Nephrol Dial Transplant (2000) 15: 517523

Nephrology
Dialysis
Transplantation

Original Article

A new method of post-dialysis blood urea sampling: the stop dialysate

flow method
Colin C. Geddes, Jamie Traynor, David Walbaum, Jonathan G. Fox and Robert A. Mactier
Renal Unit, Stobhill Hospital, Balornock Road, Glasgow, Scotland, UK

Abstract
Background. A standardized practical method of postdialysis blood sampling is required to improve the
precision of using urea kinetics in the evaluation of
haemodialysis dose and to permit comparative audit.
The methods recommended in the Renal Association
and Dialysis Outcomes Quality Initiative (DOQI )
guidelines reduce the blood pump speed to a low rate
at the end of haemodialysis before blood sampling
after 10 and 15 s respectively. However, these low
flow methods compensate only partially for cardiopulmonary recirculation and may be impractical in routine
practice because they involve sequential steps and
require accurate timing of sampling. Therefore we have
evaluated an alternative method of stopping only the
dialysate flow at the end of the haemodialysis session
before performing post-dialysis blood sampling.
Methods. The study was performed in three phases.
Serial measurements of blood urea were obtained from
arterial and venous samples taken at times 0, 30, 60,
120, 180, 240, 300 and 360 s after stopping dialysate
flow and leaving the extracorporeal blood flow rate
unchanged at the end of the haemodialysis session in
10 patients. A peripheral venous sample was also taken
from the contralateral arm at 0 s to reflect body water
urea concentration at the end of dialysis without the
effect of access recirculation and with a minimal effect
of cardiopulmonary recirculation. The same haemodialysis prescription was repeated in the same 10 patients
using the Renal Association method to permit comparison between the two methods. The practical use of
the stop dialysate flow method was then evaluated in
117 regular haemodialysis patients undergoing routine
monthly assessment of dialysis adequacy and compared
with sampling immediately post-dialysis.
Results. Within 4 min of stopping the dialysate flow
there was no difference between the blood urea concentrations of arterial and venous samples, indicating
cessation of diffusion across the dialysis membrane.
Correspondence and offprint requests to: Dr R. Mactier, Renal Unit,
Stobhill Hospital, Balornock Road, Glasgow G21 3UW, Scotland,
UK.

Also the blood urea concentrations in all of the arterial

and venous samples between 4 and 6 min were constant
and were equivalent to the blood urea concentration
of the peripheral venous sample taken at 0 s. These
data suggest that post-dialysis blood sampling may be
performed 5 min after stopping dialysate flow at the
end of the haemodialysis session. In contrast, the
blood urea concentration in the post-dialysis samples
obtained using the Renal Association method
were lower than the contralateral arm blood urea concentration taken at 0 s (0.310.42; P<0.05) and
consequently the percentage URR was higher
(1.351.84%). In 117 patients the post-dialysis blood
urea sample 5 min after stopping dialysate flow
averaged
5.492.11 mmol/1
compared
with
5.072.05 mmol/l immediately after the end of the
haemodialysis session (P<0.0001). This was equivalent
to a reduction in URR from 71.78.3% with sampling
immediately post-dialysis to 69.19.3% with the stop
dialysate flow method.
Conclusions. This study shows that there is a window
period between 4 and 6 min after stopping dialysate
flow at the end of the haemodialysis session when the
blood urea concentration in a sample taken from any
part of the extracorporeal circuit remains constantly
within the co-efficient of variation of laboratory measurement, and is equivalent to a peripheral venous
sample taken immediately at the end of the dialysis
session. A stop dialysate flow method with blood
sampling after 5 min offers several advantages over
slow flow methods, since it allows for cardiopulmonary as well as access recirculation, does not require
precise timing of blood sampling, and is simple to
perform in a busy renal unit. For these reasons the
stop dialysate flow method may be used for routine
monitoring of the adequacy of delivered haemodialysis
and for comparative audit among haemodialysis
centres.
Keywords: adequacy;
sampling; renal failure

2000 European Renal AssociationEuropean Dialysis and Transplant Association

haemodialysis;

post-dialysis

518

Introduction
The survival of patients on chronic haemodialysis is
directly related to fractional urea clearance [1,2]. For
this reason, measurements of urea removal are used to
monitor the adequacy of haemodialysis and to compare
the quality of haemodialysis delivery among dialysis
centres [3]. There are several methods of measuring
urea removal in haemodialysis and the current recommended methods are measurement of Kt/V by formal
urea kinetic modelling, calculation of an estimated
single pool Kt/V, or calculation of the urea reduction
ratio ( URR) [4,5]. All of these methods require the
measurement of the urea concentration of pre- and
post-dialysis blood samples to calculate urea clearance.
The timing of blood sampling for estimation of postdialysis urea concentration is critical since there may
be significant overestimation of urea removal if the
method of post-dialysis blood sampling does not allow
for post-dialysis urea rebound. The dilutional effects
of access recirculation and (in subjects dialysing
through arteriovenous fistulae) cardiopulmonary recirculation cease within 2 min of stopping haemodialysis
and account for 69% of total post-dialysis urea rebound
[68]. A delay in post-dialysis blood sampling for at
least 30 min is needed to allow for tissue rebound due
to intercompartmental urea dysequilibrium at the end
of haemodialysis but this approach is impractical in
routine clinical practice. There is currently no gold
standard for post-dialysis blood sampling for calculation of urea removal; however, a standardized method
that corrects for the major part of urea rebound
occurring within the first 2 min of stopping haemodialysis should be available for the routine assessment of
dialysis dose and for comparative audit.
The Renal Association ( UK ) and the National
Kidney Foundation ( USA) Dialysis Outcomes Quality
Initiative (DOQI ) guidelines currently recommend
slow flow methods of post-dialysis blood sampling
because they negate the effects of access recirculation,
but they compensate only partially for cardiopulmonary recirculation [4,5]. The slow flow method of the
Renal Association guidelines involves four steps that
require accurate timing, and sampling is performed
during the period of most rapid urea rebound [4,9]:
(i) stop ultrafiltration
(ii) reduce blood flow to 50 ml/min for 10 s exactly
(iii) clamp arterial line
(iv) draw a 3 ml sample from the arterial needle
Similarly the slow flow method recommended in
the DOQI guidelines is also a four-step procedure [5]:
(i) stop dialysate flow
(ii) set the ultrafiltration rate to 50 ml/h
(iii) reduce the blood flow rate to 50100 ml/min for
15 s
(iv) draw blood with the blood pump at the above
speed ( low-flow method ) or stop the blood pump
before drawing blood from the arterial line (stop
pump method)

C. C. Geddes et al.

The workgroup of the National Kidney Foundation

DOQI guidelines acknowledged that in busy haemodialysis units the rigor needed to execute the slow flow
method cannot always be provided by the dialysis
technicians [5]. A survey in early 1998 of the 11 adult
haemodialysis units in Scotland found that none of
the units employed a slow flow method of taking the
post-dialysis blood sample to calculate urea clearance (personal communication, S. Green, Scottish
Haemodialysis Nurses Core Group). Ten units
employed immediate post-dialysis sampling from the
arterial line. This method therefore underestimates the
post-dialysis blood urea concentration and overestimates urea removal because of access and cardiopulmonary recirculation The other haemodialysis centre
performed post-dialysis sampling after the blood in the
extracorporeal circuit had been re-infused (re-infusion
method). This method is also imprecise because the
time taken to re-infuse blood is variable and because
of dilution with infused saline. A recent survey of
North American dialysis facilities revealed that there
were at least 20 post-dialysis sampling methods in use,
and that up to 41.6% of the centres surveyed were
using a method of post-dialysis blood sampling that
did not allow for the effects of access and cardiopulmonary recirculation and were therefore overestimating urea removal by dialysis [10]. A recent
report from the Dialysis Outcomes and Practice
Patterns Study (DOPPS ) has shown that units that
obtained the post-dialysis sample more than 3 min
after the end of haemodialysis had a 7% lower estimated Kt/V than patients with blood sampling between
30 and 60 s post-dialysis, and introducing a 0.1 Kt/V
correction in all patients with post-dialysis sampling
before 3 min increased the proportion of the patients
receiving a Kt/V below 1.2 from 26% to 36% [11].
These observations emphasize both the importance of
timing and the need for standardization of the method
of post-dialysis sampling.
This study describes a new method of post-dialysis
blood sampling after stopping the dialysate flow at the
end of the dialysis session for 5 min instead of reducing
extracorporeal blood flow to a low rate for 10 or 15 s
[4,5]. By leaving the extracorporeal blood flow rate
unchanged the stop dialysate flow method should
avoid the risk of clotting in the extracorporeal circuit,
which may occur if more prolonged low blood flow
rates were recommended with a slow flow method to
allow fully for the effects of both access and cardiopulmonary recirculation. After dialysis ceased after stopping the dialysate flow at the end of the haemodialysis
session it was predicted from previous studies that
access recirculation would stop after 30 s, cardiopulmonary circulation would cease after 2 min, and a
window period may exist over the next few minutes
when the effect of rebound of urea from tissues would
be minimal [8,12] This hypothesis was confirmed. This
stop dialysate flow method was compared to the
Renal Association slow flow method [4] and was also
then compared with immediate post-dialysis sampling,

Stop dialysate flow method of post-dialysis blood sampling

which was the method in current use in most Scottish

haemodialysis units.

519

when equilibration with the dialysate remaining in the

dialyser occurred.
Results were expressed as meanstandard deviation (SD).
Statistical comparison was by paired t-test.

Subjects and methods

(ii) Comparison of stop dialysate flow and slow flow
methods

The study was conducted in three phases

(i) Evaluation of the stop dialysate flow method

The blood urea concentration in the arterial and venous
ports was analysed serially for several minutes after stopping
dialysate flow to identify if there was a window period
when the blood urea concentration in the extracorporeal
circuit may represent the post-dialysis urea concentration
without the effects of access or cardiopulmonary recirculation
[12]. Ten subjects on haemodialysis for end-stage renal failure
who had been clinically well over the preceding month and
had functioning arteriovenous fistulae (usual mean blood
flow >300 ml/min) were invited to participate and gave
written informed consent. The demographics and haemodialysis prescriptions of the subjects are summarized in Table 1.
Pre-dialysis blood was aspirated ( Urea ) from the arterial
pre
limb of the arteriovenous fistula before the introduction of
heparin or saline and at the end of the prescribed dialysis
session dialysate flow was switched off (time 0). Ultrafiltration was also stopped. The blood pump remained at the
usual prescribed flow rate. Blood was sampled from both
the arterial port (before the dialyser) and venous port (after
the dialyser) at time 0, 30, 60, 120, 180, 240, 300 and 360 s.
Blood was sampled from the contralateral arm at time 0
(contralateral t ) to obtain a representative sample of body
0
water urea concentration immediately at the end of haemodialysis without the effect of access recirculation and with a
minimal effect of cardiopulmonary recirculation. Three
investigators conducted the serial blood sampling to ensure
exact timing. Urea concentration in the arterial port samples
(A , A , A , A , A , A , A , A ) was compared
0 30 60 120 180 240 300 360
with urea concentration in the contralateral t sample to
0
identify a window period when access and cardiopulmonary
recirculation had ceased, tissue rebound was minimal, and
the blood urea concentration was relatively constant. Urea
concentrations in the samples from this time period were
compared with A . Urea concentrations in samples from the
0
venous port ( V , V , V , V , V , V , V , V ) were
0 30 60 120 180 240 300 360
compared with the corresponding arterial samples to identify

Post-dialysis blood was sampled using the slow flow method

described by the Renal Association [4] in the same 10 subjects
on a separate dialysis session and a contralateral time 0
sample was again taken to represent the post-dialysis blood
urea concentration immediately at the end of haemodialysis
without the influence of access recirculation and with a
minimal effect of cardiopulmonary recirculation. The differences between post-dialysis blood circuit urea : contralateral
t urea were determined and compared for the slow flow
0
and stop dialysate flow methods. The urea clearance was
compared for these two methods by deriving URR and Kt/V
using the following formulae:
URR=100(1Urea /Urea )%
post
pre
where Urea
is post-dialysis blood urea concentration and
post
Urea is the pre-dialysis blood urea concentration [1,3];
pre
Kt/V=Ln(R0.008t)+(43.5R)UF/W
where Ln is the natural logarithm, R is the post-dialysis
blood urea/pre-dialysis blood urea, t is the dialysis session
length in hours, UF is the ultrafiltration volume in litres,
and W is the post-dialysis weight in kilograms [13].

(iii) Comparison of the stop dialysate flow method with

immediate post-dialysis sampling
The stop dialysate flow method was compared with immediate post-dialysis sampling in the same dialysis session in all
117 patients on hospital haemodialysis in Stobhill Hospital.
All of the blood samples were taken by nursing staff to assess
if the method was applicable for routine use. The blood urea
concentration using the stop dialysate flow method was
compared with the blood urea concentration in the arterial
limb of the arteriovenous fistula at time 0. The urea clearance
was compared for these two methods by calculating URR
(see formula above).
All results were expressed as meanSD. Statistical
comparison was by paired t-test of means. For comparison

Table 1. Summary of the demographic and haemodialysis prescriptions of the 10 subjects studied
Subjects

Age
(years)

Sex

Hours HD
per week

Dialyser

Mean blood flow

(ml/min)*

Mean UF volume per

session ( litres)*

Mean post HD
weight (kg)*

1
2
3
4
5
6
7
8
9
10
Mean

64.8
71.1
54.2
51.8
44.1
22.8
50.5
28.0
47.8
47.1
48.2

m
f
m
m
f
f
m
m
f
m

12
12
12
12
12
12
12
15
12
15
12.6

F8
AM65
AM65
F8
AM65
AM65
AM65
AM65
AM65
AM65

350
350
305
350
260
233.5
350
320
300
292
311

2.5
2.0
2.8
3.3
2.3
2.6
2.7
4.2
3.0
2.6
2.8

55.2
62.6
66.3
57.9
40.9
48.8
67.8
72.5
51.3
78.6
60.2

*Mean of the two dialysis sessions comparing the stop dialysate flow method with the slow flow method.

520

C. C. Geddes et al.

between the stop dialysate flow and slow flow methods,

the results were expressed as the difference from the values
of the contralateral t sample obtained at the end of the
0
same dialysis session.

Measurement of blood urea concentration

Blood urea concentrations were measured by the kinetic
urease method (Sigma Chemical Co. Ltd ) on an Olympus
discrete analyser (Au 5200). For each dialysis session, all
samples were analysed in the same batch. The coefficient of
variation of urea concentrations in the range of urea concentrations under evaluation was 0.8% for all samples batch
tested from an individual dialysis and 2% for samples from
different dialysis sessions.
Ethical approval was obtained from the local research and
ethics committee.

Results
Identification of a window period using the stop
dialysate flow method
Post-dialysis blood urea concentrations in serial
samples from the arterial port after switching off
dialysate flow at the end of the prescribed dialysis
session in the 10 subjects are shown in Table 2. The
mean blood urea concentration in A is significantly
0
lower than A , A , and A
(5.182.41 vs
240
300
360
5.752.42, P=0.01; 5.842.43, P=0.004; 5.852.47,
P=0.003 respectively). There was no significant difference between the mean urea concentration in A ,
240
A , and A . The mean blood urea concentrations
300
360
in serial samples from the arterial and venous ports
after stopping dialysate flow at the end of haemodialysis are shown in Figure 1. There was no difference
between urea concentration in A , A , A
and
240
300
360
V , V , V respectively, indicating that by 46 min
240 300 360
the dialysate remaining in the dialyser after stopping
dialysate flow had equilibrated with blood urea.
Figure 1 shows the urea concentration in serial
samples from the arterial and venous ports expressed
as a fraction of the contralateral time 0 urea. The urea
concentrations in A , A , and A were the same
240 300
360
as in contralateral time 0 (5.852.54, P=0.58,
P=0.95, P=1.00 respectively). Mean blood urea
concentrations and URR of the contralateral arm
time 0 urea are compared with post-dialysis sampling
Table 2. Evaluation of stop dialysate flow method

Contralateral arm
0s
180 s
240 s
300 s
360 s

Mean post-dialysis blood

urea (mmol/l )

URR (%)

5.85
5.18
5.52
5.75
5.84
5.85

69.41 (8.94)
72.88 (9.00)*
71.11 (8.11)**
69.87 (8.50)**
69.40 (8.39)**
69.31 (8.83)**

(2.54)
(2.41)*
(2.36)**
(2.41)**
(2.43)**
(2.46)**

P<0.05; *P>0.05 compared to contralateral arm (paired t-test).

Fig. 1. Mean arterial and venous blood urea concentrations after

stopping dialysate flow expressed as a fraction of blood urea
concentration of contralateral arm at time zero versus time (n=10).

at sequential times after stopping dialysate flow in

Table 2. These results indicate that the blood urea
concentrations and hence the URR derived from A ,
240
A , and A are equivalent to the contralateral arm
300
360
time 0 blood urea concentration and URR.
On the basis of these data, 300 s (5 min) after
switching off dialysate flow was selected as the appropriate time to sample blood from the extracorporeal
circuit for the stop dialysate flow method.
Comparison of stop dialysate flow and slow flow
methods
The pre- and post-dialysis blood urea concentrations
using both the stop dialysate flow method with blood
sampling after 5 min and the Renal Association slow
flow method with the same haemodialysis prescriptions in the same 10 patients are compared in Table 3.
The mean post-dialysis urea concentrations were similar to the contralateral arm time 0 urea in the stop
dialysate flow method but significantly lower than
the contralateral arm time 0 urea concentration in the
slow flow method (6.292.80 and 5.982.66, P=
0.043). Consequently the derived URR with the slow
flow method was significantly higher than the URR
calculated from the contralateral arm time 0 urea
concentration (727 and 707%, P=0.045).
Differences in blood urea concentrations and URR
between the contralateral arm time 0 and the stop
dialysate flow and slow flow blood sampling methods
are summarized in Table 4. Post-dialysis blood urea
concentrations and URR in samples between 240 and
360 s after stopping dialysate flow and the contralateral
arm time 0 were equivalent, but were significantly
different from results of sampling using the slow flow
method or immediately after the end of haemodialysis
(P<0.05).
Comparison of stop dialysate flow method with blood
sampling immediately post-dialysis
The post-dialysis blood urea measured using a sample
taken immediately at the end of the dialysis session
was significantly lower than the post-dialysis urea
measured by the stop dialysate flow method at the

Stop dialysate flow method of post-dialysis blood sampling

521

Table 3. Comparison of pre- and post-dialysis blood urea concentrations by the stop dialysate flow and slow flow methods on two
separate dialysis sessions in the same 10 subjects. Urea =pre-dialysis urea; urea =post-dialysis urea; contralateral time =contralateral
pre
post
0
arm venous sample taken immediately at the end of haemodialysis session
Stop dialysate flow method

Slow flow method

Subject

Urea
pre
(mmol/l )

Urea
post
(mmol/l )

Contralateral
time
0
(mmol/l )

Urea
pre
(mmol/l )

Urea
post
(mmol/l )

Contralateral
time
0
(mmol/l )

1
2
3
4
5
6
7
8
9
10
mean
standard
deviation

15.1
19.3
21.4
21.2
21
15.4
19.3
24.1
16.1
12.6
18.55
3.59

3.1
4.2
7.8
6.6
8.9
3
7.3
9.4
4
4.1
5.84
2.43

2.9
3.7
7.6
7.9
8.9
3.2
6.8
9.5
3.8
4.2
5.85
2.54

8.5
25.2
24.2
32.5
23.9
16.6
18
25.9
16.1
16.1
20.70
6.87

1.9
4.9
8.3
10.5
4.7
4.1
6.6
9.4
4.7
4.7
5.98*
2.67

1.8
5.5
8.8
11.4
5
4.3
7.4
8.9
4.9
4.9
6.29*
2.80

*P<0.05 (paired t-test).

end of the same dialysis session in 117 hospital haemodialysis subjects who were having routine monthly urea
clearance measurements (5.072.05 mmol/l vs
5.492.11 mmol/l, P<0.0001). The corresponding
URR, calculated using immediate post-dialysis sampling, was significantly higher than the URR calculated
using the stop dialysate flow method. (728 vs
699%, P<0.0001). Cumulative URR data using
stop dialysate flow and no slow flow methods are
compared in Figure 2. The cumulative URR data in
the haemodialysis population was shifted significantly
to the left using the stop dialysate flow method instead
of immediate post-dialysis blood sampling and converted six patients (5% of the patients studied ) with
an URR above 65% to below the current minimum
standard of the Renal Association [4]. In a subset of
23 patients using double-lumen central venous catheters (six temporary and 17 permanent catheters) for
haemodialysis blood samples were taken from the
venous as well as the arterial lumen after stopping the
dialysate flow for 5 min. The mean blood urea concentrations in the arterial and venous samples were similar
(5.992.9 and 5.912.84 mmol/l respectively, P=
0.45) indicating that blood sampling with the stop
dialysate flow method can be obtained from any part
of the extracorporeal circuit.

was completed ( Table 2). This stop dialysate flow

method therefore compensates for the initial urea
rebound, which occurs during the first 2 min after
stopping haemodialysis due to access and cardiopulmonary recirculation. This study shows that for the
first 4 min after stopping the dialysate flow at the end
of a haemodialysis session dialysis still takes place into
the dialysate within the dialyser as the urea concentration in blood leaving the dialyser is initially lower than
blood entering the dialyser (Figure 1). However, this
small amount of dialysis occurring into a stagnant
pool of dialysate within the dialyser will have no
significant influence on haemodialysis adequacy. The
range and frequency of use of different dialysers and
the duration and efficiency of haemodialysis were similar in the three study phases and are representative of
current haemodialysis prescribing practice. Between 4
and 6 min after stopping dialysate flow the urea concentration from the arterial port remains relatively
constant and there is also no difference between the
urea concentrations in samples from the venous and
arterial ports, indicating equilibration with the dialysate remaining in the dialyser (Figure 1). There is also

Discussion
These data show that the stop dialysate flow method
is a useful and valid new method for post-dialysis
blood urea sampling for monitoring of haemodialysis
adequacy. Blood sampling from any part of the extracorporeal circuit 5 min after stopping the dialysate
flow at the end of the haemodialysis session was shown
to be equivalent to blood sampling from the contralateral arm immediately after the haemodialysis session

Fig. 2. Cumulative URR using no slow flow method (0 s) and stop

dialysate flow method (5 mins).

522

C. C. Geddes et al.

no difference between the dialysis circuit blood urea

between 4 and 6 min after stopping dialysate flow and
the contralateral arm venous sample at time 0 (contralateral t ) confirming that access and cardiopulmon0
ary recirculation are no longer lowering the dialysis
circuit blood urea concentration (Figure 1). Although
blood urea concentrations remained relatively constant
between 4 and 6 min it may be anticipated that the
blood urea concentration would rise slowly over the
next 30 min due to tissue rebound, although this was
not assessed in the present study [8,12].
It is not possible to compare the stop dialysate flow
and slow flow methods in the same haemodialysis
session for technical reasons. The two methods were
compared in the same subjects on different haemodialysis sessions in the same month by using the blood
urea concentration in a contralateral arm venous
sample at time 0 post-dialysis as a reference. This
showed that the blood urea concentration post-dialysis
is slightly lower when measured using the slow flow
method ( Table 3). This is expected, as the slow flow
method does not negate the effect of cardiopulmonary
recirculation, which is important since Schneditz et al.
showed that failing to take cardiopulmonary recirculation into account results in a 614% overestimation of
urea removal [14]. Obtaining a blood sample without
waiting for access and cardiopulmonary recirculation
to cease was shown in the present study to significantly
underestimate systemic urea concentration by approximately 10% ( Table 2) and therefore overestimate urea
removal when compared to the stop dialysate flow
method ( Table 4).
A contralateral arm venous sample immediately after
the haemodialysis session was used in this study to
estimate body water urea concentration at the end of
dialysis and permit comparisons between the stop
dialysate flow, slow flow, and no slow flow methods
of post-dialysis sampling. Some authors have raised
doubts regarding the validity of using samples from
the contralateral arm as an index of post-dialysis body
water urea concentration [15]. Depner et al. have
shown that blood urea concentrations in samples from
the contralateral arm during and after high-efficiency
haemodialysis were higher than the blood urea concentration in the arterial port after the fistula had been

occluded between the arterial and venous needles for

1 min to limit the effect of access recirculation [15].
This finding was attributed to haemodialysis-induced
compartment dysequilibrium in the relatively poorly
perfused contralateral arm [15] but is at least partly
due to the differential effects of cardiopulmonary recirculation in the arteriovenous and contralateral arms
[14]. Thus contralateral arm venous sampling at the
end of haemodialysis is an acceptable reference method
of obtaining a representative post-dialysis blood
sample for urea kinetics but it is not practical for
routine use.
There are several technical advantages with the stop
dialysate flow method described in this study over the
slow flow method. The stop dialysate flow method
negates the effect of cardiopulmonary as well as access
recirculation whereas recommended slow flow
methods [4,5,16 ] only allow partially for cardiopulmonary recirculation. A slow flow method would
need to delay blood sampling for at least 2 min to
obviate the effects of both access and cardiopulmonary
recirculation [17], which would confer an increased
risk of thrombosis in the extracorporeal circuit when
perfused at a low flow rate, especially since many renal
units discontinue heparin for the last 30 min of haemodialysis. The stop dialysate flow method requires only
one step by the dialysis technician or nurse before
taking the blood sample (i.e. stopping the dialysate
flow at time 0). The Renal Association slow flow
method [4] requires three steps before blood sampling
(i.e. stopping ultrafiltration and clamping the arterial
tubing exactly 10 s after slowing blood flow to
50100 ml/min). Other recommended slow flow
methods [5,16 ] also involve multiple steps requiring
accurate timing of blood sampling during the early
post-dialysis period of most rapid urea rebound, which
is likely to decrease the precision of urea measurement.
The timing of sampling with the stop dialysate flow
method is not as critical as with a slow flow method,
since the variability in blood urea concentrations at 4,
5, or 6 min after stopping dialysate flow is small and
is within the reported co-efficient of variation of urea
measurements. The sample can be taken from the
arterial or venous port with the stop dialysate flow
method, whereas it is critical that the correct tubing is

Table 4. Comparison of urea kinetics using stop dialysate flow and Renal Association slow flow methods
Time post-dialysis

0s
Renal Association
slow flow (10 s)
stop dialysate 180
stop dialysate 240
stop dialysate 300
stop dialysate 360
*P=0.01; **P=0.04.

Mean difference between

contralateral arm blood
urea conc and post-dialysis
urea concentration

s
s
s
s

Mean difference between

contralateral arm URR and
post-dialysis URR%

Mean difference between

contralateral arm Kt/V and
post-dialysis Kt/V

0.67*

3.47*

0.15*

0.31**
0.33
0.10
0.01
0.00

1.35**
1.69
0.45
0.01
0.10

0.05**
0.06
0.01
0.01
0.01

Stop dialysate flow method of post-dialysis blood sampling

clamped and sampled when using the slow flow

method. The most likely technical error with the stop
dialysate flow method is forgetting to switch off the
dialysate flow. If this is recognized during the procedure, then the dialysate flow can be switched off and
the sample can be taken 5 min later. Errors using the
slow flow method cannot be corrected once they have
occurred.

Conclusion
A standardized, practical method of post-dialysis blood
sampling which reliably measures the true post-dialysis
urea concentration would improve prescription of
adequate haemodialysis therapy and allow meaningful
comparisons between haemodialysis centres. The stop
dialysate flow method with blood sampling after 5 min
offers several advantages over slow flow methods
since it:
(i) allows
for
access
and
cardiopulmonary
recirculation
(ii) is equivalent to blood sampling from the contralateral arm at time 0
(iii) does not require exact timing of blood sampling
(iv) permits blood sampling from arterial or venous
limbs of the extracorporeal circuit
(v) provides lower risk of thrombosis of the extracorporeal circuit
(vi) is simple to perform and therefore practical in a
busy renal unit
For these reasons the stop dialysate flow method
of post-dialysis blood sampling may be used for the
routine monitoring of adequacy of delivered haemodialysis and for comparative audit in haemodialysis units.

523

2.
3.
4.
5.
6.
7.

8.
9.
10.

11.
12.
13.
14.
15.

Acknowledgements. This study was presented in part at the meeting

of the Scottish Renal Association in Dumfries, November 1998.

16.

References

17.

1. Owen WF Jr, Lew NL, Liu Y, Lowrie EG, Lazarus JM. The
urea reduction ratio and serum albumin concentration as pre-

dictors of mortality in patients undergoing haemodialysis. N

Engl J Med 1993; 329: 10011006
Held PJ, Port FK, Wolfe RA, Stannard DC, Carroll CE,
Daugirdas JT. The dose of haemodialysis and patient mortality.
Kidney Int 1996; 50: 550556
Junor BJR. Audit of quality of hospital haemodialysis in
Scotland. The Scottish Renal Registry. Nephrol Dial Transplant
1998; 13: 24262427
Treatment of Adult Patients with Renal Failure. Recommended
Standards and Audit Measures. The Renal Association. 1998.
2122
Clinical Practice Guidelines for Haemodialysis Adequacy.
National Kidney Foundation Dialysis Outcomes Quality
Initiative (DOQI ). 1997: 2561
Sherman RA, Matera JJ, Novik L, Cody RP. Recirculation
re-assessed: the impact of blood flow rate and the low-flow
method re-evaluated. Am J Kidney Dis 1994; 23: 846848
Daugirdas JT, Burke MS, Balter P, Priester-Coary A, Majika
T. Screening for extreme post-dialysis rebound using the Smye
method: patients with access recirculation identified when a slow
flow method is not used to draw the post-dialysis blood. Am
J Kidney Dis 1996; 28: 727731
Depner T. Assessing the adequacy of haemodialysis: urea modelling. Kidney Int 1994; 45: 15221535
Priester-Coary A, Daugirdas JT. A recommended technique for
obtaining the post-dialysis BUN. Semin Dial 1997; 10: 2325
Beto JA, Bansal VK, Ing TS, Daugirdas JT. Variation in blood
sample collection for determination of haemodialysis adequacy.
Am J Kid Dis 1998; 31: 135141
Young EW, Sethuraman G, Dickinson DM, Keen ML, Held
PJ. The effect of timing of the post-dialysis BUN draw on
calculated delivered Kt/V. J Am Soc Nephrol 1998; 9: 188A
Pedrini LA, Zereik S, Rasmy S. Causes, kinetics and clinical
implications of post haemodialysis urea rebound. Kidney Int
1988; 34: 817824
Daugirdas JT. Second generation logarithmic estimates of singlepool variable volume Kt/V: an analysis of error. J Am Soc
Nephrol 1993; 4: 12051213
Schneditz D, Kaufman AM, Polaschegg HD, Levin NW,
Daugirdas JT. Cardiopulmonary recirculation during haemodialysis. Kidney Int 1992; 42: 14501456
Depner TA, Rizwan S, Cheer AY, Wagner JM, Eder LTI. High
venous urea concentrations in the opposite arm. A consequence
of haemodialysis-induced compartment disequilibrium. ASAIO
Trans 1991; 7: 141143
Depner TA. Prescribing Haemodialysis: A Guide to Urea Kinetic
Modelling. Kluwer Academic Publisher, Norwell, Massachusetts,
USA: l99l; 270
Schneditz D, Daugirdas JT, Cu GA, Morris AT, Polaschegg
HD, Levin NW, Kaufman AM. Impact of cardiopulmonary
recirculation on measurement of access recirculation J Am Soc
Nephrol 1992; 3: 393

Received for publication: 22.3.99

Accepted in revised form: 4.11.99