Perfusion 1998; 13: 360–368

Clinical evaluation of setting pump occlusion by the

dynamic method: effect on flow
LB Mongero Columbia-Presbyterian Medical Center, Milstein Hospital, New York, JR Beck, TW Orr,
RM Kroslowitz Department of Clinical Perfusion, Cardiothoracic Surgery K Lee-Sensiba and MC Oz
Department of Surgery, North Shore University Hospital, Manhasset, New York

Pump manufacturers recommend setting roller pump occlusion such that

the level of a 100 cm column of crystalloid drops 2.5 cm/min (Sarns™, 8000
Modular Perfusion System, operator’s manual, roller pump software version
2.3L. May 1993; 2.1–2.14). Though this almost occlusive setting ensures
accurate pump flow, it has been shown to cause more hemolysis than
nonocclusive pumps (Noon GP, Kane LE, Feldman L et al. Reduction of blood
trauma in roller pumps for long-term perfusion. World J Surg 1985; 9:
65–71). We conducted a clinical study (n = 19) to compare the standard
occlusion method with the dynamic method and to determine the accuracy
of flow for the nonocclusive pump.
Standard occlusion was set by clamping the pump tubing distal to the
arterial line filter and timing the drop in pump outlet pressure as indicated by
a pressure transducer connected to the filter. The occlusion setting,
expressed in mmHg/s, was recorded for each roller at two specific points
along the raceway. The pump was then set nonocclusively with the dynamic
method using the Better Header™ (BH) (Circulatory Technology, Oyster Bay,
NY, USA). Readings of the change in pressure in the same two selected
points on the raceway were taken. The latter was repeated after
discontinuation of bypass. Flow was recorded throughout the procedure
from both roller pump output display and a flow meter (Model #109
Transonic, Ithaca, NY, USA).
The average drop in pump outlet pressure for the standard method
was 1.3 ± 4.0 (range 0–18 mmHg/s), and for the dynamic method was
38 ± 28 (range 1.2–89 mmHg/s). Off bypass, the average reading was

Address for correspondence: LB Mongero, Columbia- Presented at the Annual Seminar of the American Academy of
Presbyterian Medical Center, Milstein Hospital, Building 4-350, Cardiovascular Perfusion, San Diego, California, 31 January–
New York, NY 10032, USA. 3 February 1997.

© Arnold 1998 0267–6591(98)PF246OA

Downloaded from prf.sagepub.com at UNIVERSITE LAVAL on June 29, 2015

 Clinical evaluation of setting pump occlusion by the dynamic method 361

44 ± 38 (range 2.0–103 mmHg/s). Regression analysis indicates that patient

flow, when corrected for retrograde flow by the dynamic method, equals
1.003 ´ revolutions per minute + 40 ml/min (r2 = 0.964). The average error
between indicated pump flow, corrected for retrograde flow, was –1% (range
from –6.7 to 6.6%). We conclude that the BH allows nonocclusive settings
(30 times less than our standard method) without sacrificing pump flow
accuracy.

Introduction Device description

The BH is a disposable safety device for a roller
The degree of occlusiveness of tubing through a pump that incorporates a pressure relief valve con-
roller pump is an important factor influencing both nected across a segment of standard pump tubing,6
forward fluid output and hemolysis. Though a less (Figure 2). It automatically limits the pressure at the
occlusive roller pump has been determined to cause outlet of the pump to a user-settable value and pro-
less hemolysis than the almost occlusive roller vides a simple, improved means to set the occlusion
pump,2,3 there is concern that a nonocclusive setting of the roller pump. Overpressurization is prevent-
may affect the accuracy of flow. Moreover, non- ed by diverting blood from the high pressure side
occlusive settings are difficult to obtain with the (outlet) of the circuit to the low pressure side
standard ‘drop rate’ technique. Except for identify- (inlet).
ing the exact point of occlusion, it is poorly repro-
ducible.4 It is the popular opinion that the properly Occlusion setting
set roller pump is just occlusive, a setting at which Once the circuit was primed with crystalloid and
a 75 cm column of fluid will fall 1 cm/min.3,5 Pump Hespan (final concentration 3%), the first perfu-
manufacturers recommend setting pump occlusion sionist set the pump occlusion by the pressure drop
such that the level of a 100 cm column of crystalloid method, whereby pump occlusion was determined
drops 2.5 cm/min. We conducted a clinical study (n by clamping the tubing at the pump outlet, rotating
= 19) to compare the standard occlusion method the rollers to create a nominal pump outlet pressure
with the dynamic method using the Better Header™ of 300 mmHg, and then timing the decrease in that
(BH) (Circulatory Technology, Oyster Bay, NY, pressure due to back-flow past the nearly occlusive
USA) and to determine the accuracy of flow for the rollers. A faster pressure drop rate corresponds to
nonocclusive pump. We also determined the varia- a greater gap and less occlusion. Pressure was mea-
tions in occlusion between two points along the sured with a mercury manometer, and its rate of
raceway, and compared occlusion before and after change timed with a stop-watch. For each patient,
bypass. four readings were measured and recorded: roller
A at 4 o’clock and at 8 o’clock and roller B at 4
o’clock and at 8 o’clock from the perfusionist’s van-
Materials and methods tage point.
The second perfusionist then set the pump
The cardiopulmonary bypass (CPB) circuit (Figure nonocclusively by the dynamic method using the BH
1) comprised standard PVC tubing, connectors, and as per instructions of the manufacturer. Briefly, the
stopcocks, a Sarns membrane oxygenator (Turbo, pressure relief valve (PRV) of the BH was pressur-
3M, Ann Arbor, MI, USA), a hard shell cardiotomy ized to 500 mmHg, the pump outlet tubing distal to
reservoir (Baxter, Irvine, CA, USA), a collapsible the PRV was clamped, and the pump speed set to
BMR-1900 venous reservoir (Baxter), arterial line fil- 5 revolutions per minute (rpm). Occlusion was
ter (Pall Filter Corporation, Glen Cove, NY, USA), adjusted until the valve was opened 50% of the
and the BH (Circulatory Technology, Oyster Bay). cycle. Pump speed was verified by measuring the
Flow was recorded throughout the procedure from time taken for two pump revolutions. Once set, the
both roller pump output display and a flow meter degree of occlusion was measured by the pressure
(Model #109, Transonic, Ithaca, NY, USA). drop method as described above. A final recording

Downloaded from prf.sagepub.com at UNIVERSITE LAVAL on June 29, 2015

362 LB Mongero et al.

Pressure pressurized the valve to 500 mmHg. If the valve just

From patient monitor opens, that means that at the pressure required to
To patient
500 open the valve (i.e. 500 mmHg) the forward flow gen-
mmHg erated by the pump equals the retrograde flow.
Hence, during the case, the maximum retrograde
flow at a line pressure of 500 mmHg would be the
Compliance same flow used to set the pump nonocclusive. Dur-
chamber
Venous ing the case, when the pump is rotating at 100 rpm,
reservoir the maximum error would be 5 rpm and the user can
Stopcock increase pump speed by 5 rpm to compensate for that
retrograde flow. In fact, this correction could be
lower because the viscosity of blood is higher than
that of the prime solution, and, therefore, less retro-
Oxygenator grade flow would result.

In vitro pump calibration

The Sarns/3M 8000 positive displacement pump
was calibrated for 1/2² inner diameter polyvinyl
chloride (PVC) tubing ´ 3/32² wall thickness tubing,
with water pumped through the system at different
Better header pump speeds. The occlusion was set according to
in arterial pump
manufacturers specifications using the standard
Figure 1 The bypass circuit used for this study occlusion method. At each speed, the pump flow,
indicated by the pump meter, was recorded and
compared to the flow calculated from timed volu-
of the degree of occlusion was made postbypass metric collection.
using the pressure drop method, with the residual
pump volume which had a hematocrit of approxi-
mately 25%. The time to set pump occlusion with Results
the static pressure drop method and with the
dynamic method was determined by timing each Comparison of occlusion settings obtained
procedure as it was being performed. Nineteen with the pressure drop and dynamic methods
patients were studied. Table 1 summarizes the data obtained for the pres-
sure drop rates for each roller at the two positions
Correction for retrograde flow and the average of the four points taken. These data
With the BH, pump occlusion is set using the were also plotted as bar graphs. The data clearly
dynamic method. With this method, the pump out- illustrate that there is a difference in the occlusion
let is clamped off, the pump is turned on at a slow between 4 and 8 o’clock on both roller A and B,
speed (in our case 5 rpm), and occlusion is set until with 4 o’clock being less occlusive. The differences
the valve starts to open. If the valve is not opening at were not statistically significant (t-test, p > 0.05) for
all, it indicates that whatever forward flow the pump the static method (Figure 3), but were statistically
generates, is going retrograde. This means that the significant with the dynamic method (Figure 4).
gap between the roller and the raceway is too open When the equivalent pressure drop rate was mea-
(too nonocclusive). If the valve is permanently open, sured after setting the occlusion with the dynamic
it indicates that whatever forward flow the pump gen- method, the dynamic setting was over 30 times less
erates is shunting from pump outlet across the valve occlusive (higher drop rate) than that obtained with
to pump inlet. This means that the gap between the the static drop methods for each roller and at each
roller and the raceway is too closed (too occlusive), position (Figure 5) and when averaged for both
and the pressure build up by the pump overcomes the rollers and locations along the raceway (Figure 6).
pressure required to open the valve. In our case, we When comparing the occlusion setting measured

Downloaded from prf.sagepub.com at UNIVERSITE LAVAL on June 29, 2015

 Clinical evaluation of setting pump occlusion by the dynamic method 363

Table 1 Summary of occlusion settings

Static prebypass Dynamic prebypass Dynamic postbypass

Roller/location Avg SD Max Min Avg SD Max Min Avg SD Max Min
A 4 o’clock 2.1 4.2 18.1 0.0 49 22 89 20 61 32 112 6
A 8 o’clock 0.6 1.1 4.5 0.0 29 17 69 6 36 23 90 8
B 4 o’clock 1.2 2.3 8.6 0.0 43 22 86 12 55 24 97 12
B 8 o’clock 0.3 0.3 1.1 0.0 20 19 66 1.2 23 17 59 2
Avg for the points 1.0 2.5 18.1 0.0 36.5 23 89 1.2 44 29 112 2
Avg, average; SD, standard deviation; max, maximum; min, minimum; A, roller A; B, roller B; 4 o’clock, location along raceway
corresponding to 4 o’clock, etc.

2.4
Pressure 500
2.0 Occlusion setting
monitor
static method prebypass
Drop rate (mmHg/s) 1.6

1.2
Compliance
chamber 0.8 p<0.072 p<0.52

0.4

0.0
A–4 A–8 B–4 B–8
o'clock o'clock o'clock o'clock
Roller and its position

,,,,,
Figure 3 The average pressure drop rate measured with the
static method prebypass
Pset

,,
PRV
Po

50 Occlusion setting
dynamic method prebypass
40
Drop rate (mmHg/s)

30

20

10 p<0.027 p<0.003

0
A–4 A–8 B–4 B–8
o'clock o'clock o'clock o'clock
Roller and its position
Figure 4 The average occlusion settings as measured by
pressure drop rate obtained with the dynamic method
Figure 2 The Better-Header™ prebypass

Downloaded from prf.sagepub.com at UNIVERSITE LAVAL on June 29, 2015

 364 LB Mongero et al.

Static versus Dynamic occlusion setting (drop rate) Pre versus Post dynamic occlusion setting
60 80
Static method Prebypass
50 Dynamic method

Drop rate (mmHg/s)

60 Postbypass
Drop rate (mmHg/s)

40
40
30

20 20

10
0
A–4 A–8 B–4 B–8
0
o'clock o'clock o'clock o'clock
A–4 A–8 B–4 B–8
o'clock o'clock o'clock o'clock Roller and its position
Roller and its position Figure 7 The average occlusion settings as measured by
pressure drop rate obtained with the dynamic method pre-
Figure 5 The average occlusion settings as measured by bypass and postbypass
pressure drop rate obtained with the static method pre-
bypass compared with the dynamic method for each roller
at 4 and 8 o’clock

40 prebypass to postbypass there seemed to be a ten-

dency for a decrease in occlusion (higher drop rate)
35 during the case, but this difference was not statisti-
Pressure drop rate (mmHg/s)

cally significant (Figure 7).

30
Table 2 summarizes the p values obtained from
25 the t-tests comparing various occlusion positions.

20 Comparison of time to set occlusion by the two

methods
15 The data for the time to set occlusion properly
10 are summarized in Figure 8. They show that
the average time to set occlusion by the static
5 method was 161 ± 89 s as compared to 42 ± 8 s with
the dynamic method. Similar large differences
0 were observed for the minimum time (82 versus
Static Dynamic
30 s) and maximum time (377 versus 57 s) used to
Occlusion method set occlusion.
Figure 6 The average occlusion settings as measured by
pressure drop rate obtained for all four points for the static
method compared to the dynamic method

Table 2 Summary of p values from t-tests

A4 versus B4 A4 versus A8 B4 versus B8 Prebypass versus postbypass

Static method prebypass 0.079 0.0706 0.5183
Dynamic method prebypass 0.0722 0.0026 0.0034
Dynamic method postbypass 0.3002 0.0346 0.0318 0.338
A4 versus B4, Roller A at 4 o’clock versus Roller B at 4 o’clock; A4 versus A8, Roller A at 4 o’clock versus Roller A at 8 o’clock;
B4 versus B8, Roller B at 4 o’clock versus Roller B at 8 o’clock.

Downloaded from prf.sagepub.com at UNIVERSITE LAVAL on June 29, 2015

 Clinical evaluation of setting pump occlusion by the dynamic method 365

400 Effect of nonocclusive setting on true pump

flow
350 Static
Figure 9 summarizes the data collected for flows
Dynamic
as a plot of pump meter readings versus flow meter
300
Time to set occlusion (s)

readings. Two sets of points are shown: the direct

250
readings from the pump readout, and the correct-
ed pump readout (pump readout less 5 rpm or 0.23
200 l/min). The line of identity is also shown. These
data are also plotted as the error in flow (l/min)
150 (Figure 10) and as a percentage of the flow meter
(Figure 11). These data indicate that the error
100 with corrected flow usually underestimates the flow
(4 ± 4%), whereas the uncorrected flow usually
50 overestimates the flow (3 ± 4%). The largest over-
0
estimation of flow with the corrected readings was
Average Minimum Maximum 6.7% compared to 17% for the uncorrected flow. It
is of interest to note that the error increases as flow
Figure 8 The time used to set pump occlusion by the static
method and the dynamic method decreases. Table 3 summarizes the data collected for
the pump flow, flow meter, the corrected pump flow
and the corresponding errors.

Indicated versus measured patient flow

6.00

Uncorrected flow
5.00 Corrected flow

4.00
Pump meter (l/min)

3.00

2.00

1.00

0.00
0.00 1.00 2.00 3.00 4.00 5.00 6.00
Flow meter (l/min)
Figure 9 The pump meter readout (corrected and uncorrected) as a function of flow meter readings

Downloaded from prf.sagepub.com at UNIVERSITE LAVAL on June 29, 2015

366 LB Mongero et al.

Error in patient flow

0.60

Uncorrected flow
Corrected flow
0.40
Error in flow indicated by pump meter (l/min)

0.20

Overestimate

0.00

Underestimate

-0.20

-0.40

-0.60
0.00 1.00 2.00 3.00 4.00 5.00 6.00
Flow meter (l/min)

Figure 10 The errors in patient flow as indicated by the pump meter readout (corrected and uncorrected) as a function of
flow meter readings

Discussion when higher viscosity is used to measure occlusion

setting. In fact, a decrease in occlusion setting was
The reasons we use centrifugal pumps are for the measured postbypass (higher drop rate). Though
safety of not blowing a line, should it become inad- not statistically significant (Figure 7), this may have
vertently kinked or clamped, and to reduce blood been caused by relaxation, or very slight thinning of
damage.6–9 Roller pumps are simple, reliable, and the tube wall, thereby increasing the gap.
proven for extracorporeal pumping.10,11 Extreme The effect of a nonocclusive setting on pump flow
overpressurization can cause damage to the circuit is a major concern. Our data indicated that the error
and disrupt flow to the patient. The features of the with corrected flow usually underestimates the flow,
pressure relief valve incorporated in the BH limit whereas the uncorrected flow usually overesti-
line pressure and eliminate this discontinuation of mates the flow. This is most likely because flow was
flow to the patient adding a measure of safety to the corrected assuming the worst case, a line pressure
bypass run. The BH also allows setting the roller of 500 mmHg and crystalloid prime with lower vis-
pump to nonocclusive, a setting that has been shown cosity. Since the line pressure decreases as flow
by Noon et al.2 to significantly reduce hemolysis. decreases, the expected retrograde flow should also
It should be pointed out that the occlusion setting decrease.
at the end of CPB, measured with blood, should Hemolysis has been studied by various authors
have increased when compared to the preCPB set- with respect to the degree of occlusiveness.12–15 There
ting measured with the crystalloid. This is the case is interesting evidence to show that a nonocclusive

Downloaded from prf.sagepub.com at UNIVERSITE LAVAL on June 29, 2015

 Clinical evaluation of setting pump occlusion by the dynamic method 367

20% Error in patient flow

15%
Uncorrected flow
Corrected flow
Error in flow indicated by pump meter (l/min)

10%

5%
Overestimate

0%

Underestimate

-5%

-10%

-15%

-20%
0.00 1.00 2.00 3.00 4.00 5.00 6.00

Figure 11 The percentage of errors in patient flow as indicated by the pump meter readout (corrected and uncorrected) as
a function flow meter readings

Table 3 Summary of the flow measurements and respective errors

Pump flow Flow meter Pump flow Error in corrected Error in uncorrected
reading (l/min) reading (l/min) corrected (l/min) flow (%) flow (%)
Avg 4.24 4.12 4.01 –3.8 2.9
SD 0.85 0.83 0.85 4.0 4.2
Max 5.88 5.71 5.65 6.7 17.1
Min 2.10 2.18 1.87 –14 –5.3
Avg, average; SD, standard deviation; max, maximum; min, minimum.

roller pump should be used to eliminate excess Conclusion

hemolysis on CPB. The reproducibility of setting
occlusions is of concern according to a study recent- We conclude that the BH allows nonocclusive set-
ly published by Rawn et al.12 and this study has shown tings (30 times less than our standard method) with-
a variability in setting occlusion from roller to roller. out sacrificing pump flow accuracy.
Perfusionists now have the ability to quickly and The BH has the ability to achieve more
accurately set and alter occlusion and obtain reproducible, nonocclusive settings in a simple
repeatable nonocclusive settings by a new dynamic manner, thereby eliminating variations in tubing
method using the BH. wall thickness, roller extension, and raceway

Downloaded from prf.sagepub.com at UNIVERSITE LAVAL on June 29, 2015

368 LB Mongero et al.

symmetry, which all affect the accuracy of the 8 Englehardt H, Vogelsang B, Reul H, Rau G.
static drip rate technique. Hydrodynamical and hemodynamical evaluation of
rotary blood pumps. Thoma, Schima eds.
Proceedings of the International Workshop on Rotary
Blood pumps, Vienna: 1988.
References 9 Oku T, Harasaki H, Smith W, Nose Y. Hemolysis:
a comparative study of four nonpulsatile pumps.
1 Sarns™, 8000 Modular Perfusion System, operator’s Trans Am Soc Artif Intern Organs 1988; 34: 500–504.
manual, roller pump software version 2.3L. May 10 Lee-Sensiba KJ, Hymowitz D, Kay S et al. New roller
1993; 2.1–2.14. pump disposable provides safety and simplifies
2 Noon GP, Kane LE, Feldman L et al. Reduction occlusion setting. AmSECT Proceedings, 1996.
of blood trauma in roller pumps for long-term 11 Tamari Y, Lee-Sensiba K, Leonard, EF. The effects
perfusion. World J Surg 1985; 9: 65–71. of pressure and flow on hemolysis caused by
3 Lee-Sensiba KJ, Tortolani AJ, King RS et al. A new BioMedicus centrifugal pumps and roller pumps:
method for setting roller pump occlusion lowers guidlines for choosing a pump. J Thorac Cardiovasc
hemolysis. AmSECT Proceedings, 1995. Surg 1993; 106: 997–1007.
4 Reed CC, Clark DK. Bubble oxygenators. Houston, 12 Rawn DJ, Yoda DN, Harris DK et al. An under
TX: Texas Medical Press, 1975: 248–49. occluded roller pump is less hemolytic than centrifugal
5 Hessel EA. Cardiopulmonary bypass circuitry and pump [Abstract]. AmSECT Proceedings, 1995.
cannulation techniques. In: Gravlee GP, Davis RF, 13 Muller MR, Wohlfahrt A, Schima H et al.
Utley JR eds. Cardiopulmonary bypass: principles Hematological parameters for in vitro testing of
and practice. Baltimore, MD: Williams &Wilkins, blood pumps data and analysis of two different
1993: 55–92. principles. Proceedings of the International Workshop
6 Hoerr HR, Kraimer MF, Wiliams JL et al. In vitro on Rotary Bloodpumps, Vienna: 1988.
comparison of the blood handling by the 14 Bernstein EF, Gleason LR. Factors influencing
constrained vortex and twin roller blood pumps. hemolysis with roller pumps. Surgery 1967; 61: 432–42.
J ExtraCorporeal Tech 1987; 19: 316–21. 15 McCaughan JS, McMichael H, Schuder JC et al.
7 Horton AM, Butt W. Pump-induced haemolysis: The use of a totally occlusive pump as a flowmeter
is the constrained vortex pump better or with observations on hemolysis caused by occlusive
worse than the roller pump? Perfusion 1992; and nonocclusive pumps and other pump-
7: 103–108. oxygenator components. Surgery 1958; July: 210–19.

Downloaded from prf.sagepub.com at UNIVERSITE LAVAL on June 29, 2015