Reliability of C-reactive protein measurement in general practice in The Netherlands

The aim of this study was to compare the reliability of the whole blood NycoCard C-reactive protein measurement (CRP) with the erythrocyte sedimentation rate (ESR) in four general practice centres (GPCs) in The Netherlands with 443 patients indicated for ESR. During 4 successive months in 1992 NycoCard CRP was measured in capillary (whole) blood at the GPC. A venous blood sample was then transported to the local hospital laboratory in order to determine CRP turbidimetrically in plasma. Westergren's ESR was measured at the GPC and at the laboratory. The dichotomized CRP values measured at the GPCs corresponded with the laboratory values in 88% of the cases. Kappa was 65% (p < 0.00001). Using a 20 mm cut-off point, ESR values measured at the GPCs and the laboratory corresponded in 96% of cases (Kappa = 90%, p < 0.00001). Interpractice variabilities and learning effects did not influence the results. 'False normal' rates for CRP and ESR were 8% and 1%; 'false elevated' rates were 28% and 12% respectively. The reliability of the NycoCard CRP measurement in whole blood disappointed. In particular the 'false elevated' rate is unacceptably high for daily general practice. The background of the poor reliability remains unclear.