SORIN BIOMEDICA MONOSTRUT HEART VALVE NEEDS AUGMENTED

SORIN BIOMEDICA MONOSTRUT HEART VALVE NEEDS DATA FROM ADDITIONAL PATIENTS before it can be considered for approval, FDA's circulatory system devices panel concluded at an April 13 meeting in Washington, D.C. The panel voted three-to-two, with one abstention, that Sorin should gather data from additional patients under existing study protocols and reanalyze data in the Monostrut premarket approval application. The recommendation followed a unanimous six-to-zero vote that the Monostrut is not approvable. Panel chair Gabriel Gregoratos, MD, University of California at Davis, stated: "In terms of evidence" of the safety and efficacy of the valve, "there is very little. We are being asked to take this pretty much as a matter of faith" based on investigators' assertions. While the panel recommended that additional data be gathered, FDA went a step further and suggested that the company consider conducting a new study. Abhijit Acharya, director of the division of cardiovascular, respiratory and neurological devices in the device center's office of device evaluation, stated that Sorin would "probably be better off" conducting a new study because "what you gain" in data "outweighs the bother of getting an" investigational device exemption. However, Jeffrey Brinker, MD, Johns Hopkins Hospital, Baltimore, argued "it would be better for the company to work with FDA under a new IDE using [existing] data as a background." Brinker said he was worried that, if panel membership changed before the Monostrut was again reviewed, new panel members may have "qualms" with data gathered in Canada in the original studies. Diane MacCulloch, a staffer in ODE's cardiovascular division who reviewed the Monostrut PMA, estimated that the company could gather the additional data requested by the panel in as little as several months. On the other hand, panel member Ronald Weintraub, MD, Beth Israel Hospital, Boston, said that conducting a new study of the Monostrut could take three or four years. Richard DeRisio, vice president for quality and regulatory affairs at Sorin, pledged at the meeting to "work closely with FDA [and] make sure we clearly understand the requirements." Miles Warren, the company's manager of cardiovascular clinical and regulatory affairs, added that Sorin is "willing to make whatever strides are necessary to get this valve approved." Sorin acquired the Monostrut, a single-strut, single-leaflet mechanical valve, when it bought the assets of Pfizer's Shiley subsidiary in December 1991 ("The Gray Sheet" Dec. 23, 1991, p. 10). With the Monostrut, Sorin inherited 10 years of Shiley clinical data and a PMA application that has been pending at FDA since 1985; the application has been supplemented with additional data requested by FDA several times since its submission. Hans Heysmann, MD, University Hospital of Leiden, The Netherlands, a principle Monostrut investigator, presented data on 1,107 patients from three separate cohorts. Of the total patient population, 565 had valves placed in the aortic position, 425 had valves placed in the mitral position, 75 had both valves replaced, and 42 had "other" types of valve replacements. The data includes results from a five-center study outside the U.S., retrospective data from Europe and a prospective study from Canada. Aortic valve patients demonstrated a 75% survival rate after an average followup of nine years, and mitral valve recipients had 70% survival after the same period of time, Heysmann stated. Analysis of valve performance in patients with aortic valve implants found that after a followup of six years, the Monostrut improved the condition of 69.7% of patients, while the condition was unchanged in 10.1% of patients and worsened in 1.6%. The remaining percentage of patients either died or had their valves explanted. For mitral valve replacements, the condition was improved in 66.9% of the patients, unchanged in 8.7%, and worse in 1.2%. The panel based its recommendations for a PMA revamp on concerns over the adequacy, interpretation and presentation of data in the application. Panel members felt that because of disparate patient populations, it was inappropriate to pool data from the three studies. The panel recommended that Sorin base the Monostrut PMA on the Canadian study and use the other data as supplemental information. Panel member Julie Swain, MD, University of Nevada School of Medicine, Las Vegas, stated: "The problem I have is we have three studies...inclusion protocols are different, they are not concurrent." She concluded: "I do not feel scientifically whatsoever that you can group this data. . . There are three separate stand-alone studies." In addition, the panel felt that the Canadian study should have been conducted in accordance with 1986 FDA guidance documents on prosthetic heart valve PMAs. For example, the study did not meet the guidance documents' recommendation that data be gathered on at least 35 aortic and 35 mitral valves at each participating center and that cardiac catheter or echodoppler followup be conducted on at least seven patients with each valve size. The panel suggested that the revised PMA submission meet the guidance requirements. Swain also expressed concern that patient deaths in the study "really are not classified in an appropriate manner" as either valve-related or non-valve-related. The panel concluded that the deaths should be reviewed "by a panel of surgeons...to reclassify them." The group also recommended that the study results be compared to historical controls. The panel commented that many of its concerns would have been obviated if randomized trials had been conducted on the valve. Michael Domanski, MD, National Heart, Lung and Blood Institute, Bethesda, Maryland, asked Sorin if "there is a fundamental problem or design reason why you would have been uncomfortable with that kind of an approach?" Warren responded that some surgeons feel that "randomizing patients would go against their ethical standards of treatment for patients." He also stated that very large patient cohorts would have to be used to gain statistically significant comparative data. The Health Industry Manufacturers Association highlighted these types of problems in recent comments on an FDA proposal to require randomized heart valve trials ("The Gray Sheet" March 15, p. 4). Panel member George Kay, MD, University of Alabama, Birmingham, commented: "Although I think its safe to say we would all prefer to see a randomized trial, these [Monostrut] studies have been going for a long time and it is unfair to hold [Sorin] to a randomized trial" standard. However, Swain added that "if I were to ask [Sorin] to do a valve study right now, I would insist that there be randomized controls." FDA's Acharya, noting that many cardiothoracic surgeons are opposed to FDA's draft heart valve guidance for "ethical and statistical reasons," said the agency is evaluating "whether there are alternatives to randomization [in heart valve studies] in certain cases." Acharya added that FDA plans to sponsor a conference on clinical trial design in June, and that heart valves will be among the device types discussed at the meeting.