If you wanted to minimize the real-life effects of misconduct, you might note that some of the retractions we cover are in tiny obscure journals hardly anyone reads. But a new meta-analysis and editorial in JAMA today suggests — as a study by Grant Steen did a few years ago — that the risk of patient harm due to scientific misconduct is not just theoretical.
As the editorialists note, hydroxyethyl starches (HES) are “synthetic fluid products used commonly in clinical practice worldwide:”
Synthetic colloids received market approval in the 1960s without evaluation of their efficacy and safety in large phase 3 clinical trials. Subsequent studies reported mixed evidence on their benefits and harms.
There has been controversy over the use of HES for decades, with the most recent high-level review showing “no significant mortality increase.” But one of the reasons for that review — by the prestigious Cochrane Collaboration — was to see if the dozens of now-retracted studies by Joachim Boldt had an effect on the overall evidence for HES. Boldt’s retractions resulted from a lack of evidence of IRB approval, as well as the likelihood of faked data.
An internal investigation found no evidence of harm to the patients Boldt treated, and the the Cochrane review found “no change in the findings related to the inclusion or exclusion of the studies by Boldt et al.,” according to the editorial. But the new meta-analysis found something different:
After exclusion of the studies by Boldt et al, Zarychanski et al found that hydroxyethyl starch was associated with a significantly increased risk of mortality (risk ratio [RR], 1.09; 95% CI, 1.02-1.17) and renal failure (RR, 1.27; 95% CI 1.09-1.47).
In other words, there was an increased risk of death and kidney failure among those given HES:
The report by Zarychanski et al highlights the potentially important and adverse effect of scientific misconduct. With the inclusion of studies by Boldt et al, the medical community might reasonably have concluded that use of hydroxyethyl starch was not inappropriate. Yet the analyses in which these studies were excluded shifts the balance of evidence toward harm. This study also demonstrates the importance of revising and revisiting recommendations and guidelines in light of new systematic reviews and evidence.