Two papers evaluating glucose meters — used by diabetics to monitor blood sugar levels — suggested that a couple of the devices don’t work as well as they should. Perhaps unsurprisingly, the companies that sell those meters objected to how the studies were conducted. By all accounts, the companies appear to be justified in their complaints.
In both cases, researchers used blood drawn from veins to test the meters. But manufacturers of the WaveSense JAZZ and GlucoRx glucose meters said their devices are designed to work with fresh blood from a finger-prick. Both papers have now been retracted.
The retraction notice for “Technical and clinical accuracy of five blood glucose meters: clinical impact assessment using error grid analysis and insulin sliding scales,” published in 2015 in the Journal of Clinical Pathology, hints at the issue:
This article has been retracted at the authors’ request pending investigation of comments about the methodology used in the research.
(Although retraction guidelines from the Committee on Publication Ethics recommend leaving retracted articles online, but marked retracted, it appears the original article is no longer online.)
We learned how the error came to light from co-author Hannah Lise Sutheran, affiliated with Queen’s Hospital in the UK. She says the paper was retracted:
due to a [methodological] error that was clear when it was explained by the manufacturer. The methodology error that the manufacturer of the blood glucose meter outlined was not evident in any of their literature so the error was not noticed until after initial publication, the referees of the article also did not pick up on this error due to lack of evidence or literature. Since the retraction and removal of the erroneous data, the article has been re-published as a cut down version so as not to invalidate the data we reported. The authors of the paper instigated the retraction of the paper as soon as we were made aware of the problem.
We asked Sutheran for more details on the error that felled the original paper. She said:
I used venous samples preserved in EDTA but the wave sense jazz and code free meters require oxygen in sample to provide accurate results (which is ok for patients as they use capillary finger prick blood) but for purposes of laboratory evaluation of systems preserved samples were used. These were the only manufacturers not to provide suitability of sample types in their literature. When the company director of wave sense jazz met the coauthor it was explained that the problem could not have been foreseen due to their lack of evidence to the contrary and he admitted this wasn’t the first paper he had instigated retractions for due to the same issue with sample type.
Parts of the paper are still accurate, though. Co-author Tim Reynolds, also affiliated with Queen’s Hospital, explained that not all meters require a lot of oxygen in the blood sample to work properly. The authors republished the analysis, excluding the WaveSense JAZZ and CodeFree meters, which did require the fresh samples; “Technical and clinical accuracy of three blood glucose meters: clinical impact assessment using error grid analysis and insulin sliding scales” now appears in the same journal as the original article.
A spokesperson for the WaveSense JAZZ meter’s manufacturer, AgaMatrix Europe Limited, told us that he could not comment on whether the company had instigated retractions of other papers.
A similar methodological problem felled another 2014 paper, which also initially reported problems with a specific glucose meter not working after testing it on the wrong kind of samples. Again, the company objected:
After “System Accuracy Evaluation of the GlucoRx Nexus Voice TD-4280 Blood Glucose Monitoring System” was published in Disease Markers, the CEO of GlucoRx published a comment in the journal explaining that the study:
uses an inappropriately designed and uncontrolled evaluation protocol which cannot be used to demonstrate a blood glucose meter’s accuracy and draw valid conclusions regarding compliance with ISO 15197:2013 standards.
The comment suggests that there are several methodology issues with the study, including one with the blood samples used:
Although the evaluation compares “like with like” samples, it uses venous blood rather than fresh capillary whole blood samples as specified for accuracy studies in the ISO standard
After the comment, the authors retracted the study with this notice:
The paper has been retracted upon the authors’ request, as the routine hospital laboratory evaluation protocol used varied from recommended best practice study design. Although this protocol established the fact that GlucoRx Nexus Voice system was clinically accurate, it could not be used to make valid analytical accuracy statements relating to ISO 15197 standards. GlucoRx Nexus meters are now the formulary choice in the area.
The paper, published in 2014, has not been cited.
Like Retraction Watch? Consider making a tax-deductible contribution to support our growth. You can also follow us on Twitter, like us on Facebook, add us to your RSS reader, sign up on our homepage for an email every time there’s a new post, or subscribe to our new daily digest. Click here to review our Comments Policy. For a sneak peek at what we’re working on, click here.