Half of all cancer drugs approved by the European Medicines Agency (EMA) from 2014 to 2016 were the result of trials that exhibited a high risk of bias and even exaggerated treatment effects, a new research study says.
Researchers at seven institutions in the U.S. and UK examined data from 39 studies resulting in EMA approvals following concern that new drug submissions in Europe lack enough scientific evidence to support approval. According to the study, risk of bias was seen in 19 of the oncology drug trials — in 10 trials owing to missing outcome data and seven due to the way outcomes were measured.
The study was tasked with examining the design characteristics, risk of bias and reporting adequacy of trials of oncology drugs approved by EMA in the three-year period. Oncology is the single largest category of drug approval in the EU. More than one-quarter of EMA approvals in 2017 were for cancer drugs.
Randomized control trials (RCT) comprised 76 percent of the study sample. Only seven approvals were supported by two or more RCTs. Researchers found an increasing percentage of cancer drugs are approved based on findings from non-randomized or single-arm trials, which made up 24 percent of the study sample.
Of the RCTs, 23 trials demonstrated deviations from intended interventions due to either lack of blinding or risk of compromised blinding.
Researchers found that trials evaluating overall survival were at lower risk of bias than those that evaluated other endpoints, such as surrogate measures of clinical benefit for cancer patients. However, overall survival, or the length of time from either the date of diagnosis or start of disease treatment to present, was used as a primary endpoint in only 26 percent of trials.
Other trials evaluated indirect measures of clinical benefit, such as disease response, event-free survival or safety endpoints, which are not always a reliable predictor of whether a patient will live longer and have a better quality of life.
Researchers also identified 10 trials with concerns resulting from inappropriate comparators and non-preferred study endpoints.
Findings of bias differed, the study said, between trials relying on information available from scientific literature as opposed to regulatory documents, such as European Public Assessment Reports (EPAR), indicating that the information being communicated by these different sources is being interpreted differently by sponsors and can result in confusion and misconception.
For example, researchers found in the case of one study, that scientific literature was missing some outcomes data, but the related EPAR provided more information about study and treatment discontinuations. Taken together, the two sources provided a more accurate picture of study results.
“Because cancer drugs are responsible for most of the recent increases in pharmaceutical spending,” the study concludes, “the evidence base that supports their market entry warrants close scrutiny.”
