Abstract
Several screening tools for early identification of psoriatic arthritis (PsA) have been developed. While these tools had high sensitivity and specificity during their development and initial validation, it remained to be determined how they would function with widespread use. At the 2012 annual meeting of the Group for Research and Assessment of Psoriasis and Psoriatic Arthritis (GRAPPA) in Stockholm, Sweden, these tools were compared for their utility when used to screen patients for PsA in clinics other than those in which they were developed. The screening tools did not perform as well as previously published, and members suggested new tools may need to be developed. An additional study of the prevalence of PsA in a large cohort of psoriasis patients, the PREPARE study, which investigated the use of screening questionnaires, was also presented.
- PSORIASIS
- TORONTO PSORIATIC ARTHRITIS SCREEN
- PSORIATIC ARTHRITIS SCREENING EVALUATION
- PSORIATIC ARTHRITIS
- PSORIASIS EPIDEMIOLOGICAL SCREENING TRIAL
Psoriatic arthritis (PsA) is an inflammatory musculoskeletal disease that affects about 30% of patients with psoriasis. The disease may be severe, leading to joint damage, joint deformity and disability, as well as reduced quality of life and function, and may be associated with increased mortality1,2,3. Even in early arthritis, erosions were detected in 27% of the patients within an average of 10 months from onset of symptoms, and in 47% of the patients within 2 years4. It has been suggested that early diagnosis of patients with PsA would lead to appropriate treatment and avoid some of these untoward consequences5.
Many psoriasis patients have PsA that is not initially recognized by the patients or by the attending physician. In studies performed at dermatology clinics in Europe, 20%–30% of the patients were not diagnosed with PsA until the time of the research visit6,7,8, further demonstrating the importance of early diagnosis.
The ClASsification criteria for Psoriatic ARthritis (CASPAR) were developed to facilitate the classification of PsA9. These criteria are highly sensitive and specific, and function well in both early and late disease10,11,12,13; however, they can be applied only to patients who have inflammatory musculoskeletal disease — peripheral arthritis, axial disease, or enthesitis — which are difficult for nonexperts to recognize. Therefore, dermatologists and other physicians need tools to screen psoriasis patients for the presence of PsA. Several tools have been developed14,15,16,17. The Leeds group used the Psoriatic Arthritis Questionnaire (PAQ), originally developed by Paul Peloso, as a substrate, with added questions on back disease and a manikin where patients mark their affected areas. The resultant Psoriasis Epidemiological Screening Trial (PEST) demonstrated high sensitivity (92%) and very good specificity (78%)14. In Boston, a joint effort by dermatologists and rheumatologists resulted in the development of the Psoriatic Arthritis Screening Evaluation (PASE)15, with a sensitivity of 82% and specificity of 73% (at a cutoff of 47). An updated version provides a sensitivity of 70% with improved specificity at 80% (cutoff of 47), and both sensitivity and specificity of 76% (cutoff of 44)18. Another screening tool, the Psoriatic Arthritis Screening Questionnaire (PASQ), was developed in St. John’s, Newfoundland, Canada, and uses the PAQ with a manikin identifying the joint areas involved. The PASQ provided a very high sensitivity of 97% and a specificity of 75% (cutoff of 7)16.
At the 2012 annual meeting of GRAPPA, these tools were compared for their utility when used to screen patients for PsA in clinics other than those in which they were developed.
ToPAS and ToPAS 2
Dafna Gladman (Toronto, Canada) reviewed the development and validation of the ToPAS, used to identify PsA in patients with or without psoriasis17. Developed initially through input from patients and physicians, ToPAS includes 12 questions about psoriasis, nail lesions, joint pain and swelling, back pain and stiffness, and dactylitis, with pictures of the skin and nail lesions. ToPAS was validated in 5 groups of patients (from a PsA clinic, dermatology clinic, family medicine clinic, rheumatology clinic, and phototherapy center), all of whom were evaluated by a rheumatologist for the presence of PsA. The ToPAS demonstrated high sensitivity (87%) and specificity (93%) in those groups of patients17.
In the original version of ToPAS, the presence of axial symptoms did not contribute to the instrument, perhaps because questions were not specific enough. Moreover, it was thought that patients might not be able to recognize joint disease and dactylitis. Subsequently, on the ToPAS 2, questions on axial disease were modified and pictures of inflamed joints as well as dactylitis were added. ToPAS 2 was tested among individuals participating in family studies and had sensitivity of 95.9%, specificity of 98.7%, with a positive predictive value of 98.9% and a negative predictive value of 94.9% when PsA patients were compared to unaffected individuals. The ToPAS and PASE were compared; although the former had a higher area under the curve (AUC), confidence intervals overlapped, suggesting that both tools functioned well19.
Comparison of ToPAS, PASE, and PEST
Philip Helliwell and colleagues (Leeds, UK) tested ToPAS, PASE, and PEST in patients undiagnosed with PsA in the CONTEST (COmparisoN of ThreE Screening Tools) trial20. All 3 questionnaires were distributed in packages to psoriasis clinics in the UK and were randomized by instrument order. Patients were age ≥ 16 years, able to read and understand English, with a diagnosis of psoriasis made by a dermatologist, and without a previous diagnosis of PsA. Patients were each given all 3 questionnaires and asked to come back only if they screened positive in any of the questionnaires. Of 938 packages that were distributed, 657 were returned; 314 patients screened positive by at least one questionnaire and 195 were examined by a rheumatologist. Forty-seven of those 195 were diagnosed with PsA by the CASPAR criteria. Of those screening positive on the questionnaires, 15% of patients with PsA screened positive on 1 questionnaire, 33% on 2 questionnaires, and 47% on all 3 questionnaires. Results of each questionnaire were calculated according to its developers’ instructions. Based on these calculations, the sensitivity, specificity, and AUC were 76.6%, 29.7%, and 0.554 for the ToPAS; 74.5%, 38.5%, and 0.594 for the PASE; and 76.6%, 37.2%, and 0.610 for the PEST.
The results of the CONTEST trial were compared with those published by Haroon, et al8 in a Dublin study that included 200 patients: 100 consecutive patients with psoriasis with no known inflammatory arthritis from dermatology clinics, and 100 patients with PsA from rheumatology clinics. All patients were seen by a rheumatologist and completed all 3 questionnaires. Of interest, the 2 studies had opposite specificity and sensitivity results. High sensitivity and low specificity was documented in CONTEST; low sensitivity and high specificity was documented in the Dublin study. The CONTEST study did not examine all patients and therefore might have underestimated specificity. Both studies showed lower AUC than the original development studies of the questionnaires, and the PEST had the best AUC results in CONTEST. Further, screening tools identified many patients with non-PsA arthritis, because they identify other musculoskeletal diagnoses if patients have significant symptoms, and thus may not be as accurate as they were in development.
Psoriatic Arthritis Screening Questionnaire
The PASQ, discussed by Majed Khraishi (St. John’s, Newfoundland, Canada), contains 10 differently weighted questions as well as a diagram on which patients marked where they have or have had pain and/or swelling. PASQ was validated in a group of established PsA (58 patients); a group with psoriasis but not PsA (29 patients); and in patients with early PsA and patients with psoriasis but no PsA16. The original validation showed sensitivity of 86% and specificity of 89%, and a newer electronic version showed high sensitivity (93%) but not as good specificity (75%). However, the electronic version was tested only in an early-PsA cohort, and may not be as accurate in patients with early disease21. Patients found it easy to use, and the tool is available free online.
PREPARE Study
The Prevalence of Psoriatic Arthritis in Adults With Psoriasis: An Estimate from Dermatology Practice (PREPARE) study, discussed by Philip Mease (Seattle, WA, USA), is a non-interventional assessment of psoriasis patients in dermatologists’ offices22. The ToPAS, PASQ, and PEST were administered to 1000 patients, although each patient completed only one questionnaire, so no comparative data are available. Patients had one visit with the dermatologist followed by 2 visits with a rheumatologist. At the first visit with the rheumatologist determination of the patient’s PsA was based on clinical features. In a subset of sites, imaging was performed in 20% of the patients with subsequent assessment based on clinical features and radiographs. The overall prevalence of PsA was 30%; 12% (41% of the PsA cohort) were new diagnoses. Of 47 patients previously diagnosed with PsA (5% of the cohort), the current rheumatologist’s assessment found no PsA, but determined that these patients may have other diagnoses such as osteoarthritis. The majority of PsA diagnoses (99%) could be made on the basis of clinical evaluation alone. Similar results were found in both university clinics and primary practice offices.
In summary, although screening tools developed to identify PsA early were found to be highly sensitive and specific during development and initial validation, further studies suggest that they may not perform as well in the clinic. Measuring the effect of the musculoskeletal symptoms on function and quality of life may be one way of improving the specificity of these tools, thus filtering out other rheumatic disorders that may cause the screening tool to score positively. Notably, however, assessment of the effect on function was included in the PASE, but PASE did not score better than the other instruments. Therefore, the most important screening tool for use in dermatology clinics would have a high sensitivity. New instruments may be needed to address this issue.
