NDMA's PPA STUDY: "CO-PRIMARY" TARGET GROUP WILL BE WOMEN AGED 18 TO 49

NDMA's PPA STUDY: "CO-PRIMARY" TARGET GROUP WILL BE WOMEN AGED 18 TO 49 who take phenylpropanolamine in response to a request from FDA at an Aug. 25 "OTC feedback" meeting. The protocol for the case-control study, called the "Yale Hemorrhagic Stroke Project," was initially submitted by the Nonprescription Drug Manufacturers Association in March to determine whether OTC phenylpropanolamine use is associated with an increased risk of hemorrhagic stroke ("The Tan Sheet, March 15, p. 8). FDAer David Graham noted that FDA's interest in studying phenylpropanolamine in the first place was aroused because of reports of strokes in women taking diet aids containing PPA. The Yale investigators, who will conduct the study on behalf of NDMA, argued against redefining the primary study group to focus on women aged 18-49. The protocol, as initially proposed, was designed to look at all phenylpropanolamine users between the ages of 18 and 54, including users of OTC cough/cold products with PPA, to determine whether they experience an increased risk of hemorrhagic stroke compared to non-users. The reviewers' "secondary aims" were: determining whether an association exists between hemorrhagic stroke and PPA in either appetite suppressants or cough/cold remedies, and identifying whether there are other risk factors such as family history, oral contraceptive use and alcohol consumption. "We've looked at these [PPA] studies too and we find that when you look at hemorrhagic stroke indications and you look at cough/cold and diet pills, the gender effect is much less prominent than what you're observing for appetite suppressants alone," said project co-investigator Ralph Horwitz, MD, Yale University School of Medicine. "It depends on how you look at the data." However, FDA Office of Drug Evaluation I Director Robert Temple suggested that "if the real problem [of hemorrhagic stroke] is women, you want to be able to focus on that because focusing on all people just dilutes the outcome." Horwitz said the investigators would be willing to consider women aged 18-49 as a secondary aim of the study. Office of OTC Drug Evaluation Director Michael Weintraub, MD responded: "As long as the study is powered enough to answer these questions, we don't care if its primary, secondary, tertiary . . . as long as you get the answer." The investigators agreed to address what were formerly called primary and secondary aims as "co-primary aims." The NDMA case-control study will include cases of hemorrhagic stroke from 20 Connecticut hospitals via an active surveillance program. Diagnosis will be confirmed by a physician blinded to PPA use. The control portion of the study will seek to enroll two subjects per case "chosen as a representative sample of the population by random digit-dialing" and matched to cases by age, gender, race and telephone exchange. Another issue of disagreement at the feedback meeting was whether the three-day exposure window the study proposes is appropriate. FDA proposed that as a primary endpoint, a patient would be considered exposed if the onset of symptoms that leads to hemorrhagic stroke occurs within 24 hours of the first dose. Although how PPA may be associated with strokes is not well understood, Graham noted that the agency suspects that "there may be a first dose effect." Horwitz said he did not "understand the scientific rationale for counting as unexposed [those] people who took [a dose] of PPA four hours before a stroke . . . because their first use of PPA was two days ago." Moreover, he said, "I'm prepared to accept that we don't understand [what is causing hemorrhagic stroke] but not understanding the mechanism does not justify an exposure definition that excludes common use." Responding to a remark from co-investigator Lawrence Brass, MD, Yale University School of Medicine, that "most of the bleeding events [that lead to hemorrhagic stroke] appear within the first 24-48 hours," Temple acknowledged that "if there is a fair fraction of those that have absolutely no symptoms in 24 hours of taking [PPA] . . . that is a reason for using a two-day window but no more." However, both sides agreed to postpone further discussion until a later date. The Yale investigators also agreed to use surrogates for both cases and controls. In addition, NDMA and FDA agreed that an interim analysis of the study should be conducted in the event that it reveals a "devastating" effect. The other area of disagreement included sample size calculations. To calculate the sample size, the Yale group said it will "delineate" the definition of phenylpropanolamine use, define an exposure window and use the data from a survey for the use of diet pills and cough/cold remedies to estimate the proportion of the eligible population that used PPA-containing products during that defined exposure window. Because the sample size calculations were based on assumptions that have been redefined since the protocol was submitted, the project investigators agreed to re-evaluate their calculations and resubmit them to Weintraub. NDMA Senior VP William Soller, PhD, said the group expects to "run the numbers" and submit them to the OTC office "in a few weeks." In a June 17 letter to NDMA, FDA outlined its concerns with the PPA study design. The agency questioned whether the exposure window selected by the investigators is too wide, whether the sample size is too small and whether surrogate interviews should be allowed. NDMA responded to FDA's concerns in an Aug. 11 letter to the agency in anticipation of the feedback meeting. In the letter, NDMA agreed to revise the number of subjects originally proposed. To address primary aims, 330 cases and 660 controls were suggested. Noting that there is "a keen interest in several secondary aims," Horwitz said the investigators "decided to augment the case sample size by 50%" to 500 cases and 1,000 controls. Additional sites could be added to accommodate the expanded case base, NDMA explained. The Yale group indicated that it is in negotiations with hospitals in Worcester, Mass. and Providence, R.I. to generate further hemorrhagic stroke cases.