Background/Purpose: Discordance between the erythrocyte sedimentation rate (ESR) and C-reactive protein (CRP) exists in a significant minority of rheumatoid  arthritis (RA) patients. Interpretation of studies exploring the phenomenon is problematic for a number of reasons including: the dichotomization of continuous variables, failure to consider potential influence of none acute phase factors and common co-morbid conditions. To assess discordant values between the ESR and CRP we did a prospective case control study by establishing deciles of CRP and ESR values and evaluating the highest decile (i.e. the one having the most discrepant values).

Methods: Data on individual RA patients consecutively seen over a 6 month period were prospectively obtained. Serologic data included complete blood counts, albumin, liver and renal function, and serum protein electrophoresis. All co-morbid conditions and medications used were noted. Weight, sex, age, and length of time from diagnosis were recorded. The clinical disease activity index (CDAI) was used as a measure of disease activity. We established deciles and categorized participants into separate groups based on their ESR/CRP ratios (stratifying by ESR; age and gender were accounted for). The discordant group comprised participants with the top 10% elevated ESR and normal CRP values and vice versa. Each patient was matched to a non-discordant patient seen most closely to that visit and matched for age and sex (and when possible, weight). Data was consequently de-identified and evaluated. We computed between-subject (discordant vs. controls) univariate analyses of variance (ANOVA) on the above variables using age as a covariate.

Results: 251 individual RA patients were seen in the 6 month period. Of 26 patients meeting criteria 22 were in the high ESR normal CRP category. We elected not to include participants with high CRP normal ESR values because of low sample number (N = 4). Age, sex, ethnicity and weight were similar between groups, The most common co-morbid conditions included hypertension, gastric reflux disease and hyperlipidemia equally seen in each group. The only significant difference our data revealed was of hematocrit (p < 0.01), with the discordant group having significant lower mean values as compared to controls (36.30 vs. 40.15).The ANOVA failed to reveal an effect on CDAI (p > 0.05)

Conclusion: No previous study has looked at extreme discordance between CRP and ESR values in RA. We prospectively evaluated RA patients and compared those in the highest decile of discordance with matched controls (accounting for sex, weight, and ethnicity), analyzing multiple variables and found that aside from hematocrit values, there was no difference between groups.

---

« Back to 2013 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/an-analysis-of-highly-discordant-erythrocyte-sedimentation-and-c-reactive-protein-levels-in-rheumatoid-arthritis/