Background/Purpose:

Our previous research showed that baseline characteristics would enable the sub-categorization of psoriatic arthritis (PsA) into the distinct phenotypic groups of axial and peripheral disease using cluster analysis (CA). We aimed to determine if demographic and disease characteristics of PsA patients with at least five years of follow-up can be clustered into distinct groups.

Methods:

527 PsA patients from an observational cohort of patients seen from 1978 to 2013. Patients had at least 5 years of follow-up and genetic information in order to be included in the analysis. CA using Ward’s method was conducted to identify groups of patients based on the following characteristics at follow-up:  gender, type of psoriasis (type I or II), duration of PsA, race, family history of psoriasis, ESR, severe PASI, psoriasis vulgaris, nail disease, dactylitis, swollen joint count, damaged joint count, axial disease, presence of arthritis mutilans, presence of arthritis prior to psoriasis, HLA-B*27 and HLA-C*06. 7 clusters were formed and matched to non-overlapping arthritis patterns:  distal arthritis, oligoarthritis, polyarthritis, axial only, distal arthritis and axial, oligoarthritis & axial, and polyarthritis & axial. Comparisons between the clusters and arthritis patterns were conducted using t-tests and Chi-squared analysis.

Results:

The baseline characteristics of the 527 patients were as follows: 306 (58.1%) males, mean age at diagnosis of PsA 28.3 (14.2) yrs, mean age at first visit 42.5 (12.3) yrs, mean duration of PsA 6.8 (8.0) years, mean active joint count 10.6 (9.2), mean PASI 5.2 (8.1), mean Steinbrocker score 12.7 (25.9), HLA-B*27 94 (18%)]. At follow-up, mean age was 57.6 (12.6) yrs, mean PASI 4.6 (6.6), mean Steinbrocker score 28.0 (37.3) with an average follow-up of 15.0 (7.6) yrs. Two main clusters of patients were identified. One consisted of distal arthritis, oligoarthritis and polyarthritis and the other of axial only, distal and axial, oligoarthritis & axial, and polyarthritis & axial, thus clearly identifying patients into peripheral and axial disease. Comparison of the two clusters showed more patients with psoriasis vulgaris among patients with peripheral disease. Patients falling into the axial disease cluster had a higher prevalence of males and HLA-B*27, higher PASI, more dactylitis, higher clinically damaged joint and swollen joint count, more axial disease, more arthritis mutilans and more patients who developed arthritis first.

Conclusion: Based on patients’ characteristics at follow-up, CA analysis separated PsA patients into two predominant arthritis patterns coinciding with the results that looked at a the same cohort  at baseline with the exclusion of B27 and C6 as clustering variables. The study provides further evidence to classify patients into just two groups based on the presence or absence of axial arthritis not only at baseline but also at follow-up.

Table: Comparison of two clusters

ACR Meeting Abstracts - https://acrabstracts.org/abstract/cluster-analysis-of-psoriatic-arthritis-patients-at-follow-up/