The influence of enzymic distribution on lidocaine metabolism was investigated in the once-through perfused rat liver preparation. Low input concentrations of 14C-lidocaine (1-2 microM) and preformed monoethylglycine xylidide (MEGX; 2.3-2.8 microM) were delivered by normal and retrograde flow directions to the liver preparations at 10 ml/min per liver. Upon reversal of normal to retrograde delivery of lidocaine, the rates at which lidocaine, MEGX, and glycine xylidide (GX) left the liver almost doubled, whereas the rates of appearance of (total) hydroxylated lidocaine and MEGX in bile and perfusate increased to lesser extents. Upon reversal of normal to retrograde delivery of preformed MEGX, the rates of appearance of MEGX and GX were virtually unchanged. Computer simulations on lidocaine and preformed MEGX metabolism were performed on both evenly distributed ("parallel tube" model) and enzyme-distributed systems. An even or parallel distribution of N-deethylation and hydroxylation activities for lidocaine metabolism failed to predict the observed increased hepatic availability of lidocaine. Rather, the distribution of a low-affinity, high-capacity N-deethylation system anterior to a high-affinity, low-capacity hydroxylation system for lidocaine metabolism adequately predicted the increased hepatic availability of lidocaine. Further extension of these consistent enzyme-distributed models on the metabolism of lidocaine metabolites suggests that the N-deethylation and hydroxylation activities for the metabolism of lidocaine, MEGX, 3-hydroxyidocaine, and 3-hydroxy MEGX are not identically distributed. When these enzyme-distributed models were appraised with reference to the "parallel tube" and "well-stirred" models of hepatic drug clearance, predictions from these enzyme-distributed models proved to be superior to the "parallel tube" and "well-stirred" models for the present data on lidocaine metabolites with normal and retrograde perfusions. Previously published data on lidocaine and MEGX metabolism after inputting 4 micrograms/ml (17 microM) lidocaine at flow rates of 10, 12, 14, and 16 ml/min were reexamined with respect to the adequacy of these enzyme-distributed models. They were found to be superior to the evenly-distributed or "parallel tube" model in predicting hepatic availability of lidocaine and the rate of appearance of MEGX. However, the enzyme-distributed systems were not as consistent as the "well-stirred" model in predicting lidocaine hepatic availability in these flow experiments.