| |
|
| |
|
More on Columbia’s
ALS Symposium TARRYTOWN, N.Y., June 15, 2003 – The second part of MDA’s ALS Clinical Trials: The Challenge of the Next Century conference was devoted to understanding and improving the conduct of ALS trials. Sessions included presentations on markers of disease progression; determination of outcome measure (end points) in trials; statistical methodology; the merits of various trial designs, including placebo group design and other approaches to using untreated comparison groups; large and small research collaborations; and calls to action. Defining the ‘Natural History’ of ALS Nigel Leigh of King’s College Hospital in London posed the question of whether the natural progression of ALS is changing. Determining this is important, because many trials of new drugs or other therapies seek to compare the treated group with the untreated disease -- its natural history. When Leigh’s group compared ALS patients from before 1996 to those after 1996, he found there was a 14 percent increase in the probability of surviving for five years. Leigh attributed the increase in five-year survival to better general care in ALS, including more widespread use of noninvasive ventilation and more attention to nutrition with insertion of feeding tubes when needed, as well as to widespread use of the drug riluzole. He cautioned against relying on older natural history data as a baseline for comparison of data from a treated group in a trial, since the experimental treatment might in fact not be better than today’s standard care. Ben Brooks, who directs the MDA/ALS Center at the University of Wisconsin-Madison, says his studies suggest that riluzole may have more of an effect on ALS early in the progress of the disease than it does later and that much of the survival improvement since 1996 can be attributed to riluzole use. He also noted that there may be factors in ALS that determine the way patients respond to treatments, including riluzole and any experimental compounds being considered, and that it’s important to consider these factors if they can be identified. For instance, he said, age; site of onset of symptoms (whether in the bulbar muscles, which control the mouth and throat, or in the limb muscles), and the stage at which a treatment is started could have a large impact on the trial results. Edward Kasarskis of the University of Kentucky in Lexington noted that use of feeding (gastrostomy) tubes and positive-pressure ventilation are “uncontrolled variables” in ALS trials. He said that investigators have assumed that these factors would “wash out” because they would by chance be approximately equally used by those receiving the treatment and those in the untreated (control) groups, but that he was not entirely convinced of the truth of this assumption. For instance, he noted, if an experimental drug were to increase or decrease weight or appetite, it might influence a patient’s decision about using a feeding tube and thus blur the effects of the experimental medication alone. Finding Trial Endpoints Jesse Cedarbaum of Regeneron Pharmaceuticals explored the question of which outcomes to look for in clinical trials. These outcomes, known as “end points,” have traditionally been length of survival in the treated and untreated groups, muscle strength differences in the compared groups, pulmonary function differences, or quality-of-life differences. Cedarbaum suggested that other measures might be more indicative of a treatment’s success or failure. For one thing, he said, it’s becoming more difficult to define “survival” in ALS now that ventilation can prolong life to an indefinite extent. The difference in physical status between a person who chooses to use ventilation and one who doesn’t may be minimal, or there may be none at all, but the person who chooses ventilation will likely survive longer, thus obscuring the measurable effect of any tested treatment. Cedarbaum said that the timing of disease progression, the rate of functional decline, or quality-of-life measures might be good alternatives to survival as outcome measures, or end points, in an ALS trial. Biological Markers of Disease Progression Research in certain other diseases, such as cancer and AIDS, is guided by reliable biological markers of disease progression -- identifiable signs that the disease is worsening, improving or staying about the same. Cancer trials, for example, have long been guided by tumor size and extent of spread (staging) and by blood levels of various molecules, such as prostate-specific antigen, or cell counts. In ALS, such biological indicators have been scarce but some exist, the experts noted, and can be used to answer certain types of questions that arise in clinical research. Finding new and better markers, they suggested, should be a priority in current ALS investigations. The lecturers in this section included M. Flint Beal of Weill Medical College of Cornell University in New York; Clifton Gooch of Columbia University in New York; Jeremy Shefner, director of the MDA/ALS Center at Upstate Medical University in Syracuse, N.Y.; Michael Swash of Royal London Hospital in England; and Douglas Arnold of McGill University in Montreal. A “motor unit” is a muscle-controlling nerve cell in the spinal cord with the fibers that connect it to muscle cells. Counting these motor units, a technique called MUNE, for “motor unit number estimation,” can be used to measure their loss during the course of ALS. The technique requires electrophysiological studies and the ability to interpret the complex signals that emerge from these. If done correctly, MUNE can provide an accurate biological marker of at least some aspects of disease progression. Another marker that relies on electrophysiology measurements is the “neurophysiological index,” which is the result of a mathematical calculation that combines the strength and timing of electrical signals from muscles and nerves. Magnetic resonance spectroscopy (MRS), which is noninvasive and relatively comfortable for patients, can determine the status of nerve cells in the brain by measuring the density of two chemicals, NAA and creatine, and their relationship to each other. Among the future targets, according to Beal at Cornell, are 20 recently identified compounds in the blood plasma that investigators believe may be correlated with ALS progression. Designing Studies Testing a new drug or other therapy may sound easy, but designing a trial that can reliably tell the difference between disease progression in people who got the treatment and those who didn’t and provide assurance that results reflect only that difference is extremely challenging. Robert Miller, director of the MDA/ALS Center at California Pacific Medical Center in San Francisco; Yuko Palesch of the Medical University of South Carolina; Dan Moore of the University of California at San Francisco; and Stanley Appel, director of the MDA/ALS Center at Baylor College of Medicine in Houston, each gave lectures that stimulated much discussion of trial design strategies. Innovative designs included some in which each participant receives a placebo (inert substance) and then the study drug (without knowing when the switch occurs), and in which the investigators compare progress during the placebo phase and treatment phase during later analysis. Every participant has a chance to take the study drug in this design, and comparisons are made between the placebo and the treatment phase for each person, not between groups of people. One design, presented by Yuko Palesch, has been designed to screen out ineffective treatments quickly and avoid unnecessary exposure to these substances or wasting of time and resources. Informally known as a “futility study,” this design compares those taking the drug to statistics on untreated patients derived from databases. The investigators conduct a short trial to see whether the drug warrants further investigation. If the drug is not eliminated at this stage, Palesch cautioned, it’s essential that patients and doctors not interpret this to mean that the drug is in fact effective. Futility trials are designed only to eliminate poorly performing drugs and not to measure positive effects, he said.
Getting Together The third day of the conference was devoted to a discussion of collaborative research groups. The main ALS research groups in the United States are the Western ALS (WALS) Group, the New England ALS (NEALS) Group, and the Great Lakes ALS (GLALS) Group. The audience also heard about the Canadian ALS group and the European ALS Consortium. The purposes of these groups are to accumulate enough participants to achieve a statistically significant result in a clinical trial; share data and sometimes DNA or tissue samples, as well as ideas; and present convincing trial designs to funding agencies. Much of the discussion of the third day centered around when and whether all U.S. or perhaps all North American ALS groups should create one large organization that would theoretically be able to realize all those purposes most efficiently. The downside of such an endeavor was also mentioned, with some audience members and speakers expressing concern about the loss of autonomy of the regional groups within the North American group and perhaps less tolerance of junior members or innovative ideas should the new organization become too large and bureaucratic. The group voted to set up a nominating committee that will in turn propose a steering committee to develop a national or North American ALS research group Taking Action Lewis P. Rowland, founder and former director of the MDA/ALS Center at Columbia University and now a professor of neurology at that institution, closed the conference by voicing several calls to action for the conferees. These were to:
|
Stanley
Appel discussed the use of “historical controls” instead of
placebo controls as untreated comparison groups in trials. Historical
controls are extracted from databases that have recorded the progress
of untreated patients. He said that the disease progression scores taken
from a database of untreated patients at Baylor look the same as those
of patients who received a placebo in several drug trials, leading him
to conclude that historical controls are a valid tool as a placebo group
substitute. Appel’s group used the Appel ALS Scale to take disease
progression measurements for the database. He noted that survival increased
from an average of 30 months in the 1980s to an average of about 40 months
between 1997 and 2001. He said the earlier data could be problematic if
used as a control group now. Historical controls lessen the number of
patients required to conduct a trial and simplify trial execution.
