Discusión de Resultados

The CASPAR study group was established to derive new data driven classification criteria for PsA. Data were collected in 32 centres worldwide by people with acknowledged expertise in this condition. On average each centre contributed 20 cases and 20 controls, the controls being cases of other inflammatory arthritis, with at least half of these having RA. Data was collected to a standard format on consecutive clinic attendees with PsA (total included 588) and the next case of inflammatory arthritis (total 536). Altogether over 100 clinical, radiological and laboratory variables were collected. The new criteria were derived by logistic

regression and classification and regression tree (CART) analysis (as a cross check) and the performance of the new criteria compared to the other existing criteria.

The new CASPAR criteria (figure 1) include characteristic dermatological, clinical and radiological features and have both high sensitivity and very high specificity. It is also interesting to note that, with these criteria, it is possible to be RF positive and still have a diagnosis of PsA, providing other characteristic features are present. People with the features described by Moll and Wright are still classifiable as PsA. Dermatological features contribute more to the criteria than the other features: the simple combination of psoriasis and an inflammatory arthritis gave the very respectable figures of 0.96 and 0.97 for sensitivity and specificity respectively (Taylor et al. 2006).

The CASPAR criteria

Inflammatory articular disease (joint, spine, or entheseal) With 3 or more points from the following:

1. Evidence of psoriasis (one of a, b, c) (a) Current psoriasis *

Psoriatic skin or scalp disease present today as judged by a rheumatologist or dermatologist

(b) Personal history of psoriasis

A history of psoriasis that may be obtained from patient, family doctor, dermatologist, rheumatologist or other qualified health-care provider

(c) Family history of psoriasis

A history of psoriasis in a first or second degree relative according to patient report

2. Psoriatic nail dystrophy

Typical psoriatic nail dystrophy including onycholysis, pitting and hyperkeratosis observed on current physical examination

3. A negative test for rheumatoid factor

By any method except latex but preferably by ELISA or nephelometry, according to the local laboratory reference range

4. Dactylitis (one of a, b)

(a) Current

Swelling of an entire digit

(b) History

A history of dactylitis recorded by a rheumatologist

5. Radiological evidence of juxta-articular new bone formation

Ill-defined ossification near joint margins (but excluding osteophyte formation) on plain xrays of hand or foot

Specificity 0.987, sensitivity 0.914

* Current psoriasis scores 2 whereas all other items score 1 Figure 1 – The CASPAR criteria

The main limitation highlighted in the CASPAR paper concerns the applicability of the criteria to early disease as the mean duration of disease of the cases was 12.5 years. Recent research done by the Toronto group attempted to address this issue and evaluated the use of the CASPAR criteria in early disease within their clinic population (Chandran et al. 2007b). They performed a retrospective analysis of patients enrolling into the specialist PsA clinic and found a sensitivity of 99.1% in those patients with disease duration of less than 2.5 years, and a sensitivity of 100% for those with disease duration of less than 12 months (Chandran et al. 2007b). Although important, this study only included patients referred to a specialist tertiary referral clinic and did not evaluate any control population. It seems likely that only patients with secure clinical diagnoses are referred and enrolled into this clinic, possibly leading to an overestimate of the sensitivity of the criteria. As yet, the only evaluations of the CASPAR criteria in new rheumatology referrals at a secondary care level have been studies in Sweden and Italy. In the Swedish early PsA register, 134 of 183 patients fulfilled the CASPAR criteria when compared to consultant diagnosis (Lindqvist et al. 2008). A single centre study in Italy of 44 patients (D'Angelo et al. 2009), found that 34 of the patients met the criteria on their first visit (77.3%) and the majority fulfilled the criteria by having skin psoriasis and being RF negative. Only two patients satisfied the radiological criteria as their disease duration was less than one year. They concluded that the CASPAR criteria are less sensitive in early PsA, mainly because patients are unlikely to have radiological evidence of new bone formation (D'Angelo et al. 2009). There were no controls analysed in either of these studies, so only the sensitivity could be calculated. There is early data from an Italian multicentre PsA study which found a high sensitivity and specificity of 91% and 97.1% respectively (D'Angelo et al. 2008). However, as yet, there have been no properly powered studies addressing the sensitivity and specificity of the CASPAR criteria in early disease and therefore they cannot currently be recommended for use in studies of early PsA.

Perhaps the weakest aspect of the CASPAR criteria is the initial qualification criterion: inflammatory arthritis including spinal, peripheral and entheseal disease. As cases were physician diagnosed and without other stipulation in the selection process, it was impossible to be more precise with this description. In fact the majority of cases had a peripheral arthritis pattern, although 72 had a combined axial/peripheral pattern and 21 did not have any peripheral joint involvement at all. Of these 21, two had pure axial disease, eight dactylitis, 12 enthesitis, and 10 inflammatory spinal pain (in combinations) (Taylor et al. 2006).

A third concern with these criteria is located in the composition of the controls. About 70% of the controls had RA but 13% had AS so the statistical analyses were influenced against selecting spinal features as characteristic of PsA (Taylor et al. 2006). Although it has been suggested that the spondylitis of PsA is qualitatively and quantitatively different from that seen in classical AS, these differences did not appear as discriminating features. Of course, had the controls only consisted of RA cases, it is possible that the spinal features would have appeared in the final criteria set. The criteria derived, and the figures for sensitivity and specificity (and those for post hoc calculations like the likelihood ratios) are therefore very dependent on the control population (non- or alternative disease) used to derive them.