Biodel Faces Questions On Indian VIAject Trials
This article was originally published in The Pink Sheet Daily
Executive Summary
Anomalies in trial data skew results in type 1 diabetics.
|
Stock-hammering bad news from key trials with VIAject , Biodel's fast-acting mealtime insulin shot, could mean the company must redo its study in type 1 diabetes. A new study may delay filing for approval at least a year, unless the firm can persuade FDA to clear the product solely on data in type 2 patients - or unless anomalies in data from Indian patients can be ironed out to the agency's satisfaction. Biodel disclosed the trouble Sept. 8, in a poster presented at the European Association for the Study of Diabetes meeting in Rome, but could not explain the data glitch. Shares of the Danbury, Conn.-based firm plunged more than 66 percent, or $11.29, closing at $5.75 on the day, and officials slated a conference call for Sept. 9 to discuss the trial results. J.P. Morgan analyst Cory Kasimov, in a research note, called the sell-off "overdone," and wrote that the data overall "confirm the product's intended value proposition" - similar glucose control without hypoglycemia and weight gain. The trials were designed to show non-inferiority of VIAject to Lilly's recombinant human insulin, sold as Humulin , in both types of disease by measuring HbA1c levels, hypoglycemia, and weight. With VIAject - thus far unpartnered - patients would use a smaller dose of insulin to handle their glucose, thanks to the blocking of hepatic gluconeogenesis brought about by the product's fast release. Biodel's approach is to combine ingredients regarded as safe with regular insulin, and shoot for approval through the 505(b)(2) regulatory pathway. At the end of the second quarter, the firm had about $100 million in cash with a burn rate of about $40 million per year, so Biodel probably has enough to do another trial if FDA demands it. Type 2 results from 186 evaluable patients who got VIAject and 205 given Humulin fell in line with favorable data found during an interim look after six weeks of therapy. But in the type 1 study - performed, like the type 2 experiment, in the U.S., Germany and India - problems with the Indian segment led to a larger than acceptable confidence interval, so the Indian results were excluded from the analysis. That analysis included 131 VIAject patients in the U.S. and Germany, and 140 given Humulin. In the two groups, HbA1c levels dipped comparably, with an adjusted difference between the groups of -0.1 percent. A 95 percent confidence interval of -0.3 to +0.1 let researchers establish non-inferiority between the two products in the type 1 study when only the U.S. and Germany were considered. India, though, had problems. Data from patients there showed markedly hiked HbA1c levels before and after the study started, a "medically improbable doubling of intra-subject variability" in those results and much reduced reporting of hypoglycemic events. Adding the Indian HbA1c data boosts the adjusted treatment group difference to -0.3, not clinically meaningful - but it also changes the lower limit of the confidence interval, and does away with the non-inferiority finding. Apart from the trouble in India, VIAject did well, turning up fewer incidents of mild-to-moderate hypoglycemia and proving that it has little effect on weight gain, unlike Humulin. Some VIAject patients dropped out because of injection-site pain, a known effect due to citric acid in the formulation. - Randall Osborne ([email protected]) |