Patients with RA were at lower risk for developing incident type 2 diabetes mellitus (T2DM) in comparison with patients with hypertension, psoriatic arthritis (PsA), or osteoarthritis, as well as the general population without RA in a retrospective cohort study of a large, nationwide, commercial health insurance claims database.
This result goes against what the study researchers from the division of pharmacoepidemiology and pharmacoeconomics at Brigham and Women’s Hospital and Harvard Medical School, both in Boston, initially hypothesized: The “risk of incident T2DM in RA patients would be similar to or less than PsA and [hypertension] patients, but higher, compared to general non-RA and OA patients.”
Prior epidemiologic studies of the relationship between RA and incident diabetes have yielded inconclusive results suggesting a small increase or no increase in risk of T2DM in patients with RA, possibly because of differences in the risk of T2DM in comparison groups used by previous studies to calculate relative risk, first author Yinzhu Jin and colleagues noted in their report published in Arthritis Care & Research.
After mining a nationwide U.S. commercial health insurance claims database, the Optum Clinformatics Data Mart, for claims data from Jan. 1, 2005, to Dec. 31, 2017, the researchers matched a total of 108,568 patients in RA, general population non-RA, hypertension, and OA cohorts based on age, sex, and index date (the date of disease-specific medication dispensing). Overall, 77% of those patients were female and had a mean age of nearly 56 years, whereas 48% of patients with PsA were female and their mean age was nearly 49 years. (PsA patients were not matched because of smaller numbers.)
During a median follow-up period of 1.4-1.8 years across the comparison groups, the crude incidence rate for diabetes per 1,000 person-years in the cohorts was 7.0 for RA, 7.4 for general non-RA, 12.3 for hypertension, 7.8 for OA, and 9.9 for PsA. The hazard ratios and 95% confidence interval for risk of diabetes in patients with RA – after adjustment for more than 40 baseline covariates that included demographics, comorbidities, medication use, and health care utilization – was 0.72 (0.66-0.78) in comparison withh the general non-RA cohort, 0.65 (0.60-0.71) in comparison with the hypertension cohort, 0.75 (0.69-0.81) in comparison with the OA cohort, and 0.76 (0.67-0.86) in comparison with the PsA cohort. These values correspond to RA patients having a 24%-35% lower risk of incident diabetes versus the comparison groups, the researchers noted. They observed results consistent to these when they conducted a sensitivity analysis using a 1-year lag time from the index date before starting follow-up.
The lower risk of T2DM in patients with RA in comparison with patients in the non-RA cohort “may be, in part, due to the effect of biologic DMARD [disease-modifying antirheumatic drug] treatment in RA which likely modifies the risk of DM,” the researchers wrote. “Both the increasing use of biologic DMARDs for RA in the U.S. over the last decade and our cohort entry criteria for the RA cohort (i.e., at least one dispensing of a DMARD) may explain the finding of the lower risk of DM in RA.”
The results found with the other three cohorts did not surprise the researchers. The reduced risk of diabetes among RA patients versus those with OA jibes with “higher rates of obesity and other comorbidities in patients with OA” as well as findings from a recent study that found a higher incidence rate of diabetes in OA, compared with RA. Ms. Jin and colleagues also acknowledged it is well known that “hypertension and PsA are associated with metabolic dysregulation and increase the risk of diabetes.”
The researchers defined patients with RA as having at least twoinpatient or outpatient ICD-9 or ICD-10 diagnosis codes of RA, separated by 7-365 days and having at least one dispensing for DMARDs within 1 year from the first RA diagnosis date, and defined the primary outcome of incident T2DM as at least one inpatient or outpatient diagnosis of T2DM plus at least one dispensing of an antidiabetic drug. They set the general non-RA cohort by selecting patients with any inpatient or outpatient diagnosis codes and a dispensing of any medications, and the hypertension, PsA, and OA comparator groups as having at least two inpatient or outpatient disease-specific ICD-9/ICD-10 codes separated by 7-365 days and at least one dispensing of disease-specific medication within 1 year from the first diagnosis date. They excluded patients with RA, PsA, or psoriasis diagnosis or disease-specific medication dispensing any time prior to or on the index date (the date of disease-specific medication dispensing).
The researchers recognized that the conclusions that can be drawn from the study are limited by the “potential misclassification of cohorts and covariates” because they “mainly used diagnosis codes and pharmacy dispensing records in claims data,” and some “important covariates such as baseline obesity are likely underreported and not adequately captured in claims data.” The level of covariate misclassification also may have been different across the study cohorts on “unmeasured covariates such as body mass index, diet, and physical activity, as well as disease specific measures,” thus introducing residual confounding. They also could not “examine potential difference in the risk of T2DM in untreated or undertreated RA patients” because “RA and all the non-RA comparator cohorts were required to use a disease-specific drug,” they wrote.
“While systemic inflammation in RA is thought to increase the risk of [cardiovascular disease] and cardiovascular risk factors such as DM, our findings suggest having RA itself does not confer an increased risk of DM. Future study should determine whether untreated RA or undertreated RA is associated with a greater risk of developing DM,” the researchers concluded.
The study was supported by a research grant from Bristol-Myers Squibb, which “played no role in the study design, data analysis or interpretation of data or presentation of results,” the researchers said. The company was “given the opportunity to make nonbinding comments on a draft of the manuscript, but the authors retained the right of publication and to determine the final wording.” One author reported receiving research grants from Brigham and Women’s Hospital from Pfizer, AbbVie, Bristol-Myers Squibb, and Roche for unrelated topics.
SOURCE: Jin Y et al. Arthritis Care Res. 2020 Aug 4. doi: 10.1002/acr.24343.