Describing Assay Precision-Reciprocal of Variance Is Correct, Not CV Percent: Its Use Should Significantly Improve...

Area of Science:

• Analytical Chemistry
• Pharmacokinetics
• Clinical Laboratory Science

Background:

• Percent coefficient of variation (CV%) is an inadequate measure of assay error, especially for low measurements.
• This inadequacy leads to data censoring below the lower limit of quantification (LLOQ) and inaccurate pharmacokinetic modeling.
• Current methods result in suboptimal therapeutic drug monitoring and patient dosage regimens.

Purpose of the Study:

• To introduce and evaluate the reciprocal of variance (1/var) as a superior alternative to CV% for quantifying assay error.
• To demonstrate how 1/var can improve the accuracy of pharmacokinetic models and therapeutic drug monitoring.
• To advocate for the adoption of 1/var to enhance laboratory and clinical practice.

Main Methods:

• Compared CV% with the reciprocal of variance (1/var) for assay measurements.
• Developed a polynomial function to describe assay standard deviation (SD) relative to measurement values.
• Assessed the performance of 1/var in weighting data, particularly at low concentrations.

Main Results:

• Reciprocal of variance (1/var) provides accurate data weighting down to zero, unlike CV%.
• Eliminates the need for censoring low results or setting an arbitrary lower limit of quantification (LLOQ).
• Enables more precise pharmacokinetic modeling and individualized patient dosage regimens.

Conclusions:

• Reciprocal of variance (1/var) is the correct metric for assay precision, superior to CV%.
• An assay error polynomial using 1/var can be easily implemented, improving laboratory services.
• Regulatory bodies should adopt policies supporting the use of 1/var, removing the need for LLOQ.