Aripiprazole: the legal and safety record

The regulator moved first

In May 2016, the FDA warned that aripiprazole is associated with compulsive or uncontrollable urges — to gamble, binge eat, shop, and have sex — and required label warnings. The detail that gives the signal its weight: the reported urges stopped when the drug was stopped or the dose reduced [1]. An on/off pattern is what turns case reports into a warning, and that regulatory finding — not any verdict — is the evidence-grade core of this record.

Then the courts, in two tracks

The injury track: roughly more than 2,400 compulsive-behavior cases consolidated as MDL 2734 in the Northern District of Florida from October 2016, resolved by a confidential master settlement dated February 15, 2019 — a date and scope stated in Bristol Myers Squibb’s own SEC filings [2]. Confidential means adjudicated nothing; no court ruled on causation for anyone. The marketing track: in December 2016, 43 states and the District of Columbia settled for $19.5 million over promotion of Abilify to children and to elderly patients with dementia and minimization of risks — allegations resolved, no admission [3].

One circle this page refuses to complete: gambling terms are prominent in aripiprazole’s public adverse-event reports, and years of lawsuit advertising are part of why. Report counts are not evidence the claims were true, which is exactly why our aripiprazole page prints the reporting system’s limits above its numbers.

Before both tracks: the 2007 settlement

In September 2007, Bristol-Myers Squibb agreed to pay more than $515 million to resolve federal civil allegations that included promoting Abilify from 2002 through 2005 for children and for dementia-related psychosis, both unapproved uses, with a long-term-care sales force that, in the Justice Department’s words, “called almost exclusively on nursing homes.” The total also covered unrelated pricing allegations; over $25 million was disgorgement of Abilify profits. Civil, no admission [9]. How antipsychotics reached nursing homes afterwards is on antipsychotics in nursing homes.

How often does it happen? Nobody was required to count

The FDA’s own communication called these problems “rare” and gave, for context, about 1.6 million US patients filling aripiprazole in 2015; its 184 cases are thirteen years of voluntary reports, so dividing one by the other gives no rate [5]. The approval trials were not built to find it. The studies that looked afterwards point the same way but are small: in US insurance claims, aripiprazole users had about five times the rate of pathological gambling (rate ratio 5.23) [6]; in a Quebec cohort of young adults with first-episode psychosis, drugs of aripiprazole’s class carried about 4.6 times the risk of new problem gambling, and 13 of the 18 cases happened on aripiprazole [7]; a 2024 meta-analysis found the class raised the odds but rated the evidence very low quality, and for aripiprazole alone could not exclude no effect [8]. The honest summary: a real, uncommon risk several times higher than baseline, surfaced by voluntary reports years after approval rather than measured by the trials that approved the drug.

What is not in the record

• No adjudicated causation finding — the MDL settled confidentially before any bellwether verdict — a test trial whose outcome guides settlement of the rest (glossary) — resolved the question [2].
• No federal criminal case over Abilify marketing; the state AG matter was civil consumer-protection [3]. (BMS’s 2007 federal settlement — $515 million, civil — DID cover earlier Abilify promotion for pediatric use and dementia-related psychosis, 2002–2005; it was civil, so “no federal criminal case” still holds, and the earlier chapter belongs in the sentence rather than a parenthesis pretending otherwise.)
• No Class I recall — twenty-nine records, all Class II/III, within openFDA's ~2012-onward coverage window [4].

Common questions

What did the FDA actually warn about Abilify?

In May 2016 the FDA issued a Drug Safety Communication stating that compulsive or uncontrollable urges — to gamble, binge eat, shop, and have sex — have been reported with aripiprazole, that the urges stopped when the drug was stopped or the dose reduced, and that labels would be updated to warn about it. That regulatory finding, not any lawsuit, is the evidence-grade core of this page.

How did the Abilify gambling lawsuits end?

In a confidential master settlement dated February 15, 2019, resolving the more-than-2,400-case federal MDL — a multidistrict litigation, the mechanism that gathers similar federal cases before one judge (see /glossary/mdl) — (No. 2734, Northern District of Florida) and related state cases — a date and scope stated in Bristol Myers Squibb’s own SEC filings. Confidential settlements adjudicate nothing: no court ruled that Abilify caused any individual’s gambling.

What was the $19.5 million settlement?

A separate consumer-protection matter: in December 2016, 43 states and the District of Columbia settled claims that Abilify had been marketed for unapproved uses — to children and to elderly patients with dementia — and that risks were minimized. Like most state AG settlements it resolved allegations without an admission.

Is the compulsive-behavior risk real, then?

The FDA concluded the association was real enough to require a warning, and the on/off pattern — urges stopping when the drug stops — is the kind of signal regulators weight heavily. What no public number can tell you is frequency: adverse-event report counts for this drug are inflated by years of litigation advertising, which is exactly why our aripiprazole page prints the reporting system’s limits above its counts.

Should I stop aripiprazole because of this?

No — abrupt antipsychotic stops carry serious risks. The actionable version of this record is a monitoring question: if new gambling, spending, eating or sexual urges appeared after starting or increasing this drug, that is precisely the conversation the 2016 label change exists to prompt, and the on/off pattern makes it a solvable problem with a prescriber rather than a private shame.

Sources

Last verified 2026-09-29. Corrections change this date.