Customize your JAMA Network experience by selecting one or more topics from the list below.
Flibanserin (Addyi), the new female libido pill, is about desire, arousal, and satisfaction. The manufacturer, Sprout Pharmaceuticals (a division of Valeant Pharmaceuticals North America LLC), clearly had desire. They purchased the rights to the drug even after its initial rejection by the US Food and Drug Administration (FDA) and persisted through 2 contentious review cycles. Sprout worked hard to arouse support for the drug, helping create and fund “Even the Score,” an advocacy campaign pushing the message that sexism—not legitimate scientific questions—motivated the drug’s rejection.1 And within 48 hours of FDA approval, flibanserin was sold to Valeant Pharmaceuticals for about $1 billion in cash.2 Very satisfying.
Very satisfying for Sprout, but what about the women who take flibanserin? What happens to their desire, arousal, and satisfaction? Very little, according to the meta-analysis by Jaspers et al3 in this issue. Premenopausal and postmenopausal women taking the approved dose—compared with placebo—experienced 0.5 more satisfying sexual encounters a month and scored 0.3 points higher on a 5-point sexual desire scale. Jaspers et al concluded that these modest benefits did not outweigh harms. According to the FDA, about 10% more premenopausal women taking the approved dose (100 mg) of flibanserin had a meaningful benefit (ie, were “much” or “very much” improved), but flibanserin increased somonolence, sedation, or fatigue compared with placebo: 21% vs 8%.4 Perhaps most importantly, combining flibanserin with alcohol (and other common drugs) can cause dangerous hypotension and syncope—problems so serious that the FDA put a black box warning, its most serious safety alert, on the label.
Given the Benefits and Harms, Why Did FDA Get to Yes?
The story began in 2009 when Boehringer Ingelheim first applied for approval of flibanserin, a failed antidepressant, to treat hypoactive sexual desire disorder in premenopausal women. But in 2 pivotal efficacy trials,5 flibanserin failed to improve sexual desire measured with a daily electronic diary—the prespecified coprimary outcome. And it caused substantial somnolence and appeared to have dangerous interactions with alcohol and other drugs. Concluding that the drug’s uncertain benefit did not outweigh the harms, the FDA’s clinical reviewers and an external advisory committee voted unanimously (11-0) against the drug. In their letter, the FDA told the company that moving forward would require a third efficacy trial with less restrictive entry criteria. The original trials excluded women with comorbidities, including mild depression and anxiety, and prohibited many commonly prescribed medications, exclusions which the FDA believed “preclude[d] a full assessment of efficacy and safety in the target population.”6 The FDA also emphasized the need to validate any new primary outcome measure of sexual desire and requested studies to better define interactions between flibanserin, other drugs, and alcohol.
Boehringer Ingelheim sold the rights to flibanserin to Sprout Pharmaceuticals, which resubmitted it for FDA approval in 2013. The resubmission included the new efficacy trial and interaction studies.6 The efficacy trial (still with restrictive entry criteria) demonstrated a small benefit using a new, inadequately validated coprimary sexual desire measure. The alcohol study, bafflingly performed in men (23 of the 25 participants), not premenopausal women, demonstrated serious interactions. Flibanserin plus alcohol substantially increased hypotension compared with either substance alone: 4 of 23 men taking flibanserin with 2 alcoholic drinks developed symptomatic hypotension requiring intervention (placed supine or in Trendelenburg position), and 1 became unconscious. The study testing flibanserin plus fluconazole (a moderate CYP3A4 inhibitor) was stopped early after 3 of 15 women developed symptomatic hypotension, and 1 became unresponsive (requiring an emergency department visit with intravenous fluids). Concluding that flibanserin provided only “numerically small treatment differences compared to placebo which do not clearly outweigh safety concerns,”6 the FDA rejected it again. Sprout unsuccessfully appealed to the FDA.
Flibanserin Resubmission for FDA Approval: Third Time the Charm
In 2015, Sprout resubmitted flibanserin to the FDA, and the drug was approved. What changed? Nothing about efficacy. The resubmission included no new benefit data.7 A new driving study, requested by the FDA to rule out impaired driving from drug-induced somnolence, was reassuring, but new pharmacokinetic studies heightened concerns about dangerous drug interactions: 1 of 9 “poor metabolizers (people with a CYP2C19 PM genotype which increases flibanserin levels)” became hypotensive and unresponsive after 1 flibanserin dose.
But something else happened. Sprout helped launch a new advocacy group, Even the Score, which conducted an intense promotional campaign directed at journalists, women’s groups, Congress, and the FDA.1 The FDA received “more than 2 dozen letters from various groups and individuals, including advocacy organizations, clinicians, researchers, and members of Congress,”8 most pushing for approval because women had no other drugs for low sexual desire. Some accused the FDA of sexism because it had approved so many medications treating male sexual dysfunction (a claim the FDA deemed misleading because FDA-approved drugs for men are approved for erectile dysfunction, not for sexual desire).
When the FDA held a second Advisory Committee meeting, many Even the Score supporters attended, testified in favor of the drug, and repeated the sexism charge.9 The Committee voted 18 to 6 for approval. All “yes” votes were qualified—they believed safe use of flibanserin needed risk evaluation and management strategies (REMS) beyond strong labeling (eg, only allowing prescriptions from physicians and pharmacists with special training).
The FDA’s own clinical reviewers, however, still recommended rejection (Figure).10-17 The 2 medical reviewers and team leader believed—as in both prior review cycles—that the benefit-harm balance was unfavorable: “We do not believe that it is reasonable for the approximately 90% of treated patients who will not respond to the product to be exposed to the numerous serious risks posed by flibanserin therapy.”12 They also argued that there was no evidence that risk management strategies (like those proposed) ensured safe use. Based on the syncope event in the new drug interaction studies and the lack of an alcohol study in premenopausal women, even the clinical pharmacology reviewer and team leader, who had previously supported approval, now voted against it.16
The nonclinical pharmacology/toxicology team supported approval but concluded that there “was a drug-induced increase in malignant mammary gland tumors in female mice. Although the increase was small, the increases were dose-related and statistically significant…. The clinical significance of the mammary tumors in mice is unknown.”10 They recommended including this concern in the “Warnings and Precautions” of labeling to allow physicians and women to factor this uncertainty into their decisions. Red and green boxes represent the number of individual votes.11 HSDD indicates hypoactive sexual desire disorder; REMS, risk evaluation and management strategies.
The clinical review teams were overruled. While the FDA publicly downplayed the dissent (saying “not all members of the review team recommended approval”18), the score was lopsided: 2 division directors voting for approval vs 5 medical and pharmacology reviewers and team leaders who did not. The Clinical Pharmacology19 and Clinical Division Director (the official signing the approval letter)14 recommended approval. The Directors acknowledged the limited efficacy and worrisome harms, but emphasized the unmet need in women for whom other treatments failed. They believed that a black box warning and the REMS would ensure safe use by women in the narrow population for whom it was approved. They believed that the open questions could await answers from postmarketing studies. And FDA emphasized that the decision was not influenced by charges of sexism but solely on the evidence.
While it is unclear how strongly politics influenced the decision, it is clear that the science was weak. How can the FDA make strong decisions based on weak science? Two policy changes could help.
Mandatory Response to the FDA’s Advice on Major Study Design Elements
Flibanserin’s evidence base was inadequate in part because the company did not follow the FDA’s advice about study design. The additional phase 3 efficacy trial was not conducted in a broader population closer to likely users; the new desire coprimary outcome was not adequately validated; and the alcohol interaction study was performed in men. Consequently, major uncertainty remains about the benefits-harm balance of flibanserin in a real world setting. Federal regulations encourage FDA–drug company interactions during drug development, including the design of key studies, but do not require companies to follow the FDA’s advice. A recent review of new drug approvals shows that companies do not always seek FDA advice and too often ignore advice that would have led to higher-quality evidence.20 Unfortunately, companies know that they can ignore FDA suggestions and still get drugs approved. Companies should be required to seek and respond to FDA advice about major study design elements (eg, randomization, primary outcome, entry criteria, drug doses tested, comparators, trial duration).
For Non–Life-Threatening Conditions, Do Not Gamble: Address Serious Harms Before FDA Approval
For men, flibanserin plus even a modest amount of alcohol intake caused symptomatic hypotension and syncope. For women, the combination could be worse. Since flibanserin treats a non–life-threatening condition, the clinical reviewers wanted to know how often symptomatic hypotension and syncope occurred with alcohol before approval. Unfortunately, physicians and patients will not know the answer for a year to 2.5 years—until the 3 required postmarketing alcohol studies are reported. When the FDA approves drugs with important postmarketing requirements, physicians and consumers should be alerted to the required studies and the uncertainties motivating them. Currently, they can only find this information by searching FDA databases or reading approval documents. Instead, the studies should routinely be highlighted in physician and consumer drug information (building on the new warnings required in labeling for accelerated approval drugs).
The flibanserin saga is unsatisfying. The FDA approved a marginally effective drug for a non–life-threatening condition in the face of substantial—and unnecessary—uncertainty about its dangers. Women with distressing sexual desire problems need good treatments. We all need a drug approval process that delivers good decisions based on adequate evidence.
Corresponding Author: Steven Woloshin, MD, MS, Center for Medicine and the Media, Dartmouth Institute for Health Policy and Clinical Practice, 35 Centerra Pkwy, Lebanon, NH 03766 (email@example.com).
Published Online: February 29, 2016. doi:10.1001/jamainternmed.2016.0073.
Conflict of Interest Disclosures: Drs Woloshin and Schwartz are cofounders of Informulary Inc, a company that provides data about the benefits, harms, and uncertainties of prescription drugs. No other disclosures are reported.
Additional Contributions: We wish to thank Aaron S. Kesselheim, MD, JD, MPH, and Brian White, BA, for their insightful comments on an earlier draft of this article. They received no compensation for their contributions.
Woloshin S, Schwartz LM. US Food and Drug Administration Approval of Flibanserin: Even the Score Does Not Add Up. JAMA Intern Med. 2016;176(4):439–442. doi:10.1001/jamainternmed.2016.0073
Create a personal account or sign in to: