The FDA's Safety Review of Pfizer's Atypical Antipsychotic Drug
Ziprasidone (ZELDOX*)

*NOW MARKETED UNDER THE NAME 'GEODON'

The Food and Drug Administration's (FDA) Psychopharmacological Drugs Advisory Committee met on July 19, 2000 to review the safety and effectiveness of Pfizer's atypical antipsychotic drug ziprasidone (ZELDOX). The committee voted nine to one to recommend approval of the drug as a first-line treatment for schizophrenia. The final decision to approve ziprasidone, including labeling that could restrict its use to second-line treatment and warnings about potentially fatal heart rhythm disturbances, or to not approve the drug is the FDA's. The agency is expected to make its final decision about the drug sometime this year.

Pfizer submitted a New Drug Application (NDA) to the FDA for ziprasidone on March 18, 1998. Three months later on June 17, 1998 the agency issued the company a not-approvable letter based on "the judgement that ziprasidone prolongs the QTc and that this represents a risk of potentially fatal ventricular arrhythmias [heart rhythm disturbances] that is not outweighed by a demonstrated and sufficient advantage of ziprasidone over already marketed antipsychotic drug products."

The QT_c interval is the length of time it takes the large chambers of the heart (ventricles) to electrically discharge and recharge. A prolongation of the QT_c interval can lead to heart rhythm disturbances, or cardiac arrhythmias, such as torsade de pointes and other life-threatening arrhythmias. The QTc interval is measured by an electrocardiogram (EKG or ECG) in milliseconds (msec). The subscript "c" indicates that the QT interval has been corrected for the patient's heart rate. Torsade de pointes is a French phrase that means "twisted point" that describes the appearance of this type of rhythm disturbance on the EKG tracing.

The basis for this article are documents posted on the FDA's web site the day before the July 2000 Psychopharmacological Drugs Advisory Committee meeting. These are the same materials that were distributed to the committee members prior to the meeting and consist of Pfizer's briefing book and the FDA reviews of the data submitted in ziprasidone's NDA by Pfizer. These documents are now available to the public as a direct result of a lawsuit the Health Research Group filed against the FDA under the Freedom of Information Act and the Federal Advisory Committee Act in 1999. We prevailed and since January 1, 2000, the materials distributed to advisory committee members are now, for the first time, available to the public on the Internet, usually the day before a meeting.

Drug companies exercise tremendous control over the information that is reported in the news media and medical literature about their new drugs. We believe that it is no longer possible to form an objective opinion about the therapeutic value of a new drug using only the published medical literature. Any evaluation of a new drug that does not include the FDA's review is incomplete and can lead to an overestimation of the potential therapeutic value of the drug.

The documents distributed to the ziprasidone advisory committee can be found on the FDA's web site at: http://www.fda.gov/OHRMS/DOCKETS/AC/00/backgrd/3619b1.htm

A major concern expressed by the FDA in its June 17, 1998 not-approval letter to Pfizer was that the degree of QTc prolongation observed with ziprasidone may have represented an underestimate given the fact that most ECGs were likely obtained when the blood levels of the drug were the lowest. The FDA recommended that Pfizer conduct another study to determine the QT_c effect of ziprasidone when the drug's blood levels were the highest in comparison to other atypical antipsychotics and with several other standard antipsychotic drugs.

Pfizer complied and conducted a direct comparison of ziprasidone at its optimal dose to haloperidol (HALDOL), olanzapine (ZYPREXA), quetiapine (SEROQUEL), risperidone (RISPERDAL), and thioridazine (MELLARIL) with ECGs obtained when the blood levels of these drugs were at their highest. The Division of Neuropharmacological Drug Products which has the responsibility for reviewing new antipsychotic drugs asked the agency's Division of Cardio-Renal Drug Products to review this study. This later Division has experts in the effect of drugs on the heart.

The study involved 183 patients with normal ECGs and psychotic disorders. Following a screening phase, the study consisted of five different treatment periods: (1) patients who were eligible for the study had their current drugs stopped over a period of several days as out-patients; (2) the patients were admitted to a research facility to be completely withdrawn from the current drugs; (3) the patients were then randomly assigned to one of the six study drugs (ziprasidone, risperidone, olanzapine, quetiapine, thioridazine or haloperidol) with doses increased over 10 to 19 days; (4) after the maximum dose of the study drugs was reached, a potentially interacting drug selected for each study drug was given to the patients; and (5) after the blood levels of the study drugs were stabilized with the combination of the study drug and interacting drug, the patients were withdrawn from treatment. Baseline ECGs were obtained. Additional ECGs were obtained on day 2 of the study, when the blood levels of study drugs were stabilized, and with the potentially interacting drugs. The ECGs were recorded at times estimated to correspond with the maximum blood levels of the study drugs.

Table 1 below summarizes the average change in QT_c interval for the six study drugs and the range when the blood levels for the drugs were stabilized and their concentrations were estimated to be at their highest. Thioridazine clearly had the largest effect in lengthening the QT_c interval followed by ziprasidone.

The number and percentage of patients grouped according to the extent of QT_c interval prolongation is seen in Table 2 below. Again, thioridazine and ziprasidone rank one and two with none of the patients taking the other drugs experiencing a prolongation of 75 msec or greater.

The experience with the antipsychotic drugs sertindole (SERLECT) and thioridazine places ziprasidone's effect on QT_c interval prolongation in some perspective. Sertindole was recommended for approval by an FDA advisory committee and approved by the FDA on October 2, 1996, but the drug was never marketed in the U.S. Before its approval it was known to increase QT_c by about 21 milliseconds. Doctors and pharmacists were notified on July 7, 2000 about the addition of extensive new safety warnings, including a boxed warning, to the professional product labeling, or "package insert," for the thioridazine regarding QT_c interval prolongation, cardiac arrhythmias, torsade de pointes and the possibility of sudden death.

Ziprasidone increases the QT_c on average about 10 to 20 milliseconds and thioridazine approximately 36 milliseconds. Sertindole was marketed in the United Kingdom but was withdrawn on December 2, 1998 following reports of cardiac arrhythmias and sudden cardiac death associated with its use. The European Commission reached a decision on February 25, 2000 on a European wide suspension of the marketing authorization of sertindole for one year. The magnitude of the increase of the QT_c and QT interval is considered to be important; it is not predictive of the degree of risk of torsade de pointes or other serious cardiac arrhythmias.

The FDA must balance the potential risks of a cardiac arrhythmia and sudden death against the research documenting what is known about the efficacy of ziprasidone.

Ziprasidone's Efficacy

The FDA was in general agreement with Pfizer regarding the antipsychotic efficacy of ziprasidone based on short-term, fixed dose, clinical trials using an inactive dummy drug or placebo for comparison. The FDA was careful to point out that they were not "aware on any evidence from these or any other studies of any superior antipsychotic efficacy for ziprasidone compared to any other antipsychotic drugs, either in typical schizophrenic patients or in those shown refractory to standard antipsychotic therapy."

The FDA did not release their own detailed analyses of the clinical trials that had previously been submitted by Pfizer in support of ziprasidone's approval. We had to depend on Pfizer's interpretation these trials. Pfizer presented the results of five clinical trials in their briefing materials distributed to the advisory committee. Four of these tested the effect of ziprasidone in the management of the acute symptoms of schizophrenia. The studies lasted four to six weeks and involved 702 patients who received ziprasidone in doses ranging from 10 milligrams to 200 milligrams, 273 patients that received a placebo, and in one trial 85 patients who were given a fixed dose of haloperidol 15 milligrams. The fifth trial assessed the effect of ziprasidone in preventing schizophrenia relapse. This last trial ran for 52 weeks, 219 patients received ziprasidone in doses from 40 milligrams to 160 milligrams, and 75 patients received the inactive placebo.

Various rating scales were used to test the efficacy of ziprasidone. A brief description of these scales can be found in the side-bar accompanying this article. In general, the standard for showing that a drug has a treatment effect (efficacy) is that the average scores on the rating scales improve and that the difference between the averages for those patients receiving ziprasidone, for example, are statistically different from the group of patients receiving the inactive placebo. Overall, Pfizer found that ziprasidone treatment was statistically superior to no treatment at all.

In the single trial in which ziprasidone was compared to an active antipsychotic drug, ziprasidone, in several different dosages, was compared to 15 milligrams of haloperidol. Pfizer estimated that the treatment effect for haloperidol was greater on the four rating scales used in the study compared to all dosages of ziprasidone tested. Pfizer did not determine if the difference between haloperidol and ziprasidone was statistically significant and the company did not provide sufficient information for us to make this determination, nor did Pfizer discuss this finding in the narrative describing the trial.

Pfizer found in the 52-week long study testing the effect of ziprasidone compared to placebo in the prevention of schizophrenia relapse that ziprasidone was statistically superior to the placebo. At the end of the study, about 25 percent of the patients given the placebo were relapse free and approximately 60 percent of the ziprasidone patients were also relapse free.

The evidence presented and interpreted by Pfizer suggest that ziprasidone is superior to an inactive placebo in the treatment and in the prevention of schizophrenia symptom relapse. But, ziprasidone may actually be inferior to haloperidol, based on four rating scales, in the treatment of schizophrenia symptoms.

Ziprasidone and Weight Gain

Pfizer highlighted the weight gain seen with other antipsychotic drug compared to ziprasidone in its briefing materials distributed to the advisory committee. Table 3 below is based on information presented in the company's briefing materials.

Pfizer promoted ziprasidone as "weight neutral" in a July 19, 2000 press release announcing the advisory committee's recommendation to approve the drug. QT_c interval prolongation was mentioned but no explanation was provided as to its potential serious consequences. Manufacturer's make it a point to minimize potential adverse effects. It remains to be seen if this difference in weight gain claimed by Pfizer will persist between ziprasidone and other antipsychotic drugs if the drug is approved and prescribed to large numbers of patients outside of a controlled experimental setting.

The law will not allow the FDA to require that a manufacturer show a new drug is either safer or more effective than currently marketed drugs as a condition of approval. In fact, new drugs may be less safe or effective than drugs already on the market. When Congress decides to "raise the bar" for approving new drugs, the American public will have better drugs, not just more drugs.

Based on the FDA's review of Pfizer's study showing the extent of ziprasidone's effect on QT_c interval compared to other antipsychotic drugs and the company's failure to demonstrate an advantage over currently available drugs, the Psychopharmacological Drugs Advisory Committee's recommendation to approve ziprasidone as a first-line treatment was irresponsible. If approved, ziprasidone should be recommended as a second line antipsychotic. If its relative lack of weight gain is confirmed in longer-term studies, it may be useful for patients on antipsychotics for whom weight gain is a major problem. However, the cardiac effects of this drug are troublesome. Long-term follow up and monitoring of patients taking it will be necessary to insure that the risks of taking it do not outweigh any possible benefits.