The European Alliance of Associations for Rheumatology developed new classification criteria for hand osteoarthritis, including separate criteria for overall hand osteoarthritis, interphalangeal osteoarthritis, and thumb base osteoarthritis.

The requirements, published in Annals of the Rheumatic Diseases, were developed through a three-phase process involving data-driven and consensus-based approaches.

Among key features of the new European Alliance of Associations for Rheumatology (EULAR) criteria were:

• Mandatory criteria of hand symptoms present on most days for 6 weeks, with no other disease or injury explaining symptoms
• Classification based on age, morning stiffness duration, number of joints with osteophytes and joint space narrowing, and symptom-structure concordance
• Radiographic features required, in contrast to the 1990 American College of Rheumatology (ACR) criteria
• Separate criteria and cut-offs for overall hand osteoarthritis (OA), interphalangeal OA, and thumb base OA
• Similar or improved sensitivity and specificity compared with ACR criteria in validation cohorts.

The criteria were intended for use in clinical trials and observational studies to assemble homogeneous patient populations. They allowed the classification of thumb base OA, which is poorly captured by existing ACR criteria.

To apply the criteria, patients had to have pain, aching, and/or stiffness in target hand joints on most days for 6 weeks, with no other condition explaining their symptoms. Psoriasis was an exclusion. Points were then assigned based on age, morning stiffness duration, number of joints with radiographic osteophytes and joint space narrowing, and concordance between symptoms and radiographic findings.

For overall hand OA, 9 or more points on a 0 to 15 scale were required for classification. For interphalangeal and thumb base OA, the cut-off was 8 points or greater. The criteria could be applied without morning stiffness data using modified cut-offs.

The development process involved three phases:

• Phase 1: A multicenter observational study of 128 persons with hand OA and 70 with other conditions identified discriminating features through logistic regression.
• Phase 2: A multidisciplinary expert panel of 21 members from 4 continents used consensus methods to refine and weight the criteria.
• Phase 3: The scoring system was refined, cut-offs determined, and criteria validated in external cohorts.

In validation cohorts, sensitivity of the EULAR criteria ranged from 0.57 to 0.74 for overall hand OA and 0.66 to 0.78 for interphalangeal and thumb base OA. This was similar to or exceeded sensitivity of ACR criteria. Specificity was comparable between EULAR and ACR criteria.

The EULAR criteria demonstrated good discrimination of hand OA cases vs controls, with an area under the curve of 0.81 for overall hand OA criteria.

In the phase 1 dataset, 40.8% of the participants fulfilled both EULAR and ACR criteria, whereas 36.7% fulfilled neither. Among those fulfilling only one set, EULAR-only cases had more joints with osteophytes (7.2 vs 2.0) and joint space narrowing (5.8 vs 1.6) compared with ACR-only cases.

The investigators noted several advantages over existing ACR criteria:

• Ability to classify thumb base OA separately
• Use of radiographic rather than clinical features, allowing central reading
• Requirement for concordance between symptoms and radiographic findings
• Exclusion of patients with other conditions causing hand pain.

However, the criteria were not intended for early disease classification or clinical diagnosis. They pushed toward classification of more established disease as a result of radiographic requirements.

The investigators concluded these new classification criteria would enable the inclusion of more homogeneous populations in hand OA research and clinical trials, including studies of targeted therapies for specific phenotypes like thumb base OA. They acknowledged that further validation in populations without OA was still needed.