You’ve probably heard about Fanapta (iloperidone), a new antipsychotic marketed by Vanda Pharmaceuticals. No, it hasn’t been approved officially yet, but it is likely to win approval in July, and the company has already been taking out teaser ads in the journals and has even sponsored a touring promotional CME program run by Stephen Stahl.

Is it likely to represent an advance in antipsychotic treatment? Well…maybe. Is it likely to represent a low point in the naming of new drugs? Definitely. And in fact, Fanapta is a slight improvement over iloperidone’s original trade name, which was – get ready – “Fiapta.”

Before Vanda purchased the right to develop iloperidone in 2004, the drug was owned by Novartis, which conducted the first three controlled trials. Apparently, they were not dazzled by the results, because they promptly sold the drug to Vanda. These studies were summarized in a recent industry-supported supplement to the Journal of Clinical Psychopharmacology (Potkin SG, et al., J Clin Psychopharmacology 2008;28 (Suppl 1):S4-S11).

In Study 1, schizophrenic patients were randomized to five treatment arms: Fanapta at one of three doses (8, 10, 12 mg/day), Haldol 15 mg/day, or placebo. At the 6 week endpoint, Fanapta 8-12 mg/day did not statistically outperform placebo (this was the prespecified outcome variable) whereas Haldol 15 mg/day was robustly superior to placebo. In both Studies 2 and 3, patients were randomized to Fanapta, Risperdal or placebo. Risperdal yielded more robust improvements over placebo than Fanapta, but at least Fanapta was statistically superior to placebo in three of the four doses tested. None of these studies reported the statistical significance of the numerical advantages of Haldol and Risperdal over Fanapta. This omission might be because the active comparator drugs were included for assay sensitivity only – that is, to ensure the integrity of the overall study design. Less charitably, the data may have been left out because Fanapta would have looked bad. Only the company knows for sure….

The Vanda-funded authors went to great lengths to explain why Fanapta did so poorly in comparison to Haldol and Risperdal. Mostly, they blamed the dosing schedule, which specified a gradual ramping up of Fanapta over a two week period. In contrast, the patients on the comparator medications achieved the therapeutic dose within a couple of days, putting Fanapta at a therapeutic disadvantage. (True, but in the real world, a drug that takes two weeks to ramp up is, in fact, less desirable than others.)

In order to correct for this imbalance, the authors did a revised analysis in which they deleted data on all patients who had dropped out of the study before two weeks. This led to a better result for the company – now, all dose ranges of Fanapta beat placebo, although Fanapta’s performance was still numerically inferior to Haldol and Risperdal.

Which brings us to the most recent Fanapta data. Once Vanda took over the drug, they revamped their research methods, either to more accurately test Fanapta’s efficacy, or to simply make certain that it would look good this time. The cynicism is not capricious: one study showed that in 90% of industry-funded comparisons of atypical antipsychotics, the sponsor’s drug came out ahead due in part to manipulations of the research design (Heres S et al., Am J Psychiatry 2006;163:185-194).

At any rate, in the new studies, the company chose to compare Fanapta with Geodon. Why Geodon? In an email response, Vanda said they made the choice because of “similar tolerability profiles” between the two drugs, thereby preventing researchers from guessing which patients were assigned to which drug – an important issue in double-blind trials. We suspect that this is only part of the story.  Fanapta has two distinct disadvantages: it requires a gradual dose titration, and it can widen the QT interval in the EKG. From a marketing perspective, what better way to distract attention from these disadvantages than to compare the drug with Geodon, a medication sharing these liabilities?

In the only Fanapta vs. Geodon study to be published thus far, a total of 593 patients were randomized to either Fanapta 24 mg/day, Geodon 160 mg/day or placebo (Cutler AJ et al., J Clin Psychopharmacology 2008;28 (Suppl 1):S20-S28). After four weeks of treatment, Fanapta and Geodon improved symptoms by the same amount (Fanapta, 12 point PANNS (Positive and Negative Syndrome Scale) improvement; Geodon 12.3 points) and significantly more than placebo (7.1 points).

In terms of side effects, Geodon caused more akathisia, EPS, and sedation than Fanapta, while Fanapta caused more dizziness, orthostatic hypotension, weight gain, and tachycardia. The two drugs lengthened the QT interval by about 11 milliseconds after 14 days of treatment, but this lessened by nearly half after four weeks.

Thus far, then, we have an antipsychotic which appears to work as well as Geodon and which appears to cause less EPS and sedation, but more orthostasis, tachycardia, and weight gain, and the same EKG liability. Not much to get excited about. What Vanda is hoping we will get excited about is the possibility that it may have developed a way of predicting patients who will respond.

An article in the journal Pharmacogenomics reported that patients who had two copies of a specific gene responded differently in the study than patients with only copy (Lavedan C et al., Pharmacogenomics 2008;9:289-301). The gene in question encodes for a protein called Ciliary Neurotrophic Factor (CNTF), which is a neuroprotectant molecule for nerve cells and may theoretically improve responses to antipsychotics. About 75% of Caucasians are homozygous for CNTF (meaning that they carry two copies of the gene).

The researchers compared the response of patients who were homozygous to those who were heterozygous (only one copy of the CNTF gene). As it turned out, there was no difference in response to Fanapta between these two groups, with both showing about a 12 point improvement on the PANNS score. So how is the gene test valuable? The patients who were homozygous responded much more poorly to placebo (5.7 point improvement) than to Fanapta, while heterozygous patients had identical responses to placebo and active drug (see table).

Thus, it’s not really accurate to say that the genetic test predicts a better response to the medication; rather, it predicts a worse response on placebo. While this is intriguing data, it’s not clear how clinicians would actually use it. If my patient is homozygous for CNTF, I would go ahead and prescribe Fanapta. But what if my patient is heterozygous? According to this paper, such a patient will still respond to Fanapta, as well as to placebo. Since it is unethical to prescribe placebos in clinical practice, I’ll still be likely to prescribe a medication. Or, it might be that these genetically identified placebo-responders will end up doing fine with no treatment at all – but the study did not include a “no treatment” arm, so we do not know if this is true.

Finally, does this genetic test predict placebo response to antipsychotics other than Fanapta?  While the Pharmacogenomics paper did not discuss this crucial question, Vanda told me that, yes, “the same trend appears to hold for Geodon.” Because of this, the test appears to provide no rationale for choosing Fanapta over Geodon, or, by extension, over any other atypical. The company says it plans to test other genetic markers that may be specific predictors of response to Fanapta.
TCPR Verdict:

Fanapta: Genomic Advance or Genomic Snow Job? Crucial data is yet to come.