OP-0138 – Risk of Active Tuberculosis among Biologic-Naïve RA Patients: Retrospective Cohort Study in a U.S. Claims Database
Mines, Gu, Liu, Horne, Harrison; Wyeth Research, United States
Objectives
TNF inhibitors have the potential to reactivate latent tuberculosis. However, it is unclear whether there is an increase in the background rate of active tuberculosis in U.S. RA patients not receiving TNF inhibitor therapy.
Methods
The Pharmetrics Patient-Centric Database was used to retrospectively compare incidence rates for incident tuberculosis, based on claims data for the prescription of at least two anti-tuberculosis medications including isoniazid, for biologic-naïve adult RA patients compared to non-RA patient matched on gender, age, health plan, calendar time, and time of enrollment in the insurance plan. Up to twenty controls were matched to each RA patient. Patients with prior biologic exposure, prior tuberculosis, and HIV or history of organ transplantation were excluded.
Results
Among approximately 55 million enrolled individuals, 59,996 were identified as having RA based on coding for RA and DMARD use. These were compared to 945,258 non-RA matched controls. Patients were mostly female (76%) with a median age of 52 years. The mean duration of follow-up in the cohort was 1.8 years. There were 4 cases of active tuberculosis among the 100,414 person years of follow-up in the RA group, resulting in an incidence rate of 4 per 100,000 patient-years. There were 9 cases of tuberculosis among the 1,759,302 person years in the non-RA controls, resulting in an incidence rate of 0.5 per 100,000 patient-years. The RR of tuberculosis was 7.79 (95% CI 2.40 - 25.3). Sensitivity analyses in which the diagnosis of tuberculosis was restricted to different case definitions did not change the significance of the results.
Conclusions
The background rate of tuberculosis may be increased in RA patients.
Editorial Comment
The frequency of tuberculosis here is exceedingly small, with too few cases to really make valid comparisons between the groups—hence the greatly inflated relative risk and broad confidence interval. Important confounders of active tuberculosis are prior exposure and country of origin for the participants, neither of which were part of the dataset used for this study. Any increase in prior exposures for only one or two RA patients could entirely explain the differences in tuberculosis observed here. Presumably, this research was undertaken for purposes of product placement, since only biologic naïve patients were investigated; yet on balance, the best interpretation of these data is that the risk of tuberculosis is exceedingly small in both RA and non-RA Americans with commercial health insurance.
