Falls in gastric intramucosal pH (pHi) are associated with morbidity and mortality in patients admitted to intensive-care units (ICU). We tested the hypothesis that ICU outcome can be improved by therapy guided by changes in pHi and aimed at improving systemic oxygen availability. We studied 260 patients admitted to ICUs with APACHE II scores of 15-25. After insertion of a gastric tonometer, each patient was randomly assigned to a control or protocol group within the admission pHi category (normal = 7.35 or higher; low = below 7.35). The control groups were treated according to standard ICU practices. The protocol groups received, in addition, treatment to increase systemic oxygen transport or to reduce oxygen demand, whenever the pHi fell below 7.35 or by more than 0.10 units from the previous measurement. The protocol was used, because pHi fell, in 67 (85%) of the protocol group with normal pHi on admission. There were no significant differences between protocol and control groups in demographic characteristics, admission blood gases or haemoglobin concentration, number or type of organ system failures, or the intensity of ICU care. For patients admitted with low pHi, survival was similar in the protocol and control groups (37% vs 36%), whereas for those admitted with normal pHi, survival was significantly greater in the protocol than in the control group (58% vs 42%; p less than 0.01). Therapy guided by pHi measurements improved survival in patients whose pHi on admission to ICU was normal. pHi-guided resuscitation may help improve outcome in such patients by preventing splanchnic organ hypoxia and the development of a systemic oxygen deficit.