in vitro performance analysis of a novel pulsatile diagonal pump in a simulated pediatric mechanical...

In Vitro Performance Analysis of a Novel Pulsatile DiagonalPump in a Simulated Pediatric Mechanical Circulatory

Support System

*Shigang Wang, †Allen R. Kunselman, and *‡§Akif Ündar

*Pediatric Cardiovascular Research Center, Penn State Hershey Children’s Hospital, Penn State Milton S. Hershey MedicalCenter, Department of Pediatrics, Penn State Hershey College of Medicine; †Department of Public Health Sciences, Penn

State Hershey College of Medicine; ‡Department of Surgery, Penn State Milton S. Hershey Medical Center, Penn StateHershey College of Medicine, Pennsylvania State University, Hershey; and §Department of Bioengineering, College of

Engineering, Pennsylvania State University, University Park, PA, USA

Abstract: The objective of this study was to evaluate thepump performance of the third-generation Medos diagonalpump, the Deltastream DP3, on hemodynamic profile andpulsatility in a simulated pediatric mechanical circulatorysupport (MCS) system. The experimental circuit consistedof a Medos Deltastream DP3 pump head and console(MEDOS Medizintechnik AG, Stolberg, Germany), a14-Fr Terumo TenderFlow Pediatric arterial cannula anda 20-Fr Terumo TenderFlow Pediatric venous returncannula (Terumo Corporation, Tokyo, Japan), and 3 ft oftubing with an internal diameter of 1⁄4 in. for both arterialand venous lines. Trials were conducted at flow ratesranging from 250 mL/min to 1000 mL/min (250-mL/minincrements) and rotational speeds ranging from 1000 to4000 rpm (1000-rpm increments) using human blood(hematocrit 40%). The postcannula pressure was main-

tained at 60 mm Hg by a Hoffman clamp. Real-timepressure and flow data were recorded using a Labview-based acquisition system. The pump provided adequatenonpulsatile and pulsatile flow, created more hemodynamicenergy under pulsatile mode, and generated higher positiveand negative pressures when the inlet and outlet of thepump head, respectively, were clamped. After the conver-sion from nonpulsatile to pulsatile mode, the flow rates andthe rotational speeds increased. In conclusion, the novelMedos Deltastream DP3 diagonal pump is able to supplythe required flow rate for pediatric MCS, generateadequate quality of pulsatility, and provide surplus hemo-dynamic energy output in a simulated pediatric MCSsystem. Key Words: Pulsatile flow—Diagonal pump—Pediatric mechanical circulatory support—Pediatrics.

Left/right ventricular mechanical supports are life-saving devices for terminal heart failure patients.They serve as a bridge to recovery or cardiac trans-plantation, or as “destination therapy” in somepatients who cannot undergo transplantation (1–4).The mechanical circulatory support (MCS) systemincludes a blood pump, arterial/venous cannulae, and

tubing. Shorter tubing and a more compact pump canbe used in a MCS system. Because of the absence ofthe hollow-fiber membrane oxygenator in the MCSsystem, lower circuit resistance allows a wide varietyof blood pumps to be used in MCS systems, such aspneumatic pumps, rotary pumps (including centrifu-gal pumps, diagonal pumps, and axial pumps), andeven roller pumps (5,6). Different pumps have differ-ent characteristics (4). Pneumatic pumps have inher-ent pulsatile flow for single or biventricular support,different stroke volumes suitable for different bodyweights, and excellent durability and mobility forshort/long-term support. However, the mechanicalvalves used to maintain unidirectional flow carry apotential risk for valve-related thrombosis. Axial-flow pumps are suitable for ventricular support due

doi:10.1111/aor.12181

Received May 2013; revised July 2013.Address correspondence and reprint requests to Dr. Akif

Ündar, Penn State Hershey College of Medicine, Department ofPediatrics—H085, 500 University Drive, P.O. Box 850; Hershey, PA17033-0850, USA. E-mail: [email protected]

Presented in part at the 9th International Conference on Pedi-atric Mechanical Circulatory Support and Pediatric Cardiopulmo-nary Bypass held, May 8–11, 2013 in Hershey, PA, USA.

bs_bs_banner

© 2013 Wiley Periodicals, Inc. and International Center for Artificial Organs and Transplantation

Artificial Organs 2014, 38(1):64–72

mailto:[email protected]

to their smaller size, implantability, and high flowoutput for continuous nonpulsatile blood flow. Theirdisadvantages, including hemolysis, thrombosis, and apotential risk for creating negative intraventricularpressure, limit their widespread use in long-term cir-culatory support. There are various types of centrifu-gal pumps available for MCS systems. Advantages ofcentrifugal pumps include low priming volumes,smaller size, better transportability, lower blooddamage, and higher safety. Technical improvementsin centrifugal pumps allow patients to undergo longersupport with fewer bleeding and thromboemboliccomplications. Nevertheless, few of these pumps canprovide pulsatile flow. Pulsatile flow has been consid-ered to be able to generate significantly higherhemodynamic energy compared with nonpulsatileflow during long-term support (7).

The Medos Deltastream DP3 is the latest-generation diagonal pump of the Medos Companyfor extracorporeal life support (ECLS) and MCS.Theredesigned DP3 diagonal pump has a separatedpump head, a magnetic coupling of forces betweenthe impeller and the driver, a high-tech ceramicbearing to improve durability and reduce bloodtrauma, low priming volume (16 mL) to minimizehemodilution and reduce the blood–artificial surfaceinterface, a large application spectrum ranging fromneonates to adults, zero-flow mode to preventbackflow, a flow-measuring sensor with integratedbubble detector for safety, and an optional pulsatileflow mode to mimic physiological flow conditions(8). It has been used in clinical venovenous extra-corporeal membrane oxygenation (ECMO), invenoarterial ECMO, and as short-term circulatorysupport in Europe (9,10). Our previous study dem-onstrated that this pump can provide adequatequality of pulsatility and generate more hemo-dynamic energy under pulsatile mode in a simulatedpediatric ECLS system (11,12). The objective of thisstudy was to evaluate the pump performance of theMedos Deltastream DP3 on hemodynamic profileand pulsatility in a simulated pediatric MCS system.

MATERIALS AND METHODS

Experimental circuitThe experimental circuit consisted of a Medos

Deltastream DP3 pump head and console (MEDOSMedizintechnik AG, Stolberg, Germany), a 14-FrTerumo TenderFlow Pediatric arterial cannula(Terumo Corporation, Tokyo, Japan), a 20-Fr TerumoTenderFlow Pediatric venous return cannula, and 3 ftof tubing with an inner diameter of 1⁄4 in. for botharterial and venous lines (Fig. 1). The circuit was

primed with lactated Ringer’s solution and packedhuman red blood cells.The hematocrit of the primingsolution was 40%. The total priming volume of thecircuit was 500 mL. A 300-mL-capacity soft bagserved as a pseudopatient. An open 1000-mL-capacity soft bag was used to control the venous linepressure. A Hoffman clamp was placed downstreamof the arterial cannula to maintain a steady arterialpressure during all trials.

Experimental design

Flow rate testTrials were conducted at flow rates ranging from

250 to 1000 mL/min (250-mL/min increments) in pul-satile and nonpulsatile modes at room tempera-ture. The postcannula pressure was maintained at60 mm Hg during all trials. The venous pressure washeld at 10 mm Hg.

FIG. 1. Experimental design of the MCS circuit.

NOVEL PULSATILE DIAGONAL MCS PUMP 65

Artif Organs, Vol. 38, No. 1, 2014

Pump testTrials were conducted at rotational speeds ranging

from 1000 to 4000 rpm (1000-rpm increments) in pul-satile and nonpulsatile modes. Two Hoffman clampswere used at both sides of the pump. For the firststep, tubing upstream and downstream of theDeltastream DP3 was totally clamped, and thenthe Hoffman clamp was gradually released to reducethe tubing resistance until the clamp was completelyopened.

When used in pulsatile mode, the Deltastream DP3pump was set to a frequency of 90 beats per minute(bpm), a systole/diastole ratio of 70/30 in one cycle,and speed differential values of 500 rpm (P500),1500 rpm (P1500), or 2500 rpm (P2500). Speed differ-ential refers to the maximum change in rotationalspeed relative to the average rotational speed.

Data acquisitionTwo Transonic ultrasound flow probes (Transonic

Systems, Inc., Ithaca, NY, USA) were placed at thepostpump and postcannula sites. Five Maxximdisposable pressure transducers (Maxxim Medical,Inc., Ithaca, NY, USA) were placed at the prepump,postpump, precannula, postcannula, and venoustubing sites, respectively. The pressure transducersand flowmeter outputs were connected to a signalconditioning unit (SC-2345, National Instruments,Austin, TX, USA), linked with a data acquisitiondevice (NI USB-6521, National Instruments), andfinally connected to a computer via USB port. Acustomized user interface based on Labview 7.1software for Windows (National Instruments) wasdesigned to record real-time data at 1000 samples persecond. A 20-s segment of pressure and flow wave-forms was recorded at all sites. The flow rate test wasrepeated six times for each unique combination,yielding a total of 96 trials. The pump test was con-ducted for a total of 220 trials.

The hemodynamic energy created by pulsatile flowwas quantified using related mathematical formulas.With the help of the Shepard’s energy equivalentpressure (EEP) formula and simultaneous blood flow(f) and pressure (p) recorded by Labview software,EEP, surplus hemodynamic energy (SHE), and totalhemodynamic energy (THE) were calculated pertime interval (time between t1 and t2) as follows (13).The constant 1332 converts pressure from units ofmm Hg to dyn/cm2 (1 mm Hg = 1332 dyn/cm2).

EEP mm Hg( ) = ∫ ∫fpdt fdtt

t

t

t

1

2

1

2

SHE erg cm EEP mean pressure3 1332( ) = ∗ −( )

THE erg cm EEP3 1332( ) = ∗

Statistical analysisAnalysis of variance (ANOVA) models were fitted

to the continuous outcomes (e.g., mean pressure) tocompare flow rate (250, 500, 750, 1000 mL/min) andmode (nonpulsatile, P500, P1500, and P2500). Alinear mixed-effects model was fit to the continuousoutcomes (e.g., mean pressure, EEP, SHE, and THE)to compare the flow rate, mode, and location (e.g.,prepump, postpump, postcannula) (14). The linearmixed-effects model is an extension of linear regres-sion that accounts for the within-subject variabilityinherent in repeated-measures designs. In this study,the repeated factor is the location. For each outcome,P values were adjusted for multiple comparisonstesting using the Tukey–Kramer procedure. Allhypothesis tests were two-sided, and all analyseswere performed using version 9.3 of the SAS Systemfor Windows (SAS Institute, Inc., Cary, NC, USA).

RESULTS

Flow ratesThe Deltastream DP3 could easily provide flow

rates of 250–1000 mL/min in nonpulsatile mode atrotational speeds of 3150–4600 rpm in our simulatedpediatric MCS circuit. It is worth noting that rota-tional speeds and flow rates automatically increasedwhen the pump was switched from nonpulsatilemode to pulsatile mode. To maintain consistent flowrates between the two modes, rotational speed forpulsatile mode was manually adjusted to the samerotational speed that was observed in nonpulsatilemode. After this adjustment was made, when thepump was switched from pulsatile mode back tononpulsatile mode, rotational speed decreased.

Our results showed that the flow rates and rota-tional speeds increased by 6–95.8% and 4.3–58.7%,respectively, after the pump was switched to pulsatilemode, with speed differential values of 500 rpm,1500 rpm, and 2500 rpm (Table 1). The highest per-centage increase in pump flow rate (95.8%) occurredat 250 mL/min flow and 2500 rpm in pulsatile mode.Meanwhile, the highest percentage increase in rota-tional speed (58.7%) was also at 250 mL/min in pul-satile mode. Interestingly, when rotational speeds inpulsatile mode were reduced to the same speed asobserved in nonpulsatile mode, flow rates werealmost the same as in nonpulsatile mode. However,after the pump was switched back to nonpulsatilemode, the flow rates and rotational speeds decreasedby 5.6–51.8% and 4.3–38.0%, respectively (Table 1).

S. WANG ET AL.66


Figure 2 shows flow–pressure waveforms at750 mL/min of flow rate in pulsatile (P500 and P2500)and nonpulsatile modes. Increased speed differentialvalues could create more pulsatility. In the presentstudy, all data related to mean pressure (MP), EEP,SHE, and THE in pulsatile mode were gathered afterrotational speeds were adjusted to match theobserved rotational speed in nonpulsatile mode. Nobackflow was found during any of the trials.

Mean pressures and energy equivalent pressuresTable 2 presents the MP and EEP at different flow

rates using different perfusion modes. The rotationalspeeds in pulsatile mode were reduced to match theobserved rotational speed at a given flow rate innonpulsatile mode. In nonpulsatile mode, EEP wasthe same as MP at the postpump and postcannulasites. When the pump was switched to pulsatile flow,EEP was always higher than MP. Higher EEP wasrecorded at flow rates of 250 mL/min and 500 mL/min, with a speed differential value of 1500 rpm. Atflow rates of 750 mL/min and 1000 mL/min, EEP washigher, with a speed differential value of 2500 rpm.With increased flow rates, MP and EEP increased,while the values of the prepump pressures becamemore negative.

Hemodynamic energyTable 3 presents hemodynamic energy changes

for pulsatile (P500, P1500, P2500) and nonpulsatilemodes. No SHE was observed in nonpulsatile mode.The highest SHE was generated at the postpumpsite, with 250 mL/min of pump flow in pulsatilemode using a speed differential value of 1500 rpm.Meanwhile, the lowest SHE was presented at1000 mL/min in pulsatile mode (P500). At lowerflow rates (250 mL/min and 500 mL/min), pulsatileflow with a speed differential value of 1500 rpmgenerated the highest SHE at the postpump andpostcannula sites. Meanwhile, at higher flow rates(750 mL/min and 1000 mL/min), pulsatile flow witha speed differential value of 2500 rpm generated thehighest SHE. When compared with nonpulsatileflow, total THE increased by 2.9–8.8% (P500), 7.8–9.7% (P1500), and 7.6–9.8% (P2500) in pulsatilemode at the postpump site. Pulsatile flow with aspeed differential value of 2500 rpm generatedhigher THE at the postpump and postcannula sitesexcept at a flow rate of 500 mL/min.

Pump evaluationFigure 3 shows the postpump flow–pressure curves

at different rotational speeds with pulsatile and

TAB

LE

1.C

hang

esin

rota

tiona

lsp

eed

and

flow

rate

betw

een

puls

atile

and

nonp

ulsa

tile

mod

es

Per

fusi

onm

ode

250

mL

/min

500

mL

/min

750

mL

/min

1000

mL

/min

Rot

atio

nal

spee

d(r

pm)

Flo

wra

te(m

L/m

in)

Rot

atio

nal

spee

d(r

pm)

Flo

wra

te(m

L/m

in)

Rot

atio

nal

spee

d(r

pm)

Flo

wra

te(m

L/m

in)

Rot

atio

nal

spee

d(r

pm)

Flo

wra

te(m

L/m

in)

NP

3150

261.

436

0051

4.4

4050

760.

546

0010

19.6

P50

033

50(6

.3%

)29

0.5

(11.

1%)

3800

(5.6

%)

555.

4(8

.0%

)42

50(4

.9%

)81

4.2

(7.1

%)

4800

(4.3

%)

1081

.2(6

.0%

)P

500→

RP

M31

5025

6.9

(−1.

7%)

3600

516.

4(0

.4%

)40

5076

5.2

(0.6

%)

4600

1023

.3(0

.4%

)P

500→

NP

2950

(−6.

3%)

235.

8(−

9.8%

)22

50(−

5.6%

)47

5.3

(−7.

6%)

3850

(−4.

9%)

711.

3(−

6.5%

)44

00(−

4.3%

)96

2.2

(−5.

6%)

P15

0041

50(3

1.7%

)39

4.6

(51.

0%)

4500

(25.

0%)

698.

9(3

5.9%

)49

00(2

1.0%

)98

1.1

(29.

0%)

5400

(17.

4%)

1251

.1(2

2.7%

)P

1500

→R

PM

3150

260.

7(−

0.3%

)36

0051

9.0

(0.9

%)

4050

766.

7(0

.8%

)46

0010

30.2

(1.0

%)

P15

00→

NP

2000

(−36

.5%

)12

6.1

(−51

.8%

)25

50(−

29.2

%)

311.

8(−

39.4

%)

3050

(−24

.7%

)51

6.2

(−32

.1%

)37

00(−

19.6

%)

761.

6(−

25.3

%)

P25

0050

00(5

8.7%

)51

1.8

(95.

8%)

5350

(48.

6%)

867.

3(6

8.6%

)57

50(4

2.0%

)11

91.1

(56.

6%)

6150

(33.

7%)

1461

.0(4

3.3%

)P

2500

→R

PM

3150

263.

2(0

.7%

)36

0051

0.8

(−0.

7%)

4100

775.

6(2

.0%

)46

0010

24.6

(0.5

%)

P25

00→

NP

2000

(−36

.5%

)12

6.0

(−51

.8%

)23

00(−

36.1

%)

264.

4(−

48.6

%)

2550

(−37

.0%

)39

8.5

(−47

.6%

)28

50(−

38.0

%)

528.

0(−

48.2

%)

All

perc

enta

ges

inpa

rent

hesi

sar

ere

lati

veto

nonp

ulsa

tile

mod

e.N

P,no

npul

sati

lem

ode;

P,pu

lsat

ilem

ode;

P→

RP

M,r

educ

edto

NP

valu

e;P

→N

P,sw

itch

edto

nonp

ulsa

tile

mod

e.



nonpulsatile perfusion. At low rotational speed(1000 rpm), increased speed differential values hadminimal effect on pressures and flow rates, whileat high rotational speed (4000 rpm), increased

speed differential values did create higher pres-sures and flow rates. At 4000 rpm, pulsatile flowwith a speed differential value of 2500 rpm gene-rated a mean postpump pressure of 235.0 mm Hg

FIG. 2. Flow-pressure waveform at 750 mL/min in pulsatile (P500 & P2500) and nonpulsatile modes. NP, nonpulsatile mode; P, pulsatilemode; P→RPM, reduced to NP value; P→NP, switched to nonpulsatile mode.

TABLE 2. Mean pressures and energy equivalent pressures at different flow rates in pulsatile and nonpulsatile modes

Flow rate Mode

Postpump site Postcannula site

Prepump MPMP EEP MP EEP

250 mL/min NP 69.2 ± 0.1 69.2 ± 0.1 61.0 ± 0.1 61.0 ± 0.1 −4.1 ± 0.1P500 70.6 ± 0.2* 75.3 ± 0.2* 62.4 ± 0.2* 66.8 ± 0.2* −4.2 ± 0.1P1500 69.3 ± 0.3 75.6 ± 0.2* 60.7 ± 0.3 66.5 ± 0.2* −4.3 ± 0.1P2500 69.9 ± 1.0* 76.0 ± 0.6* 61.5 ± 0.8 67.1 ± 0.5* −4.4 ± 0.1*

500 mL/min NP 77.4 ± 0.0 77.4 ± 0.0 59.9 ± 0.0 59.9 ± 0.0 −13.9 ± 0.0P500 78.9 ± 0.1* 81.5 ± 0.1* 61.5 ± 0.1* 63.7 ± 0.1* −14.1 ± 0.1P1500 80.3 ± 0.1* 84.9 ± 0.1* 62.6 ± 0.1* 66.5 ± 0.1* −14.4 ± 0.1*P2500 78.7 ± 0.2* 83.2 ± 0.3* 61.2 ± 0.2* 65.0 ± 0.2* −14.1 ± 0.1

750 mL/min NP 88.5 ± 0.0 88.5 ± 0.0 60.5 ± 0.1 60.5 ± 0.1 −23.9 ± 0.0P500 89.8 ± 0.1* 91.8 ± 0.2* 61.2 ± 0.1* 62.7 ± 0.1* −24.3 ± 0.1*P1500 90.9 ± 0.3* 95.3 ± 0.3* 62.1 ± 0.2* 65.3 ± 0.2* −24.7 ± 0.1*P2500 92.4 ± 0.2* 96.9 ± 0.2* 63.2 ± 0.1* 66.5 ± 0.1* −25.1 ± 0.1*

1000 mL/min NP 103.1 ± 0.0 103.1 ± 0.0 60.7 ± 0.0 60.7 ± 0.0 −38.5 ± 0.0P500 104.5 ± 0.2* 106.1 ± 0.2* 61.7 ± 0.1* 62.7 ± 0.1* −38.8 ± 0.1*P1500 107.2 ± 0.3* 112.0 ± 0.2* 63.7 ± 0.2* 66.8 ± 0.1* −39.4 ± 0.2*P2500 106.8 ± 0.3* 111.9 ± 0.3* 63.7 ± 0.2* 67.0 ± 0.2* −38.8 ± 0.2*

* P < 0.05 vs. nonpulsatile mode.NP, nonpulsatile mode; P500, pulsatile speed differential value 500 rpm; P1500, pulsatile speed differential value 1500 rpm; P2500, pulsatile

speed differential value 2500 rpm.

S. WANG ET AL.68


after clamping downstream of the pump head.Nonpulsatile flow generated a mean postpumppressure of 116.1 mm Hg. Figure 4 presents prepumpflow–pressure curves at different rotational speeds

with pulsatile and nonpulsatile perfusion. Increasedrotational speeds and increased speed differentialvalues generated more negative prepump pressures.At 4000 rpm, pulsatile flow with a speed differential

TABLE 3. Hemodynamic energy at different flow rates in pulsatile and nonpulsatile modes

Flow rate Mode

SHE (erg/cm3) THE (erg/cm3)

Postpump Postcannula Postpump Postcannula

250 mL/min NP 0.0 ± 0.2 0.2 ± 0.3 92 168.8 ± 68.7 81 275.4 ± 121.0P500 6279.2 ± 115.0* 5863.4 ± 111.7* 100 296.2 ± 203.0*(8.8%) 88 986.2 ± 217.5* (9.5%)P1500 8310.1 ± 140.8* 7628.9 ± 136.6* 100 633.1 ± 331.2*(9.2%) 88 545.8 ± 267.4* (8.9%)P2500 8046.8 ± 575.5* 7411.6 ± 508.0* 101 212.5 ± 743.9*(9.8%) 89 382.3 ± 617.3*(10.0%)

500 mL/min NP 0.0 ± 0.1 0.1 ± 0.1 103 055.9 ± 29.2 79 840.2 ± 42.4P500 3534.9 ± 24.4* 2985.5 ± 22.1* 108 594.9 ± 122.7*(5.4%) 84 909.6 ± 114.3* (6.3%)P1500 6149.7 ± 57.0* 5175.7 ± 46.6* 113 055.6 ± 113.8*(9.7%) 88 596.7 ± 101.4*(11.0%)P2500 5963.7 ± 27.9* 5018.1 ± 30.1* 110 858.2 ± 339.0*(7.6%) 86 581.9 ± 224.9* (8.4%)

750 mL/min NP 0.0 ± 0.1 0.1 ± 0.0 117 817.9 ± 19.2 80 645.0 ± 80.1P500 2670.6 ± 51.2* 1939.5 ± 34.2* 122 264.9 ± 220.9*(3.8%) 83 462.1 ± 170.9* (3.5%)P1500 5869.3 ± 60.7* 4272.0 ± 47.0* 126 991.2 ± 384.3*(7.8%) 86 944.9 ± 248.0* (7.8%)P2500 6105.4 ± 50.9* 4455.9 ± 41.5* 129 136.7 ± 255.2*(9.6%) 88 637.2 ± 175.5* (9.9%)

1000 mL/min NP 0.9 ± 0.2 0.8 ± 0.2 137 354.5 ± 47.0 80 878.5 ± 50.3P500 2092.7 ± 21.0* 1304.9 ± 13.4* 141 269.5 ± 212.9*(2.9%) 83 490.9 ± 130.2* (3.2%)P1500 6446.5 ± 49.9* 4078.9 ± 37.5* 149 211.1 ± 294.9*(8.6%) 88 933.1 ± 169.0*(10.0%)P2500 6734.0 ± 49.5* 4315.8 ± 29.8* 149 033.8 ± 403.7*(8.5%) 89 205.2 ± 202.2*(10.3%)

All percentages in parentheses are relative to nonpulsatile mode.* P < 0.05 vs. nonpulsatile mode.NP, nonpulsatile mode; P500, pulsatile speed differential value 500 rpm; P1500, pulsatile speed differential value 1500 rpm; P2500, pulsatile

speed differential value 2500 rpm

FIG. 3. Flow–pressure curves (postpump site) at different rotational speeds in pulsatile (P) and nonpulsatile (NP) modes.



value of 2500 rpm generated a mean prepump pres-sure of −206.5 mm Hg after clamping upstream of thepump head. Nonpulsatile flow generated a meanprepump pressure of −93.8 mm Hg.

Figure 5 shows the pressure waveforms after sepa-rately clamping both sides of the pump head at dif-ferent rotational speeds in pulsatile (P2500) andnonpulsatile mode. The momentary pressures werefar above mean positive postpump or negativeprepump pressures in pulsatile mode.

DISCUSSION

The Medos Deltastream DP3 rotary pump can becategorized as a diagonal or mixed-flow pump. Theblood path is thus at an angle between that of cen-trifugal pumps (90°) and axial flow pumps (180°). Interms of clinical applications, it is suitable for a widerange of patients and procedures due to its abilityto provide flow rates ranging from 0 to 8 L/min,pressure differences ranging from 0 to 600 mm Hg,and rotational speeds ranging from 500 to10 000 rpm, as well as the ability to function in pulsa-tile or nonpulsatile mode. Safety features of theDeltastream DP3 include zero-flow mode, integrated

bubble detector, and flow/pressure sensors. This new-generation diagonal pump combines a bearingsystem with magnetic coupling of forces and is suit-able for medium-term applications up to 7 days (8).In clinical practice, the Deltastream DP3 has beenused as a blood pump in venovenous or venoarterialECMO for heart and/or lung support in children andadults (9,10).

Our results showed that the Deltastream DP3 canprovide nearly physiological pulsatile flow in oursimulated pediatric MCS circuit and pure continuousnonpulsatile flow, as opposed to roller pumps, whichproduce a small amount of pulsatility at all times.When the pump was switched from nonpulsatile modeto pulsatile mode, rotational speed and flow rateincreased automatically. Using the pulsatile settingsdescribed above, the flow rates increased from261.4 mL/min to 511.8 mL/min (95.8% increase),and from 1019.6 mL/min to 1461 mL/min (43.3%increase).The reason may be that the rotational speedincreased from the initial setup during systole (70% ofthe cycle) and decreased during diastole (30% of thecycle).Therefore, when pump flow was switched fromnonpulsatile to pulsatile mode, a manual adjustmentof rotational speed was needed to maintain the same

FIG. 4. Flow-pressure curves (prepump site) at different rotational speeds in pulsatile (P) and nonpulsatile (NP) modes.

S. WANG ET AL.70


flow rate as in nonpulsatile mode. In the samemanner, when pulsatile mode was switched back tononpulsatile mode, the flow rates decreased. Thus,rotational speed must be increased to keep the flowrate identical with that observed in pulsatile mode.Forsimilar reasons, pulsatile flow using a speed differen-tial value of 2500 rpm produced more SHE than pul-satile flow using speed differential values of 500 rpmand 1500 rpm.After the rotational speed was reducedto the same value as in nonpulsatile mode, SHEdecreased by nearly one-half at 2500 rpm and aboutone-third at 1500 rpm, but did not change at a speeddifferential value of 500 rpm.THE also had 2.9–8.8%(P500), 7.8–9.7% (P1500), and 7.6–9.8% (P2500) oforiginal amplitude at the postpump site at the sametime.Therefore,the higher the speed differential valueis, the higher the SHE and THE. Therefore, higherspeed differentials in pulsatile mode generate morehemodynamic energy, while the rotational speed andflow rate automatically increase. Clinicians shouldmanually adjust the rotational speed to the same valueobserved in nonpulsatile mode in order to maintainidentical flow rate.

Our results confirmed that pulsatile flow generatedhigher pump head pressure when the outlet of the

pump head was clamped, especially at higher rota-tional speed. Most hollow-fiber membrane oxygen-ators are able to endure more than 600 mm Hg ofpositive pressure. There is thus less of a chance ofhollow fiber breakdown at rotational speeds less than5000 rpm when the downstream oxygenator is closed,although the instantaneous pressure was nearly400 mm Hg. However, the instantaneous negativepressure could reach −300 mm Hg. At this time, airbubbles may be separated from the bloodstream bythe negative pressure and then delivered into thepatient’s body in the closed MCS circuit (15). There-fore, clamping downstream of the pump head posesless potential risk than clamping upstream of thepump head. In order to prevent hemolysis and pumpcavitation, clamping downstream of the pump headand stopping the pump is suggested if the blood flowmust be interrupted for a long period of time athigher rotational speed.

A better optimization algorithm is needed to main-tain identical flow rates during conversion betweenpulsatile and nonpulsatile mode. In addition, devel-oping a synchronous pulsatile perfusion mode wouldfurther optimize pump flow to provide potentialbenefit for patients’ recovery (16,17).

FIG. 5. Pressure waveforms at prepump site (inlet of pump head clamped) and postpump site (outlet of pump head clamped) in pulsatilemode at 2500 rpm rotational speed differential (P2500) and nonpulsatile (NP) mode.



LIMITATION

Our findings are limited by the fact that this is an invitro study and thus may not completely representthe behavior of this pump in a clinical setting. Inaddition, our claim that no backflow was observed inpulsatile mode is limited by the absence of arterialcompliance in our pseudopatient.

CONCLUSIONS

The novel Medos Deltastream DP3 diagonal pumpis able to supply the required flow rate for a pediatricpatient, generate adequate quality of pulsatility, andprovide surplus hemodynamic energy output as evi-dence of pulsatility in a simulated pediatric mechani-cal circulatory support system. However, furtherimprovements in the performance of this rotarypump through the use of optimization-based algo-rithms are needed to achieve the best quality ofpulsatility.

Acknowledgments: Special thanks go to Dr.Jürgen O. Böhm and Andreas Spilker from MedosMedizintechnik AG, Stolberg, Germany, and IvoSimundic and Georg Matheis from Novalung GmbH,Heilbronn, Germany, for lending the DP3 pumpconsole and sending all disposables for this study.

Conflict of Interest: None.

REFERENCES

1. Ündar A. Facts and myths surrounding pediatric mechanicalcardiovascular circulatory support research: a personal per-spective. Artif Organs 2012;36:467–9.

2. Clark JB, Pauliks LB, Myers JL, Ündar A. Mechanical circu-latory support for end-stage heart failure in repaired and

palliated congenital heart disease [Invited Review]. CurrCardiol Rev 2011;7:102–9.

3. Pauliks LB, Ündar A. Heart transplantation for congenitalheart disease [Invited Review]. World J Pediatr Congenit HeartSurg 2011;2:603–8.

4. Pauliks LB, Ündar A. New devices for pediatric mechanicalcirculatory support [Invited Review]. Curr Opin Cardiol2008;23:91–6.

5. Undar A, Wang S. Current devices for pediatric extracorporeallife support and mechanical circulatory support systems in theUnited States. Biomed Mater Eng 2013;23:57–62.

6. Nezic D, Knezevic A, Cirkovic M, Milojevic P, Jakovljevic M,Jagodic S. Postcardiotomy assisted circulation with rollerpump—early and late results. Acta Chir Iugosl 2003;50:69–71.

7. Ündar A. Myths and truths of pulsatile and nonpulsatileperfusion during acute and chronic cardiac support. ArtifOrgans 2004;28:439–43.

8. MEDOS Medizintechnik AG. User Guide—MedosDeltastream III Blood Pump System. Stolberg, Germany:MEDOS Medizintechnik AG, 2011.

9. Schmid C, Philipp A, Hilker M, et al. Venovenousextracorporeal membrane oxygenation for acute lung failurein adults. J Heart Lung Transplant 2012;31:9–15.

10. Tiedge S, Optenhöfel J. First uses of a new diagonal pump inextracorporeal support systems for children and infants[German]. Kardiotechnik 2011;20:72–6.

11. Wang S, Kunselman AR, Ündar A. Novel pulsatile diagonalpump for pediatric extracorporeal life support system. ArtifOrgans 2013;37:37–47.

12. Krawiec C, Wang S, Kunselman AR, Ündar A. Impact of pul-satile flow on hemodynamic energy in a Medos DeltastreamDP3 pediatric extracorporeal life support system. Artif Organs2013. doi: 10.1111/aor.12117.

13. Shepard RB, Simpson DC, Sharp JF. Energy equivalent pres-sure. Arch Surg 1966;93:730–40.

14. Fitzmaurice GM, Laird NM, Ware JH. Applied LongitudinalAnalysis. Hoboken, NJ: John Wiley & Sons, 2004.

15. Simons AP, Ganushchak YM, Teerenstra S, Bergmans DC,Maessen JG, Weerwind PW. Hypovolemia in extracorporeallife support can lead to arterial gaseous microemboli. ArtifOrgans 2013;37:276–82.

16. Reesink KD, Sauren LD, Dekker AL, et al. Synchronouslycounterpulsating extracorporeal life support enhances myo-cardial working conditions regardless of systemic perfusionpressure. Eur J Cardiothorac Surg 2005;28:790–6.

17. Gazzoli F, Viganò M, Pagani F, et al. Initial results of clinicaltrial with a new left ventricular assist device (LVAD) providingsynchronous pulsatile flow. Int J Artif Organs 2009;32:344–53.

S. WANG ET AL.72


in vitro performance analysis of a novel pulsatile diagonal pump in a simulated pediatric mechanical...

Documents