Uncertainty in liver function assessment on the basis of single-point galactose concentration.

The uncertainty in liver function assessment based on single-sample galactose levels after galactose injection, in comparison to the standard procedure using the galactose elimination capacity (GEC), was assessed in 905 tests performed in a wide range of liver functions. The 45-min galactose levels significantly correlated with GEC, the correlation being better in subjects with a good liver function. In cirrhosis, the prediction using the 60-min galactose value was better than when the 45-min value was used. In the whole series, the 95% confidence interval of GEC predicted by 45-min galactose was as large as +/- 1.55mg x kg(-1) x min(-1) as absolute value, corresponding to a range from -41 to +47% of measured GEC. In cirrhosis, the 95% confidence interval was +/- 1.42 mg x kg(-1) x min(-1) and between -40 and +46% of measured GEC. The 60-min values were more predictive, but in Child class C patients the average error was 0.42 mg x kg(-1) x min(-1) (95% confidence interval -0.64 to +1.49, corresponding to -39 to +76% of measured GEC). The uncertainty was maintained within +/- 10% of measured values only in 50% of the tests. We conclude that the single-point galactose test introduces a considerable error, mainly in patients with more advanced liver disease, which may bias the decision-making process.