The galactose elimination capacity test: a study of the technique based on the analysis of 868 measurements.

Abstract

Our objective in this study was to analyze the correspondence of galactose concentration-on-time-decay curve to theoretical assumptions and the confidence limits of the determination of galactose elimination capacity. We analyzed a retrospective series of 868 galactose elimination tests, performed on subjects with and without liver disease. Zero-order kinetics of galactose elimination was tested by comparison of the residual variance of linear regression with that obtained after quadratic transformation. The uncertainty in determination of galactose elimination capacity was calculated on the regression line by computing the 95% confidence limits of the estimate. The time-course of galactose concentration suggested an initial uneven distribution, and the first (20-min) data point deviated significantly from the regression. The galactose decay curve in plasma rejected linearity in 13% of tests; after exclusion of the first data-point, linearity was rejected in only 3% of cases. The 95% confidence interval of galactose elimination capacity was on average +/- 16%, but in individual tests it was as large as +/- 60-80%. The uncertainty of the test was not affected by linearity. It was larger, with poor fitting of the experimental data on the regression of galactose concentration on time, low number of data points, and low galactose elimination. It was maintained within +/- 20% only when residual variance was > or = 2% of total variance (nearly 50% of tests). The methodology for the determination of galactose elimination capacity leads to considerable uncertainty as to the final result, which must be considered whenever the test is used for clinical purposes in the decision-making process. It tends to be larger in patients with advanced disease and can be accurately calculated so as to contribute to a proper evaluation of the test result.