A Hypothetical Case

A high school student with severe, treatment-resistant atopic dermatitis was brought to a dermatology clinic by her parents for evaluation of a rash. During the consultation, the patient was crying, reportng an unbearable itch and rash causing sleep-deprivation. She has previously failed many topical and biologic agents. The dermatologist recommended starting a JAK inhibitor and informed them of the list of boxed warnings found in the package insert regarding serious infection, mortality, malignancies, major adverse cardiovascular events (MACE), and thrombosis. The patient expressed a strong desire to proceed with the JAK inhibitor given the impact on her quality of life, but her parents were afraid of these side effects. In the end, they left without starting a JAK inhibitor; the patient appeared dejected with no hope for relief in sight.

What are FDA “Class Label” Warnings?

For any practicing dermatologist or dermatology provider, it is important to understand what FDA “Class Label” Warnings mean. The FDA is allowed to apply warnings on any medication in the same class as ones which had a side effect, even if they are not supported by evidence for the other medications. Interestingly, a majority of the time, the warnings contraindicate the findings of the actual clinical trial data on individual medications; however, as long as the FDA considers the drug to be “in the same class” as another medication which has such proven side effects, the FDA is able to apply the same warnings to that drug. Therefore, they are “Class Label” Warnings. Moreover, the FDA is not required to be transparent; they do not have to inform clinicians or consumers that these warnings are not supported by evidence. For example, all of the boxed warnings for upadacitinib and abrocitinib, commonly used for atopic dermatitis, came from the Oral Surveillance Study which was conducted on tofacitinib for treatment of rheumatoid arthritis. This study involved only older patients with at least one serious cardiovascular risk factor; the FDA has conducted the Oral Surveillance Study in this way to “enrich the risks” because even with tofacitinib for rheumatoid arthritis, all of the risks spelled out in the current boxed warnings could not be substantiated with data when ordinary rheumatoid arthritis patients were allowed to be enrolled. Moreover, this critical study did not include dermatology patients, nor tested upadacitinib or abrocitinib.

The above hypothetical case is particularly unfortunate, as the dermatologist may have been unaware that all five of the boxed warnings were not evidence-based; in fact, all of them were refuted by the actual worldwide Phase 3 Clinical Trial test results, except for herpes which was increased in risk. However, the FDA does not typically issue boxed warnings solely for increased risk of herpes. This is proven as a fact when the FDA originally approved upadacitinib for rheumatoid arthritis in August 2019; despite the increased risk for herpes, upadacitinib did not have any boxed warnings in 2019. As of December 2021, all five of the boxed warnings were applied together to upadacitinib based on the Oral Surveillance Study which was concluded around that time. The fact about upadacitinib having no boxed warnings when it was originally approved by the FDA in 2019 is brought up here to dispel any notion that the FDA would have given upadacitinib a boxed warning for serious infections based on an increase in herpes risk, because that in fact did not take place.

What is the Actual Evidence or Lack of Evidence for Boxed Warnings for JAK inhibitors for Atopic Dermatitis?

In order to determine if there is a real risk of side effects, three approaches are usually employed: 1) The result of “head-to-head” direct comparison against placebo at the end of the placebo-controlled period. 2) What serious side effects were documented during the “one-year period”, which is the first year of the Worldwide Phase 3 Clinical Trial. The results from the first year is particularly important because the data from that time period is what the FDA uses to approve or reject the medication.1,2,3 3) What happens beyond the first year of the Worldwide Phase 3 Clinical Trial.4,5 The FDA is particularly interested in any serious, unexpected side effect that might emerge during this long term extension phase of the study.

The Actual Side Effect Data for Upadacitinib and Abrocitinib in Atopic Dermatitis Trials

There are five categories of boxed warnings on JAK inhibitors for atopic dermatitis. In this section we will examine if these warnings are evidence-based or not for upadacitinib and abrocitinib.1,2,3

1) Serious infection

The first line in the boxed warning is quite intimidating. It states, “increased risk of serious bacterial, fungal, viral and opportunistic infections leading to hospitalization or death, including tuberculosis.” Based on 2,960* upadacitinib patients and 2,856* abrocitinib patients followed for approximately** one year, there was one death from serious infection in the first year of the Worldwide Phase 3 Study for upadacitinib due to bone tuberculosis. Apart from the above case of bone tuberculosis, there were no other clinical cases of tuberculosis (see the case below under “Mortality” for detail). There was one death in an abrocitinib patient with COVID-19, but this case was not considered related to the study medication (see below under “Mortality”). Moreover, at the end of the placebo-control period (up to 16 weeks of clinical trials), for the highest dose of both upadacitinib and abrocitinib, the rate of serious infection was notably lower than the placebo group’s serious infection rate (serious infection rate for upadacitinib for 30 mg/day was 0.4% compared to placebo group rate of 0.6%) (for abrocitinib, the serious infection rate for the highest dose of 200 mg/day was 0.4% while the placebo rate was 0.7%). Interestingly, for both medications, the lower dose serious infection rate was slightly higher than the placebo rate. Thus, there was no dose response relationship. With regards to herpes, as mentioned previously, the rate of herpes simplex, herpes zoster and eczema herpeticum are increased compared to placebo with these agents (only for abrocitinib, the highest dose group had no case of eczema herpeticum at the end of placebo-controlled period, unlike the placebo group and low dose group). Lastly, for upadacitinib, there were 3 cases of pneumonia but all patients recovered. Two of the three patients continued upadacitinib while for one 14-year-old pediatric patient, following recovery, the patient and the family elected to discontinue the study. In abrocitinib-treated patients, there were 6 cases of pneumonia.

2) Mortality

For upadacitinib, there were 2 deaths during the first year out of 2,960 researched subjects, both of which were judged to be unrelated to the study medication. One 71-year-old man died of multiple medical problems including esophageal fistula, COPD, cardiac arrhythmia, pericarditis and sepsis. Another 69-year-old woman died from bone tuberculosis; upadacitinib was discontinued one month before her death. Given its half-life of 14–15 hours, upadacitinib was no longer present in her system when this death occurred.

For abrocitinib, there were four reported deaths out of 2,856 research subjects, none of which were considered related to the study medication. One case involved a 78-year-old woman who died of gastric adenocarcinoma diagnosed on Day 43; the patient had symptoms prior to joining the study. Another case involved a 69-year-old African American woman who died of COVID-19, as stated above. A third case involved a 73-year-old woman who died of sudden death, which occurred after 22 days of discontinuation with abrocitinib. The half-life of abrocitinib is only 3-5 hours, and once again, because of a short half-life, abrocitinib was no longer in her system when this event happened. Lastly, a 42-year-old man died of cardiac failure.

3) Malignancies

During the first year of Worldwide Phase 3 testing of upadacitinib, a 53-year old woman was thought to have CTCL however multiple skin biopsies failed to confirm CTCL. She was allowed to continue the upadacitinib study until Day 194, when another skin biopsy proved the diagnosis of CTCL and was reported as lymphoma.

Six cases of malignancy other than non-melanoma skin cancer were reported out of 2,960 Phase 3 patients and were not considered related to the study medication. For abrocitinib, there were 3 internal malignancies during the first year, two of which were prostate cancers in elderly men. One patient was 73 years old and the other was 68 years old. Lastly, a 78-year-old woman died of gastric adenocarcinoma as described above, which was preexisting her study start date. No lymphoma or any other cases of malignancy, excluding non-melanoma skin cancer, were documented out of 2,856 patients in approximately** the first year of the Worldwide Phase 3 Study.

4) Major Adverse Cardiovascular Events (MACE)

For upadacitinib (n=2,960), there were three cases of MACE in the first year, none of which were judged to be related to the study medication. For abrocitinib (n=2,856), there were four cases of MACE, all of whom had multiple cardiovascular risk factors. Three of the four cases were listed in available publications, including a 73-year-old woman with sudden death 22 days after discontinuing the study medication (as reported above), an obese 60-year-old man who died of MI on Day 104, and a 63-year-old man with many risk factors who died of MI on Day 159. Lastly, a 42-year-old man died of cardiac failure on Day 334, was not captured in the previously published manuscripts because the event occurred several months after the publication of the initial data.

5) Thrombosis

For upadacitinib, the placebo group had the highest rate of thromboembolic events. There was a case of pulmonary embolus within the first year in the placebo group, occurring on Day 55. For the upadacitinib group, the lower dose group who took 15 mg of upadacitinib per day had no case of thromboembolic events in the first year. For the higher dose upadacitinib group who took 30 mg per day, there was one case of DVT documented on Day 365. The patient had a history of DVT and pulmonary embolus prior to joining the upadacitinib clinical trial. Moreover, this patient also had a family history of DVT in his mother. In addition, this patient also had chronic lower leg swelling, obesity, hypertension, and was a smoker. Even though there was one case of pulmonary embolus in the placebo group and one case of DVT in the upadacitinib group during the first year of Worldwide Phase 3 Study, because there were much more patients on the active arm taking upadacitinib compared to placebo arm which not only had a much less number of patients but of shorter duration, the rate of thromboembolic event turned out to be the highest with placebo group compared to either the 15 mg or 30 mg dose groups for upadacitinib.

For abrocitinib during approximately** the first year of the Worldwide Phase 3 Clinical Trial, out of 2,856 patients, there were 2 cases of pulmonary embolism, both in older women, one of whom was on estrogen replacement therapy which also increases the risk of thrombosis. In addition, there were 2 cases of DVT, one of which was thought to be unrelated to abrocitinib because one DVT occurred in association with arthroscopic knee surgery. The other case was a 50-year-old woman with a BMI of 33.7 kg/m2 and a history of hypertension who has experienced superficial thrombophlebitis confirmed with Doppler ultrasound.

Lastly, as of the writing of this manuscript, both upadacitinib and abrocitinib, the open-label long-term extension data—now approaching half a decade since the start of the clinical trials—have not revealed any significant support for the boxed warnings.2 However, the space limitation for this publication precludes presenting the year-by-year data in great detail.

Returning to the Hypothetical Case

Although the patient and her family initially declined treatment with a JAK inhibitor due to concerns about the boxed warnings, the severity of her treatment-resistant atopic dermatitis, which significantly impaired her quality of life, ultimately led them to reconsider and contact the dermatologist again. Meanwhile, prior to recontacting, the dermatologist came across this article explaining the difference between the FDA’s “Class Label” Warnings versus evidence-based warnings. This new understanding alleviated some of the dermatologist’s own concerns regarding JAK inhibitors and enabled a more informed discussion with the patient and her family. The dermatologist was able to explain that, based on available clinical trial data, JAK inhibitors, like upadacitinib and abrocitinib, represent a reasonably safe therapeutic option. With a better understanding of the actual risks, the patient and her family felt reassured to pursue treatment without undue fear.

Conclusion

As the results section of this article shows, out of almost 6,000 total patients (approximately 3,000 for each of the two JAK inhibitors) followed for approximately one year, there are only a few documented events that pertain to the boxed warnings. However, the reality of medical practice is that most clinicians face significant time restraints and have difficulty finding the time and effort to independently review original published clinical trial data to appropriately judge whether boxed warnings issued by the FDA are supported by evidence or not. Moreover, the fact that the FDA is not required to reveal to clinicians that even the most serious warnings may be not supported by evidence makes it almost impossible for a busy clinician to figure out which one of the serious warnings are evidence-based and which of them are not. The authors are not ruling out the possibility that some unknown serious risk may still exist, and therefore understand the importance of the currently ongoing long-term open-label extension study to ascertain such a possibility. While the authors acknowledge that it is within the FDA’s mandate to issue safety warnings as it deems appropriate, more transparency would be critically helpful in guiding clinical decision-making and ensuring that patients are not denied desperately needed care. It is noteworthy that government regulatory agencies far from the US FDA such as the FDA equivalent for the European Union (EU) or Japan do not subscribe to the practice of the US FDA to issue even the most serious warnings that are refuted by evidence; therefore, these five categories of boxed warnings do not exist in the EU or Japan for JAK inhibitors for atopic dermatitis as in the US.6 Lastly, it is noteworthy that the way the US FDA issues boxed warnings are not always consistent or logical. For example, topical ruxolitinib has all five boxed warnings discussed here, but oral ruxolitinib with much higher blood levels has no boxed warnings.7 Therefore, the authors hope this article helps clinicians better understand the data-driven reality of these medications, in contrast to the FDA’s boxed warnings, which may be speculations and could provoke undue fear.

Asterisks

*The total number of research subjects may differ from publication to publication, depending on when the data cut was conducted.

**The original data cut was not necessarily analyzed exactly at one year because of the nature of rolling enrollment to the study. Therefore, there can be some slight variations in the time period of the subjects.