Background/Purpose:  Joint pain is one of the most common reasons for referral to pediatric rheumatology clinics throughout the world.  Typically, these patients ultimately fall into one of 3 categories of diagnosis: (1) inflammatory, (2) non-inflammatory, or (3) mechanical, each requiring different management. Our objective was to determine what factors on clinical history best predict the final etiology of the joint pain in the 3 categories. 

Methods: New patients between the ages of 2 and 22 seen by the pediatric rheumatology clinic at Hackensack University Medical Center between 1/1/2013 and 12/31/2014 with joint pain were included in our retrospective chart review. Data were collected regarding demographics, symptom history, and physician assessment. Final diagnosis was categorized as one of the 3 stated categories. Univariate statistics were performed for all covariates in the sample. Significant differences were tested using a Fisher’s Exact test. A multivariate multinomial regression model was estimated to assess whether the characteristics associated with Mechanical or Non-inflammatory joint pain were significantly different than Inflammatory joint pain. A multinomial logistic regression was performed as there was not an inherent ordering of severity of the groups of joint pain.

Results:   288 charts were included in the analysis.  31.6% were inflammatory, 25.35% non-inflammatory, and 28.47% mechanical (table 1).  42 patients, or 14.58% of the patients did not have a clear final diagnosis, with either resolution of their symptoms or failure to follow up.  These patients were not included in the multivariate multinomial regression analysis.  An abnormal CRP, presence of swelling, a lower pain score and the absence of a history of anxiety or depression increased the odds of inflammatory joint pain as compared to non-inflammatory.  Compared to patients who have symptoms present for <6 weeks, patients who have symptoms for 3-12 months are 6.89 times more likely to have mechanical pain (OR= 6.893, p=0.0370) and 8.85 times more likely to have non-inflammatory pain than to have inflammatory pain (OR=8.854, p=0.0772) (table 2).

Conclusion:   Our results indicate that while patients with non-inflammatory pain have higher pain scores, a longer duration of pain and perhaps more subjective pain, the objective data (swelling and increased CRP) are more consistent with inflammatory joint pain.  This information is especially useful for the general pediatrician when triaging patients with joint pain and deciding upon referrals.  Educating pediatricians could result in better management and therefore improved outcomes.