The Food and Drug Administration sent Novartis an unambiguous signal this week, indicating the drugmaker could face civil or criminal penalties for manipulating data used to support approval of its gene therapy Zolgensma. 

That warning was also directed at other pharmaceutical companies, former agency officials say, a reminder that’s particularly relevant for those working on gene therapies and other cutting-edge treatments. 

“This is a rapidly evolving field where there are going to be accelerated approvals. The quality of the data is critical,” said former FDA chief Robert Califf in an interview with BioPharma Dive. 

“People who are working in a lab on a cure for a disease in children should be aware that falsifying data is a crime and it’s huge betrayal of public trust.” 

The FDA’s decision to clear Zolgensma was a landmark moment for Novartis, which invested $8.7 billion in buying the treatment’s biotech developer, AveXis. 

The Swiss pharma company needed data from only three dozen infants to convince the regulator of Zolgensma, which treats a rare neurodegenerative disease called spinal muscular atrophy. Results were dramatic, showing the treatment helped keep alive infants who would otherwise have almost certainly died or seen their disease progress. 

Yet that approval is now under the microscope, after revelations Novartis knew of problems with some of the preclinical data in its application but didn’t tell regulators until after Zolgensma’s approval. 

The faulty data, according to the FDA, was confined to results from an assay used to test in mice two different versions of Zolgensma. Clinical data was not manipulated, and the agency affirms Zolgensma should remain on the market. 

Even preclinical data can be important, though, and more so for drugs with limited data from human testing.

“Ensuring truthful, complete and accurate data in product applications, regardless of priority review status, is a critical component of industry’s responsibility as they work to demonstrate the safety, purity, and potency of biological products,” the FDA wrote in an emailed statement, referring to the accelerated evaluation granted Zolgensma. 

“The submission of such truthful, complete and accurate data is also critical for the FDA to be able to protect the public health, and the law requires it.” 

Zolgensma is only the second gene therapy for an inherited disease cleared for market in the U.S., but others — with varying sizes of clinical evidence — are nearing regulators’ desks.

As Novartis tells it, the company acted quickly to substantiate and fully investigate the allegations of data manipulation. At no point did the company believe the inaccurate data risked patient harm, informing its decision to wait until it had findings in hand before informing the FDA. 

“I would say it was normal course of business and pretty typical for how we handle these items,” said Novartis’ head of regulatory affairs, Rob Kowalski, on a call with investors and Wall Street analysts Wednesday. It’s worth noting that Novartis has also informed health agencies in the European Union and Japan, where Zolgensma is not approved, of the data issues. 

Normal course of business or not, the FDA’s response has been anything but. In an agency memo, an FDA official indicated that, had the regulator known of the data issues beforehand, it would have delayed its decision to approve Zolgensma. 

“I suspect there will be wrongdoers here. And consequences,” tweeted former FDA Commissioner Scott Gottlieb on Tuesday.

Novartis has blamed a “few individuals” for the manipulated data — which involve discrepancies in time stamps and the recorded number of survival days for treated mice — and is in the process of “exiting” those employees. Novartis CEO Vas Narasimhan, meanwhile, described it as an “isolated incident.” 

It’s not clear how the FDA would determine whether to pursue civil or criminal sanctions against AveXis or Novartis. The agency plans to continue its assessment and could require supplemental applications for Zolgensma, a process that could take some months. 

Historically, settled violations between drugmakers and the government most often involve the overcharging of health programs or unlawful promotion.

A few in recent decades, however, have involved concealed data or poor manufacturing practices, according to data compiled by Public Citizen. Those centered on marketed or generic products, however, rather than manipulated data in a drug approval application. 

The FDA declined to respond to BioPharma Dive’s questions on historical precedents for Novartis’ situation. 

“The FDA frequently finds problems with manufacturing when scale up [from clinical testing] happens,” said Califf, adding “there’s always a judgment about whether you require that everything be fixed.” 

“I think it’s good for the FDA to make a statement for the world to say you’d better button down the quality of your information,” he added. 

Patti Zettler, an assistant professor of law at The Ohio State University and a former associate chief counsel at the FDA, echoed Califf’s comments. 

“There’s benefit to the agency publicly expressing how serious its concerns are, using its enforcement authorities when it discovers violations to send that signal,” Zettler said.