8

THE COMPOSITION OF DIALYSIS FLUID

WILLIAM K. STEWART

Historical 199 Production of bicarbonate-containing dialysis fluid 208

Individual concentrations of electrolytes and dextrose 200 Acetate and chloride 209
Sodium 200 Dextrose 209
Current usage 201 Fluoride 210
Relationship to intra-dialytic upset and osmolal change 201 Contaminants and other substances 210
Limitations of increased sodium dialysis and reduced Particulate 210
osmolal change 202 Chemical 210
Sodium modelling and future trends 203 Bacterial 211
Potassium 203 Other substances 211
Acute renal failure 203 Constitution of dialysis fluid 211
Chronic renal failure 203 Proportionating machines 211
Calcium 204 Temperature control 212
Magnesium 206 De-aeration of dialysis fluid 212
Hard-water syndrome 207 Acknowledgements 212
Acetate and bicarbonate 207 References 213
Relevance to rapid dialysis 208

HISTORICAL gave calcium gluconate intravenously post-dialysis. To

Kolff (1) anticipated most of the problems that have been
encountered in formulating a suitable dialysis fluid for hae-
modialysis. After many trials he advocated a low sodium,
high potassium fluid (Table 1). The major problem was the
high pH of the solution and the presence of bicarbonate
which affected the solubility of calcium salts. Kolff attempt-
ed to adjust pH by bubbling CO 2 into the dialysis fluid. He
finally used tap water containing calcium 1.0 mmolll and
avoid haemolysis, dextrose monohydrate was added.
For early coil-type dialysers immersed in a fixed volume
tank of dialysis fluid (e.g. Kolff-Travenol Tank) it was es-
sential to remove the dialysate and renew the dialysis fluid
every 2 h in order to achieve higher clearances (6). With
these systems each 'bath' of dialysis fluid was adjusted to
near pH 7.4 with lactic acid and bubbled CO 2 , following
which the calcium chloride was added. Any citra ted and
heparinised blood used for 'priming' the extracorporeal cir-

Table 1. Composition of extracellular fluid and of former and contemporary hacmodialysis fluids (references in parenthesis).

Serum Serum Interstitial Earliest dialysis 1960's dialysis 1986 dialysis fluid* *
(2) water (2) fluid (1) fluid (3-5) Extreme range majority
(2)

Sodium mmol/l 142 152.7 145 126.5 ]30-135 109-148 128-140

Potassium mmol/l 4.0 4.3 4.0 5.4 f)-1.5 0-4.0 1.0--2.0
Calcium mmol/l 2.5 1.5 1.5 1.0 1.25 0-1.75 1.5-1.75
Magnesium mmolll 0.8 0.5 0.5 not stated 0.5 O--1.5 0.5-0.85
Chloride mmol/l 101 108.5 114 10 100.5 87-117 95-109
Bicarbonate mmol/l 27B 29.3B 31B 23.9B 3S-40A 0-46.7 A 35-4OA
1961 (3) 1965 (4)
Glucose mmolll 5.0 83-111 or 166* III II 0-30 II
g/dl (l.09 1.5-2.0 or 3.0* 2.0 02 (l--054 O.2X

* = for fluid removal

B = bicarbonate; A = acetate
** = Courtesy of Macarthys Medical Ltd.
X = 78% at 0.2 ± 0.02, 16% at nil.
J. F. Maher (ed.), Replacement of Renal Function by Dialysis
© Kluwer Academic Publishers 1989
200 William K. Stewart
cuit was dialysed against the adjusted bath fluid before being
connected to the patient so as to avoid hypocalcaemic effects
(7).
The concentration of the principal electrolytes in extracel-
lular fluid is given in Table 1 (2). Empirically, the composi-
tion of dialysis fluid should probably be similar to that of
normal interstitial fluid appropriately corrected for the small
protein component of the latter. In practice considerable
variation has been advocated in both cation and anion con-
centrations. The common composition of dialysis fluid used
in the early 1960's is listed (Table 1). The content of dextrose
was reduced in later years, once ultrafiltration by pump-
generated hydrostatic pressure had been shown to be superi-
or to the earlier transmembrane osmolar gradient method of
achieving water shift (3). Concentrations of other compo-
nents have changed considerably since. The wide range of
concentrations currently commercially-available (late
1980s) reflects disparate fluids used for particular clinical
circumstances.
Since a main aim of haemodialysis is to restore physiolog-
ical amounts and concentrations of univalent and divalent
ions in the patient's intracellular and extracellular fluids as a
result of diffusional transfers between dialysis fluid and
blood plasma, the level of individual ions in dialysis fluid
might be set arbitrarily at the mean levels found in plasma
water. In this way sub- or hyper-normal concentrations in
patients would tend to be corrected given time by equili-
bration along the concentration gradient. Accordingly, dial-
ysis fluid sodium and potassium levels might be set at the
respective median values for these cations in plasma water,
while calcium and magnesium could likewise be set at the
median levels for the diffusable fraction in plasma water of
each divalent cation. This simplistic ideal is departed from in
particular circumstances, such as acute renal failure, after
parathyroid surgery and when the degree of urgency of an
individual electrolyte abnormality calls for rapid correction.
Before 1975 the popularity of centralised dialysis fluid
delivery systems, each linked to up to 30 dialysis positions
and using commercially-manufactured dialysis fluid concen-
trate, led to more standardisation of dialysis fluid composi-
tion. A typical example of the dialysis fluid used for all 30
patients in such units might be Na140, K1.5, Ca1.87,
MgO.5 (all mmol/l). Departures from such dialysis fluid
formulations are increasingly adopted now to meet individu-
al patient needs. Individualised prescribing of dialysis con-
tent has become possible now that bedside (single patient)
proportionating pump-monitors are widely available. Using
these single-patient machines, the proportion of diluent wa-
ter to concentrate can be varied within certain limits and,
additionally, selected single components can be added to the
concentrate prior to dilution provided that adequate care is
taken to ensure complete mixing with the concentrate be-
fore dilution and use. In this way the level in dialysis fluid of
sodium, potassium, calcium and magnesium can be individ-
ually increased according to need. Conversely, an increase
of the dilution factor (conventionally 1:35 by volume) can
decrease all concentrations if necessary.
When the stratagem of adding concentrated individual
electrolyte solutions is employed, it is essential that the
content of the solutions so added is carefully calculated,
prepared in advance and distinctively labelled. The act of
addition to the concentrate and mixing must be properly
witnessed and confirmed by formal labelling of the concen-
trate container. Unless these additions are made with estab-
lished formality there is a serious risk of confusion or worse
through unintentional double additions. Since the resultant
composition cannot be easily checked analytically before
use in clinical circumstances, it is vital that all the calcula-
tions, pharmaceutical preparation and labelling are under-
taken with deliberation and in circumstances where double-
checking is possible. These additives, for example high con-
centration potassium salt solution, are probably best stored
in separate locked cupboards.
The role of the final concentration of each electrolyte and
dextrose will be examined in detail.
INDIVIDUAL CONCENTRATIONS OF
ELECTROLYTES AND DEXTROSE
Sodium
In dialysis fluid, as in extracellular fluid, sodium is the major
determinant of osmolality. The early 1970's saw moves away
from the hitherto unquestioned dialysis fluid with a low
sodium concentration relative to plasma and especially to
plasma water (Table 1) (8-10). The use of dialysis fluid with
sodium as low as 115 mmolll and as high as 155 mmoIll (11,
12) has been described, but the extremes are infrequently
used. In 1976 most United Kingdom and continental Eu-
ropean dialysis centres still chose a sodium concentration of
less than 135 mmoIll, with a mean level of 130mmolll (13).
The low sodium concentration in maintenance haemodialy-
sis was carried forward from the earlier requirement (14) for
a sodium gradient of about 20 mmoIlI between dialysis fluid
and plasma water. The rate of removal of sodium from the
patient with early non-convective techniques was deter-
mined by the concentration difference across the dialysis
membrane. Extraction of water in early haemodialysers was
effected by osmotically-induced ultrafiltration, osmolality
of dialysis fluid being increased relative to blood plasma by
using unphysiologically-high concentrations of dextrose
(> 1.5 g/dl) which more than out-weighed the otherwise low
dialysis fluid osmolarity due to low sodium concentrations.
Current haemodialysers are much more robust than earli-
er models and rupture of membranes is uncommon. Such
dialysers are capable of withstanding higher hydrostatic
pressures within the blood compartment. With these phys-
ical changes in dialyser characteristics, ultrafiltration nowa-
days is brought about predominantly by a hydrostatic pres-
sure difference across the dialysis membrane. Ultrafiltration
is capable of inducing net convective shifts of water, sodium
and other micromolecular solutes from plasma to dialysate
even in the absence of those concentrations gradients neces-
sary for passive diffusion-induced transfers. Hydrostatical-
ly-effected movement of water, and accompanying solute by