Polyethylene glycol (PEG) has been used as a poorly absorbable marker in intestinal perfusion studies, but there is controversy about the absorbability of PEG, particularly when glucose-sodium cotransport is occurring. Total intestinal perfusion studies were done in five normal humans using three solutions containing 1 g/liter PEG 3350 and designed to produce low rates of water absorption, high rates of water absorption, or high rates of glucose-sodium cotransport. Water absorption rates were calculated by traditional nonabsorbable marker equations and by a novel balance technique in which absorption was taken as the difference between the volumes of solution infused and recovered during steady-state conditions. Effluent PEG recovery was 99 +/- 4%, 109 +/- 2%, and 104 +/- 6% of the amount infused with each solution. Water absorption rates measured by use of PEG concentrations were similar to those calculated by the balance technique (r = 0.99). The complete recovery of PEG confirms the poor absorbability of PEG 3350, and the excellent agreement between techniques validates PEG as a poorly absorbed marker, even when glucose-sodium cotransport is occurring.