This case study describes the use of phenylpropanolamine-containing medications in the last several decades, both in prescribed and over-the-counter formulations. The initial case reports and preliminary epidemiologic results indicating risk of stroke in PPA users led to a multi-center epidemiologic study sponsored by the Consumer Health Products Association and the Food and Drug Administration in the mid-1990s. During this same time period, a series of lawsuits was filed in various states and these were eventually consolidated into a Multi-District Litigation (MDL) managed by Judge Barbara Rothstein, in Seattle, WA. The epidemiologic study was completed and published in the New England Journal of Medicine in 2000.
This study became a key point of contention in the MDL, with defendants claiming the study was improperly done, provided no evidence of increased risk of stroke in males, had few statistically significant findings, among other issues raised in discovery depositions and Court hearings. Judge Rothstein reviewed the arguments and ruled in 2003 that the epidemiologic study was valid evidence to put before juries and that the findings were relevant to both females and males who were exposed to PPA and suffered strokes. Similarly, a three-judge panel in Texas ruled that the epidemiologic study should be allowed in as evidence in Texas cases. Thus, two attempts to keep peer-reviewed epidemiologic evidence from being allowed in PPA lawsuits have failed, and the Courts have exercised proper judicial review. Finally, a Korean study was completed in 2004 which had similar findings to the U.S. multi-center study and is being prepared for publication.
This case study on phenylpropanolamine (PPA) and the multi-district litigation that ensued after it was associated with multiple cases of stroke is an example of how epidemiologic evidence was presented in a court case, how the evidence was challenged and supported by various parties, and how the Court decided to handle one particularly relevant epidemiologic study. The Court decision was after a Daubert hearing in Seattle, Washington where plaintiffs and defendants provided oral and written testimony before Judge Barbara Rothstein and several other judges in televised proceedings. Judge Rothstein's decision to allow the "Yale study" into evidence was important because it reflected a good understanding of the strengths and limitations of epidemiologic data. What follows is a chronologic summary of the main developments in this case, with references and links to key documents.
Initial Case Reports
Phenylpropanolamine (PPA) has been used for decades in various medications, including diet pills (Dexatrim, and others), and cough and cold medications (Ornade, Dimetapp, and others) that were sold both with and without prescriptions. Beginning in the 1970s, case reports began to appear in the medical literature documenting hemorrhagic strokes in people who had taken various medications containing PPA. Because of this concern, an initial epidemiologic study was published in 1984, looking at the incidence of stroke in persons who had been prescribed PPA-containing medications in a large group health plan. This study, which included only one case of stroke, was inconclusive (Jick et al 1984).
In the 1980s, additional case reports of stroke in PPA users appeared in the medical literature, and spontaneous reports accumulated in a voluntary reporting system maintained by the Food and Drug Administration (an FDA summary report was completed in 1991). Many of the case reports were in young women between the ages of 18 and 49 who had taken diet pills, and some case reports described stroke patients who had taken PPA in combination with drugs of abuse. Additional studies of the effects of PPA on human volunteers, plus studies of pharmacologic effects in laboratory animals accumulated throughout the decade. By the 1990s, enough information had accumulated to persuade the Consumer Health Products Association (CHPA) and the Food and Drug Administration (FDA) to sponsor an independent study.
Yale Multi-Center Case-Control Study
A group of investigators at Yale University Medical Center proposed a multi-center case-control study to evaluate the association between PPA and hemorrhagic stroke, in collaboration with colleagues in 43 hospitals in various parts of the U.S. The study was approved by an independent scientific committee, with FDA consultation, and the data collection proceeded over a period of four years. The final results were reviewed by external advisors, consultants to the CHPA, and scientists at the FDA and approved for publication. A manuscript summarizing the main results was published in the New England Journal of Medicine in 2000 (Kernan et al 2000). The results indicated a significant increased risk of hemorrhagic stroke in persons who had taken PPA-containing appetite suppressants and cough and cold medications, especially in women, although there was some indication of increased risk in men, as well. The risk of stroke for women who used PPA-containing appetite suppressants was greater than 16-fold compared to controls, and the risk for men and women after first use of PPA-containing medications was greater than 3-fold compared to their matched controls.
Lawsuits Filed by Stroke Victims; Daubert Hearing and Decision
After the submission of the results of the study as a final report to the FDA, a warning was issued limiting sales of PPA-containing medications. In the next two years, a series of lawsuits were filed by stroke victims in different states. One of the earliest such lawsuits was in Maine (referred to as the Newenham case) on behalf of a teenaged girl who had taken over the counter cold medicine containing PPA and subsequently had a stroke. The large number of cases filed led to the formation of a multi-district litigation approach, consolidated in the United States District Court Western District of Washington at Seattle (In Re Propanolamine (PPA) Product Liability Litigation - MDL 1407). As part of the pre-trial discovery, attorneys for defendant pharmaceutical companies deposed the Yale investigators, including Drs. Horwitz, Brass, Kernan and Viscoli at length. The defense attorneys also deposed Dr. Richard Clapp, an epidemiologist who had be retained by attorneys for the plaintiff in the Newenham case, about the validity and reliability of the “Yale study” published in the New England Journal of Medicine.
Much of the dispute by defense attorneys in the pre-trial discovery was over the details of case definitions, data collection procedures, changes in protocol once the study was underway, and subsequently over the magnitude and statistical significance of the odds ratios. The Yale investigators explained their methods and the fact that all proposed alterations were approved by an independent advisory committee. Furthermore, the results were peer-reviewed by independent experts identified by the New England Journal of Medicine editorial staff and modified as requested in the final published article. One requirement that the journal routinely makes, which the Yale investigators complied with, is that all significance tests should be “two-sided,” which is a more conservative approach than the authors had used in their final study report. Dr. Clapp reviewed and endorsed the study as a valid scientific result that met the requirements for consideration by the Court.
Defense experts all had various criticisms of the study, but none could point to fatal flaws. As a final challenge, the defense attorneys maintained that the Yale study only supported an increased risk of stroke in patients between the ages of 18 and 49, the age group that comprised all the cases, and only in females, although there were results for males in the study. The Court, after a three-day Daubert hearing in Seattle in June, 2003, ruled that the results could apply to males as well as females and to patients beyond the 18-49 range (Rothstein ruling 2003). The Court also agreed that the Yale study was reliable and should be allowed as evidence in the multi-district litigation.
Additional Judicial Review of Yale Study
Another judicial review of the Yale study was held in Texas before a three-judge panel in the Bexar County Court. This hearing was held because of the requirements of the Havner decision, which essentially reprised the Daubert decision for Texas State Courts. Following a complaint filed by plaintiffs, the Havner hearing was held in San Antonio in January, 2004. Some of the same experts who had testified in the Multi-District Litigation, including defendants' experts and plaintiffs' expert Dr. Clapp, testified at the hearing on the validity of the Yale study. Arguments against the Yale study included questions about individual cases, lack of control for confounding by use of illegal drugs, and lack of statistical significance of some of the findings. Plaintiffs rebutted these arguments and explained how statistical significance was calculated, and noted that some odds ratios were greater than 2.0, even though that was not an epidemiologic criterion for drawing a causal inference.
The three-judge panel denied the defendants’ petition to disallow the Yale study. The Texas cases are proceeding or being settled. In both the Multi-District Litigation and the Havner hearing, the Courts viewed the Yale study as relevant and reliable scientific evidence that should be allowed in to the proceedings. Efforts by defendants to disparage the Yale study, the choices made by the investigators themselves, and the interpretation of the results by plaintiffs’ experts ultimately failed. Since the MDL and Havner decisions, a Korean study has been completed which replicates the essential findings of the Yale study. The Korean study (Yoon et al 2004) was sponsored by the Korean Pharmaceutical Manufacturers and the Korean FDA and followed a protocol quite similar to the Yale study and likewise found an increased risk of stroke among patients who had taken PPA-containing medications. This study is not yet published in a scientific journal but has been reviewed by experts for plaintiffs and defendants in the MDL and will be offered as further evidence in future proceedings.
Links and Resources
Stier BG, Hennekens CH. Phenylpropanolamine and hemorrhagic stroke in the Hemorrhagic Stroke Project: a reappraisal in the context of science, the Food and Drug Administration, and the law. Ann Epidemiol. 2006 Jan;16(1):49-52.
Clapp R, Williams ML. (Letter) Regarding "phenylpropanolamine and hemorrhagic stroke in the hemorrhagic stroke project". Ann Epidemiol. 2006 Jul;16(7):580-2.
Michaels D. (Letter) Regarding "phenylpropanolamine and hemorrhagic stroke in the hemorrhagic stroke project": mercenary epidemiology-data reanalysis and reinterpretation for sponsors with financial interest in the outcome. Ann Epidemiol. 2006 Jul;16(7):583-5.
Stier BG, Hennekens CH. Response to letters by Clapp and Williams and by Michaels. Ann Epidemiol. 2006 Jul;16(7):586.
Los Angeles Times report on PPA (3/28/04).
About the Author
This case study was written by Richard Clapp, DSc, MPH. Dr. Clapp served as a plaintiffs' expert witness on methodological issues in the PPA multidistrict litigation. He also served as a rebuttal witness in the Havner hearing in Texas.