Do TNF Inhibitors Really Worsen RA-Associated Lung Disease?

Worries that tumor necrosis factor (TNF) inhibitors may exacerbate or even cause lung disease in patients with rheumatoid arthritis (RA) are unfounded, at least for U.S. military veterans, a new study suggested.

Serious respiratory complications were no more common among 237 RA patients in the Veterans Affairs (VA) health system who started TNF inhibitors compared with 237 otherwise similar patients treated with other “advanced” drug types, according to Bryant England, MD, PhD, of the University of Nebraska Medical Center in Omaha, and colleagues.

In fact, rates of adverse events such as respiratory hospitalization, death from lung disease, and all-cause mortality were numerically higher in the non-TNF inhibitor users — hazard ratios ranged from 1.15 to 1.38 — but these increases fell far short of statistical significance, the group reported in Lancet Rheumatology.

Comparator drugs in the analysis included Janus kinase (JAK) inhibitors such as tofacitinib (Xeljanz) and biologics with non-TNF targets, including abatacept (Orencia), tocilizumab (Actemra), and rituximab (Rituxan), among others.

“These findings do not support the systematic avoidance of TNF inhibitors in the management of patients with rheumatoid arthritis-associated ILD [interstitial lung disease],” the researchers wrote.

In an accompanying commentary, two independent experts noted several limitations to the analysis, but agreed with the study authors’ conclusion that TNF inhibitors seem safe enough for RA patients with lung complications.

Additionally, wrote Gregory McDermott, MD, and Jeffrey Sparks, MD, MMSc, both of Brigham and Women’s Hospital in Boston, the results “raise important questions about the complex interplay between rheumatoid arthritis disease activity, immunosuppression, and rheumatoid arthritis-associated ILD outcomes.”

At issue is the balance between the effects of treatment on ILD and those of RA itself; “choosing an effective therapy could plausibly be more important than the specific medication choice,” the pair observed. “Future research addressing this question is needed.”

Prompting the new study were previous reports, mostly anecdotal, indicating that risk of ILD incidence and progression appeared to be higher in patients receiving TNF inhibitors as opposed to other medication types. However, England and colleagues noted, these had serious drawbacks: “patients who receive TNF inhibitors typically have more severe rheumatoid arthritis than patients who receive conventional DMARDs [disease-modifying anti-rheumatic drugs].”

One more systematic study from Japan, with 163 RA patients, found that ILD was more prevalent among patients on TNF inhibitors and that biomarkers for lung disease increased more with these agents.

Thus, there were suggestions that perhaps TNF inhibitors should be avoided in RA patients with lung involvement, even though some studies had looked for differences and failed to find them.

For their effort, the researchers checked VA system records for individuals with diagnoses of RA-associated ILD, focusing on those receiving either TNF inhibitors or an alternative targeted drug for RA. Those also receiving medications for ILD, such as azathioprine or nintedanib (Ofev), were excluded. Ultimately, this yielded 237 eligible patients on anti-TNF agents, who were matched via propensity scoring with 237 others who were given other targeted therapies.

As in most VA studies, the included patients were overwhelmingly men (92%). Their mean age was 68, and they had lived with RA for an average of 7 years; some 80% were white. Nearly half were current smokers and another 36% were ex-smokers. More than three-quarters were overweight or obese.

Mean ILD duration was 3 years, and about one-quarter used supplemental oxygen at home. Among those with forced vital capacity values, most showed impairment; it was less than 60% of predicted in just over 20% of patients, and 37% had values of 60% to 80%.

A composite outcome of death or respiratory hospitalization occurred in 43% of the TNF inhibitor group over 3 years of follow-up, versus 47% of those using other medications (HR 1.21, 95% CI 0.92-1.58). Hazard ratios for individual outcomes were as follows:

• All-cause mortality: 1.15 (95% CI 0.83-1.60)
• Respiratory hospitalization: 1.27 (95% CI 0.91-1.76)
• Respiratory death: 1.38 (95% CI 0.79-2.42)

Limitations to the analysis included, above all, the predominance of male patients (typically, 65%-85% of RA patient samples are women) and their history of military service. The study also relied on administrative data that could be faulty, and some data (such as forced vital capacity) were missing for many patients. McDermott and Sparks also noted that the sample size was relatively small, introducing substantial uncertainty in the hazard estimates.

Please enable JavaScript to view the comments powered by Disqus.