SAN DIEGO — Rosiglitazone has shown no hint of excess liver toxicity in 10 years of safety monitoring by GlaxoSmithKline, Alexander R. Cobitz, M.D., Ph.D., reported at the annual scientific sessions of the American Diabetes Association.
The withdrawal of the first thiazolidinedione, troglitazone, from the market in 2000 because of liver toxicity prompted concern about the entire thiazolidinedione class of glucose-lowering agents, and resulted in strict label requirements for frequent liver monitoring in patients using the second-generation agents rosiglitazone and pioglitazone. These requirements have since been relaxed to simply measuring liver function at the time of initiation of therapy and as clinically indicated.
Now, after 10 years of safety monitoring in more than 7,000 patients, no cumulative evidence of hepatotoxicity has been seen in nearly 9,000 patient-years of exposure. “In the shadow of troglitazone, there were requirements put on the other two. … However, each of these molecules is markedly different,” said Dr. Cobitz, director of metabolism, clinical development, and medical affairs at GlaxoSmithKline, King of Prussia, Pa.
Included in his analysis were the phase III data on 4,327 patients with 2,493 patient-years of exposure that were submitted to the U.S. Food and Drug Administration for the new drug application for rosiglitazone (Avandia) in 1998, along with subsequent postmarketing surveillance data comprising an additional 256% in drug exposure since it entered the U.S. market. In all, the data include 38 double-blind and open-label clinical studies of rosiglitazone treatment in North America and Europe.
Patients in all the trials had baseline liver function test results at or below 2.5 times the upper limit of normal. Liver function was assessed at screening, at baseline, every 4 weeks for the first 3 months, and at 6- to 12-week intervals thereafter.
Among 7,429 patients enrolled in the trials—including 3,194 on rosiglitazone monotherapy and the rest on rosiglitazone in combination with metformin, sulfonylurea, or insulin—a total of 0.3% experienced a serum alanine aminotransferase (ALT) level more than 3 times the upper limit of normal. That same proportion was also seen among the 2,792 patients who received comparators without rosiglitazone in the trials—placebo, metformin, sulfonylurea, or insulin.
The number of person-years of exposure among the rosiglitazone subjects ranged from 892 with rosiglitazone plus insulin to 8,851 with monotherapy. Among the comparator groups, the range was from 186 person-years of exposure to placebo up to 1,186 person-years of exposure to sulfonylurea.
“The rate of hepatic events per patient-year is similar, if not better, than for the comparators,” Dr. Cobitz pointed out.
There were no differences between rosiglitazone and the comparators in threefold elevations of aspartate aminotransferase or alkaline phosphatase levels, or for elevations of 1.5 times the upper limit of normal in total bilirubin levels.
In all groups, those proportions ranged from 0% to 0.8%, with 0.8% being the percentage of 1.5-fold total elevations in bilirubin levels in the placebo group. Among the rosiglitazone recipients, 0.5% had that amount of elevation in bilirubin levels.
Further analysis showed no differences in ALT levels between three and five times the upper limit of normal (0%–0.3% for all rosiglitazone and comparator groups) or in ALT between five and eight times the upper limit of normal (0%–0.2%).
These data support the idea that although the chemical structures of all thiazolidinediones are identical on the right-hand side, their different left-hand structures result in very different biochemical and metabolic features, Dr. Cobitz said.