A Cellulose Acetate Electrophoretic Procedure

0 downloads 0 Views 771KB Size Report
109-116 (1980). A Cellulose Acetate Electrophoretic Procedure Evaluated for Quantitation of ..... (11), using agarose gel electrophoresis, reported interassay.
of norepinephrine in the centralnervous system. J. Pharmacol. Ther. 163, 147-149 (1968).

Exp.

4. Karoum, F.,Ruthven,C. R. J.,and Sandier, M., Gas chromatographic assay of phenolicalcoholsin biological materialusing electron capture detection. Biochem. Med. 5,505-514 (1971).

5. Dekirmenjian,

H., and Maas, J. W., An improved

3-methoxy-4-hydroxyphenylethylene

liquid chromatography.

Anal. Biochem.

procedure

of

glycoldetermination by gas-

35,113-122 (1970).

6. Wilk, S., Gitlow, S. E., Clark, D. D., and Paley, D. H., Determination of urinary 3-methoxy-4-hydroxyphenylethylene glycol by gasliquid chromatography and electron capture detection. Clin. Chim. Acta 16,403-408 (1967). 7. O’Keefe, R., and Brooksbank, B. W. L.,Determination of 3-methoxy-4-hydroxyphenylethylene glycol, a noradrenaline metabolite,

CLINCHEM27/1,

175-178

in cerebrospinal fluid and urine. Gun. Chem. 19,1031-1039(1973). 8. Karoum, F., Lefevre, H., Bigelow, L. B., and Costa, E., Urinary excretion of 4-hydroxy-3-methoxyphenylglycol and 4-hydroxy-3methoxyphenylethanol in man and rat. Clin. Chim. Acta 43,127-137 (1973).

Langlais, P. J.,McEntee, W. G.,and Bird,E. D.,Rapid liquidchromatographic measurement of 3-methoxy-4-hydroxyphenylglycol and other monoamine metabolites in human cerehrospinai fluid. Gun. Chem. 26,786-788(1980). 10. Krstulovic, A. M., Matzura, C. T., Bertani-Dziedzic, L., et al., Endogenous levels of free and conjugated urinary 3-methoxy-49.

hydroxyphenylethyleneglycol in control subjects and patients with pheochromocytoma determined by reversed-phase liquid chromatography with electrochemical detection. Clin. Chim. Acta 103,

109-116 (1980).

(1981)

A Cellulose Acetate Electrophoretic Procedure Evaluated for Quantitation of High-Density Lipoprotein Cholesterol Ron B. Schifman, Peter F. Brumbaugh, Douglas Grover, Paul R. Finley, and Edward J. Harrow We evaluated the performance of a commercially available Burstein (2) described a precipitation technique involving cellulose acetate electrophoretic method for quantitating divalent cations and polyvalent anions for effectively separating other serum lipoproteins from HDL. This method, with high-density lipoprotein cholesterol (I) in serum by commodifications, has largely replaced more cumbersome ultraparing it to a method involving precipitation with dextran centrifugation methods for HDLC isolation and was the suIfate500/Mg2+. In both methods, enzymic reagents are method used in the Framingham heart epidemiologic study used for cholesterol measurement. For electrophoretic that demonstrated the inverse relation between HDLC and measurement of I the mean intramembrane CV was 4.1 % coronary artery disease (3). (at 220 to 360 mg/L) and the intermembrane CV ranged Several important factors must be considered in quantifrom 12.2 to 21.0% (at 220 to 880 mg/L). Interassay tating HDLC. Its concentration in serum is generally fourfold precision was significantly better for the precipitation lower than the value for total serum cholesterol, so methods method (CV = 3.9% at 390 mgIL). The electrophoretic for measuring HDLC must therefore have good sensitivity and procedure demonstrated significant measurement bias, linearity. In addition, unlike total cholesterol measurements, both at high and low I concentrations. However, low-denmajor changes in interpretation are associated with a narrow sity lipoprotein cholesterol, measured electrophoretically, range of HDLC values, so that analytical precision is critical. correlated wellwithits calculated concentrations obtained Finally, individual laboratories cannot independently deby the precipitation method. Measurements of I by this termine interpretive criteria for HDLC but must rely on electrophoretic procedure did not achieve the accuracy previously reported longitudinal studies. Therefore, good and reproducibility that have been demonstrated for premethod comparison with reference laboratories is essential, cipitation methods and that are necessary for reliable Several types of precipitation methods coupled with enclinical interpretation of resultsforI. zymic and nonenzymic cholesterol quantitation have been The concentration of high-densitylipoproteincholesterol (HDLC) in serum isan independent negative riskfactorfor the development of coronary artery disease and has become an important component of the lipid profile (1).’ As a result, HDLC assays have received increased attention and evaluation. Measurement of HDLC involves two steps: high-density lipoprotein (HDL) is separated from other serum lipoproteins and cholesterol is measured in the isolatedHDL fraction.

Department of Pathology, University of Arizona Medical Center and Veterans Administration Hospital, Tucson, AZ 85710. 1 Nonstandard abbreviations used: HDL, HDLC, high-density Iipoprotein(s), high-density lipoprotein cholesterol; LDL, LDLC, low-density lipoprotein(s), low-density lipoprotein cholesterol; TC, total cholesterol; and TG, triglycerides (triacylglycerols). Received Aug. 7, 1980; accepted Oct. 10,1980.

reported; nearly all show the necessary accuracy and precision needed for meaningful interpretation (4). Recently, a different electrophoretic approach has become available for HDLC separation and quantitation (5). After serum electrophoresis, an enzymic cholesterol reagent is applied to the membrane and HDLC is measured with a scanning densitometer as a percentage of the total cholesterol. This approach has several advantages, including small sample size, direct total cholesterol/HDLC results, and compatibility with other clinical laboratory electrophoretic techniques. However, HDLC quantitation by electrophoresis is a major departure from precipitation methods, and results should be critically compared to those obtained by other generally used methods before this technique is recommended for routine use. We describe here our comparison of two methods for measuring HDLC: a well-standardized precipitation method involving dextran sulfate-500/Mg2 with enzymic cholesterol reagents CLINICALCHEMISTRY, Vol.27,No. 1,1981

175

Table 3. lnterassay Precision of HDLC Measurements for Two Lyophilized Control Pools

Table 1. Summary of Repeated HDLC

Measurements by Electrophoresis a Mean HDLC,

Intramembrane

mg/L

Mean

Range

220 260 290 360

3.4 4.8 3.6 4.3

0-8.4 0-18.8 0.5-10.1 1.4-5.9

n

mg/L

16

130 320

20

Fourpatients sera; n = 8 for each.

a

Prechpltatlon

Electrophoresis Mean Interassay HDLC, CV,

CV, %

Mean HDLC,

n 57 59

16.7 8.0

mg/L 180 390

Interassay CV,

%

6.7 3.9

measuring

the HDLC concentration on four sets of paired samples, using eight membranes (four pairs per membrane) (Table 1). We further evaluated intermembrane precision by measuring the HDLC concentration on five additional patient samples, using 15 membranes (Table 2). Using lyophilized control sera at two different concentrations, we compared the interassay precision of the electrophoretic method with that for the precipitation method (Table 3). The former showed greater measurement variability than the latter (pool I, F-ratio 3.36, p < 0.001; pool II, F-ratio 2.88, p < 0.001). Accuracy. We used 99 serum samples to compare the electrophoretic method with the precipitation method. The TC concentrations of these samples ranged from 910 to 3520 mgfL (mean, 2150 mg/L). The TG concentrations ranged from 420 to 3220 mg/L (mean, 1338 mg/L). The mean HDLC concentration was 445 (SD 131) mg/L by the precipitation method and 476 (SD 209) mg/L by the electrophoretic method. A wider range of HDLC values was obtained by electrophoresis (110-1020 mgIL) than by precipitation (240-810 mg/L). LDLC concentrations as measured by electrophoresis ranged from 660 to 2790 mg/L, with a mean of 1508 (SD 459) mg/L. Calculated LDLC concentrations ranged from 460 to 2710 mg/L with a mean of 1501 (SD 442) mg/L. Figure 1 summarizes the comparison between HDLC results patient

and an electrophoretic method available reagents.

cellulose

acetate

membranes

involving commercially and enzymic cholesterol

Material and Methods HDLC and LDLC by electrophoresis: Serum HDLC and LDLC were separated and quantitated on cellulose acetate membranes by the electrophoretic HDL-cholesterol method available from Helena Laboratories, Beaumont, TX 77704. We performed the assays using the procedure outlined by Helena. Membranes were scanned with the Cliniscan densitometer (Helena Laboratories). This method is described in detail by Cobb et al. (5). HDLC and LDLC by precipitation: HDLC was quantitated by a modification (6) of the method of Kostner (7), in which dextran sulfate-500/Mg2 is the precipitating agent,and cholesterol is quantitated with enzymic reagents. LDLC was calculated by the formula LDLC = TC - (HDLC+ TG/5) (8). None of the samples had the creamy layer indicative of chylomicrons.

Intramembrane precision was determined by using sera from four patients with low concentrations of HDLC (360 mg/L). The samples were paired and analyzed on eight membranes during four days. Intermembrane precision was determined by using sera from five patients, four of whom had low HDLC concentrations (410 mg/L) and one of whom had an HDLC concentration of 880 mg/L. Each sample was analyzed on 15 different membranes within four days of collection. We also evaluated interassay precision, using Ortho normal lyophilized control serum, unassayed (Orthd Diagnostics, Raritan,NJ 08869). This controlserum was diluted twofold in normal salineand designatedas pool I. UndilutedOrtho control serum was designatedas pool II. We did 57 precipitation and 16 electrophoretic HDLC assays on pool I, and 59 precipitation and 20 electrophoretic assays on pool II. Sera from 99 consecutive patients were analyzed for HDLC and LDLC by both electrophoresis and precipitation. The data were analyzed statistically with the aid of a 9845T computer (Hewlett-Packard, Fort Collins, CO 80525).

Results Precision.

We

evaluated

intramembrane

Table 2. Intermembrane

precision

Precision of HDLC

Measurement by Electrophoresisa Mean

a

176

HDLC

CV, %

270 310

18.7 21.0

350

12.2

410 880

16.4 14.3

Five patients sera; n

CLINICAL

Intermembrane

mg/L

=

15 for each.

CHEMISTRY,

Vol. 27. No. 1, 1981

by

obtained by electrophoresis and by preipitation. The slope of the linear regression line is 1.42 (95% confidence interval, 1.27 to 1.56)and they-intercept is-150 mg/L (95% confidence interval, -220 to -90 mg/L). There was a significant difference between means by the paired t-test (t = -2.85,p 0.05).

Discussion Electrophoretic serum lipoproteins,

techniques are widely used to separate which are then detected by staining with

various

dyes,

lipophilic

but

lipoprotein

electrophoresis

is

semiquantitative and may not be as useful in providing an assessment of a patient’s lipid status as measurement of total cholesterol and triglyceride (9). Quantitation of cholesterol in lipoprotein fractions separated by ultracentrifugation or precipitation methods provides more sensitive and reliable diagnostic and prognostic information (10). Recently, serum electrophoresis coupled with lipoprotein cholesterol quantitation by enzymic staining was developed as an alternative to precipitation and ultracentrifugation methods (5). We compared the precision and accuracy of this electrophoretic method with a standard dextran sulfate-500/Mg2 precipitation procedure. The reliability of the precipitation method we used for this

1000

2500

r=.91 77 y=.95x + 76

-J -J

EU)

800

2000

600

I500

*

Cl)

.-

0

_J

00

‘H

F-

400

r=-.8925

oiii

y=l.42x

-j

I

-

154

200

500

a

a a

(U

HDL

a a

a a

(0

CHOLESTEROL

a a 0

a a a

(mg/L)

0

a

a a v

a a a

a a

LDL CHOLESTEROL

(mg/L)

(CALCULATED) Fig. 1. Comparison of serum HDLC and LDLC as measured by cellulose acetate electrophoresis and dextran sulfate500/Mg2+ precipitation assays (PRECIPITATION)

study

Warnick et al. (4) compared six precipitation methods and demonstrated that the sulfate-500/Mg2 method had precision equal to or better than all other precipitation procedures evaluated. Finley et al. (6) demonstrated excellent agreement between results by the dextran sulfate-500/Mg2 and Cooperative Lipoprotein Phenotyping Study methods. The electrophoretic method we used is commercially available and cellulose acetate membranes are used. The overall correlation between the electrophoretic and precipitation procedures demonstrated a significant nonlinearity and bias that was a function of the HDLC concentration. At low HDLC concentrations the electrophoretic method gave consistently lower values than the precipitation method. At high HDLC concentrations electrophoretic values were consistently higher. The linear regression line had a slope and y-intercept significantly differently from 1 and 0, respectively, at the 95% confidence level. In addition, the difference between method results was significant by the paired t -test (t = 2.85 p