'Probable Misconduct' in Fracture Risk Trials in Stroke, PD

Deborah Brauser

November 18, 2016

An investigation published online in Neurology suggests that findings in 33 randomized controlled trials from one group of researchers in Japan that investigated various interventions to reduce hip fractures in patients with stroke, Parkinson's disease, or Alzheimer's disease are unreliable because of concerns regarding data integrity and outcomes that are "remarkably positive."

Using statistical analysis and comparing differences in P values, the researchers found that, altogether, the treatment groups were "improbably similar." Those patients who received any of the interventions, including vitamin B or risedronate, were 78% less likely to have a hip fracture than those receiving placebo or no treatment. However, other meta-analyses have shown benefit rates of only 40% or less.

"There were multiple examples of inconsistencies between and within trials, errors in reported data, misleading text, duplicated data and text, and uncertainties about ethical oversight," write the investigators.

"The more we dug, the more problems that we found," lead author Mark J. Bolland, PhD, University of Auckland, New Zealand, told Medscape Medical News.

The original studies were published in various journals between 1997 and 2012, with Yoshihiro Sato, MD, Department of Neurology, Mitate Hospital, Tagawa, Japan, listed as the lead author on 26 articles and as one of the researchers on the remaining seven. Three of the studies were published in Neurology, with the last one appearing in 2007.

In an accompanying editorial, Neurology Editor-in-Chief Robert A Gross, MD, PhD, writes that during the 11-month vetting process before publishing the new review, Dr Sato contacted the journal and "admitted that the work reported in Neurology was fraudulent, relieved those he listed as coauthors of any wrongdoing, and requested retraction." The journal published retractions for all three studies on July 12, 2016.

Dr Robert Gross

Dr Gross told Medscape Medical News that seeing faulty data on so many trials in the review "was disheartening." The use of statistics across all of the studies led to the new conclusions.

"There would have been no easy way to determine on an individual case that the data didn't make sense," said Dr Gross.

Nine other retractions have been published so far, including two from JAMA Internal Medicine (formerly the Archives of Internal Medicine), one from JAMA , one from Parkinsonism and Related Disorders , and one from Movement Disorders , owing to "concerns about data integrity and inappropriate assignment of authorship."

Detailed Digging

Dr Bolland reported that the idea for taking on such an investigation started with a conversation he had with coinvestigator Allison Avenell, MD, Health Research Unit, University of Aberdeen, Scotland.

"She mentioned that she had come across some trials by Dr Sato and his group with almost identical data," recalled Dr Bolland. "So I had a look and it was pretty obvious that there were a number of issues. So we decided to dig into them in a lot more detail."

Beginning in 2012, they assessed 33 trials in humans published by Dr Sato and colleagues within a 15-year period. The 18 journals that published the studies were as follows:

  • Bone: four studies

  • Neurology: three studies

  • Journal of Neurology, Neurosurgery, and Psychiatry: three studies

  • Yonsei Medical Journal: three studies

  • Archives of Internal Medicine : two studies

  • Journal of Stroke and Cerebrovascular Diseases: two studies

  • Journal of the Neurological Sciences: two studies

  • American Journal of Physical Medicine and Rehabilitation: two studies

  • Osteoporosis International: two studies

  • Journal of Bone and Mineral Research: two studies

  • JAMA : one study

  • Stroke: one study

  • Parkinsonism and Related Disorders: one study

  • Cerebrovascular Diseases: one study

  • Journal of Musculoskeletal and Neuronal Interactions: one study

  • Movement Disorders: one study

  • Aging Clinical and Experimental Research: one study

  • Kurume Medical Journal: one study

Together, the trials included 5894 older patients "with substantial comorbidity [who were] recruited over very short periods," write the investigators.

For example, it took only 2 months to recruit 280 men with hemiplegic stroke who lived in the community and were at least 66 years of age, 4 months to recruit 374 women with the same characteristics, and 2 months to recruit 500 women with Alzheimer's disease – all within the same 5-month period in 2003.

The researchers also found that one 55-bed hospital, with only four clinicians, was said to have seen, over 3 years, 40 patients with Parkinson's disease with neuroleptic malignant syndrome after treatment with levodopa had been reduced or withdrawn.

"A correspondent found this astonishing because clinicians at their institution, which has a special interest in Parkinson disease, could only 'recall 2 such cases in living memory,' " write the investigators.

In addition, the outcomes for the trial patients were "generally extremely positive and inconsistent with contemporaneous literature," with very low mortality rates and only an 8.8% withdrawal rate.

The relative risk for hip fracture for the patients receiving one of the interventions was only 0.22 (95% confidence interval, 0.15 - 0.31; P < .0001).

There was also a strong suggestion of a "systematic failure of randomization...because the randomized groups are much more similar than would be expected by chance, if randomization had occurred successfully," report the investigators.

Overall, the review "raises serious concerns about the integrity and validity of the reported results," they write, noting that the "implausible productivity" of the researchers, not keeping ethics committee members separate from study involvement, and self-plagiarism also contributed to the unreliability of the data. Also at issue was the question of authorship.

"Honorary" Authors

Dr Sato was the lead author of 26 of the 33 studies, and Jun Iwamoto, MD, Keio University School of Medicine, Japan, was listed as the lead for the remaining seven. However, Dr Sato admitted to granting "honorary" authorship, and stated that none of his colleagues engaged in any misconduct.

According to an interview with the editor of Movement Disorders, as reported by RetractionWatch, Dr Iwamoto and other listed authors confirmed in writing that they were only honorary authors.

Still, "having ghost authors is really against most current editorial standards," noted Dr Gross. "This is something that would give us worry in and of itself. We want authors to attest to what they did on the paper, and it has to be substantive."

In his editorial, Dr Gross writes that although they received the review in December 2015, it demonstrated "probable scientific misconduct (fraud) on a large scale" while using complicated methodology.

Therefore, a thorough vetting process was needed, including contacting the trial authors, Mitate Hospital, and the other journals, as well as asking for help from three statisticians and former Editor-in-Chief Robert Daroff, MD.

"The findings...had implications for the authors and their institutions, for other journals and editors who published work by these authors, and for clinicians, guideline committees, and policymakers," he writes.

He adds that detecting fraud in a single article can be difficult, so it was important that the investigation by Dr Bolland and colleagues examined many trials together.

However, "simply because this group analysis suggested the likelihood of fraudulent activity," it cannot be assumed that all 33 studies are fraudulent, he noted.

Next Steps?

So, what are the ramifications of all of this and what can be done in the future? Dr Bolland said he would hope that peer-reviewed journals would catch major mistakes in manuscripts, but minor errors can be more difficult.

"They could request all patient data for a trial in order for it to be verified. But that begs the question: would a journal be expected to do a complete analysis before they publish? That would create a huge amount of work for them," he said.

"Also, if someone is really smart and determined to manipulate data, which I'm not suggesting happened here, but for someone like that, I'm not sure you could pick it up even if you do have access to the raw data. It's just a really difficult situation."

When asked the same question, Dr Gross said that this case has led to a review of their internal peer review process to see whether there is any room for improvement. "It's a human process and open to error. To that extent, it's not perfectible. But you want policies in place that, at a minimum, are helpful."

Currently, that includes putting all submitted papers through an electronic process called iThenticate, which searches for overlap with published articles.

"This type of technology wasn't available when these papers were originally submitted," he noted, "so standards are always evolving and improving."

However, he does not agree that asking for all patient data from a submitted study is the answer. "Having data in repositories that are freely available is a great idea in the interest of science. But I don't think journals can be expected to be the repositories."

The study was funded by the Health Research Council of New Zealand. The study authors have disclosed no relevant financial relationships. Dr Gross is supported for educational endeavors by a Clinical and Translational Science award by the National Institutes of Health to the University of Rochester Medical Center and receives an honorarium from the American Academy of Neurology as editor-in-chief of Neurology.

Neurology. Published online November 9, 2016. Abstract, Editorial

Follow Deborah Brauser on Twitter: @MedscapeDeb . For more Medscape Neurology news, join us on Facebook and Twitter .


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.