The levonorgestrel-releasing intrauterine device potentiates stress reactivity

Highlights

• LNG-IUD is associated with a robust potentiation of HPA axis responsivity.

• LNG-IUD potentiates the autonomic heart rate response during psychosocial stress.

• The mechanism of LNG-IUD sensitization of the HPA axis is centrally-mediated.

Abstract

Background

The levonorgestrel-releasing intrauterine device (LNG-IUD) is currently recommended as a first-line contraceptive with an exclusively local intrauterine influence. However, recent clinical trials have identified side effects of LNG-IUD that appear to be systemically mediated, including depressed mood and emotional lability.

Methods

We performed two experimental studies and a cross-sectional study. For each study, women were included from three groups: LNG-IUD (0.02 mg/24 h), oral ethinylestradiol/levonorgestrel (0.03 mg/0.15 mg; EE30/LNG) and natural cycling (NC). Study 1–Salivary cortisol was measured at baseline and at defined intervals following the Trier Social Stress Test (TSST). Heart rate was monitored continuously throughout the TSST. Study 2–Salivary cortisol and serum total cortisol were evaluated relative to low-dose (1 μg) adrenocorticotropic hormone (ACTH) administration. Study 3–Hair cortisol was measured as a naturalistic index of long-term cortisol exposure.

Results

Women using LNG-IUD had an exaggerated salivary cortisol response to the TSST (24.95 ± 13.45 nmol/L, 95% CI 17.49–32.40), compared to EE30/LNG (3.27 ± 2.83 nmol/L, 95% CI 1.71–4.84) and NC (10.85 ± 11.03 nmol/L, 95% CI 6.30–15.40) (P < 0.0001). Heart rate was significantly potentiated during the TSST in women using LNG-IUD (P = 0.047). In response to ACTH challenge, women using LNG-IUD and EE30/LNG had a blunted salivary cortisol response, compared to NC (P < 0.0001). Women using LNG-IUD had significantly elevated levels of hair cortisol compared to EE30/LNG or NC (P < 0.0001).

Conclusions

Our findings suggest that LNG-IUD contraception induces a centrally-mediated sensitization of both autonomic and hypothalamic-pituitary-adrenal (HPA) axis responsivity. LNG-IUD sensitization of HPA axis responsivity was observed acutely under standardized laboratory conditions, as well as chronically under naturalistic conditions.

Introduction

Since the launch of the first hormonal contraceptive in 1960, providing women with convenient and effective protection from pregnancy, continuous progress has been made towards both reducing side effects and improving compliance while maintaining high efficacy (Mansour et al., 2010). Long-acting reversible forms of contraceptives, including the levonorgestrel-releasing intrauterine device (LNG-IUD), are currently among the most widely used forms of birth control in North America and Europe (Buhling et al., 2014, Joshi et al., 2015, Shoupe, 2016, United Nation, Department of Economic and Social Affairs, 2015). The National Institute for Health and Care Excellence and the American College of Obstetricians and Gynecologists have both recently endorsed the use of the LNG-IUD as a first-line contraceptive option (American College of Obstetricians and Gynecologists (ACOG), 2015; ACOG, 2012; NICE, 2013; Winner et al., 2012). This recommendation has also been advocated by the American Academy of Pediatrics, encouraging pediatricians to recommend the LNG-IUD to sexually active adolescents for prevention of unintended pregnancies (American Academy of Pediatrics, 2014).

In addition to providing long–acting protection (5 years per device), the LNG-IUD is a highly effective but rapidly reversible contraceptive method (Kailasam and Cahill, 2008). Furthermore, the LNG-IUD can be used by women of any age or parity, requires minimal to no maintenance, has extensive evidence supporting its safety, and has an added value for a range of gynaecological conditions (Fraser, 2013, ACOG, 2012, Prager and Darney, 2007). Moreover, and central to its popularity, the LNG-IUD is widely claimed to have no systemic physiological effects (Attia et al., 2013, Beatty and Blumenthal, 2009, Lähteenmäki et al., 2000).

The most widely held model for the hormonal mechanism of action of LNG-IUD is through local intrauterine progestin release that results in extensive decidualization of the endometrium, an environment unsuitable for fertilization and implantation (Kailasam and Cahill, 2008). The LNG-IUD has been shown to have little influence on ovarian activity, leading to widespread consensus that the release of LNG into the systemic circulation is below the physiologically-active level with a consequently negligible risk of adverse systemic effects (Attia et al., 2013, Jensen, 2005, Lähteenmäki et al., 2000, Prager and Darney, 2007, Xiao et al., 1990). However, several studies have recently cast doubt upon the claim that LNG-IUD functions with an exclusively local intrauterine influence (Halmesmäki et al., 2007, Halmesmäki et al., 2004, Lethaby et al., 2005), due to side effects including depressed mood and emotional lability (Elovainio et al., 2007, Lethaby et al., 2005). A recent Danish population-based epidemiologic study established an association between progestin-containing hormonal contraceptives, including the LNG-IUD and other progestin-only contraception, with both a significantly elevated risk of diagnosis for depression and a higher rate of antidepressant use (Skovlund et al., 2016).

Although these findings might suggest a systemic influence of intrauterine LNG-IUD progestin release, direct physiological evidence has never been established. Extensive studies of baseline endocrine measurements have been performed without any significant alterations identified (Nilsson et al., 1980, Mishell et al., 1977). However, baseline serum measurements might be insufficient to evaluate alterations in stress reactivity. Therefore, the aim of the current studies was to directly investigate whether the LNG-IUD influences the physiological responses to stress by examining autonomic and hypothalamic-pituitary-adrenal (HPA) axis responsivity in women using the LNG-IUD, oral combination estrogen-progestin contraception, or naturally cycling. In particular, women using oral combined estrogen-progestin contraception were included on the basis of their clinical relevance given that oral combination contraceptives are the most widely used form of hormonal contraception in most countries, as well as the potential mechanistic relevance with regard to how the influence of synthetic progestin on HPA axis stress responsivity might be moderated by co-administration of synthetic estrogen. Moreover, to evaluate the possible influence of the IUD itself − independent of levonorgestrel − on cortisol responsivity during the TSST, we also recruited women using a copper IUD.

In Study 1, we applied the TSST to induce moderate psychosocial stress in a laboratory setting. In Study 2, we performed the low-dose (1 μg) Synacthen test to distinguish between central and peripheral mechanisms of HPA axis functioning. In Study 3, we examined long-term cortisol exposure under naturalistic conditions using hair cortisol measurements.