A recent article from CNN discussed Nuplazid® (pimavanserin), a medication approved by the FDA in 2016 for treatment of hallucinations or delusions in Parkinson’s disease (a.k.a. PDP, for Parkinson disease psychosis). The article made some efforts to present a balanced perspective, but the title and overall tenor of the article focused on the possibility that this new drug might be killing people after being rushed to market.
It may eventually turn out that people taking pimavanserin have a higher death rate than people taking a placebo. For any new drug, it takes far more experience to reliably estimate how common side effects are than it does to know whether it works. But the data so far do not suggest that pimavanserin is dangerous. In fact, the FDA reached that same conclusion. But I think most readers of the CNN article would not realize why I (and the FDA) reached that conclusion, so I offer the following information.
- People taking pimavanserin are old and sick. Patients in the pivotal clinical trial were (on average) 72 years old, had had Parkinson’s disease for 10-12 years, had experienced psychotic symptoms for 3 years, and were on at least one antiparkinsonian medication. More than half have other psychiatric diagnoses, such as major depression or panic attacks. Based on my own patients, I’d guess the average patient in that study was on 5-10 other medications, as well. Their UPDRS scores suggest that some would have mild symptoms (while on medications) but others would be in a wheelchair. This group is not unusual for people with PDP. The mortality rate in PDP has been estimated as 25% at 2 years and 80% at 7 years. It can’t all be blamed on treatment for psychosis, because more than 5% of untreated PDP patients had died when followed up at only 6 months (see Figure). In trials of other drugs in PDP, deaths occurred during open-label treatment or when on placebo. The point is, at some point in life, death comes to everyone, and it’s a lot more likely when you’re in your 8th decade of life and have a complex medical history.
- Alternatives? As of early 2018, someone with PDP has 5 main treatment options, listed below. None of the alternatives to pimavanserin is very appealing. Weintraub and colleagues reviewed the VA’s treatment database just before the FDA approval of Nuplazid®, and reported annual mortality rates in PD patients diagnosed with psychosis and taking antipsychotic drugs. The mortality rates in the next few bullet points come from that paper.
- Most antipsychotics will worsen your parkinsonism. If that weren’t enough of a problem, the mortality rate in PDP patients on these medications was 29-49% per year in the VA study. Experts agree that these medications should not be used in PDP.
- Quetiapine is still widely used in PDP, but 5 studies have compared it to placebo, and none of them met their predefined criterion for superiority. Only one showed any hint of efficacy. Why take a medicine that doesn’t work? The mortality rate in PDP was 19% per year.
- Clozapine is the only medication other than pimavanserin to be proven more effective than a placebo in PDP. It works well, but taking it is a hassle, requiring weekly blood draws for months, fading after a year to monthly blood draws. The reason for the blood draws is for early identification of a side effect that happens in 1-2% of people who take it and can be lethal if missed. Too few patients in the VA study were on clozapine for its mortality rate to be reported separately. But in one of the first open-label clozapine studies in PDP, 6 of 53 patients died in 3 months, for a raw mortality rate of 45% per year.
- No treatment. Most patients who are having enough PDP symptoms to be prescribed a medication for them are really bothered by their symptoms. This is a point that the CNN article missed. Seeing something that’s not there is usually very disturbing. When symptoms are at their worst, patients may run out of the house in winter, tear apart their house looking for surveillance devices, or even get out their gun to defend themselves. And the symptoms don’t usually get better on their own; in fact, the severity of psychosis tends to increase as the disease progresses, even if it starts out very mild. We have fewer data on patients with symptoms severe enough to treat who are untreated, but in the VA study, more than 5% of untreated patients were dead in 6 months, for an estimated annual mortality rate of >10%.
- Pimavanserin. We know pimavanserin works better than a placebo. ACADIA probably has more accurate numbers, but using the total sales number from the CNN article and the current pharmacy price of the drug, one can estimate at least 3,582 patient-years of exposure to the drug over the time frame in which the 244 reports of death occurred, for a raw mortality rate of 6.8% per year. Even if the true rate is twice that high, it compares well to the mortality rates for the other options listed above.
- Pimavanserin is a treatment, not a cure, and PDP is usually chronic if untreated. We have no cure for PDP at this point in time, and if the drug works, you’ll want to stay on it. So the plan for many patients will be to take it as long as they live, and of course that means patients will be taking it when they die.
- Mandatory reporting. The specialty pharmacy and ACADIA’s patient support program check in with patients from time to time and they are required to report every illness they hear about, whether related to pimavanserin or not, and of course every death. However, ACADIA will not necessarily have information on all the other drugs patients are taking, or on all their other illnesses. Makers of other drugs generally will not have such information, only cases that patients or doctors spontaneously report.
So for now, pimavanserin is still one of the best treatment options for these patients. A recent review article agreed, recommending pimavanserin and clozapine as the only PDP treatments supported by good evidence.
Disclosure: ACADIA Pharmaceuticals, the manufacturer of Nuplazid®, has paid me for consultation and speakers bureau presentations, and reimbursed Washington University for research studies performed here for which I was the site investigator. They had no input into this blog post. This post represents my personal opinions, not necessarily those of my employer.