procedure employed, which affected ac- curacy (7). This may be important also for digoxin RIAs. The low analytical recovery of digox- in in extracts of standards.
difficulties (6). The differences found may have several other explanations, such as use of impure digoxin in standards. Furthermore, the commercial standards were prepared in a matrix of human serum but, like our routine samples, our standards were in heparinized plasma. The kit manuals do not suggest any effect of this difference. In an earlier
study (6) we found that the
standard curve (bound cpm vs concerttration) for standards prepared in heparinized plasma was steeper than it was when serum was used. In commercial kits for cortisol assay the native matrix used as standard medium in the kits was altered by the preparation procedure employed, which affected accuracy (7). This may be important also for digoxin RIAs. The low analytical recovery of digoxin in extracts of standards after the first 20 days of storage was unexpected. Free fatty acids increase during
storage of serum, even at low temperatures (8). Digoxin and other steroids can be entrapped in micelles of free fatty acids, which may disturb direct RIA (9, 10) and also possibly decrease extraction into dichloromethane. Bacterial contamination of the standard preparations may influence digoxin RIA (11), but we detected no bacterial growth in the affected standards. Reliable standards are a prerequisite for useful monitoring of digitalized patients, as is also the use of the same
therapeutic range in different hospitals. Our results indicate a need for better control by the kit supplier. We suggest that digoxin be calibrated against
the internationally
approved
substance ftilfilling the specifications in the international pharmacopoeia and urge the use of specified standard media resembling as closely as possible the samples collected. Standards should not be used after refrigerated storage for longer than one week.
Clin
Chem Clin Biochem
capacity (3-, 5-, and 10-mL), acidified with 50 mg of sulfosalicylic acid, cen6. Molin L, Bergdahl B, Dahlstr#{246}m G, Letrifuged, and then chromatographed ander E. A method for preparation of repro50 zL of the supernate. Each sample ducible standards in digoxin radiogave the unknown peak. Chromatograimmunoassay (RIA) and influence of choice phy of aqueous saturated solutions of of plasma or serum as the matrix. To be tripotassium or disodium EDTA did published. not demonstrate any ninhydrin-posi7. Lantto 0, Bjorkhem I, Blomstrand R, tive peak, but when these solutions Kallner A. Interlaboratory evaluation of were kept 175 #{176}C for 2 h in a sealed four RIA kitsfor determination of plasma cortisol, with special reference to accuracy: tube the unknown interferent was produced in great quantity, along with Influence of matrix in calibration standards. Clin Chem 26, 1899-1902 (1980). several other unidentified ninhydrinpositive compounds. Some a-amino 8. Forbes AL, Camlin JA. Effects of storage products of the thermal decomposition on serum non-esterfied fatty acid concentrations. Proc Soc Exp Biol Med 102, 709-710 of EDTA and its salts in aqueous solu(1959). tion were previously described (2), and 9. O’Leary TD, Howe LA, Geary TD. Imone of them was also produced during provement in a radioimmunoassay for disterilization by gamma-ray irradiation goxin. Clin Chem 25, 332-334 (1979). of Vacutainer Tubes containing tripo10. Rash JM, Jerkunika I, Sgoutas D. tassium EDTA in aqueous solution. Mechanisms of interference of nonesterifled Probably this product is present in all fatty acids in radioimmunoassays of stethe manufacturer’s sterile tubes that roids. Clin Chim Acta 93, 283-294 (1979). 11. Boone J, Griffin C, Shaw W. Effect of bacterial contamination on apparent digoxin values. Clin Chem 23, 2180 (1977). Lilian Molin. Bj#{246}rn Bergdahl Dept.
of Clin.
Pharrnacol
and Dept. of Intern. Med. Linkoping Univ. Univ. Hospital S-581 85 Linkoping,
1. Kubasik NP, Brody BB, Barold SS.Problems in measurement of serum digoxin by commercially available radioimmunoassay kits. Am J Cardiol 36, 975-977 (1975). 2. Hansell JR. Three-years’ experience in interlaboratorytesting of commercial digoxin kits. Clin Pathol 66, Suppl 1, 234-237
(1976). 3. MUllerH, Graul EH, Koop EA. Unterschiedliche Standardkonzentrationen als Ursache fttr divergierendeSerumdigoxininhestimmungen mit Radioimmunoassays. Fortschr Geb Roentgenstr
Nuclearrned
124,
67-69(1976). 4. Bergdahl B, MolinL, LindwallL, et al. Four
kits
for plasma digoxin radioimmunoassay compared. Clin Chem 25, 305-308 (1979). 5. Lindwall L, Molin L, Bergdahl B. An evaluation method providing confidence intervals applied to radioimmunoassay. J
Sweden
contain tion.
EDTA
salts in aqueous solu-
References 1. Parvy P. Huang Y, Kamoun P. Age related reference values for urinary free amino acids: A simple method of evaluation. J Clin Chem Clin Biochem 17,205-210 (1979). 2. Venezky DL, Moniz WB. Nuclear magnetic resonance study of the thermal decomposition of ethylenedinitrilotetraacetic acid and its salts in aqueous solutions. Anal Chem 41, 11-16 (1969). Philippe Jacqueline Lab. Biochimie
G#{233}n#{233}tique
Groupe Hospitalier
To the Editor: Screening plasma and urine for amino acid abnormalities
R. Parvy I. Bardet
Pierre P. Kamoun
EDTA in Vacutainer Tubes Can Interfere with Plasma Amino Acid Analysis
is an essential
Necker-Enfants Malades 149 rue de S#{232}vres 75743 Paris Ceder 15
France
part of the biochemical investigation of patients suspected of having an inborn error of metabolism.
References
18, 879-883
(1980).
Recently,
we were
involved in the investigation of a child; ion-exchange chromatography of the plasma amino acids showed an unusually increased ninhydrin-positive peak, which eluted like authentic methionine sulfoxide with our citric acidsodium citrate buffer program (1), but with a different 570/440 nm ratio (methionine sulfoxide 5.4, unknown peak 4.7). The pattern for urinary free amino acids showed no abnormalities, and the patient was receiving no medication. On two different days, plasma amino acids of blood samples collected in sterile Vacutainer Tubes (Becton Dickinson) containing tripotassium EDTA solution demonstrated the same abnormality. The possibility of an in-
terfering substance present in this type of tube was then tested as follows. We added 1 mL of distilled water to each of six such Vacutainer
Tubes of different
Analytical Differences In
Measurementof Plasma Catecholamines To the Editor: Measurement
of catecholamines
in
human plasma offers a considerable analytical challenge. In recent years this has been met by an almost exponential growth in publications, most of them representing modifications to the liquid chromatography with electrochemical detection (LCEC) approach (1, 2) or to the radioenzymatic assays (REA) (3, 4). Despite the extensive literature, there have been surprisingly few comparisons between results by these entirely different analytical procedures (2, 5). There is a noticeable variation in catecholamine results re-
CLINICAI CHEMISTRY, Vol. 29, No. 4, 1983
735
