Downplaying Drug Toxicity in Oncology Clinical Trials

Kristin Jenkins

November 15, 2018

The "poor reporting practice" regarding toxicities associated with new cancer drugs needs to be improved, say researchers.

A review of more than 100 randomized drug trials published in five top-tier clinical journals shows that 43% used words such as "tolerable," "favorable," "acceptable," "manageable," "feasible," and "safe" to describe adverse events.

This language downplays treatment-related harms, which often appeared among data in tables but not in the narrative of the report, say the authors.

Some studies did not include data on serious adverse events (SAEs) or fatal adverse events (FAEs).

The report was published online November 1 in the BMJ.

"We consider the lack of harms reporting and the use of subjective terms to describe harms to be poor reporting practice," say the authors, led Bishal Gyawali, MD, PhD, who was affiliated with Nagoya University Hospital, Japan, at the time the study was conducted and is now at Brigham and Women's Hospital, Harvard Medical School, Boston, Massachusetts.

"Not fully reporting the harms of cancer drugs is of particular concern because cancer drugs usually provide modest benefits at high costs &mdash in terms of both price and toxicities," they comment. "Downplaying harms can suggest a better risk-benefit profile than actually exists."

Downplaying harms can suggest a better risk-benefit profile than actually exists. Dr Bishal Gyawali

However, an expert who was approached for comment pointed out that the review authors did not talk to the patients themselves.

Had the authors talked to patients, "they would have been surprised to learn the degree to which patients are able and willing to endure side effects for the chance of gaining benefit from treatment," commented Jeffrey S. Weber, MD, PhD, deputy director of the Perlmutter Cancer Center and professor of medicine at New York University Langone Medical Center in New York City.

"When you compare the [cancer drug trial] outcome for patients to death from cancer, then it becomes a little clearer what the term 'acceptable' or 'tolerable' means," he said.

Patients will endure significant side effects in return for modest benefits. Dr Jeffrey Weber

"In my experience of 30 years as an investigational oncologist, patients will endure significant side effects in return for modest benefits," Weber told Medscape Medical News.

Downplaying Harms

The authors of the review assessed 122 cancer drug trials published in 2016 in the New England Journal of Medicine, Lancet, Lancet Oncology, the Journal of the American Medical Association, and the Journal of Clinical Oncology. They found that 14 drug trial reports did not include data on SAEs, 22 presented no data on serious events, and two presented no data on deaths.

In the 39 trials that did report harms data, the investigators found that the rates of SAEs were higher in the experimental arm than in the control arm in 30 trials (77%), SAEs were higher in 26 of 31 trials (84%), and deaths were higher in 34 of 51 trials (66%).

"Thus, despite using terms such as 'favorable' and 'tolerable' to describe the harms profiles of new treatments, the trials often showed a greater number of harms than in the control arms," the authors say.

When asked if the findings came as a surprise, Gyawali told Medscape Medical News that after contacting one of the corresponding authors for data on the incidence of FAEs in a drug trial, he was told the information was confidential and could not be shared.

"How can anyone publish a trial report of a drug, claim acceptable toxicities, not report data on FAE, and claim that toxicities data are confidential? That was the biggest surprise of all," he said.

Before the study began, Gyawali said the reviewers were familiar with the terms used to downplay toxicities associated with drug therapy, having heard them used at conferences and having read them in abstracts. They assumed this kind of language, which is discouraged in the CONSORT guidelines for reporting harms, would not survive the editorial process at leading clinical journals.

Instead, they found that in an alarming number of cases, benefits information, even with surrogate endpoints such as improved responses, was overly emphasized, whereas harms data, such as the number of drug-related deaths, were downplayed or not reported.

For example, in a published clinical trial on ribociclib (Kisqali, Novartis and Astex Pharmaceuticals), a CDK46 inhibitor for hormone receptor–positive advanced breast cancer, the discussion section included the sentence, "Most patients had an acceptable adverse-event profile."

On assessing the data, Gyawali and colleagues found that more than twice as many patients in the ribociclib arm as in the control arm experienced SAEs of grade 3 or higher.

"The difference in treatment related serious adverse events (leading to death, life threatening condition, hospital admission or prolonged admission, disability or permanent damage, congenital anomaly or birth defect, or that required medical or surgical intervention to prevent one of the other outcomes) was nearly five times higher," they write.

In another example, the concluding paragraph in the NAPOLI-1 trial report on the use of nanoliposomal irinotecan (Onivyde, Baxter/Merrimack) in pancreatic cancer refers to the drug's "manageable and mostly reversible safety profile." But the data show that there were five FAEs in the intervention arm and none in the control arm.

"To claim that toxicities were tolerable is a big disservice and disrespectful to the oncology community," said Gyawali. "All patients deserve complete information on benefits and harms from cancer drugs to make delicate treatment decisions."

The authors propose that patient data be gathered as to whether the toxicities associated with drug therapy are acceptable. Any drug-related impact on quality of life should be objectively assessed using validated tools, say the authors, who have published another review showing that quality-of-life reporting in cancer drug trials is also poor.

"By mandating quality-of-life information for every trial, concerns of clinicians and patients about the benefit-harm trade-off with cancer drugs can be allayed," said Gyawali. "Journals and regulatory agencies can make this happen."

When approached for comment, Weber agreed that more patient reporting would be useful.

"Incorporation of validated instruments to assess patient-reported quality of life is increasingly seen in large randomized cancer trials," he pointed out, adding that this is "highly informative" and "should be encouraged."

Weber also noted that in most clinical drug trials, toxicities are reported for fixed periods during treatment. When patients remain on drug therapy for longer periods or finish therapy and experience late side effects, these may not be reported.

Similarly, the dose-limiting toxicity periods that are used to decide whether or not drugs are toxic are often relatively short &mdash 4 to 6 weeks &mdash he explained. Because a journal report may include only those side effects that occur more than 5% to 10% of the time in a given category, rare side effects may be excluded.

"That being said, grade 3-4 toxicities and deaths that might possibly, probably, or certainly be due to treatment are always reported carefully in published cancer clinical literature, in my view," Weber told Medscape Medical News.

He has encountered very few clinicians who complained of unexpected toxicity in the cancer drugs they were using or who said that clinical trial reports were misleading or provided insufficient warning about side effects, Weber said.

"A brief perusal of an average informed consent form for a typical cancer treatment trial lists literally dozens of side effects that are uncommon, more common, or even likely to occur," he noted. "Most trial reports include detailed information about side effects, by grade if not by SAE vs AE. I personally would rather see the graded toxicities and the likelihood of stopping therapy for toxicity, which is an important parameter."

Details Should Be Included

In their article, Gyawali and colleagues emphasize that the incidence of severe, serious, and fatal adverse events should be included in all trial reports and that vague terms used to downplay toxicities should be avoided completely. "Proper risk-benefit assessment of any cancer drug should be made with actual harms and efficacy data, and not based on general concepts of safe, tolerable, or intolerable," they say.

"Every trial must transparently and clearly report on severe, serious, and fatal adverse events in addition to the routine reporting of other adverse events," Gyawali told Medscape Medical News. He noted that FAEs in particular should be described in detail, perhaps in an appendix.

The authors recommend that editors and reviewers at medical journals ask for detailed, quantified harms data and actively discourage the use of subjective terms, especially in abstracts and conclusions. When reading clinical journal reports on drug trials, clinicians should make a point of looking at the tables that present data on toxicities.

Dr Gyawali is an adviser to the BMJ analysis team. Coauthor Yuichi Ando, MD, PhD, has relationships with industry that are not related to the current report. Dr Weber has relationships with Altor BioScience, Celldex, Biond, CytomX Therapeutics, Bristol-Myers Squibb, Merck, Genentech, AbbVie, AstraZeneca, Daiichi Sankyo, GlaxoSmithKline, Eisai, Altor BioScience, Amgen, Ichor Medical Systems, CytomX Therapeutics, Nektar, Novartis, Array, WindMIL, Takeda, Sellas, Pieris Pharmaceuticals, Eisai, CNektar, Astellas Pharma, Incyte, Roche, and Novartis.

BMJ. Published online on November 1, 2018. Abstract


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.