The illustration describing the experimental preparation and the accompanying text in the methods section (5), make it clear that the venous return from the periphery was taken up from the right atrium by a pump and forced into the pulmonary artery. The authors emphasized that the left heart faithfully pumped out what was pumped into it. The flowmeter from which they obtained their Qr data recorded the outflow of the pump so it is equally clear that what they recorded was Qo -- they had no separate measure of Qr. Of course, since the focus of the paper was on venous return and because the data points were labeled as Qr, it was (and is) easy for readers to overlook or forget the fact that Qo was identical with Qr at the times when Qr and Pcv data were recorded.

Less clear in the original papers is the arrangement for control of Pcv, but careful reading supplies all the necessary detail. A complicated tubing system connected the intake of the pump and the cannulated right atrium. The complication consisted of a Starling resistor -- a section of flexible tubing that would collapse at any point along its length where intramural pressure was slightly subatmospheric. By moving this section up and down, relative to the level of the right atrium, pressure at the right atrium could be set at desired levels. Figure 2 is a schematic diagram of this arrangement for setting Pcv, contributed by Loring Rowell.

The key to understanding the data is the fact that the pumping rate was adjusted to keep the the Starling resistor at the collapse point. In that condition the pressure at the inflow end of the resistor (the end on the heart side, not the pump side) would be known to be equal to atmospheric pressure. By measuring the height of the hydrostatic column between the level of the resistor and the level of the heart, the pressure at heart level is known. For example, if the resistor is 10 cm below heart level, then pressure in the atrium is -10 cm of water (ignoring the difference between the specific masses of blood and water).

Most of the time during an experiment, the resistor was positioned well below heart level, making the pressure at the right atrium, i.e., Pcv, subatmospheric by several cm of water. This was probably a practical necessity, dictated by the animal's need for adequate cardiac output. Naturally, the experimenters wished to have cardiac output normal or supernormal for most of the time; otherwise, the animal's condition would have deteriorated. Guyton and coworkers emphasized that venous return (i.e., pump output) was maximal^* see footnote 1 under conditions of subatmospheric Pcv, and, in the range of subatmospheric pressures, virtually independent of Pcv, "…regardless of how low the pressure falls" (5, p 611). This phenomenon of independence between Pcv and flow on the flat portion of a venous return curve was interpreted on the basis of the same phenomenon seen in the Starling resistor -- the variable resistance of vessels in the portions of the animal's venous system where pressures were subatmospheric or nearly equal to atmospheric.

Data points for Qr at prescribed levels of Pcv were obtained by elevating the Starling resistor for 8-10 seconds to establish the desired Pcv. We know from the description of the experiments that a data set of points that described a venous return curve like that shown schematically in Figure 3 was obtained in a sequence of such runs with no blood volume added or removed. We also know that the pump was adjusted as necessary to maintain collapse in the Starling resistor and that the pump was of the positive displacement type that squeezes out a steady flow and develops powerful suction against any resistance imposed in the supply line. Therefore, we know that , when Pcv was adjusted upward to a particular target level, the steady level of Qr that was recorded was obtained by reducing the pumping rate. Why was it necessary to turn the pump down?

The short answer is that slowing the pump was necessary to achieve the reduced Qo that would come into equilibrium with the new level of Pcv. In order to push equilibrium Pcv upward, Qo must be set downward.

To look at this in a little more detail, recall that each run was preceded by a relatively long period with maximal levels of Qr and with Pcv subatmospheric by several mm Hg. This means that, immediately upon elevating the Starling resistor to, say, 5 cm above the heart, pressures in the column of blood between the resistor and the right atrium and throughout the vena cavae began to move from their original subatmospheric level toward the eventual equilibrium consistent with + 5 cm H2O pressure in the right atrium.

Where is the volume to come from that distends the vessels in the vicinity of heart level to bring their pressure up to the new equilibrium level? It could not possibly have accumulated had pump rate continued at the original level of Qo (thus relentlessly continuing to suck up Qr equal to that original maximal Qo of the rest period). In order to permit the volume to accumulate, pump rate had to be reduced.

It must have required a few seconds for sufficient returning venous volume to accumulate to bring pressures up to the new equilibrium level throughout the segments of the venous vasculature which had been pumped down to a minimal volume (the central compartment in Figures 1 and 2) due to the low right atrial pressure maintained during the rest period. The investigators did not add blood volume from their reservoir for (1) they stipulated they did not, and (2) that would have altered mean circulatory pressure (discussed below).

As it turns out, lower Qo is consistent with elevated Pcv, so, by reducing Qo to the right level, a stable equilibrium could be achieved. With reduced Qo, pressures fall in the peripheral venular system (peripheral compartment in Figures 1 and 2). This reduction of distending pressure results in proportionate reduction of volume contained within the peripheral venular system. As this volume moves out of the venules, it augments the venous return. It was this displacement of blood volume from peripheral toward central segments of the venous vasculature that provided the volume that distended central veins to achieve the elevated pressure in equilibrium with the new level of the Starling resistor. The only period in the experiments in which Qo was not identical to Qr was during the brief transient period following the setting of pump output to a new level.

Because the pressure in the tubing between heart and Starling resistor must have tended to fall immediately upon elevation of the resistor, the flaccid tubing of the resistor would be seen to collapse altogether, or flutter. This physical change in the resistor would be taken as a signal that it was necessary to reduce pumping rate. In other words -- the disparity between the rate at which the pump was attempting to suck flow through the Starling resistor and the rate at which blood was flowing into the input of the Starling resistor would have resulted in reduction of volume within the flexible segment of tubing in the resistor itself. Presumably, operators learned how to adjust pump output quickly to keep the resistor at equilibrium with atmospheric pressure.

Venous return curves indicate that Qo and Pcv obey a functional relationship; for a given level of one of these variables, the other is determined through the mechanical properties of the vasculature. Nonetheless, calling one or the other the independent variable is not a mere semantic distinction. In a vascular system with fixed total volume and resistance and compliance properties, nothing happens unless mechanical energy is supplied via a pump. Pump output is what the experimenters varied to discover the functional relationship and only through control of pump output can this be accomplished. Thinking of central venous pressure as a variable with an independent influence on cardiac output in this context is absurd.

In summary, Guyton, et al., emphasized that no volume was added or removed from the system during collection of data for a particular venous return curve. All data points on a venous return curve are obtained from periods in which cardiac output and venous return were equal. Cardiac output was set by adjusting the pump to obtain the flow rate through the vasculature consistent with the Pcv established by the level of the Starling resistor. The flowmeter from which the data points logged for Qr were obtained recorded the output of the roller pump, i.e., Qo (accepting the authors' statement that Qo did not differ from the rate of flow the pump forced into the pulmonary artery). Actual venous return differed from Qo only during the brief transient period following the physical movement of the Starling resistor and the adjustment of pump output, but this disparity was not recorded. During these transient periods, the difference between Qr and Qo was accommodated as the volume in the compliant segments of the peripheal vasculature changed in relation to the new profile of pressures.

Venous return curves fall to zero at a Pcv identified as the mean circulatory pressure (Pmc). What this means is, for the investigators to be able to obtain that particular value of Pcv, they had to stop the pump. The pump did not stop itself. Venous return went to zero because cardiac output was set at zero.

Footnote 1