The limitations of the drug approval process in fully characterizing a new drug's safety profile are clearly illustrated in the dermatologic therapeutics arena and clinicians should keep these limitations in mind when prescribing and counseling patients about a relatively new therapeutic agent, according to Dr. Joel M. Gelfand.
Previously unknown risks of new drugs are routinely identified after Food and Drug Administration approval, when "rare" adverse events are more likely to be detected, said Dr. Gelfand at the Hawaii Dermatology Seminar sponsored by Skin Disease Education Foundation (SDEF). Examples of rare adverse events – those that occur at a rate of less than 1 per 1,000 – include fatal arrhythmias associated with terfenadine and astemizole, lymphomas associated with biologics, and progressive multifocal leukoencephalopathy (PML) associated with efalizumab.
Dr. Joel M. Gelfand
The clinical trials that are the basis of drug approvals evaluate short-term safety only and lack the statistical power to detect rare events, he said, pointing out that a study of 3,000 patients can only detect adverse events that occur at a rate of more than 1 per 1,000.
The types of adverse effects detected pre-approval are pharmacologic side effects, which are common and dose dependent, such as isotretinoin-induced cheilitis, said Dr. Gelfand, of the department of dermatology at the University of Pennsylvania in Philadelphia.
Adverse effects that are detected after approval fall into two categories: idiosyncratic or allergic reactions, which are rare and "occur in close proximity to exposure," he said, such as dapsone-induced agranulocytosis, and new morbidities, which are delayed and uncommon. An example of the latter is skin cancers associated with psoralen and ultraviolet A (PUVA) treatment, which was associated with melanoma in 1997, more than 2 decades after it was found to be effective for treating psoriasis.
There is a need for ongoing risk assessment throughout the life cycle of a drug, Dr. Gelfand said, which includes MedWatch, the FDA's voluntary adverse event reporting program that relies on spontaneous reporting of adverse events. MedWatch's advantages include that it is inexpensive and can identify safety signals; however, it is limited by under-reporting and MedWatch reports usually cannot be used to determine causality, he noted.
The association between efalizumab (Raptiva) and PML, an untreatable and often fatal central nervous system infection that occurs primarily in the setting of immunosuppression, is an example of "signal detection in action," Dr. Gelfand said. Efalizumab was approved in 2003 for psoriasis, based on studies of about 2,700 people, including 218 treated for more than 1 year. By 2008, after 46,000 people had been treated, including 3,000 for at least 2 years, there had been 3 confirmed cases and one suspected case of PML. All four were spontaneous reports, which "can be useful for very rare diseases such as PML," he said.
With efalizumab treatment, the overall estimated risk of PML is 1 in 15,000 patients per year and one in 1,000 patients treated for more than 2 years. These are likely underestimates because of incomplete reporting, and is a relationship that is "likely causal," he said. In the case of efalizumab, the risk was "judged unacceptable given treatment alternatives and disease indication" and the drug was withdrawn from the market in 2009.
A causal association between isotretinoin and inflammatory bowel disease (IBD) has not been established, but the drug's prescribing information includes a warning about the association.
The data on isotretinoin and IBD include a large administrative claims database of over 8,000 cases of IBD, and 21,832 controls published in 2010 (Am. J. Gastroenterol. 2010;105:1986-93). Analysis of the database found that the risk of IBD within 12 months of being treated with isotretinoin was 1.7; for ulcerative colitis, the risk was increased fourfold; for Crohn's disease the risk was slightly reduced.
In that study, which controlled for age, sex, and geographic region, the dose response was evident for ulcerative colitis only, Dr. Gelfand said.
In a review of 85 spontaneous reports of IBD in isotretinoin patients to the FDA between 1997-2002, isotretinoin was considered "highly probable" as the cause in 4 cases (5%), "probable" in 58 cases (68%), and "possible" in 23 cases (27%). These included three cases with a positive dechallenge and rechallenge.
When in comes to ulcerative colitis, the overall data available "suggest a specific association" between isotretinoin and the disease, but the data are conflicting, he said. Studies have not addressed the possibility that patients with severe acne are at an increased risk of ulcerative colitis, or of confounding variables including oral antibiotics and smoking, he said. If the risk is real, it is small, with a number needed to harm that exceeds 3,300.