Patients with clinically diagnosed hand OA not fulfilling the ACR classification criteria are in an earlier disease phase and more often have thumb base OA.

Objective: To investigate the performance of the American College of Rheumatology (ACR) classification criteria for hand osteoarthritis.

Design: Longitudinal data up to four years from a cohort of consecutive patients with primary hand osteoarthritis diagnosed by their rheumatologist (Hostas study) were used to classify presence or absence of hand osteoarthritis according to the 1990 ACR criteria (traditional format: one major and 4 minor ACR criteria) (ACR/ACR). Demographics, Australian/Canadian osteoarthritis hand index (AUSCAN) pain and function were obtained. Hand radiographs were scored according to Kellgren-Lawrence; radiographic osteoarthritis was defined as Kellgren-Lawrence ≥2 in ≥1 CMC1 joint or ≥2 DIP/PIP/MCP joints.

Results: Of 538 patients (mean age 61 years, 86.1% women) 485 (90.1%) fulfilled ACR criteria at baseline. Except for the minor criterion swelling of <3 MCP joints, all criteria differed between the groups. ACR patients were younger, with higher BMI, a shorter time since diagnosis, and less bony enlargements, joint deformities and radiographic osteoarthritis, except for radiographic CMC1 osteoarthritis which was seen more often in ACR patients. No difference in AUSCAN pain or function was seen between ACR versus ACR patients. After follow-up 37/53 (69.8%) converted to ACR, 2/53 (3.8%) did not, and 14/53 (26.4%) were lost to follow-up.

Conclusions: In clinical practice the majority of patients fulfill the ACR classification criteria, but those in an earlier disease phase, with less signs of hand osteoarthritis or with primarily thumb base osteoarthritis are less likely to fulfill them. New classification criteria also including earlier disease stages and with attention for hand osteoarthritis subtypes are required.