Of the various pieces of equipment and devices used in cardiopulmonary bypass (CPB), roller pumps are particularly important, because they serve to propel the blood through various devices until reaching the patient. In cardiac surgery, two types of pump are currently used: the roller pump (peristaltic pumping) and the centrifugal pump, the former being more frequently used.1 The double-headed DeBakey roller pump with a pulsatile flow profile has come to be adopted universally for CPB. A segment of elastic tube is mounted on the rigid, horseshoe-shaped bed, forming a circular segment with parallel extensions on which move two cylinders (rollers) equidistant from a central axis. As the central axis turns, the rollers compress the tube and propel its contents forward. An excessively occluded roller increases damage to blood cells and may lead to accentuated hemolysis. An excessively loose roller permits retrograde blood flow and, additionally, leads to variable volumes of blood being pumped in accordance with the resistance of the devices used and the vascular system of the patient. Studies have confirmed the effect of occlusion on hemolysis rates2,3 and on cardiotomy suction return.4,5 Errors in the measurement of blood flow rate have been reported during CPB,6 and its influence with the resistance of the devices added to the circuit.7
Drop rate is a method for regulating pump rollers before CPB surgeries, and its measurement is normally carried out at the lowest point of the raceway or midway between three randomly selected points close to the lowest point.8,9
Another roller pump calibration method is the so-called dynamic calibration,8 where the pump circuit is filled with saline and a pressure monitor is installed at the pump outlet and the pump is set to a slow speed at a constant 5–10 rotations per minute (rpm). The tubing at the outlet is clamped beyond the pressure monitoring location and the rollers are regulated until the desired pressure is achieved.
In our studies of the hydrodynamic profile of roller pumps, we have observed a reduction of drop rate with time that could not be explained by silicone tube water or saline absorption. We have hypothesized that this reduction could be related to residual stress in silicone tubes at the measuring points.
The objective of this was to evaluate the role of residual stress of silicone tubes used in CPB procedures by measurements carried out during drop rate and dynamic calibration methods.
Part of this study was conducted at the Nucleus of Medicine and Experimental Surgery of the School of Medical Sciences, University of Campinas (UNICAMP) and part at the Laboratory of Mechanical Properties of the School of Mechanical Engineering, UNICAMP.
Silicone tubes normally used in CPB procedures with 3/8 × 1/16, 3/8 × 3/32, and ½ × 3/32 inches in diameter were tested. The tubes were supplied by two manufacturers.
Dynamic Calibration Method
For the measurements of dynamic calibration, the method proposed by Tamari et al.8 was used. Figure 1 illustrates the methodology used for the tests. The test was carried out using tubing with a diameter of 3/8 × 1/16 inches with pump 2. In short, one should 1) fill the pump circuit with saline solution; 2) place the pressure sensor at the pump outlet; 3) retract the rollers to a nonocclusive position; 4) set the pump at a constant 10 rpm; 5) clamp the outlet tubing at a point distal to the pressure sensor; and 6) set the occlusion of the rollers to the desired average pressure.
A data acquisition system model PCI-9112 was used, manufactured by Adlink, Chungho, Taiwan, together with a pressure sensor (Ashcroft Willy Instrumentos de Medição Ltda, São Paulo, Brazil) calibrated to a pressure range of −1.105 to 2.105 Pa (i.e., −750 to 1,500 mm Hg). We have developed a program for acquisition, reading and archiving of data.
Each value of the average pressure in dynamic calibration (DCPave) (Table 1) was calculated as the mean value from 1,000 points recorded on file with confidence level of 95% (total record time equals to 20 s and time interval between samples 20 ms). DCPave was calculated at 0, 5, 10, 15, 20, 25, 30, 40, 50, 60, and 70 minutes. Between measurements, the rollers were maintained at 90-degree angle. Figure 1 shows the measurement setup and part of the pressure signal recorded for each point in Table 1.
Drop Rate Method
Drop rate versus time was measured for a fixed variation of 50 mm in a column of 0.9% saline solution standardized at 1,000 mm of height in a ¼-inch polyvinyl chloride (PVC) tube, as shown in Figure 2. A digital chronometer accurate to one-hundredth of a second was used to record elapsed time.
Drop rate tests were performed at predetermined time intervals in two models of tubes: 3/8 × 1/16 inches (manufacturer 1) and 3/8 × 3/32 inches (manufacturer 2) and in two samples of tubes of each model. We have used these tubing because each manufacturer provides only one type of tube. All tubes were new. The rollers positions for the measurements are as follow:
Tube 1, 3/8 × 1/16 inches in diameter, roller A in position 0-degree angle, pump 2, saline;
Tube 2, 3/8 × 1/16 inches in diameter, roller B in position +20-degree angle, pump 1, saline;
Tube 3, 3/8 × 3/32 inches in diameter, roller A in position 0-degree angle, pump 1, saline;
Tube 4, 3/8 × 3/32 inches in diameter, roller B in position +20-degree angle, pump 2, saline.
These two roller positions (0 and +20-degree angle) and the use of two different pumps were chosen to see whether possible effects on drop rate were due to roller position, to the pump used or to residual stress of the silicone tube. Measurements were taken at 0, 5, 10, 15, 20, 25, 30, 35, 40, 45, 50, 55, 60, 65, and 70 minutes.
With the pump adjusted to the desired occlusion point, the roller was moved manually until the column of saline solution reached a height of 1,050 mm. Roller was initially placed at the standard angle and tube was clamped at its entry point. After waiting a few seconds to allow pressure stabilization to occurs, the tubing was then released and the elapsed time between the 1,050 and 1,000 mm positions was recorded (first measure). Then, the roller was moved manually again until the column of saline solution reached a height of 1,050 mm at the same angle, the tube was clamped at its entry point and a second elapsed time measurement was performed. Drop rate was calculated as 50 mm divided by the measured elapsed time for roller positions A and B. Then, drop rate was taken as the mean of the two calculated values. This procedure was repeated for the other silicone tubes, standard angles and pump. The temperature of the saline used in the tests was 24.0 ± 0.5°C, whereas room temperature was maintained at 25.0 ± 1.0°C.
Residual stress was measured in silicone tubes of 3/8 × 1/16, 3/8 × 3/32, and ½ × 3/32 inches in diameter. A constant force was applied and displacement was measured for 70 minutes. Values were recorded every 30 s with a total of 140 records per sample. A cylindrical load tip 45 mm in diameter was placed perpendicularly to the tube length, simulating the position of the rollers.
To perform the measurements, the Model 810—TestStar II servo-hydraulic test bench for mechanical tests (MTS, Minneapolis, MN, 1978/1996) was used, with a load cell of 100 N and precision of 1%. Figure 3 illustrates the measurement setup.
Silicone tubes were occluded with a constant 27 N load leaving approximate gaps of 3.3, 1.4, and 3.9 mm for the 3/8 × 1/16,3/8 × 3/32, and ½ × 3/32 inches, respectively. The gaps were calculated from the initial position of the load cell in contact with the external diameter of the tube. Visual inspection has guaranteed that the internal walls were not occluding the tube.
Dynamic Calibration Method
The results of the DCPave (Figure 4) show normal distribution and mean pressure remained stable over time (p > 0.79). Table 1 shows the measured pressure average values.
Drop Rate Method
Figure 5 shows the results of drop rate values over time with the four silicone tubes. Table 2 shows mean and standard deviation of the calculated drop rate. The data shown in Table 2 present normal distribution (p < 0.01). Covariance analyses of the four linear regressions have shown that there are no statistical differences for the angular coefficients (p > 0.56).
Figure 6 provides results of the compression tests. The force applied to the three tubes was maintained constant at 27.0 ± 0.1 N for the entire 70-minute duration of the test. The variation in displacement (n = 140) measured during 70 minutes was 0.35, 0.20, and 0.35 mm in tube 1, 2, and 3, respectively. The data did not present normal distribution and Spearman's correlation test has shown significant reduction in displacements values (p < 0.0001).
The adjustment of the roller pumps before CPB is fundamental, because very occlusive roller adjustment causes severe hemolysis but guarantees a condition of flow near the measuring conditions of the pumps.2,8 Nonocclusive adjustments promote retrograde flow and complicate the relationship between real flows and those established by the rotation of the pump.7,9
The drop rate measurements (Figure 5) showed decreasing values over time (angular coefficient β <0). Comparison between regressions demonstrates equality in angular coefficients by analysis of covariance test (p > 0.56), i.e., reductions in drop rate were independent of the thickness of the tubes (β = −0.17), which is a strong indication that the variations are related to the properties of the material. Other researchers have shown that variations around 0.1 mm between the rollers and the raceway can cause changes from 2 to 3 times in drop rate, depending on the diameter of the tube and the initial occlusion.8 We have not found in the literature any information regarding the reduction in drop rate measurements besides a comment by Tamari et al.8
Our tests were made with two different roller pumps with measurements in rollers A and B in two positions 0-degree angle and 20-degree angle, as shown in Figure 2. The temperature of the saline was kept in ±0.5°C. In this condition dimension variations on raceway and in the rollers may be disregarded as our results have shown that the angular coefficients of the linear regression of curves of the drop rate measurements in different conditions are statistically the same (β = −0.17). We then hypothesized that the only possible explanation to our results was the reduction of the tube area caused by the residual stress of the tube material.
This was confirmed by the compression tests that have shown variations in the displacement between 0.2 and 0.35 mm during 70 minutes. Half of the values of displacement were obtained in the first 10 minutes of load application for the three types of tubes analyzed. If pump adjustment is made using drop rate and calibration is performed 10 minutes after the positioning of the roller at the measuring point the new measured value will be different.
The compression test (Figure 6) shows the effect of the residual stress relaxation in silicone tubes over time. This relaxation means that the force that maintains the occlusion (spring effect of the tube wall) decreases with time and the available area for the saline go through decreases with increasing relaxation. This explains the decrease in drop rate over time shown in Figure 5. Residual stress did not influence the adjustments made by the method of dynamic calibration as we can observe that there were no changes in the mean pressure absolute values over time (Figure 4).
Our results point out that drop rate calibration method should not be used in combination with silicone tubes due to the nonrepeatability of the calibration measurements caused by the silicone tube residual stress. This was not observed in the dynamic calibration method. The possible use of different tube material (for instance PVC) should be carefully studied.
Drop rate measurements with silicone tubes of 3/8 × 1/16 and 3/8 × 3/32 inches in diameters presented a decrease in measured values over time (p < 0.002) and angular coefficients were statistically equal (p > 0.56). The technique to measure drop rate was very susceptible to errors in their absolute values over time and presented great difficulty to reproduce the settings using this method. The cause of variations in drop rate over time was the residual stress in the silicone tubes tested.
These changes undermine the settings of the pumps for surgical procedures and do not guarantee the repeatability of the calibration. They also complicate comparisons between devices used in CPB that should be made with caution when the drop rate in silicone tubes is used as adjustment method.
The measurements of occlusion using the dynamic calibration method showed no significant changes in their values over time (p > 0.79) and are presented as a great approach to regulation.
In our opinion, drop rate method should not be used as an adjustment in roller pumps with silicone tubes and settings by static methods using different fluids may also change over time. It should be replaced by dynamic methods that do not demonstrate variations over time with the residual stress.
1.Boettcher W, Merkle F, Weitkemper HH: History of extracorporeal circulation: The invention and modification of blood pumps. J Extra Corpor Technol
35: 184–191, 2003.
2.McCaughan JS Jr, McMichael H, Schuder JC, Kirby CK: The use of a totally occlusive pump as a flowmeter with observations on hemolysis caused by occlusive and nonocclusive pumps and other pump-oxygenator components. Surgery
44: 210–219, 1958.
3.Bernstein EF, Gleason LR: Factors influencing hemolysis with roller pumps. Surgery
61: 432–442, 1967.
4.Edmunds LH Jr, Saxena NC, Hillyer P, Wilson TJ: Relationship between platelet count and cardiotomy suction return. Ann Thorac Surg
25: 306–310, 1978.
5.Morris KN, Kinross FM, Stirling GR: Hemolysis of blood in the pericardium: The major source of plasma hemoglobin during total body perfusion. J Thorac Cardiovasc Surg
49: 250–258, 1965.
6.Hargrove M, O'Donnell A, Aherne T: Differences in displayed pump flow compared to measured flow under varying conditions during simulated cardiopulmonary bypass. Perfusion
23: 227–230, 2008.
7.Tayama E, Teshima H, Takaseya T, et al
: Non-occlusive condition with the better-header roller pump: Impacts of flow dynamics and hemolysis. Ann Thorac Cardiovasc Surg
10: 357–361, 2004.
8.Tamari Y, Lee-Sensiba K, Leonard EF, Tortolani AJ: A dynamic method for setting roller pumps nonocclusively reduces hemolysis and predicts retrograde flow. ASAIO J
43: 39–52, 1997.
Copyright © 2010 by the American Society for Artificial Internal Organs
9.Mongero LB, Beck JR, Kroslowitz RM, et al
: Clinical evaluation by the dynamic method: Effect on flow. Perfusion
13: 360–368, 1998.