Endpoints for Clinical Trials - Case H2Alonzo Garcia, M.D., entering his fellowship in pulmonary medicine, was assigned to carry out an approved study designed by Dr. Bruce Sedgwick, his training director. The study entails the use of the recombinant enzyme, DNAse, to improve pulmonary function and to reduce the incidence of new infections in patients with cystic fibrosis. The study was double-blinded, meaning that neither he nor his patients and their families know whether the inhaler the patient received contained the active enzyme or a placebo. Experiments conducted in vitro and in small animals indicated great success for the agent, though they also pointed to some potential problems. These included development of an allergic reaction to the enzyme or to the other materials in the inhaler, and a direct and adverse chemical effect on the lung passages.
The clinical study for which Dr. Garcia was responsible consisted of a three-month trial of enzyme or placebo, a one-month drug-free period, then a three-month trial of the other arm of the study. Forty patients were to be entered by the completion of the trial. They were to be evaluated by a lung function test and, since most of the patients were children, by a standardized questionnaire completed by a parent.
The experimental drug was so effective that shortly after the trial began, Dr. Garcia found it easy to know who was receiving active enzyme and who was receiving placebo. Even though the questionnaire filled out by the patients' parents was uniform, he discovered himself encouraging respondents to comment about the beneficial effects of the enzyme.
After 20 patients were entered in the trial, one of the parents, who happened to be a scientist, said to Dr. Garcia, "The quality of our daughter's life has greatly improved since she was entered in this protocol. Clearly, the drug is having an enormous impact that cannot be ignored and the blinding must be stopped. Won't you ask the company to terminate the experiment and make DNAse available to us? It will save the lives of our children."
Dr. Garcia, empathizing with the parent, asked Dr. Sedgwick about ending the protocol. Dr. Sedgwick responded, "We are not only testing efficacy here; we are also testing for adverse effects that may be uncommon but quite serious. What would you think about the agent if one percent of recipients had a bout of life-threatening anaphylaxis? Furthermore, we are also going to follow the patients for a reduction of the rate of infection, hospitalization, and need for antibiotics."
Dr. Garcia responded that the statistical power of the study was not sufficient to determine whether an effect would occur in one percent of subjects, nor the duration long enough to obtain meaningful data on infections. For that, a different experiment, monitoring the treatment of large numbers of subjects over a longer period of time would be required. Since the degree of improvement was sufficient to demonstrate efficacy after only 20 subjects, there was no need to continue to employ the placebo controls and to continue with the current study.
Dr. Sedgwick pointed out that the FDA required 40 cases from their institution for the efficacy study and their requirements took priority over the statistical analysis in most cases.
- What values are in conflict in this case? How would you approach their resolution?
- The grossly apparent effectiveness of DNAse in cystic fibrosis seems to have ruined the blinding of subject and investigator that protects against biased reporting of efficacy. What can or should be done about that in the context of this experiment? What about in the broader context of clinical trials?
- The FDA plays a critical role in the design of studies intended to achieve approval of a new therapeutic agent. In fact, companies negotiate in advance with the agency to ensure that, if the study is successful, the agent will be approved. Is this in the best interest of the patient, the company, and society?