There’s an unhelpful retraction notice online in the journal Antioxidants & Redox Signaling, a Mary Ann Liebert publication. The paper, “Inhibition of LXRalpha-dependent steatosis and oxidative injury by liquiritigenin, a licorice flavonoid, as mediated with Nrf2 activation,” has been removed from the site, except for the abstract, which now has this in front of it:
THIS WORK HAS BEEN RETRACTED BY THE AUTHORS
That, as we’ve said before in exasperation, certainly clears things right up.
But we found out the reason for the retraction from Paul S. Brookes, an associate professor of anesthesiology at the University of Rochester Medical Center. Here’s the letter he sent the editors of Antioxidants & Redox Signaling and Free Radical Biology and Medicine, an Elsevier title:
I wish to report data duplication in two (maybe three) publications in your journals. The citations are as follows:
(1) Kim YW et al. (2011) Antiox Redox Signal 14(5): 733-745.
(2) Kim YW et al. (2010) Free Radic Biol Med 48(4): 567-578.
(3) Kim YW et al. (2011) Free Radic Biol Med 49(11): 1722-1734.
Specifically, Figure 1 of paper (1) and Figures 4 and 5 of the paper (2) contain duplicated data. In some cases, the data is presented as originating from the same experimental groups – this is annoying but not strictly wrong. However, there is at least one case where identical data is represented as originating from two independent and unrelated experiments.
Figure 1A of paper (1) shows body weight gain in mice fed different diets, below on the left. For comparison, Figure 4A of paper (2) is shown on the right. (View two figures here.)
The HFD and ND control data from these two papers are identical. This is not particularly troublesome, although from an experimental perspective, representing that the two drug treatments (Sauchinone in paper 2 and Liquiritigenin in paper 1) were normalized to controls, when in fact the same controls were used for both studies, is not very rigorous. The submission dates of the papers were September 2009 and April 2010 -7 months apart, so arguably the experiments could have been done at the same time with one group of controls. This should have been stated upon submission.
But this is not the biggest problem…
Figure 1D from paper 1 shows cross sectional micrographs to indicate steatosis in the livers. These data are shown on the left below. Figure 5B from paper 2 is shown below on the right. (View two figures here.)
In the ARS figure (left), the bottom 2 panels show the effects of liquiritigenin on steatosis. In the FRBM figure (right), the bottom 2 panels show the effects of sauchinone on steatosis. These are the same figures duplicated to represent two completely unrelated treatments. This could possible be an error on the part of the authors – they could have simply mixed up the images? The fact that the images have been adjusted differently (the ones on the left have the brightness and color saturation enhanced) suggests this is not a simple mix up.
There is also the underlying issue that many of the data in paper 1 (ARS) are overlapping with another FRBM paper, listed as #3 above. For example Figure 3B of paper 1 shows release of the liver enzyme ALT in high and low fat diet fed animals exposed to LQ. Figure 7A of paper 3 shows the same thing… the numbers match exactly. The same data has been published twice… (View two figures here.)
It is also evident that many of the data in Figure 2 of the ARS paper (#1) are repeated in Table 1 of the FRBM paper (#3).
I trust that both of you will be looking into these matters shortly, and taking appropriate action. At the very least, a correction/erratum regarding the mis-labeling of the cross-sectional images is warranted.
The editors of Antioxidants & Redox Signaling moved very quickly, retracting the paper within two weeks of Brookes’ letter. There are apparently no plans to retract the Free Radical Biology and Medicine papers. We’ve contacted the editors of both journals, and will update with anything we learn.