To determine the role of acid hydrolysis on the gastrointestinal absorption of erythromycin, six healthy subjects received erythromycin as a 240 mg intravenous dose, a 250 mg oral solution administered via endoscope directly into the duodenum and bypassing the stomach, and an enteric-coated 250 mg capsule. Blood samples were collected for 6 hours and serum erythromycin quantified by a microbiological method. The time to achieve maximum serum concentrations for the solution was 0.25 +/- 0.08 (mean +/- SD) hours and for the capsule was 2.92 +/- 0.55 hours. The absolute bioavailability of erythromycin from the capsule was 32 +/- 7% and for the duodenal solution 43 +/- 14%. The ratio of the areas under the serum erythromycin concentration-time curve of capsule to solution was 80 +/- 28% (range 38 to 110%). There is substantial loss of erythromycin apart from gastric acid hydrolysis, which cannot be accounted for by hepatic first-pass metabolism. Attempts to further improve the oral bioavailability of erythromycin beyond 50% by manipulation of formulation are likely to be futile.