EDTA inthec-GMP radioimmunoassay. (NEN) by adding of 0.5, 1.0 ... EDTA per liter; at higher concentrations the apparent c-GMP ... acetic anhydride in aqueous ...
Clearly, . EDTA interferes positively in the c-AMP radioimmunoassay of NEN.
Next we tested the interference of EDTA inthec-GMP radioimmunoassay (NEN) by adding of 0.5, 1.0,2.5,5.0, and 10.0 mmol of EDTA per liter to 12.5 nmol of standard c-GMP per liter. We saw interference up to 2.5 mmol of EDTA per liter; at higher concentrations the apparent c-GMP increased tremendously. We concluded that when plasma cyclic nucleotide concentrations are measured by means of a radioimmunoassay not involving acetylation, the interference of EDTA must be reckoned with. When blood is collected, an EDTA concentration of 6 mmol/L is often recommended as anticoagulant, because it also efficiently prevents degradation of cyclic nucleotides. In the subsequently isolated plasma the EDTA concentration will be about doubled (12 mmol/L). It is well known that plasma normally contains about 15 nmol of c-AMP per liter and about 3 to 4 nmol of c-GMP per
liter. For optimal assay the 100-jzL sample should contain about 1.5 pmol. Thus after dilution (for c-AMP assay) or concentration (for c-GMP assay), EDTA still would be expected to interfere. This may explain the different
values found by several authors. We advise that plasma samples be purified by column chromatography. A less time-consuming alternative, however, is to use a competitive proteinbinding assay or an acetylated radio-
immunoassay. The latter assay involves further dilution of the sample, preventing the unwanted interference. References 1. Gilman, A. G., A protein binding assay for adenosine 3’,5’-cyclic monophosphate.Proc. Nati. Acad. Sci. USA 67, 305-312 (1970). 2. Steiner, A. L., Pagliara,A. S., Chase, L. R., and Kipnis, D. M., Radioimmunoassay for cyclic nucleotides. J. Biol. Chem. 247,
a radioimmunoassay kit in which similar buffer media are used as in the NEN kit, obtained similar results. This interference is, they say, “described in the manual as a precaution.” They also agree with the authors’ conclusions in the last paragraph above. Dr. Goldberg (4), to whom the Letter was also referred, says that “the use of acetylated derivatives is preferable [to column chromatography] for two reasons. First, it is much simpler, easier, and faster than column chromatography. Second, the acetylated derivative is closer in structure to the succinylated derivatives used as a hapten to induce the antibodies used for assay. On general principles it would seem better to use an antigen as close as possible to the original inducing antigen.”
acetic anhydride
in aqueous solution.
by J. Cy-
clic Nucleotide Res. 1, 207-218 (1975). 4. Goldberg, M. L., Radioimmunoassay for adenosine 3’,5’-cyclic monophosphate and guanosine 3’,5’-cyclic monophosphate in human blood, urine and cerebrospinal fluid. Clin. Chem. 23, 576-580 (1977).
Dept. of Pulmonary Disease University Hospital Catharijnesingel 101 3500 CG Utrecht The Netherlands Ed. note: NEN, in some experiments in which they used their standards with
by gallbladder puncture during cholecystectomy, were immediately frozen and stored at -20 #{176}C until tested. The phospholipid concentration in 25 bile samples was determined with a
puncture.
Calibration
To the Editor:
curves prepared with use
Quantitation of phospholipids in bile, along with cholesterol and bile acids, is essential to a “bile lithogenic index” (1) in establishing and monitoring the pharmacological treatment of cholesterol gallstones (2). Cholesterol and bile acids are easily estimated by enzymic assays (3, 4), but determination of phospholipids is not easy. All methods described hitherto are complex and time-consuming, requiring
of various concentrations
preliminary extraction and digestion of phospholipids by strong oxidizing
Within-run
agents. Recently a new enzymic assay has been proposed (5) for serum choline-containing phospholipids. The aim of this study is to evaluate the possibility of its use in phospholipid bile estimation. The principle of the assay is: + 2 H20
Choline-phospholipids phospholipase
D
choline
phosphatic acid
H20 oxidase
betaine + 2 H202 2 H202 +
phenol 4-aminoantipyrine
peroxiclase
Quinone may be estimated by meathe absorbance at 500 nm, and the results for an unknown by use of a calibration curve prepared by using serial dilutions of phospholipid standards. Phospholipase D (EC 3.1.4.4.), choline oxidase (EC 1.1.3.17, from Arthrobacter globiformis), peroxidase (EC 1.11.1.7.), and 4-aminoantipyrine were supplied by suring
clinicalvalues usually encountered. The tabulation below summarizes precision
data obtained by running two different biological samples. Bile
n
Mean SD Phospholipid,
CV,%
g/L
A B
30 30
1.48
0.01 0.67
2.98
0.02
0.67
30 30
1,47 2.96
0.02 0.09
1.36 3.0
Between-day
A B
As can be seen,
the proposed method to the other
precise as compared
commonly used methods. We evaluated interference
+ choline
Choline +202+
of phospho-
lipid standards are linear in a range (1.5 to 7.5 gIL) sufficient to encompass
is fairly
quinone chromophore PieterL. B. Bruynzeel Maartje L. Hamelink Willem van den Bogaard
after stimulation with (1 unit/kg) or directly
each sample in saline five- or 10-fold, respectively, for bile obtained from duodenal intubation and gallbladder
Enzymlc Assay of CholineContaining Phosphoiiplds in Bile
Brooker, G., Femtomole radioimmunoassay for cyclic AMP
and cyclic GMP after 2’-O-acetylation
denal intubation cholecystokinin
Gilford 2400-S spectrophotometer. The test procedure was performed according to Takayama et al. (5) after diluting
1114-1120 (1972). 3. Harper, J. F., and sensitive
latron Laboratories Inc. (Poli Industria Chimica S.p.A., Italy) in kit form (batch no. 0179). Lecithin, sphingomyelin, and lysolecithin, used as standards, were purchased from Sigma Chemical Co., and their purity was checked by thinlayer chromatography (6, 7). Standard solutions (1.5, 3.0, 4.5, 6.0, and 7.5 g/L) were prepared by diluting the abovementioned phospholipids in isotonic saline containing 5 g of Triton X-100 surfactant per liter. Specimens of bile, obtained by duo-
by cho-
lesterol, bile acids, and bilirubin, which are present in a much higher concentration than in serum, by analytical recovery. The results (89.5 to 108,12% recovery, average 97.4%) ensure that there
is no such interference. We compared results obtained by this method (x) and that of Svamborg and Svennerholm (8), previously used by our group. The resulting equation was y = 0.96 x + 1.81 (r = 0.994, p 19.20 (>23.32) 1.40(1.57)22.66(25.42) 88.9 0.96(1.04) 92.0 1.0 (0.83) >19.20 120 (>16) 154 1.41 (0.91) >19.20 (>12.46) Mean cross 2.32% 40.05% reactivi’ ‘ a $ apparent digoxin -
a
DIgoxin8 nmoIIL
Mean dlgoxln b
5. Takayama, M., Itoh, S., Nagasaki, T., and Tanimizu, I., A new enzymatic method for determination of serum choline-containing phospholipids. Clin. Chim. Acta 79, 93
phospholipids
Added
digitoxln
Poisoning
W.,
Bestimmung des Gesamtin Serum. J. Clin. Chem. Clin. 12, 226 (1974).
Table 2. Cross Reactivity of Digltoxln in Digoxin RIA
Table 1. Dlgoxln Error In Plasma of Patients Being Treated with
The (3anvna Coat 1251Digitoxln AlA
-
kit was used
(Clinical Assays,DMslon ofTravenolLaboratories, Inc., Cambridge,MA 02139).b Duplicate assays. ________________________________
ity of the antisera used in the numerous kit procedures for digoxin. Does the digoxin
come from metabo-
lites or only from cross reactivity with digitoxin? To answer this question, we assayed digitoxin standards in the range of the previous concentrations. The digitoxin-free plasma used to prepare the digitoxin standards was also tested and the apparent digoxin (0.35 nmol/L) due to nonspecific binding was subtracted from results for each digitoxin standard. The results (Table 2) show values very similar to those of Table 1, with a mean percentage cross reactivity of 2.55%. This confirms the cross reaction with unchanged digitoxin and a very low rate of digoxin appearance during the metabolism step. However, two other studies report errors caused by anomalous metabolites. Two hospitalized patients for toxic cardiac symptoms under digitoxin medication gave normal values for digitoxin (15 and 6.5 nmol/L), but toxic values for digoxin (4A and 2.75 nmol/L, respectively). These digoxin concentrations are produced from digitoxin by a particular metabolic pattern and are very different from those described by Muller et al., who did not compare their results with the digitoxin standard test to determine the nature of interferences. In any case, we think that it is difficult to estimate the amount of digoxin due to metabolites of digitoxin (0.2 or 0.3 nmol/L) because of the limitation of the main quality criteria for RIA procedures. We conclude that the amount of digitoxin that might bind to the digoxin antibodies in different kit procedures is very variable and depends greatly on the cross reaction of unchanged digitoxin. This discrepancy and the variable cross reactions between methods used to measure the two cardiotonics must be .
To the Editor:
Muller et al. (1) recently reported the cross reactivity of digitoxin in two radioimmunoassay (RIA) kit procedures for digoxin. Both
tests showed
similar
average percentages of cross reactivity with digitoxin (7.2 to 11.9%). They concluded that the amount of cross reaction is directly proportional to the concentration of digitoxin in serum, giving false
therapeutic
values of digoxin (0.82 to
1.90 g/L)
for concentrations of digitoxin of 10 to 25 tg/L. They raised two questions: does the digoxin RIA measure
the digoxin from metabolized digitoxin, and should it be postulated that there is no cross reactivity with the digitoxin medication? The kit we currently use for digoxin, “Phadebas#{174} Digoxin RIA” (Pharmacia Diagnostics AB, Box 17, S-75103 Uppsala 1, Sweden) was not one of those evaluated. We have checked its cross reactivity with digitoxin and also compared some values with an ELISA en670
CLINICAL CHEMISTRY,
zyme immunoassay proposed by Boehringer (Boehringer Mannheim GmbH, Mannheim, West Germany). We assayed all samples in duplicate exactly as instructed by the kit manufacturers. We assayed 21 plasmas, from patients receiving maintenance doses orally and from subjects with massive digitoxin poisoning. Table 1 gives the concentrations of digoxin found in plasmas containing digitoxin, with the percentages of cross reactivity. Our RIA procedure shows lower values than the RIA methods used by Muller et a!. All concentrations are below the therapeutic range or near the low limit, with a mean cross reactivity of 2.32%. The ELISA technique, however, has very important interferences, with a mean cross reactivity of 40.05%, and its use results in values corresponding to toxic concentrations, even for low concentrations of digitoxin. These results confirm that the extent of cross reactivity depends on the qual-
Vol. 26, No. 5, 1980
considered, to avoid clinical interpretation.
great
errors
in
Reference 1. Muller, H., Brauer, H., and Breach, B.,
