Health Letter, November 2018
By Azza AbuDagga, M.H.A., Ph.D.
Although disagreements remain, there is a general consensus that widely used treatments proven to be beneficial should not be withheld from human subjects in clinical trials when doing so would expose these subjects to an increased risk of death or irreversible morbidity.
A recent study by Public Citizen’s Health Research Group showed that use of placebos only, as opposed to existing beneficial treatments, in control groups was common in clinical trials supporting U.S. Food and Drug Administration (FDA) approval of new medications for nine life-threatening diseases from January 1, 2006, to December 31, 2011. The study cautioned that this practice potentially puts human subjects at increased risk.
Our study was published online on April 24, 2018, in Open Access Journal of Clinical Trials.
First, we identified all original new drug applications (NDAs) and biologic license applications (BLAs) approved by the FDA from January 1, 2006, and December 31, 2011. We obtained these applications from the Drugs@FDA database, a publicly available online source that comprises information on all drugs and some biological products approved by the FDA.
Second, we narrowed down the applications to those that pertain to nine life-threatening diseases that had the largest number of approvals during our study period: diabetes mellitus (type 1 or 2), hypertension, asthma, schizophrenia, chronic obstructive pulmonary disease (COPD), partial seizures, osteoporosis, Parkinson’s disease and bipolar disorder. We defined a life-threatening disease as a disease for which long-term mortality risk is elevated, including risk attributable to secondary complications.
Third, we identified all clinical trials that were discussed in the FDA medical review documents associated with the NDAs and BLAs of the nine life-threatening diseases that we selected. These clinical trials were conducted by drug manufacturers to evaluate the safety and effectiveness of their products. We also searched additional sources, such as clinicaltrials.gov, the medical literature and other public sources to obtain key information about these trials if such information was not included in relevant FDA clinical review documents.
For each trial, we collected information that helped us to determine the type of interventions provided to control group subjects. Based on the type of interventions provided to control group subjects at the point of randomization, we classified each trial into one of two mutually exclusive categories: (1) active-controlled and (2) placebo-only-controlled. We considered active interventions to include pharmacological as well as dietary or lifestyle interventions (such as vitamin D and calcium supplements for osteoporosis and diet and exercise counseling for diabetes). We classified trials as active-controlled only if all subjects in all control groups received an active intervention.
We also collected information about certain trial characteristics, such as the duration of the intervention phase of these trials (trial duration) and, where available, the size, start year, and country location(s) of the trials and the disease severity among the enrolled subjects.
Findings and discussion
A total of 508 clinical trials supported the FDA approval of new medications for the nine life-threatening diseases with the largest number of approvals during our study period. Of those, 201 trials (40 percent) were placebo-only-controlled — meaning that subjects in at least one control group in each trial did not receive an active intervention. This practice potentially placed over 19,000 human subjects, who were randomly provided with placebos only, at increased risk of harm.
The use of placebo-only control groups varied significantly by disease type. The proportions of trials that included these groups were 76 percent for COPD, 64 percent for schizophrenia, 50 percent for bipolar disorder, 47 percent for asthma, 43 percent for hypertension, 16 percent for diabetes, 8 percent for Parkinson’s disease and zero percent for partial seizures and osteoporosis.
Many trials enrolling the most severely ill human subjects were placebo-only-controlled: We identified placebo-only-controlled designs for 77 percent of trials for severe COPD, 56 percent of trials for severe hypertension, 44 percent of trials for moderate or severe persistent asthma and 7 percent of trials for severely uncontrolled diabetes.
Trials of long duration were significantly less likely to be placebo-only-controlled. Although use of placebo-only controls among trials lasting longer than 52 weeks was rare, 16 percent of the trials lasting between 26 and 52 weeks were placebo-only-controlled.
Certain factors can explain some of our findings. For example, the lack of use of placebo-only controls in trials involving partial seizures is likely due to the fact that new anti-seizure drugs are typically assessed by adding the new drug to an existing medication because denying effective seizure treatment can have serious consequences in almost all subjects. Additionally, placebo-only controls may have been common in COPD trials because related treatment has not been established to prolong survival, meaning failure to provide active pharmacological treatment may not increase the risk of mortality for subjects (although treatment does reduce the risk of symptom exacerbations, which can lead to life-threatening complications).
Although our study assumed a potential for risk exposure among placebo-only subjects based on trial design at randomization, it did not assess the degree of risk, including the risk of mortality, or the actual harm, if any, to these subjects post-randomization.
However, our findings about the widespread withholding of active interventions that are known to be beneficial from control subjects in trials involving the nine life-threatening diseases considered in our study may present risks for subjects, even among those who are considered to have low-risk disease. For example, we found that investigators withdrew therapy in 15 diabetes trials from subjects who were previously taking at least one oral agent, with three trials involving subjects previously taking up to two oral agents. This practice is consistent with FDA guidance, which advises that it is ethically appropriate to withdraw therapy from diabetic subjects previously controlled with one low-dose oral agent. However, doing so is concerning because these subjects may have been partially responsive to prior treatments and may have experienced harm from treatment withdrawal.
Our study noted a few limitations, including the fact that it relied on data available in FDA medical review documents and other public sources that varied in comprehensiveness, raising the possibility that important information related to trial intervention was not reported.
Moreover, trials with placebo-only control groups should require much greater scrutiny by sponsors, researchers and institutional review boards.
Our findings should alarm any patient who is seriously ill and considering enrolling in a clinical trial; when human subjects with life-threatening diseases are given placebos and are therefore deprived of available approved treatments that have been shown to be beneficial, they may be placed at significant risk of harm, especially if they have severe disease or the trial has a long duration.
We also called for future research to assess how the use of placebo interventions in trials supporting FDA approval of medications adversely affects placebo-group subjects.