Fierce Biotech recently focused on a report by The BMJ, a British science journal, on the validity of clinical trials which are critical to decision making by patients and clinicians.
The BMJ reported that about half of the European studies of cancer drugs were possibly biased. These reports involve drugs that were approved by the European Medicines Agency (EMA) between 2014 and 2016.
For example, only nine out of thirty-two cancer drugs that the EMA approved between 2014 and 2016 used one or more randomized control trials and were therefore considered at minimal risk with respect to bias.
There is also a strong indication that treatment results were exaggerated.
The BMJ also reported on European cancer drug trials conducted around 2010 with major flaws that question whether the trials were actually beneficial.
Bias is Not New
Research began in earnest in 1995 showing that substandard design, analysis, and reporting of clinical studies have an effect on their overall results. These shortcomings can result in biased findings and exaggerate the actual benefit of the treatment being investigated.
Surrogate Endpoints Versus Overall Survival
Cancer drugs represent the largest class of drug approvals and product development. What matters most to cancer patients is how long they can remain on a drug and be cancer-free.
For this reason, there is cause for concern because many cancer drugs are evaluated based on their surrogate endpoints. That may include the reduction of tumors or lower levels of biomarkers (measurable substances), not survival or quality of life.
A cancer drug that meets surrogate endpoints may not improve overall survival. There are many recent examples but a dramatic example involves the BELLINI trial and patients with multiple myeloma.
In this trial, the patients who were treated with the oral drug venetoclax had a shorter period of survival than the patients in the control group.
Yet venetoclax seemed to be more effective based on the results of common surrogate endpoints such as the patients’ initial response to the drug.
Another factor that clouds results of the trials is the small number of participants and the short follow-up. The drug olaratumab that targets soft tissue sarcoma is an excellent example because its phase 2 trial provided data for its initial regulatory approval. However, its phase 3 trial did not show benefits of overall survival.
It should be noteworthy to point out that these issues are in stark contrast to the hype surrounding new cancer therapies that are reported to the media or are presented at medical conferences.
Some Bias is Inevitable
Due to the complexity of drug trials, it is difficult to avoid some bias. But either intentional or unintentional lack of data for trial endpoints will inevitably result in biased judgments.
The possibility of bias may also enter the realm of double-blind trials when an investigator recognizes side effects that are associated with a particular treatment.
Reports and regulatory documents seldom mention measures that are taken to adhere to trial protocol.
Reports are Silent as to Outcome Data
A certain number of patients withdraw their consent and discontinue their participation from cancer trials. Outcome data about these patients is rarely discussed.
Some trials have inappropriately censored patients after changing their treatment. This was usually not disclosed in trial reports.
Clinical Trials: Gauging Validity
The BMJ authors claim that it is almost impossible to ascertain the validity of clinical studies due to the lack of pertinent detail. The authors also claim that critical details are included in supplementary appendices and that it is difficult to decipher statistical analysis plans or trial protocols.
BMJ’s Recommendation
BMJ recommends:
- Testing innovative strategies to communicate the validity of clinical studies
- Assessment of that evidence so that these trials accurately portray treatment effects
- Include a risk of bias assessment in Public Assessment Reports for pivotal studies supporting drug approvals
- The European Clinical Trials Register might mandate submission of risk bias assessments together with trial results
- An assessment of risk of bias could be submitted by trial investigators in their peer-reviewed publications
- A website might be developed by academic researchers that would report the risk of bias in studies supporting drug approvals.
About Future Efforts
The EMA has recently emphasized regulatory transparency. Future efforts by the EMA would greatly improve trial validity through enhanced reporting standards.
Input from clinicians and patients can ensure that risk of bias information leads to informed decisions by patients and clinicians.
Have you or anyone you know participated in clinical trials?
