Recent research has suggested that tumor necrosis factor inhibitors can significantly reduce diabetes risk in people with rheumatoid arthritis.
Dr. Siri Lillegraven of Diakonhjemmet Hospital in Oslo presented results from the CORRONA (Consortium of Rheumatology Researchers of North America) study, which Dr. Lillegraven, its lead author, called "the first large study to find the same association" at the annual European Congress of Rheumatology. The study used CORRONA registry data from 22,943 patients and about 22,000 RA treatment regimens with a mean duration between 1.5 and 2.4 years.
Dr. Lillegraven and her colleagues found an adjusted hazard ratio for type 2 diabetes of 0.35 for TNF inhibitors (95% confidence interval, 0.13-0.91), compared with a reference group of non-methotrexate, non-hydroxychloroquine, nonbiologic disease-modifying antirheumatic drugs such as cyclosporine, sulfasalazine, and leflunomide. The DMARDs hydroxychloroquine and methotrexate were separately compared with this reference group.
"It was a statistically significant finding, and the model was adjusted for differences between patients who received TNF inhibitors and the patients who received the comparator drugs," Dr. Lillegraven said in an interview.
One of Dr. Lillegraven’s coauthors on the study, Dr. Daniel Solomon of Brigham and Women’s Hospital in Boston earlier reported a lower risk of type 2 diabetes for individuals taking TNF inhibitors or hydroxychloroquine, compared with nonbiologic DMARDs (JAMA 2011;305:2525-31).
In that study, which enrolled about 14,000 patients and evaluated about 22,000 treatment episodes, the multivariate adjusted hazard ratio for diabetes was 0.62 (95% CI, 0.42-0.91) for TNF inhibitors, compared with other nonbiologic DMARDs. The effect was even greater for hydroxychloroquine, compared with other nonbiologics (hazard ratio, 0.54; 95% CI, 0.36-0.80).
Dr. Lillegraven and her colleagues saw a similar effect size for hydroxychloroquine, compared with the nonbiologic DMARDs, but this did not reach statistical significance.
As both studies were observational in design, Dr. Lillegraven noted, the results do not carry the weight of randomized, controlled trial findings. "We would have loved to have a clinical trial that confirmed the findings," she said, adding that designing such a trial would be difficult. "The outcome is relatively rare, and you will not likely get enough diabetes outcomes to be able to conclude whether an exposure had an effect." In Dr. Lillegraven and her colleagues’ study, for example, only 84 incident cases of diabetes occurred.
Last year, investigators reported that TNF inhibitors were associated with a halving of diabetes risk, compared with RA patients who had never used them (HR, 0.49; 95% CI 0.24-0.99), in a cohort of 1,587 RA patients without diabetes at enrollment who were followed for 3-4 years (Arthritis Care Res. 2012;64:215-21), and several studies have suggested a relationship between the biologic pathways that TNF inhibitors affect and diabetes.
Dr. Lillegraven also analyzed the impact of body-mass index (BMI) and steroid dosage on diabetes incidence in these patients. Those with a BMI of more than 30 kg/m2 had a statistically significant sixfold increased rate of incident diabetes, compared with patients with a BMI of less than 25 kg/m2. Patients with a BMI of 25-30 kg/m2 had a significant, nearly twofold increased rate. Patients who received a steroid dose of at least 7.5 mg/day had a statistically significant, twofold increased diabetes incidence, compared with patients who did not receive any steroid treatment.
Dr. Lillegraven said that her study’s findings, added to the earlier findings, support "a potential for tailoring treatment in high-risk individuals." But it is still too early to draw any definite conclusions regarding how this should be carried out in the clinic, she cautioned.
Dr. Lillegraven declared no conflicts of interest relevant to her study. Dr. Solomon declared unpaid consultancies for Pfizer and Novartis. Three other coauthors reported financial relationships with pharmaceutical firms and CORRONA, a database registry for rheumatologic diseases; one is an employee of CORRONA.