Pathophysiology and clinical implications of perioperative fluid excess.

EditorÐWe read with interest the review concerning the pathophysiology and clinical implications of perioperative ̄uid excess. However, we were surprised and have to disagree with the conclusion that there is a need for randomized, prospective clinical studies to compare `high' vs `low' ̄uid regimens. We were `surprised' because the authors highlighted two excellent studies where appropriate plasma volume expansion resulted in signi®cantly better gastrointestinal perfusion, and reductions in morbidity and hospital stay, 3 and commented that ` ̄uid loading to optimize cardiac function should therefore be guided by the Starling curve, and may have bene®cial effects on postoperative organ function'. We suggest that a better conclusion would be to call for more randomized, prospective clinical studies comparing ̄uid regimens guided by appropriate perioperative haemodynamic monitoring with current practice. Current practice tends to rely on ̄uid therapy being guided by changes in heart rate and arterial blood pressure, which are insensitive of occult hypovolaemia, and are able to detect only major circulatory losses. A recent randomized controlled trial in elderly patients undergoing surgery for hip fracture, again demonstrated reductions in postoperative morbidity and a faster postoperative recovery in patients receiving ̄uid regimens where cardiac function was `optimized'. `Optimization' was guided by utilizing the Starling curve, with either central venous pressure or oesophageal Doppler cardiac output measurement. No patient who received additional ̄uids as a result of these ̄uid challenge techniques, experienced ̄uid overload. The wide range of ̄uid given to the protocol groups (500±3500 ml) in this study emphasizes the fact that giving a single volume load is not appropriate when applied to individuals. Fluid must be individually titrated to dynamic changes in appropriate monitored variables, and studies of arbitrary `high' vs `low' ̄uid regimens would be dangerous. Holte and colleagues have elegantly reviewed the dangers of inappropriate excessive ̄uid administration, and there are many studies demonstrating detrimental effects on morbidity and mortality as a consequence of occult hypovolaemia. Therefore any emphasis about ̄uid therapy in the perioperative period should be on additional ̄uid therapy guided by appropriate, relatively simple and minimally invasive, monitoring techniques. The dangers of perioperative ̄uid excess are avoided by these monitoring techniques, 3 6 and we would not wish for future studies to include `low ̄uid regimens', returning to the days of the dehydrated patient undergoing surgery in a critical physiological state, with all the inevitable morbidity and mortality this incurs.