Comparison of Dermatologic and Neuropsychological Side Effects of Oral Contraceptives, 2016–2020

Introduction

Oral contraceptives have been shown to be safe and efficacious for more than half a century, with the first oral contraceptive being approved by the Food and Drug Administration (FDA) in 1960. As with many medications, side effects noted came through the study of medication use in the general population. Despite long term use and myriad options for oral contraceptives, there is still much unknown about the side effect profile, particularly when comparing formulations.¹ Each medication in this class has nuances which may prompt a prescriber to favor a particular formulation. However, there are limited data on head-to-head comparisons of side effects between oral contraceptive medications, particularly in dermatologic and neuropsychological manifestations, which are more subjective and less likely to be accurately captured in medical records.²

Research indicates that 4–10% of COC users report mood-related issues such as increased anxiety, irritability, or depressed mood.³ In some US-based analyses, a subset of oral contraceptive users met criteria for major depression: for example, a study using NHANES data (2005–2012) found a 4.6% prevalence of major depressive symptoms among current OCP users (age 18–55) compared with higher rates in former or never users.⁴

These findings coexist with conflicting evidence. Some epidemiological work suggests that OC users may have lower levels of depressive symptoms compared with non-users, potentially reflecting selection bias, prior mental health status, or other confounding factors.⁵ Other studies highlight the difficulty in establishing causality due to methodological limitations, including small sample sizes, non-randomized designs, or focus on single contraceptive formulations.^3,6,7 This contradictory landscape—where some users experience worsening of acne or mood, while others report improvements or no change—underscores the complexity of hormonal contraception’s effects and the inadequacy of current data to draw definitive conclusions.

Given these ambiguities and methodological gaps, the present study seeks to examine multiple COC formulations in a large and diverse sample, focusing specifically on both dermatological and neurological/psychiatric outcomes. By doing so, we aim to provide a more comprehensive assessment of the prevalence, severity, and real-world impact of side effects such as acne, migraines, anxiety, and depression.

Materials and Methods

We sourced data from the FDA’s Adverse Event Reporting System (FAERS) to determine the incidence of side effects for each medication in the United States.⁸ The Medical Expenditure Panel Survey (MEPS) was used to determine the total number of patients prescribed each medication.⁹ Both databases are upkept by the United States government and data was analyzed from January 2016 to December 2020. Incidence rates of each side effect for each medication were calculated and comparisons made by utilizing binary logistic regression to determine odds ratios.

We selected ethinyl estradiol (EE)/norethindrone (NET) as our reference medication, as the most long-standing oral contraceptive, containing a first-generation progestin.¹⁰ We queried both databases for medication brand and generic names. Medications investigated include EE/levonorgestrel (LNG), EE/desogestrel (DSG), EE/norgestimate (NRG), EE/drospirenone (DRSP), norethindrone-only pill (NET) and the reference EE/NET. All patients receiving prescriptions for oral contraceptives were included, regardless of indication. Table S1 shows the progestin generation, brand names for each formulation included, along with total prescriptions during the study period. Dermatologic and neuropsychological outcomes included acne, migraine, anxiety, and depression. Side effects were reported by either healthcare providers or the patient. We determined side effect rate using reports over total prescriptions, subsequently calculating odds ratios. Significance was determined at p<0.05. We used Minitab (v19) and R (version 4.0.3; https://www.r-project.org/) statistical software for analysis.

Results

Table 1 shows the rate of dermatologic and neuropsychological side effects for each medication formulation per 100,000 users. Statistically significant higher relative rates of acne were seen with EE/LNG (OR 1.6), EE/NRG (OR 2.3), and EE/DRSP (OR 2.8). Higher rates of migraine were seen with EE/DRSP (OR 2.7). Anxiety was noted at higher rates with EE/LNG (OR 1.7) and EE/DSG (OR 2.0), and lower rates with NET alone (OR 0.3). Depression was noted at higher rates with EE/DRSP (OR 3.2).

Table 1 Rate of Side Effects per Medication Formulation with Comparison to Reference Group Ethinyl Estradiol (EE)/Norethindrone Acetate (NETA)

Discussion

Acne

Earlier generations of progestins such as the first and second generations tend to be more androgenic, whereas third- and fourth-generation progestins tend to be less androgenic.^10,11 As a result, less acne would be expected with third- and fourth-generation progestins. Despite this, several previous studies demonstrate mixed results.¹² For example, Samanta and Maiti (2022) reported that COCs can either improve or exacerbate acne depending on the formulation, highlighting the contradictory evidence.¹³ As expected, desogestrel, a third-generation progestin demonstrated no difference in acne. However, other third- and fourth-generation progestins, drospirenone and norgestimate showed higher reported rates of acne. Drospirenone has been regarded as the least androgenic progestin and even antiandrogenic due to its structural similarity to spironolactone.^10,11,14 The increased rate of acne may be due to increased prescription of drospirenone containing pills for patients with a history of acne. Of note, norethindrone alone showed no difference, signifying that the estrogen product may not have much of an effect on acne, however given the structure of database reporting this may be confounded by selection bias.

Migraine

To date, the characteristics of oral contraceptives which may cause migraines are not fully understood. It is believed that withdrawal of estrogen may contribute to headaches, but this mechanism has not been fully elucidated.¹⁵ In our study, we noted that drospirenone demonstrated increased rates of migraine compared to the reference. This observation can be compared with prior studies worldwide; for example, Ciarcia & Huckins (2024) also report higher migraine rates with drospirenone-containing pills, suggesting that specific progestin components may contribute to neurological side effects. Another proposed mechanism of why this might be is due to drospirenone’s mineralocorticoid antagonistic activity that may promote cerebral vasodilation and trigger migraines as a result.¹⁶ All other medications showed no difference. NET alone did not show any difference signifying that the estrogen component may not play as big of a role as was previously thought.

Anxiety

Currently, there is no consensus on progestin’s contribution to anxiety symptoms in patients on oral contraceptives.¹⁷ We noticed less anxiety reported with the NET only pill formulation. This may be indicating that the estrogen component may increase anxiety in patients on combined oral contraceptives. Levonorgestrel and desogestrel also showed increased anxiety; the reasons remain unclear, but similar findings have been reported by Jaafar et al (2024), who noted variable anxiety effects depending on formulation and patient baseline characteristics.¹⁸ Further, there was no difference with the drospirenone, which is the only medication approved for premenstrual dysphoric disorder (PMDD).

Depression

Much like anxiety, little is known about what characteristics of oral contraceptives may contribute to depressive symptoms. The only medication FDA approved for PMDD, which contains drospirenone, was associated with higher depression rates in our study.¹⁵ This finding corresponds with prior reports indicating that drospirenone-containing COCs may be associated with depressive symptoms, though confounding by indication and prescription bias with an increase in prescriptions to patients with baseline depression is possible.^16,19

Limitations

Strengths of this study include its investigation of specific side effects on a population-level. Limitations of the study include reliance on voluntary post-marketing reporting which introduces bias. Although millions of patients are prescribed these medications each year, only a few hundred instances for each of the studied side effects were identified. Also, dose stratification was not completed which may play a role in the rate and severity of each side effect. Likewise, the structure of the database did not allow us to control for confounders such as certain medications being chosen more frequently in a patient with a certain comorbidity along with the inability to control for additional demographic confounders including age and BMI. Newer formulations may be more likely to be prescribed for patients who have had previous side effects from other formulations. Additional studies will be needed to further characterize the post-marketing trends we have identified.

Conclusion

This study supported previous studies on rates of side effects by different oral contraceptive formulations. Selecting medications for a patient when there is a multitude of options often leads to the provider choosing a medication they are most familiar with prescribing. However, better understanding oral contraceptive products, their components, and potential side effect profiles allow for improved personalization of a medication regimen and may help to limit exacerbations of patients’ comorbid conditions.