Since most of you seem to have not gone to the source of the CNN article here it is :
©Institute for Safe Medication Practices 2017 Q1 QuarterWatch – Page 7 of 18 www.ismp.org/QuarterWatch/
In 2017 Q1 data, the volume of reports of adverse drug events to the FDA remained relatively stable, as it has for the preceding eight quarters. For the latest quarter the FDA received 297,010 new adverse drug event reports with 79,669 (26.8%) describing events that were fatal or serious and occurred in the U.S. Overall, 95.4% of reports were prepared and submitted by industry, with just 4.6% of cases reported directly to the FDA by consumers and health professionals. The quartertly total included 16,307 reported patient deaths in the U.S. and 11,944 death cases from abroad. In previous reports, we have described the limitations of reports of patient deaths from manufacturers, who report patient deaths they learn about irrespective of whether the drug was suspected of being a contributing factor.
Pimavanserin (NUPLAZID) and Hallucinations
Pimavanserin (NUPLAZID) was a new kind of antipsychotic drug with a novel mechanism of action. It received expedited consideration at the FDA because it targeted an important unmet medical need. While the initial target population of psychosis in Parkinson’s disease was of modest size, the company is seeking expand its approved use in the future into much larger patient populations with schizophrenia and Alzheimer’s. However, the FDA medical officer responsible for evaluating the company’s application recommended against approval and was overruled. Additionally, in 3 out of 4 clinical trials for efficacy, the drug failed to demonstrate a benefit in reducing hallucinations and other symptoms of psychosis. And now the first substantial group of adverse drug event reports reinforces the concerns of those who warned that pimavanserin might do more harm than good.
Hearing, Seeing Things Not There
Patients with hallucinations hear and see things that are not there. Visual hallucinations can be as subtle as a mysterious shadow glimpsed from the corner of the eye to fully-formed persons, strange objects, or animals. Patients might hear sounds ranging from background crowd noise to clear voices speaking understandably. The individuals experiencing these effects may realize what they are seeing or hearing is not real. And they may not. Hallucinations are part of a broader spectrum of symptoms of psychosis, which include paranoia, delusions, and other indications of a loss of touch with reality. Hallucinations commonly occur in schizophrenia. They can be induced by illegal drugs such as LSD and high doses of amphetamines, methamphetamines, or cocaine. Hallucinations are also induced by levodopa (usually given in combination with carbidopa), the initial and primary treatment for Parkinson’s disease.
Hallucinations in Parkinson’s
In Parkinson’s disease, the dopamine-producing neurons are progressively lost. The U.S. patient population is estimated to be approximately 1 million persons, mostly over age 65. The most immediate consequences of Parkinson’s are movement disorders, including tremors, muscle weakness, slowed movement, and later muscle rigidity. One primary early treatment is to increase the supply of dopamine through administering levodopa, a precursor of dopamine. An additional treatment is to add dopamine agonists, drugs that bind to dopamine receptors in the brain and central nervous system. But these drug interventions do not replicate the complex signaling that occurs in the nearly instantaneous neurotransmission of dopamine, and dopamine involves many different processes in addition to movement, notably mood, impulse control, memory, and attention. The consequence of the drugs used to treat Parkinson’s, combined with the progression of the disease, is that 20%-70% will develop hallucinations and other symptoms of psychosis. For years the scientific literature held that the hallucinations seen in Parkinson’s were entirely drug induced; later studies concluded that some cases occurred in the absence ©Institute for Safe Medication Practices 2017 Q1 QuarterWatch – Page 8 of 18 www.ismp.org/QuarterWatch/
of the dopamine drugs. The standard but off-label treatment of hallucinations in Parkinson’s seems a medical contradiction in terms. Customary treatment is to administer antipsychotic drugs, notably quetiapine. But these drugs work by blocking the normal function of dopamine receptors (primarily a subfamily called D2 receptors). Not only did this make no sense in patients already losing dopaminergic function, but in addition, antipsychotic drugs independently induce movement disorders similar to those occurring in Parkinson’s in approximately 35% of treated patients. A significant additional drawback is that antipsychotic drugs carry a Boxed Warning that they are not recommended for use in elderly patients. This is because most Parkinson’s patients fall into the elderly category, and the drugs have been found to approximately double mortality.
A California pharmaceutical startup called Acadia Pharmaceuticals pursued a different path. If blocking dopamine transmission in patients with already impaired dopamine neurons did not make sense, that did not exclude the possibility of manipulating other neurotransmitters. Hallucinations are not exclusively linked to impaired dopamine function. Another contributor is an important subfamily of serotonin receptors, called type 2 or 5-HT2a. These type 2 serotonin receptors mediate many functions, and notably have a central role in learning, memory, and cognition. However, the likelihood that type 2 serotonin receptors might also contribute to hallucinations was partly supported by these findings: 1) Some illegal hallucinogenic drugs also bind and stimulate type 2 serotonin receptors; 2) The antipsychotic quetiapine blocked both dopamine and serotonin type 2 receptors.
The idea behind pimavanserin was that it would not block or stimulate dopamine receptors, and instead would target type 2 serotonin receptors. Not only would it prevent signal transmission with the brain’s own serotonin, it had additional effects suppressing the activity of the neurons with these receptors. If it worked, potential medical uses were numerous: Existing antipsychotics are poorly tolerated and marginally effective ; combining them with pimavanserin might increase beneficial effects, or permit lower doses of the antipsychotics, and therefore reduce drug-induced movement disorders. The Parkinson’s disease population was also substantial, and even larger was the aging population with Alzheimer’s. So, on one hand, here seemed to be an entirely new antipsychotic agent for a potentially large patient population where more effective and humane treatments were needed. However, on the other hand, effectively shutting down a key serotonin system central to learning, memory, and cognition could have substantial adverse effects.
The Testing of Pimavanserin
Early testing of pimavanserin in patients began in 2004. The first major trial to establish benefit and assess safety began August 2005 in 400 patients with schizophrenia. It was designed to test whether adding pimavanserin treatment to approved antipsychotic drugs would improve outcomes measured on a standard scale of psychosis. It was completed in 2007. Since no final published results are shown on the official clinical trials web site, presumably few if any benefits were found. The next year the company began testing pimavanserin in Parkinson’s disease psychosis, launching a Phase 2 (or proof of concept) trial, and then two Phase 3 trials (demonstrating an effect that can be replicated in the likely treatment population). One of the Phase 3 trials was terminated early for futility; the other two trials did not provide evidence of benefit in treating hallucinations and other symptoms of psychosis in Parkinson’s patients.
A Meeting at the FDA
After three trials that had failed to show a treatment benefit, the manufacturer met with the FDA seeking an agreement that the agency would approve the drug on the basis of a single new trial. Despite the previous failed trials, the agency agreed to judge the drug on a single trial (rather than the normal 2 trials) provided it could demonstrate "strong" effects. In addition to reducing the required level of evidence, the agency provided additional incentives, declaring it a "breakthrough drug" and agreeing to a faster "priority review." To grant pimavanserin "breakthrough" status was a stretch, since breakthrough status requires preliminary clinical evidence of a "substantial improvement" over available therapies. At that point, pimavanserin ©Institute for Safe Medication Practices 2017 Q1 QuarterWatch – Page 9 of 18 www.ismp.org/QuarterWatch/
hadn’t demonstrated improvement, although no other treatments had been approved for comparison purposes.
A Pivotal Trial Redesigned
Seeking a better chance of demonstrating a benefit, Acadia redesigned its next trial. The accepted measurement scale for measuring changes in psychotic symptoms contained 20 items. The company reduced the scale to just 9 items, a new scale subset that had not been previously validated. It excluded patients with milder psychosis. It eliminated doses lower than 34 mg, twice daily. Because results had been weaker in the foreign clinical trial sites, the new trial was conducted only in North America. And all the assessments of possible benefit were conducted via video link by a single blinded, central rating team. An initial two-week period of non-pharmacological treatment was added to eliminate placebo responses.
The trial results in 199 patients provided evidence of benefit using the new psychosis rating scale. The results also showed it was not making the other Parkinson’s symptoms worse (a major risk of antipsychotics). It was hard to evaluate the clinical relevance of the new scale, but the later FDA assessments suggested an improvement of about 23%.
The FDA Medical Reviewer Objects to Approval
A hallmark of the FDA’s long tradition of critical scientific evaluations is not only that it allows mid-level and senior staff dissent, but also that their reports are usually made public. The medical reviewer, a psychiatrist named Paul J. Andreason, outlined numerous concerns in a 160-page review.
While noting that the results indicated a statistically significant benefit, he had other concerns. While two trials (evidence that can be replicated) were normally required, the FDA had agreed to act on a single trial. The primary measurement scale had not been used before with other drugs. The patient population was "enriched" to focus on patients most likely to respond, rather than typical patients. And 13.4% of the pimavanserin patients had protocol violations, notably use of banned antipsychotic drugs. In addition, he questioned whether the change of 3 points on a 45-point scale was clinically significant. Using one accepted assessment, a change of the effect size seen in the trial was qualitatively described as "minimal." An agency expert on clinical assessments was also consulted, and declared "We also conclude that a 3-point change (out of 45) does not clearly represent a clinically meaningful change." This expert thought it would take a benefit about twice that reported to achieve a clinically meaningful change. But notably, the medical officer concluded that despite these marginal benefits he would have still recommended approval because, he said, the agency had approved numerous other psychiatric drugs with this level of evidence. His primary grounds for rejection of approval were on the other side of the drug balance: safety.
The Safety Profile
The first concern was that despite 18 clinical trials conducted with pimavanserin, the size of the safety database was small; just 1,096 patients were exposed to the active drug (1,500 is the usual standard); and only 202 patients with Parkinson’s were exposed to the recommended 34 mg dose.
What moved the medical reviewer toward rejecting the drug was that "There is a disproportionate death and serious adverse event risk in pimavanserin 34-mg daily treatment versus placebo." A drug that harms patients measured by death and serious adverse events meets an elementary definition of an unsafe drug, even if it improves some symptoms. While the numbers of patients were small, excesses of deaths or serious injuries were seen across different measures. Lumping together all the pimavanserin treated patients, 5.3% died, compared to 0.5% on placebo. Limiting the comparison to controlled trials at the recommended dose, treated patients were 2.4 times more likely to experience a serious adverse drug event (including death). Just measuring dropouts for adverse drug effects in its one pivotal trial, four times as many taking pimavanserin were discontinued for adverse effects (9.5% vs 2.1%) compared to patients on placebo. ©Institute for Safe Medication Practices 2017 Q1 QuarterWatch – Page 10 of 18 www.ismp.org/QuarterWatch/
These findings, while consistent across trials and different measures, had a limitation. The specific adverse reactions found were diverse, occurring in numerous body systems. The drug had no signature risk, such as liver toxicity or serious infections. In addition, while deaths were more numerous in the treatment group, mortality rates are substantial in the Parkinson’s disease population. On the other hand, a similar diverse but substantial increase in mortality had also occurred in antipsychotic drugs, leading to an FDA Boxed Warning against use in patients 65 years and older.
The safety data for the pivotal trials at the recommended dose included one additional finding of concern, illustrated in Table 2. When hallucinations were directly reported to study investigators as an adverse event, they occurred more frequently among pimavanserin patients than among those taking a placebo. This raised further questions about the accuracy of the new psychosis rating scale and the use of assessments through video links. Event TermPimavanserinPlacebon = 202n=231Hallucination5%3%Confusional state6%3%* Reproduced from prescribing informationTable 2. Psychiatric adverse reactions in 6-week placebo controlled studies*
The Approval Decision
Neither an FDA advisory committee nor senior FDA management was moved by the medical officer’s concerns. The committee voted 12-2 to recommend approval, which soon followed. In a published explanation, agency officials (including the medical officer as coauthor) discounted the safety evidence. It noted that 51 deaths had occurred in open label trials, but said "no drug-related cause was apparent." As for the clinical trials where a comparison group was present, the authors concluded "the number of events was too small to reach a firm conclusion."
The Adverse Event Report Data
The first substantial group of adverse event reports provides signals that support the known safety concerns, and contribute an additional one. As noted in Table 1, the leading reported adverse events were remarkably similar to the concerns raised by the medical officer’s report and the marginal benefits seen in three clinical trials, notably: numerous reports of hallucinations (n = 487) and confusional state (n = 258) as well as complaints that the drug was ineffective (n = 333) and numerous patient deaths (n = 244). It was also notable that 74% of the pimavanserin reports came from health professionals, who would be expected to be familiar with the symptoms and mortality of patients with Parkinson’s disease psychosis, and therefore less likely to report these events unless a drug role was suspected. If there was an overall message in these adverse event reports, it was that hundreds of health professionals were trying this new drug in their patients and reporting that either it didn’t work, or in some cases made the patients worse.
However, the adverse event reports revealed a new safety concern: concomitant use of antipsychotic drugs, notably quetiapine (SEROQUEL). The reports revealed 318 cases of heavily medicated Parkinson’s patients (a median of 10 drugs) where pimavanserin had been added to antipsychotic drugs (all but 7 cases were quetiapine). The health outcomes were somewhat worse comparing patients also taking antipsychotics to those who did not: deaths in 13% versus 11%, hospitalizations in 15% vs 11%. No antipsychotic drugs have been approved for use in Parkinson’s patients, and the FDA explicitly warns against their use in dementia-related psychosis in patients over 65 years age. However, it was an open secret that quetiapine and other antipsychotics were widely used anyway.* On its face, combination therapy with quetiapine and pimavanserin appears unsound. Patients with already impaired dopamine function are given quetiapine,
* In the FDA review article explaining the approval of pimavanserin, senior management noted that although antipsychotic use was not recommended, "the FDA, in public statements, did not suggest the use was unreasonable."  ©Institute for Safe Medication Practices 2017 Q1 QuarterWatch – Page 11 of 18 www.ismp.org/QuarterWatch/
which not only blocks normal dopamine signaling, but also blocks the same type 2 serotonin receptors targeted by pimavanserin. Further, the combination therapy has never been tested. And both drugs are suspected of increasing mortality.
Interpreting the Adverse Drug Event Reports
The adverse event reports for pimavanserin presents challenges that not only include but go beyond the normal limitations of these data. Hallucinations are the leading symptom of psychosis and so a possible adverse effect of the drug might be mistaken for a symptom of a worsening disease. However, we considered this a bona fide signal for several reasons: A) The number of reports was so large that it outnumbered similar reports for all other drugs, including antipsychotics. Sheer numbers of reports have scientific weight because it is unlikely that so many observers were uniformly wrong. B) The reports came predominantly from health professionals who apparently suspected a drug effect. C) An excess of hallucinations as an adverse drug reaction was also observed in clinical trials. On the other hand, these adverse event report totals do not support useful direct comparisons of event counts for other drugs, including antipsychotics. This is because of very large differences in patient exposure, and likely reporting rates. Drug manufacturer reports of patient deaths are challenging to interpret because if a company learns that a patient taking its drug has died, the FDA requires an early report, even if a possible drug role has not been investigated. However, the large number of reported deaths remains of concern in a setting where increased mortality was at least suspected if not proven in the clinical trials assessments at the FDA.
The Company Response
We shared our preliminary findings with Acadia Pharmaceuticals, which provided a response. It noted that the total volume of reports was expanded because of the company’s extensive direct interactions with health professionals and consumers because of its distribution of the drug through a network of specialty pharmacies rather than retail pharmacies.* The company also said it believed that the reports of hallucinations and confusion could have occurred during early treatment "while patients were becoming accustomed to the drug." It noted that it takes 12 days to reach steady state concentrations, and benefits were not seen in the clinical trial until after 4 weeks. With regard to mortality, the company noted that in a new trial (that had been completed but not yet published) in Alzheimer’s psychosis, "we observed no difference in deaths."
The company’s reported contacts with health professionals and patients likely did expand the number of reports, essentially because its postmarket surveillance was more extensive than for many other drugs. On the other hand, that should mean that this group of reports more accurately reflects the experience of health professionals and consumers in a real-life postmarket setting.
The numerous reports that the drug was ineffective underline the limited benefits seen in the clinical trials. Even the best trial result, a 23% decline in a sensitive measurement scale, observed only after 4 weeks of therapy, might be too small for either health professionals or consumers to observe.
We were also concerned about adverse event reports indicating off-label use in combination therapy with quetiapine, a powerful antipsychotic that was banned in the pimavanserin clinical trials. The FDA and manufacturer should consider additional warnings and other measures to deal with inappropriate combination therapy.