Optimal fluid management is one of the main challenges in the care of the critically ill. However, the physiological parameters that are commonly monitored and used to guide fluid management are often inadequate and even misleading. From 1987 to 1989 we published four experimental studies which described a method for predicting the response of the cardiac output to fluid administration during mechanical ventilation. The method is based on the analysis of the variations in the arterial pressure in response to a mechanical breath, which serves as a repetitive hemodynamic challenge. Our studies showed that the systolic pressure variation and its components are able to reflect even small changes in the circulating blood volume. Moreover, these dynamic parameters provide information about the slope of the left ventricular function curve, and therefore predict the response to fluid administration better than static preload parameters. Many new dynamic parameters have been introduced since then, including the pulse pressure (PPV) and stroke volume (SVV) variations, and various echocardiographic and other parameters. Though seemingly different, all these parameters are based on measuring the response to a predefined preload-modifying maneuver. The clinical usefulness of these 'dynamic' parameters is limited by many confounding factors, the recognition of which is absolutely necessary for their proper use. With more than 20 years of hindsight we believe that our early studies helped pave the way for the recognition that fluid administration should ideally be preceded by the assessment of "fluid responsiveness". The introduction of dynamic parameters into clinical practice can therefore be viewed as a significant step towards a more rational approach to fluid management.