Background The last decade brought major advances in treatment choices for rheumatoid arthritis (RA) patients. The first choice in the treatment of RA patients is methotrexate (MTX) with a rapid step up to alternative conventional disease modifying antirheumatic drugs (DMARDs) or biologicals according to established treatment algorithms. Following these strategies, treatment retention can be seen as a surrogate marker for treatment response in clinical practice.
Objectives We aimed to identify patient or treatment characteristics or comorbidities that are associated with treatment retention in an observational cohort of RA patients.
Methods We identified patients starting DMARD treatment for RA patients from a longitudinal observational database. We considered the first 6 sequential treatments in our analysis, because the number of patients with >6 treatment was small. We used a two step forward conditional Cox regression model. In the first step we tested clinical variables and the impact of the sequence of the treatment as well as DMARD category (eg. csDMARD monotreatment, Non-TNFi biological DMARD, etc); and in the second step we tested the effects of comorbidities using hemoglobin levels, leukocyte counts, serum creatinine and glomerular filtration rate (GFR), alanin-aminotransferase (ALT) and gamma-glutamyl-transferase (GGT) as additional predictors. We finally repeated the analyses using only patients of an inception cohort.
Results 695 patients were identified in our longitudinal observational database for analyses: 78% female; mean ± SD disease duration: 4.7±7.9 years; CDAI: 17.6±11.1; age: 54.3±12.9; 61% rheumatoid factor positive). The inception cohort comprised 260 patients starting their first DMARD: 74% female; disease duration: 0.09±0.46 years; CDAI: 17.6±10.9; age: 54.3±13.5; 56% rheumatoid factor positive.
In analyses of the inception group we found that CDAI, age and treatment segment function as clinical predictors for treatment retention. Function, rheumatoid factor status, anti-citrullinated antibody or disease duration did not contribute significantly in explaining the outcome. The more previous treatments the patient had received, the higher the disease activity, and the younger the patient, the higher the risk for discontinuation of therapy. Considering comorbidities in step 2, only ALT could add significantly in step 2 of the analysis, with higher ALT levels being associated with lower risk of treatment discontinuation (Table 1). Whether the high baseline ALT is a consequence of chronic use of pain killers needs to be explored. In this case, the subjective benefit of DMARDs may also contribute to longer retention.
When using the complete cohort the same clinical variables were found significant. Concerning laboratory surrogates of comorbidities, only higher hemoglobin levels were associated with a lower risk of treatment discontinuation (Table 1). ALT did not associate with treatment retention in the complete cohort.
Conclusions Patients on their first DMARD treatment show higher retention rates than on subsequent treatments, independent of DMARD category. Anemia is an important independent risk factor of DMARD discontinuation.
Disclosure of Interest None declared