Avoiding all adverse drug reactions not realistic

0331-Drug Reactions_Murad

Murad Alam, MD: ‘When the FDA approves a product, the label is usually whatever the manufacturer wants. The label is based on trials and the manufacturer sets the criteria.’

Adverse drug reactions are unavoidable. The current system of drug approvals cannot catch any but the most common side effects.

“Most drugs are quite safe,” said Joel Gelfand, MD, associate professor of dermatology and epidemiology, medical director of the clinical studies unit and senior scholar at the University of Pennsylvania Center for Clinical Epidemiology and Biostatistics. “But when drugs are used by millions of patients, even rare side effects can affects thousands of people and make headlines.”

Dr. Gelfand was the course director for a March 22 session, “Adverse Drug Reactions: The Evidence and the Controversies.” He also explored the potential sources drug safety problems. A key area is the growing use of drug therapy.

Patients in the U.S. filled nearly 4 billion prescriptions in 2009, a 40 percent increase from annual drug usage in the prior decade. One result is improved health outcomes from more effective drug therapies. Another result is 1.5 million hospitalizations for adverse drug reactions and 100,000 deaths each year.

“When you have 3,000 patients in your study, you can only begin to detect problems that occur in more than one patient per 1,000,” Dr. Gelfand said. “So it should be no surprise that safety issues routinely appear after a drug is approved. We should never expect approval trials to catch all, or even most, side effects.”

More than half of approved drugs have serious adverse effects not detected before approval, he added. Some 7.5 percent of drugs have black-box warnings added after approval, and 2.7 percent are withdrawn for safety reasons.

Stevens-Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN) are almost exclusively adverse drug reactions. The highest risk drug culprits — antibacterial sulfonamides, anticonvulsants, oxicam NSAIDs, allopurinol, lamotrigine and nevirapine — have been long been recognized. But the incidence is low, between 1 and 2.6 per million person years for SJS and 0.4 to 1.2 per million person years for TEN.

“This is the one dermatologic disease that other doctors know about,” said Misha Rosenbach, MD, assistant professor of dermatology & internal medicine and director of inpatient dermatology, Perelman School of Medicine at the University of Pennsylvania, said of TEN. “You have to intervene early and intensively.”

SJS involves less than 10 percent of epidermal detachment and has about 5 percent mortality. TEN has more than 30 percent epidermal detachment and 30 percent mortality. The two overlap in patients who have lost between 10 percent and 30 percent of their skin.

The most important step in treatment is to withdraw the culprit drug as early as possible. The best results follow drug withdrawal no later than the day blisters first appear. The longer treatment is delayed, the higher the morbidity and mortality, Dr. Rosenbach said.

Patients should immediately be moved to an ICU or burn center for supportive care and systemic therapy, she said.

Supportive therapy is important to prevent long-term sequelae. Patients need temperature control, aseptic and sterile handling, non-adhesive bandaging, hydration control with moisture-retentive ointments on skin erosions, nutritional support, and a daily ophthalmic exam to minimize ocular complications.

There are conflicting data on systemic treatment options. European clinicians typically rely on cyclosporine while U.S. clinicians favor high-dose intravenous immunoglobulin (IVIG). Plasmapheresis holds promise and thalidomide is contraindicated. There are limited case reports and case series suggesting newer biologic TNF-inhibitors may be useful.

“Prevention is far better than treatment,” Dr. Rosenbach said. Clinicians should be alert for SJS and TEN when any of these problematic drugs are used.”

Safety also is an issue in cosmetic dermatology. All three forms of botulinum toxin A approved for use in the U.S. are packaged in single-use vials that the label says should not be stored for more than 24 hours after reconstitution. But a recent survey of American Society for Dermatologic Surgery members found that two-thirds of respondents store the reconstituted drug for a month or longer and use vials for multiple patients.

“What many clinicians may not realize is that the evidence base for the label is very thin,” said Murad Alam, MD, professor of dermatology, otolaryngology, and surgery, and chief of cutaneous and aesthetic surgery at Northwestern University Feinberg School of Medicine. “When the FDA approves a product, the label is usually whatever the manufacturer wants. The label is based on trials and the manufacturer sets the criteria.”

A recent ASDS consensus statement noted that botulinum toxin A products can be refrigerated for at least one month after reconstitution or frozen and thawed before use. There have been no reports of adverse events from using product that has been stored for more than 24 hours after reconstitution, Dr. Alam said, and no difference if the product included preservative or was preservative-free.

Could a patient registry provide more data? It depends.

Registries, case cohort studies, and spontaneous reports all can provide safety data, said Carl Herzog, MD, professor of dermatology at Harvard Medical School and chair of dermatology at Beth Israel Deaconess Medical Center. And all must be used cautiously.

Medwatch, the FDA spontaneous reporting system, provided useful safety signals for progressive multifocal leukoencephalopathy in early biologics, he noted. Medwatch has been less useful in assessing the risk of IBD in isotretinoin use because more than 90 percent of adverse-event reports were submitted by attorneys with a financial stake in the outcome.

He expressed similar concerns over conflicts of interest in registries and cohort studies. Both can be helpful in assessing safety when executed and documented well. And both can be misleading if designed or executed poorly.

“There may or may not be sufficient drug exposure or adequate follow-up to give relevant results,” Dr. Herzog caution. “Registries and cohort studies can help assess safety, but they are not foolproof.”

Return to index