Adjusting for bias in C-reactive protein levels when using a vitros slide method in infants

Low total serum protein levels may cause a positive bias on C-reactive protein (CRP) detected by the Vitros 250 Chemistry System (Ortho-Clinical Diagnostics, Inc., Johnson & Johnson Co., Raritan, NJ). Low total serum protein levels are observed in some infants. Our objective was to define a cutoff value for normal levels of CRP measured on the Vitros System that is comparable to the cutoff value of 1.0 mg/dL measured by rate nephelometry on a Beckman Array System (Beckman Instruments Inc., Fullerton, CA).

CRP was prospectively measured on the same serum sample on Vitros and Beckman systems. Using a result of ≥1.0 as the "gold standard" definition of an abnormal CRP, measures of association were calculated.

CRP was measured in 981 blood samples that were collected from 361 infants. A cutoff CRP level using the Vitros system at 1.5 mg/dL had the highest sensitivity and negative predictive value comparable to 1.0 mg/dL measured by nephelometry. By regression analysis, each increase by 1 mg/dL by nephelometry caused an increase by 1.5 mg/dL on the Vitros system (R(2) = 0.94; p < 0.001; slope = 0.66; 95% confidence intervals, 0.65, 0.67).

In infants, when measuring CRP levels by Vitros CRP slide system, a normal reference level of 1.5 mg/dL instead of 1 mg/dL should be used.