Nephrol Dial Transplant ( 1997) 12: 184–186
Nephrology
Dialysis
Transplantation
Technical Report
Interference of creatinine measurement in CAPD fluid is dependent on
glucose and creatinine concentrations
T. W. L. Mak1, C. K. Cheung1, C. M. F. Cheung1, C. B. Leung2, C. W. K. Lam1 and K. N. Lai2
Departments of 1Chemical Pathology and 2Medicine, The Chinese University of Hong Kong, Prince of Wales Hospital,
Shatin, Hong Kong
Abstract
Background. High glucose concentration in CAPD
fluid is known to interfere with creatinine measurement, which is required for assessment of peritoneal
membrane permeability and adequacy of dialysis.
Correction formulae have been proposed but they may
be method/analyser-dependent. We studied such interference in detail.
Methods. CAPD fluid was diluted to prepare six specimens with glucose concentrations ranging from 9.1 to
154.4 mmol/l. To each specimen, creatinine standard
was added to give five different concentrations from
50 to 800 mmol/l. The 30 specimens were assayed for
creatinine with six routine clinical chemistry analysers
( Hitachi 911 and 747, Technicon RAXT and SMAC3,
Beckman CX7, and Kodak Ektachem 700). Creatinine
interference was calculated by subtracting the apparent
creatinine concentration with corresponding baseline
creatinine concentration (measured at glucose=
9.1 mmol/l ) in the same series.
Results. At constant creatinine concentration, interference increased with increasing glucose concentration
to varying extents (up to 200%) amongst the six
analysers. At constant glucose concentration, interference decreased with increasing creatinine concentration
in analysers using the alkaline picrate reaction but
increased in the Kodak analyser using enzymatic assay.
Conclusion. Interference of creatinine measurement in
CAPD fluid is dependent on glucose and creatinine
concentrations, and each centre should derive specific
correction formulae for its analytical system.
Key words: CAPD fluid; creatinine measurement; glucose interference
Introduction
Creatinine measurement in biological fluid usually
employs the alkaline-picrate reaction or specific enzymCorrespondence and offprint requests to: Professor C. W. K. Lam,
Department of Chemical Pathology, The Chinese University of
Hong Kong, Prince of Wales Hospital, Shatin, Hong Kong.
atic assay [1]. Both methods are known to suffer from
interference by high glucose concentration in the
sample [1,2]. In physiological concentrations of glucose, this interference is insignificant.
Measurement of creatinine concentration in continuous ambulatory peritoneal dialysis (CAPD) fluid is
required to assess the permeability of the peritoneum
or for calculation of peritoneal creatinine clearance in
assessing the adequacy of dialysis [3]. The very high
glucose concentration of CAPD fluid, as much as
125 mmol/l, poses a difficult analytical problem for
creatinine measurement. Erroneous results can be misinterpreted if the glucose interference has not been
taken into consideration.
Several correction formulae have been proposed for
glucose interference of creatinine measurement in
CAPD fluid. For example, Twardowski et al. [3 ]
suggested the following correction equation:
Corrected creatinine=Apparent creatinine–F
×glucose (F=constant factor)
However, the indiscriminant use of such formula for
all analytical systems has been criticized [1 ] as being
unreliable. As experimental conditions such as temperature, substrate concentration, buffer pH, and mode
of measurement are not identical for different analytical
systems, Farrell and Bailey [1 ] suggested that a specific
formula must be derived for each individual system.
We conducted a detailed study of such interference
at different glucose and creatinine concentrations in
six commonly used analytical systems.
Materials and methods
Creatinine standard was purchased from BDH Chemicals
Ltd, Poole, UK. CAPD fluid containing 4.25% (w/v) dextrose
was obtained from Baxter Healthcare Corporation, Deerfield,
IL, USA.
CAPD fluid was diluted with distilled water to prepare
specimens with six different glucose concentrations of 9.1,
18.3, 35.5, 67.6, 105.6 and 154.4 mmol/l. For each series of
glucose concentration, creatinine standard was added to form
a range of five different creatinine concentrations of 50, 100,
© 1997 European Renal Association–European Dialysis and Transplant Association
Interference of creatinine measurement in CAPD fluid by glucose
200, 400, and 800 mmol/l. Thirty specimens with different
combinations of glucose and creatinine concentrations were
thus prepared.
The specimens were assayed in duplicate for creatinine
with six commonly used routine clinical chemistry analysers:
Hitachi 911 and 747 (Boehringer Mannheim, Mannheim,
Germany), Technicon RAXT and SMAC3 (Technicon,
Tarrytown, NY, USA), Beckman CX7 (Beckman Instrument
Corp. Brea, CA, USA), and a Kodak Ektachem-700
( Eastman Kodak Co. Rochester, NY, USA) The first five
systems employed the alkaline-picrate reaction. The Kodak
analyser used creatinine specific enzymes creatinine aminohydrolase and creatine aminohydrolase. Interassay coefficients
of variation (CV ) of these methods were <5% at creatinine
concentrations ranging from 100 to 153 mmol/l/
For each series of specimens with the same creatinine but
different glucose concentrations, the creatinine concentration
measured in the sample containing 9.1 mmol/l of glucose
was taken as the baseline creatinine concentration. Glucose
at this concentration or below did not cause significant
interference in all the analytical systems being studied. The
creatinine interference was calculated by subtracting the
apparent creatinine concentration from the corresponding
baseline creatinine concentration in the same series.
Glucose was measured with a hexokinase method on a
Hitachi 911 analyser using commercial reagents (Boehringer
Mannhein, Mannhein, Germany). Interassay CV were 2.1
and 2.9% at glucose concentrations of 2.1 and 10.1 mmol/l,
respectively.
Results
Interference of creatinine measurement varied with
both glucose and creatinine concentrations to different
extents in the different analytical systems. These variations are shown in Figure 1a–f.
It was observed that creatinine interference increased
with increasing glucose concentration in all six analytical systems. For a given creatinine concentration, the
relationship between creatinine interference and glucose concentration was approximately linear at glucose
concentrations above 18.1 mmol/l.
It was also observed that the slopes of the regression
lines were affected by the concentration of creatinine
present in the specimens. In systems using the alkalinepicrate reaction, interferences by the same concentration of glucose were smaller for specimens with higher
creatinine concentrations. This phenomenon was most
apparent in the Hitachi 747 analyser. The effect in the
Kodak Ektachem analyser was in the opposite direction such that interference increased with increasing
creatinine concentration. For illustrative purpose, the
185
Fig. 1a–f. Relationship of creatinine interference and glucose concentration in six analytical systems. The baseline creatinine concentrations were 50 mmol/l ($), 100 mmol/l (&), 200 mmol/l (+),
400 mmol/l (,), and 800 mmol/l (2).
multiple correction formulae for glucose interference
at different creatinine concentrations in the Kodak
Ektachem are tabulated in Table 1.
Discussion
The alkaline-picrate reaction for creatinine assay is
one of the oldest analytical method still in common
use nowadays. It is inexpensive and reliable in most
clinical situations. The principle involves the formation
of a chromogen by creatinine with picrate in an alkaline
medium, and the end-product is measured spectrophotometrically. When adapting this reaction to routine
analytical systems, many minor variations in reaction
condition have been made. These include the buffer
pH, concentration of picrate, time of measurement,
end-point or kinetic mode, and time of reaction. These
variations do not give rise to significant differences in
creatinine measurement in most clinical specimens.
Table 1. Formulae for correcting creatinine interference by glucose on the Kodak Ektachem-700 analyser
Measured creatinine (mmol/l )
<100
100–200
200–400
400–800
800–1000
Intercept
Slope
−1.9
0.048
−1.1
0.098
−2.2
0.26
−4.7
0.58
−6.9
0.72
Creatinine interference=[Glucose concentration]×Slope+Intercept.
Corrected creatinine concentration=[Measured creatinine concentration]– [Creatinine interference].
186
Although creatinine measurement has been known
to be affected by high glucose concentration for a long
time [2], this is not usually problematical except with
severe diabetic complications. However, when measuring creatinine in CAPD fluid, most if not all analytical
systems have been incapable of producing accurate
results without correction. If the interference was
dependent on glucose concentration alone, like Farrell
and Bailey [1 ] had suggested, a single correction formula could be used. In this study we demonstrated
that interference is dependent both on glucose and
creatinine concentrations. Hence, multiple correction
formulae, each applicable for a range of creatinine
concentrations, are required to make appropriate
adjustments.
We believe that the glucose interference is due to the
formation of an interfering chromogen by glucose with
picrate. This chromogen might have a much smaller
absorption in the measuring wavelength. At physiological concentrations of glucose, this effect is insignificant.
In specimens with very high glucose concentration, like
the CAPD fluid, this effect becomes important. As
reaction conditions differ among different analytical
systems, the magnitude of this interference varies.
Our observation of the creatinine dependent phenomenon is more difficult to explain. We postulate
that this might be due to the competition of creatinine
and glucose for the limited amount of picrate. For any
given concentrations of glucose and picrate, the higher
the creatinine concentration, the more target chromogen will be formed. This would reduce the formation
of glucose-picrate interfering chromogen.
Positive interference by glucose was also observed
with the enzymatic creatinine method on the Kodak
Ektachem analyser. The pattern of interference was,
T. W. L. Mak et al.
however, quite different from the alkaline-picrate
method. The magnitude of interference was actually
higher at higher creatinine concentrations. This is in
contrast to what was observed in the picrate method,
but is in agreement with the positive interference
reported for a similar enzymatic method [2 ]. The cause
of this interference is uncertain but it may be related
to the detection system and/or the enzyme [4 ].
It is important to adjust for the glucose interference
of creatinine measurement in CAPD fluid. Without
such correction, other derived indices would also be
inaccurate and misleading.
As the pattern of interference varies with the method
and also the instruments used, a common correction
formula for all the systems evaluated cannot be drawn.
It is thus recommended that each centre performing
creatinine assay should conduct our simple experiment
to derive specific correction formulae for the
interference.
Acknowledgements. We thank Dr Anthony Shek and Miss Judy Lai
for their assistance in measuring some of the specimens with their
analysers.
References
1. Farrell SC, Bailey MP. Measurement of creatinine in peritoneal
dialysis fluid. Ann Clin Biochem 1991; 28: 624–625
2. Boyne P, Beilby J, Hickman P, Knight J, Williams D. Scientific
and Technical Committee: Technical report No. 27 Creatinine
Interferences Working Party. Clin Biochem Rev 1988; 9: 54–57
3. Twardowski ZJ, Nolph KD, Khanna R et al. Peritoneal equilibration test. Perit Dial Bull 1987; 7: 138–147
4. Gerard S, Khayam-Bashi H. Negative interference with the
Ektachem ( Kodak) enzymic assay for creatinine by high serum
glucose. Clin Chem 1984; 30: 1884
Received for publication: 29.4.96
Accepted in revised form: 10.9.96