The use of total parenteral nutrition (TPN), in the critically ill, may be associated with the translocation of bacteria and their products from the lumen of the gastrointestinal tract (GIT) to the systemic circulation. We report a study comparing the effects of TPN and enteral nutrition (EN) on GIT function. Twenty-four critically ill patients were randomly allocated to receive TPN or EN. GIT absorption was measured by urinary recovery of D-xylose and 3-O-methyl-D-glucose (3O MG) after enteral administration. The ratio between urinary recovery of lactulose and L-rhamnose (L/R) was used to measure GIT mucosal permeability. Results are expressed as the percentage of enterally administered saccharide recovered. Measurements were performed at entry to the study and on every subsequent third day. Baseline recovery of D-xylose (5.95 +/- 1.61%, EN; 6.56 +/- 3.38%, TPN) and 3O MG (12.35 +/- 4.06%, EN; 7.96 +/- 4.19%, TPN) was significantly lower than for the controls (35.03 +/- 1.40% for D-xylose, p < 0.05 compared to both study groups; 49.20 +/- 1.98% for 3O MG, p < 0.05 compared to both study groups). Baseline L/R ratio was increased (0.292 +/- 0.072%, EN; 0.463 +/- 0.118%, TPN) compared with the controls (0.038 +/- 0.006%, p < 0.05 compared to both study groups). In the EN group, the L/R ratio displayed a progressive, significant fall. In the TPN group, no significant change in the L/R ratio occurred. This study demonstrates that GIT dysfunction is evident in critically ill patients and suggests that loss of GIT mucosal integrity is reversed by the institution of EN.