But, for all the grief it’s causing, Protocol 321 may be the least of Upjohn’s problems. Far more critical to the U.S. Food and Drug Administration’s 1982 approval of Halcion was a clinical trial called Protocol 6045. Unlike Protocol 321, a six-week tolerance study conducted on 58 healthy prison inmates, Protocol 6045 included four 28-day studies involving 129 true insomniacs. It was the largest of three trials the FDA deemed “pivotal” to Halcion’s approval. The problem, overlooked by FDA officials until NEWSWEEK brought it to their attention last week, is that most of the evidence from Protocol 6045 is worthless. More than half of the study’s participants were enrolled, treated and examined by a confessed fraud.
Dr. William C. Franklin, an internist at the Clinic of the Southwest in Houston, helped a number of pharmaceutical companies test new drugs during the 1960s and ’70s. But in the fall of 1982, Franklin’s career as a clinical researcher came to a halt. An FDA investigation showed that he had fabricated patient records or failed to report adverse reactions in several drug studies. No one accused him of falsifying his Halcion data, but he agreed in a statement dated Oct. 20,1982, that in light of his violations he would not participate in any study of “investigational new drugs or devices.” On Nov. 24, nine days after it approved Halcion for marketing, the FDA sent letters to Upjohn and nine other companies, notifying them that Franklin had been disqualified. “FDA will not accept data from Dr. Franklin’s studies in support of claims for products within FDA jurisdiction,” the letter read. “Please determine whether, after his data are removed from consideration, those data remaining are adequate to justify all claims for safety and efficacy.”
Five months later, Upjohn wrote back to say that Franklin’s role in the Halcion trials had been negligible. E. L. Schumann, director of regulatory affairs, noted that Franklin had participated in three studies. But he didn’t mention that those included Protocol 6045–even though it was the largest of three pivotal studies and Franklin had enrolled 52 percent of the patients. He simply noted that Franklin had participated in two other studies (Protocol 6040 and Protocol 6041) that generated data on adverse reactions. Protocol 6041 was in fact the second largest of Upjohn’s three pivotal studies, and Franklin had enrolled 18 percent of the patients. But Schumann didn’t mention that either. Instead, he asserted that since Franklin was responsible for just 4 percent of the patients in a large group drawn from various unspecified studies, “it is unlikely that the deletion of his data would have a meaningful impact on the overall incidence of reactions.” Upjohn stands by that account. “When we got notification from FDA, we analyzed the safety profile extensively,” says Upjohn spokesman Philip Sheldon. “Franklin’s data didn’t affect it at all.”
FDA officials were astonished last week to discover the true impact of Franklin’s work. They’re now scrambling to determine whether Halcion’s 1982 approval was even warranted. Luckily for Upjohn, the agency won’t look only at the studies it deemed pivotal at the time. The FDA is already reviewing the drug’s entire record in light of the controversy over Protocol 321. And Dr. Robert Temple, director of new drug evaluations, guesses there is now enough additional evidence on Halcion to justify the drug’s license. But Temple, like other FDA officials, feels duped. “This letter is just stunning,” he says. “It’s obviously incomplete, inaccurate in part and generally misleading.”
Upjohn dismisses that charge. “We’ve reviewed all our communications with the FDA,” says Sheldon. “There was no attempt on our part to be deceptive.” He adds that since FDA has all the Halcion studies on file, nothing they contain should come as a surprise. Why didn’t anyone at FDA glance at the records back in 1983? How did the omissions in Upjohn’s account go unnoticed for nearly a decade? “Upjohn’s letter came to us,” says Dr. Alan Lisook, chief of the FDA’s clinical investigations branch. “We saw it and said it looked reasonable. We felt there were problems with an investigator. We had no indication there might be problems with the sponsor.” Those problems are now unmistakable.
