The age at onset in early arthritis (EA) may influence the disease activity and its evolution. The aim of the current study is to identify possible differences regarding the "old" and the "new" classification criteria between patients with early-onset and late-onset early arthritis. The study included 64 patients. They were divided in two groups, according to the mean age: early-onset EA--less or equal than 45 years old (group A) and late-onset EA--over 45 years old (group B). The "old" criteria as well as the "new" ones were assessed for all patients, at the time of the first visit to the rheumatologist. The initiation of treatment with Methotrexate was used as "gold standard" to calculate the sensitivity and the specificity of both criteria. "New" criteria were fulfilled in 51% (A) and 72% of cases (B), while "old" criteria were fulfilled in 37% of patients (A) and 62% (B). Methotrexate was initiated in 82% of patients (B) and in 51% (A), p = 0.01. "New" criteria demonstrated a sensitivity of 77.7% (A) and 83.3% (B), while "old" criteria had a sensitivity of 50% (A) and 66.6% (B). Patients with late onset had significantly higher disease activity scores: 76% (B) versus 40% (A), p = 0.04. The sensitivity and the specificity of the "new" criteria for RA are comparable in patients with early-onset and late-onset EA, and the sensitivity of these criteria is increased compared to the "old" criteria. Patients with late onset fulfilling the "old" criteria had poor prognostic factors and higher disease activity at the time of diagnosis, which may have possible implications for the disease course.