To assess avoidance of hepatic first‐pass effect of drugs, we undertook in situ experiments using rectal administration of lidocaine in the rabbit. We also employed in situ duodenal route to estimate first‐pass metabolism across the gastrointestinal mucosa. Rabbits were administered lidocaine HCl intravenously (i.v., 50 mg/20 min) and portally (i.p.v., 33.3, 16.7, 8.3 mg/20 min) and avoidance of hepatic first‐pass effect (Fh) was calculated from the area under curve (AUC). Fh was about 30% and did not vary with increasing i.p.v. dose. Intravenous and i.p.v. administration was followed by duodenal (i.d.) or rectal (i.r.) administration and the absorption (fa), Fh, and avoidance of first‐pass effect in the duodenal mucosal membrane (Fm) were determined. With i.d. administration, lidocaine was absorbed completely with negligible first‐pass effect in the mucosa (Fm=1). On the other hand, while lidocaine was also absorbed almost completely via the i.r. route, avoidance of first‐pass effect was 60%, representing twice the bioavailability via i.d. administration. On the basis of these data, assuming that the first‐pass effect in the rectal mucosa was negligible, we estimate the fraction of rectal venous drainage bypassing the portal circulation and thus hepatic metabolism (fnh) to be about 40%. Copyright © 1998 John Wiley & Sons, Ltd.