Serum Cholesterol Methodology: 100 Years of Development*

Transcription

1 ANNALS OF CLINICAL AND LABORATORY SCIENCE, Vol. 20, No. 1 Copyright 1990, Institute for Clinical Science, Inc. Serum Cholesterol Methodology: 100 Years of Development* BRADLEY E. COPELAND, M.D. Department of Pathology and Laboratory Medicine, University o f Cincinnati College o f Medicine, and Laboratory Service, Medical Center, U.S. Department o f Veterans Affairs, Cincinnati, OH ABSTRACT Serum cholesterol is the third nationally recom m ended health screening test. It is the first test which depends upon the chemical analysis of a serum compound. Over the past 100 years, the methods of m easurem ent and methods of standardizing these measurements have improved owing to the combined efforts of clinical research, governmental and non-governm ental actions, and commercial producers. The future constancy of serum cholesterol m easurem ent will be based on: (1) National Bureau of Standards (NBS) pure cholesterol standard,21 (2) NBS D efinitive Isotope D ilution Mass Spectrography M ethod,5 (3) C e n te rs for D isease C ontrol (CDC) L ipid L aboratory R eference M ethod,1011 and (4) CD C National Reference System for Serum Cholesterol.4 Introduction In 1885, L ieberm an16 reported the green-blue color produced by the reaction of cholesterol with concentrated sulfuric acid and acetic anhydride. Burc h a r d 3 fo llo w e d in w ith a modification to increase the color developed by dissolving the cholesterol in chloroform; hence, the Lieberman-Burchard color reagent. The m odern m easurem ent of serum c h o le s te ro l was in flu e n c e d by th e * Address reprint requests to: Bradley E. Copeland, M.D., Department of Pathology and Laboratory Medicine, University of Cincinnati College of Medicine, 231 Bethesda Avenue, Cincinnati, OH research of Bloor2 at Harvard in 1915 to 1922 a n d h as in v o lv e d illu s tr io u s m em bers of the chemical pathology discipline over th e past 70 years. The im portance of identifying hypercholesterolem ia in patients was recognized as a signal of out-of-control diabetes mellitus. In turn, this led to the postulation of a relationship betw een serum cholesterol and atherosclerosis as morphologic, and chem ical analyses o f a th e ro m a to u s plaques in the aorta and coronary vessels revealed a high concentration of cholesterol. Long-term studies dating back to the 1930s began to provide an epidemiologic background for the use of serum cholesterol as a p redictor of coronary artery /90/ $01.80 Institute for Clinical Science, Inc.

2 2 COPELAND thrombosis. The first conclusive study which had a sufficient n u m b er of subjects carefully followed over first a sixand then a 20-year period was the Fram ingham study (1951) conducted by the now ren o w n ed staff of th e H arvard School of Public H ealth9,1314 on the citizens of an affluent Boston suburb, Fram ingham, Massachusetts. In addition to a yearly physical examination, chest x-ray, blood pressure and serum cholesterol w ere m onitored. Having a stable population with a high degree of follow-up from year to year, the prediction of coronary throm bosis in this population by e le vated serum cholesterol was indisputable, and risk categories w ere e sta b lished for clinical application. Since 1961, when the main Framingham data were reported by Kannel, Dawber, et al.,14 repeated trials have confirmed the original finding. Several international trials have confirmed this observation as well. By 1984, there was sufficient evidence to establish the fact that modification of elevated serum cholesterol could influence the m orbidity and mortality from c o ro n a ry th ro m b o s is. T his led th e National Institutes of Health Consensus Development Conference on Lowering Blood Cholesterol to Prevent Heart Disease20 to make a national screening recommendation with the anticipation that over a period of 20 to 30 years, the high incidence of death from coronary throm bosis could be significantly altered. The result was the publication of a document which established serum cholesterol as th e th ird n atio n a lly reco m m e n d e d health screening tool for all citizens of the USA on a repeated five-year basis. Therefore, in 1987 the NIH National H eart, Lung and Blood Institute established the National Cholesterol Education Program 19 which convened a symposium on the current status of serum cholesterol as a public health screening tool. D u rin g th e p ast two years, an entirely new literature has sprung into place regarding the im plem entation of this national recom m endation. The two previously nationally recom m ended national public health screening tools were (1) blood pressure and its predictive value in the prevention of the com plications of hypertension and (2) cytologic screening of cells desquamated from the uterine cervix and its predictive value for cancer of the uterine cervix. Both of these screening measures have dem onstrably influenced death rate figures for the respective areas of medical mortality. Serum cholesterol is the first national screening tool which relies upon the chem ical analysis of a substance in human serum. The experience over the past 100 years dem onstrates the evolutionary pitfalls which have been encountered. It also is a good example of the m aturation of a specific analytical m ethodology to include Reference M ethodology11 and a D efinitive M ethodology.5 Through the combined efforts of clinical researchers, laboratory scientists and governm ental and non-governm ental lead e rs, th e fu tu re decision-m aking backbone of the National C holesterol E ducation Program 19 of the N ational Institutes of H ealth is well established. The Im portance of Constancy of Serum C holesterol M easurem ent If reliable medical decisions are to be m ade, th e serum cholesterol m easurem ent m ust be held constant. The p u r pose of this presentation is to explore the background and progress in this regard. The present question is: Do we now have the tools to keep cholesterol m easurem ents comparable for the next 100 years? It is interesting to note that in 1950 Dr. Shields W arren,22 then D irector of

3 SERUM CHOLESTEROL METHODOLOGY 3 the Division of Biology and Medicine of the Atomic E nergy Commission, im plem en ted a study of com parative cholesterol methods by four leading research groups in San Francisco, C leveland, Pittsburgh, and Boston. This was necessary since the observed systematic bias b e tw e en m ajor rese a rc h groups was excessive. As with each of the m ethods for measurem ent of compounds in the hum an serum, serum cholesterol m easurem ent evolved (1) with respect to photometric analysis (Sunderman24); (2) with respect to p u re standards (F ieser12); (3) w ith respect to specific extraction isolation of cholesterol from the serum (Schoenh e im e r and S p e rry,23 K e n d a ll and Abell1); and finally (4) through the efforts of CD C and NBS to control the explosion of autom ated methodology (1960-to present) using non-extraction procedures and many uncontrolled secondary standards. This u n e x p ected m elange of methods of automated analysis affected both research and clinical efforts and almost prevented the final determination of cholesterol screening efficacy. It is clear that the single institution which brought order out of chaos and m ade possible the confidence in the Lipid Research Clinics Cholesterol Program s15,17 upon w hich the N IH conferees acted was the Communicable Disease Center Lipid Laboratory led by Dr. G erald Cooper. Im p o rtan t additions w ere (1) the National B ureau of Standards definitive m ethod (19805), and (2) the N IH Manual of Laboratory O peration for the Lipid Research Clinics Program Coronary Diet and Coronary Drug studies requiring CD C cholesterol m easurem ent supervision. The constancy of the cooperating Lipid Clinics cholesterol analyses has been all important. The history of the developm ent of cholesterol m ethodology has several im portant lessons which follow Sunderm an s7 basic principles of clinical chem istry analysis that (1) p u re chemical standards must be the basis for a method; (2) each step in the procedure m ust be analyzed for its effect on the final result; (3) reproducibility of results from day to day m ust be achieved; and (4) hum an reference risk values m ust be used for decision-making. Cholesterol methodology will be discussed from the following viewpoints: (1) Instrum entation, (2) Sample preparation, color reaction, enzym e specific m ethods, calibrator decisions, (3) Pure standards, (4) Secondary reference m aterials, and (5) Present status. Instrum entation The first instrum ent used for quantitation was the hum an eye which could detect concentration differences on a (1) spot test; (2) a visual comparison with a set of calibrated standards in sealed tubes, and (3) the visual split field comparator. The split field comparator was the tool of choice from 1910 to In this device, a standard solution was viewed at a set depth, and the unknown sample was com pared with the known standard solution on the split screen. The observer matched the color of the two half circles by changing the depth of the unknow n sam ple. The m olecular c o n c e n tra tio n of th e unknow n was inversely related to the linear depth of the standard. At the equality point, the same n u m b er of m olecules w ere b e tw een the p lu n g er face and th e cup depth for both standard and unknown. Sunderman and Razek24 in 1939 at the U niversity of P ennsylvania, using a unique state of the art recording spectrop h o to m eter, im proved p recisio n by introducing the concept of a tim e-related observation. The tim e-related observation was based on th e recording spectroscopy data w hich show ed th a t th e

4 4 COPELAND spectral transm ittance color developm ent curves of the cholesterol acetic an h y d rid e sulfuric acid m ixture coincided at 10 to 40 m inutes of color developm ent at 540 millimicrons. By 1945 photoelectric cells were well known. Simple as well as sophisticated photometric-measuring instrum ents had becom e available. The two serious problems which had led to unreliable photocell in stru m en ts w ere the inconstant output of the photovoltaic cell itself and the m inute-to-m inute variability of the line voltage. Instability of the reading of the same tube could be the result of either input fluctuation or output fluctuation. K lett stabilized his com parator photom eter by balancing the working photocell output against a reference photocell output using a calibrated resistance. The reference photocell cancelled the effect of line voltage changes. An ingenious logarithmic unit scale on the resistance made possible an easy linear calculation of concentration. Evelyn stabilized the power source by using a large six-volt battery with a low c u rre n t draw, and th is in stru m e n t achieved wide acceptance because of its stability. The third instrum ent which appeared at this tim e was the Beckman quartz prism spectrophotom eter which com b ined fine discrim ination of the wavelength of incident light, a photom ultiplier for detection of em ergent light and a balanced reference circuit. This instrument, the Beckman DU, was renowned for 30 years and set a standard for photom etric analysis. L ater, Beckm an and Carey introduced an advanced m odel which, until recently, was the U nited States National Bureau of Standards reference instrum ent. In the 1960s and 1970s, two additional instrum ents were developed separately and then combined to make the present day ultimate measuring instrum ent, the mass spectrom eter for com ponent m easurem ent w ith liquid chrom atography for component isolation. Molecules are isolated by the gas chromatography and are analyzed at the molecular or submolecular level by the mass spectrograph. The introduction of stable isotope dilution in which a labelled isotope is clearly id e n tif ie d c o m p le te s th e c y c le o f im provem ent. Against this background of steadily improving m ethods of analytical quantitation, there also developed a plethora of auto m ated analyzers and rea g e n t systems. The objective for these analyzers was (1) fast sample processing, (2) autom ated reagent additions, (3) automated sample reading, (4) simplified output, (5) small sample size, and (6) ability to process whole serum or whole blood. Many problems of systematic bias developed, usually giving a positive 10 to 30 mg per dl bias. The eventual successes with the autom ated instrum ent screening relate to im provem ents related to standardization and quality control programs to be considered. Sample Preparation, Color Reaction and Enzyme Specific Reagent Methods and C alibration Decisions Early work in the microchemical analysis of hum an blood and blood serum in d ic a te d th a t re d cell p ro te in s and serum proteins were major interferences in th e analytical procedure for cholesterol color development. A protein-free filtrate of whole blood was developed as well as a protein-free filtrate of plasma. B etw een 1910 and 1940,1>2-23 it was found co n v en ien t to p re c ip ita te the serum p rotein and extract the cholesterol and cholesterol esters using several solvent combinations chloroform, ethyl ether, acetone and ethyl alcohol. After evaporation of the solvent, th e choles

5 SERUM CHOLESTEROL METHODOLOGY 5 terol and cholesterol esters were reacted with acetic anhydride and concentrated sulfuric acid to produce a green-blue color. This color was related directly to the cholesterol concentration. It was noted that the color for cholesterol esters developed faster than the color for unesterified cholesterol. Schoenheimer and Sperry23 in 1934 published a m ore specific m eth o d for to tal ch o lestero l in which the cholesterol esters w ere first hydrolyzed. The total cholesterol was precipitated as the digitonide. The p recipitate was w ashed, redissolved, and reacted in the acetic anhydride-h 2S 0 4 system. D uring the 1940s and early 1950s, this was considered the reference m ethod by the clinical research community. Copeland8 noted (1954) that when the digitonide cholesterol crystalline precipitate was washed, the crystalline material becam e a paste. A second wash of the pasty m aterial was called for w hich seem ed unnecessary. Several years later, Sperry affirmed this observation. The elim ination of the effect of the cholesterol ester from the color developm ent and the general shift from visual colorim eter to stabilized photovoltaic cell detector took place in the 1940s. Although not carefully docum ented, a significant reduction in the average cholesterol levels took place, approximately minus 15 to 30 mg per dl. The introduction in 1952 of the Abell- Kendall Total Cholesterol m ethod1 substituted alkaline hydrolysis of the ester and extraction with petroleum ether for the digitonide precipitation and precipitate washing steps of the Schoenheimer- Sperry m ethod.23 Abell-Kendall results w e re sh o w n to b e c o m p a ra b le to Schoenheim er-sperry results, and the tedious digitonide p recipitation and washing steps w ere avoided. I t was th e A b e ll-k e n d a ll m eth o d which was selected by the scientists at the Harvard School of Public Health22 as the basis for the Framingham Study in the 1950s and 1960s. It has continued to be the reference m ethod used nationally in clinical research circles. T h e C D C L ip id L a b o ra to ry, th e National Bureau of Standards Lipid Laboratory, and the Food and Drug Administration (FDA) collaborated in the 1970s to develop a Definitive M ethod for Cholesterol in S erum -nam ely the Isotope Dilution-Mass Spectrography m ethod.5 The definitive m ethod then was used to confirm the validity of the Reference M ethod11 which was the Abell-Kendall M ethod as modified at the CDC Lipid Laboratory. S everal in te rn a tio n a l com parative trials have been conducted using the Definitive M ethod w ith reference laboratories in Sweden and Norway. In this way, the international comparability of cholesterol standardization has been m aintained. In the mid 1970s, methods were introduced based on specific enzym es for splitting the ester linkage and oxidase specific reaction with cholesterol. These methods were shown to approximate the Abell-Kendall m ethod. They received im m ed iate accep tan ce ow ing to the avoidance of the caustic chemicals, acetic anhydride, and concentrated sulfuric acid. The FD A w ith its authority in the medical appliance and instrum ent area undertook licensing of all com m ercial products for clinical testing in early Cholesterol was among the many kit tests licensed according to the basic principle that the test should m easure cholesterol and be free of interfering compounds. Many of the early automated reagent system s designed for autom ated use w ere found to be badly calibrated, and a significant n u m b er of unfortunate incidents occurred. This was due to two fac

6 6 COPELAND tors: (1) faulty calibration, and (2) normal reference ranges which were incorrectly d efin e d. T he u su al sy stem atic bias ranged from plus 15 to 30 mg per dl. At the present time, FDA requires that all acceptable kits m ust be traceable to C D C L ipid L aboratory or NBS stan dards. The most recent study6 of 41 curren t m ethods summarizes the 1988 College of Pathologists Chem istry Survey. This indicates that 30 percent of the m ethods have less than a five mg bias maximum error with respect to the CDC referen ce m ethod value, and th at 45 percent of the methods in use have less than a five to 10 mg bias with respect to the CD C value. Seventeen percent of the m ethods showed a bias outside 10 mg. These 17 percent would require parallel standardization. P u re C holesterol Standard M aterial The liver has a specialized function for purification of cholesterol in th e gallstone. Until 1969, analysts recrystallized cholesterol for standardization purposes. No pure, stable cholesterol was available from the early 1920s to Cholesterol was known to deteriorate and contain im purities and always req u ired recryst a l l i z a t i o n. F i e s e r ( )12 r e ported that oxidation of pure cholesterol occurred on standing and stated that a s ta b le c h o le s te r o l c o m p o u n d was needed. Cholesterol dibrom ide was recom m ended by Fieser (1953) as a stable cholesterol material. In the 1950s and early 1960s, the cholesterol dibrom ide was the reference m aterial of choice; however, it was not commercially available, and each laboratory had to purify the commercial product or prepare its own dibrom ide compound. Lack of a readily obtainable stable cholesterol standard led the Standards Com m ittee of the College of American Pathologists (founded in 1948 by Sunderman) to organize a two-day seminar in 1966 on the subject of serum cholesterol standardization. This m eeting was held at the H arvard School of Public H ealth, June 16, 1966,18 and included Dr. Fieser and Dr. Stare from Harvard, and representatives from the National B ureau of Standards, th e C ollege of A m erican Pathologists, the Am erican Association of Clinical C hem ists, and other interested scientists. The conclusion of the m eeting was that a pure, stable cholesterol standard was needed at the national level. The second conclusion was that no commercial products m et these specifications. A second m eeting was held by the CAP S tandards C om m ittee Septem ber 1966 at which m anufacturers and reagent com panies w ere requested to consider the production of the recom mended standard. No action was forthcoming. The Standards Com m ittee of the College of American Pathologists at this point decided to produce the needed cholesterol standard itself unless the National Bureau of Standards decided to undertake this responsibility. Fortunately, under the leadership of Dr. Thomas Mears and Dr. Wayne W. M einke, th e National B ureau of Standards Cholesterol Standard Reference M aterial SRM #91 la was developed and m arketed (1969).21 This was the first of a new NBS series of SRM s for laboratory m edicine and clinical chemistry. D evelopm ent of Secondary Serum R eference M aterials In the 1970s, as part of the CDC Lipid Laboratory Standardization program,15-17 a series of secondary serum sam ples were produced for calibration of m ethodology in the long-term coronary D iet Trial and the long-term coronary D rug Trial and for the Lipid Research Clinics.

7 SERUM CHOLESTEROL METHODOLOGY 7 These samples w ere verified by the reference m ethod (Abell-Kendall) and had the advantage of the traceability to the CDC Lipid Laboratory allowing all the d a ta from a p p ro x im a te ly 12 to 15 research centers to be pooled. W ithout th e cen tral standardization, the data would have shown extensive m ethodology system atic bias which could have obscured the effect of these im portant studies. Fortunately, under the leadership of C ooper at the Com m unicable Disease C enter Lipid Laboratory, a focal point of standardization was m aintained. These secondary reference materials from C D C w ere calibrated by th e Isoto p e D ilu tio n M ass S p e c tro g rap h y (IDMS) m ethod as well as by the Reference M ethod (CDC). T h ere was no question about the validity of the CDC secondary standard material. F resh Parallel Secondary Calibration Several m anufacturers whose in stru m en ts show ed a negative bias w ith lyophilized secondary reference m aterials showed that when fresh serum samples are ru n by the C D C R eference M ethod (Abell-Kendall) and then are used as parallel secondary calibrators, the analyses based on the fresh parallel secondary calibrators, check very closely with CDC analyzed values. This is the current status. Both the College of American Pathologists and the National Bureau of Standards have m arketed lyophilized secondary reference material whose concentration has been determ ined by NBS, IDM S, and CDC. Extensive studies are now underway to extend the usefulness of the current crop of secondary calibrators. One approach has been to balance the cholesterol ester ty p e s in th e ly o p h iliz e d m a te ria l. A nother approach is to prepare the secondary standard as a liquid material. As shown in a recent analysis by Copeland,6 32 of the 41 m ethods (80 percent) used by participants in the College of American P ath o lo g ists C h e m istry S urvey (1988) can use the present lyophilized m aterial as an accuracy reference m aterial. However, seven instrum ents show a distinct negative lyophilized serum bias and must be calibrated indirectly by parallel serum measurements using the Abell-Kendall CDC reference method. This is very inconvenient, and it p re vents the user from having direct control over the accuracy status of her/his cholesterol analyses. Currently, the Food and D rug Adm inistration has accepted this m ethod of reference calibration. However, further research is in progress to elim inate this problem atic bias. Present Status of Serum Cholesterol Standardization and C om parability Verification 1. All analyses of cholesterol m ust be traceable to th e C enters for D isease Control and National Bureau of Standards and M ethods. 2. A Standard Reference Material of p u re c h o le s te r o l is a v a ila b le (NBS) Standard Reference M aterials of cholesterol in serum lyophilized and frozen liq u id are available (NBS and CAP) A conjoint array of laboratories called th e National Reference System for Cholesterol is available for parallel fresh serum analysis (by the C D C R eference M ethod) of split liquid serum samples allowing verification of individual laboratory calibration (CDC).4 5. A nationwide program for comparative evaluation of cholesterol in lyophilized serum is available in which the national perform ance in 4,000 laboratories is m onitored four times a year (CAP Survey).25

8 100 Years of Evolution in Cholesterol Measurement Color Development acetic anhydride acetic anhydridesulfuric acid Lieberman-Burchard Lieberman Burchard Lieberman Burchard Lieberman Burchard Dawber-Mann Framingham Study selected Abel 1-Kendal 1 method as the "four clinics" analytical method. 00 COPELAND TABLE I Method Lieberman^ Burchard3 Date Pubii shed Measurement Ester Standard Instrument Hydrolysis Extraction Developed the color reaction cholesterol plus sulfuric acid plus Improved color by chloroform solution of cholesterol Bloor Laboratory purified cholesterol Vi sual comparator None Chloroform Schoenheimer Sperry23 Laboratory purified cholesterol Visual Comparator (Zeiss) alkaline hydrolysi s Digitonide precipitate wash acetone ethanol Sunderman2^ 1939 Laboratory purified cholesterol Recording Timed Spectophotom- Color deeter velopment Chloroform Two shifts (1) Visual Comparator to photocell detector - (2) from no ester hydrolysis to ester hydrolysis before extraction, reduced total cholesterol by mg/dl. Technical Group22 - Committee on Lipoprotein-Comparative Studies - First Major Comparability Study Abell-Kendall Laboratory Photoelectric alcoholic purified photometer alkaline cholesterol hydrolysis petroleum ether hexane Stare9»22

9 Many automated, rapid methods introduced significant positive systematic bias mg/dl, screening procedures and in the other automated systems monitored by CAP Surveys. SERUM CHOLESTEROL METHODOLOGY CD Cholesterol^ 1966 Standard Conference National Bureau 1969 of Standards^! NIH-CDC-NBS CDC - Lipid 1977 Laboratory^ College of American Pathologists Standards Committee - meetinq at Harvard School of Public Health: Recommended that a pure reference cholesterol material be marketed. SRM 911a Cholesterol: First NBS Standard Reference Material for clinical laboratory measurement purposes. Major cholesterol projects used CDC Lipid Laboratory as the Quality Control Monitor Coronary Drug Project Coronary Diet Project Lipid Research Clinics Selected Abell-Kendall Method as reference method for Lipid Standardization Laboratory. Reference Method 1976 Duncan C o o p e r ^» l l NBS-SRM High quality spectrophotometer alcoholic hexane alkal i Lieberman, Burchard (optomized) Definitive Method Cohen-SchafferS 1980 NBS SRM Isotope dilution Gas chromatography Mass spectrometer hexane Particle size Mass spectrometry National Refer ence Method Laboratory Network^ Cooper (CDC) NBS - SRM High quality spectrophotometer alcoholic alkali hexane Lieberman, Burchard (optomized) Significant progress in controlling systematic bias in both large volume small sample

10 10 COPELAND 6. T he previous five program s are necessary to keep the proliferation o f c h o le s te r o l m e a s u r e m e n t methods for human blood or serum within the comparability and transferability lim its needed over the next 50 years for reliable medical decision-making. Discussion In retrospect of the past 100 years, the research and clinical evolution of cholesterol m easurem ent has been the result of im portant contributions by individual scientists: Bloor, Lieberman, Burchard, K endall, Schoenheim er, Sunderm an, Sperry, Castelli, Dawber, Stare, Keys, Cooper, and others; by governm ental agencies: Communicable Disease Center Lipid Laboratory, National Bureau of Standards Reference M aterial Section, Food and Drug Administration Section on Clinical Testing Devices, and NIH Com m ittee on Cholesterol Education; by national organizations: including College of American Pathologists Standards Com m ittee and the CAP Comparative Survey Program, and National Com m ittee on Clinical Laboratory Standards, American Association of Clinical C hem ists, and by many m anufacturers who have developed and m arketed cholesterol m easuring systems. There has been a constant striving for accuracy and precision. Comparability was preserved by the CD C and NBS. These first principles m ust be continually review ed and redefined into the twenty-first century at least until 2,030 when the effect of cholesterol monitoring of th e USA p opulation can be d efinitively m easured in term s of m ortality statistics. The N IH com m ittee has been very proactive in its support of universal screening. Various subcom m ittees on quality control and accuracy have made m ajor contributions to a disciplined approach to bringing this new health protection effort to fruition. Conclusion T he r e c e n t im p le m e n ta tio n o f a National C holesterol Inform ation Program has b een b u ilt on 100 years of research. It appears that prospects for precise and accurate measurements for the present and future generations are guaranteed by the careful work of many investigators who have established basic principles which m ust be continued for the next 100 years. References 1. Ab e l l, L. L., L evy, B. B., Br o d ie, B. B., and K e n d a l l, F. E.: Sim plified m eth o d for th e estim atio n o f to tal ch o lestero l in serum an d d e m o n stratio n o f its specificity. J. Biol. C hem. 195:357, B lo o r, W. R. : The determination of cholesterol in blood. J. Biol. Chem. 2 4 : , B u r c h a r d, H.: Increased color intensity when cholesterol is in chloroform. B eitrage zur Kenntnis des Cholesterins. Chem. Zentralbl. 61:2 5-27, CDC National Reference System for Cholesterol Atlanta. 5. C o h e n, A., H er t z, H. S., M a n d e l, J., Pa u le, R. C., Sn eg a sk i, L. T., Su n, T., W e l s h, M. J., W h it e, E., C a l i, P., and Sc h a f f e r, R.: Total serum cholesterol by isotope dilution mass spectrography: A candidate definitive method. Clin. Chem. 26: , C o p e l a n d, B. E.: Analysis o f cholesterol method performance in the eight lyophilized samples in the chemistry survey College of American Pathologists, Chicago, Illinois. Manuscript in preparation, COPELAND, B. E.: Personal communication with Dr. F. W. Sunderman, C o p e l a n d, B. E.: Personal communication with Dr. Warren Sperry, D a w b e r, T. R., M o o r e, F. E., and M a n n, G. W.: Coronary heart disease in Framingham study. Amer. J. Pub. Health 47:4, D u n c a n, I. W. a n d C o o p e r, G. R. : D e v e lo p m en t o f a candidate referen ce m eth o d for cholesterol. C lin. C hem. 22:1193, D u n c a n, I. W., M a t h e r, A., and C o o p e r, G.R. : The procedure for the proposed Cholesterol Reference Method, Atlanta, GA: Centers for Disease Control, 1982.

11 SERUM CHOLESTEROL METHODOLOGY FlESER, L.: Cholesterol and companions. VII. Steroid bromides. J. Am. Chem. Soc. 75: , Ka n n e l, W. B., C a st e l l i, W. P., G o r d o n, T., et al.: Serum cholesterol, lipoprotein, and the risk of coronary heart disease: The Framingham Study. Ann. Intern. Med. 74:1, Ka n n e l, W. B., D a w ber, T. R., Ka g a n, A., Rev o t sk ie, N., and St o k e s, J.: Factors of risk in the development of coronary heart disease: Sixyear follow-up experience. Ann. Intern. Med. 55:33-50, Lab M ethods Com m ittee of Lipid Research Clinics Program of National Heart, Lung, and Blood Institute. Serum and plasma cholesterol. Clin. Chem. 23:60-63, L ie b e r m a n n, C.: U eber das Oxychinoterpen. Deutsch. Chem. Geselsch. 7S: , Lipid Research Clinics Program. Manual of Laboratory Operations. 1. Lipid and Lipoprotein Analysis. Departm ent of Health, Education and Welfare Publication #N IH , Government Printing Office, Washington, D.C., M u e l l in g, R. J. and C o p e l a n d, B. E.: Two meetings on selection criteria for a pure cholesterol preparation to be used for standardizing serum cholesterol measurements for medical diagnosis and therapy. Am. J. Clin. Path. 47: , National Cholesterol Education Program, U.S. D epartm ent of Health and Human Services, Public Health Service, National Institutes of Health, Washington, D.C., National Institutes of Health-Consensus Development Conference. Lowering blood cholesterol to prevent heart disease. J. Am. Med. Assoc. 253: , Office of Standard Reference Materials, Cholesterol SRM #911a. National Bureau of Standards, Washington, D.C Technical Group Committee on Lipoprotein and A therosclerosis of National Advisory H eart Council. Cooperative lipoprotein study. Circulation 14:691, Sc h o e n h e im e r, R. and Sperry, W.: A micromethod for the determination of free and combined cholesterol. J. Biol. Chem. 106: , S u n d e r m a n, F. W. and Ra zek, J. Spectrophotometric studies of the color development in the analysis of sugar by the Benedict method and of cholesterol by Lieberman-Burchard Reaction. J. Biol. Chem. 118: , Survey Program Chemistry, College of American Pathologists, Chicago, Illinois, 1988.