Pfizer Ready For Fablyn Vs. Evista Match, But FDA May Not Play Along

The long-term data from Pfizer's PEARL study are the firm's best bet for eking out a competitive distinction for its osteoporosis drug Fablyn against Lilly's Evista . FDA, however, seems less inclined to put those findings in labeling, particularly the breast cancer data.

When the data were discussed at the Sept. 8 advisory committee review of Fablyn, FDA's Division of Reproductive and Urologic Products Director Scott Monroe was clear that the agency intended to be very careful about what study findings would be included in labeling to avoid problems with misinterpretation, if Fablyn is approved.

With breast cancer prevention in particular, he was explicit that FDA would not allow a back-door claim through data in labeling. Such a claim would need to go through a review by the Office of Oncology Products.

If approved, Fablyn (lasofoxifene) would be the second selective estrogen receptor agent approved for osteoporosis treatment, providing direct competition to Evista (raloxifene), the only SERM currently approved for that indication. Wyeth is also seeking approval for a SERM: Viviant (bazedoxifene) received its second "approvable" letter in May (¹ (Also see "Wyeth’s Viviant “Approvable” For Osteoporosis" - Pink Sheet, 23 May, 2008.)).

FDA's Advisory Committee for Reproductive Health Drugs supported Fablyn's use in high-risk women and those unable to tolerate other therapies, though it left the question of the significance of a mortality trend in a clinical trial unanswered (² (Also see "Fablyn Panel Finds Mortality Trend Inconclusive, But Pfizer Ready With A Plan" - Pink Sheet, 15 Sep, 2008.), p. 3).

Since January, Pfizer has been touting its three-year, 8,500-patient Postmenopausal Evaluation And Risk-reduction with Lasofoxifene (PEARL) study, saying it could provide evidence of Fablyn's advantages over Evista (³ (Also see "Pfizer Files Lasofoxifene NDA For Treatment Of Osteoporosis" - Pink Sheet, 15 Jan, 2008.)).

At the advisory committee meeting, Pfizer presented five-year data from an extension of PEARL. In addition to garnering more safety information, the PEARL extension expanded the key endpoints to include risk of estrogen-receptor positive breast cancer, and secondary endpoints such as major coronary events.

The long-term data and additional endpoints give Pfizer points of comparison versus Evista.

Building A Competitive Product

As part of Pfizer's presentation to the advisory committee, the firm had Steven Cummings, a professor at the University of California at San Francisco and chair of the PEARL Scientific Advisory Committee, describe an ideal treatment for postmenopausal women, and then compare the available osteoporosis drugs to that ideal.

The ideal treatment for postmenopausal women would be a drug that not only decreases vertebral and non-vertebral fractures, but also coronary heart disease, stroke, breast cancer, vulvar and vaginal atrophy and hot flashes.

The efficacy for fracture reduction is there for most marketed treatments, including bisphosphonates, calcitonon, hormone therapy, anabolic medication and Evista. But efficacy in other areas is lacking, and the side effect profiles introduce significant disadvantages (or opportunities for competitors).

The highly popular bisphosponate class, which includes Roche/GSK's Boniva (ibandronate), for example, can cause gastrointestinal symptoms, atrial fibrillation, acute phase reactions and rarely, osteonecrosis and femoral shaft fractures.

Committee members and Pfizer representatives cited a trend of an increasing number of women resisting the therapy due to negative media attention, and difficulties with reimbursement (⁴ 'The Pink Sheet,' July 28, 2008, In Brief).

"I have an ever shrinking number of choices to give them. Women do not want to take estrogen; increasingly they do not want to take bisphosphonates. That leaves me with SERMs, and in that category there is only raloxifene," said Steven Goldstein, New York University, another Pfizer speaker.

While Evista decreases vertebral fractures 35 percent to 40 percent, it does not decrease non-vertebral fractures; the three-year results from PEARL show Fablyn reducing the risk of new/worsening vertebral fractures by 42 percent and non-vertebral fractures by 22 percent.

Evista's adverse event profile is similar to that of other SERMs approved for different indications, and includes a two- to three-fold increase in VTEs and a two-fold increase in endometrial polyps and endometrial biopsies. Similar events were seen with Fablyn.

Evista gets a distinct advantage, however, for its 50 percent to 70 percent decrease in invasive breast cancer.

In September 2007, FDA approved Lilly's NDA for Evista for the reduction in risk of invasive breast cancer in postmenopausal women with osteoporosis and for postmenopausal women at high risk for invasive breast cancer (⁵ (Also see "Evista Gains Breast Cancer Risk Reduction Claim But Adds Black Box Warning" - Pink Sheet, 24 Sep, 2007.), p. 16).

Evista's label cites the MORE trial, in which the incidence of breast cancer was assessed as a secondary endpoint. After four years, a 60 mg daily dose of raloxifene reduced the incidence of all breast cancer by 62 percent and invasive breast cancer by 71 percent compared to placebo. Labeling notes that the invasive breast cancer reduction was primarily driven by the 80 percent reduction in estrogen receptor-positive breast cancer.

Evista was then studied in the CORE trial for four additional years, where similar results were found.

Pfizer Ready With PEARL Breast Cancer Data

Both the 0.25 mg and 0.5 mg doses of Fablyn tested in PEARL significantly reduced the risk of ER+ breast cancer (84 percent and 67 percent, respectively) through three years. Through five years, the 0.5 mg dose also significantly reduced the risk by 81 percent compared to placebo; the 48 percent reduction in risk observed with the 0.25 mg dose versus placebo was not statistically significant.

Fablyn also significantly reduced the risk of ER+ invasive breast cancer through three years (82 percent in the 0.25 mg dose, 73 percent in the 0.5 mg dose). The larger dose also significantly reduced the risk by 83 percent through five years, but the risk reduction observed with the lower dose through five years was not significantly different from placebo.

Furthermore, the 0.5 mg dose significantly reduced the overall risk of invasive breast cancer (not just ER+) by 75 percent compared to placebo through three years and by 85 percent through five years.

Pfizer also presented an analysis of stroke from the five-year PEARL dataset, showing a significant reduction in the risk of all stroke events for both doses compared to placebo.

And while the 0.5 mg dose was not associated with a reduction in risk for major coronary events at three years, it did have a significant reduction at five years. After five years, the reduction in risk for major coronary events in the pooled Fablyn treatment groups compared to the placebo was associated with a statistically significant 32 percent.

"Fablyn offers a huge advantage over raloxifene. I think that a take-home message to me as a clinician is that the safety is very much like raloxifene," said Goldstein. "More importantly, this is the first SERM that decreases non-vertebral fracture that is also statically significant."

Breast Cancer Claims? Not So Fast, Says FDA

FDA threw a wrench in Pfizer's plans when it said at the advisory committee meeting that it does not agree with Pfizer's conclusion that Fablyn reduces the risk of developing breast cancer, and also questioned claims about decreasing risks in all the secondary endpoints.

Pfizer argued that the breast cancer endpoint was a primary endpoint for the five-year extension of PEARL with approximately 90 percent power to assume a 70 percent difference. In addition, Pfizer said, all of the amendments to primary endpoints made on the five-year extension were before the three-year results were unblinded.

But FDA held its ground. "They were secondary endpoints," said Monroe. "The study wasn't really designed to look at those, so if they had failed, we just wouldn't be hearing about them today. As we know, if you look at enough endpoints, some are going to win and some aren't going to win. So we've heard a lot about the ones that have won today."

When a confused committee member asked FDA to clarify that breast cancer was a primary endpoint for the five-year study, Monroe said, "I don't think the agency would agree with that interpretation. ... For somebody to get a claim that their product prevents breast cancer or reduces it, there are a number of additional criteria that [Office of Oncology Drug Products] would like to see."

Pfizer admitted they had, at one point, met with the oncology office to assess what it would need to do in order to obtain a formal breast cancer prevention claim, but opted not to go forward with development for that.

"I believe the numbers the sponsor has shown are correct, but it's more complicated than just having correct numbers," Monroe said. "There are a lot of issues in terms of having correct design, the way it was put together ... and I think this is where we get into a lot of gray areas from the folks that presented earlier." The difference between getting an indication and getting data in the labeling "as sort of descriptive" is in "how one interprets it. Sometimes that description gets very gray," he added.

Committee members took the agency's lead. "The study was designed to look at a primary endpoint, which was radiographic vertebral fractures, and that's exactly what I'm going to be basing my opinion on," said voting member Daniel Gillen, University of California, Irvine.

Regarding the other secondary endpoints, he noted that "the study was not powered for the inferences that have been made and has not been adjusted to look at the multiple comparisons with the other endpoints."

Others touched on a possible marketing ploy by adding a description about breast cancer risk reduction. "I'm starting to see the marketing campaign that comes when people start talking about things like cancer, risks and prevention and all those things that we really don't have the data around," said patient representative and temporary voting member Natalie Portis.

With the findings on breast cancer, hot flashes and vaginal changes, "I think we get into murky territory when we throw that into our response because we don't have the information to look at that yet and make a decision," she stated.

Monroe said that the agency is cautious about what it would allow in Fablyn's label, if approved, in terms of descriptive material and data that may lead to misconceptions. "At this point I can't even tell you - if this drug gets approved - what the labeling would say vis-à-vis the findings of this study. Clearly we would not want in the labeling anything that would over-represent what one could interpret."

Regardless of labeling, Pfizer presumably could rely on reprints of a PEARL publication to get the data before prescribers.

- Lauren Smith ([email protected])