The Value of the Black Box Warning in Dermatology

majority of drugs are approved based on data from relatively short clinical trials conducted in specific patient populations that may not represent their ultimate market and clinical use. In addition, initial study data on novel therapeutic agents may change as newer, longer-term data demonstrates differing results. There is always the potential for new adverse events (AEs) to emerge, and post-marketing safety studies and passive surveillance using Adverse Event Reporting System (AERS) data collection are employed to monitor this risk.¹ This computerized database combines MedWatch, which is based on voluntary reporting by health care providers, and FDA-mandatory reports from pharmaceutical companies, that despite being mandatory are nonetheless comprised of spontaneously reported AEs. Additional safety information may come from FDA-conducted analyses of databases with information linking drugs to adverse events, commercial databases purchased by the FDA, as well as new case reports and clinical trials in the medical literature.⁶

It is well-known that AEs are underreported by both health care providers and pharmaceutical sponsors, yet AERS relies solely on spontaneous reporting.⁷ The majority of boxed, or “black box” warnings, however, are based on this postmarketing surveillance, rather than on data from randomized controlled clinical trials. Today, this term has evolved to describe the most serious type of marketed prescription drug warning, the boxed warning.^8,9

Approximately 400 drugs in the United States have a boxed warning and the number is growing, despite a relatively stable number of drug withdrawals.^10,11 It is estimated that the likelihood that a drug will acquire a new boxed warning or be withdrawn from the market over 25 years is 20%.¹² For dermatologists, the topical calcineurin inhibitors (TCIs) are a well-known class of drugs that carries a boxed warning, primarily based on animal toxicity and carcinogenic potential.13 A decade after their initial approval, the manufacturers of onabotulinumtoxinA (Botox® and Botox Cosmetic®) were required to add a new boxed warning in 2009, notifying prescribers and patients of the potential for “distant spread of toxin effect,” which was based on data in children with cerebral palsy being treated off-label for spasticity. ¹⁴ Table 1 lists common drugs used in dermatology that carry a boxed warning.¹⁵

As evidenced by the preceding examples, there is no clear metric to determine how and when a boxed warning is applied. Inconsistencies in the review process, language, timing, and dissemination of these warnings therefore impact both dermatologists and their patients.

The boxed warning criteria were released by the FDA in 1979. Despite the vague nature of these criteria, they remain valid to this day: “The boxed warning ordinarily shall be based on clinical data, but serious animal toxicity may also be the basis of a boxed warning in the absence of clinical data.”⁹ Of note, a causal relationship between the drug and the adverse event does not have to be proven.⁸ In 2006, further guidelines were released that provided “nonbinding recommendations” for the application and use of boxed warnings, but did not provide any specific criteria for their issue.^16,17 The FDA describes the boxed warning as a means to alert health care providers of three general situations: 1) when the risk of an AE is so serious that it may outweigh the benefits of a drug (eg, life-threatening, fatal, or can cause permanent disability), 2) when serious AEs may be prevented or their risk decreased with appropriate prescribing considerations including lab monitoring and suitable patient selection, and 3) the FDA approved the drug with mandatory restrictions or guidelines for safe use.¹⁷ Boxed warnings may also be issued in other unspecified circumstances and in instances in which there may be an anticipated adverse reaction, such as in the contraindication of drugs during pregnancy due to evidence in humans or animals. ¹⁷ The boxed warning about the potential risk of cancer with the use of TCIs was based on their mechanism of action, animal studies, and postmarketing surveillance demonstrating 20 case reports of lymphoma in patients who had used TCIs worldwide.¹⁸

When a drug is identified to be high risk, the FDA commissioner may refer it to the Center for Drug Evaluation and Research (CDER), which may convene an advisory committee for further review.¹⁹ However, boxed warnings are not required to undergo this process prior to implementation and are often issued without this type of evaluation. In the three-year time period from 2004-2006, 77 new black-box warnings were released, of which only 11 were discussed by advisory committees.²⁰ Advisory meeting transcripts often indicated confusion regarding the boxed warning and an emphasis on the potential impact of the warning rather than the content.²⁰ Same-class drugs do not all share the same safety information including a boxed warning.²¹ Furthermore, the median time for a black box warning to appear on the label of another drug in the same class is 66 months (2-170 months).²¹

The language of boxed warnings is also variable. The wording of each warning is negotiated between the multidisciplinary FDA group, which includes pharmacologists, chemists, medical officers, and statisticians, who review scientific data and negotiate specific drug labeling with drug manufacturers.²² A study evaluating the informativeness of warnings for drugs causing