Characteristics and outcomes in sero-positive late-onset rheumatoid arthritis patients who start a new disease modifying anti-rheumatic drug

Background

Controversy exists in delineating the characteristics and therapeutics involved in late-onset rheumatoid arthritis (LORA) and younger-onset rheumatoid arthritis (YORA). Many studies have suggested that the disease process of the LORA patient is distinctly different, and thus requires different treatment. Our preliminary work using a rigidly defined, rheumatoid factor (RF)-positive rheumatoid arthritis prospective cohort of 263 patients suggests that LORA and YORA patients are similar after correcting for age-related factors, for example erythrocyte sedimentation rate and C-reactive protein. In the same cohort of patients, Wu and colleagues recently demonstrated that the presence of both shared epitope-containing DRB1*04 alleles and novel RANKL polymorphisms was associated with an 18–20 year earlier onset of rheumatoid arthritis in YORA patients and did not predict increased severity of disease.

Objective

Aim 1

To demonstrate that RF-positive YORA and LORA patients with the same disease duration have similar baseline disease-related characteristics, after adjusting for age-related processes.

Aim 2

To compare outcomes of RF-positive LORA and YORA patients with the same disease duration, after accounting for age-related processes.

Hypothesis 2A

Responses to disease-modifying anti-rheumatic drug treatments will be similar between LORA and YORA patients with equivalent disease duration.

Hypothesis 2B

RF-positive LORA and YORA patients with equivalent disease duration will have comparable side effects to disease-modifying anti-rheumatic drugs.

Methods

This study will use the Consortium of Rheumatology Researchers of North America and the Rheumatoid Arthritis Disease-Modifying Anti-Rheumatic Drug Intervention and Utilization Study 1&2 databases, consisting of 15,000 rheumatoid arthritis patients. Patients with known RF positivity will be included in the cohort, with LORA ≥ 60 and YORA < 60 years of age. We will control for duration by stratification. We will use two methods to demonstrate similarity between YORA and LORA subjects: equivalence testing and confidence intervals. Alternate statistical analyses include generalized matching by the propensity scores method and regression analyses.

Significance

LORA patients have been treated inconsistently and with much apprehension, due to uncertainty about the diagnosis of LORA and concern about prescribing toxic medications to the elderly. With strictly defined cohorts showing similar characteristics/outcomes, treatment of the LORA and YORA patients by physicians should be similar.