Objectives: This study aimed to evaluate the impact of different comorbidities on thereflecting its safety profile persistence of biological disease-modifying anti-rheumatic drugs (bDMARDs) in rheumatoid arthritis (RA), taking advantage of a retrospective analysis of administrative healthcare databases (AHDs).Method: A retrospective observational study was conducted on AHDs of the Lombardy region, Italy (2004-2013). Among RA patients treated with bDMARDs, drug survival was estimated using Cox proportional hazard models [hazard ratio (HR), 95% confidence interval (CI)], crude and adjusted for prespecified confounders (gender, age, disease duration, concomitant use of non-steroidal anti-inflammatory drugs, glucocorticoids, conventional DMARDs, specific bDMARDs), in first-line and subsequent lines of treatment. The role of comorbidities in administration of specific bDMARDs was analysed through multinomial logistic models.Results: The study included 4657 RA patients. In the first-line treatment strategy, the Charlson Comorbidity Index (CCI) (RA excluded) was significantly associated with an increased rate of bDMARD failure (CCI = 1: HR 1.28, 95% CI 1.13-1.46; CCI ≥ 2: HR 1.26, 95% CI 1.03-1.53). Among selected comorbidities, chronic obstructive pulmonary disease (HR 1.38, 95% CI 1.01-1.91), diabetes (HR 1.18, 95% CI 1.01-1.37), and previous-year bacterial infections (HR 1.18, 95% CI 1.07-1.30) were slightly associated with risk of bDMARD failure, while acute myocardial infarction (HR 1.30, 95% CI 0.97-1.75), mild liver disease (HR 1.21, 95% CI 0.91-1.60), and solid tumours (HR 1.19, 95% CI 0.93-1.53) were not. In the following treatment lines, neoplasms were associated with reduced risk of failure (HR 0.64, 95% CI 0.41-0.99). Multiple comorbidities were associated with first-line abatacept and rituximab administration.Conclusions: Comorbidities affect treatment decisions in RA and influence bDMARD failure, and should be considered when analysing the persistence of biological therapy.