Scientists alert Lancet to ‘data inconsistencies’ in the Russian vaccine trial paper

A group of 16 scientists has written an open letter to Dr Richard Horton, editor of The Lancet and authors of the Russia’s Sputnik V novel coronavirus vaccine trial paper, raising questions about the integrity of the data published in The Lancet on September 4.

Just four days after the results of the Phase-1/2 trials of Russia’s Sputnik V vaccine for novel coronavirus was published in The Lancet, a group of 16 scientists has written an open letter to the authors of the paper and Dr Richard Horton, editor of the journal, raising questions about the integrity of data published in the paper.

The open letter by Enrico M. Bucci, adjunct Professor at Temple University, who is well known for his work in biomedical scientific integrity, along with 15 other scientists from Italy, Japan, German, France has raised concerns about “potential data inconsistencies” in the published paper.

“With a group of colleagues, I found the data presented in the recent paper on The Lancet a bit puzzling. We need access to the original data,” Dr Bucci tweeted.

Stating that the scientific community should “pay even more attention to the scientific evidence and the underlying data” they point out that the importance of making data “fully available for close scrutiny”. “In lack of the original data… the research described in the article published by The Lancet presents several different points of concern,” they write.

“The points raised [by the 16 scientists] are valid. Without access to the actual data, the points and patterns on the graphs look too similar at multiple places. This raises concern about the data, and therefore the way this vaccine was tested,” virologist Dr Shahid Jameel, CEO of DBT-Wellcome Trust India Alliance says in email.

“Whenever unusual results are encountered and patterns look too similar, there are always concerns,” Dr Gagandeep Kang, Professor of Microbiology at CMC Vellore says in an email. “Having done many observational studies and clinical trials, including regulatory trials, I know that the quantity and details of data required for regulatory submission is greater at the individual and cumulative level than is required for a publication (even where data sharing is required). Therefore, given the small number of participants in the Russian vaccine trials, it should be possible to rapidly respond and share individual level clinical and laboratory data, so that anyone with concerns can review the data and re-create the quite simple figures in the manuscript. This should allay any concerns, and will, in no way, affect any related intellectual property of Gamaleya, because this is participant level data and not related to the vaccine and manufacture.” 

In three figures published in the paper, they point out several data patterns that appear identical in different experiments. For instance, in the figure on humoral immune responses, all the nine participants challenged with Ad26-S frozen vaccine appear to have identical antibody titres on days 21 and 28. Similarly, identical antibody titres are seen in the case of seven of nine volunteers challenged with rAd5-S freeze-dried vaccine. Data inconsistencies are also seen where the experimental points “differ for a constant value in two completely unrelated experiments” and are “completely preserved among other two completely unrelated volunteer groups”.

Despite the variable under study being discrete, it “seems to us that on the ground of simple probabilistic evaluations the fact of observing so many data points preserved among different experiments is highly unlikely,” they write.

In another figure on the results of the cellular response to the different formulations, experimental point patterns are seen to be repeated. “In this case, the variable under study (cell proliferation percentage) is continuous in nature, rendering the coincidence of data points among different experiments even less likely,” they write. 

Published in The Hindu on September 8, 2020