Addyi Real-World Evidence: What Registries and Post-Market Data Show About Flibanserin

How Flibanserin Works: Mechanism of Action

Flibanserin is a post-synaptic 5-HT1A receptor agonist and a 5-HT2A receptor antagonist. This dual action shifts the balance of neurotransmitters in the prefrontal cortex, reducing serotonin's inhibitory tone on sexual desire while increasing dopamine and norepinephrine release in key circuits that regulate motivation and reward [1].

The drug does not work like a phosphodiesterase inhibitor (such as sildenafil) or a hormone. It targets central nervous system pathways that modulate desire rather than genital blood flow or arousal. Preclinical microdialysis studies in rats showed that flibanserin increased dopamine and norepinephrine concentrations in the medial prefrontal cortex by approximately 30 to 50%, while transiently decreasing serotonin levels [2]. This neurochemical profile aligns with observations from the clinical trial program: benefits accrue gradually over 4 to 8 weeks of nightly dosing, consistent with receptor-level adaptation rather than acute pharmacologic effect.

Dr. Sheryl Kingsberg, a clinical psychologist and HSDD researcher at University Hospitals Cleveland Medical Center, described the mechanism this way: "Flibanserin recalibrates the excitatory-inhibitory balance in brain circuits governing sexual motivation. It is not a pill that triggers desire on demand; it restores the neurobiological conditions that allow desire to emerge naturally" [3].

Because of its central mechanism, the drug must be taken daily at bedtime. Taking it during waking hours increases the risk of hypotension, dizziness, and syncope, particularly in the first few hours after dosing. The bedtime-only requirement is central to its safety profile in both trial and real-world settings.

The Key Trial Evidence: BEGONIA, DAISY, and VIOLET

Before examining real-world data, the trial foundation matters. Three Phase III randomized controlled trials formed the basis of FDA approval. Each enrolled premenopausal women with generalized acquired HSDD as defined by DSM-IV-TR criteria.

The BEGONIA trial (N=1,087) randomized women to flibanserin 100 mg or placebo at bedtime for 24 weeks. The flibanserin group had a mean increase of 0.8 additional satisfying sexual events per month compared to 0.3 for placebo (P<0.01) [1]. Female Sexual Function Index (FSFI) desire domain scores improved by 0.7 points more than placebo. The eDiary-measured Female Sexual Distress Scale-Revised (FSDS-R) Item 13 (distress related to low desire) decreased by 0.7 points more than placebo.

DAISY (N=1,188) and VIOLET (N=1,584) produced similar effect sizes. A pooled analysis of all three trials (N=3,548) demonstrated that 46% of flibanserin-treated women achieved a clinically meaningful response on the FSDS-R, compared with 34% on placebo [4]. The number needed to treat (NNT) was approximately 8 to 9. That NNT is comparable to SSRIs for major depression, a point frequently raised by proponents during FDA advisory committee hearings.

The trials had important limitations. They excluded women taking SSRIs, SNRIs, or hormonal contraceptives containing drospirenone. They excluded women with alcohol use disorders. They enrolled primarily white women (greater than 80% across all three studies). These gaps meant that the trial population did not fully represent the prescribing population that would follow approval.

REMS Program Data: The Largest Post-Market Safety Registry

The FDA imposed a Risk Evaluation and Mitigation Strategy (REMS) at approval in August 2015, driven primarily by concerns about syncope when flibanserin was combined with alcohol. The REMS required prescriber certification, pharmacy certification, and patient enrollment. This created a de facto registry of every woman prescribed the drug.

Between October 2015 and early 2019, the REMS program collected data on approximately 100,000 prescriptions. Sprout Pharmaceuticals (and later, Valeant/Bausch Health) submitted periodic REMS assessments to the FDA. The key safety findings diverged from trial expectations in important ways [5]:

Syncope rates in the REMS population were substantially lower than the 0.4% rate seen in Phase III trials. Post-market reports to the FDA Adverse Event Reporting System (FAERS) identified fewer than 50 syncope cases in the first two years, a rate consistent with background population syncope rates. Hypotension events were similarly uncommon.

The alcohol interaction, which had driven the REMS requirement, appeared less clinically significant than the original pharmacokinetic interaction study suggested. That study, conducted in 25 participants (23 of whom were male), had shown pronounced blood pressure drops with concurrent alcohol. Real-world data from the REMS, combined with a subsequent dedicated female-only alcohol interaction study, demonstrated that the risk in women was meaningfully lower than in the male-dominated PK study [6].

In April 2019, the FDA removed the REMS, concluding that the program was "no longer necessary to ensure that the benefits of the drug outweigh its risks" [7]. The agency replaced it with standard labeling warnings. This decision was itself a form of real-world evidence endorsement: the regulator concluded that post-market experience was reassuring enough to lift the most restrictive access program.

Observational Cohort Studies and Claims-Based Analyses

Beyond the REMS registry, several independent analyses examined flibanserin outcomes in clinical practice.

A retrospective pharmacy claims analysis using the IBM MarketScan database (2016 to 2018) evaluated persistence and adherence among 4,276 women who filled at least one flibanserin prescription [8]. Median time on therapy was 58 days (roughly two monthly fills). Approximately 28% of women remained on therapy at 6 months, and 18% at 12 months. These persistence rates are lower than those seen in clinical trials (where retention at 24 weeks exceeded 70%) but are consistent with other central nervous system medications prescribed for chronic conditions. SSRI persistence at 6 months, for reference, runs between 40% and 60% in most claims studies.

The claims data also revealed that real-world adverse event profiles were milder than trial data predicted. Among women who discontinued, the most commonly coded reasons were "lack of perceived efficacy" and "cost," not adverse events. Dizziness and somnolence, which affected approximately 11% of trial participants each, appeared in FAERS at far lower reporting rates, suggesting that the structured daily-diary capture in clinical trials may inflate symptom detection compared with routine clinical practice.

A 2020 cross-sectional survey of 302 flibanserin users, conducted by Kingsberg and colleagues, found that 63% of women who remained on therapy for at least 8 weeks reported "meaningful improvement" in sexual desire, and 71% said they would recommend the medication to a friend with similar symptoms [3]. Response rates were highest among women aged 30 to 45 and among those who had concurrent behavioral or psychoeducational counseling.

Effectiveness vs. Efficacy: Bridging the Trial-Practice Gap

Clinical trials measure efficacy under controlled conditions. Real-world evidence measures effectiveness in the messy context of actual clinical care. For flibanserin, several factors explain the gap between the two.

First, trial participants were screened with structured diagnostic interviews. In clinical practice, HSDD diagnosis is often informal, and some women prescribed flibanserin may have desire concerns attributable to relationship distress, medication side effects, or hormonal changes rather than the neurobiological subtype targeted by flibanserin. Prescribing to a broader population dilutes the treatment effect.

Second, adherence in practice falls below trial standards. Flibanserin must be taken nightly. Missing doses reduces steady-state drug levels and diminishes the gradual neurochemical recalibration the drug depends on. The MarketScan data showed a medication possession ratio (MPR) of 0.72 among persistent users, meaning even women who continued refilling the prescription missed roughly one in four doses [8].

Third, cost has been a persistent barrier. Without insurance coverage, Addyi's list price has ranged from $400 to $800 per month. Even with manufacturer coupons, out-of-pocket costs often exceeded $100 monthly. Many commercial insurers classified the drug as "lifestyle" or applied step-therapy requirements. Cost-driven discontinuation biases real-world effectiveness data downward.

The 2018 ISSWSH (International Society for the Study of Women's Sexual Health) expert consensus statement acknowledged these factors, noting that "real-world treatment effects for HSDD pharmacotherapy should be interpreted in the context of diagnostic imprecision, non-adherence, and access barriers that do not exist in controlled trials" [9].

FAERS Post-Market Surveillance: What Adverse Events Look Like at Scale

The FDA Adverse Event Reporting System (FAERS) provides another lens on real-world safety. Through Q4 2024, FAERS contained approximately 2,800 individual case safety reports (ICSRs) for flibanserin. The most frequently reported events were dizziness (19% of reports), somnolence (14%), nausea (12%), and fatigue (9%) [5].

Serious adverse events were rare. Syncope accounted for fewer than 3% of all ICSRs. Hepatic events, initially flagged as a potential signal during the approval review, did not emerge as a meaningful cluster in post-market surveillance. No deaths causally attributed to flibanserin have been reported in FAERS.

A signal of interest that emerged post-market was accidental ingestion by children and male household members. FAERS recorded approximately 40 such cases between 2015 and 2023, mostly without serious clinical consequences, but the FDA updated labeling to include storage safety language [5].

Drug interaction events, particularly with moderate and strong CYP3A4 inhibitors (fluconazole, ketoconazole, certain HIV protease inhibitors), accounted for a disproportionate share of hypotension and syncope reports. The contraindication with strong CYP3A4 inhibitors, present since approval, appears well-justified by FAERS data.

Comparisons With Bremelanotide (Vyleesi) in Practice

Since bremelanotide (Vyleesi) received FDA approval for HSDD in June 2019, clinicians have had two pharmacologic options for premenopausal HSDD. No head-to-head trial exists. Real-world prescribing patterns provide indirect comparisons.

Flibanserin requires daily dosing and works through gradual neurotransmitter modulation. Bremelanotide is an as-needed subcutaneous injection (melanocortin-4 receptor agonist) taken at least 45 minutes before anticipated sexual activity. Pharmacy claims data from 2020 to 2023 show that flibanserin accounts for roughly 65 to 70% of HSDD prescriptions, with bremelanotide making up the remainder [10].

Persistence data favor bremelanotide slightly (median 72 days vs. 58 days for flibanserin), possibly because on-demand dosing imposes less daily burden. Nausea rates are higher with bremelanotide (approximately 40% in trials vs. 10% for flibanserin), but women using bremelanotide report that nausea diminishes after the first several doses.

Dr. Anita Clayton, Professor of Psychiatry at the University of Virginia and lead investigator on multiple HSDD trials, has stated: "The choice between flibanserin and bremelanotide should be individualized. Women who prefer a daily medication and want continuous baseline improvement in desire may prefer flibanserin; those who want on-demand flexibility may prefer bremelanotide" [10].

Ongoing and Future Real-World Data Collection

Several data streams continue to generate flibanserin RWE. Bausch Health maintains post-marketing commitments to the FDA, including periodic safety updates. The FORWARD Registry (Female Sexual Function Registry), launched by ISSWSH, is enrolling women with HSDD and other female sexual dysfunctions across U.S. academic and community sites, with plans to capture treatment patterns, patient-reported outcomes, and long-term safety for both flibanserin and bremelanotide [9].

Electronic health record (EHR) studies using de-identified databases (TriNetX, Optum, Veradigm) have begun publishing on flibanserin prescribing demographics. Early findings confirm underrepresentation of Black and Hispanic women in both prescribing and research, a gap that mirrors the clinical trial enrollment limitations noted during the original FDA review.

Pharmacoepidemiologic work also continues on the CYP3A4 interaction question. A 2023 nested case-control study within a large managed-care population found that concomitant use of moderate CYP3A4 inhibitors (diltiazem, erythromycin, grapefruit juice) approximately doubled the odds of an ER visit for dizziness or presyncope within 30 days of a new flibanserin prescription (OR 2.1, 95% CI 1.3 to 3.4) [6]. This finding reinforces current labeling but suggests the interaction risk is manageable with proper screening.

What the Evidence Means for Prescribing Decisions

Real-world data does not contradict trial findings. It refines them. Flibanserin works for a subset of premenopausal women with HSDD, producing modest gains in desire and satisfying sexual events. The REMS removal confirmed that the safety profile in practice is more favorable than early pharmacokinetic interaction studies implied.

Clinicians prescribing flibanserin should set expectations clearly: response takes 4 to 8 weeks, the drug must be taken nightly, alcohol should be limited (though no longer contraindicated), and CYP3A4 inhibitor co-administration remains contraindicated. Women who have not responded after 8 weeks of consistent nightly use should discontinue the medication, per FDA labeling [7].

The NNT of 8 to 9 from pooled trial data likely applies to well-selected patients in practice. For the broader prescribing population, the effective NNT may be higher, given diagnostic imprecision and adherence realities. Shared decision-making, including discussion of behavioral interventions, hormonal evaluation, and alternative pharmacotherapy with bremelanotide, produces the best outcomes according to ISSWSH consensus guidelines [9].