The degree to which the oral rheumatology drug tofacitinib (Xeljanz) seems to promote new malignancies is now fleshed out in the latest, and at least somewhat reassuring, report from a landmark safety study.
Tofacitinib at either 5 or 10 mg twice daily in the so-called ORAL Surveillance trial was associated with incident cancers at a rate of 1.13 per 100 patient-years of treatment, compared with 0.77 per 100 patient-years with tumor necrosis factor (TNF) inhibitors, according to Jeffrey Curtis, MD, of the University of Alabama at Birmingham, and colleagues.
That worked out to a hazard ratio of 1.48 (95% CI 1.04-2.09) for tofacitinib versus TNF inhibitor therapy, the group reported in Annals of the Rheumatic Diseases.
Some of the data had been reported earlier — in a press release from tofacitinib maker Pfizer and in a New England Journal of Medicine paper this past January — but the new publication includes substantially more detail, including 50 pages of supplementary tables and figures on top of the 13-page main paper.
ORAL Surveillance began in 2014 under orders from the FDA, following the drug’s initial approval for rheumatoid arthritis in 2012. (It has since gained approvals for psoriatic arthritis, ankylosing spondylitis, juvenile idiopathic arthritis, and ulcerative colitis.) The registration trials showed signs that tofacitinib increased risks of certain malignancies, as well as infections, and its first label included boxed warnings about them. However, the agency wanted Pfizer to produce a clearer picture of the risks with a dedicated safety study. After ORAL Surveillance data first became available in 2021, the warnings were beefed up.
The new report now provides additional details on the malignancy risk, including rates of specific cancer types and numbers needed to harm as compared with TNF inhibitors.
Curtis and colleagues randomized 4,362 patients at a 1:1:1 ratio to tofacitinib at 10 or 5 mg twice daily or to one of two TNF inhibitors. Patients were enrolled at 323 sites in 30 nations. In North America, the assigned TNF inhibitor was adalimumab (Humira); etanercept (Enbrel) was used elsewhere. After an interim analysis in 2019 showed increased mortality for the higher tofacitinib dose, that arm was terminated and its patients were shifted to the lower dosage. Adalimumab and etanercept doses were the customary standards, and all patients continued on methotrexate.
Only individuals 50 and older with rheumatoid arthritis not responding to methotrexate were enrolled (approvals for the other previously mentioned indications all came after the trial was designed). Patients were also required to have at least one cardiovascular risk factor beyond age.
In total, 122 malignancies (not counting non-melanoma skin cancers) developed in the two tofacitinib groups — 62 with the lower dose and 60 with the higher dose — versus 42 in the TNF inhibitor arm.
Lymphomas were among the first cancer types to draw scrutiny, and the new data back that up: Curtis and colleagues calculated a hazard ratio of 5.09 for the two tofacitinib doses combined relative to TNF inhibitors. However, the 95% confidence interval was extremely broad and spanned 1.0 (0.65-39.78, to be precise) because there were only 10 cases total with tofacitinib and just one in the TNF inhibitor group. Lung cancer also appeared to be substantially more common with tofacitinib (HR 2.17, 95% CI 0.95-4.93), as did cutaneous squamous cell tumors (HR 2.32, 95% CI 1.08-4.99) and non-melanoma skin cancers overall (HR 2.02, 95% CI 1.17-3.50).
For cancers overall, the researchers found a number needed to harm of 55 with 5 years of tofacitinib treatment — that is, treating 55 patients for 5 years would lead to one extra cancer case relative to what would be expected with TNF inhibitors.
Kaplan-Meier curves showed that tofacitinib’s extra cancer risk became clearest after about 3 years of treatment. At year 5, some 97% of the TNF inhibitor group remained cancer-free, versus 94% of the tofacitinib groups combined. There was no obvious dose-relationship for tofacitinib and cancer rates.
Also noteworthy were findings that the following were associated with increased cancer risk, based on all three treatment arms combined:
- White race versus non-white race
- North American residence
- Current or past smoking versus never-smokers
- Higher versus lower cardiovascular risk score
- History of cardiovascular disease versus no history
- Increasing age
These factors generally applied to lung cancers and non-melanoma skin cancers, as well as to all malignancies excluding non-melanoma skin cancers. Curtis and colleagues cautioned that some of these findings were interrelated (e.g., participants in North America came into the study with generally higher cardiovascular risk scores) or could have been influenced by use of medications such as non-steroidal anti-inflammatory drugs, which may carry cancer risks of their own.
Limitations cited by the authors included that the trial was not designed to assess differential risks for specific cancer types, and that cardiovascular risk scores included a fixed 1.5-fold increase in the statistical analysis to account for the presence of rheumatoid arthritis, which Curtis and colleagues called “a crude approximation.”
The trial was sponsored by Pfizer and required by the FDA.
Study authors reported extensive relationships with pharmaceutical companies including Pfizer and other commercial entities.