Researchers have retracted a highly-cited 2003 Nature Genetics paper which identified mutations underlying some types of epilepsy after discovering some blatantly erroneous results that negate the study's main finding, they report in the September issue of the journal.

Generic family pedigree Image: Wikimedia commons, Armin Kübelbeck

"I think it's appalling, to be perfectly honest," said Massimo Pandolfo of the Hôpital Erasme of the Université Libre de Bruxelles in Belgium, who was not involved in the original publication, but cited the 2003 paper in his own work. "This goes beyond any degree of acceptable mistake."

The 2003 Nature Genetics article, which has been cited 187 times, according to ISI, claimed that four common subtypes of an inherited form of epilepsy were associated with three different mutations in a single gene coding for a chloride-gated ion channel. The work involved analysis of three different families and identified an autosomal dominant pattern of inheritance.

"We thought [the gene] must really have a dominant and strong effect on this [type of] epilepsy," said Holger Lerche of the University of Ulm in Germany, a corresponding author on both the original paper and the reanalysis of the data accompanying the retraction notice in the September issue of Nature Genetics. But after reexamining the data, "this cannot be held as a hypothesis now."

Two of the three pedigrees included several individuals who were marked as having an epileptic disorder, when in reality they had never had a seizure. Furthermore, some of the DNA samples used to identify the mutations appeared to have been used for multiple individuals.

The conflicting information came to the attention of Christian Elger, the head of the Clinic of Epileptology at the University of Bonn, early in 2007, said Vice Dean of Research Markus Nöthen. After confronting Armin Heils, the last author of the original publication and the only author who did not agree to the retraction, with the inconsistent data, Heils agreed to leave the clinic, and the university recruited the help of researchers who had not participated in the original work to re-examine the data and contact the families again for further analysis.

"We really wanted to know in detail what was really correct with the publication because that was a major impact publication," Nöthen said.

In the first family, they found, only three relatives had the mutation (instead of the originally reported five), and of those three, only one had epilepsy (instead of all five known carriers, as originally reported). Additionally, three of the original DNA samples with the mutation had come from just one individual. In the second family, only two individuals (instead of eight) apparently suffered from epilepsy, but because the family did not agree to give another blood sample, mutation carriers could not be identified. In this family, 16 DNA samples appeared to have come from just 10 individuals. (The third pedigree was confirmed as originally reported.)

These inconsistencies may explain why researchers have been unable to replicate the study's results. When geneticist Alexander Zimprich of the Medical University of Vienna in Austria and his colleagues were unable to identify any of these mutations in epileptic patients, he thought the mutations must be very rare. "But now," he said, "it seems to be clear why this thing has not been found in other replication studies" -- it does not have the obvious relationship to epilepsy that was claimed.

Nature Genetics editor-in-chief Myles Axton commended the authors for bringing the error to the public's attention. "This is probably the best example I've seen of authors correcting the record in a thorough way both through their own work and in collaboration with outside experts," he said. "It's just a shame that it took them so long."

Heils did not reply to requests for comment sent to his University of Bonn and personal email addresses.

The blog DrugMonkey, which posted the retraction notice, expressed confusion as to whether it reflected mistakes or misconduct. While no formal investigation of misconduct was conducted, Lerche said that these results are "very difficult to explain by an error that might have occurred."

"If this is just sloppiness, this is unacceptable sloppiness," Pandolfo said.


Related stories:

Nature to retract plant study
[9th December 2008]

Rumblings over Science retractions
[26th March 2008]

Authors retract Science paper
[26th July 2007]