Background/Purpose: A high proportion of patients who present with recent-onset arthritis have undifferentiated arthritis (UA), as these patients do not fulfill the classification criteria for a specific diagnosis. Several cohort studies in UA have used predictive score models to screen patients at risk to develop rheumatoid arthritis (RA). Our objective is to validate the Leiden Prediction Rule (LPR) in an inception cohort of patients with UA.

Methods: We included patients with UA diagnosed from 2008 to 2011, mostly from a community based epidemiological study (COPCORD), who met the following characteristics: >18 years old, at least 1 swollen joint, with a symptom duration of >1 week to 1 year. We excluded patients with pregnancy or with well-defined inflammatory disease. During one year the patients were followed every 2 months. We applied the LPR at the first visit, which includes sex, age, morning stiffness, disease distribution, painful and swollen joints, C-reactive protein (CRP), rheumatoid factor assay (RF) anti-cyclic citrullinated peptide (anti-CCP) antibodies; in each visit they were evaluated for clinical signs of inflammation including the metacarpophalangeal (MCP)/metatarsophalangeal (MTP) squeeze test (ST). Subsequently we reclassified patients a year after their diagnosis of UA. The diagnosis of RA was made according to 1987 ACR classification criteria. Statistic Analysis. The results were expressed as means and standard deviations. The groups were compared with Mann Whitney U test. Chi-Square test was applied for categorical variables. The association between MCP/MTP ST and LPR was established by relative risk (RR) with a confidence interval (CI) of 95%.  

Results: We included 47 patients with UA with a mean age of 51.6 years + 9.5 SD, 98% were females. One year after diagnosis, the patients were reclassified as follows: RA 20(43%), and non- RA which included: persistent UA 12(25%), other inflammatory arthritis 9(19%) and spontaneous remission 6(13%). There were no statistical differences between the groups on the number of swollen or painful joints ( p=0.576 and p =0.564, respectively) or titers of CRP, RF and anti-CCP antibodies (p=0.754, p=0.97, p=0.29, respectively). In patients who progressed to RA, LPR was ≥ 8 points in only 15% (3 patients), while 85% (17 patients) had <8 points in LPR. When we compared the score obtained from LPR between groups, we did not find any significant difference (p=0.940). We found an association between the MCP/MTP squeeze test (ST) and patients who progressed to RA, RR 1.74; 95% CI [0.92 – 3.28] on MCP ST and RR 2.99; 95% CI [1.12 – 4.70] on MTP ST.

Conclusion: A proportion of 43% of our patients with UA had progression to RA after 1 year of follow-up; 100% patients who scored >8 points in the LPR progressed to RA. These proportions are similar to the data published previously from other UA cohorts. Notably, an important difference is the 56% of patients who progressed to RA with a score <6 points on LPR compared to the Leiden cohort were it was less than 10%. We did find that the presence of pain during the MCP/MTP squeeze test is associated with the progression of RA.

---

« Back to 2012 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/a-prediction-rule-for-the-progression-to-rheumatoid-arthritis-applied-in-a-mexican-mestizo-cohort-with-undifferentiated-arthritis/