The terminology used for describing intervention groups in randomised controlled trials (RCT) on the effect of intravenous fluid on outcome in abdominal surgery has been imprecise, and the lack of standardised definitions of the terms 'standard', 'restricted' and 'liberal' has led to some confusion and difficulty in interpreting the literature. The aims of this paper were to clarify these definitions and to use them to perform a meta-analysis of nine RCT on primarily crystalloid-based peri-operative intravenous fluid therapy in 801 patients undergoing elective open abdominal surgery. Patients who received more or less fluids than those who received a 'balanced' amount were considered to be in a state of 'fluid imbalance'. When 'restricted' fluid regimens were compared with 'standard or liberal' fluid regimens, there was no difference in post-operative complication rates (risk ratio 0.83 (95% CI 0.49, 1.39, P = 0.47) [corrected] or length of hospital stay (weighted mean difference (WMD) - 1.77 (95% CI - 4.36, 0.81) d, P = 0.18). However, when the fluid regimens were reclassified and patients were grouped into those who were managed in a state of fluid 'balance' or 'imbalance', the former group had significantly fewer complications (risk ratio 0·59 (95% CI 0·44, 0·81), P=0·0008) and a shorter length of stay (WMD -3·44 (95% CI -6·33, -0·54) d, P=0·02) than the latter. Using imprecise terminology, there was no apparent difference between the effects of fluid-restricted and standard or liberal fluid regimens on outcome in patients undergoing elective open abdominal surgery. However, patients managed in a state of fluid balance fared better than those managed in a state of fluid imbalance.