Trisha E. O`Hehir, RDH
The recent 1996 Joint Symposium on Clinical Trial Design and Analysis in Periodontics attracted a large group of researchers, clinicians, academicians, and federal government bureaucrats to Washington, D.C. It was an interesting four-day meeting - directed toward university-based research teams and corporate entities with new products they hope to bring to the market.
Despite the focus of the meeting, several questions were raised which may be of interest to clinicians. Which endpoints should researchers be measuring, and are patients` views and opinions being considered in these endpoints? Are control groups ethical? And a new term was introduced: nocebo.
The meeting was hosted on the campus of the National Institute of Health. Employees of the Food and Drug Administration made several presentations. As you might expect, the bureaucratic approach was clearly evident during the discussion of the process necessary for products to receive FDA approval. Every step and office involved in this process is represented by an acronym or a form number! A representative from the ADA discussed a similar process to be followed for products to receive the ADA Seal.
Had the books of rules and regulations been presented to symposium participants, we would have needed little red wagons to carry them around!
Outcomes after the beginning and before the end
True endpoint evaluations are those which directly measure tangible benefits of therapy - such as reduced tooth loss, improved digestion, or increased comfort while chewing - for the patient. However, these endpoints are rarely measured in clinical trials.
Periodontal disease is generally slow moving, and years go by before tooth loss is seen. Retrospective studies looking back several decades at tooth loss among patients in a periodontal practice have actually demonstrated a very slow rate. Studies comparing new treatments to standard treatment would need to run the lifetime of each patient in order to actually measure tooth loss.
In contrast, surrogate endpoints measure disease progression with an assumption that these indices are predictive of tangible benefits for the patient. Examples include bleeding on probing, attachment level changes, bacterial counts, plaque, and calculus. Because of the slow progression of this disease, surrogate outcomes have replaced true outcomes, since they can be evaluated over a period of six to 12 months.
The problem with this approach is the lack of correlation between surrogate endpoints and true endpoints. A desirable outcome for a surrogate endpoint may actually demonstrate an undesirable true outcome as well.
An example of this in medicine was seen during a study of heart medication. Since heart arrhythmias are associated with an increased risk of death, drugs which suppress cardiac arrhythmias were viewed as desirable. The surrogate outcome of suppressed arrhythmias assumed longer lives.
One study documented the effectiveness of an arrhythmia-suppressing drug, confirming the surrogate outcome. The drug, however, was also shown to actually increase mortality. Simply relying on the surrogate endpoint of suppressed arrhythmias led to an estimated 8,000 deaths.
Evaluating tooth loss would involve long and expensive studies, but simply settling for surrogate endpoints to expedite research may lead to incorrect assumptions. In some cases, it may actually be dangerous. It was suggested that, in addition to clinical indices, tangible benefits such as comfort and tooth function be measured for the patient. Such evaluations are subjective, but they may be more reflective of clinical success than periodontal indices.
When evaluating the findings of research, it is essential to know if patients have been removed or OexitedO from the study. It is not unusual for patients who show more than 2 millimeters of bone loss, or who lose a tooth, to be removed from the study in order to receive adequate therapy. Unfortunately, final evaluation of the data does not usually include these people.
Has someone left?
In medical research, if patients are removed from a study because of death, statistics may not accurately reflect the study findings. The same is true in periodontics. If those with active disease are exited from a study, the severity of negative changes may not be presented in the final report.
Concerns were raised about the criteria used for selecting a study group, as well as the ethical dilemma posed by the control group. Many patients with periodontal disease who volunteer for research also use tobacco, have diabetes or another systemic disease, may be pregnant, have lost some teeth, take medication, or have received some type of therapy ? all considered restrictions for study subjects. With so many restrictions, it is difficult to find enough patients with adequate disease.
Conversely, establishing a control group of patients who will not receive therapy raises ethical questions. Since we know surgical as well as non-surgical therapy can stop the progress of periodontal disease, can we ethically stand back and watch disease progression in the control group? Some feel strongly that control groups are not needed, while others, committed to the scientific approach, support the need for control groups.
My thought was: What about the 40 to 50 percent of the population who choose no therapy by never visiting a dentist or dental hygienist? They would be the perfect control group (except for the fact that they are never seen by a clinician to be monitored).
Narrow focus means no broad outlook
Extrapolating research results to the general population also poses a problem when the study group has been so narrowly restricted. Age, race, and gender must be considered when drugs are tested. This is due to the variability of drug metabolism between adults and children, men and women, and between various races. Since drugs are not tested on children, pregnant women, or lactating women, no generalizations about drug effectiveness can be made to include groups which have not been tested.
The issue of clinical significance was raised by Drs. Steve Garrett and Erwin Barrington, who both voiced the need for guidelines in this area. At this time, clinical significance seems to be an arbitrary decision by those conducting the research. Despite past discussions which have not resulted in consensus, the need for guidelines is still an important issue.
In addition to the questions raised at the symposium, a new term, nocebo, was introduced by Dr. Michael Newman. It?s the opposite of a placebo. When patients receive a placebo, they think it is the medication being tested. Significant improvement is often measured due to psychological reasons. If, on the other hand, patients know they are not getting the test treatment, they actually get worse ? again due to psychological reasons. In this day and age of informed consent, it is sometimes difficult to keep patients in the dark about test or control group assignments, making research that much more difficult.
It was an interesting meeting. But it reminded me how wide the gap is between research and day-to-day clinical practice. Clinicians are more concerned about long-term results than a half of a millimeter improvement for one therapy or drug over another during a 12-month period. Clinicians deal with real people, while many researchers focus entirely on the numbers.
Let?s hope a few bridges can be built in the future that span the gap between research and real life. Each side has a great deal to offer the other.
Trisha E. O`Hehir, RDH, is a senior consulting editor of RDH. She also is editor of Perio Reports, a newsletter for dental professionals that addresses periodontics.