Usefulness of CRP and ESR in Predicting Septic Joints

Abstract

To determine whether erythrocyte sedimentation rate (ESR), C-reactive protein (CRP), or a combination of both was better in diagnosing a septic joint. A retrospective chart review was done in all patients who had serum assays for ESR or CRP as well as joint fluid analysis over a two-year period. Based on lab and operative findings, patients (cohorts) were categorized as having normal, inflammatory, or septic joints. Sensitivities (SEN), specificities (SP), positive and negative predictive values (PPV, NPV) were obtained using our lab's positive cutoffs of 15 for ESR and 0.8 for CRP. Contingency tables were used for comparisons between predictor variables and the presence of septic joints. Receiver operator curves (ROC) were obtained for CRPs and ESRs. Of 163 patients, 72 had inflammatory joints, 44 had septic joints, and 47 were normal. Fifteen admitted to drug use and 43 to alcohol consumption. There were 120 males and 42 females. The mean CRP for septic joints was 13, 8.5 for inflammatory joints, and 6 for normal. The mean ESR for septic joints was 57, 48 for inflammatory joints, and 43 for normal joints. By univariate analysis, drug use and elevated CRPs were significantly associated with septic joints while alcohol use, ESRs, and gender were not. A regression model with 4 variables indicated that drug use and CRP were predictive of septic joint; alcohol and ESR were not. CRP is helpful in determining the presence of a septic joint; ESR is not.