Separating the wheat from the bias


Ami Banerjee blogs on how the Catalogue of Bias can help us sift out the good information when reading about the next ‘blockbuster’ drug

Heart failure (HF), when the heart’s output is not meeting the body’s demands, is increasing internationally, and the race is on to find new treatments and preventive measures. “HF with preserved ejection fraction” (HFpEF), where the heart’s relaxation rather than pumping function is defective, has growing importance, but currently lacks any effective treatment.

Entresto, the trade name for sacubitril-valsartan (an angiotensin-neprolysin inhibitor), made shockwaves in the cardiology world following the results of the PARADIGM-HF trial in 2014, which showed a 20% reduction compared with enalapril, the longest-established treatment (angiotensin converting enzyme inhibitor), for the primary outcome of death or hospitalisation in patients with HF with reduced ejection fraction (reduced pumping function). These results led to the first fast-track approval of a non-cancer drug by NICE. On the 29th July 2019, Novartis, Entresto’s manufacturer, released the hotly anticipated results of the PARAGON-HF trial and presented findings at the European Society of Cardiology (ESC) Annual Scientific Congress in Paris a month later. The trial focused on HFpEF and showed negative results for Entresto against valsartan (an angiotensin-receptor blocker).

Why did Novartis release the results early (before the ESC congress embargo)? Perhaps this is a genuine effort to restore trust in the pharmaceutical sector and offset “positive results bias” and “industry sponsorship bias”, where sponsors typically promote positive results. However, one could argue publication bias is present because the early statement was not accompanied by full publication or presentation of results (at least not for a further month).

Novartis’ statement said the trial “narrowly missed statistical significance for its composite primary endpoint” and a senior Novartis official said, “sacubitril/valsartan may result in clinically important benefits in HFpEF”, which constitute “spin bias”, overplaying what are negative findings. Recent emphasis on improving outcomes in HFpEF and on the wide portfolio of Entresto trials makes “hot stuff bias” a possibility.

Metascience is outpacing science with more new systematic reviews being published than new trials. Both in volume and velocity, blogs and tweets have overtaken trials. The day after the early release of the trial results, hundreds of tweets and links appeared in my feed, at least a month before scientific publications from PARAGON-HF. It has never been so important for patients, public, health professionals and scientists to be aware of the limitations of science. The Catalogue of Bias is our multidisciplinary effort to address this need, with examples, by systematically describing biases which may occur in biomedical research, so that we can all ask the right questions.

Ami Banerjee, Associate Professor in Clinical Data Science and Honorary Consultant Cardiologist, UCL

Conflicts of interest: Advisory boards for Pfizer, Astra-Zeneca and Boehringer Ingelheim and Trustee for South Asian Health Foundation.