Research gaps in evidence-based contraception guidance: US Medical Eligibility Criteria for Contraceptive Use, 2024, and US Selected Practice Recommendations for Contraceptive Use, 2024

Abstract

The U.S. Centers for Disease Control and Prevention publishes evidence-based contraception guidelines to assist healthcare providers when counseling patients about safe and effective contraceptive use. The US Medical Eligibility Criteria for Contraceptive Use (US MEC) and the US Selected Practice Recommendations for Contraceptive Use (US SPR) have been updated on a regular basis, with the most recent full update published in 2024. This paper provides an overview of research gaps identified throughout the update process and discusses in further depth evidence gaps on several topics, including risk of thrombosis with hormonal contraception, medications for intrauterine device placement, and bleeding irregularities during implant use. Expanded study on these research gaps is needed to bolster guidelines and to improve safe and effective delivery of contraception.

1. Introduction

Evidence-based contraception guidelines are critically important for healthcare providers when counseling patients on safe and effective contraceptive use. US Medical Eligibility Criteria for Contraceptive Use (US MEC) provides guidelines for safe use of contraception among patients with certain characteristics and medical conditions, and US Selected Practice Recommendations for Contraceptive Use (US SPR) provides guidelines for common yet sometimes challenging issues in contraceptive management [1, 2]. During 2022–2023, the US Centers for Disease Control and Prevention (CDC) undertook a formal process to update these two guideline documents; this process involved identifying priority updates, conducting systematic evidence reviews, and presenting and discussing evidence at a meeting of family planning experts and stakeholders. CDC determined the final recommendations and published the updated US MEC and US SPR in August 2024 [1, 2].

Throughout the US MEC and US SPR update this process, research gaps were identified through evidence reviews, expert discussion, and feedback from health care providers and public comment. This paper lists those research gaps (Tables 1 and 2) and explores three topics in further depth: 1) risk of thrombosis with use of hormonal contraception, 2) medications for intrauterine device (IUD) placement, and 3) bleeding irregularities during implant use. These topics were selected by the authors of this paper because they highlight concepts that can be applied across many questions in contraception research, including a need to expand studies to include all relevant study populations and to better assess patients’ values and preferences related to contraception.

Table 1.

Research gaps identified in development of the US Medical Eligibility Criteria for Contraceptive Use, 2024

Condition	Research gap
Anticoagulation	• Risk for recurrent VTE with prophylactic dose (after acute phase) and hormonal contraceptive use (by type and formulation)
Chronic kidney disease	• Risk of adverse health effects associated with hormonal contraception among women with chronic kidney disease, including thrombosis, bone health and fracture, and hyperkalemia (with drospirnone-containing contraception)   ○ Risk by CKD subtypes and severity of disease
Hepatitis/cirrhosis	• Risk of thrombosis and other adverse health effects associated with hormonal contraceptive use (by type and formulation) among women with hepatocellular dysfunction   ○ Risk by type and severity of liver disease • Drug-drug interactions between antiviral treatments for hepatitis and hormonal contraception • Alternative estrogens (other than ethinyl estradiol) in contraceptive methods and effects on liver metabolism
Hepatocellular adenoma	• Among women with hepatocellular adenoma, risk of developing hepatocellular carcinoma associated with hormonal contraceptive use (by type and formulation)
Migraine with aura	• Risk of stroke associated with combined hormonal contraception among those with migraine with aura and variation by:   ○ Estrogen type and dose (e.g., < 20 mcg ethinyl estradiol)   ○ Type of migraine (with or without aura, menstrual migraine)   ○ Cyclic vs continuous use of hormonal contraception   ○ Remote history of migraine with aura
Postpartum	• Rates of IUD expulsion by IUD type • Rates of IUD expulsion by timing postpartum • Rates of IUD expulsions by mode of delivery
Sickle cell disease	• Risk of adverse events associated with hormonal contraceptive use by SCD disease types, e.g., genotypes (SC and SS) and sickle trait
Thrombogenic medical conditions (e.g. chronic kidney disease, sickle cell disease, etc.)	• Baseline risk of thrombosis among women of reproductive age with thrombogenic medical conditions (without hormonal contraceptive use)   ○ Are there disease subtypes that have differential risk of thrombosis? • Risk of thrombosis associated with use of hormonal contraception (by formulation and type) among women of reproductive age with thrombogenic medical conditions   ○ Estrogen-containing methods by estrogen type (are alternative estrogens less thrombotic than ethinyl estradiol?)   ○ Progestin-only methods by progestin type • Biologic mechanisms of thrombosis in chronic conditions such as obesity • Risk of thrombosis associated with hormonal contraceptive use among women with multiple risk factors (e.g., obesity plus other risk factors)
Thrombophilia	• Risk of thrombosis associated with hormonal contraceptive use   ○ by different thrombophilia types, including inherited and acquired (e.g., obstetric antiphospholipid syndrome)   ○ with anticoagulation

CKD, chronic kidney disease; IUD, intrauterine device; SCD, sickle cell disease

Table 2.

Research gaps identified in development of the US Selected Practice Recommendations for Contraceptive Use, 2024

Topic	Research gap
Interventions for IUD placement	• Pharmacologic interventions with positive but limited evidence on patient and provider outcomes (e.g., patient pain and satisfaction, placement success, and provider ease of placement)   ○ Lidocaine with intracervical block   ○ Intrauterine instillation of a local anesthetic   ○ Isonicotinic acid hydrazide (i.e., isoniazid)   ○ Comparative effectiveness trials of interventions
	• Patient satisfaction with intervention and IUD placement procedure
	• Patient values and preferences about pain management with IUD placement
	• Variability of results by patient characteristics, e.g.,
	○ Age   ○ Parity   ○ Mode of delivery   ○ Recent failed IUD placement
	• Effectiveness of non-pharmacologic interventions (e.g., acupuncture,
	breathing techniques)
Interventions for bleeding irregularities among LNG-IUD and implant users	• Pharmacologic interventions for bleeding irregularities with LNG-IUD • Pharmacologic interventions with positive but limited evidence on bleeding outcomes (e.g., time to cessation of bleeding and duration of bleed-free interval)   ○ SERMS/SPRMs (e.g., tamoxifen, mifepristone, ulipristal acetate)   ○ Comparative effectiveness trials of interventions
	• Patient satisfaction with intervention and resulting bleeding patterns
	• Patient values and preferences about management of bleeding irregularities, including duration of effectiveness
Self-administration of injectable contraception	• Effective strategies to improve access (e.g., patient training) • Patient satisfaction with self-administration
Hormonal contraception after the use of UPA as emergency	• Drug-drug interactions between UPA and non-COC hormonal contraception
contraception	• Drug-drug interactions between UPA and OCs after missed pills, by OC formulation, timing in menstrual cycle, and number of pills missed

COC, combined oral contraceptive; IUD, intrauterine device; LNG, levonorgestrel; OC, oral contraceptive; SERM, selective estrogen receptor modulator; SPRM, selective progesterone receptor modulator; UPA, ulipristal acetate

2. Risk of thrombosis with use of hormonal contraception

While hormonal contraception (HC) is safe for most women in the general population, use of combined hormonal contraception (CHC) is associated with a small but increased risk for thrombosis, including venous thromboembolism (VTE) and arterial thromboembolism (ATE) (i.e., myocardial infarction [MI] and stroke) [3–6]. The US MEC contains recommendations for women with thrombogenic medical conditions [1]. Women with these conditions are already at elevated risk of thrombosis, which might be further increased with use of HC. Advances in contraception development have introduced combined oral contraceptives (COCs) with alternative estrogen formulations to ethinyl estradiol (EE), i.e., estradiol valerate (E2V) and estetrol (E4) [7]. Little is known about thrombosis risk with COCs containing these alternative estrogens, particularly among women with medical conditions, and recommendations in the US MEC, 2024 are based on those for COCs containing EE. However, both of the alternative estrogens have decreased estrogenic potency compared with EE, which may result in less thrombogenic risk compared with EE-containing contraceptive methods [7]. Studies on these alternative estrogens have generally shown limited effects on hemostatic markers [7], while studies on VTE incidence have been mixed with some showing 33–54% decreased risk [8–10] and others showing similar risk when compared with EE-containing COCs [11, 12]. Further study is needed that examines thrombosis risk among women in the general population using E2V and E4. Trials of E4-containing COCs included some women with cardiovascular risk factors, despite trial exclusion criteria, and found that E4 might not substantially increase risk for developing hypertension but further research examining this and other clinical outcomes with E4 pills is needed [13].

Whether progestin-only contraception (POC) (i.e., progestin-only intrauterine devices [IUDs], implants, injectables, and progestin-only pills) increases risk for thrombosis among women with medical conditions was highlighted as a key research gap during development of the US MEC, 2010 and the updated US MEC, 2016 [14, 15]. Emerging evidence has found that use of the injectable depot medroxyprogesterone acetate (DMPA) by women in the general population (i.e., without risk factors for thrombosis) is associated with a 2 to 3-fold elevated risk for VTE compared with non-use [16–18]. An updated systematic review on risk for thrombosis with use of POC, conducted for the US MEC, 2024, identified 22 studies on women in the general population and 14 studies in women with thrombogenic conditions, including thrombophilia or history of VTE (7 studies), smoking (6 studies), hypertension (5 studies), postpartum (2 studies), lupus (2 studies), and diabetes (1 study) [19]. The studies generally found an increased risk for VTE with use of DMPA, but not other POC, among women in the general population and with certain medical conditions. However, there is a significant gap in information on these conditions as well as a lack of evidence on many medical conditions included in the US MEC. Future research can further elucidate risks of thrombosis among women with specific medical conditions for each of the POC methods.

Two medical conditions, sickle cell disease (SCD) and chronic kidney disease (CKD), are highlighted here, both of which are associated with an increased risk for thrombosis and for which the US MEC, 2024, recommendations are updated (SCD) or newly added (CKD) [1]. An increasing body of evidence demonstrates the elevated risk of thrombosis, both ATE and VTE, with SCD [20–22]. Almost 25% of SCD patients will have a clinically apparent stroke by age 45 and 11% will experience VTE by age 40 [23, 24]. For the US MEC, 2024, an updated systematic review was conducted on risk for thrombosis among women with SCD who use HC and identified only three studies [25]. These studies included small numbers of women with SCD using HC and did not provide any information on SCD severity. CKD is also associated with higher risk for thrombosis [26], with increasing VTE risk as kidney function declines and levated risk among CKD patients who experience nephrotic syndrome or are undergoing dialysis [26–32]. A systematic review was conducted for the US MEC, 2024 to examine whether there is additional risk with use of HC; no studies reported on thrombosis among women with CKD using HC [33].

For both SCD and CKD, there is a lack of information on background risk of thrombosis among women of reproductive age with these conditions. Most studies on these conditions report an aggregated thrombosis risk and do not report separately by sex or by age group. In addition, there is a lack of evidence on thrombosis risk by disease subtype or severity. Clinical course varies widely among patients with the same medical condition. Studies could examine whether risk of thrombosis varies by certain disease characteristics, e.g., laboratory parameters, organ dysfunction, comorbid conditions, or complications. Administrative, usually insurance claims-based, datasets have been increasingly used for clinical research and they can be particularly helpful when examining diseases or outcomes that are relatively rare. However, they have limited information on disease subtypes. Use of datasets that are linked to medical records could enhance information on clinical severity and could allow for better understanding of differential thrombosis risk.

Furthermore, these reviews demonstrate a continued gap in evidence on whether HC increases risk for thrombosis in women already at elevated risk due to a medical condition (Table 1). Additional studies may be helpful that include women with medical conditions, such as SCD and CKD, using HC compared with women using non-hormonal or no contraception. Future studies could seek to identify larger numbers of affected women to allow for more robust results. Studies could use patient populations ascertained from sources such as specialty clinics or disease registries.

3. Medications for IUD placement

The US SPR, 2024, contains updated recommendations for certain interventions to ease challenges in contraceptive provision and use, including provision of medications for IUD placement [2]. Concern about pain with IUD placement is a barrier to IUD initiation and has led to patients choosing other contraceptive methods [34]. The experience of pain during IUD placement is individual and might be affected by patient-specific factors such as anxiety or history of trauma [35, 36]. A broad search for randomized controlled trials (RCTs) examining pharmacologic interventions for routine IUD placement was conducted to update the evidence for the US SPR, 2024, leading to a series of intervention-specific systematic reviews and meta-analyses. The reviews examined effects of different medications (i.e., misoprostol, local anesthetics, analgesics, smooth muscle relaxants, and vaginal dinoprostone) on outcomes such as patient pain and satisfaction, placement success, and provider ease of placement [37].

For misoprostol, 14 RCTs were identified [37]. One RCT among women with a recent failed IUD placement found that vaginal misoprostol resulted in higher IUD placement success. Otherwise, in meta-analyses of trials among general populations of women (i.e., women without a recent failed IUD placement), differences were generally not found for misoprostol compared with placebo or no treatment for other outcomes (e.g., placement success, patient pain during IUD placement, patient satisfaction, provider ease of placement, and need for adjunctive placement measures such as cervical dilation). Other meta-analyses found that misoprostol was associated with increased side effects, specifically pre-placement abdominal pain or cramping and diarrhea.

For local anesthetics, 13 RCTs were found that examined topical lidocaine, 6 examined lidocaine as a paracervical block, 1 examined intracervical block, and 4 examined intrauterine instillation of a local anesthetic [37]. In some RCTs, lidocaine given by paracervical block or topically reduced patient pain; however other RCTs found no reduction in patient pain [37]. Differences were generally not found in other outcomes (e.g., placement success, patient satisfaction, need for adjunctive placement measures such as cervical dilation, or adverse events). Evidence on lidocaine as an intracervical block and intrauterine instillation of a local anesthetic was limited but generally found no differences between study groups in the outcomes examined (e.g., placement success or provider ease of placement) [37]. Lidocaine as an intracervical block and intrauterine instillation of mepivacaine might reduce patient pain, but these findings were based on only one study per intervention. There is a lack of evidence on certain outcomes, including whether: lidocaine as a paracervical block affects provider ease of placement; topical lidocaine affects the incidence of side effects; lidocaine as an intracervical block affects patient satisfaction or the incidence of side effects; and intrauterine instillation of a local anesthetic affects the incidence of side effects or adverse events.

For analgesics, 17 RCTs were found that examined six different interventions (nonsteroidal anti-inflammatory drugs [NSAIDs], NSAID plus lidocaine, NSAID plus smooth muscle relaxant, tramadol, acetaminophen, and nitrous oxide); all analgesic interventions had limited evidence (i.e., ≤2 RCTs examining the medication) except for NSAIDs which were examined by 12 RCTs [37]. Evidence on NSAIDs, either alone or in combination with lidocaine, generally suggested no positive effect on patient pain or satisfaction with the IUD placement procedure, and no effect on the other outcomes examined (e.g., placement success, provider ease of placement, need for cervical dilation, side effects, or adverse events). Some positive effects on select outcomes were observed for an NSAID plus a smooth muscle relaxant (i.e., reduced patient pain during IUD placement), tramadol (i.e., improved provider ease of placement and increased patient satisfaction), and nitrous oxide (reduced nausea and increased patient satisfaction); however, findings were based on very limited evidence [37]. There is a lack of evidence on certain outcomes, including whether: an NSAID plus lidocaine affects need for adjunctive placement measures or side effects; an NSAID plus smooth muscle relaxant affects provider ease of placement, need for adjunctive placement measures, patient satisfaction, or the incidence of side effects and adverse events; tramadol affects need for adjunctive placement measures; acetaminophen affects provider ease of placement, need for adjunctive placement measures, patient satisfaction, or the incidence of side effects and adverse events; and nitrous oxide affects need for adjunctive placement measures or adverse events.

There is limited evidence for additional interventions (i.e., topical, oral, or vaginal smooth muscle relaxants and vaginal dinoprostone); however, two trials suggested that isonicotinic acid hydrazide reduced patient pain without side effects [37].

The updated search for evidence sought to examine the effect of medications compared with control on a core set of pre-specified patient and provider outcomes (e.g., patient pain, patient satisfaction, and provider ease of placement). Depending on the medication, the number of outcomes with available evidence varied. For example, evidence was available for each outcome of interest for misoprostol and NSAIDs, but evidence for certain outcomes was lacking for other medications. Measuring and reporting on a core set of outcomes in research increases consistency across trials, reduces reporting bias, facilitates conducting meta-analysis, and improves the overall quality of the evidence base. For example, two pre-specified outcomes for the updated evidence search were medication side effects and adverse events occurring before clinic discharge. Since many RCTs reporting on these outcomes did not mention their assessment as key outcomes in the methodology section of the published report, it was unclear whether data collection for side effects and/or adverse events was systematic or non-systematic [37]. Clear documentation of pre-specified outcomes of interest assures readers that results are not reported selectively based on the findings. Some RCTs in the evidence base had incomplete data (e.g., missing numerator and/or denominator data or standard deviations around means) or reported data in a non-comparable measurement format (e.g., medians), preventing those trials from being included in meta-analysis [37]. The quality of the evidence could be improved with complete reporting of results, preferably in a format conducive to pooling findings. If continuous data (e.g., patient pain as measured using a visual analog scale) have skewed distributions, researchers might consider reporting findings as both medians and means. Many meta-analyses had considerable statistical heterogeneity, suggesting variability across studies [37]. In many situations, we were unable to conduct subgroup analyses by potential effect modifiers (e.g., prior vaginal delivery status, IUD type) due to insufficient numbers of trials examining each group in a planned subgroup analysis. Future research might consider conducting studies among subgroups with less evidence (e.g., women with a prior vaginal delivery, women using smaller IUD device types, women with a recent failed IUD placement). Most RCTs generally found that medications for IUD placement had no effect on patient pain; however the visual analog scale may not completely capture women’s pain experiences. Future research might consider using other pain measures to provide complementary information (e.g., pain acceptability). Last, some medications showed beneficial effects on at least one outcome of interest, but the evidence was too limited to make a recommendation for the medication in the US SPR, 2024. An additional RCT examining intrauterine instillation of mepivacaine was identified since presenting the evidence informing the US SPR, 2024 [38]; this trial also found beneficial effects (i.e., reduced pain during IUD placement and increased patient satisfaction). Researchers might consider future studies examining lidocaine as an intracervical block, intrauterine instillation of a local anesthetic, tramadol, nitrous oxide, and isonicotinic acid hydrazide.

4. Bleeding irregularities during implant use

The US SPR, 2024, also provides recommendations for treatments to improve bleeding irregularities while using contraceptive implants [2]. Bleeding irregularities with implants are common and a major reason for implant discontinuation [39–41]. In a systematic review of contraceptive user values and preferences, side effects and safety were the most common issues considered by users [42]. Users desired minimal side effects or adverse events (e.g., bleeding, weight gain, and acne) and if side effects were unavoidable, users desired to be able to anticipate, manage, and tolerate side effects. Similarly, a systematic review of contraceptive values and preferences among adolescents and young adults found that side effects and safety were major concerns [43]. Menstrual changes (e.g., heavy or irregular bleeding or amenorrhea) were of concern for implant users, as well as IUD and DMPA users, and could lead to method discontinuation. An updated systematic review, conducted for the US SPR, 2024, examined clinical trials (randomized and non-randomized) of several treatments for bleeding irregularities with implants and reported outcomes such as changes in bleeding (e.g., change in number of bleeding days or time to achieving amenorrhea) and patient satisfaction (e.g., patient satisfaction with intervention, patient satisfaction with bleeding patterns, and patient perception of reduction of bothersome bleeding) [44]. Trials found that certain treatments had only on-treatment effects (e.g., decreased number of bleeding/spotting days while on treatment) while others had both on- and off-treatment effects (e.g., increased duration of bleed-free interval after treatment cessation) [44]. Treatments with primarily on-treatment effects included estrogen-containing treatments, such as COCs or higher-dose oral EE alone (>50 mcg), and tranexamic acid. Bleeding typically resumed after treatment cessation. Treatments with both on- and off—treatment effects included non-steroidal anti-inflammatory drugs (NSAIDs: celecoxib, mefenamic acid, ibuprofen) and tamoxifen. A limited number of trials found that certain treatments (i.e., tamoxifen, ulipristal acetate, celecoxib, and mifepristone) led to improvement in patient satisfaction [44]. Treatments that did not show consistent on- or off-treatment effects included mifepristone alone, doxycycline alone, lower-dose oral EE (20 mcg), transdermal estradiol patch, aspirin, and vitamin E.

Similar to the evidence for medications for IUD placement, evidence for management of bleeding irregularities during implant use is sparse for each individual treatment, and even for trials investigating similar treatments, varying regimen lengths, treatment doses, and heterogeneity in outcome measurements restricted the ability to perform meta-analyses. The evidence is conflicting or of low certainty for commonly available and commonly used treatments such as ibuprofen. Several studies investigated non-readily available doses or formulations of medications including mifepristone, ulipristal acetate, and tranexamic acid. For example, although one RCT with low risk of bias showed promising on and off-treatment improvements in bleeding and increased user satisfaction with bleeding pattern with a short course of ulipristal acetate, the dosage studied is not commercially available in the United States, and no trials have replicated this finding [44].

Future studies could specifically focus on further investigating medications in the selective estrogen receptor modulator (e.g., tamoxifen) and selective progesterone receptor modulator (e.g., ulipristal acetate) categories, as these treatments appear to be effective, but data are limited. These treatments have the potential to interfere with the contraceptive effect of the implant. Limited evidence from two trials suggests that there is no increased biochemical ovulatory activity (based on progesterone or progesterone metabolite levels) for tamoxifen or ulipristal acetate when compared with placebo, but the studies were not powered to detect clinical differences in pregnancy outcomes [44]. As with most other treatment types, there is also limited evidence on the efficacy and safety of repeated treatment courses if bleeding irregularities return. Additionally, studies could investigate comparative effectiveness, safety, and acceptability of interventions such as NSAIDs compared with tamoxifen or tamoxifen compared with ulipristal acetate to further inform patient choices.

Finally, evidence on patient satisfaction with the individual interventions and their effects on bleeding is limited [44], and it is unclear whether statistically significant but small improvements in bleeding are important and meaningful to the patient. Additionally, studies did not elicit patient goals or preferences prior to initiating treatment for bleeding irregularities, which may affect satisfaction scores. Understanding these values and preferences is key to incorporation of patient-centered outcomes into research. Future studies might be helpful that examine what values and preferences are important for patients when seeking care for bleeding irregularities with implants or other contraceptive methods – and more broadly, among those seeking to initiate contraception or who face challenges while using contraception. Patient-centered contraceptive counseling prioritizes the tenets of patient values, preferences, and lived experiences in selecting, switching, or discontinuing a contraceptive method [45]; improved understanding of these tenets can support women to obtain their desired contraceptive methods [45].

5. Conclusion

This paper highlights several gaps in understanding contraceptive safety among women with certain characteristics and medical conditions, identified through the recent process to update the US MEC and US SPR. Further research might fill information gaps on baseline risk for thrombosis in women of reproductive age who have thrombogenic conditions, particularly among subgroups with different disease severity, and on whether that risk is increased with use of HC. This paper also highlights gaps in evidence on what measures may support initiation and continued contraceptive use, such as certain interventions for IUD placement or for bleeding irregularities during implant use. Future research is needed that incorporates patient-centered outcomes in contraceptive initiation and continuation. These gaps highlight a need for expanded inclusion of patients who have traditionally been underrepresented in contraceptive research, including those with chronic medical conditions and those at risk for challenges in accessing and using contraception. High quality research to address these gaps could strengthen the recommendations in the US MEC and US SPR and improve clinical practice.