Estimation of uncertainty in kinetic-colorimetric assay of bacterial endotoxins

The relative standard deviation (R.S.D.) of measurements is estimated in the kinetic-colorimetric assay of bacterial endotoxins without recourse to the usual repeated experiments. The measurements are the slopes of kinetic curves and two major factors are considered to cause the uncertainty of the measurements: (1) the pipetting of the sample and color development reagent; and (2) noise in the detection unit. The measurement R.S.D. is formulated as a function of endotoxin concentration. Two parameters (S.D. of the pipetted volumes and S.D. of the detector noise) are also required in the uncertainty equation, but no arbitrary coefficients are included. Since the S.D. values for pipettes and detector noise can be determined independently of the endotoxin assays, the measurement R.S.D. can be estimated by the above equation without repeating the assays. However, the calibration curve is necessary. The theoretical estimation is shown to be in good agreement with the experimental R.S.D. (n = 12) over a wide concentration range.