Agreement between erythrocyte sedimentation rate and C-reactive protein in hospital practice.

BACKGROUND: Erythrocyte sedimentation rate (ESR) and C-reactive protein (CRP) are frequently prescribed jointly. The usefulness of this practice is uncertain.
METHODS: All patients with ESR and CRP measured at the same time in an academic tertiary hospital during a 1-year period were included. Concomitant measures of serum creatinine, hematocrit, and anti-Xa activity were recorded to study noninflammatory cause of increased ESR. Level of agreement between ESR and CRP was assessed with kappa coefficient, and their accuracy was determined in a medical chart review of 99 randomly selected patients with disagreement between both markers.
RESULTS: Among 5777 patients, 35% and 58% had an elevated CRP and ESR, respectively. ESR and CRP were in agreement in 67% of patients (both elevated in 30%, both normal in 37%). A disagreement was observed in 33% (elevated ESR/normal CRP in 28%, normal ESR/elevated CRP in 5%). The kappa coefficient showed poor agreement (k=0.38) between both markers. Review of medical chart showed that 25 patients with elevated CRP and normal ESR had an active inflammatory disease (false-negative ESR). Conversely, 74 patients had elevated ESR and normal CRP-32% had resolving inflammatory disorders, 28% disclosed a variable interfering with the ESR measure (false-positive ESR), 32% had unexplained discrepancies, and 8% had an active inflammatory disease (false-negative CRP).
CONCLUSION: In hospital practice, joint measurement of ESR and CRP is unwarranted. Because of slow variation and frequent confounding, ESR is frequently misleading in unselected patients. When an inflammatory disorder is suspected, priority should be given to CRP.