Kidney International, Vol. 55 (1999), pp. 1131–1149
NEPHROLOGY FORUM
The optimization of continuous ambulatory peritoneal dialysis
Principal discussant: Dimitrios G. Oreopoulos
The Toronto Hospital and University of Toronto, Toronto, Canada
CASE PRESENTATION
A 55-year-old Greek Canadian, a restaurant owner, developed end-stage renal disease secondary to proliferative glomerulonephritis of 3 years’ duration. Because he was fatigued, had
a poor appetite with occasional morning nausea, and could not
sleep well, he began continuous ambulatory peritoneal dialysis
(CAPD) 5 years ago through a surgically implanted Toronto
Western peritoneal catheter. At that time his weight was 71 kg
and body surface area 1.76 m2. He had substantial residual renal
function, with a 24-hour creatinine clearance of 9.0 ml/min, a
urea clearance of 4.3 ml/min, and a urine volume of 1100 ml/
day. Initial biochemical investigations showed a serum albumin
of 3.5 g/dl (Bromcresol green method); hemoglobin, 10.1 g/dl;
and hematocrit, 33%. He was started on a “standard” CAPD
regimen of 2-liter exchanges four times daily. At the end of
the second week, clearance measurements with the Adequestt
program were as follows: D/P creatinine, 0.85; KT/V (urea),
2.57 (1.07 renal and 1.5 peritoneal); weekly creatinine clearance
(WCC) corrected for 1.73 m2 BSA, 113 liters (65 liters renal
and 48 liters peritoneal); and normalized equivalent of protein
nitrogen appearance (NPNA), 0.8 g/kg/day. The blood urea
was 20 mm/liter; serum creatinine, 416 mm/liter (4.7 mg/dl); and
blood pressure, 150/80 mm Hg.
For the next two years he did well on CAPD. He felt well
and continued working in his restaurant. His appetite was good,
he slept well, and he had an active sexual life. For the first
year, his blood pressure was normal without antihypertensive
Presentation of this Forum is made possible by grants from Merck &
Company, Incorporated; Astra Pharmaceuticals; Hoechst Marion
Roussel, Incorporated; Dialysis Clinic, Incorporated; and R & D Laboratories, Incorporated.
Key words: chronic renal failure, peritoneal dialysis, nutrition, adequacy.
 1999 by the International Society of Nephrology
medication, but he required medication thereafter. He had no
episodes of peritonitis and gained 5 kg. However, his urine
output gradually declined despite large doses of furosemide.
At the beginning of the third year, despite an increasing dose
and number of antihypertensive medications, he complained
of progressive fatigue, anorexia, and ankle swelling. He was
hypertensive (blood pressure, 180/100 mm Hg), and his urine
output was negligible. His serum albumin had decreased to 3.0
g/dl, the blood urea had decreased to 12 mm/liter, and the
serum creatinine was 1100 mm/liter (12.5 mg/dl). On repeat
assessment, his clearances were KT/V, 1.55 (1.45 peritoneal
and 0.1 renal), and WCC, 48 liters (42 liters peritoneal and 6
liters renal). His NPNA had decreased to 0.65 g/kg/day. His
weight had increased by 3 kg and he was edematous.
His CAPD prescription was increased to 2.5 liters 4 times
daily, after which his KT/V increased to 1.75 (1.65 peritoneal
and 0.1 renal) and his weekly creatinine clearance to 58 liters
(52 liters peritoneal and 6 liters renal). After 3 months, his
NPNA remained low at 0.60 g/kg/day. His uremic symptoms
and hypertension became worse, and he could not achieve
his target weight despite three hypertonic exchanges per day.
Average ultrafiltration with 2.5% glucose at four hours was 25
ml, and with 4.25% glucose was 180 ml. Dialysate protein losses
were 12 g/day and his serum albumin fell to 2.8 g/dl. Because his
condition was getting worse, he was transferred to automated
peritoneal dialysis (APD) with 4 3 2.8 liter exchange over nine
hours each night and a three-hour 2.5 liter exchange each
evening.
On this regimen, his KT/V was 2.1 and creatinine clearance
60 liters/week. Furthermore, his net ultrafiltration increased,
he attained his target weight and, after six months, his NPNA
had increased to 0.8 g/kg/day, but his serum albumin remained
low at 3.1 g/dl. Clinically he felt better. He was able to sleep
well and he resumed working part-time work in his restaurant.
Three months later he received a successful transplant with a
cadaveric kidney.
DISCUSSION
Dr. Dimitrios G. Oreopoulos (Director, Peritoneal
Dialysis Program, The Toronto Hospital—Western Division, and Professor of Medicine, University of Toronto,
Toronto, Ontario, Canada): A Nephrology Forum given
in 1983 by Dr. Stephen Vas [1] was devoted to the diagnosis and treatment of peritonitis—at that time, the main
complication of CAPD. Since then, the introduction of
several disconnect systems has decreased the rate of peritonitis substantially. Although it is still a serious complication, peritonitis is no longer the main reason resulting
1131
1132
Nephrology Forum: Optimization of CAPD
in discontinuation of CAPD and transfer to hemodialysis
(HD). Instead, the main cause for failure of CAPD in
Canada currently is inadequate dialysis; 12% of CAPD
patients fail for this reason versus 8.6% because of peritonitis [2]. Underdialysis is now common among CAPD
patients because an increasing number have been treated
with this technique for more than three years, an interval
over which residual renal function usually declines to zero.
Today’s patient, a 55-year-old man with high peritoneal permeability (defined by D/P creatinine . 0.80) was
started relatively early on dialysis. That is, his endogenous creatinine clearance (CCr) was 9.0 ml/min. The low
serum albumin (3.5 g/dl) at the start of dialysis might
have been due to a combination of factors such as proteinuria, about which we have no information, and an
inadequate protein intake. His clinical symptoms improved and his elevated blood pressure was controlled
by CAPD. In addition, his appetite improved and he
gained weight. By the end of the second year, however,
he again became uremic and hypertension returned; a
further decrease in his serum albumin and a decrease in
blood urea indicated low protein intake and confirmed
the low NPNA of 0.65 g/kg/day. These changes primarily
reflect the decline in his residual renal function with the
resulting inadequate clearance of urea and creatinine.
An increase in the dialysis solution volume to 10 liters/
day did not improve his symptoms or correct his significant ultrafiltration failure secondary to his high transport
state (that is, fast glucose absorption causing dissipation
of the osmotic gradient responsible for ultrafiltration).
His clinical condition improved, however, after his
dialysis technique was converted to APD, an enhanced
type of dialysis with short night exchanges of large volumes, and one relatively long (3-hour) exchange during
the daytime. His urea and creatinine clearances improved, and his fluid status was better controlled. After
six months, his NPNA increased but his serum albumin
level remained low. This man’s history demonstrates all
the points about adequacy of peritoneal dialysis that I
will discuss during this Forum, specifically, the clearance
of small molecules (KT/V urea, WCC), the importance
of residual renal function, the central role of malnutrition, the phenomenon of high peritoneal permeability
and, finally, the importance of fluid/volume overload and
its control.
When physicians devise regimens for patients with
chronic renal failure, the treatment is optimal if it eliminates all signs and symptoms of the disease and ensures
the same life expectancy as they would have in the absence of the disease. With few exceptions, however [3],
the life expectancy of dialysis patients is much shorter
than that of an aged-matched population, and the term
“adequate dialysis” is a compromise that implies an “acceptable outcome,” which one hopes will be close to
optimal [4].
In the 1960s and 1970s, clinicians based their definitions of adequacy of dialysis on their clinical acumen
and a review of classic biochemical parameters such as
blood urea, creatinine level, and hematocrit. These parameters remain valid, albeit insufficient, criteria for determining the adequacy of dialysis. Symptoms and signs
of underdialysis such as nausea, vomiting, fatigue, weakness, insomnia, and restless legs also should be rigorously
sought in assessing the adequacy of dialysis. These latter
criteria can be due to other causes, however, and it is
difficult to quantitate them. For these reasons, these
symptoms and signs are not sufficient for defining adequacy and predicting outcome. Furthermore, one might
not recognize underdialysis early if one bases the diagnosis only on clinical and biochemical parameters, and it
might not be possible to “catch up” if the dose of dialysis
is not increased soon after underdialysis begins. We
therefore need now objective, quantifiable parameters
by which we can assess adequacy of dialysis and detect
underdialysis before clinical symptoms and signs appear.
In this presentation I will discuss only three major
dialysis-related factors that can predict outcome, and I
will discuss the relationship among them: small-solute
clearances, nutrition, and fluid/volume control. These
three factors probably are interrelated, and achieving
adequate dialysis requires that we address all three together. Optimization of all three clearly will improve
outcomes for all our CAPD patients. I will not examine
the effect of cardiovascular disease, lipid abnormalities
and their control on dialysis outcome, or issues concerning quality of life, not because they are not important,
but because I believe their impact on patient survival
deserves a separate discussion.
Small-solute clearances
The National Cooperative Dialysis Study (NCDS), the
first attempt at providing objective parameters of adequacy [5], showed that parameters derived from urea
kinetic modeling (UKM) best describe the adequacy of
hemodialysis. According to that study, indicators of a
low morbidity are an NPNA . 1 g/kg/day, and a timeaveraged concentration of BUN <50 mg/dl. Based on
the results of the NCDS study, Gotch and Sargent developed the concept of KT/V that integrated efficiency of
solute clearance (K), treatment time (T), and patient
size (V) [6]. This dimensionless measure of fractional
clearance of body water for urea has been accepted and
validated as an index of adequacy for hemodialysis. After
establishing that KT/V was a reliable predictor of outcome in hemodialysis, it was reasonable to extrapolate
these same concepts to the prediction of outcome in
patients on peritoneal dialysis. In hemodialysis, the rate
of solute removal changes during dialysis because the
concentration of blood urea decreases during dialysis.
In CAPD, because the blood urea concentration is rela-
Nephrology Forum: Optimization of CAPD
tively constant, clearance and solute removal stay about
the same and are related in a linear fashion. In most
CAPD patients, the urea in dialysate and plasma achieve
virtual equilibrium. Thus, given the stable blood urea
concentration, the drain volume is equivalent to urea
removal (KT) [4]. This constant relationship between
blood levels and removal rate allows continuous treatments like CAPD to remove the same volume of solute
per week as does thrice-weekly hemodialysis sessions,
even though the clearance rates are substantially lower:
5 to 10 ml/min for CAPD versus 200 to 300 ml/min for
HD [7, 8]. Teehan et al, who first applied the UKM
principle to CAPD, showed that measurements of BUN,
NPNA, and KT/V provide a reasonable basis for uniform
prescription of peritoneal dialysis and allow comparison
of treatment between patients and centers as well as
optimization of dialysis and nutrition during therapy [9].
Measuring KT/V in PD patients. After repeated measurements of KTrp/V, which includes both the contribution of the renal (KTr/V) and the peritoneal clearance
(KTp/V), the coefficient of variation is 8% [10]. The
greatest variability is seen with KTr/V (35%); only moderate variability is seen for KTp/V (7%) [10]. The renal
component varies even more widely when the amount
of urine excreted in 24 hours is very low (1–2 voidings/
day), in which case calculations of clearance should be
based on a 48-hour urine collection.
The peritoneal component (KTp/V) of KTrp/V varies
chiefly because of variations in the calculation of V,
that is, the volume of distribution of urea. For practical
purposes, this is considered equal to the total-body water, although studies using isotopes of urea found that
urea distribution space can be smaller than total-body
water space by approximately 12% to 14% [11]. Total
body water (V) can be estimated as a fixed fraction of
body weight (60% of BW for men and 55% of BW for
women) or an average of 58% for any individual, or by
anthropometric formulas such as the Watson formula,
which most investigators now consider the method of
choice [12].
Because no agreement has been reached concerning
the weight to be used in estimating V, different results
of KT/V have been reported. In general, estimates of V
based on actual weight should be most accurate when
actual and desired (ideal) weight do not differ greatly
[12]. In obese patients (wt . 20% of ideal BW), V should
be calculated using ideal BW [12]. Otherwise, the V will
be overestimated and KT/V underestimated. In wasted
patients, KT/V will be misleadingly high if one uses actual body weight for the calculation of V; instead, to
derive a KT/V indicative of a dialysis dose related to a
healthier weight, one should use standard body weight,
which can be obtained after calculating V with the Watson formula and dividing it by 0.58 [9]. The recommendation of the Dialysis Outcome Quality Initiative (DOQI)
1133
Fig. 1. Equivalence of CAPD and HD KT/Vs based on the peak concentration hypothesis. (From Ref. 15.)
work group for the malnourished patient, in whom no
other cause of malnutrition is discovered, is to increase
the dose of dialysis by multiplying the target KT/V by
the ratio V desired/V actual [13].
Comparing KT/V targets in PD and HD: A paradox.
Although it is logical to attempt to extrapolate UKM
data from patients treated with HD to those treated with
PD, it is paradoxical that the typical KT/V delivered to
CAPD patients is between 1.2 and 2.4 per week [14]. By
HD standards, this figure is well below the accepted level
of 3.0 to 3.6/week; however, numerous comparisons of
the two modalities have shown no difference in morbidity and mortality. Keshaviah and colleagues have tried to
resolve this paradox by invoking the peak concentration
hypothesis [15], which asserts that the success of CAPD
is related to its continuous, steady-state nature. If uremic
toxicity is associated with the peak concentration of BUN
rather than its time-average value, then in hemodialysis
patients, for approximately one-half of the week the
BUN will exceed the time-average urea concentration
(TAC). Therefore, for hemodialysis the KT/V would
have to be higher than that of CAPD to keep the peak
BUN at the same level and achieve equivalent control
of uremia [16]. On the basis of this hypothesis, Keshaviah
calculated that, for a hemodialysis KT/V of 1.3 per treatment, the corresponding weekly CAPD KT/V is 2.0.
(Fig. 1) [15]. Other possible reasons why CAPD patients
appear to need less urea clearance compared to those
on hemodialysis include the lack of changes in hydration
and blood pH, the high clearance of large or middle
molecules, a less catabolic environment, and better preservation of residual renal function (RRF).
The rate of decline in residual renal function appears
slower in PD than in HD [17, 18]. Possible reasons for
this difference are: (1) stable high BUN levels in CAPD;
(2) absence of marked fluid shifts permitting hemodynamic stability and avoidance of glomerular ischemia; (3)
reduced dietary protein intake combined with increased
protein losses in dialysate; (4) absence of cytokine-medi-
1134
Nephrology Forum: Optimization of CAPD
Table 1. Summary of studies investigating KT/V targets for CAPDa
Targets based on
theoretical constructs
Recommended
KT/V target
Popovich et al
Teehan
Keshaviah
2.0
2.25
2.0
Prospective cohort studies/
univariate analysis
Teehan
DeAlvaro
1.83
2.0
Lameire
Brandes
1.85
.2.5
Blake
,1.50
Arkouche
.2.3
Lameire
.2.4
a
Comment
Improved survival
Better survival than
KT/V 1.7
Survival . 5 years
Better survival than
KT/V , 1.5
Increased
mortality
Better clinical outcome
than KT/V , 1.6
Fewer hospitalizations
From Ref. 19
ated responses to extracorporeal exposure to dialysis
membrane; and (5) the continuous volume-expanded
state that often is associated with CAPD [17, 18]. I will
return to the importance of residual renal function in
small solute clearance and nutrition in a few moments.
What level of KT/V (urea) gives an acceptable outcome? Churchill recently reviewed all studies that attempted to determine adequate targets of weekly KT/V
urea [19]. He divided the studies into three groups: those
that base KT/V targets on theoretical constructs, prospective studies using univariate analysis, and prospective studies using multivariate analysis. Table 1 summarizes the results of the theoretical constructs and those
based on cohort studies with univariate analysis. With a
few exceptions, most studies seem to indicate that a
weekly KT/V $ 1.9 provides better survival and a better
clinical outcome.
Four prospective controlled studies have used multivariate analysis to assess the relationship between KT/
V and outcome and to define other independent variables that predict increased risk of death [20–23]. Teehan
et al, who studied 51 patients over a period of five years,
found that decreased serum albumin, older age, longer
time on CAPD, and decreased KT/V were associated
with decreased survival [20]. Genestier and colleagues
found that older age, cardiovascular disease, diabetes,
worse comorbidity score, and lower initial KT/V and
WCC were associated with a worse survival in their study
of 201 patients [21]. Maiorca and coworkers, who studied
68 prevalent patients over three years, determined that
older age, peripheral vascular disease, dyslipidemia, arrhythmia, initial serum albumin , 3.5 g/dl, and weekly
KT/V , 1.7 were detrimental factors [22]. The presence
of residual renal function and a KT/V . 1.98 were associated with significantly better survival.
Although not a randomized controlled trial, the CAN-
USA study provides the best support for a correlation
between small-solute clearance and patient outcome
[23]. This study followed for a minimum of two years
680 patients who started CAPD in 14 centers in Canada
and the US. Analysis of survival showed that, during these
two years, both dialysis dose and non-dialysis-related
factors were related to the likelihood of death. When
the Cox proportional hazards model was used with dialysis dose and nutritional status as time-dependent covariants, the relative risk of death increased with increasing
age, insulin-dependent diabetes mellitus (IDDM), cardiovascular disease, decreasing serum albumin, poor nutritional status, and being dialyzed in the US. Higher
clearance, inclusive of residual renal function, was associated with better patient survival, ability to continue treatment with CAPD, and less frequent hospitalization. The
expected two-year patient survival based on sustained
KT/V and creatinine clearance steadily increased over
the range of values studied; no plateau effect was seen
at the highest doses of KT/V (2.3) or creatinine clearance
(95/liters/week/1.73 m2 BSA); hence no minimum target
could be set. The higher the weekly KT/V urea and creatinine clearance, the better the survival.
One should be cautious when interpreting the CANUSA results; these data are inferences from a statistical
model and are not derived from a prospective controlled
study. Although the study showed a clear association
between dose and outcome, it does not ensure causality.
Furthermore, the CANUSA and other studies that suggest that targets of dialysis adequacy correlate with patient outcome base their conclusions on adequacy values
that include both dialysis and residual renal clearance
[20, 23]. However, in all of them the decline of the renal
component of KT/V (KTr/V) over time is responsible
for most of the variation in total clearance because the
peritoneal clearance remains relatively stable [20, 23–25].
Recent analysis of the CANUSA results indicated that
changes in residual renal function, but not those in peritoneal clearance, were significantly correlated with outcome [26]; in other words, the CANUSA and similar
studies show that, as residual renal function declined,
mortality increased. Today we have no study that indicates whether one can improve outcome by increasing
peritoneal clearance to compensate for declining renal
function. Recently, however, we retrospectively reviewed
115 anuric CAPD patients in our program and correlated
peritoneal KT/V with outcome (Fig. 2; unpublished data).
Those who had a KTp/V of ,1.75 had a significantly
higher mortality rate than did patients whose KTp/V exceeded 1.75 (Fig. 2). In other words, there is a minimum
peritoneal urea clearance (KTp/V of 1.75) below which
mortality is increased.
Not all studies have confirmed the correlation between
KT/V and patient outcome [24, 27]. This variability in
the published results is due partially to differences in
Nephrology Forum: Optimization of CAPD
Fig. 2. Comparison of anuric CAPD patients with KTp/V values above
(j) or below (s) 1.75. (P , 0.05.)
the size of the patient groups, the heterogeneity of the
population, the length of study periods [21], differences
in weight used to calculate V, and the inclusion in the
analysis of the risk factors I have noted, the hazards of
which are not proportional. Furthermore, as I will discuss
later, inclusion in the analysis of “high transporters,”
who though they have high small-solute (especially creatinine) clearances, still have a high mortality rate, makes
the interpretation of the correlations more difficult.
Based on the results of these studies, the Dialysis Outcome Quality Initiative working group [13] and an ad hoc
committee of experts [28] recommended rather arbitrary
minimum clearance targets for CAPD patients, namely,
KT/V . 2.0 and weekly creatinine clearance . 60 liters/
1.73 m2. Until our previously mentioned experience with
anuric CAPD patients is further investigated with prospective controlled studies, it is reasonable to consider
as equivalent the renal and peritoneal clearances and to
compensate for the declining residual renal function by
increasing the peritoneal dose to achieve the recommended target [13]. Are the same target levels appropriate for patients receiving intermittent peritoneal dialysis? If, as postulated by the peak concentration
hypothesis, control of the peak urea nitrogen concentration is critical in preventing uremic toxicity, then intermittent peritoneal dialysis requires a higher dose to
maintain the peak concentration below the levels
achieved in CAPD. The DOQI recommendation for nocturnal intermittent peritoneal dialysis is 2.2 per week for
KT/V and 66 liters/1.73 m2 for WCC [13].
Let’s look for a moment at creatinine clearance targets. Using retrospective comparisons, Boen et al recom-
1135
mended in the late 1970s a weekly creatinine clearance
of 100 liters for hemodialysis and 50 liters for peritoneal
dialysis [29]. Subsequently, Twardowski and colleagues
recommended 40 to 50 liters/1.73 m2 BSA based on their
clinical experience [30]. Arkouche and coworkers found
these levels inadequate and proposed higher levels, 60
liters/week/1.73 m2 BSA [31]. Maiorca et al concluded
that, for a KT/V of 1.96, above which survival was excellent, the equivalent WCC was 58 liters [22]. According to
the CANUSA study, the higher the creatinine clearance
(sum of renal and peritoneal), the better the outcome,
and a clearance of 70 liter/1.73 m2 was associated with
a 78% survival at 2 years, the same as a KT/V of .2.1
[23]. The targets recommended by DOQI for continuous
treatments such as CAPD and CCPD (WCC of 60 liters/
1.73 m2) are slightly lower than the target of WCC . 70
liters, which one would deduce from the CANUSA
study. Selgas et al found that WCC was a much weaker
predictor of outcome than urea clearance [32]. Similarly,
Blake et al noted no relationship between total WCC
and mortality except for WCC values , 48 liters/1.73 m2
[33]. No consensus exists as to whether urea or creatinine
clearance is a better marker of adequacy. In fact, in
about 20% of patients, one finds a distinct dissociation
between these two markers [34].
Discrepancy between KT/V and weekly creatinine
clearance can be due to (1) the different levels of residual
renal function, (2) the difference in the rate of equilibration for urea and creatinine between plasma and the
dialysis solution, and (3) the utilization of different normalizing constant (that is, V for urea and body surface
area for creatinine). Thus in a group of CAPD patients
with some residual renal function, renal KTr/V urea contributed on average only 17% to the total KTrp/V,
whereas the renal creatinine clearance contributed 35%
to the overall creatinine clearance [35]. This difference
was due to the reabsorption of urea and the secretion
of creatinine in the renal tubules; therefore, early in
CAPD, when patients have some residual renal function,
it is easier to achieve creatinine clearance targets than
KT/V urea targets. On the contrary, when patients become anuric, it is easier to achieve KT/V urea targets
because of the slower equilibration of creatinine compared to urea. For the same reason, during APD at night,
when exchanges dwell in the peritoneal cavity over short
periods, urea clearance targets are easier to achieve than
creatinine targets [36]. The use of a different normalizing
constant (that is, V for urea and body surface area for
creatinine) is an additional reason for the discrepancy
between KT/V and WCC. Thus for an obese patient, an
overestimation of V can result in a low KT/V. Body
weight changes have less effect on BSA and therefore
on normalized WCC [37].
1136
Nephrology Forum: Optimization of CAPD
Table 2. Formulas calculating dietary protein intake from urea
levels in CAPD patients
Randerson [43]
Bergström [44]
Kopple [46]
PNA (g/day) 5 8 1 0.15 UA (mM/day)a
PNA (g/day) 5 19 1 0.2134 UA (mM/day)b
PNA (g/day) 5 6.8 1 0.97 UNA (g/day)c
a
PNA, protein equivalent of nitrogen appearance; UA, urea in dialysate and
urine
b
Formula cited here is Dr. Bergström’s revision [45] of the original version [44]
c
UNA, urea nitrogen appearance
Nutrition
The observation that the dose of dialysis is related to
the NPNA (an index of the patient’s protein intake and
appetite) led Lindsay and Spanner to suggest that in
hemodialysis, good nutritional status, as indicated by a
protein equivalent of nitrogen appearance (PNA) . 1
g/day, is the major determinant of low morbidity [38];
however, good nutrition cannot be obtained without adequate dialysis, as indicated by a KT/V . 1.0 per hemodialysis session. One can assess nutrition by clinical evaluation, dietary history, anthropometric measurements, and
various biophysical and biochemical methods. Among
the anthropometric indices, one of the simplest and most
useful is subjective global assessment, which includes
four items: weight loss in the previous six months, anorexia, loss of subcutaneous fat, and assessment of muscle wasting. According to this method, patients are characterized as well nourished, mildly malnourished, or
severely malnourished. This method has been validated
in CAPD patients [39].
The PNA. In addition to providing the KT/V, urea
kinetic modeling can be used to calculate daily protein
intake based on the amount of urea excreted (urea appearance), assuming that the patient is in neutral balance. The protein intake, thus calculated, used to be
called the protein catabolic rate. However, since the actual catabolic rate is 5 to 6 times that of daily protein
intake [40], and we do not know whether the patient is
catabolic, a better term is “protein equivalent of nitrogen
appearance.” Some investigators have found a high correlation between actual dietary protein intake, estimated
by food records, and the PNA [41]; others have noticed
a marked discrepancy between the two [42].
Table 2 shows the various formulas proposed for the
calculation of protein intake for urea appearance. The
Randerson equation, derived from hemodialysis patients, has been validated in CAPD patients [43]; to this
result, one should add the value of estimated protein
losses. Bergström’s formula, based on nitrogen balance
data among CAPD patients, takes into account the
greater non-urea nitrogen losses in peritoneal dialysis
[44] and therefore yields significantly higher values than
the Randerson formula. The formula presented here [45]
differs from that in the original publication, which con-
tained a small error (corrected by Dr. J. Bergström, personal communication). Kopple’s formula, which also is
founded on balance studies, gives values between those
obtained with the Randerson and Bergström formulas
[46]. Although the Randerson formula has been the preferred method to date, I agree with the conclusion of
Mandolfo et al [47] that the Bergström formula, which
is based on balance studies, should become the standard
method.
Because larger patients should have a higher protein
intake than smaller patients, it has been a common practice to adjust the PNA to the patient’s body weight to
create a normalized PNA (NPNA) that can be used for
comparison and correlation purposes. However, this normalization process creates significant problems because
clinicians have not agreed what weight to use for this
calculation. None of the various proposals, or for that
matter the nomalization concept itself, has been clinically
validated [5, 48]. Thus, if the present, actual body weight
is used, normalization creates a mathematical bias in
obese patients, in whom the NPNA is underestimated,
and in the wasted individual in whom NPNA is overestimated. Alternatives that have been suggested are: (1)
using standard body weight (5V/0.58), where V is calculated by the Watson formula, which takes into account
age, gender, height, and weight and (2) desirable body
weight, which is the midpoint of the range associated with
greatest longevity for normal individuals of the same age
range, gender, and skeletal frame as the patient. These
weights are published in the actuarial tables of the Metropolitan Life Insurance Company. Although normalization using these techniques provides higher values of
NPNA for the obese and lower for the wasted, none of
these alternatives has been validated in any published
studies; hence we should test their usefulness in prospective studies. Canaud et al recently proposed normalizing
the PNA to the lean body mass, because muscle mass
constitutes the major reserve of endogenous proteins
and nitrogen fuel within the body, thus establishing a
link between PNA and lean body mass [49].
The NPNA is an important reflection of protein intake
but, measured by various formulas, the correlation coefficient between dietary protein intake and NPNA is about
0.6, which is less than impressive [43, 50]. In CAPD, in
contrast to hemodialysis, in which NPNA is an important
predictor of outcome [14], NPNA frequently correlates
with technique failure but not with other outcomes [20–
24, 48, 50, 51]. The PNA is inversely related to age, being
lower in older patients [52] and in diabetics. In Selgas’
study, patients with NPNA . 1 g/kg/day had a lower
rate of hospitalization compared to those with a value ,
1.0 g/kg/day [32]. Harty et al have questioned the validity
of adjusting PNA [48] in light of the failure of NPNA
to reflect conventional parameters of nutrition in crosssectional analyses of hemodialysis [52] and CAPD pa-
1137
Nephrology Forum: Optimization of CAPD
tients [24]. Their study of 147 CAPD patients showed
no correlation between NPNA and measures of nutrition
or measures of body composition and disclosed no significant differences between well-nourished and malnourished patients. Until a proven method for normalizing PNA is developed and validated, the clinician should
be aware of the risks of normalization as currently applied with regard to obese and wasted patients.
Dialysis dose and nutrition. The precise relationship
between adequacy of dialysis and nutritional status has
remained elusive and is a continuing subject of debate
[53]. Uremia commonly leads to anorexia, which generally is thought to be due to the accumulation of toxic
compounds that inhibit appetite [44, 54]. In elegant experiments in rats, Bergström and colleagues demonstrated
that uremic toxins with a molecular weight between 1.0
and 1.5 kD (extracted from both uremic serum and normal urine) suppress appetite [55]; when these molecules
were injected intraperitoneally into normal rats, their
food intake decreased in a dose-related fashion. The
CANUSA study recently confirmed that anorectic uremic patients who reach end-stage disease commonly regain their appetite after initiation of dialysis [53]. This
observation suggests that uremic toxins that cause anorexia are removed by dialysis. It also is common clinical
experience that inadequately dialyzed patients have a
poor appetite and occasionally a special aversion to meat,
and that sometimes their appetite improves when the
dialysis prescription is appropriately modified [56]. On
the other hand, well-dialyzed patients feel good and eat
well. For years now, but especially for the last 5 to 10
years, investigators have attempted to validate these clinical impressions by establishing a relationship between
dialysis dose and the amount of dietary protein intake
[57]. Establishing such a relationship would be of particular importance because it would unify two factors, dialysis dose and nutritional status, both of which are strong
predictors of outcome. Such a relationship would give
nutrition a central role and thus would have important
therapeutic implications, because one could improve nutrition by increasing dialysis dose.
In 1989 Lindsay and Spanner established the relationship between KT/V and NPNA for hemodialysis and
CAPD patients. They observed that the slope of the
relationship was lower among CAPD than hemodialysis
patients. Among hemodialysis patients dialyzed with a
cellulosic membrane, the mean slope was 0.61, whereas
among those dialyzed with polyacrylonitrile membranes,
the slope was higher (1.08) [38]. Subsequently Lysaght
et al [17] and other investigators [20, 58] confirmed the
positive relationship between NPNA and KT/V in CAPD
patients. Contrary to the findings of Lindsay and Spanner
[38], other investigators found that the slope between
KT/V and NPNA is higher (1.5 to 2.0 3) in CAPD patients
than in hemodialysis patients [17, 58]. This observation
Table 3. Shared components are responsible for the correlation
between KT/V and NPNA (mathematical coupling)a
KT/V
NPNA
Dialysate urea
3 dialysate vol.
Plasma urea
Dialysate urea 3 dialysate vol.
V (Const 3 wt)
V/0.58 (wt)
a
From Ref. 35
suggests that increasing the dialysis dose has a more
salutary effect in CAPD. Bergström et al, who analyzed
pooled data of balanced studies—their own and those
of Blumenkrantz et al [59]—found a good correlation
between NPNA and KT/V in both hemodialysis and
CAPD but a steeper slope in CAPD [44].
The finding of a steeper slope for CAPD than hemodialysis supports the peak concentration hypothesis [15]
and the role of middle molecules as a uremic toxin.
Bergström and Lindholm also suggested that the release
of catabolic cytokines, the development of acidosis between dialysis sessions, and the loss of amino acids during
hemodialysis has an even greater impact on catabolism
than do protein losses in CAPD patients [60]. Furthermore, glucose absorption can have a protein-sparing effect [35, 60], although Bergström and Lindholm could
find no relationship between the amount of glucose absorbed and the protein or energy intake, nor any difference in protein/energy intakes, whether the abdomen
was empty or filled with dialysis solution.
Harty and colleagues challenged the validity of this
relationship [35, 48]. First they reported that, although
they observed a correlation between KT/V and NPNA,
this relationship did not exist between unadjusted PNA
and KT/V. Instead they found only a weak correlation
between unadjusted PNA and unadjusted urea clearance
(KT). Paradoxically, malnourished patients had the highest values for normalized urea clearances (KT/V), producing a significant but inverse correlation between KT/V
and lean body mass [35]. This paradox could be partially
resolved when, instead of normalizing to actual weight,
the clearance was normalized to ideal body weight obtained from the National Health and Nutrition Examination Survey (NHANES) tables; then a weakly positive
correlation emerged between KT/V and some nutritional
parameters. Interestingly, in the latter case, the contribution of residual renal function, not the level of peritoneal
clearance, was responsible for the positive correlation.
These observations led Harty and colleagues to argue
forcefully and convincingly that this relationship is conceptually and mathematically flawed [35]. They pointed
out that, because KT/V and NPNA share three components (Table 3), any relationship between the two must
be due, in part, to mathematical coupling. They supported
this contention further when they showed that randomly
1138
Nephrology Forum: Optimization of CAPD
Fig. 3. Correlation between KT/V and NPNA levels off over a certain
level of KT/V (P , 0.005). (Courtesy of Dr. Peter Blake.)
generated KT/Vs and NPNAs were strongly correlated
with r values between 0.6 and 0.74. In 1996 Uehlinger
showed that this correlation between randomly generated
numbers is apparent only when V shows a big variation;
it is completely absent when V is constant [61]. This suggests that it is not the normalization with a common component per se that introduces the statistical artifact, but
rather the variation of the component. The lack of a strong
relationship between KT/V urea and other formal measures of nutrition contrasts significantly with the strong
correlation between KT/V and NPNA, and supports the
view that, in large part, the correlation (between KT/V
and NPNA) is mathematical. I should emphasize, however, that these findings do not negate the existence
of a physiologic relationship between dialysis dose and
protein intake; Harty et al also accept this possibility
[25]. Mathematical coupling increases the magnitude of
such a correlation. Unfortunately we have no simple
statistical procedure to test whether an observed relationship exceeds the purely statistical association [61].
The following observations support the existence of a
physiologic relationship between NPNA and KT/V: (1)
In 1993 Blake et al asserted that the relationship between
KT/V and NPNA is not linear but flattens out after a
weekly KT/V of 1.83 [62] (Fig. 3). Nolph et al [63] and
Ronco et al [7, 64] subsequently confirmed this finding.
These observations suggest that there is a dialysis dose
above which protein intake is dialysis-dose-independent
and instead will depend on the patient’s dietary choices
[7, 63, 65]. Below this point, however, protein intake
should strongly be correlated with dialysis dose. This
observation probably explains the lack of correlation
that some investigators have observed between NPNA
and KT/V in well-dialyzed patients. (2) Differences exist
between the slopes in hemodialysis and CAPD and the
difference in the slope of patients dialyzed with cellulosic
and those with polyacrylonitrile membranes [38, 44]. (3)
This relationship was confirmed in balance studies when
protein intake was measured. (4) A dissociation in the
relationship occurs when PNA is increased or decreased
by changes in protein intake [66]. (5) KT/V and NPNA
are more closely related in diabetics than in non-diabetics [67]. More than any other evidence, however, longitudinal studies support a physiologic basis for the relationship between NPNA and KT/V. Such studies are not
affected by a spurious correlation, because an increase
in urea generation due to an increase in KT/V would
provide unequivocal evidence of increased production
and removal of urea nitrogen. Under these circumstances, an increase in NPNA in the absence of weight
loss or acute stress could result only from an increase in
protein intake [68]. Of course the strongest support for
the influence of dialysis dose on nutritional status would
be a correlation between changes in dialysis dose and
changes in various nutritional estimates, such as subjective global assessment (SGA) and lean body mass (or
serum albumin).
Table 4 provides a summary of several prospective
longitudinal studies on the effect of KT/V on NPNA and
other nutritional parameters. Lindsay et al, who increased
KT/V from 0.59 to 0.9 among eight CAPD patients for
three months, observed an improvement in NPNA by
0.282 g/kg/day [69]. Heimburger and colleagues increased
weekly KT/V from 1.53 to 1.89 in 19 CAPD patients over
three months and observed an increase in NPNA and
an increase in serum albumin from 2.91 6 0.41 to 3.05 6
0.43 g/dl, which was not statistically significant [70]. After
an increase in weekly KT/V in 17 CAPD patients from
1.44 to 1.78 for one month, Burkart et al observed an
increase in NPNA from 0.616 to 0.695 g/kg/day (P ,
0.05) and in serum albumin from 3.5 (mean of 3 monthly
albumin values prior to change) to 3.69 g/dl (3 monthly
albumin values after change) (P , 0.035) [71]. Uribarri
and coworkers increased weekly KT/V from 1.85 to 2.3
in 23 patients for 6 months and observed an increase in
serum albumin from 3.55 to 3.83 g/dl, but not of NPNA
(0.98 to 1.0 g/kg/day) [72]. By increasing KT/V by 25%
in 40 malnourished CAPD patients, Davies et al halted
or partially reversed their nutritional decline (improvement in lean body mass), but it took at least 6 months
to see the benefit of intervention [73]. Lynn et al increased KT/V among 10 patients from the previously
accepted targets of .1.6 to a new KT/V of 2.81 [74].
They demonstrated significant increases in NPNA levels
of .1.0 g/kg/day in 9 of 10 patients [74]. These patients
reverted to baseline NPNA levels four weeks after returning to their original dialysis prescription. Williams
et al prospectively studied 10 patients after change of
KT/V from 1.37 to 1.68 for three months. Neither NPNA
nor serum albumin changed significantly [75]. Harty et
al increased weekly KT/V from 1.5 to 1.8 in 33 patients
for 3 months but could show no increase in NPNA [76].
Prospectively, Malhotra et al increased the KT/V from
1139
Nephrology Forum: Optimization of CAPD
Table 4. Summary of prospective longitudinal studies of the effect of changes in KT/V on NPNA, serum albumin,
or other nutritional parameters
KT/V
initial/final
Lindsay et al [69]
Heimburger et al [70]
Burkart et al [71]
Uribarri et al [72]
Davies et al [73]
Lynn [74]
Williams [75]
Harty et al [76]
Malhotra et al [77]
a
b
0.59/0.9
1.53/1.89
1.44/1.78
1.85/2.3
D↑25%
.1.6/2.81
1.37/1.68
1.5/1.8
1.4/2.1
Duration
months
N
3
3
1
6
.6
8
19
17
23
40
10
3
3
865
10
33
37
Changes in NPNA or DNPNP
g/kg/day
0.282
0.81/0.87
0.616/0.695a
0.98/1.0
0.86/0.91
Sign. incr. in PNA in 9/10
patients .1.0 g/kg/day
0.88/0.89
No changes
0.74/0.96a
Change in mean serum albumin
before/after
2.91/3.05 g/dl
3.5/3.7 g/dl
3.5/3.8 g/dl
Improvement in LBMb
No changes in mean serum albumin
3.4/3.3 (decreased in 18, increased
in 15, unchanged in 4)
P , 0.05
LBM, lean body mass; PNA, protein equivalent of nitrogen appearance
1.4 to 2.1 in 37 patients [77]. Appetite improved and
NPNA increased from 0.74 to 0.96 g/kg/day, but the mean
serum albumin of the group did not change. In individual
patients, however, the serum albumin decreased in 18 and
increased in 15; those in whom serum albumin increased
were younger and had a higher dietary protein intake
and BUN after the increase in dialysis dose.
Obviously these results are inconclusive. Analysis of
the CANUSA results provides the best evidence of the
effect of urea clearance on nutritional parameters [53].
Thus, after initiation of dialysis, weekly total KT/V increased by 1.5 and weekly total CCr by 35.6 liters/1.73 m2.
During the first six months, there was improvement in
the mean subjective global assessment, lean body mass,
and percentage of lean body mass. The subjective global
assessment increased by 0.72 U (P , 0.001), the lean
body mass by 1.6 kg (P , 0.001), and the percentage
of lean body mass by 1.4% (P 5 0.01). However, no
statistically significant changes were noted in serum albumin, PCR, or normalized PCR. This improvement in
several unrelated markers of nutritional status during
the first six months of CAPD supports the contention
that improved toxin clearance contributes to improved
nutritional status [53]. However, the correlations were
weak, so non-dialysis factors such as comorbidity likely
have an important influence on nutritional status. Serum
albumin levels, although not influenced by dialysis dose,
remained a strong predictor of outcome. This observation suggests that serum albumin level is influenced more
strongly by non-nutritional factors, such as comorbidity
and peritoneal albumin losses.
Malnutrition. Malnutrition has a complex pathogenesis and does not depend only on adequacy of dialysis.
Hence two major types (or causes) of malnutrition frequently occur together in dialysis patients (Fig. 4). The
first, uremic malnutrition, is related to the uremic disorder per se [50]. Another contributing factor in the malnu-
trition of dialysis patients is metabolic acidosis which,
even when slight, can increase the catabolism of branchchain amino acids (valine, leucine, and isoleucine) and
lead to valine depletion that limits protein synthesis. A
portion of this uremic malnutrition would respond to
increases in small-solute clearance and correction of acidosis, but a further portion might be related more to
defective protein anabolism, secondary to dysfunction
of the growth hormone/IGF-1 endocrine axis. This latter
portion might respond better to the administration of
agents such as rHGH or even anabolic steroids than to
increases in dialytic dose [78].
The second major type of malnutrition, which might
be termed comorbid malnutrition, that is, secondary to
diabetes, heart failure, chronic inflammation, malignancy,
etc., would not respond to alterations in dialytic dose or
even protein intake. Infections are important stimulants
of protein catabolism; these, combined with increased
protein losses, which vary between 5 and 15 g/day and
increase significantly with peritonitis, can lead to malnutrition and hypoalbuminemia in CAPD patients [60].
Malnutrition in individual patients can be of one type
or the other or even a mix of these [50].
Fluid/volume control
Low urea clearance and malnutrition/hypoalbuminemia increase the risk of death, while cardiovascular
deaths are the main cause of mortality in CAPD patients.
Therefore, it is logical for us to speculate that low KT/V
and malnutrition increase the risk of death from cardiovascular causes. How might low KT/V values and malnutrition modify cardiovascular risk? Harty et al suggest
that insufficient clearance returns the patient to the predialysis uremic milieu and thus increases the likelihood
of cardiac dysfunction [25]. However, another possible
relationship ties dialysis adequacy targets and cardiovascular mortality: the relationship between fluid overload
1140
Nephrology Forum: Optimization of CAPD
Fig. 4. Various pathogenetic mechanisms of
malnutrition in CAPD patients. (From Ref. 50.)
and hypertension, which leads to atherosclerosis and leftventricular hypertrophy [79]. Thus, volume overload can
directly affect cardiovascular mortality through the development of hypertension and left-ventricular hypertrophy [80], which in turn identify the dialysis patients who
are at risk of sudden death, stroke, and congestive heart
failure [81, 82]. Fluid volume control becomes more difficult when anuria supervenes two to four years after the
initiation of CAPD. By that time, peritoneal ultrafiltration has declined in as many as 30% of CAPD patients,
who then develop fluid retention [83, 84]. Most CAPD
patients suffer from fluid overload, as indicated by central wedge pressure measurements and measurements
of plasma cGMP levels—a marker of the hydration state
[18, 85, 86]. Charra et al have achieved success in managing hypertension through control of extracellular fluid
volume [3]; they believe this approach is superior to
pharmacologic control. Also, they have noted that even
small increases in fluid volume increase the patient’s
resistance to antihypertensive medication.
Patients have been classified into four groups on the
basis of peritoneal transport: low (L), low average (LA),
high average (HA), and high transporters (HI) [87]. The
high transporters (D/Pcreat . 0.80) need special attention from the nephrologist. These patients have significantly lower serum albumin levels and NPNA, and lower
body mass than do the other groups [88]. The low albumin level probably is the result of increased protein
losses across the peritoneum. Also, these patients have
a low NPNA, disproportionate to the level of KT/V,
which might be due to increased use of high-glucose
solutions that, in turn, decrease appetite.
In the CANUSA study, the two-year patient survival
probabilities were 91%, 80%, 72%, and 68% for L, LA,
HA, and HI, respectively [89]. All groups have low serum
albumin and low lean body mass, but we do not know
whether (in these patients) these indices have the same
predictive value as if they were the result of malnutrition.
Harty et al found no evidence of excess malnutrition
[assessed by anthropometric and serum protein values
and subjective global assessment (SGA) categorization,
composite nutritional index] in high transporters [35],
whereas Nolph and colleagues did [88]. Nolph et al recommended that high transporters may warrant increased
clearance and protein intake targets. An alternative explanation for the increased mortality rate among high
transporters is the low ultrafiltration that leads to fluid
overload, which (when chronic) can lead to left-ventricular hypertrophy and cardiovascular disease [80, 90]. We
do not know the best treatment for high transporters,
but volume control likely is important. High transporters
thus are candidates for nocturnal intermittent PD with
short exchanges. This technique leads to better ultrafiltration, slightly reduced protein losses, lower glucose
absorption, and better protein intake [88]. Whether conversion to NIPD will lead to higher survival rates is not
known. In an outstanding recent review, Bergström and
Lindholm described a hypothesis integrating malnutrition, cardiac disease, and mortality [91].
Improving KT/V in underdialyzed patients
The need for an increase in KT/V becomes critical
when the patient becomes anuric; special attention must
be paid to those of large body size, especially patients
who (with respect to peritoneal permeability) are low
transporters. Before one institutes a change in the dialysis prescription, one must ensure that the patient is
compliant with the dialysis instructions. Blake et al noted
that patients ignore 20% to 30% of all prescribed changes
in their dialysis treatment plan, and usually the staff
Nephrology Forum: Optimization of CAPD
cannot detect this [92]. Furthermore, some patients run
the risk of reducing the effective exchange volume when
they flush larger volumes (than they have been instructed) during the flush of the lines, either because of
ignorance or because they are in a hurry to finish the
exchange. Noncompliance that can lead to inadequate
dialysis appears to be more frequent in the US than in
Canada, and this factor might account, at least in part,
for the observed difference in the mortality rate between
the two countries [92].
Of the various ways of enhancing peritoneal dialysis,
increasing the exchange volume is the most critical, least
expensive, and least intrusive into the patient’s life [93].
Schoenfeld et al found that clearance of small molecules
increases linearly with intraperitoneal volumes [94]. Keshaviah et al confirmed this observation and found that
the fill volumes needed to achieve peak clearance increased with increasing body surface area: approximately
2.5 liters for a person of average size and 3.0 to 3.5
liters for an approximate BSA $ 2 m2 [95]. The net
ultrafiltration is increased in response to higher volumes
because of better maintenance of the dialysate-to-plasma
glucose concentration gradient. Concerns about the use
of increased volume include risk of hernia, increased
glucose absorption from the larger intraperitoneal volume, increased costs, and diminished quality of life because of the extra time required for dialysis, which leads
to decreased patient acceptance of the procedure [28].
Finally, urea and creatinine clearances were significantly
higher (up to 1.73) in the supine than the upright position [94, 96].
In the majority of anuric CAPD patients, the “standard” CAPD protocol of four two-liter exchanges per
day is not adequate to achieve the recommended smallsolute targets. Accordingly, the physician should modify
the protocol, taking into account the patient’s size and
peritoneal permeability. Anuric patients who want to
remain on CAPD require a fifth exchange that can be
accomplished in the night with the “night exchanger”
Quantumt (Baxter). In most cases, however, the patient’s
dialysis treatment will have to be converted to automated
peritoneal dialysis, that is, either continuous cyclic peritoneal dialysis (CCPD), with two liters left in the peritoneal cavity during the daytime or with one or two exchanges during the daytime (enhanced CCPD).
Certain principles should govern the prescription of
automated peritoneal dialysis. (1) The target for KT/V
on intermittent treatments should be higher than that
for patients on CAPD. The recent DOQI group recommended a KT/V of 2.2 and creatinine clearance of 66
liters/week/1.73 m2. (2) If the creatinine clearance target
is maintained when one is switching from CAPD to NIPD,
urea clearance targets automatically will be achieved.
(3) The same quantity of the total dialysate delivered
in larger volumes of fewer exchanges produces better
1141
clearances. (4) The longer a patient stays on automated
peritoneal dialysis at night, the better the clearance. (5)
Increasing the number of exchanges over a given period
does not always increase clearances and, in patients with
low peritoneal transport, can even decrease clearances.
(6) Many large patients require two daytime exchanges.
(7) Tidal PD, in which the peritoneal cavity is not drained
completely but a volume of 1.0 to 1.5 liters is left in
the peritoneal cavity between exchanges, theoretically
is better but has been somewhat disappointing in practice
[97]. In trying to optimize dialysis, we should not forget
that whatever measures we recommend should be compatible with costs and the patient’s time availability and
quality of life. As Nolph said, our goal must be to offer
dialysis so that our patients will live comfortably rather
than merely live so they can do dialysis [98].
Although the topic of my discussion has been adequacy
of CAPD, let me briefly comment on the utility of this
concept in patients with severe, chronic renal failure. If
maintaining targets of small-solute clearance and good
nutritional status is critical for the patient on dialysis, it is
reasonable to assume that these same targets, especially
those established for a continuous therapy like CAPD,
are also critical for the patient with declining renal function. In this case dialysis should start when residual renal
function falls below the targets. This view is supported
by the observation that initial KT/V is inversely related
to hospital admission rates, number of hospitalization
days [99], and patient survival [53, 100]. Similarly, nutritional status at the start of dialysis is predictive of longterm outcome [23]. The DOQI group recommends that
patients be advised to initiate some form of dialysis when
the weekly renal KT/V urea falls below 2.0 unless the
patient’s edema-free weight is stable, NPNA is above
0.8 kg/day, and the patient has no clinical symptoms.
One has to balance the benefits of this approach against
the patient’s desire to postpone dialysis as long as possible, and the increased costs and potential risks of dialysis.
Prospective, controlled studies should establish the potential benefits of the early start before it is widely accepted and practiced.
Before I conclude, let me make a comment on the
standardization of guidelines. There is no doubt that
the practice of peritoneal dialysis varies among centers
within a country and also among various countries. In a
recent survey, Ganz et al revealed a significant variation
in dialysis practice and, most important, in various laboratory parameters (that is, hemogloblin, albumin, PTH)
among various centers in British Columbia [101]. The
DOQI guidelines on adequacy of peritoneal dialysis represent the first attempt to develop an approach to decrease discordance among centers and improve patient
outcome. I consider this important effort only the beginning, and I envision regular revisions based on new evidence in the years to come.
1142
Nephrology Forum: Optimization of CAPD
QUESTIONS AND ANSWERS
Dr. Nicolaos E. Madias (Chief, Division of Nephrology, New England Medical Center, Boston, Massachusetts): You addressed the paradox that patients maintained
on peritoneal dialysis have a life expectancy similar to
those maintained on hemodialysis despite markedly lower
KT/V values, and you accounted for this paradox by invoking the peak concentration hypothesis. Although attractive, this is merely a hypothesis and, after all, there
are ample differences between these two modalities.
Could you please offer other potential explanations for
this phenomenon?
Dr. Oreopoulos: Other possible explanations include
the stable hydration and blood chemistry in CAPD patients, the high clearance of large molecules, a better
preservation of residual renal function, and a less catabolic environment. The weight of hemodialysis patients
often fluctuates by 3 to 6 kg between dialyses, while that
of CAPD patients remains constant. Their biochemical
values increase to a maximum before hemodialysis and
decrease dramatically at its end. On the contrary, the
weight and blood chemistry of CAPD patients do not
fluctuate. The CAPD patients definitely have less acidosis than do the hemodialysis patients; serum bicarbonate
levels in the former group range from about 24 to 28
mm/liter. In hemodialysis, the interaction between blood
and the dialysis membrane produces various cytokines
that can be nephrotoxic. Finally, in CAPD, residual renal
function reportedly declines at a much slower rate than
it does in patients treated by hemodialysis. All these
differences taken together might explain the similar mortality rate between the two treatment modalities, despite
lower clearances of small molecules with CAPD.
Dr. Madias: Why does residual renal function decline
at a slower rate in CAPD than it does in hemodialysis?
Dr. Oreopoulos: There are various potential reasons.
In addition to the release of cytokines during the hemodialysis procedure, CAPD patients have a lower protein
intake than do hemodialysis patients. This difference in
protein intake is accentuated further by peritoneal losses
of protein during CAPD. Furthermore, CAPD patients
are in a continuous state of fluid expansion when compared with hemodialysis patients [85, 86]. Finally, CAPD
patients do not experience the decreases in blood pressure that are so frequent during hemodialysis. These
factors taken together might contribute to the better
maintenance of residual renal function in CAPD patients.
Dr. Madias: You did not say much about hypoalbuminemia in peritoneal dialysis patients. Is it as ominous a
predictor of mortality in peritoneal dialysis as it is in
hemodialysis? Is it related to underdialysis?
Dr. Oreopoulos: I spoke chiefly about other nutritional indices of malnutrition and did not discuss albumin
levels because many factors other than malnutrition af-
fect serum albumin levels. There is no doubt that, in
large series, CAPD patients have lower mean serum
albumin levels than do hemodialysis patients. However,
low serum albumin does not always reflect nutritional
status. In an excellent study, Kaysen and Schoenfeld
showed that CAPD patients have reduced catabolism
and increased anabolism and therefore their total albumin mass does not differ from that in controls [102].
According to these authors, the main reason for hypoalbuminemia in CAPD is an increase in the volume of
distribution of the albumin mass. It is almost universally
accepted that, both in hemodialysis and peritoneal dialysis, hypoalbuminemia is a strong predictor of mortality
[20, 33, 40]. However, it seems that the level of albumin
does not have the same effect on outcome in peritoneal
dialysis as in hemodialysis. Thus, whereas in hemodialysis the relative risk (RR) of death was significantly increased when serum albumin was 3.0 to 3.5 g/dl, in CAPD
such an increase in mortality appeared when the albumin
concentration was less than 3.0 g/dl [103]. The strongest
single predictor of low serum albumin is high peritoneal
permeability that leads to low albumin levels either
through the increased protein (albumin) losses or through
overhydration and hemodilution of serum albumin [104].
Recently Han et al [105] and Yeun and Kaysen [106]
showed that serum albumin is strongly influenced by
its role as an inverse acute-phase reactant; both groups
hypothesized that the same patterns of cytokine activation associated with acute-phase reaction also could increase vascular and thus peritoneal permeability. Additional factors that can lead to hypoalbuminemia are
diabetes and cardiovascular disease, which by themselves
lead to increased mortality rates. In patients with these
disorders, hypoalbuminemia is an index of the severity of
the disease and not necessarily of malnutrition. Finally, I
would like to stress that the method of measuring albumin can affect serum albumin levels; the most frequently
used method—the colorimetric test using Bromcresol
green—markedly overestimates albumin levels, whereas
the Bromcresol purple method gives values that slightly
underestimate the true level compared with the “gold
standard,” which is the nephelometry measurement [40].
Because albumin is affected by so many factors, it is not
as good an index of nutrition as are other nutritional
parameters.
Dr. Madias: Would you please comment on the effect
of intraperitoneal amino acids on the level of malnutrition in general, and on serum albumin in particular?
Dr. Oreopoulos: Very early in the history of CAPD,
circa 1978, when we recognized that malnutrition was
an early complication of CAPD, we suggested replacing
glucose as an osmotic agent with amino acids [107]. In
this way we could dialyze the patients with amino acid
solutions and simultaneously replace protein losses. Although this was a great idea, it took the industry 15 years
Nephrology Forum: Optimization of CAPD
to come up with commercial amino-acid-containing solutions. Such solutions are not yet available in the US
but have been used for the last 4 to 5 years in Europe
(Nutrinealt with 1.1% amino acids, Baxter). Dialysis
with one or two daily exchanges of Nutrineal for as long
as four weeks produces a positive nitrogen balance [108].
However, serum albumin increased only minimally. Subsequently, in a large study 71 patients received one to
two exchanges with Nutrineal for three months and were
compared with 63 patients who were dialyzed with standard solutions [109]. Even though the differences between the two groups were not great, a subgroup of
malnourished patients treated with amino acids showed
an increase in serum albumin and improvement in other
nutritional parameters. Thus I believe that the use of
amino-acid-containing solutions has a place in severely
malnourished patients on CAPD.
Dr. John T. Harrington (Dean, Tufts University
School of Medicine, Boston, Massachusetts): Dimitrios,
thank you for a wonderful review. In your comments
about Peter Blake’s study, did you say that patients with
a KT/V less than 1.5 had a good outcome? That finding
is at variance with other observations. Could you tell us
more about your interpretation of Blake’s study?
Dr. Oreopoulos: No, I did not say that, John. Indeed,
I might not have been clear about that point. Blake
found that mortality increases after the KT/V falls below
1.5. He could not find a difference in mortality with high
levels, however. Actually, what he found makes sense; if
you keep decreasing dialysis dose, eventually the patients
will suffer increased morbidity and mortality due to uremia. However, Blake’s cut-off point was lower than that
in other studies. All other studies showed increased mortality if KT/V falls below 1.8 to 2.0. In Blake’s series,
the cut-off point was below 1.5. The difference between
his and other studies might be due to the fact that he
did not measure KT/V but merely calculated it; yet, when
he re-analyzed the data using the measured KT/V, his
original findings remained unchanged. The most likely
explanation is that the number of patients followed was
too small for Blake to detect the benefits of higher clearances on mortality.
Dr. Ronald D. Perrone (Division of Nephrology,
New England Medical Center): You spoke of the mathematical relationship between KT/V and NPNA because
of the shared variables. Have any good studies looked
at the relationship between weekly creatinine clearance,
an independent measure of adequacy, and NPNA?
Dr. Oreopoulos: The only study I can quote is the
CANUSA study [23]. During the first six months, because of the initiation of dialysis, total weekly creatinine
clearances rose by approximately 30 to 40 liters (provided by dialysis), and there was an increase in the subjective global assessment and in lean body mass but not
in serum albumin or NPNA. During the subsequent 12
1143
months, when total weekly creatinine clearance did not
change, the nutritional parameters did not change either.
Dr. Madias: Mathematical bias notwithstanding, you
indicated that in two different populations the relationship between KT/V and NPNA was steeper in CAPD
than in hemodialysis. What might be the basis for this
difference?
Dr. Oreopoulos: This observation supports the peak
concentration hypothesis in that the CAPD patients have
constant biochemical values. An intermittent therapy
requires a higher weekly KT/V to maintain the BUN
peaks at or below the steady BUN of CAPD, assuming
a similar protein intake and urea excretion rates. Thus,
if elevated BUN peaks are associated with appetite suppression (directly or indirectly), a higher weekly KT/V
would be required with an intermittent therapy to yield
a given NPNA. The observation also emphasizes the
importance of toxic substances of middle or large molecular weight that can be removed more effectively by
peritoneal dialysis than by hemodialysis.
Dr. Geetha Narayan (Division of Nephrology, St.
Elizabeth’s Medical Center, Brighton, Massachusetts): I
believe that you said that both the well-nourished and
the malnourished groups had the same NPNA when adjusted for standardized weight. Could you explain that?
Dr. Oreopoulos: Harty et al compared a group of
patients who were well nourished and another group of
malnourished patients (who had lost weight) and found,
as you would expect, that the calculated daily PNA—that
is, an expression of the daily protein intake expressed
in grams per day—was higher in the well-nourished
group than in the malnourished group [48]. However,
when they normalized the PNA with the patients’ present weight (NPNA, that is, g/kg/day), they found the
reverse; that is, the NPNA was higher in the malnourished patients, who had lost body weight, than in the
well-nourished ones. This is a misleading paradox.
Dr. Ajay K. Singh (Director of Clinical Nephrology,
Brigham & Women’s Hospital, Boston, Massachusetts):
A consensus appears to be emerging in the management
of ESRD patients that hemodialysis rather than peritoneal dialysis is the preferable dialytic option for larger
patients. In a recent study by your group addressing this
issue [110], I believe you compared patients weighing 80
kg or more with a group whose body weight was 60 to
80 kg. Could you elaborate on the findings in that study
and tell us whether you agree with this emerging consensus?
Dr. Oreopoulos: In that study, we compared the survival of CAPD patients whose weight was over 80 kg
with those whose weight was less than 60 kg and found
no difference in mortality rate between the two groups
[110]. Unfortunately, we did not have data on their KT/V,
but you would expect, would you not, that because the
daily dialysis volumes were similar in the two groups,
1144
Nephrology Forum: Optimization of CAPD
there probably were differences in their KT/Vs. These
findings are similar to those of Kopple et al, who found
that overweight hemodialysis patients, despite lower percent urea reduction rates, had better survival rates [111].
Heavier patients might have an advantage in that they
are better nourished to begin with, and this might compensate for the fact that they receive relatively less dialysis. Interestingly, in our study, we found a significantly
higher mortality rate in both groups among those patients who had lost weight. This brings us back to the
central role of nutrition. Another explanation for our
findings would be that the patients with the higher body
weights might have had better residual renal function. I
believe that if there is some residual renal function, as
during the first year or two of renal replacement therapy,
you can treat with CAPD almost any individual irrespective of his or her weight. All problems start when the
patient becomes anuric. In our experience, however, the
required KT/V targets need not necessarily be as high
in the anuric patient; we found that, after you provide a
KT/V over 1.75, the survival rate is excellent: 80% the
first 2 years. Is it possible that the anuric patient does
not need a clearance as high as that in the patient with
residual renal function? Confirmation of our observation
with prospective studies is required.
Dr. Singh: There appears to be a striking correlation
between membrane permeability characteristics and mortality. This relationship would suggest that peritoneal
equilibration test (PET) studies should become an integral and early part of the evaluation and management
of virtually every PD patient. Can you tell us about your
practice?
Dr. Oreopoulos: I will tell you what we are doing in
our unit and what are the DOQI recommendations. We
do an Adequestt, which provides measurement of PET,
KT/V, weekly creatinine clearance (WCC), and NPNA
in each of our new patients; we do not do it during
the first two to three weeks of CAPD, because a PET
performed within the first week after initiation of peritoneal dialysis can yield a higher transport result than
would a PET performed a few weeks later [112]. The
DOQI recommends that we obtain baseline measurements of KT/V and creatinine clearance two or three
times during the first six months, and thus obtain a solid
baseline value, and then measure these parameters every
6 to 12 months thereafter. However, all one needs to
measure in the followup is the creatinine and urea in
peritoneal effluent and urine. Repeating the PET is unnecessary because usually it does not change. You should
perform a formal equilibration test only if there is an
unexpected change in either peritoneal creatinine equilibration or ultrafiltration. In our unit, we ask the patients
to bring a 24-hour urine specimen every second visit,
and we analyze this for creatinine and urea clearances
because we know that patients who have adequate resid-
ual renal function have adequate KT/V and weekly creatinine clearance. After all, 1 ml of GFR provides 10 liters
of weekly creatinine clearance and 0.25 of weekly KT/V.
When the creatinine clearance falls below 1 ml/min or
if the 24-hour urine volume drops below 100 ml/day on
at least two successive occasions, we adjust the dialysis
dose and repeat the adequacy test. I want to emphasize
again the importance of following up the residual renal
function.
Dr. Singh: Peter Blake presented some attractive data
on compliance with the peritoneal dialysis prescription
when he visited Boston about a year ago. In light of
the disturbing data on mortality of peritoneal dialysis
patients in the United States, it seems crucial to ensure
compliance. Blake used questionnaires in his study, but
in the individual patient, how do you assess compliance
other than, of course, the “look-them-in-the-eye” test?
Dr. Oreopoulos: You should suspect non-compliance
every time you find an unexplained increase in serum
creatinine in the presence of the same residual creatinine
clearance. In cases like that, you can measure creatinine
in the effluent; if it is 15% greater than the effluent
creatinine value established in the baseline period, you
should suspect non-compliance. It would be likely that
the patient had been skipping dialysis for some time
prior to collection and that on the day of the collection
was washing out the accumulated creatinine. I believe
that the “look-them-in-the-eye” test is a good and reliable measure. The question, “Do you do all dialysis exchanges?” emphasizes the importance of complying with
the prescription. Even if patients lie, at least they get
the message. It is interesting that Blake relied on an
anonymous questionnaire [92]. He assumed that the patients would tell the truth anonymously, but that assumption is open to criticism. Interestingly, he found a greater
degree of non-compliance in the US than in Canada;
this difference might explain some of the observed difference in mortality rates between the two populations [113,
114]. In this area, Bernardini and Piraino and colleagues
provided the gold standard [115]. They went to their
patients’ houses and measured the dialysate inventory.
They compared the actual number of bags with the number that the patient should have on hand if the patient
is compliant with the dialysis regimen. They also found
that a large percentage of their patients were noncompliant. I believe that the companies that deliver the dialysis
solutions might be able to help us in this respect by doing
a monthly inventory.
Dr. Suphamai Bunnapradist (Renal Fellow, Division
of Nephrology, New England Medical Center): Would
you please comment on target KT/V in nocturnal intermittent peritoneal dialysis? Is it different from the target
in CAPD?
Dr. Oreopoulos: If control of the peak urea nitrogen
concentration is critical in preventing uremic toxicity, as
Nephrology Forum: Optimization of CAPD
suggested by the peak concentration hypothesis, then
any intermittent form of peritoneal dialysis requires a
higher “dose” to maintain the peak concentration below
the steady state [14, 15]. The DOQI recommendation for
NIPD is 2.2 for weekly KT/V and 66 liters/week/1.73 m2
for weekly creatinine clearance.
Dr. Madias: What do you recommend when you encounter discordance between the two indices—KT/V and
creatinine clearance?
Dr. Oreopoulos: In responding to this difficult question, I will tell you what the DOQI recommends and what
I am doing in my unit. The DOQI recommends that you
try to achieve both targets. If you achieve only one of
the two, continue with treatment but watch the patient
carefully. In anuric patients and those on short exchanges
with APD, it is easier to achieve KT/V targets than
creatinine clearance targets. On the contrary, in those
with some residual renal function, you can achieve creatinine clearance easier than KT/V. The hemodialysis community always works with urea targets, so I tend to be
satisfied that I am doing a good job if my patients have
an adequate KT/V, and I do not worry too much about
the creatinine clearance targets.
Dr. Madias: You mentioned that the survival of CAPD
patients is similar to that of hemodialysis patients. But
analysis of the USRDS data showed that the mortality
rate was higher among CAPD patients. Also, undergoing
CAPD in the US was a risk factor for mortality in the
CANUSA study. Could you please comment on these
issues?
Dr. Oreopoulos: Nick is referring to the article by
Bloembergen et al [113]. These authors studied a cohort
of CAPD patients during the period of 1987 to 1989 and
indeed found a higher mortality rate among the CAPD
patients than in hemodialysis patients, especially among
elderly diabetics. This study had a rather unusual methodology in that it treated each patient as a new patient
each year and defined their therapeutic modality as the
one in place on January 1st of that year. In other words,
it was a mixture of incident and prevalent patients, some
of whom must have been anuric. Also, the authors did
not compare the relative incidence of co-morbidity. On
the contrary, Fenton from the Canadian Registry studied
all new patients on CAPD and on hemodialysis and found
that all groups of patients—diabetics, non-diabetics, young
and old—had a higher mortality rate with hemodialysis
than with CAPD for at least the first three years of
dialysis [114]. I believe that any method of analysis that
omits, as the Bloembergen study did, the first year of
treatment for many patients will tend to show hemodialysis in a better light than peritoneal dialysis.
Many other reasons account for the difference between the US and Canadian reports. Canadian patients
seem to be more compliant with their treatment than
are patients in the US [92]. Another big difference is that
1145
in Canada approximately 100 new patients per million
population begin dialysis every year; in the United States,
more than 250 new patients per million start dialysis
every year. This difference suggests that the case mix of
the American patients varies from that in Canada. I
believe, however, that a very important reason for the
higher mortality rate among CAPD patients in the US is
that during the years 1987 to 1989, very few nephrologists
were concerned about dialysis dose. All patients had the
“standard” scheme of 4 3 2-liter exchanges per day,
irrespective of their residual renal function and peritoneal permeability. The concepts about dialysis dose
started to enter our regular practice only in the early
nineties. Indeed, when Vonesh and Moran looked at
the USRDS results in subsequent years, they found an
improvement in the results [116]. They also noted that,
at least for male patients, mortality rates for CAPD
during recent years were similar to hemodialysis.
Dr. Harrington: You used the phrase, as does everyone in the field, “adequacy” of CAPD. Rather than aiming for “adequate” CAPD therapy, what alterations
would you envision in CAPD solutions, connections, etc.,
if we were aiming to optimize CAPD?
Dr. Oreopoulos: Even though we should continue to
work hard to achieve adequate clearances for our patients, we need better dialysis solutions, and a tremendous effort has been made in the last few years to develop
them. We have great difficulty in maintaining our peritoneal dialysis patients over long periods. We all have
patients who have been on CAPD for four to five years.
However, only a few patients have been on CAPD for
more than 10 years. Only one of my patients has been
on CAPD for almost 20 years. In any event, these longterm survivors are an exception. I believe this is because
the peritoneal membrane undergoes deleterious alterations when it is exposed to the solutions currently used
over long periods. These changes lead on the one hand
to poor ultrafiltration and on the other to impairment
of the defense mechanisms of the peritoneal cavity, probably because of severe damage to the mesothelial cells.
As a result, peritonitis in patients who have been on
CAPD for five or six years or more is much more severe
than the peritonitis early in CAPD. I dread the threat
of peritonitis in a long-term patient of mine. I am convinced that the long-term exposure of the peritoneal
membrane to unphysiologic, that is, bioincompatible, solutions makes an important contribution to these chronic
changes. The present solutions are acidic (pH 5.5) and
hyperosmotic. The dialysis industry is under tremendous
pressure to come up with new, more “peritoneal
friendly” solutions, and I believe that within the next
few years we are going to see new peritoneal solutions.
In Europe they have developed and use solutions that
contain, instead of glucose as the osmotic agent, a large
glucose polymer (Icodextrin), which achieves ultrafiltra-
1146
Nephrology Forum: Optimization of CAPD
tion by an increased oncotic pressure instead of the increased osmotic pressure caused by dextrose [117]. Over
the long term, such solutions probably will produce ultrafiltration without causing significant peritoneal damage.
Another possible benefit of such solutions is a lower
caloric load and a lower generation of advanced glycosylation end products. Also, new solutions will have a neutral pH and will have bicarbonate or a mixture of bicarbonate and lactate instead of lactate as a buffer. Overall,
I believe that the new solutions will probably enable us
to maintain the peritoneal membrane for longer periods,
but this can only be demonstrated in long-term studies.
Dr. Harrington: Peritoneal dialysis still is used much
less frequently in the United States than is hemodialysis.
What factors do you believe account for its relative neglect by the US nephrology community?
Dr. Oreopoulos: Alan Nissenson has reviewed the
factors that contribute to the wide variation in the application of CAPD worldwide and found that reimbursement policies are critical [118]. However, I understand
that in the US, the compensation is similar for the two
modalities. There must be other reasons. I believe that
peritoneal dialysis, especially now that CAPD is no
longer a simple procedure, poses a challenge to the inexperienced nephrologist, who has to work much harder
to provide exemplary care to the CAPD patient. Complications such as serious peritonitis, fluid control, and malnutrition call for greater commitment to the management of these patients. Distance from the center is also
important. In Canada, where some patients live hundreds of miles away from the treatment centers, it becomes much easier to use CAPD in the patient’s own
home. The patient’s attitude also is critical, and CAPD
provides a good alternative for those who want to be
independent and do not want to come to the hemodialysis facility three times each week. Undoubtedly, CAPD
is easier than hemodialysis for children, especially the
very young. Other groups who may benefit from CAPD
are the elderly, especially those living at home with an
assistant or those living in a nursing home. Also, younger
diabetics may do better on CAPD. In conclusion, many
factors are responsible for the different levels of utilization of CAPD in different centers and in different nations. But I want to emphasize that if a center has only a
few patients on CAPD, the nephrologist and the nursing
staff will never gain enough experience to give the patient the best possible care. Under these circumstances,
the more complications and failures physicians encounter, the less enthusiastic they become in promoting the
advantages of CAPD to new patients. You need a critical
fraction, which I would set at around 20% to 30% of
the total ESRD population of a given center, to achieve
and maintain the required expertise. I believe that peritoneal dialysis is complementary to hemodialysis and trans-
plantation in our efforts to provide our ESRD patients
with the best possible renal replacement treatment.
Reprint requests to Dr. D. G. Oreopoulos, Division of Nephrology,
The Toronto Hospital, Western Division, 399 Bathurst Street, Toronto,
Ontario, Canada M5T 258.
REFERENCES
1. Vas S: Nephrology Forum: Microbiologic aspects of chronic ambulatory peritoneal dialysis. Kidney Int 23:83–92, 1983
2. Canadian Registry: Canadian Organ Replacement Register. Annual Report, 1997
3. Charra B, Calemard E, Ruffet M, Chazot C, Terrat J-C,
Laurent G: Survival as an index of adequacy of dialysis. Kidney
Int 41:1286–1291, 1992
4. Diaz-Buxo JA: Is continuous ambulatory peritoneal dialysis adequate long-term therapy for end-stage renal disease? A critical
assessment. J Am Soc Nephrol 3:1039–1048, 1992
5. Laird NM, Berkey CS, Lowrie EG: Modeling success or failure
of dialysis therapy: The National Cooperative Dialysis Study.
Kidney Int 23(Suppl 13):S101–S106, 1983
6. Gotch F, Sargent JA: A mechanistic analysis of the National
Cooperative Dialysis Study (NCDS). Kidney Int 28:526–534, 1985
7. Ronco C, Conz P, Agostini F, Bosch JP, Lew SQ, La Greca
G: The concept of adequacy in peritoneal dialysis. Perit Dial Int
14(Suppl 3):S93–S98, 1994
8. Chen TW, Huang T-P, Liu M-C, Wang M-L: The removal index
for evaluation of dialysis. Perit Dial Int 16(Suppl 2):128–134, 1996
9. Teehan BP, Schleifer CR, Brown J: Assessment of dialysis
adequacy and nutrition by urea kinetic modeling. Perit Dial Int
14(Suppl 3):S99–S104, 1994
10. Rodby RA, Firan CA, Cheng Y-G, Korbet SM: Reproducibility
of studies of peritoneal dialysis adequacy. Kidney Int 50:267–271,
1996
11. Ilstrup K, Hanson G, Shapiro W, Keshaviah P: Examining the
foundation of urea kinetics. Trans Am Soc Artif Intern Organs
31:164–167, 1985
12. Tzamaloukas AH: In search of the ideal V. Perit Dial Int
16(Suppl 4):345–346, 1996
13. DOQI (Dialysis Outcome Quality Initiative): Clinical Practice
Guidelines. New York, National Kidney Foundation, 1997
14. Blake PG: Is quantitation of dialysis by urea kinetics important
in continuous ambulatory peritoneal dialysis? No. Perit Dial Int
13(Suppl 2):S212–S214, 1993
15. Keshaviah PR, Nolph KD, Van Stone JC: The peak concentration hypothesis: A urea kinetic approach to comparing the adequacy of continuous ambulatory peritoneal dialysis (CAPD) and
hemodialysis. Perit Dial Int 9:257–260, 1989
16. Keshaviah P: Adequacy of CAPD: A quantitative approach.
Kidney Int 42(Suppl 38):S160–S164, 1992
17. Lysaght MJ, Pollock CA, Hallet MD, Ibels LS, Farrell PC:
The relevance of urea kinetic modeling to CAPD. Trans Am Soc
Artif Intern Organs 35:784–790, 1989
18. Lameire NH: The impact of residual renal function on the adequacy of peritoneal dialysis. Nephron 77:13–28, 1997
19. Churchill DN: Outcome studies in adequacy of dialysis. Perit
Dial Int 17(Suppl 3):S42–S45, 1997
20. Teehan BP, Schleifer CR, Brown JM, Sigler CR, Raimondo
J: Urea kinetic analysis and clinical outcome in CAPD: A five
year longitudinal study, in Advances in Peritoneal Dialysis, edited
by Khanna R. Nolph KD, Prowant BF, Twardowski ZJ, Oreopoulos DG, Toronto, Perit Dial Bull 6:181–186, 1990
21. Genestier S, Hedelin G, Schaffer P, Faller B: Prognostic factors in CAPD patients: A retrospective study of a 10 year period.
Nephrol Dial Transplant 10:1905–1911, 1995
22. Maiorca R, Brunori G, Zubani R, Cancarini GC, Manili L,
Camerini C, Movilli E, Pola A, d’Avolio G, Gelatti U: Predictive value of dialysis adequacy and nutritional indices for mortality and morbidity in CAPD and HD patients: A longitudinal
study. Nephrol Dial Transplant 10:2295–2305, 1995
23. Churchill DN, Taylor DW, Keshaviah PR: Adequacy of dialysis
Nephrology Forum: Optimization of CAPD
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
and nutrition in continuous peritoneal dialysis: Association with
clinical outcomes. CANADA-USA (CANUSA) Peritoneal Dialysis Study Group. J Am Soc Nephrol 7(Suppl 2):198–207, 1996
Blake PG, Sombolos K, Abraham G, Weissgarten J, Pemberton
R, Chu GL, Oreopoulos DG: Lack of correlation between urea
kinetic indices and clinical outcomes in CAPD patients. Kidney
Int 39:700–706, 1991
Harty J, Venning M, Gokal R: Does CAPD guarantee adequate
dialysis and nutrition? Nephrol Dial Transplant 9:1721–1723, 1994
Bargman JM, Thorpe K, Churchill DN, for the CANUSA
Peritoneal Dialysis Study Group: The importance of residual
renal function for survival in patients on peritoneal dialysis. (abstract) J Am Soc Nephrol 8:185A, 1997
Pederson JA, Smith MV: Changing empiric (ERx) to kinetic
(KRx) CAPD prescription: No short term effect. (abstract) Kidney Int 37:332, 1990
Blake P, Burkart JM, Churchill DN, Daugirdas J, Depner
T, Hamburger RJ, Hull AR, Korbet SM, Moran J, Nolph KD,
Oreopoulos DG, Schreiber M, Soderbloom R: Recommended
clinical practices for maximizing peritoneal dialysis clearances.
Perit Dial Int 6(Suppl 5):448–456, 1996
Boen S, Haagsma-Schouter W, Birnie R: Longterm peritoneal
dialysis and peritoneal dialysis index. Nephrol Dial Transplant
7:377–378, 1978
Twardowski ZJ, Nolph KD, Prowant BF, Moore HL: Efficiency
of high volume low frequency continuous ambulatory peritoneal
dialysis (CAPD). Trans Am Soc Artif Intern Organs 24:53–57, 1983
Arkouche W, Delawari E, My H, Laville M, Abdullah E,
Traeger J: Quantification of adequacy of peritoneal dialysis. Perit
Dial Int 13(Suppl 2):S215–S218, 1993
Selgas R, Bajo MA, Fernandes-Reyes MJ, Bosque E, Revuelta-Lopez K, Jimenez C, Borrego F, de Alvaro F: An analysis of adequacy of dialysis in a selected population on CAPD for
over 3 years: The influence of urea and creatinine kinetics.
Nephrol Dial Transplant 8:1244–1253, 1993
Blake PG, Balaskas EV, Izatt S, Oreopoulos DG: Is total
creatinine clearance a good predictor of clinical outcomes in continuous ambulatory peritoneal dialysis? Perit Dial Int 12:353–358, 1992
Chen HH, Shetty A, Afthentopoulos IE, Oreopoulos DG:
Discrepancy between weekly Kt/V and weekly creatinine clearance in patients on CAPD, in Advances in Peritoneal Dialysis,
edited by Khanna R. Nolph KD, Prowant BF, Twardowski
ZJ, Oreopoulos DG, Toronto, Perit Dial Bull 11:83–87, 1995
Harty J, Boulton H, Heelis N, Uttley L, Venning M, Gokal R:
Limitations of kinetic models as predictors of nutritional dialysis
adequacy in continous ambulatory peritoneal dialysis patients.
Am J Nephrol 13:454–463, 1993
Nolph KD, Twardowski ZJ, Keshaviah PR: Weekly clearance
of urea and creatinine on CAPD and NIPD. Perit Dial Int 12:298–
303, 1992
Vonesh EF, Burkart J, McMurray SD, Williams PF: Peritoneal
dialysis kinetic modeling: Validation in a multicenter clinical
study. Perit Dial Int 16:471–481, 1996
Lindsay RM, Spanner E: A hypothesis: The protein catabolic
rate is dependent upon the type and amount of treatment in
dialyzed uremic patients. Am J Kidney Dis 13:382–389, 1989
Enia G, Sicuso C, Alati G, Zocalli C: Subjective global assessment of nutrition in dialysis patients. Nephrol Dial Transplant
8:1094–1098, 1993
Heimburger O, Bergström J, Lindholm B: Albumin and amino
acid levels as markers of adequacy in continuous ambulatory
peritoneal dialysis. Perit Dial Int 14(Suppl 3):S123–S132, 1994
Panzetta G, Tessitore N, Faccini G, Maschio G: The catabolic
rate as a measure of protein intake in dialysis patients: Usefulness
limits. Nephrol Dial Transplant 5(Suppl1):125–127, 1990
Bingham SA, Cummings JM: Urine nitrogen as an independent
validatory measure of dietary intake: A study of nitrogen balance
in individuals consuming their normal diet. Am J Clin Nutr 42:1276–
1289, 1985
Keshaviah PR, Nolph KD: Protein catabolic rate calculations in
CAPD patients. ASAIO Trans 37:M400–M402, 1991
Bergström J, Furst P, Alvestrand A, Lindholm B: Protein and
energy intake, nitrogen balance and nitrogen losses in patients
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
63.
64.
65.
66.
1147
with continuous ambulatory peritoneal dialysis. Kidney Int 44:1048–
1057, 1993
Bergström J, Heimbürger O, Lindholm B: Calculation of the
protein equivalent of total nitrogen appearance from urea appearance: which formulas should be used? Perit Dial Int 18:467–473,
1998
Kopple JK, Xui-Lin G, Qing DP: Dietary protein, urea nitrogen
appearance and total nitrogen appearance in chronic renal failure
and CAPD patients. Kidney Int 52:486–494, 1997
Mandolfo S, Zucchi A, Cavalieri D’Oro L, Corradi B, Imbasciati E: Protein nitrogen appearance in CAPD patients: What is
the best formula? Nephrol Dial Transplant 11:1592–1596, 1996
Harty JC, Boulton H, Curwell J, Heelis N, Uttley L, Venning
MC, Gokal R: The normalized protein catabolic rate is a flawed
marker of nutrition in CAPD patients. Kidney Int 45:103–109, 1994
Canaud B, Leblanc M, Garred LJ, Bosc J-Y, Argiles A, Mion
C: Protein catabolic rate over lean body mass ratio: A more
rational approach to normalize the protein catabolic rate in dialysis patients. Am J Kidney Dis 30:672–679, 1997
Blake PG, Oreopoulos DG: Answers to all your questions about
peritoneal urea clearance and nutrition in CAPD patients. Perit
Dial Int 16:248–251, 1996
Lameire NH, Vanholder R, Veyt D, Lambert MC, Ringoir S:
A longitudinal, five year survey of urea kinetic parameters in
CAPD patients. Kidney Int 42:426–432, 1992
Krediet RT, Koomen GCM, Struijk DG, Van Olden RW, Imholz ALT, Boeschoten EW: Practical methods for assessing
dialysis efficiency during peritoneal dialysis. Kidney Int 46(Suppl
48):S7–S13, 1994
McCusker FX, Teehan BP, Thorpe KE, Keshaviah PR,
Churchill DN, for the Canada-USA (CANUSA) Peritoneal
Dialysis Study Group: How much peritoneal dialysis is required
for maintenance of a good nutritional state? Kidney Int 50:(Suppl
56):S56–S61, 1996
Ikizler TA, Greene JH, Wingard RL, Parker RA, Hakim RM:
Spontaneous dietary protein intake during progression of chronic
renal failure. J Am Soc Nephrol 6:1386–1391, 1995
Bergström J, Mamoun H, Anderstam B, Sodersten P: Middle
molecules (MM) isolated from uremic ultrafiltrate (UF) and normal urine induce dose-dependent inhibition of appetite in the
rat. (abstract) J Am Soc Nephrol 5:488, 1994
Bergström J: Nutrition and mortality in hemodialysis. J Am Soc
Nephrol 6:1329–1341, 1995
Gokal R, Harty J: Are there limits for CAPD? Adequacy and
nutritional considerations. Perit Dial Int 16:437–441, 1996
Nolph KD, Moore HL, Prowant B, Meyer M, Twardowski
ZJ, Khanna R, Ponferrada L, Keshaviah P: Cross-sectional
assessment of weekly urea and creatinine clearance and indices
of nutrition in continuous ambulatory peritoneal dialysis patients.
Perit Dial Int 13:178–183, 1993
Blumenkrantz MJ, Kopple JD, Moran JK, Coburn JW: Metabolic balance studies and dietary protein requirements in patients
undergoing continuous ambulatory peritoneal dialysis. Kidney Int
21:849–861, 1982
Bergström J, Lindholm B: Nutrition and adequacy of dialysis:
How do hemodialysis and CAPD compare? Kidney Int 43(Suppl
40):S39–S50, 1993
Uehlinger DE: Another look at the relationship between protein
intake and dialysis dose. J Am Soc Nephrol 7:166–168, 1996
Blake PG, Lindsay RM, Spanner E, Heidenheim P, Baird J,
Allison M, Oreopoulos DG: Factors modifying the relationship
between Kt/V urea and normalized protein catabolic rate (PCRN)
in CAPD. (abstract) J Am Soc Nephrol 4:398, 1993
Nolph ND, Keshaviah P, Emerson P, Van Stone JC, Twardowski ZJ, Khanna R, Moore HL, Collins A, Edwards A: A new
approach to optimizing urea clearance in hemodialysis and continuous ambulatory peritoneal dialysis. ASAIO J 41:M446–M451, 1995
Ronco C: Adequacy of peritoneal dialysis is more than Kt/V.
Nephrol Dial Transpl 12(Suppl 1):68–73, 1997
Panzetta G: Protein intake does not depend on the dose of
dialysis delivered—provided Kt/V is adequate. Nephrol Dial
Transplant 10:2286–2289, 1995
Lindsay RM, Spanner E, Heidenheim P, Kortas C, Blake PG:
1148
67.
68.
69.
70.
71.
72.
73.
74.
75.
76.
77.
78.
79.
80.
81.
82.
83.
84.
85.
86.
87.
88.
89.
Nephrology Forum: Optimization of CAPD
PCR, Kt/V and membrane. Kidney Int 43(Suppl 41):S268–S278,
1993
Bernardini J, Piraino B: Urea kinetics and nitrogen appearance
rates in diabetic patients on peritoneal dialysis. Perit Dial Int 15:259–
263, 1995
Gotch F: A positive correlation of PCRN to Kt/V in cross sectional studies is not a proof of causal relationship. Perit Dial Int
15:274–275, 1995
Lindsay RM, Spanner E, Heidenheim RP, Lefebvre JM, Hodsman A, Baird J, Allison MEM: Which comes first, Kt/V or PCR:
Chicken or egg? Kidney Int 42(Suppl 38):S32–S36, 1992
Heimburger O, Tranaeus A, Bergström J, Lindholm B: The
effect of increased PD on Kt/V, protein catabolic rate (PCR) and
serum albumin. (abstract) Perit Dial Int 12(Suppl 1):S19, 1992
Burkart J, Jordan J, Garchow S, Jones M: Using a computer
kinetic modeling program to prescribe PD. (abstract) Perit Dial
Int 13(Suppl 4):S77, 1993
Uribarri J, Dimaano F, London RD, Dowling J, Marcus RG:
Increase in Kt/V increased serum albumin but not nPCR in a
group of patients on continuous peritoneal dialysis. Perit Dial Int
17:511–513, 1997
Davies SJ, Phillips L, Griffiths A, Naish PF, Russell GI:
Impact of increasing Kt/V in malnourished CAPD patients. (abstract) J Am Soc Nephrol 8:262A, 1997
Lynn RI, Fishbane S, Ginsberg NS: The effect of Kt/V urea on
nitrogen appearance and appetite in peritoneal dialysis. Perit Dial
Int 15(Suppl 5):S50–S52, 1995
Williams P, Jones J, Marriott J: Do increases in dialysis dose
in CAPD patients lead to nutritional improvements? Nephrol
Dial Int 9:1841–1842, 1994
Harty J, Faragher B, Venning M, Gokal R: Urea kinetic modeling exaggerates the relationship between nutrition and dialysis
in CAPD patients. (The hazards of cross-sectional analysis). Perit
Dial Int 15:105–109, 1995
Malhotra D, Tzamaloukas AH, Murata GH, Fox L, Goldman
RS, Avasthi PS: Serum albumin in continous peritoneal dialysis:
Its predictors and relationship to urea clearance. Kidney Int
50:243–249, 1996
Dombros NV, Digenis GE, Soliman G, Oreopoulos DG: Anabolic steroids in the treatment of malnourished CAPD patients:
A retrospective study. Perit Dial Int 14:344–347, 1994
Lameire N, Vanholder RC, Van Loo A, Lambert M-C, Vijt
D, Bockstaele VL, Vogeleere P, Ringoir SM: Cardiovascular
diseases in peritoneal dialysis patients: The size of the problems.
Kidney Int 50(Suppl 56):S28–S36, 1996
Coles GA: Have we underestimated the importance of fluid balance for the survival of PD patients? Perit Dial Int 17:321–326, 1997
Foley RN, Parfrey PS, Harnett JD, Kent GM, Murray DC,
Barre PE: Impact of hypertension on cardiomyopathy, morbidity
and mortality in end-stage renal disease. Kidney Int 49:1379–1385,
1996
Ritz E, Koch M: Morbidity and mortality due to hypertension
in patients with renal failure. Am J Kidney Dis 21(Suppl 2):113–
118, 1993
Heimburger O, Waniewski J, Verynski A, Tranaeus A, Lindholm B: Peritoneal transport in CAPD patients with permanent
loss of ultrafiltration capacity. Kidney Int 38:495–506, 1990
Coles GA, Williams JD: The management of ultrafiltration failure in peritoneal dialysis. Kidney Int 46(Suppl 48):S14–S17, 1994
Rottemburg J: Residual renal function and recovery of renal
function in patients treated by CAPD. Kidney Int 43(Suppl
40):S106–S110, 1993
Lauster F, Heim JM, Drummer C, Fülle H-J, Gerzer R, Schiffe
H: Plasma cGMP level as a marker of the hydration state in renal
replacement therapy. (abstract) Kidney Int 43(Suppl 41):S57, 1993
Twardowski ZJ, Nolph KD, Khanna R, Prowant BF, Ryan
LP, Moore HL, Nielsen MP: Peritoneal equilibration test. Perit
Dial Bull 7:138–147, 1987
Nolph KD, Moore HL, Prowant BF, Twardowski NJ, Khanna
R, Gamboa S, Keshaviah P: Continuous ambulatory peritoneal
dialysis with high flux membrane. ASAIO J 39:904–909, 1993
Churchill DN, Thorpe KE, Nolph KD, Keshaviah PR, Oreopoulos DG, Page D, for the Canada USA (CANUSA) Perito-
90.
91.
92.
93.
94.
95.
96.
97.
98.
99.
100.
101.
102.
103.
104.
105.
106.
107.
108.
109.
neal Dialysis Study Group: Increased peritoneal membrane
transport is associated with decreased patient and technique survival for continuous peritoneal dialysis patients. J Am Soc Nephrol
9:1285–1292, 1998
Blake PG: What is the problem with high transporters? Perit
Dial Int 17:317–320, 1997
Bergström J, Lindholm B: Malnutrition, cardiac disease, and
mortality: An integrated point of view. Am J Kidney Dis 32:
834–841, 1998
Blake PG, Korbet S, Blake R, Bargman J, Burkart J, Dasgupta
M, Delano B, Fine A, Finkelstein F, McCusker F: Admitted
noncompliance (NC) with CAPD exchanges is more common in
US than Canadian patients. (abstract) J Am Soc Nephrol 8:278A,
1997
Diaz-Buxo JA: Enhancement of peritoneal dialysis: The PD plus
concept. Am J Kidney Dis 27:92–98, 1996
Schoenfeld P, Diaz-Buxo J, Keen M, Gotch FA: The effect
of body position (P), surface area (BSA), and intraperitoneal
exchange volume (Vip) on the peritoneal transport constant
(KpA). (abstract) J Am Soc Nephrol 4:416, 1993
Keshaviah P, Emerson PF, Vonesh EF, Brandes JC: Relationship between body size, fill volume, and mass transfer area coefficient in peritoneal dialysis. J Am Soc Nephrol 4:1820–1826, 1994
Brandes JC, Packard WJ, Watters SK, Fritsche C: Optimization of dialysate flow and mass transfer during automated peritoneal dialysis. Am J Kidney Dis 25:603–610, 1995
Piraino B, Bender F, Bernardini J: A comparison of clearances
on tidal peritoneal dialysis and intermittent peritoneal dialysis.
Perit Dial Int 14:145–148, 1994
Nolph DK: Has peritoneal dialysis peaked? The impact of the
CANUSA Study. ASAIO J 42:136–138, 1996
Tattersall JE, Greenwood R, Farrington K: Urea kinetics and
when to commence dialysis. Am J Nephrol 15:283–289, 1995
Tattersall JE, Doyle S, Greenwood RN, Farrington K: Kinetic modelling and underdialysis in CAPD patients. Nephrol
Dial Transplant 8:535–538, 1993
Ganz G, Singh S, Djurdiev O, Levin A: Current dialysis practices
in British Columbia, a cross-sectional study. (abstract) J Am Soc
Nephrol 8:193A, 1997
Kaysen GA, Schoenfeld PY: Albumin homeostasis in patients
undergoing continuous ambulatory peritoneal dialysis. Kidney Int
25:107–114, 1984
Winchester JF: The albumin dilemma. Am J Kidney Dis 20:76–77,
1992
Blake PG, Flowerdew G, Blake RM, Oreopoulos DG: Serum
albuminemia in patients on continuous ambulatory peritoneal
dialysis—predictors and correlations with outcomes. J Am Soc
Nephrol 3:1501–1507, 1993
Han DS, Lee SW, Kang SW, Choi KM, Lee HY, Cho EY, Lee
JH: Factors affecting low values of serum albumin in CAPD
patients, in Advances in Peritoneal Dialysis, edited by Khanna
R. Nolph KD, Prowant BF, Twardowski ZJ, Oreopoulos DG,
Toronto, Perit Dial Publications 12:288–292, 1996
Yeun JY, Kaysen GA: Acute phase proteins and peritoneal dialysate albumin loss are the main determinants of serum albumin
in peritoneal dialysis patients. Am J Kidney Dis 30:923–927, 1997
Oreopoulos DG, Crassweller P, Katirtzoglou A, Ogilvie R,
Zellerman G, Rodela H, Vas SI: Amino acids as an osmotic
agent (instead of glucose) in continuous ambulatory peritoneal
dialysis, in Continuous Ambulatory Peritoneal Dialysis, edited by
Legrain E, Amsterdam, Excerpta Medica, 1979, pp 335–340
Kopple JD, Bernard D, Messana J, Swartz R, Bergström J,
Lindholm B, Lim V, Brunori G, Leiserowitz M, Bier DM,
Stegink LD, Martis L, Boyle CA, Serkes KD, Vonesh E, Jones
MR: Treatment of malnourished CAPD patients with amino acid
based dialysate. Kidney Int 47:1148–1157, 1995
Jones M, Hamburger R, Charytan C, Sandronis S, Bernard
D, Piraino B, Schreiber M, Gehr T, Fein P, Friedlander M,
Burkart J, Ross D, Zimmerman S, Swartz R, Knight T, Kraus
A, McDonald L, Hartnett M, Weaver M, Hagen T, Vonesh
E, Algrim Boyle C, Piscopo D, Martis L, Moran J: Treatment
of malnutrition in peritoneal dialysis (PD) patients with 1.1%
Nephrology Forum: Optimization of CAPD
110.
111.
112.
113.
amino-acid (AA) dialysate solution. (abstract) Perit Dial Int
15:S42, 1995
Afthentopoulos IE, Oreopoulos DG: Is CAPD an effective
treatment for ESRD patients with a weight over 80 kg? Clin
Nephrol 47:389–393, 1997
Kopple JD, Lew NL, Zhu X, Lowrie EG: Body weight predicts
mortality in maintenance hemodialysis (MHD) patients. (abstract) J Am Soc Nephrol 8:198A, 1997
Rocco MV, Jordan JR, Burkart MJ: Changes in peritoneal
transport during the first month of peritoneal dialysis. Perit Dial
Int 15:12–17, 1995
Bloembergen WE, Port KF, Manger A, Wolfe RA: A comparison of mortalities between patients treated with hemodialysis and
peritoneal dialysis. J Am Soc Nephrol 6:177–183, 1995
1149
114. Fenton SSA, Schaubel DE, Desmeules M, Morrison HI, Mao
Y, Copleston P, Jeffrey JR, Kjellstrand CM: A comparison
of adjusted mortality rates. Am J Kidney Dis 30:334–342, 1997
115. Bernardini J, Piraino B: Measuring compliance with prescribed
exchanges in CAPD and CCPD patients. Perit Dial Int 17:338–342,
1997
116. Vonesh E, Moran J: Further comparisons of mortality between
hemodialysis (HD) and peritoneal dialysis (PD). (abstract) Perit
Dial Int 18(Suppl 1):S56, 1998
117. Peers E, Gokal R: Icodextrin: Overview of clinical experience.
Perit Dial Int 17:22–26, 1997
118. Nissenson AR, Prichard SS, Cheng IKP, Gokal R, Kubota M,
Maiorca R, Riella MC, Rottenbourg J, Stewart JH: Nonmedical factors that impact on ESRD modality selection. Kidney
Int 43(Suppl):S120–S127, 1993