Some API industry professionals claim it’s common sense to discard meaningless data generated when these “gross laboratory errors” happen.
▸ Incorrect sample
▸ Incorrect standard preparation
▸ Incorrect sample weighing
▸ Wrong diluent used
▸ Wrong diluent volume
▸ Mobile-phase preparation error
▸ Improper mixing of mobile phase
▸ Incorrect pH adjustment
▸ Incorrect HPLC/GC setup and run
▸ Wrong injection volume
▸ Wrong detector wavelength
▸ Incorrect injection sequence
▸ Uncalibrated instrument used
▸ HPLC/GC auto-injector malfunction
▸ Lamp intensity failure
▸ Column leak, poor plate count, poor resolution
▸ Power failure
▸ Plumbing problem, leak
Over the past two years, the U.S. Food & Drug Administration has slammed several foreign drug suppliers for compromising the integrity of the data they record during drug manufacturing. In several cases, the result was a U.S. ban of pharmaceuticals made by those firms. Some industry players are now saying the focus on data integrity may be overdone and is unnecessarily reducing drug availability in the U.S.
“Data integrity issues are viewed as fraudulent acts,” said Peter J. Werth, chief executive officer of ChemWerth, an importer of active pharmaceutical ingredients (APIs) made at factories located mostly in China. “We feel that if FDA continues on the current path, this will severely impact the amount of APIs available for sale to the U.S. market and could very well lead to further drug shortages.” Werth made the remarks in June at CPhI China, a drug ingredients trade fair held annually in Shanghai.
FDA inspectors routinely visit drug manufacturers both in the U.S. and overseas, looking at the general state of their facilities and at the data they generate during manufacturing and quality control. More so than they do in the U.S., FDA officials conducting inspections abroad rely on data that companies supvply to determine whether they comply with current Good Manufacturing Practices (cGMP), a set of drug manufacturing regulations.
FDA has sent warning letters to and instituted bans on several manufacturers in India and China in recent years. Companies have failed inspections for undocumented deletions of test results or for not controlling who has access to data history. The list of companies caught in FDA’s dragnet includes ones otherwise considered to be proficient drug manufacturers.
Werth argued that FDA is going overboard on data integrity. In many cases, he alleged, technicians delete data from instruments such as gas or liquid chromatographs for perfectly innocent reasons. Lab instruments often record wild results because of “gross laboratory errors” that occur when technicians improperly prepare samples, fail to calibrate instruments correctly, or experience equipment failure. Technicians who delete such results aren’t trying to commit fraud but just want to discard useless, distracting data, Werth claimed.
C&EN showed Werth’s claim to FDA and the agency replied that he “appears to misinterpret applicable cGMP requirements.”
Despite its criticism, ChemWerth claims that none of the plants it imports from have received warning letters. A ChemWerth partner, Jinan Jinda Pharmaceutical, failed a European inspection in 2015, but at a plant that ChemWerth doesn’t represent.
Peter Saxon is president of Saxon International Associates, a New Jersey-based firm that has advised dozens of Indian and Chinese drug companies on U.S. drug manufacturing regulations. He agrees that FDA has become overzealous when it comes to data integrity. “There can be several very valid reasons that files in the [quality-control] instrumentation audit trails have been deleted,” he says. “The problem is not that these irrelevant files were deleted but rather not documenting why they were.”
Data integrity only emerged as a major concern of FDA in 2013, and some foreign manufacturers that previously hadn’t experienced regulatory problems may not have seen a reason to change their operating procedures, Saxon says.
But other drug industry veterans take issue with Werth’s assertions. For instance, Jaeyon Yoon, chief marketing officer at SK Biotek, a South Korean API manufacturer, argues that FDA should demand data integrity to prevent fraud. Otherwise, he says, “when an API does not meet specifications, a simple fix for the manufacturer is to retest it until it passes.”
Dinesh S. Thakur, head of Medassure Global Compliance, a pharmaceutical supply chain consulting firm, also argues for data integrity vigilance. Thakur was the whistle-blower who, in 2007, alerted FDA to systematic manipulation of manufacturing records at India’s Ranbaxy Laboratories to mask quality issues. FDA eventually fined Ranbaxy $500 million.
“What happens on the ground, I am afraid, is far from what should happen in principle,” Thakur says. “Even if a batch of API fails, there has been evidence uncovered by FDA which showed that formulation development companies blend two batches to make them pass specifications.” Moreover, he says, some API manufacturers will risk damage to their reputations and ship failed batches—with doctored quality-control results—rather than shoulder the expense of producing a new one.
Companies should have procedures in place to prevent lab technicians from deleting data, even off-results generated after gross laboratory errors, Thakur says. “Who makes a final call on what a gross laboratory error is and what is fraud?” he asks. “We need to take subjectivity out of the process and make it transparent so that an independent third party can assess what a true error is compared to a breach of data integrity.”
Although it’s a hot topic now, data deletion may soon become a thing of the past. Increasingly, leading instrument makers such as Waters and Shimadzu are launching instruments that don’t allow users to modify recorded data. “The FDA will likely make this mandatory, perhaps within the coming year,” Saxon expects. Drugmakers will have to explain every flawed test result, but the controversy over data integrity will likely fade away.