Background Rheumatic diseases, such as rheumatoid arthritis and spondylarthritis, are chronic autoimmune diseases that can be treated with new biological drugs.

Purpose The purpose of this study was to evaluate drug retention as a surrogate for the effectiveness of biologic therapy.

Material and methods In a rheumatological centre in Apulia, Italy, 210 patients with rheumatoid arthritis (RA) and 336 patients with spondylarthritis (SpA) were evaluated, following treatment with ‘biologics’ from 1 January 2001 to 31 May 2016. The effectiveness of the biological drugs was evaluated using Kaplan–Meier drug retention curves.

Results For RA patients, we analysed data after 2, 4 and 5 years of therapy. At 2 years, the drug persistence rates were: 75% for etanercept, 63% for rituximab, 60% for adalimumab, 50% for abatacept with a rapid decrease during continuation of therapy; and 42% for infliximab at 2 years and 25% at 4 years. At 5 years the values were: 57% for tocilizumab and 35% for rituximab, adalimumab, etanercept and certolizumab.

For SpA patients, we analysed the data after 2 and 4 years of therapy. At 2 years, the drug persistence rates were: 50–56% for etanercept, infliximab and adalimimab, 35% for certolizumab and 31% for golimumab. Ustekinumab was not statistically evaluable because marketing authorisation was less than 2 years. At 4 years the values were: 44% for etanercept, 38% for adalimumab, 34% for infliximab and 13% for golimumab. Certolizumab was not statistically evaluable because it was commercially available for less than 4 years for psoriatic arthritis.

Conclusion A comparison of RA and SpA showed that RA patients had a higher persistence after 2 years of treatment (75–60% vs 55–50%) than SpA patients; at 4 years the drug persistence rate was 50% for RA compared with 35% for SpA. In this study, RA patients probably responded better to ‘biological drugs’ because the biological targets were defined; in SpA, targets are less known, having recently discovered the IL-23/IL-17 (interleukin-23/interleukin-17) aetiopathogenetic way. Therefore, the different clinical responses and frequent change in drugs shows that it cannot predict either therapeutic response or appearance of adverse events.

References and/or acknowledgements We would like to thank the UO of Rheumatology at ASL Taranto and pharmacists at the Hospital Pharmacy of ASL Taranto

No conflict of interest