Investors have recently scrutinized Byetta’s potential to cause pancreatitis, a serious side effect associated with significant morbidity and mortality. In his review of the FDA MedWatch Adverse Event Reports System (AERS) database for Byetta, Lehman Brothers analyst Jim Birchenough found 24 unique cases of pancreatitis associated with Byetta from August 2005 through January 2006. Of these, 15 cases had no apparent confounding factors such as gallstones or other known causes of pancreatitis. Birchenough estimates that this translates into an incidence rate of 65/100,000 overall, and 40/100,000 once patients with confounding factors are eliminated. He argues that this rate is higher than the estimated baseline rate in the general population (17/100,000) as well as the AERS reported rate for Lantus (0.1/100,000) and TZDs (0.4/100,000 for Actos and Avandia) over the same time period. Furthermore, the reported rate for Byetta is in between that associated with atypical anti-pyschotics, which have some mention of pancreatitis in their label, and valproic acid, which has a bolded black box warning about pancreatitis.
Birchenough also argues that there is a strong temporal relationship between initiation of Byetta therapy and pancreatitis, which further supports a causal relationship. In 9 of the cases, pancreatitis occurred within one week of initiation of Byetta therapy. In 15 cases, it occurred within 1 month of Byetta initiation. In 8 cases, symptoms improved after discontinuation of Byetta. Furthermore, in 23 cases, Byetta was listed as the primary suspect drug by at least 1 report.
Birchenough also points out that it may be biologically plausible for Byetta to cause pancreatitis via a direct cellular effect since the pancreas is one of the primary sites of Byetta action. Since Byetta is also associated with antibody formation, it could theoretically also potentiate an immune response in the pancreas. (We are less impressed with Birchenough’s argument that pancreatitis is a theoretical concern based on Byetta’s origin in the saliva of the Gila Monster and documented pancreatitis associated with the Gila Monster bite.)
While Birchenough makes several compelling arguments, there are always other sides to examine. Hmm, where to start! Well, first, it is difficult to ascertain a direct causal role for Byetta since the AERS database contains limited information regarding the clinical history of the patients with pancreatitis, the accuracy of the diagnosis, and concomitant medications. For example, we don’t even know whether pancreatitis diagnoses are based on elevated pancreatic enzymes or just on clinical symptoms of abdominal pain, which can often be confused with other diagnoses. Second, adverse events tend to be reported much more frequently early in a product launch, and this reporting bias makes comparisons between different drugs in AERS flawed. Furthermore, patients taking Byetta are likely to have more severe diabetes than those taking other diabetes medications since Byetta is indicated for patients who have failed oral drugs and are taking several concomitant medications. In fact, in its own review of a large payor database, Amylin reported a lower reported rate of pancreatitis with Byetta than the rate in type 2 diabetics, contrary to Birchenough’s findings. In addition, when the FDA updated the Byetta label in April 2006 (after they presumably had access to all of these reported cases), there were no additional Warnings or Precautions added to the label and there was no direct mention of pancreatitis under the Adverse Reactions section. This seems to imply that the FDA did not find a causal relationship between Byetta and pancreatitis. And, remember, Byetta got through the FDA without a panel review precisely because its safety profile was so pristine!
Also of note, of the 24 cases of pancreatitis that Birchenough found, only one of them was reported in 1Q06, which implies a decreasing incidence rate. Taken alone, the rate in 1Q06 is actually much lower than the rate of pancreatitis in the general population. This is contrary to what we would have expected since more people continue to start on Byetta over time. If pancreatitis is actually associated with initiation of Byetta treatment, we would have expected the rate of pancreatitis to increase over time as more people initiate therapy. This decrease in incidence rate may reflect better drug introduction since it coincides with the release of more information about adjusting meal size and composition when initiating Byetta. Alternatively, it may reflect more accurate diagnosis of pancreatitis. In any case, this issue clearly needs to be further analyzed.
Thomas Wei of PiperJaffray points out, “there are rare serious side effects associated with almost all diabetes drugs, including lactoc acidosis (and pancreatitis) with metformin, hypoglycemia with sulfonylureas and insulin, and edema/heart failure with TZDs, but all have been successful due to the overall benefit outweighing the risks. We continue to believe that Byetta’s differentiated clinical profile, including weight-loss, post-prandial glucose control, and restoration of first phase insulin release, support continued adoption.”
We will continue to monitor the situation and speak to thought leaders - the first one we called, Dr. Irl Hirsch gave us his bottom line tonight, “Rare side effects occur and aren’t seen until many people are on the drug. I can’t speak for anyone else, but it won’t stop me from using the drug in people who can benefit.” We welcome the focus on patients ...