AIM: Age at disease onset has been implicated as an indicator of disease activity and severity in rheumatoid arthritis (RA). This study aimed to investigate how old age at disease onset affects patient treatment and prognosis in early RA. METHODS: Data from the Kyoto University Rheumatoid Arthritis Management Alliance (KURAMA) cohort was analyzed. From 2011 to 2015, a total of 2182 patients with RA were enrolled in the cohort; 239 patients were newly diagnosed with RA and were followed up for 2 years. The patients were divided into the following two groups: the young-onset RA (YORA) which included patients <60 years old (n = 117) and elderly-onset RA (EORA) which comprised patients ≥60 years old (n = 122). The clinical and laboratory data were compared at baseline, at 1 year, and at 2 years after onset. RESULTS: Disease activity was higher in EORA than in YORA at baseline. Although disease activity was equivalent between EORA and YORA at 1 or 2 years, more EORA patients had bone erosions at baseline and at 2 years. More than 25% of the anti-citrullinated protein autoantibody (ACPA)-positive EORA patients without erosions at baseline had bone erosions even if they attained clinical remission at 1 or 2 years, while ~10% of YORA patients had erosions. CONCLUSION: Bone erosions were more frequently found in EORA. Clinical remission at 1 or 2 years was not enough to protect bone erosions in the ACPA-positive EORA patients. Optimal treatment strategies preventing radiological damage should be considered for EORA.