark	A nationwide prospective cohort study in Denmark studying the association between hormonal contraceptive and first diagnosis of depression or prescription of anti-depressants
Eisenberg et al. 2015	16–45 Years range	–	3 Years follow up	USA	1751 Women	Clinical sites in the USA, with in-person or phone follow up	Open-label, partially randomised trial, comparing LNG efficacy in different age groups
Heikinheimo et al. 2014	38.7 4.5sd	No controls	5 Years, yearly follow up	Sweden	170 Women	Multicenter study	Prospective multicenter study assessing data in second consecutive LNG IUD
Tazegül Pekin et al. 2014	36.5 18–50 Range	–	6 Months	Turkey	120 Menopausal women	Women attending clinics	Women referred for menorrhagia were recruited into the study and followed up for 6 months
Enzlin et al. 2012	17–55	IUD	None	Belgium	402	Multicenter, gynecological follow-up	Analysis of national survey data testing the associations of hormonal contraceptives and depressive symptoms
Toffol et al. 2012	25–64, Mean age 39.8 years, SD 8.6	Hormonal contraception	None	Finland	8586	National FINRISK Study Survey, a cross-sectional population-based health survey carried out in Finland every 5 years since 1972.	Analysis of national survey data testing the associations of hormonal contraceptives and depressive symptoms
Toffol et al. 2011	18–29 Years vs 30–54 years	Young vs. older groups	None	Finland	3223	National FINRISK Study Survey, a cross-sectional population-based health survey carried out in Finland every 5 years since 1972.	Analysis of national survey data testing the associations of hormonal contraceptives and depressive symptoms
Daud and Ewies 2008	43 (6.8), Mean (SD)	None	None	UK	145	Women with IUS for DUB at The Ipswich Hospital NHS Trust	Analysis of cross-sectional survey assessing the associations of hormonal contraceptives and depressive symptoms
Robinson et al. 2008	43.74 Years (range 20–81 years)	None	None	UK	1056	Women with IUS in Queens’ Medical Centre, Nottingham, England.	Analysis of cross-sectional survey assessing the associations of hormonal contraceptives and depressive symptoms
Elovainio et al. 2007	35–49	Hysterectomy	6m, 12m, 5y	Finland	236	Women treated for menorrhagia with IUS or hysterectomy	Non-blinded RCT comparing LNG-IUS to hysterectomy
Backman et al. 2000	36 (Mean)	None	None	Finland	17914	Women with IUS in Finland	Cross-sectional
Cox et al. 2000	18–45	None	12m	UK	689	Family planning clinics	Ongoing, open, single group, multicenter phase III study in out-patients.
Andersson et al. 1994	30.3 (SD 4.9)	Copper IUD	3m, 12m and 5 y	Sweden	2758	Women attending clinics	Open randomised comparative multicenter trial of two parallel groups, LNG IUS and Copper IUD users
Nilsson et al. 1982	18–38	IUD	3m, 6m, 12m	Finland	483	Women attending clinics	Open randomised comparative multicenter trial between LNG-IUS and copper IUD

LNG: Levonorgestrel; IUD: intrauterine device; IUS: intrauterine device; FAERS: FDA Adverse Event Reporting System; FDA: U.S. Food and Drug Administration; FINRISK: Finnish population survey on risk factors on chronic; noncommunicable diseases; DUB: dysfunctional uterine bleeding; MEDDRA: medical dictionary for regulatory activities; WHO: World Health Organisation; ROR: reporting odds ratio; OR: odds ratio; HR: hazard ratio; ICD: International Classification of Diseases; BDI: Beck's Depression Inventory; GHQ: General Health Questionnaire; DASS: Depression Anxiety Stress Scales; SSFS: Short Sexual Functioning Scale; QSF: Questionnaire on Sexual Functioning; VAS: Visual Analogue Scale; EQ-5D-3 L: descriptive system comprises the following five dimensions: mobility; self-care; usual activities; pain/discomfort and anxiety/depression; Cu-IUD: copper intrauterine device; TSST: Trier Social Stress Test; HMB: heavy menstrual bleeding; RCT: randomised controlled trial.

Table 3. Characteristics of the included study (Part 2).

Study	Definition of exposure	Definition and measurement of outcome	Inclusion criteria	Exclusion criteria	Summary of results
Langlade et al. 2019	Mirena IUD usage	Description of the AEs coded according to the MedDRA classification	All case reports involving Mirena® recorded in the database according to the accountability (WHO) criteria.	Non mentioned	Depressive disorders were reported more frequently following the media coverage in France
Horibe et al. 2018	LNG-IUS use	Coded according to the Medical Dictionary for Regulatory Activities	Only the drug code of primary suspected drugs was included	Duplicate reports of the same patient from different reporting sources from the analysis and extracted reports	LNG-IUS was associated with higher risk of postpartum depression (ROR: 12.5, 95%CI 8.7–18)
Skovlund et al. 2018	Comparison between never users and current users of hormonal contraceptives were made.	Diagnosis of suicide attempt or a contact with a somatic or psychiatric hospital defined as ‘suicide attempt’	Woman who turned 15 with no prior contraceptive use	Women with prior suicide attempts, antidepressant use, psychiatric diagnoses, cancer diagnoses, or venous thrombosis diagnoses	Risk of first attempt of suicide was higher in LNG-IUS users compared to never users (HR 2.86, 95%CI 2.06–3.97), however this was true for other hormonal contraceptives as well
Slattery et al. 2018	Seven codes for LNG-IUDs and 31 for non-hormonal IUDs in the THIN product dictionary	The first Read Code or prescribed treatment for panic attacks, anxiety, sleep disorders or restlessness after prescription of the study IUD	Incident prescription for LNG-IUDs and non-LNG-releasing IUDs in women with at least 1 year of follow-up after registration	Women with a history of previous IUD insertion were excluded	LNG-IUD was associated with higher risk of first-time development of depressive symptoms (HR 1.17, 95%CI 1.08–1.26), however no significant effect on persons with past history of depression (HR 1.08, 95%CI 0.97–1.21)
Aleknaviciute et al. 2017	LNG-IUS and, Trier Social Stress Test, ACTH stimulation test	Heart rate, salivary cortisol, Hair cortisol as response to stress	–	Psychiatric disorders, pregnancy or lactation, use of medications other than contraceptives, thyroid disorders, gynaecologic or pelvic diseases	Women using LNG IUD had exaggerated salivary cortisol response to TSST compared to oral LNG and normal controls, potentiated heart rate was recorded during the test
Roberts and Hansen 2017	Medicare Severity-Diagnosis Related Group codes	ICD-9 Codes were used for the definition of depressive symptoms during 12 months following delivery	–	Women diagnosed with depression or used antidepressants in 24 months prior to the study	Women with LNG-IUD showed lower risk of healthcare visit related to depression (HR 0.65, 95% CI 0.52–0.82), however no significant effect on start of antidepressant use (HR 1.01 95%CI 0.87–1.18). Women in younger ages had a higher risk of using antidepressants in the first postpartum year (HR: 1.64, 95% CI 1.28–2.09)
Skovlund et al. 2016	A first redeemed prescription of hormonal contraception	Depressive symptoms were defined as: first prescription of an anti-depressants or according to ICD-10 codes F32 to F33.9	Females aged 15–34 years using hormonal contraceptives with no prior diagnosis of depression or any major psychiatric disorders	Previous diagnosis of depression or antidepressant use, previous psychiatric disorders, previous cancer diagnosis and women with contraindications to hormonal contraceptives were excluded	Compared to non-users of hormonal contraceptives, users of LNG IUD experienced a rate ratio of 1.4 (95%CI, 1.31–1.42) for first use of antidepressants and 1.4 (1.22–1.50) for first diagnosis of depression. Higher age was associated with lower risk of first use of antidepressants
Eisenberg et al. 2015	Levonorgestrel IUS placement	Adverse events were organised into standardised terms using the MedDRA (Medical Dictionary for Regulatory Activities)	Healthy, non-pregnant, sexually active women aged 16–45 years with regular menstrual cycles were recruited	Pregnancy, pelvic inflammatory disease, malignancy, previous psychiatric disorders, and drug allergies	Depression or depressive symptoms was reported in 5.4% of LNG IUS users over 3 years
Heikinheimo et al. 2014	Second consecutive LNG IUS	Not mentioned	Women using LNG-IUS for contraception or treatment of HMB and planning to replace it with new one	Menopausal symptoms impairing the subject’s quality of life or current oestrogen therapy for menopausal symptoms were exclusion criteria	The authors reported 2 discontinuations of the LNG-IUD due to depression, however whether the IUD was the cause of depression is not clear
Tazegül Pekin et al. 2014	LNG IUD for 6 months	13-item Beck’s Depression Inventory	Pre-menopausal women aged 18–50 years	Patients who desired removal of LNG-IUS before 6 months period, pelvic inflammatory disease, malignancy, previous psychiatric disorders, and drug allergies.	No significant difference in depressive symptoms scores from baseline and after 6 months follow up
Enzlin et al. 2012	LNG-IUS	BDI-II, DAS, WHO well-being scale, SSFS, QSF	Over 18 years, in a stable heterosexual relationship for > 1 year, using Cu-IUD for at least 6 months.	–	Non-significant difference was observed between LNG-IUD and Cu-IUD users regarding depressive symptoms (Beck depression inventory score) and sexual dysfunction
Toffol et al. 2012	Hormonal contraception	BDI-21, BDI-13	–	–	A negative association between the current use of OCs and Beck Depression Inventory-13 (BDI-13) score was found
Toffol et al. 2011	Hormonal contraception	BDI-21, GHQ-12, CIDI	–	–	The current use of the LNG-IUS was inversely associated with ‘earlier waking’ (BDI) and with ‘impaired concentration’ (GHQ), while the length of LNG-IUS use was inversely associated with ‘strain’ (GHQ)
Daud and Ewies 2008	LNG-IUS	VAS, adverse events, satisfaction	LNG-IUS inserted for DUB regardless of whether it was also needed as a contraceptive	–	Some women reported discontinuation of the LNG-IUS dure to progestogenic adverse effects which included depression among other causes
Robinson et al. 2008	LNG-IUS	Adverse events, satisfaction	All women who had been fitted with a LNG-IUS in the Menstrual Disorders Clinic at the Queen’s Medical Centre	–	Depressive symptoms were reported in 60 out of 772 women, however it was not clear whether it was the cause or not
Elovainio et al. 2007	LNG-IUS	Pain relief, BDI-13,	Women between 35 and 49 years who were menstruating, who had completed their family size, eligible for a hysterectomy according to commonly used criteria	Previous malignancies, endometrial and ovarian pathologies, urinary or bowel diseases	Depressive symptoms were related to discontinuation of LNG-IUS use (OR: 3.7, 95%CI 1.55–8.82) during a 5 year follow up. The result remained significant after adjusting for other confounders
Backman et al. 2000)	LNG-IUS	Subjective report of depressive symptoms	All women in Finland who had a LNG IUS inserted below age 48	Planning pregnancy or age over 48	Premature removal of IUS was associated with 45% increase in depressive symptoms, (HR:1.45, 95%CI 1.32–1.60). The result remained significant after adjusting for other factors
Cox et al. 2000	LNG-IUS	Subjective reports of mood swings or depression	According to the doctor’s own clinical judgement	–	3 Out of 689 women reported that the cause for LNG-IUS removal was mood swings or depression
Andersson et al. 1994	Insertion and follow up of LNG-IUS	Subjective reports of depression	Between 18 and 38 years with at least one prior pregnancy	History of ectopic pregnancy, currently breast-feeding, steroid hormones contraceptives 12 months before the study	Subjective reports of depressive symptoms were more in women with LNG-IUS compared to copper IUD
Nilsson et al. 1982	Insertion and follow up of LNG-IUS	Subjective reports of depression	Between 18 and 38 years with one prior pregnancy	Pregnancy or breast feeding, history of pregnancy, pelvic inflammatory disease, no long-acting hormonal contraceptives since last pregnancy	Nervousness and depression were reported in 23.8–22.4% and 11.2–14.4% in IUS, respectively. However, it was not clear whether the LNG-IUS was the main cause for the depressive symptoms

LNG: Levonorgestrel; IUD: intrauterine device; IUS: intrauterine device; FAERS: FDA Adverse Event Reporting System; FDA: U.S. Food and Drug Administration; FINRISK: Finnish population survey on risk factors on chronic; noncommunicable diseases; DUB: dysfunctional uterine bleeding; MEDDRA: medical dictionary for regulatory activities; WHO: World Health Organisation; ROR: reporting odds ratio; OR: odds ratio; HR: hazard ratio; ICD: International Classification of Diseases; BDI: Beck's Depression Inventory; GHQ: General Health Questionnaire; DASS: Depression Anxiety Stress Scales; SSFS: Short Sexual Functioning Scale; QSF: Questionnaire on Sexual Functioning; VAS: Visual analogue scale; EQ-5D-3 L: descriptive system comprises the following five dimensions: mobility; self-care; usual activities; pain/discomfort and anxiety/depression; Cu-IUD: copper intrauterine device; TSST: Trier Social Stress Test; HMB: heavy menstrual bleeding; RCT: randomised controlled trial.

LNG-IUD and depressive symptoms

Positive association

A total of 10 studies showed increased depressive symptoms among LNG-IUD users.

In a pharmacovigilance database study comprising 3224 case reports of women with LNG-IUD, 38.8% of the reports were related to anxio-depressive disorders after the media coverage peak (Langlade et al. 2019).

A retrospective study based on data from the food and drug administration adverse event reporting system investigated the association of postpartum depression and several drugs including contraceptive devices and implants between 2004 and 2015 in spontaneously reported adverse events. 34 reports of postpartum depression were detected from a total of 75,500 cases among LNG-IUD users (Horibe et al. 2018).

A retrospective study conducted via a questionnaire by Backman et al. demonstrated a strong association between early removal of LNG-IUD and depressive symptoms (Backman et al. 2005). This study investigated 17914 women in Finland with the insertion of LNG-IUDs between April 1990 and December 1993. A multivariate analysis was conducted via Cox Hazard Model and premature removal of the LNG-IUDs was strongly associated with depression in some women.

A phase 3 multicenter study conducted in the United Kingdom by Cox et al. in a clinical setting studied the performance and effectiveness of LNG-IUD in 678 women (Cox et al. 2002). Of those, 13 women presented with depressive symptoms.

An open randomised multicenter study was conducted by Andersson et al. to examine the effectiveness of 20 mcg/24 h LNG-IUD compared to Nova T (a copper-releasing device) for the duration of its course (5 years) (Andersson et al. 1994). In this study, 1821 women were included in the LNG-IUD group and 937 women in the Nova T group. In terms of hormonal disturbances, the LNG-IUD group had a termination rate of 2.9 compared to the Nova T group (2.0, p < 0.001). A study conducted by Nilsson et al. revealed significant results for Nervousness (Nilsson et al. 1982).

Absence of an association

Four studies revealed no association between the use of LNG-IUD and depressive symptoms.

In particular, Toffol et al. performed a cross-sectional population-based health survey in Finland that concluded the absence of an association between LNG-IUD use and Beck Depression Inventory-13 (BDI-13) scores (Toffol et al. 2012). Enzlin et al. performed a multicenter study on more than 350 women using LNG-IUD and no difference in psychological symptoms was observed in comparison to users of Cu-IUD (Enzlin et al. 2012). Besides, Tazegül et al. performed a study on 120 premenopausal women that received LNG-IUDs in which depression scores did not show a significant difference from the baseline (Tazegül Pekin et al. 2014). Toffol et al. observed no association between the use of LNG-IUDs and depressive symptoms (Toffol et al. 2011).

Negative association

Two of the studies showed a negative association between LNG-IUD and depressive symptoms/depression. A retrospective study by Roberts et al. consisting of 75,528 postpartum women with depression and subsequent antidepressant use revealed that postpartum women using LNG-releasing contraceptives had a lower risk of being diagnosed with depression (Roberts and Hansen 2017). A reduction of depressive symptoms with hormonal IUD was observed in Keyes et al. population-based, longitudinal analysis study (Keyes et al. 2013).

LNG-IUD and anxiety

A cohort study conducted by Slattery et al. studied the effects of hormonal IUDs on new users of LNG-IUDs compared to non-hormonal IUDs and their association with anxiety and sleep disturbances in subjects with no medical psychiatric history (Slattery et al. 2018). This study revealed a significant association between LNG-IUD with anxiety and sleep disturbances indicating a hazard ratio of 1.18 with a 95% confidence interval of 1.08–1.29 for anxiety. Regarding sleep disturbances, they observed a hazard ratio of 1.22 with a 95% confidence interval of 1.08–1.38. The results for restlessness or panic attacks were not significant.

LNG-IUD and suicide

A nationwide prospective cohort study conducted by Skovlund et al. in Denmark studied the relative risk of suicide and suicide attempts in 475,802 women over the age of 15 using hormonal contraceptives from 1996-2013 (23). The women in this study did not have a psychiatric or medical history (Skovlund et al. 2018). Numerous confounding variables were extracted. Approximately half a million women were followed for 8.3 years with a mean age of 21 years. Results indicated 6999 initial suicide attempts and 71 suicides. Relative risk was the highest in adolescents. The relative risk of suicide attempt of hormonal contraceptive use was 2.06 (95% CI = 1.92–2.21) in women between the age of 15-19 years (Skovlund et al. 2018). In the 20-24 year-old age group, the relative risk was 1.61 (95% CI = 1.39–1.85), and 1.64 (95% CI = 1.14–2.36) for the 25–33 age group (Skovlund et al. 2018). In addition, the relative risk of initial suicide attempts increased considerably following the insertion of hormonal contraceptives for one year and decreased thereafter in comparison to the non-users (Skovlund et al. 2018).

Other related studies

Several studies have also been conducted in the past to study the effects of LNG-IUD (Heikinheimo et al. 2014; Eisenberg et al. 2015; Aleknaviciute et al. 2017). A partially randomised trial conducted by Eisenberg et al. studied the three-year efficacy and safety of 52mg LNG-IUD in 1751 women between the ages of 15 and 45; 1600 women were between the ages of 16 and 35 and 151 women were 36–45 years of age. Depressive symptoms were one of the adverse effects noted in 94 (5.4%) of the women among all participants in this study (Eisenberg et al. 2015). However, it was also here unclear in the Eisenberg et al. study or the other studies whether depressive symptoms were associated with the LNG-IUD or other confounding variables (Heikinheimo et al. 2014; Eisenberg et al. 2015; Aleknaviciute et al. 2017).

A cross-sectional study and two experimental studies were conducted by Aleknaviciute et al. that investigated each woman from three different groups: women with LNG-IUD, women taking oral ethinyl oestradiol and levonorgestrel and women with natural cycling and no hormonal contraceptives (Aleknaviciute et al. 2017). In study 1, repeated measurements (including baseline) of salivary cortisol levels were conducted in all groups following the established Trier Social Stress Test (TSST). Study 2 investigated salivary cortisol and serum total cortisol between these groups following the administration of 1 mcg ACTH. Study 3 assessed hair cortisol to draw any conclusion on the prolonged or long-term effects of cortisol. A significant increase in salivary cortisol was observed in women with LNG-IUD following the TSST compared to the other groups (Aleknaviciute et al. 2017). Heart rates, ACTH challenge, and cortisol hair levels were also significantly higher in the LNG-IUD group. In summary, these results were interpreted as a hypersensitivity of the autonomic HPA axis in presence of psychosocial stress owing to the use of LNG-IUD.

Quality assessment

A total of 22 studies were assessed for their quality of evidence. Four RCTs were assessed by RoB 2 tool, 3 prospective studies with no control group were assessed by the NIH tool, 3 case-control studies were assessed by NOS, and 12 observational and cross-sectional studies were assessed by the respective NIH tool. The included RCTs showed an overall low risk of bias with some concerns regarding the blinding of personnel in Andersson et al. (1994) and Sivin and Stern (1994) (Table 4). As for the prospective studies, Heikinheimo et al. (2014) and Elovainio et al. (2007) were deemed of high quality while Cox et al. (2002) was of fair quality (Table 5). For the case-control studies, Slattery et al. (2018) was of high quality while Aleknaviciute et al. (2017) and Enzlin et al. (2012) were deemed of fair quality (Table 6). As for the remaining studies, Skovlund et al. (2018), Skovlund et al. (2016), Tazegül Pekin et al. (2014), Toffol et al. (2012), Daud and Ewies (2008), Robinson et al. (2008), and Backman et al. (2005) were assessed to be of high quality while Langlade et al. (2019), Horibe et al. (2018), Roberts and Hansen (2017), Keyes et al. (2013), and Toffol et al. (2012) were deemed as of moderate quality (Table 7).

Table 4. Cochrane ROB2 tool for assessment of randomised trials.

Study	Risk of bias arising from the randomisation process	Allocation concealment	Blinding of participants and personnel	Blinding of outcome assessment	Incomplete outcome data	Selective reporting	Other bias
Eisenberg et al. 2015	Low risk	Low risk	Low risk	Low risk	Low risk	Low risk	Low risk
Andersson et al. 1994	Low risk	Low risk	Some concerns	Some concerns	Low risk	Low risk	Low risk
Sivin and Stern 1994	Low risk	Low risk	Low risk	Some concerns	Low risk	Low risk	Low risk
Nilsson et al. 1982	Low risk	Low risk	Low risk	Low risk	Low risk	Low risk	Low risk

Table 5: NIH prospective studies with no control group.

Study	1: Clear research question	2: Clear definition of study population	3. Representative population	4. Enrolment of all eligible subjects according to predefined criteria	5. Large enough sample size	6. Clear definition of intervention	7. Clear definition of outcome measures	8. Assessors blinded to exposure	9. 20% or less loss to follow-Up	10. Analysis of change in outcome measures (Before and after)	11. Outcome measures examined multiple times	12. Analysis co nducted at a group level?	Score
Heikinheimo et al. 2014	Y	Y	Y	N	Y	Y	Y	CD	N	Y	Y	NA	9
Elovainio et al. 2007	Y	Y	Y	Y	Y	Y	Y	N	Y	Y	Y	NA	10
Cox et al. 2000	Y	Y	Y	N	Y	Y	Y	N	N	Y	Y	NA	8

Y: yes; N: no; CD: cannot determine; NA: non-applicable.

Table 6. New castle Ottawa scale for non-randomised studies.

Study ID	Newcastle–Ottawa Scale
	Selection (maximum:★★★★)				Comparability (maximum:★★)	Exposure (maximum:★★★)			Total score (out of 9)
	Is the case definition adequate? (maximum:★)	Representativeness of the cases (maximum:★)	Selection of controls (maximum:★)	Definition of controls (maximum:★)	Comparability of cases and controls on the basis of the design or analysis	Ascertainment of exposure (maximum:★)	Same method of ascertainment for cases and controls (maximum:★)	Non-Response rate (maximum:★)
Slattery et al. 2018	★	★	★	★	★★	★	★	–	8
Aleknaviciute et al. 2017	★	–	★	★	★	★	★	–	6
Enzlin et al. 2012	★	–	★		★	★	★	–	5

Note: The stars represent points for each study.

Table 7. NIH for observational and cross-sectional studies.

Study	1: Clear research question	2: Clear definition of study population	3: 50% eligible participation rate	4: All subjects recruited from similar populations and time period	5: Sample size and power justification	6: Exposure measured before outcome	7: Adequate timeframe for association between exposure and outcome	8. Different levels of exposure were examined	9: Same exposure for all participants	10: Exposure assessed multiple times	11: Outcome measure was clear	12: Outcome assessors blinded to exposure	13: Loss to follow-up 20% or more	14: Adjustment for confounding factors	Overall score
Langlade et al. 2019	Y	Y	Y	Y	N	Y	Y	NA	Y	N	N	Y	NA	Y	10
Horibe et al. 2018	Y	Y	NA	Y	N	Y	Y	NA	Y	N	Y	Y	NA	N	8
Skovlund et al. 2018	Y	Y	Y	Y	NA	Y	Y	Y	Y	Y	Y	Y	Y	Y	12
Roberts and Hansen 2017	Y	Y	NA	Y	NA	Y	Y	N	Y	Y	Y	Y	NA	Y	9
Skovlund et al. 2016	Y	Y	NA	Y	NA	Y	Y	Y	Y	Y	Y	Y	NA	Y	11
Tazegül Pekin et al. 2014	Y	Y	NA	Y	N	Y	Y	NA	Y	Y	Y	Y	Y	Y	12
Keyes et al. 2013	Y	Y	NA	Y	N	Y	Y	Y	Y	N	Y	Y	NA	Y	10
Toffol et al. 2012	Y	Y	NA	Y	N	Y	Y	N	Y	N	Y	Y	NA	Y	9
Toffol et al. 2011	Y	Y	Y	Y	Y	Y	Y	Y	Y	N	Y	Y	NA	Y	11
Daud and Eweis 2008	Y	Y	Y	Y	N	Y	Y	Y	Y	Y	Y	Y	Y	N	11
Robinson et al. 2008	Y	Y	Y	Y	N	Y	Y	NA	Y	Y	Y	Y	NA	Y	11
Backman et al. 2000	Y	Y	Y	Y	NA	Y	Y	N	Y	Y	Y	Y	Y	N	11

Y: yes; N: no; CD: cannot determine; NA: non-applicable.

Discussion

We conducted a systematic review to study the hypothesis that LNG-IUD might be associated with a higher risk of psychiatric symptoms, such as depressive and/or anxiety symptoms.

Following the PRISMA criteria as an evidence-based minimum set for reporting systematic reviews, twenty-two studies were included in the final analyses and showed ambiguous results, but a greater number of them reported a positive association with depressive symptoms related to LNG-IUDs.

According to a pharmacovigilance database study for Mirena® reports, 38.8% of the reports after the media coverage peak were related to symptoms of anxiety and depression (Langlade et al. 2019). A questionnaire on the reproductive and contraceptive history of over 17,000 users of LNG-IUD revealed that 36% of the LNG-IUD users reported depressive symptoms. An open-label, partially randomised trial on 1600 women who received LNG-IUD observed that 5.4% reported depression or depressed mood (Eisenberg et al. 2015).

Anxiety, mood, sleep disturbances, restlessness, and suicidal attempts were also reported in the previous literature (Tazegül Pekin et al. 2014; Bitzer et al. 2018; Slattery et al. 2018). Despite these aforementioned adverse effects, LNG-releasing IUDs are broadly used due to patient satisfaction and effectiveness (Committee on Adolescent Health Care Long-Acting Reversible Contraception Working Group TACoO, Gynecologists 2012). However, previous research showed that depression and mood swings were among the leading causes of discontinuation of LNG-releasing IUDs (Backman et al. 2000; Cox et al. 2002; Elovainio et al. 2007; Daud and Ewies 2008; Eisenberg et al. 2015; Hall et al. 2015). This is in line with the results of Elovaino et al. (2007) and Heikinheimo et al. (2014). Large doses of synthetic oestrogens and progestins in combined oral contraceptive tablets (COCs) (e.g. Enovid: 5 mg norethynodrel, 75 mcg mestranol) were thought to interact with mood-related neurotransmitters and neurotransmitter metabolism by contraceptive researchers in the 1960s and 1970s (Hall et al. 2015). Female reproductive physiology and hormonal disturbances can be attributed to the higher risk of depression in women compared to men (Soares and Zitek 2008). During puberty, pregnancy, postpartum, and menopause, women are more vulnerable to experiencing fluctuations in their mood, such as depressive manifestations (Soares and Zitek 2008).

The pathophysiological mechanism of depression as an adverse event of LNG-IUD use might be linked to the hypothalamic-pituitary-adrenal axis response to LNG (Aleknaviciute et al. 2017). Furthermore, central sensitisation of the autonomic nervous system is thought to lead to raised heart rate and higher cortisol levels, even in comparison to oral contraceptives, the copper spiral, or the natural cycle (Aleknaviciute et al. 2017). We believe that the reported studies of women who complained of mood swings after using LNG-IUDs should be correlated with the evidence that the LNG-IUD potentiates increased but relatively low serum levels of levonorgestrel, which could exert systemic adverse effects. Furthermore, women with a positive history of depression might be more vulnerable to these adverse events. Therefore, LNG-IUD might be not the first choice in this risk group. It is worth mentioning that Bosco-Lévy and colleagues matched a cohort of 9318 LNG-IUS users with 10,185 copper-IUD and analysed new initiation of psychiatric drugs or a new psychiatric visit among other adverse effects. They found that the LNG-IUD cohort had a tendency to initiate anxiolytic drugs more than the copper-IUD counterparts (HR 1.08, 95%CI 1.01–1.15), probably because of having more anxiety and depressive symptoms than the copper-IUD users. Furthermore, a slight increase in device removal was observed in the LNG-IUD users (Bosco-Levy et al. 2019).

Noteworthy, the possible predictive value of affective temperamental dysregulation might be a possible contributor to negative clinical outcomes in depressed patients. Importantly, patients with higher total scores on affective dysregulated temperaments are more likely to have higher hopelessness which is an important predictor of suicidality. Importantly, 67% of patients in the dysregulated group had Beck Hopelessness Scale total scores indicating high levels of hopelessness (Serafini et al. 2012). Despite several assumptions, the underlying etiopathological mechanisms of the association between psychiatric symptoms and hormonal contraception, in particular with LNG-IUD remain unclear and are still the subject of research.

Considering the pharmacokinetic properties of LNG-IUDs, in particular, the insertion is followed by a significant increase in serum LNG levels that may contribute to the negative effects of LNG-IUDs on mental health (Hall et al. 2015). Thus, one can assume that also the replacement of LNG-IUD is accompanied by a higher risk of (mental) side effects. In line with this assumption, we recently reported a case of developing psychiatric symptoms, e.g. depressive mood, anxiety, and suicidal ideation in association with the replacement of the LNG-IUD (Healthline 2022).

In contrast, some studies showed the opposite direction in findings. A cross-sectional population-based health survey in Finland showed no association between the use of LNG-IUD and depressive symptoms assessed by Beck Depression Inventory-13 (BDI-13) scores (Toffol et al. 2012). A completed Study Survey investigated the effects of current use and the duration of hormonal contraceptives including LNG-IUD on mood and psychological symptoms and found that hormonal contraception is not associated with a negative influence on mental health (Toffol et al. 2012). Enzlin et al. did not observe a difference in psychological symptoms among LNG-IUD users versus Cu-IUD users (Enzlin et al. 2012). Also, in line with this, Tazegül et al. observed no significant difference in depressive scores among 120 premenopausal women through the first six months after LNG-IUD insertion; there were no significant differences in depression scales in women that received LNG-IUD compared to the time before insertion (Tazegül Pekin et al. 2014). Moreover, some studies reported a negative relationship between depressive symptoms and the use of LNG-IUDs. Toffol et al. (2011) reported a negative association between LNG-IUD and Beck Depression inventory total score. In addition, Keyes et al. performed a study on more than 6000 sexually active women in the United States and observed that women that used IUDs for contraception, did not have an association with depressive symptoms, however no information if the IUDs contained hormones were available; Keyes et al. also reported a reduction of depressive symptoms with hormonal IUD (Keyes et al. 2013).

Böttcher et al. performed a survey on MEDLINE/PubMed and found that depression was not a common side effect of hormone-based contraception (Bottcher et al. 2012). A previous review of the literature showed that the majority of women using hormonal contraception did not experience mood-related alterations (Smiech and Rabe-Jabłońska 2006). The rates of mood symptoms were similar, or even lower than women who did not use hormonal contraception (McCloskey et al. 2021). Worly et al. performed a systematic review of 26 articles and found no evidence of LNG-IUD-associated depression (Worly et al. 2018). This correlates with the results of Toffol et al. who performed a cross-sectional approach using data from previous surveys and also did not find an association between LNG-IUD and depression (Toffol et al. 2012). A systematic review of the literature from Ti et al. found no association between hormonal contraception and postpartum depression (Ti and Curtis 2019). In addition, an observational study of 120 premenopausal women who received LNG-IUD did not show a significant increase in the depression scores (Short-Form Health Survey) (SF-36) and Beck's Depression Inventory after six months follow-up (Tazegül Pekin et al. 2014). A secondary analysis of insurance records in more than 75,000 postpartum women showed that the use of LNG-IUS was associated with a lower risk of depression (Roberts and Hansen 2017). Here it is important to emphasise that postpartum women that have used antidepressants or were diagnosed with depression in the two years preceding delivery, were excluded from the study (Roberts and Hansen 2017). Schaffir et al. performed a MEDLINE review on combined hormonal contraception in the past 30 years and found that their effect on mood was rather infrequent (Schaffir et al. 2016). In addition, Bürger et al. concluded through a systematic review that LNG-IUD use improves the quality of life and sexual function (Burger et al. 2021). Moreover, Pagano et al. found that women with pre-existing depression did not suffer from the worsening of clinical course of depression following the use of LNG-IUD (Pagano et al. 2016). Furthermore, Enzlin et al. performed a cross-sectional study and observed that depressive symptoms did not differ between women using LNG-IUD versus copper IUD (Enzlin et al. 2012).

Gopimohan et al. analysed a cohort of women aged 30–50 years who were referred for abnormal uterine bleeding. The women had an insertion of LNG-IUD and were followed up for 6 months. The main outcomes were menstrual blood loss and quality of life (assessed using the EQ-5D-3 L questionnaire). The questionnaire measures several quality-of-life aspects including depression. A score improvement was found pre and post-intervention, however, it was not clear whether the domain of mental health is the one responsible for the statistically significant improvement (Gopimohan et al. 2015).

In addition, Hurskainen et al. conducted a randomised trial in which 236 women were enrolled for treatment of menorrhagia and followed up for 5 years. One-hundred-and-nineteen had insertion of LNG-IUD while the rest received a hysterectomy. Depressive and anxiety symptoms were improved in both groups after 5 years of follow-up, indicating that the main cause of these symptoms was the disease (whether the indication of hysterectomy or the cause of menorrhagia). Furthermore, no significant differences in depression or anxiety symptoms were found between both cohorts pre or post-intervention (Hurskainen et al. 2004).

Hence, counselling women who have been recognised as being at higher risk of developing psychiatric symptoms is essential. Finally, it should be noted that major depressive disorder is a multifactorial disease, where the hormonal imbalance imposed by the LNG-IUD is only part of the aetiology. This means that the association depicted in the studies included in this review does not mean direct causation of illness in all patients. This can be explained by the wide heterogeneity in the included studies.

Limitations

This work attempted to describe and discuss the current literature on LNG-IUD and its association with psychiatric disorders. Some limitations of our analysis are: The study is solely based on published articles. In concern to progestin contraceptives, previous literature did not show evidence of increased depression through the use of progestin contraceptives (Attia et al. 2013). Larger blood samples should be investigated during the luteal phase. Furthermore, affective symptoms were not assessed in some studies (Aleknaviciute et al. 2017). Potential confounding factors, such as major life events, sexual relationships, and previous mental health problems, are all possible issues that could influence the results of the studies of the LNG-IUD in relation to psychiatric symptoms. Besides, the included studies had various age groups, in which the relative risk of depression might vary between different age groups. In addition, the time since the placement of the LNG-IUD probably varied between the studies. Furthermore, the focus of this study was only on psychological adverse events. Other adverse events related to LNG-IUDs, for example, breast cancer (Conz et al. 2020), were not assessed.

In the meantime, further research is required to further assess the potential association between mood disorders and hormonal contraception. Differentiation of LNG-IUD from oral contraception use and its potential adverse effects should be thoroughly examined in future studies to clearly understand the respective psychiatric complications each may ensue.

Conclusion

Despite controversial information in the literature, an association between LNG-IUD use and depressive/anxiety symptoms or disorders seems to exist considering the selected studies, even if not all studies report such a connection. The causative relationship between symptoms and LNG-IUD is not fully understood yet but is probably due to the sensitisation of central structures that regulate the stress axis of the brain. This review might be interpreted with care and could not be generalised, keeping in mind that the number of reports of depression or anxiety is still considered very low, explaining the current controversial debate regarding LNG-IUD, which is known to be one of the most effective, safe, widely used contraceptive methods worldwide (Bednarek and Jensen 2010). The findings of this review have some main implications, such as the importance of better communication between physicians and patients about potential adverse events, including depression or mood disorders, which might help reduce the fear of women from using levonorgestrel-releasing IUDs. Moreover, we believe it is mandatory to accurately pay attention to the psychiatric history before insertion of LNG-IUD and to follow up on potential psychiatric adverse events after counselling the patient. Insights into the timing and response to removal measures for symptom management other than removal, such as treatment of underlying mood disorders (both pharmaceutical and psychological techniques), and the potential benefits of early follow-up visits are required. Nonetheless, patients should be counselled about the risks and benefits of the potential LNG-IUD options apart from pregnancy prevention. This literature review can further aid healthcare providers in sufficiently informing their patients of the known adverse effects to make sound decisions and prevent unforeseen circumstances. Finally, further research that employs randomised controlled trials is needed to allow a more definitive assessment of the effects of contraception on women's mental health.

Author contributions

ME developed the idea. ME and HG performed the database search. KT, ME, NK, IP, and CS-L screened the articles for inclusion, extracted the data, and performed the quality assessment. KT performed the data analysis. All authors wrote the manuscript and approved the final version of the manuscript before submission for publication. The study process was monitored by ME.

Acknowledgments

No one other than the listed authors contributed to this work.

Statement of interest

All authors certify that they have no affiliations with or involvement in any organisation or entity with any financial interest (such as honoraria; educational grants; participation in speakers’ bureaus; membership, employment, consultancies, stock ownership, or other equity interest; and expert testimony or patent-licensing arrangements), or non-financial interest (such as personal or professional relationships, affiliations, knowledge, or beliefs) in the subject matter or materials discussed in this manuscript. None to declare.

Data availability statement

All the data used for the conduction of this study are readily available from the articles included in the analysis. However, the data collected can be obtained by contacting the first author (Dr. Mohmed Elsayed at [email protected]).

Additional information

Funding

No funding was received for the conduction/publication of this study.