By Marilynn Larkin
NEW YORK (Reuters Health) - 9/5/2019
Seventeen percent of randomized controlled trials that led to new anticancer drug approvals in the U.S. had suboptimal control arms and did not prove superiority to standard-of-care treatments, a quality-improvement study shows.
"Physicians should be skeptical of drug approvals and read the seminal trials carefully to be well-informed of the limitations of some of these results," Dr. Talal Hilal of Mayo Clinic in Phoenix, told Reuters Health. "Just because a drug is granted FDA approval does not automatically make it better than what is already available/practiced."
"Neither patients nor physicians want to use a drug that has not been proven to be superior to standard of care," he said by email. "This means weighing risks and benefits of the available standard of care and what the new drug may add."
Coauthor Dr. Vinay Prasad of Oregon Health and Science University in Portland added, in a separate email, "The question that faces any new drug is whether it makes life better beyond, in addition to, or after the current best medical care. If you show a new drug is better than something less than this, you haven't answered the question that faces patients and doctors."
"For example," said Dr. Hilal, "after eight years of using triplet therapy for relapsed multiple myeloma, the FDA continues to approve new agents that have been compared to doublet therapy," he noted. "Would they have been shown to be more effective than a triplet therapy control arm? We will never know."
"What does the new drug add beyond current best medical practice? We find a sizable proportion of trials shirk this question," Dr. Prasad said.
Drs. Hilal, Prasad, and Mohamad Bassam Sobol, also of Mayo Clinic, Phoenix, analyzed 145 randomized controlled trials (RCTs) that led to 143 anticancer drug approvals granted by the FDA from January 2013 through July 2018.
Control arm quality was deemed suboptimal if the choice of control agent was restricted to exclude a recommended agent; the control arm was specified but the recommended agent was not; and if prior RCT data had demonstrated that the control agent was inferior to an available alternative.
As reported online May 2 in JAMA Oncology, 47 single-arm studies were excluded, leaving 98 studies that led to 96 drug approvals for analysis. Sixteen (17%) of the approvals were based on RCTs with suboptimal control arms - including two accelerated approvals.
Four trials with suboptimal controls (25%) omitted active treatment in the control arm by limiting investigators' choice; 10 (63%) omitted active treatment in the control arm by using a control agent known to be inferior to other available agents or not allowing combinations; and two (13%) used a treatment in the control arm that had been taken previously and was associated with a known lack of benefit with re-exposure.
Dr. Hilal said, "The best way to ensure that control arms are optimal is for the FDA to set the bar closer to the reality of clinical practice in the US. There is probably a lag time between when therapies become commonly used in practice (i.e., become standard of care) and when the FDA decides that this should be the new control arm standard."
Dr. Jeff Sharman, Medical Oncologist at Willamette Valley Cancer Institute and Research Center in Eugene, Oregon and Medical Director of Hematology Research at US Oncology, echoed Dr. Hilal's concern in an email to Reuters Health, "In this quickly evolving landscape, studies may have obsolete control arms by the time they are completed, and interpreting the impact of a novel therapy against suboptimal control arms presents challenges to practicing clinicians."
Dr. Sharman, who was not involved in the study, coauthored a 2016 letter in The New England Journal of Medicine about the inappropriate use of chlorambucil monotherapy as a comparator regimen in studies of anti-chronic lymphocytic leukemia drugs (https://bit.ly/2VmwgTn).
JAMA Oncol 2019.