Placebo Effect Could Skew Tests Of New Therapies

In a unique study, investigators at Duke Clinical Research Institute Durham, N.C., looked at whether a patient's assessment of his or her own health status -- a frequent "endpoint" in clinical trials of novel therapies -- bore any relationship to the actual clinical outcome they had, based on the treatment they received.

In a study of 1,189 patients with severe coronary artery disease, researchers found that the treatment patients received -- aggressive or not -- had no impact on their sense of well-being. All patients in the study rated their health status as similarly improved since coming under a doctor's care.

But such a bright outlook was not always reflected in mortality figures. Death rates for patients for whom only drug therapy was available were twice as high compared to those who received heart bypass surgery or an angioplasty procedure to clear clogged heart passages.

The study was a surprise to researchers, who say it points out a prominent "placebo effect" that can be slanting the results of clinical trials that judge success based on quality-of-life indicators.

"Even the sickest patients in this study, those for whom revascularization wasn't an option, felt better during the time they were being followed by physicians," said Duke cardiologist Dr. James Jollis, who led the study. "When you take care of patients over time and follow their progress, the placebo effect can be strong."

He prepared his study for presentation Aug. 28 at the 22nd annual congress of the European Society of Cardiology. The study was funded by Genentech Inc., of South San Francisco, Calif.

Jollis noted that trials of several new techniques, such as myocardial laser perforation and the use of vascular growth factors, have relied on general health status instead of "harder endpoints such as death or costs."

For example, recent tests of vascular growth factors have been negative because patients in both the test and control groups both have shown improvement. The unexpected improvement among placebo patients could lead to falsely negative findings, Jollis said. "Based on the Duke analyses, such trials may have been positive if more patients had been included," he added.

"There are many more things that affect the way patients feel other than just the status of their hearts," he said. "Variables can range from non-cardiac illnesses to non-medical issues, such as social, psychological and economic factors.

"Our data suggest that qualify-of-life measures vary little, such that many patients would be required to detect a treatment benefit, while mortality and cost vary greatly, such that fewer patients would be required to detect a benefit," Jollis said.

This is important because the number of patients with "end-stage" coronary artery disease -- those for whom bypass surgery or angioplasty is no longer an option -- "continues to grow" and, consequently, more therapies are being designed to treat them, Jollis said. "It's important that these new therapies be tested and measured in such a way that it demonstrates their effectiveness."

The Duke study specifically picked a group of patients with severe coronary disease to study because they expected that patients who received revascularization procedures (bypass surgery or angioplasty) would report a higher quality of life, compared to patients who could only be treated with medicine. Optimal treatment for such disease is revascularization, but in this study, 487 patients (41 percent) could not undergo the therapy due to their coronary anatomy or excess risk of death.

Patients were asked to complete two questionnaires after treatment and then at one and two years of follow-up study. One questionnaire consisted of a 36-item health survey on such topics as general health perceptions, physical functioning, bodily pain and general mental health. The other asked patients to rank the ease with which they are able to perform common activities of daily living.

Results show that patients who only received drug therapy had a similar pattern of positive change in their self-assessment of physical and emotional limitations. Functional status also improved for all patients, independent of their treatment.

But when data on death rates and cost of hospitalization were tallied on the same group of 1,189 patients, it was clear that those who had not qualified for a revascularization procedure did much worse, Jollis said.

After two years, mortality in patients who were treated with only medicine was 38 percent, compared to 15 percent in patients who had an angioplasty and 19 percent for bypass surgery patients. During that follow-up period, hospital costs of $40,471 was also substantially higher in the group that didn't have revascularization, compared to $34,785 for angioplasty patients and $24,005 for bypass patients.

"Remarkably, despite greater disease burden and mortality, patients with end-stage coronary artery disease do not perceive their health status to change in a significantly different fashion compared to patients still eligible for conventional revascularization," Jollis said. "These findings should inform the design of trials involving novel therapies, suggesting that mortality and cost be included as primary endpoints when considering effectiveness."