Paying respect at the burial of infant remains that came to light as a result of the Alder Hey inquiry. Credit: C. Lane/Reuters/Corbis

"Fundamentally flawed": that's how a January 2001 report described ten research papers based on body parts taken from hundreds of dead children without their parents' consent. Yet more than a decade on, Nature has learned, only one of those papers has been removed from the scientific record.

The lack of action, even in what seems to be a clear-cut case, highlights the reluctance of institutions and journals to retract papers when the authors stand by the results.

A series of inquiries by the government, the police and the General Medical Council (GMC) revealed that until the late 1990s, pathologists in the United Kingdom had routinely been taking organs and other samples from children's bodies without asking parents for permission. The findings caused shock and disbelief across the country, and led to an overhaul of human-tissue legislation.

The scandal centred on Alder Hey Children's Hospital in Liverpool. A government-commissioned report chaired by the lawyer Michael Redfern concluded that Dutch pathologist Dick van Velzen routinely removed every organ from the bodies that he examined at the hospital, storing the tissues for research. Yet van Velzen often did not carry out proper post-mortems on these children, filing incomplete autopsy reports or making them up completely.

Van Velzen published several papers on these tissue samples with collaborators at the University of Liverpool, where he held a post. Much of this work aimed to investigate the causes of unexplained cot death, or sudden infant death syndrome (SIDS), by comparing infants who had died from SIDS with those who had died from other causes.

The Redfern report concluded that because van Velzen did not carry out proper post-mortems, it was impossible to be sure how any of the infants — in either the SIDS group or the control group — had actually died. In the light of this and other problems, the report highlighted ten peer-reviewed papers110 that it concluded were "fundamentally flawed". The GMC subsequently banned van Velzen from practising medicine in the United Kingdom, because of his serious professional misconduct in the matter. No action was taken against van Velzen's co-authors. Nature has been unable to contact van Velzen.

In 2009, Christian Holscher, a neuroscientist at the University of Ulster, decided to find out what had happened to the papers. Vyvyan Howard, a colleague of van Velzen's at Alder Hey and co-author on all ten papers, was by then working in Holscher's department.

Holscher was surprised to find that none of the papers had been retracted. "It's not okay that these papers are still in the public domain," he says. "If you don't know the background, you assume this is a scientific study with proper results."

Holscher wrote to the editors of several journals involved to ask why the papers had not been retracted. Victor Chernick of Pediatric Pulmonology retracted the paper in his journal3 almost immediately — so far, it is the only paper on the list to have been withdrawn. Elia Maalouf of Early Human Development replied to say that he would retract that journal's paper5, but this has not happened. Miguel Reyes-Múgica of Pediatric and Developmental Pathology said he saw no need to retract the paper in his journal4.

Reyes-Múgica told Nature that it is not up to him to judge whether the paper is invalid, despite the findings of the Redfern report. "The editor's job does not have to do with judicial matters," he says.

Five of the ten papers are currently available on PubMed2,47, and three are included on the University of Ulster's online publication database2,4,5. Most of the papers have not been highly cited, but one7 — which shows that growth-retarded fetuses have poor kidney development — receives a steady stream of PubMed citations, and has been cited six times since 2010.

Clinical contribution

Howard says that he and his co-authors considered whether to retract their papers after the Redfern report was published, but ultimately decided that all of their research remains valid. "We think we've made a contribution there," he says. "The outcome of that research has affected clinical practice," Howard says, in that growth-retarded fetuses are now delivered as soon as possible to improve their chances of survival.

The editors of these journals often had no idea that these papers were found to be invalid. ,

He and the other co-authors argue that Redfern and his colleagues misunderstood their work. Howard and Paul Sibbons, now director of operations at Northwick Park Institute for Medical Research in Harrow, UK, and co-author on seven of the criticized papers, say that several papers do not rely on a comparison between SIDS and non-SIDS cases, and so the lack of proper diagnosis would not affect the conclusions. Sibbons adds that even in the studies that did rely on a comparison, his analysis of the tissue samples provided diagnoses that were adequate for the research. He also says that he was "astounded" that the Redfern report contained "so many inaccuracies", and says that his research should not have been classified as "flawed".

"My findings of fact have remained unchallenged for 11 years," Redfern says. He declined to comment on the authors' decision not to retract the criticized papers. "That is their choice," he told Nature. "It is for you, I think, to draw conclusions."

Holscher, who describes the Redfern report as "detailed and meticulous", argues that the University of Liverpool should have requested that the papers be retracted, rather than leaving it up to the authors. "The editors of these journals often had no idea that these papers were found to be invalid," he says. "It's the university's job to tell them. It's not in the interests of the authors."

In a statement, a spokeswoman for the University of Liverpool said that the university has strict guidelines and standards for research integrity. "Following publication of the Redfern report, the university prevented further publication of work based on samples that were the subject of the inquiry." She added that the university's Research Governance Committee is now considering the papers that were not retracted. "We will request the papers are withdrawn if we feel this is appropriate."

The three papers currently included on the University of Ulster's online database are "legitimate published output from a member of staff", according to Norman Black, pro-vice-chancellor for research and innovation at the university. "The university is not aware of any request or recommendation by the Alder Hey inquiry to retract the papers," he says.

Tremendous variation

Ivan Oransky, a US science journalist who runs the blog Retraction Watch, questions the wisdom of leaving retractions either to authors or their institutions, noting that both are likely to be reluctant. He cites the case of Harvard psychologist Marc Hauser, who resigned from his position last month after an internal investigation, which has not been published. According to Harvard, Hauser was guilty of eight counts of scientific misconduct relating to three papers (see Nature 466, 908–909; 2010). "There was only one retraction," says Oransky. "The investigation took Harvard three years, and we still don't know what it says."

Instead, he argues that journal editors should play a more active role. "Editors should stand by what they publish," he says. He complains of "tremendous variation" in the attitudes of journal editors, but cites as a positive example Steven Shafer, editor-in-chief of Anesthesia & Analgesia, who led a group of editors of 11 affected journals after the German anaesthesiologist Joachim Boldt was found to have published research that had been conducted without proper ethical approval. Working together, the editors have since retracted around 90 of Boldt's papers.

Sibbons, who is himself editor of the journal Comparative Clinical Pathology, says that for an editor to judge properly whether papers should be retracted in cases such as these would require a retrospective review process, something that is unlikely to be worth the cost. He argues that follow-up experiments already provide a mechanism to check the validity of studies. "Replication of experiments is usually what highlights anomalies, not retraction," he says.

But Oransky points out that according to guidelines published by the Committee on Publication Ethics, a group of journal editors, it is within the scope of the editor to retract a paper if an investigation finds it to be flawed.

When the authors do not agree to the retraction, he suggests that an editor could ask them for a rebuttal and print that as part of the retraction notice. In the absence of a full retraction, he says that a warning label alerting others to any concerns would be "a great thing". The aim, he says, is to increase transparency and "treat science as a process rather than a product".