1066
Anal. Chem. 1984, 56, 1066-1070
Analysis of Blood Serum for Essential Metals by Simultaneous Multielement Atomic Absorption Spectrometry with Flame Atomization Susan A. Lewis and Thomas C. O’Haver*
University of Maryland, College Park, Maryland 20742 James M. Harnly
Nutrient Composition Laboratory, U.S. Department of Agriculture, Beltsville, Maryland 20705
single dilution, due to this wide concentration range. The SIMAAC system can cover 4 to 6 orders of magnitude of concentration by the use of multiple analytical curves (9). We report the simultaneous analysis of serum for Na, K, Ca, Mg, Fe, Cu, and Zn by using the SIMAAC system with airacetylene flame atomization.
Blood serum Is analyzed for Na, K, Ca, Mg, Fe, Cu, and Zn by using a simultaneous, multielement, atomic absorption spectrometer with a continuum source (SIMAAC). Because of the extended range capabilities of the SIMAAC system, analyses can be carried out on a single 5-fold dilution. Standards and samples are matched In appropriate concentrations for the major, minor, and trace elements. The SIMAAC results show excellent agreement with the certified values for the National Bureau of Standards serum reference material and the manufacturer’s values of a commercially available calibration material, Cation-Cal (American Dade). The relative standard deviations of day-to-day precision studies range from 1.4% for Na to 11.1% for Fe. Good correlations are obtained between the values obtained by SIMAAC and comparative clinical methods.
a
EXPERIMENTAL SECTION Instrumentation. The SIMAAC system has been described
in detail elsewhere (10). It consists of a continuum source, an atomizer (either flame or furnace), a 20-channel Echelle polychromator, which produces a two-dimensional spectral array, a quartz refractor plate for wavelength modulation, and a dedicated minicomputer. Wavelength modulation provides double-beam performance and background correction at all wavelengths. Reagents. Two stock, custom-made, mixed standards (Spex Industries, Metuchen, NJ) containing 6000 mg/L Na, 300 mg/L K, 200 mg/L Ca, and 100 mg/L Mg and 100 mg/L Fe, Cu, and Zn, respectively, were combined to give five standards that bracket the physiological ranges of the metals under study in blood serum. The five standards have base concentrations of: Na (3000 mg/L), K (150 mg/L), Ca (100 mg/L), Mg (50 mg/L), Fe, Cu, and Zn (2 mg/L), with dilution factors of 0.1, 0.4, 0.8,1.2, and 1.6. These standards were then diluted 5-fold with a diluent containing 50 ml/L glycerol (11) (Sigma Chemical Co., St. Louis, MO) and 1 g/L lanthanum (12) as chloride (Fisher Scientific Co., Silver deionized water (Millipore Corp., Bedford, Spring, MD) in 18 MA). Viscosity differences between the standards and samples are minimized by use of a glycerol diluent for the standards. The potential chemical interference of phosphate for calcium and magnesium determinations is reduced by the addition of lanthanum to the sample and standard diluents. The National Bureau of Standards human serum SRM 909 and a commercially available, nonhuman, protein-based calibration reference material, Cation-Cal (American Dade, Miami, FL) were used as quality control materials. Procedure. The NBS SRM 909, Cation-Cal, and serum samples of 0.5 mL were diluted 5-fold with 1 g/L aqueous lanthanum. Samples and diluted standards were analyzed in triplicate with blanks of 1 g/L La and 1 g/L La plus 50 mL/L glycerol, respectively. By use of flame atomization and the SlMAAC system, samples can be analyzed at a rate of 80-90 per hour. Optimum results were obtained at observation heights 1.5 mm 12.5 above the burner head with an air-acetylene flame (air
The analysis of serum for essential metals is of major importance in the clinical laboratory. Evenson (1) has stated that the trend in clinical chemistry, with regard to essential metals, has moved from disease states and toxicology into the area of nutrition, reference values, and health maintenance. Mertz (2) has emphasized the necessity for methods that are specific, precise, and sensitive for essential metals. In his review, Faulkner (3) noted the importance of analytical techniques in the determination of nutritionally essential trace metals. Multicomponent analyses are common in the clinical laboratory and are preferred for their efficient use of time and sample and their cost effectiveness. However, many of the essential metals are determined by single-element techniques. Multielement techniques such as neutron activation analysis and X-ray fluorescence are impractical in a routine clincal laboratory for reasons of time, cost, specimen size, preparation, and handling. The three optical techniques upon which multielement techniques can be based—emission, fluorescence, and absorption—have been compared in several reviews (4-7). Emission is generally considered the most fundamentally suited for multielement analysis; with the introduction of inductively coupled plasmas, emission techniques are entering a renaissance (8). Atomic fluorescence is not widely used for routine analysis; however with the advent of the first commercially available instrument (Baird Corp., Bedford, MA), this technique will probably be more fully utilized. However, atomic absorption spectrometry (AAS) is by far the most widely used and well accepted of these techniques in the clinical laboratory, but it is not generally considered a multielement technique. We describe the use of a prototype simultaneous multielement atomic absorption spectrometer (SIMAAC) to determine the seven most often determined metals in serum. The average concentrations of these metals in serum range from 1 to 3500 mg/L. The use of conventional AAS would not permit the determination of these metals on 0003-2700/84/0356-1066S01.50/0
=
L/min, acetylene
=
2.6
L/min)
(13).
Comparative Methods. The methods used for comparison to SIMAAC were as follows. In the first part of the study we used methods that were currently in use in a local clinical laboratory. Na and K were determined with ion-selective electrodes (ISE) by the Ektachem analyzer (Kodak, Rochester, NY). Colorimetric methods were employed for the determination of Ca with Arsenazo 3 on the Ektachem analyzer, Mg with Calmagite (Pierce Chemical Co., Rockford, IL) on a Cobas centrifugal analyzer (Hoffman La Roche, Nutley, NJ), and Fe with Ferrozine (Harleco, Gibbstown, NJ) on a spectrophotometer (Perkin-Elmer, Norwalk, CT). Zn was determined by air-acetylene AAS against glycerol-based standards (11) using a Perkin-Elmer 303 and Cu by furnace AAS (14) using a Perkin-Elmer 603. ©
1984 American Chemical Society
ANALYTICAL CHEMISTRY, VOL. 56, NO. 7, JUNE 1984
Results of Analysis of Quality Control Materials by SIMAAC (n
Table I.
10)
-
Cation-Cal
SRM 909 %
%
RSD
Na, mmol/L
K, mmol/L
mmol/L
Ca,
Mg, mmol/L Fe, mg/L Cu, mg/L
Zn, mg/L Value given
133 3.47 3.00 1.23 1.74 1.04 1.28
as mean
run
1.5 1.8 1.0 0.9 9.8 2.8 7.8
certified value0
day
day-to- manufactu
c c
c b
1.4 1.6 1.7 1.7 9.0 2.5 7.2
153 5.66 2.91 2.83 1.94 2.05 2.75
Value given for emission only.
c
value
day
run
mean
134.1 ( + 4.6) (-2.8) 3.52 ( + 0.19) (-0.06) 3.02 (+0.17) (-0.06) 1.21 (±0.07)
1.4 2.9 1.8 4.0 11.1 5.7 8.0
(95% confidence limits).
RSD
within-
within- day-tomean
1067
·
156b 6.00' 2.85 2.90 2.02 2.17 2.78
1.4 3.7 1.7 2.1
11.0 4.8 8.4
Not certified.
Table II. Regression Statistics Comparing Results Obtained by the SIMAAC System with Those by Routine Clinical Methods
Deming regression line
analyte
clinical method
n
Na, mmol/L
Ektachem (ISE) Ektachem (ISE) Ektachem (Arsenazo-3) Cobas (Calmagite) colorimetric (Ferrozine) furnace AAS flame AAS (glycerol based stds)
20 20 20 20 20 20 20
K, mmol/L Ca,
mmol/L
Mg, mmol/L Fe, mg/L Cu, mg/L
Zn, mg/L Table III.
= = =
= =
=
r
SE
0.899 0.949 0.962 0.978 0.857 0.959 0.966
1.59 0.18 0.06 0.09 0.31 0.12 0.11
Regression Statistics Comparing Results Obtained by the SIMAAC System with “Reference” Methods
reference method
analyte Na, mmol/L
K, mmol/L Ca,
y y y y y y y
1.02* -1.71 0.22 + 0.92* 0.12 + 0.93* -0.16 + 1.38* -0.11 + 1.31* -0.11 + 0.98* -0.09 + 0.92* +
=
mmol/L
Mg, mmol/L Fe, mg/L Cu, mg/L
Zn, mg/L
flame flame flame flame flame flame flame
n
photometry after 200-fold dilution photometry after 200-fold dilution AAS after 50-fold dilution AAS after 50-fold dilution AAS after deproteinization AAS after deproteinization AAS after deproteinization
In the second part of the study, other well-accepted or reference methods were used. Na and K were determined by flame photometry (Beckman Instruments, Carlsbad, CA), Ca and Mg by flame AAS, after a 50-fold dilution (12), and Fe, Cu, and Zn by flame AAS after deproteinization (12,15), using a Perkin-Elmer
Deming regression line
44 44 44 44 40 46 56
y y y y y y y
=
= = =
= = =
-0.19 -0.08
0.99* 1.00* 0.04 + 0.97* -0.02 + 0.98* -0.04 + 0.99* -0.01 + 1.02* -0.09 + 1.10* + +
r
SE
0.989 0.995 0.975 0.998 0.987 0.992 0.975
1.75 0.10 0.06 0.02 0.10 0.05 0.10
Table IV. Comparison of Performance Limits (95% Confidence Level) max
603.
RESULTS AND DISCUSSION
analyte
ref range
The results obtained for the determination of Na, K, Ca, Mg, Fe, Cu, and Zn in NBS SRM 909 and Cation-Cal, using the SIMAAC system, are shown in Table I, expressed as means, and within-run and day-to-day relative standard deviations (RSD). The means of the NBS SRM 909 obtained by SIMAAC are all within the 95% confidence levels set by NBS. The linear regression equations, correlation coefficients (r), and standard errors (SE) of the mean values obtained by SIMAAC vs. the certified and manufacturer’s values for the NBS SRM 909 and Cation-Cal are y = -0.08 + 0.99* (r = 0.999, SE = 0.53) and y = 0.37 + 0.98* (r = 0.999, SE = 2.05), respectively. The relative standard deviations are low (