Embed Size (px)
Citation preview
●
Rmpttmt1ptsss©
I
Rlnoaiarpdflubmpt
F
M
S
s2a
4
Effect of Differing Blood Lines on Delivered Blood FlowDuring Hemodialysis
Junaid Ahmed, MD, Anatole Besarab, MD, Tom Lubkowski, MS, and Stanley Frinak, MSEE
Background: The influence of differing blood tubing sets on delivered dialyzer blood flow (QB) was studied usingeadySet (RS; Medisystem Inc, Seattle, WA) and CombiSet (CS; Fresenius, Walnut Creek, CA). Methods: In vitro, weeasured QB and pre-pump arterial pressures (PA; in millimeters of mercury) during 4 hours at a prescribed blood
ump flow (QP) of 450 mL/min, using a glycerin-water mixture for RS and CS blood tubing sets (n � 10 each). Wehen performed 3 clinical studies to determine the difference between QP and QB as a function of PA for both tubingypes and any effects of tubing type on delivered Kt/V in patients. Results: In vitro, 4-hour time-averaged QB was 451L/min with RS versus 422 mL/min with CS. QB was approximately 40 mL/min greater using RS than CS blood
ubing sets at QP of approximately 445 mL/min with [QP � QB] increasing linearly with PA for CS, but not RS (study). At a mean QP of approximately 408 mL/min, CS blood tubing significantly increased Kt/V by approximately 0.1 (90atients; study 2). In study 3, [QP � QB] increased with treatment time with CS (8 to 29 mL/min), but not RS (from �4o 6 mL/min). For patients achieving equal QP of 400 mL/min or greater for all treatments (n � 11), paired analysishowed greater values for urea reduction ratio (�3.5%; P � 0.039) and Kt/V (�0.14; P � 0.04) with RS than CS tubingets. Conclusion: Differences in [QP � QB] occur both in vitro and in vivo, are much smaller with RS than CS tubingets, and result in a greater delivered Kt/V of up to 0.14 for RS. Am J Kidney Dis 44:498-508.2004 by the National Kidney Foundation, Inc.
NDEX WORDS: Dialysis; blood flow; tubing; kinetic; clearance.
dtomeDdcddb
o�(ewir6ftacmmiis
EMOVAL OF a solute like urea is quanti-tated as clearance that, in turn, varies non-
inearly with blood pump flow (QP).1 Urea ki-etic modeling often is based on clearancesbtained from in vitro manufacturer’s data. Themount of dialysis delivered, expressed as Kt/V,s an important determinant of patient outcomes2
nd is a quality indicator tracked by end-stageenal disease networks and payers.3 The dialysisrescription includes QP because of the strongependency of urea solute clearance on bloodow.4,5 Delivery of adequate dialysis in terms ofrea kinetics thus is dependent on 3 variables:lood flow delivered to the dialyzer (QB), treat-ent duration, and amount of solute initially
resent in the dialyzed individual, a function ofotal urea volume of distribution.
From the Division of Nephrology and Hypertension, Henryord Hospital, Detroit, MI.Received December 17, 2003; accepted in revised formay 5, 2004.Supported in part by a grant from Medisystem Inc,
eattle, WA.Address reprint requests to Anatole Besarab, MD, Divi-
ion of Nephrology and Hypertension, Henry Ford Hospital,799 West Grand Blvd, Detroit, MI 48202.E-mail:[email protected]© 2004 by the National Kidney Foundation, Inc.0272-6386/04/4403-0013$30.00/0
odoi:10.1053/j.ajkd.2004.05.028
American Journal of Kidn98
In the clinical setting, the blood flow rateisplay on dialysis machines is used to “dial in”he patient’s prescribed QP. This display is basedn revolutions per minute and is assumed byost practitioners to be accurate and uninflu-
nced by the blood tubing sets used for dialysis.ifferences between QP and true QB have beenescribed, particularly with the use of dialysisatheters.6 Accurate measurement of blood flowuring hemodialysis is essential to avoid under-ialysis, particularly at the high flows demandedy high-performance dialyzers.7
Previous studies have shown that over a rangef pre-pump arterial pressures (PAs) from �50 to400 mm Hg and venous outflow pressures
PVs) of �50 to �300 mm Hg, only PA influ-nced flow when measured volumetrically orith an ultrasonic flow probe, and this effect was
ndependent of hematocrit over the clinicallyelevant range.7,8 During routine hemodialysis in4 patients, QB was 9% less when PA variedrom �180 to �220 mm Hg.7 Teruel et al8 foundhat the difference between nominal flow andctual (ultrasonically determined) blood flow in-reased with increasing flow: 11.6% at 300 mL/in, 13.1% at 350 mL/min, and 14.8% at 400L/min. Schmidt et al9 added complexity to the
ssue by showing that the difference [QP � QB]ncreased during 5 hours of pump operation,uggesting that blood tubing fatigue also devel-
10
ps. Subsequently, Mehta et al showed that useey Diseases, Vol 44, No 3 (September), 2004: pp 498-508
osip
eecbesbHIaattdQad
pDsctptsvtttt(
a(CLsIwD
omssaoD
TtTitDb1tpbamtp
dTSnehmpt41
dt(gpaTgc1dagbBwoabmmatmtma
(a(a
BLOOD TUBING EFFECTS ON DIALYZER BLOOD FLOW 499
f a larger bore needle (14 G) could help manyubjects achieve a QP of 500 mL/min, therebymproving the percentage of urea reduction by 5oints.Previous studies focused primarily on the influ-
nce of the needle,7,8,10-12 type of access (cath-ter versus arteriovenous [AV] shunt), or ma-hine delivery system.6 Only 1 study examinedlood tubing sets made by different manufactur-rs, but provided no details.9 The studies de-cribed in this report arose out of an observationy 1 of the authors (A.B.) that the Transonicemodialysis Flow Monitor (Transonics System
nc, Ithaca, NY) was reading significant discrep-ncies in [QP � QB] at 1 locale, West Virginia,nd not another, Michigan. Both needles andubing sets being used at the time differed be-ween the 2 sites. Studies were designed toetermine whether the noted difference in [QP �
B] resulted from these blood tubing set factorsnd determine the clinical significance of such aifference.
METHODS
We postulated that differences in elasticity of the bloodump tubing segment exist among manufactured brand sets.ifferences in elastic recoil would, in turn, influence the cross-
ectional area in the roller pump segment (the degree of tubingollapse) and, therefore, the stroke volume pumped per revolu-ion. Lower stroke volumes would result in less QB. We alsoroposed that the elasticity of the tubing segment changes withhe duration of use (“fatigue”) and can lead to reduced cross-ectional area of roller pump segment, producing a lower strokeolume resulting in an even lower time-averaged QB. We testedhe hypothesis that differences in elasticity among manufac-ured tubing sets that exist affect actual QB and thus influencehe delivery of dialysis to patients.All studies were approved byhe Institutional Review Board of West Virginia UniversityMorgantown, WV) and Henry Ford Hospital (Detroit, MI).
Two blood tubing sets and 2 fistula needle sets were in uset the time of the initial observation. Fresenius CombiSetCS) tubing set (code: 2HR218; Fresenius, Walnut Creek,A) combined with 15-G JMS AV fistula (AVF; JMS C0td, Hiroshima, Japan) needles (code: 820-2519) and Medi-ystem ReadySet (RS) tubing set (code: 2319L2, 2507L2,Y27L2, 2318L5; Medisystem Inc, Seattle, WA) combinedith 15-G ReadySet Twin Pack AVF fistula needles (code:9-2005MG).For the in vitro studies, a Fresenius blood pump (as used
n 2008 D, E, and H machines) was calibrated to theanufacturer’s specifications and used for testing. The test
olution was a mixture of 36% glycerin and 64% water (toimulate the dynamic fluid properties of blood) maintainedt 37°C. Each blood tubing set was set up in the laboratoryn the blood pump to simulate the configuration of the 2008
, E, and H machines and was primed with the test solution. she arterial tubing was routed directly into the venousubing without a dialyzer in the path to avoid ultrafiltration.he matching fistula needles were attached to the arterial
nput and venous output tubings. Each blood tubing set wasested during a 4-hour period, with QP set at 450 mL/min.elivered QB was measured volumetrically over 30 secondsy using a 250-mL graduated cylinder and stopwatch in all0 experiments with each tubing set. Flow was measured inhe first 5 minutes and hourly for 4 hours. Although aressure/vacuum gauge was placed on the monitor lines ofoth the arterial and venous tubing in all experiments,rterial and venous pressure corresponding to flow measure-ents were made only in the last 5 experiments with each
ubing set. Pump revolutions per minute were checkederiodically using a tachometer.All clinical studies were conducted using Fresenius hemo-
ialysis delivery systems and F80 dialyzers (Fresenius). Theransonic Blood Flow monitor (HD01 or HD02; Transonicystems Inc) was used to measure QB.13 Under conditions ofo ultrafiltration, the 2 sensors were within 2 mL/min ofach other. Arterial and venous transducers on all Freseniusemodialysis delivery systems were calibrated to within �5m Hg at 200 mm Hg. In the first clinical study, blood
umps were volumetrically calibrated within 20 mL/min ofhe nominally display flow rate of 500 mL/min (accuracy of%). At a blood flow of 300 mL/min, accuracy was within �0 mL/min.All subjects gave informed consent and were receiving
ialysis 3 times/wk. Hemoglobin levels of study patients hado be within the Kidney Dialysis Outcomes Quality InitiativeK/DOQI)–recommended range of 11 to 12 g/dL (110 to 120/L) to be eligible for further screening. We then selectedatients with permanent fistulae (native AVFs or grafts) withvascular access blood flow, measured by means of the
ransonic blood flow monitor,14 5 minutes into dialysisreater than 500 mL/min; patients with catheters were ex-luded. For all study patients, access was cannulated using a5-G needle. Intra-access pressures (using AcessAlert; Me-isystems, Seattle, WA)15 at the time of fistula cannulationlso had to be within normal limits for native AVFs andrafts. Arterial and venous ultrasound sensors were cali-rated by the manufacturer before the studies. The Transoniclood Flow monitor provided a continuous display of flowithin both the arterial and venous tubing. Under conditionsf no ultrafiltration, the 2 sensors agree within 1 to 2 mL/mint QB of 300 to 500 mL/min. QP was maximized as limitedy arterial and venous pressure limits of �280 and �300m Hg, respectively. QB was measured within the first 10inutes of initiating hemodialysis and as soon as QP was
chieved by the dialysis nurse, at the beginning of each hourhereafter, and within the last 15 minutes of the dialysis treat-ent. To minimize tubing fatigue before starting the patient on
he dialysis treatment, recirculation of the prime was not startedore than 5 minutes before needle insertion into the vascular
ccess.The first clinical study was conducted at 2 sites: 21 patients
13 men, 8 women) were studied using CS blood tubing setsnd 15-G JMS AVF needles (West Virginia), and 34 patients21 men, 13 women) were studied using RS blood tubing setsnd 15-G Medisystem AVF needles (Michigan). QB was mea-
ured by means of a Transonic HD01 system.
ownwAd
win
(stasiAbbdMbowD(
ttmolwp
mwt
S
fwctaataAhl
filadecy(sCiS
I
(
AHMED ET AL500
The second clinical study was conducted at a satellite unitf Greenfield Health Systems (Michigan). Ninety patientsere dialyzed using CS blood tubing sets and JMS AVeedles for 1 month. The following month, the same patientsere dialyzed using RS blood tubing sets and RS Twin PackVF needles. Urea reduction ratio (URR) and Kt/V wereetermined. Kt/V was calculated from the equation16:
�ln�R � 0.008t� � �4 � 3.5R� * UF ⁄ W,
here R is the post/preurea ratio, UF is ultrafiltration, and Ws the patient weight in kg. In this study, the gauge of theeedles was not controlled.The final clinical study was performed at the same site
Michigan). We selected 23 patients, as in the first clinicaltudy. Inclusion criteria were stable adult hemodialysis pa-ients, no history of hypotensive episodes during dialysis,nd no history of clotting dialyzers during dialysis. Exclu-ion criteria were age younger than 18 years and mentalmpairment. All fistula needles were 15-G and RS Twin PackVF. During the first week, patients were dialyzed with CSlood tubing sets. The dialysis machines then were cali-rated for the change in blood tubing sets. Patients then wereialyzed the following week using RS blood tubing sets andedisystem AVF needles. During the midweek treatment of
oth weeks, a new dialyzer was used to minimize any effectsf dialyzer reuse on measured clearances, and blood samplesere collected to quantify the midweek URR and Kt/V.uring the midweek treatment, blood water urea clearance
CUrea) was calculated.17
During all sessions, PA, PV, and QB were measured withinhe first 10 minutes, hourly, and within the last 15 minutes ofreatment. Blood flow displayed by the machine, QP, was
aintained constant to the greatest extent possible through-ut the dialysis treatment, with the exception of the first andast 5 minutes of treatment. After completing measurementsith the second set of blood tubing, static intra-access
ressure and access flow were measured to ensure that no aajor changes had occurred in vascular access functionithin the 2-week period. Hemoglobin levels were main-
ained within the K/DOQI recommended range.
tatistical MethodsDescriptive statistics (mean, SD, or SEM) were calculated
or each measured parameter. Time-averaged treatment flowsere determined by integrating the area under the flow time
urves of each patient for both QP and QB. Sample size forhe last clinical study was calculated as follows: with annticipated difference in mean for the quantity [QP � QB] oft least 25 to 30 mL/min (clinically important) at QP greaterhan 400 mL/min and for values for � of 0.05, � of 0.9, andn anticipated SD of 40 mL/min, 18 subjects were needed.nticipating at least a 25% dropout rate from unanticipatedospitalizations or missed treatments, 23 patients were se-ected for the study.
For in vitro studies, unpaired t-tests were used. For therst 2 clinical studies, an unpaired t-test also was used. In the
ast study, a paired t-test was used, with each patient actings his or her own control. The CS blood tubing set wasesignated as control, and the RS blood tubing set, asxperimental. For all comparisons, P less than 0.05 isonsidered statistically significant. Repeated-measures anal-sis was used to discern effects of time on QB. StatViewSAS Institute, Cary, NC) was used for all statistical analy-es, and MathCad (Mathsoft Engineering & Education Inc,ambridge, MA) was used for nonlinear curve fitting and
ntegration procedures. Variation is presented as either SD orEM.
RESULTS
n Vitro Comparison of Blood Tubing Sets
Data from in vitro studies are shown in Fig 1n � 10 for each blood tubing set). QP was set
Fig. 1. Results of the invitro study showing reducedflow with time for the CSblood tubing set with theblood pump setting fixed at450 mL/min. Error bars rep-resent SD.
t 450 mL/min for the studies. Initial QB for
RQtd2tsiCpmstia(
tFe1dbt
C
vVbdtl
T
idot
BLOOD TUBING EFFECTS ON DIALYZER BLOOD FLOW 501
S was 457 � 5 (SD) mL/min and exceeded
P equal to 450 mL/min (P � 0.01), whereashe value for CS, 448 � 9 mL/min, did notiffer from QP. During 4 hours, QB averaged9 mL/min greater with RS than CS bloodubing sets. PAs and venous pressures mea-ured during the 5 experiments with each tub-ng are listed in Table 1; decreasing flows withS blood tubing sets were accompanied byrogressively less negative PA (�283 to �249m Hg). Conversely, with RS blood tubing
ets, PA remained within a tight range of �266o �260 mm Hg during the 4 hours. Decreas-ng flows with CS blood tubing sets also wereccompanied by a 23–mm Hg reduction in PV
198 to 175 mm Hg), whereas with RS blood
Table 1. Temporal Changes in
CS
ime (h)
ArterialPressure(mm Hg)
VenousPressure(mm Hg)
MeasurRate (m
0 �283 � 1 198 � 1 4481 �270 � 3 186 � 1 4232 �259 � 1 181 � 1 4163 �255 � 2 178 � 1 4114 �249 � 1 175 � 2 408
NOTE: N � 5 for each tubing type. Results are expresse
Fig. 2. Photograph show-ng the change in diameteruring continuous operationf the CS and RS blood pump
ubing segments.ubing sets, PV decreased by only 5 mm Hg.igure 2 presents a visual display of the differ-nce in blood pump tubing segment diameters0 minutes after the start of the study. Theegree of recoil of the pump tubing segmentack to its normal diameter is clearly less withhe CS than RS pump tubing segment.
linical Study 1
Table 2 lists findings in 21 patients using CSersus 34 patients using RS blood tubing sets.ascular access flow averaged more than 1 L inoth groups, and even the minimal access flowelivered a QP of 500 mL/min without exceedinghe pressure limits. PA was independent of vascu-ar access flow: CS blood tubing sets, PA �
res and Delivered Flow in Vitro
RS
ArterialPressure(mm Hg)
VenousPressure(mm Hg)
Measured FlowRate (mL/min)
�266 � 7 188 � 3 456 � 2�266 � 4 190 � 4 456 � 2�264 � 2 189 � 4 452 � 2�263 � 4 188 � 3 451 � 2�260 � 4 183 � 2 448 � 2
an � SEM.
Pressu
ed FlowL/min)
� 2� 2� 2� 2� 2
d as me
�t0mw4as
scp[m
tvw
C
aottfdse
APMM(
AHMED ET AL502
220 to 0.002 QA (R2 � 0.001), and RS bloodubing sets, PA � �217 to 0.008 QA (R2 �.017). A negative pressure greater than �270m Hg was seen only once with CS and twiceith RS blood tubing sets. Mean QP (471 and74 mL/min) and pre-pump PA pressures (�222nd �226 mm Hg) did not differ, although thetudies were performed at 2 different sites.
However, QBs measured directly by the ultra-ound Transonic flow probe differed signifi-antly: the flow delivered did not differ from thatrescribed using RS blood tubing sets (8 � 5SEM] mL/min; P � 0.4), whereas it was 48 � 5L/min lower (P � 0.0001) using CS blood
Table 2. Comparison of Prescribed an
CS (n � 21
ccess flow (mL/min) 1,095 � 108 (55re-pump pressure (mm Hg) �222 � 8 (160–ean QP (mL/min) 471 � 9 (400–ean QB (mL/min) 424 � 7 (239–
QP � QB) (mL/min) 48 � 5 (48�2
NOTE. Values are expressed as mean � SEM (range).Abbreviation: NS, not significant.
ubing sets. Furthermore, [QP� QB] varied in-ersely with PA (Fig 3) for CS tubing, whereas itas independent of PA with RS tubing.
linical Study 2
In this portion of the study, we attempted tossess the clinical impact of any changes in QB
n indicators of dialysis adequacy. All patients athe Greenfield Health Systems center were ini-ially dialyzed using CS and JMS AVF needlesor 1 month, and monthly URR and Kt/V wereetermined. After all machines underwent conver-ion of their air detectors and pressure transduc-rs were calibrated, the same patients were dia-
ered Blood Flow by Blood Tubing Sets
RS (n � 34) P
0) 1,128 � 88 (526–2,600) NS�226 � 6 (150–300) NS
474 � 10 (400–500) NS466 � 8 (311–510) �0.05
8 � 5 (40�57) �0.05
Fig. 3. [QP � QB] versusPA for CS and RS blood tub-ing sets. Greater negative PAresults in a much greater in-crease in [QP � QB] for the
d Deliv
)
0–2,15290)500)527)75)
CS blood tubing set.
lAwtct1o9masuF
eu(bcscKbtK
igm
C
tpctuefwiUsu
1tcmtQm
PUPTPUK
BLOOD TUBING EFFECTS ON DIALYZER BLOOD FLOW 503
yzed using RS blood tubing sets and MedisystemVF needles the next month. URR and Kt/Vere determined. We did not preclude changes in
he dialysis prescription by the attending physi-ians. Patients using catheters (1/3 of subjects) orhose with hemoglobin levels outside the 11- to2-g/dL (110- to 120-g/L) range were included;nly transient patients were excluded. Thirty of0 patients had a prescription change in treat-ent time between the 2 months: 15 subjects had
n increase (mean, 15.8 � 1.5 minutes), and 15ubjects had a decrease (mean, 18.5 � 2.1 min-tes). Two patients had a dialyzer change from80 to F200A (Fresenius).Table 3 lists the results. There were no differ-
nces in postdialysis weight, ultrafiltration vol-mes removed, predialysis blood urea nitrogenBUN) level, treatment time, and prescribed QP
etween the 2 months. Because of the use ofatheters, mean QP was lower than in clinicaltudy 1. There were small, but significant, in-reases in URR from 66.9% to 69.4%, and int/V, from 1.31 to 1.40 after the switch to RSlood tubing sets. With the membranes used andhe treatment time, an increase of 0.09 units int/V required that the mean delivered clearance
Table 4. Comparison of Prescribed and Delivered BlQP Greater than 450 mL/min for
Time (min)
CS
QP (mL/min) QB (mL/min) QP (mL/min
10 483 � 7 477 � 7 485 � 660 480 � 4 463 � 4 471 � 5
120 478 � 4 457 � 4 472 � 4Last 15 min 474 � 5 447 � 5 466 � 5
Table 3. Comparison of Tubing Types on U
CS
ostdialysis weight (kg) 79.0 � 1.8ltrafiltration (L) 2.4 � 0.3redialysis BUN (mg/dL) 55 � 3reatment time (min) 197 � 3ump blood flow (mL/min) 402 � 9RR (%) 66.9 � 0.8t/V 1.31 � 0.0
NOTE. Results are expressed as mean � SEM. To convAbbreviation: NS, not significant.
NOTE. Number of observations � 27 for each tubing set. Resul
ncrease by approximately 21 mL/min, a valuereater than the insignificant difference of 12L/min noted in QP.
linical Study 3
Although delivered QB, PA, and in vitroests performed indicated that the blood tubingump segment is the main factor in the clini-ally observed Kt/V differences, other subtlereatment factor differences may have contrib-ted. In addition, we had not controlled for theffect of differences in AV graft needles. There-ore, we performed a careful prospective studyith direct measurement of tubing blood flow,
n vivo urea clearances, and measurements ofRR and Kt/V during the same treatment
ession in which the same AVF needles weresed and only blood tubing was varied.Of 23 patients originally enrolled in the study,
9 patients eventually completed the study, al-hough not all aspects per protocol. Variation inannulation or needle position precluded attain-ent of the prescribed QP during many of the
reatments during the 2 consecutive weeks. EqualP greater than 400 mL/min during the criticalidweek sessions was not achieved in 8 of 19
ows by Blood Tubing Sets in 9 Subjects Achieving aessions With Each Tubing Set
CS RS
B (mL/min) [QB � QP] (mL/min) [QB � QP] (mL/min)
489 � 6 �3.5 � 2.2 �4.3 � 1.4473 � 5 �16.6 � 1.7 �2.0 � 2.4467 � 5 �21.5 � 2.2 �4.9 � 1.8459 � 5 �26.0 � 2.5 �7.0 � 1.9
Delivered Kt/V in 90 Hemodialysis Patients
RS P
78.7 � 2.5 NS2.5 � 0.2 NS57 � 3 NS
194 � 3 NS414 � 8 NS
69.4 � 0.7 �0.051.40 � 0.03 �0.05
in mg/dL to �mol/L, multiply by 0.357.
ood FlAll 3 S
RS
) Q
RR and
3
ert BUN
ts are expressed as mean � SEM.
sgstv1g4Rm3sstudttho
tslPtoa
amuascKweQmsstflTCg
slpat
so
BBBPMMMMMMM
m
AHMED ET AL504
ubjects; only 9 patients were able to achievereater than 450 mL/min during all 6 treatmentessions. Table 4 lists the results. During allreatments, mean initial QP did not differ (483ersus 485 mL/min). Average treatment time was89 � 8 minutes with both tubing sets. Inte-rated during the entire treatment, QP was 479 �mL/min with CS and 477 � 4 mL/min withS. QB within the first 10 minutes of the treat-ent was 4 mL/min greater than QP with RS andmL/min lower than QP with CS blood tubing
ets. During the first 2 hours of dialysis, a progres-ively increasing difference was observed be-ween the blood pump setting and QB measuredltrasonically for CS blood tubing sets. Thisifference reached 26 mL/min at the end of thereatment session. Conversely, with RS bloodubing sets, there was no change during the first 2ours and only a 7-mL/min difference at the endf dialysis (Table 4). All these differences be-
Table 5. Comparison of Negative PA by BloodTubing Sets in 9 Subjects Achieving a QP Greater
Than 450 mL/min for All 3 Sessions WithEach Tubing Type
Time (min) CS RS P
�10 �248 � 2 �250 � 3 Not significant60 �248 � 2 �257 � 2 �0.05
120 �249 � 2 �260 � 2 �0.05Last 15 min �252 � 3 �260 � 2 �0.05
NOTE. Number of observations � 27 for each tubinget. Results are expressed as mean � SEM in millimetersf mercury.
Table 6. Comparison of Clinical Dialysis Paramete
CS
UN (predialysis) (mg/dL) 58.7 � 4UN (arterial) (15 min) 41.9 � 3UN (venous) (15 min) 11.8 � 1ostdialysis BUN 16.6 � 1ean QB (mL/min) 470 � 6ean QD (mL/min) 795 � 2ean CUrea (mL/min) 301 � 4ean URR (%) 72 � 1ean Kt/V 1.50 � 0ean ultrafiltration (kg) 3 � 0ean weight (kg) 84.2 � 6
NOTE: Critical t value is greater than 2.0. Differencesean � SEM. To convert BUN in mg/dL to �mol/L, multiply
Abbreviation: QD, dialysate flow rate.ween the 2 blood tubing sets were statisticallyignificant (P � 0.05). Consistent with our ear-ier results in clinical study 1 and in vitro study,As recorded hourly were more negative with RS
han CS blood tubing sets (Table 5). The limitsn the machine were set to �280 mm Hg beforelarms were activated.
Effects on dialysis delivery for all 19 subjectsre listed in Table 6. There were no differences inean QB, dialysate flow, postdialysis weight,
ltrafiltration volume, predialysis BUN level,rterial and venous BUN during clearance mea-urements, and postdialysis BUN level (unpairedomparisons). Differences in URR, CUrea, andt/V, although tending toward greater valuesith RS, did not achieve significance. However,
xamination of data showed unequal prescribedP values during the critical midweek treat-ents. We therefore examined results only in
ubjects who either consistently maintained atable flow rate of 500 mL/min during all dialysisreatments (n � 9) or those who achieved equalows at both midweek treatments. As listed inable 7, treatment time and QP did not differ;Urea, URR, and Kt/V were all significantlyreater with RS than CS tubing.
DISCUSSION
Accurate knowledge of QB during hemodialy-is is essential to avoid underdialysis, particu-arly at the high flows demanded by high-erformance dialyzers.7,18 Practitioners usuallyre unaware that the nominal value displayed byhe dialysis delivery system frequently is inaccu-
9 Subjects Dialyzed With the 2 Blood Tubing Sets
RS t
58.2 � 4 0.141.7 � 3.5 0.0511.8 � 1.1 015.2 � 1.4 0.7456 � 9 1.3789 � 4 1.58308 � 4.4 1.1174 � 1 0.9
1.56 � 0.06 0.7472.5 � 0.3 1.19
84.4 � 6.6 0.008
en groups are not significant. Results are expressed as57; (CUrea) in mL/min to mL/s, multiply by 0.01667.
rs in 1
.6
.5
.7
.6
.05
.8
.5
betweby 0.3
rgastwmltit
faflcamKrdImciomgbbfpli
mttWwpd
roatssvta1cedpdtStcagiqiteccdnaslsotitrn
o
N
TQCUK
0
BLOOD TUBING EFFECTS ON DIALYZER BLOOD FLOW 505
ate in terms of actual QB, particularly whenreater blood flows are used with high-efficiencynd high-flux membranes. The blood flow pre-cribed in a hemodialysis session corresponds tohat indicated by the blood roller pump display,hether digital or analog. In the clinical setting,ost calibration is based on the number of revo-
utions per minute of the blood roller pump byachometry, and not by volumetric calibration. Its presumed that the manufacturers took care ofhe details related to the blood tubing.
The delivery of adequate blood flow is vitalor the optimal performance of these dialyzersnd optimizing the clearances attained. Bloodow, dialysate flow, and surface area have signifi-ant and independent effects on urea, creatinine,nd phosphorus clearance19 and potassium re-oval.20 The importance of delivering adequatet/V cannot be overstated. Since the landmark
eport by Gotsch and Sargent,21 adequacy ofialysis has been measured in terms of Kt/V.deally, the delivered dose of dialysis should beeasured on-line.22 However, in most centers, it
urrently is measured indirectly though URR orndirect formulas for Kt/V. Each increase in Kt/Vf 0.1 unit is associated with a decrease inortality of 7%.4 At single-pool Kt/V values
reater than 1.4 (double-pool, 1.223), additionalenefit does not accrue.24 When Kt/V is noted toe repeatedly decreased, a search for a variety ofactors usually is initiated. These include noncom-liance, vascular access recirculation, and dia-yzer dysfunction.25 Reduced blood processing
Table 7. Comparison of the Difference in CUrea andDialysis Adequacy Parameters Between the 2 TubingSets in Patients Who Maintained a Stable Flow Rate
Greater Than 450 mL/min DuringAll Dialysis Treatments
Difference(RS � CS) P
o. of pairedobservations
9
reatment time (min) 1.4 0.7
P (mL/min) 2.8 0.6
Urea �6 0.02RR �0.033 0.033t/V �0.14 0.024
NOTE. To convert CUrea in mL/min to mL/s, multiply by.01667.
s found because of a low QP or reduced treat- t
ent time in less than half the cases. Recircula-ion greater than 12% is found in another 25%. Inhe remaining third, no apparent reason is found.
e believe that in many of these residual casesith persistent abnormality, collapse of the bloodump segment may account for the inadequateelivery of dialysis.Type of vascular access influences QB. The
elationship between blood flow and inflow andutflow pressures has been examined previouslynd is known to be nonlinear and dependent onhe type of vascular access used for hemodialy-is. When QB is measured accurately by ultra-ound, both pre-pump pressure and postdialyzerenous pressures are complex exponential func-ions of QB and hematocrit when fistula needlesre in the circuit.26,27 For needle cannulas of 15,6, or 17 G, the pressure decrease across themannot be described by the Hagen-Poiseuille law,ven when corrected for inlet or outlet effects,escribed by Bernoulli’s law.28 The connectingiece between cannula and tubing, as well as theirection of flow through the needle, influenceshe pressure decreases through such needles.imilar second-order mathematical relations be-
ween pressure and flow also are found withatheters.29 As catheters capable of higher flowsre developed, Twardowski and Haynie29 sug-ested that pressure flow relationships also bencorporated into final design. Catheter flow fre-uently is limited by the high resistance inherentn the long catheter length (�30 cm) relative tohat of dialysis needles (2.54 cm), promotingxcessive pre-pump negative pressures and recir-ulation. Flow-dependent recirculation throughircuits other than the peripheral AV access re-uces dialysis efficiency. With catheters, the in-er diameter of the catheter has major effects onccess flow; slight changes in diameter at theame degree of negative PA are associated witharge increases in flow.30 However, even withmall decreases in orifice area at the inflow, asccurs with fibrin sheaths or microthrombi, nega-ive pressure at the inflow generated by the pumps magnified as partial collapse of the pumpubing segment leads to erroneous blood floweadings by the pump flow indicator (the nomi-al value displayed).The magnitude of this error was clearly pointed
ut by Sands et al,6 who noted that catheters gave
he largest error (an error of 86 mL/min at a QP of
4ptsrtgspnniee
bpcmQfl1rtgfedfopPsd
visndtnbObtmius
mtfl
cdshntsraticapiobocu
telmtppQtrr
iptmbrclCasg
AHMED ET AL506
00 mL/min). However, even in patients withermanent accesses and fistula needle use, pa-ients receiving therapy on Fresenius deliveryystems (2000H) had greater QBs than thoseeceiving therapy on Baxter 550 machines (Bax-er Healthcare, Deerfield, IL); 29 and 38 mLreater at 300- and 400-mL/min blood pumpettings. In these studies, degree of error wasroportional to the QP setting and could beeutralized partially by changing to a larger boreeedle, QB increasing by 35 mL/min when chang-ng from a 15-G to a 14-G fistula needle.6 Similarffects of needle gauge were found by Hasbargant al.11
Collapse of the blood pump segment of thelood tubing sets with high negative pre-pumpressures has been recognized since 1980.31 In aarefully conducted study of 9 consecutive treat-ents in 20 patients, Teruel et al8 measured trueB by using ultrasonic transit time. Ultrasonicow was always lower than QP; 11%, 13%, and5% lower at QP of 300, 350, and 400 mL/min,espectively. Although most investigators foundhe deviation of QB from QP to begin at flowsreater than 200 mL/min, some investigatorsound that QB at a QP setting of 300 mL/min wasqual to or sometimes greater than the nominalisplay.6 This raises the issue of whether manu-acturers of the roller pumps, aware of this effectf negative inflow pressure on QB in occlusiveumps, actually calibrate their pumps so that atA pressures typically attained in the clinicaletting (�150 to �250 mm Hg), QP is accuratelyelivered.In our own studies, we were surprised, both in
itro and during the clinical studies, that thenitial QB (measured within 10 minutes of thetart of dialysis) was within 5 mL/min of theominal values between 450 and 500 mL/min, aeviation of only 1% for CS tubing. CS tubing ishe tubing manufactured by the makers of Frese-ius pumps and suggests that calibration now iseing performed at the high end of blood flows.f course, with the use of the more elastic RSlood tubing sets, initial blood flows were greaterhan those displayed. However, QB deteriorateduch more rapidly with CS than RS blood tub-
ng sets, even as the blood pump display contin-ed to show the same value. Results in this study
how for the first time that blood tubing sets tade by different manufacturers are not equal inheir ability to deliver sustained prescribed bloodows at identical blood pump settings.The marked deviation of QB from QP at in-
reasingly more negative PAs has been wellescribed.7-9 Depner et al7 found that flow mea-ured with an ultrasonic probe during routineemodialysis averaged 9% less than that nomi-ally displayed. One could theoretically continueo increase the QP by redesigning the deliveryystems to accept greater pre-pump and venouseturn pressures because these pressures no longerre germane to ultrafiltration control. Some inves-igators believe that high venous or arterial tub-ng pressures during hemodialysis are of nolinical concern.32 However, we are concernedbout hemolysis at extreme negative pre-pumpressure, particularly when no pre-pump displays used, rather a “pillow.” Hemolysis does notccur because of the negative pressure per se,33
ut from the shearing of cells through the narrowrifice.34 Deaths have occurred from hemolysisaused by pump malocclusion or inappropriatese of pillows.34
Although the pump segment of the bloodubing is specifically made from very resilient/lastic materials, restoration to its full diameter isess complete as pre-pump pressures becomeore negative. A smaller diameter cross-sec-
ional caused by incomplete recoil will result in 2henomena, both observed in this study. The firsthenomenon is a lower flow than expected forB because stroke volume is decreased despite
he same pump revolutions per minute. As aesult, the pump display, QP, no longer accu-ately records blood flow.
The second phenomenon is a less negativenlet pressure. The pump generates a negativeressure, but it is the elastic recoil of the pumpubing segment that has to maintain it and is aajor part of the driving force for inflow of
lood into the tubing (the remainder is the arte-ial pressure at the needle). Lower elastic recoilorresponds to less negative pressure, leading toower inflow. Thus, it is not surprising that withS blood tubing sets, the difference between QP
nd QB correlated inversely with pre-pump pres-ure (Fig 3), becoming 0 only as PA increased toreater than �180 mm Hg. Conversely, with RS
ubing, there was no effect of PA on [QP � QB],
eFwshe[w
meslooaatisI6
ilmpaslbpiptq
oatAdatecKt$o
d
ms1
h
AmHM
Stw1
o6
bd
rM
brs
IK
Mau2
oi
a
BLOOD TUBING EFFECTS ON DIALYZER BLOOD FLOW 507
ven to PA values of �290 mm Hg. As shown inig 2, this difference in elastic recoil can be seenith the naked eye. Both blood tubing sets also
howed some fatigue during the course of 4ours in vitro and 2.5 to 5 hours in vivo. How-ver, the degree of fatigue, represented by theQP � QB] versus time curves, was much smallerith RS than CS blood tubing sets.Ultrasonic transit time currently is the bestethod for measuring blood flow rate in the
xtracorporeal hemodialysis circuit, and its usehould be expanded as a calibration tool, particu-arly as tubing is acquired from different vendorsr lot numbers change. In this study, various lotsf both blood tubing sets were used. This mayccount for the variation in degree of time-veraged QB difference between the 2 bloodubing sets among the studies; 29 mL/min in then vitro study, 40 mL/min in the first clinicaltudy, and 17 mL/min in the third clinical study.n the latter study, this increase in QB produced a-mL/min increase in urea clearance.Clinicians often are challenged with increas-
ng dialysis adequacy in selected patients, particu-arly those of larger size, to maintain the mini-um Kt/V suggested by the K/DOQI clinical
ractice guidelines. In many cases, this requiresn increase in treatment time and/or, if pre-cribed flow is already maximized, use of aarger dialyzer. Our study indicates that use oflood tubing sets that maintain their viscoelasticroperties and deliver greater blood flows canncrease URR by 4.8% and Kt/V by 0.14 units,roviding clinicians with an alternative strategyo augment dialysis adequacy that does not re-uire changes in treatment time.Patients most at risk from inadequate delivery
f QB are those with persistently very negativerterial inlet pressures. Frequently, these are pa-ients who have catheters or dysfunctional nativeVFs. These patients also might benefit fromialysis using tubing sets that have larger flownd maintain their elastic recoil. Use of bloodubing sets that maintain their viscoelastic prop-rties and provide better delivered blood flowan be a useful strategy to improve URR andt/V without extra effort. Because the 2 blood
ubing sets have the same average sales price,2.50, the increase in dialysis delivered can
ccur without additional cost. bGLOSSARY
K, clearance of a specified solutet, duration of dialysisV, urea distribution volumeKt/V, adequacy of dialysis parameterQB, dialyzer blood flowQP, prescribed blood pump flowQP � QB, difference between prescribed and
elivered blood flowPA, pre-pump arterial pressuresPV, venous outlet or drip chamber pressureURR, urea reduction ratio
REFERENCES1. Levine J, Bernard DB: The role of urea kineticodeling, TACurea, and Kt/V in achieving optimal dialy-
is: A critical reappraisal. Am J Kidney Dis 15:285-301,9902. Held PJ, Port FK, Wolfe RA, et al: The dose of
emodialysis and mortality. Kidney Int 50:550-556, 19963. Centers for Medicare and Medicaid Services: 2002
nnual Report. End-Stage Renal Disease Clinical Perfor-ance Measures Project. Baltimore, MD, Department ofealth and Human Services, Centers for Medicare andedicaid Services, Center for Beneficiary Choices, 20024. Clark WR, Leypoldt JK, Henderson LE, Mueller BA,
cott MK, Vonesh EF: Quantifying the effect of changes inhe hemodialysis prescription on effective solute removalith a mathematical model. J Am Soc Nephrol 10:601-609,9995. Ward RA: Blood flow rate: An important determinant
f urea clearance and delivered Kt/V. Adv Ren Replace Ther:75-79, 19996. Sands JJ, Glidden D, Jacavage W, Jones B: Difference
etween delivered and prescribed blood flow (QB) in hemo-ialysis. ASAIO J 42:M717-M719, 19967. Depner TA, Rizwan S, Stasi TA: Pressure effects of
oller pumps during hemodialysis. ASASIO Trans 36:M456-459, 19908. Teruel JL, Fernandez LM, Marcen R, et al: Differences
etween blood flow as indicated by the hemodialysis bloodoller pump and blood flow measured by an ultrasonicensor. Nephron 85:142-147, 2000
9. Schmidt DF, Schniepp BJ, Kurtz SB, McCarthy JT:naccurate blood flow rate during rapid dialysis. Am Jidney Dis 17:34-37, 199110. Mehta HK, Deaubrea D, McDougall JG, GoldsteinB: Correction in the discrepancy between prescribed and
ctual blood flow rates in chronic hemodialysis patients withse of larger gauge needles. Am J Kidney Dis 39:1231-1235,00211. Hasbargan JA, Weaver DT, Hasbargan BJ: The effect
f needle gauge on recirculation, venous pressure and bleed-ng from puncture sites. Clin Nephrol 44:322-324, 1995
12. Polaschegg H. Pressure drops in cannulas for hemodi-lysis. Int J Artif Organs 24:614-623, 2001
13. Mankus RA, Ash SR, Sutton JM: Comparison of
lood flow rates and hydraulic resistance between the
Mc
mK
m(c2
aTL
U2
s1
GpJ
GKp
N5
cT
L
pK
hH
pc8
uv1
YA
a
a2
4
dE
l2
Ea
wh
AHMED ET AL508
ahurkar catheter, the Tesio twin catheter, and the Ash splitatheter. ASAIO J 44:M532-M534, 1998
14. Krivitski NM: Theory and validation of access floweasurement by dilution technique during hemodialysis.idney Int 48:244-250, 199515. Besarab A, Frinak S, Aslam M: Pressure measure-ents in the surveillance of vascular accesses, in Gray R
ed): A Multidisciplinary Approach for Hemodialysis Ac-ess. Philadelphia, PA, Lippincott, Williams & Wilkins,002, pp 137-15016. Daugirdas JT, Van Stone JC: Physiologic principles
nd urea kinetic modeling, in Daurgirdas JT, Blake PG, IngS (eds): Handbook of Dialysis (ed 3). Philadelphia, PA,ippincott, Williams & Wilkins, 2001, pp 15-4517. Depner TA: Prescribing Hemodialysis: A Guide to
rea Modeling. Norwell, MA, Kluwer, 1991, appendix G, p6518. Bosch JP, Barlee V, Valdecasa JG: Blood flow mea-
urements during hemodialysis. Adv Ren Replace Ther:83-88, 199419. Mandolfo S, Malberti F, Imbasciati E, Cogliati P,
auly A: Impact of blood and dialysate flow and surface onerformance of new polysulfone hemodialysis dialyzers. IntArtif Organs 26:113-120, 200320. Gutzwiller JP, Schneditz D, Huber AR, Schindler C,
arbani E, Zehnder CE: Increasing blood flow increasest/V (urea) and potassium removal but fails to improvehosphate removal. Clin Nephrol 59:130-136, 200321. Gotsch FA, Sargent JA: A mechanistic analysis of the
ational Cooperative Dialysis Study. Kidney Int 28:526-34, 198522. DiFilippo S, Andrulli S, Manzoni C, Corti M, Lo-
atelli F: On-line assessment of delivered dialysis dose.echnical note. Kidney Int 54:263-267, 199823. Depner T, Beck G, Daugirdas J, Kusek J, Eknoyan G:
essons from the Hemodialysis (HEMO) Study: An im- 1
roved measure of the actual hemodialysis dose. Am Jidney Dis 33:142-149, 199924. Cheung AK, Levin NW, Greene T, et al: Effects of
igh-flux hemodialysis on clinical outcomes: Results of theEMO Study. J Am Soc Nephrol 14:3251-3263, 200325. Coyne DW, Delmez J, Spence G, Windus DW: Im-
aired delivery of hemodialysis prescriptions: An analysis ofauses and an approach to evaluation. J Am Soc Nephrol:1315-1318, 199726. Besarab A, Sullivan KL, Ross R, Moritz M: The
tility of intra-access monitoring in detecting and correctingenous outlet stenoses prior to thrombosis. Kidney Int 47:364-1373, 199527. Frinak S, Zasuwa G, Dunfee T, Shetty A, Besarab A,
ee J: Computerized measurement of intra-access pressure.m J Kidney Dis 40:760-768, 200228. Polaschegg H: Pressure drops in cannulas for hemodi-
lysis. Int J Artif Organs 24:614-623, 200129. Twardowski ZJ, Haynie JD: Measurement of hemodi-
lysis catheter blood flow in vivo. Int J Artif Organs 25:276-80, 200230. Depner TA: Catheter performance. Semin Dial 14:425-
31, 200231. Stragier A, Wenderickx D, Jadoul M: Blood flow
isplayed by dialysis machines: Is it accurate? EDTNARCA J 22:3-6, 199632. Twardowski ZJ: Safety of high venous and arterial
ine pressures during hemodialysis. Semin Dial 13:336-337,00033. Chambers SD, Ceccio Sl, Annich GA, Bartlett RH:
xtreme negative pressure does not cause erythrocyte dam-ge in flowing blood. ASAIO J 45:431-435, 1999
34. Francos GC, Burke JF, Besarab A, Martinez J, Kirk-ood RG, Hummel LA: An unsuspected cause of acuteemolysis during hemodialysis. ASAIO Trans 29:140-146,
983
