When Novartis said it had fired two top AveXis execs in August for their role in the Zolgensma data fiasco, it wasn’t clear whether they doctored the data themselves or simply failed to catch the problem.
Now, in an official FDA filing, the company alleges the two “altered or instructed others to alter” the data, and it blames them for a delay in notifying the FDA about the questionable data, saying they slowed an internal probe of the whistleblower's accusations.
In its 59-page long response to a recent FDA Form 483 citation, the Novartis gene therapy unit said a whistleblower accused the two senior execs of doctoring raw data derived from a mice assay used at the time to test Zolgensma samples.
The first phase of the internal investigation—before AveXis officially opened a so-called nonconformance report at the company, and months before Novartis notified the FDA—“was significantly drawn out due to the implicated senior executives’ lack of cooperation and categorical denial of the allegations, which continues to this day,” the company says.
Former AveXis CSO Brian Kaspar, in a previous statement from law firm Heuston Hannigan, “categorically denies any wrongdoing,” and said he is “prepared to assert his rights and defend his conduct accordingly.” And contradicting AveXis’ description, he said he has “cooperated with Avexis’s internal investigation.”
Novartis has been under fire for its delay in reporting the data problems to the FDA, even after it confirmed the allegation internally in May. According to the drugmaker, it was running a two-phase investigation, and in Tuesday's FDA response letter, AveXis boasted it finished both stages each in less than 40 working days.
By AveXis’ account, the whistleblower hadn’t been available for an interview 12 days after the company launched its internal probe on March 14. After reviewing what it described as “tens of thousands of documents and electronic records,” and after the two execs—Brian Kaspar and his brother Allan Kaspar, who served as AveXis senior vice president of R&D—“could not offer a credible explanation for revisions to and inconsistencies in the data,” the investigation team decided the allegations were credible.
On May 8, the nonconformance report was started to dig out more details, launching the investigation's second phase at Novartis. It lasted 36 working days, the document states.
It was during that 36-working-day period that the FDA granted Zolgensma the green light in spinal muscular atrophy.
In the Form 483 response, AveXis stressed it was following “standard industry practice, operating procedures and precedent, all of which supported completion of the preliminary [non-conformance] investigation before submission of a report to the FDA given that there was no evidence of impact on patient safety, product efficacy or quality.”
Novartis disclosed the data manipulation to the FDA on June 28. In its response to the Form 483, AveXis blamed the timing on the complexity of data integrity issues, including the significant amount of data involved in the Zolgensma application and the difficulty in distinguishing between discrepancies that were the result of deliberate wrongdoing or a lack of clear instructions.
Still, it said it “fully appreciates that the circumstances presented by a new gene therapy [are] something that should be taken into account with regard to the timing of notifying FDA.”
The company recently promised to notify the agency within five business days after receipt of “any credible allegation” related to data integrity during a filing. Also, in light of the Form 483, Novartis said it has decided to "completely integrate" AveXis into the group's global quality network and "will accelerate several key integration actions to ensure that this integration occurs quickly and effectively."
“All major biopharmaceutical companies have […] cases where there is an impact on safety or product quality, there’s very clear guidance on how we need to report. When there’s not […] it’s our judgment on how best to manage the situation,” CEO Vas Narasimhan said during a recent investor’s event in London.
“What we’ve realized through this situation is that during a filing it’s difficult for us to exercise that judgment without it being later considered maybe not the best judgment. So we’re just taking judgment out of the equation,” he said.
The Swiss drugmaker has said the small proportion of implicated preclinical data doesn’t affect clinical evidence on Zolgensma’s efficacy or safety, and the FDA also said it believes the drug should remain on the market.