A half-dozen biopharma companies are seeking further clarity from the US Food and Drug Administration (FDA) regarding draft guidance from August on using a placebo in certain oncology trials.
The draft features FDA’s thoughts on the ethical challenges
of using a placebo in randomized controlled clinical trials for therapies to treat hematologic malignancy and oncologic disease. The draft also discusses blinding and unblinding, patient reported outcomes (PROs) and practical and ethical concerns.
Roche’s Genentech said that it believes the draft could be “further refined to fully reflect” FDA’s increased recognition of the value of patient input regarding their experiences with disease and treatment.
“We urge the Agency to clarify that, although collection of PROs from patients blinded to treatment may be preferred to remove the risk of assessor bias, there are inherent difficulties in oncology and hematology settings to design trials as double-blinded studies. For instance, when treatments are being compared to active entities with differing schedules or routes of administration, particularly intravenous (IV), addition of placebo is not ethical and blinding of the patients not achievable.”
Merck, meanwhile, notes that the draft indicates that a placebo-controlled trial would be appropriate in circumstances, such as where surveillance is standard of care (SOC), or with certain trial design features, for example, if the trial uses an add-on design or when the endpoint intended to support a labeling claim has a high degree of subjectivity, such as with PROs. However, Merck cautions that the use of a placebo control “may raise problems of ethics, acceptability and feasibility in subsequent trials or other scenarios, and may not be acceptable once some benefit has been shown.”
Merck also says that it would be helpful if FDA could provide additional guidance on what a suitable comparator would be in situations where there is no standard of care (SOC) and a placebo control would be inappropriate.
“We also ask that FDA clarify when/which type of external controls might be adequate to substantiate the relevance of single-arm trial data. For example, would real world evidence (RWE)-generated historical controls or model-based meta-analyses of literature be acceptable, and in what situations?” Merck questions.
Regeneron also notes that the draft offers advice on the impact of unblinding on study design, endpoints and statistical analysis.
“Specifically, if patients are unblinded and receive alternative treatment upon progression as suggested in the draft guidance, evaluation of overall survival (OS) may be confounded by post-study treatments, and using OS as a primary endpoint may not be feasible. Therefore, we recommend that the Agency consider alternative primary endpoints to support registration for oncology drugs and incorporate these discussions in the final version of this guidance,” Regeneron said.
Bristol-Myers Squibb, meanwhile, said the draft is “a positive step” in recommending the use of placebos in blinded randomized controlled studies for the treatment of hematologic malignancies and oncologic diseases, but the company called for the scope of the guidance to be refined to include considerations for blinding in placebo control studies only.
“In certain sections of the proposed guidance, it is vague whether the recommendations apply to any blinded randomized controlled study or just the ones that use placebos,” B-MS says.
And AstraZeneca noted that the draft provides guidance on the considerations for patient and investigator blinding, but it does not address considerations for sponsor-level blinding of placebo-controlled trials. “It would be helpful if the Agency could provide information to be considered for maintenance of the sponsor blind and/or unblinding of the sponsor,” the company says.
As progression-free survival is a common primary endpoint in oncology trials, AstraZeneca further says that the draft does not address concerns with bias in “somewhat subjective” endpoints. “It would be helpful for the FDA to provide guidance on whether blinded central adjudication is required for endpoints other than overall survival,” the company adds.
AbbVie also sought more detail on the risks of unblinding/running open label studies, with descriptions on how potential biases should be managed, “including how much bias is acceptable.”