1547 - Substituting Joint Swelling and Radiographic Erosions in the 1987 ACR Criteria for Rheumatoid Arthritis with Synovitis and Erosions Defined by Ultrasonography Improves the Sensitivity with Little Loss in Specificity

Tove Lorenzen1, Jens Kristian Pedersen1, Lis Smedegaard-Andersen1, Steen Nielsen1, Kim Hørslev-Petersen1, Mikkel Østergaard2, Marcin Szkudlarek2. 1King Christian Xth Hospital for Rheumatic Diseases, Graasten, Denmark; 2Copenhagen University Hospital, Hvidovre, Denmark
Presentation Number: 1547

PURPOSE: In epidemiological studies it is often difficult to confirm the diagnosis in patients with self-reported rheumatoid arthritis (RA). In this study, we examined if the performance of the list format of the 1987 ACR criteria for RA could be improved by substituting clinical and radiological findings with findings defined by ultrasonography (US) in persons with self-reported RA.
METHODS: From a population-based survey, we recruited 53 persons with self-reported RA and from case records it was ascertained that 19 had previously fulfilled the 1987 ACR criteria (“ever RA”). Their mean age was 60 years (range 30-83) and mean duration of joint symptoms was 14 years (range 1-53). At the clinical examination an independent assessor evaluated 40 joints for swelling, and a test for rheumatoid factor, and radiographs of the hands and wrists were performed. On the same day, blinded to the results of previous or current classifications, US was performed on the wrists, metacarpophalangeal, proximal interphalangeal and metatarsophalangeal joints (32 joints) by one of two examiners. Synovitis and erosions were evaluated using the semi-quantitative US scoring system developed and later validated by Szkudlarek et al1, 2, 3. In a pilot study, kappa-values for inter-observer agreement of the US examiners were 0.80 for US synovitis and 0.84 for US erosions.
RESULTS: Out of the 19 “ever RA” cases, 10 (53%) fulfilled the ACR list criteria at the day of observation. When clinically defined joint swelling was substituted by synovitis defined by US (score > 1) and erosions on radiographs were replaced with bone erosion (score > 1) on US, 15 of the 19 (79%) with “ever RA” fulfilled the criteria and two persons without “ever RA” were classified as having RA. None of the two were known with other specific joint condition. With “ever RA” as gold standard, the diagnostic performance (95% confidence limits) of the two criteria set is shown in the table.
Cases fulfilling criteriaCases not fulfilling criteriaSensitivitySpecificityPositive predictive valueNegative predictive value
Ever RA1934----
Current ACR list criteria1000.53
(0.32-0.73)
1.00
(0.90-1.00)
1.00
(0.72-1.00)
0.80
(0.65-0.87)
Current US ACR list criteria1520.79
(0.57-0.92)
0.94
(0.81-0.98)
0.88
(0.66-0.97)
0.89
(0.75-0.96)

CONCLUSIONS: In this study substitution of clinical and radiographic criteria with US findings improved the sensitivity of the 1987 ACR criteria for RA with little loss in specificity.
Reference: 1Arthritis Rheum 2003; 48:955, 2Arthritis Rheum 2004;50:2103-12, 3Arthritis Res Ther. 2006 ;8:R52.

 T. Lorenzen, None.