Proficiency testing of creatine kinase and creatine kinase-2: the experience of the Ontario Laboratory Proficiency Testing Program

¹ Laboratory Proficiency Testing Program, 250 Bloor St. East, Suite 501, Toronto, ON, Canada M4W 1E6.

² Department of Biochemistry, University of Western Ontario, London, ON, Canada.

³ Department of Laboratory Medicine, The Credit Valley Hospital, Mississauga, ON, Canada.

⁴ Department of Laboratory Medicine, Sunnybrook Health Science Centre, North York, ON, Canada.

⁵ Department of Laboratory Medicine, University of Toronto, Toronto, ON, Canada.

⁶ Department of Clinical Chemistry, Hamilton General Campus, Hamilton Health Sciences Corporation, Hamilton, ON, Canada.

⁷ Department of Pathology, McMaster University, Hamilton, ON, Canada.

⁸ Analyzer types and manufacturers are: aca, Dimension, Paramax, and Stratus systems (Dade International); Access (Sanofi Diagnostics Pasteur); ACS 180 (Chiron); Beckman CX systems (Beckman Instruments); Cobas Mira systems (Roche Diagnostic Systems); Hitachi systems (Boehringer Mannheim); ICON (Hybritech); RA systems and Immuno I (Bayer Corp.); Spectrum, IMx, and AxSYM (Abbott Labs.); Vitros DT, 250, 500, and 700 (Johnson and Johnson Clinical Diagnostics).

⁹ LPTP requires that proficiency testing challenges are processed as patient specimens. Accordingly, outliers reported in these surveys have not been subjected to statistical trimming.
^a Address correspondence to this author at: 19 Linksgate Rd., London, ON, Canada N6G 2A6. E-mail ahenders{at}julian.uwo.ca.

	Abstract

Top Abstract Introduction Materials and Methods Results Discussion References

 
The Ontario Laboratory Proficiency Testing Program has regularly monitored the analytical performance of total creatine kinase (CK) (230 participants) and CK isoenzyme-2 (CK-MB) (160 participants) throughout the entire province. Consistently, a wide dispersion of results has been observed not only between different analyzer systems but also among identical analyzers. Accordingly, the results of the last three proficiency surveys for these analytes were examined statistically to establish both the extent of these variations and the range of values reported for the male upper reference ranges. The results of many of the analyzer systems were significantly different from each other, as were many of the reference ranges. This unsatisfactory situation may only be remedied by the use of reference materials as shown by others. The consequences of these findings also effect the reliability of epidemiological surveys such as the WHO MONICA Project (Circulation 1994;90:583–612), which monitors deaths due to heart disease and includes cardiac enzyme results in its criteria.

	Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

 
The healthcare system in Ontario served, in 1996, a population of 10.8 million people (38% of Canada's population), rather similar to that of Michigan (9.5 million), Ohio (11.2 million), or Illinois (11.8 million), and results of proficiency testing obtained from such a system are therefore likely to be of interest to other North American healthcare organizations.

Mandatory external quality assurance for Ontario laboratories comprises licensing and inspection by the Ontario Ministry of Health and, since 1974, mandatory proficiency testing by the Ontario Medical Association under a contract from the Ontario Ministry of Health. The Laboratory Proficiency Testing Program (LPTP), a unit of the Ontario Medical Association (1), has as its primary functions (a) the external quality assessment of the performance of all Ontario patient-testing laboratories, whether owned by hospitals or commercial enterprises, and (b) the provision of educational assistance to all such laboratories when there are proven deficiencies in their preanalytical, analytical, and postanalytical performance. Testing is assigned to several working committees, one of which is the Enzymes, Cardiac Markers, and Lipids Committee. This Committee has tested all Ontario laboratories licensed to determine creatine kinase (CK; EC 2.7.3.2) and CK isoenzyme-2 (CK-2 or CK-MB) since 1989 (2) with lyophilized human-based serum, and, from 1991 onwards, with fresh frozen human serum supplemented with purified human CK-3 (CK-MM) and CK-2 (3).

Of all enzyme activity assays, the determination of CK is one of the most investigated (4) and has been the subject of IFCC standardization (5). This assay, known as the NAC-EDTA formulation, is used by 90% of the 230 Ontario laboratories so licensed, and 99% of all these laboratories can regularly achieve replications of 6.5% at all levels of CK activity (6). Although intralaboratory performance is satisfactory, the Committee has noted consistent differences in results between the same type of analyzer when used by different laboratories even though reference ranges are similar. Also, there appear to be marked differences in the results produced by different analyzers. Similarly the intralaboratory assay performance for CK-2 challenges is satisfactory, with >90% of the 160 participants achieving replication of 10.0%. Interlaboratory performance, however, often shows marked discrepancies in both activity and mass assay results for this isoenzyme.

With the consolidation of tertiary care services throughout Ontario, enzyme estimations made at a peripheral hospital may diverge from those made at the receiving hospital, with obvious consequences for patient care. Furthermore, the cardiac enzyme criteria of the WHO MONICA Project (7) for the diagnosis of myocardial infarction code as abnormal a result more than twice the "upper limit of normal" for that test in each (our italics) laboratory. If there are biases in analytical performance or in the reference range, the resulting diagnostic classification may be erroneous. Accordingly, the Committee statistically assessed these analytical differences in three recent surveys (June 1996, Z-9606; October 1996, Z-9610; and January 1997, Z-9701) to ascertain the extent of analyzer differences and the effect of the reported male upper reference limit on them.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

 
Male upper reference ranges
were obtained from the LPTP archives for each laboratory for total CK and CK-2 for an adult male. Every laboratory is required to report its methods, analyzers, and reference ranges to LPTP.

Survey materials.
Fresh frozen human serum was obtained from the London branch of the Canadian Red Cross Society. Several individual bags from different individuals were pooled to provide sufficient material for each survey. Endogenous CK and CK-2 activities were uniformly at the lower ends of their respective reference ranges. Purified human CK-3 (from skeletal muscle) and CK-2 (from heart) were obtained from Scripps Labs., where they had been prepared by sequential ammonium sulfate precipitation, ion-exchange chromatography, and proprietary affinity chromatography. Scripps Labs. established homogeneity by the appearance of a single protein zone after sodium dodecyl sulfate electrophoresis and by a single zone of CK activity after electrophoresis on agarose gel. These enzymes were provided in liquid form suspended in 500 mL/L glycerol, 5 mmol/L succinate, 1 mmol/L EDTA, 5 mmol/L 2-mercaptoethanol, and 10 mmol/L NaCl. Enzyme activities were 700 U/L at 37 °C. Survey materials (Table 1 ) were prepared by supplementing the fresh frozen serum with these purified enzymes; the extent of contamination of the survey materials by the enzyme preservatives was minimal because of the addition of microliter quantities to liter volumes of the fresh frozen serum. Survey materials were prepared and distributed as previously described (3).

Surveys.
The total CK and CK-2 activities and CK-2 mass assay concentrations used in the three LPTP surveys are listed in Table 1 together with the analyzers used by participants in each survey (Table 2 ). LPTP requests that all vials be analyzed within 24 h of receipt and that they may also be analyzed in the same test run. Experience suggests that results are therefore subject only to within-run variance.

For total CK, results from different analyzers from a single vendor were aggregated only if there were no statistically significant differences between them (see below); analyzer groups of less than five were not assessed and there were insufficient numbers to examine the effects of different vendor kits on the reported results. For CK-2 the results of all methods were included in the assessment, although statistical analyses could not be performed on the results from a single analyzer group.

Statistical analysis.
Statistical assessment was conducted with SigmaStat version 2.0 or with SYSTAT version 7.0 (SPSS). Analytical results and upper reference limits were first tested for normality (with the Kolmogorov–Smirnov test with Lilliefors' correction set at P = 0.05 (8)) and then for equal variances before testing either parametrically, by an unpaired t-test, or nonparametrically, by the Mann–Whitney rank sum test. A significant difference was defined as P <0.05 and all such values are reported exactly unless P <0.001. Linear regression and correlation were determined for all survey results for individual analyzer groups and the male upper reference limits reported for these analyzers.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

 
Point plots are shown for the results for survey Z-9606 for total CK activity (Fig. 1 A), CK-2 activity (Fig. 2 A), and CK-2 mass (Fig. 3 A); for survey Z-9610 for total CK activity (Fig. 1B ); and for survey Z-9701 for total CK activities (Figs. 1C , vial 1, and D, vials 2–5), for CK-2 activities (all vials, Fig. 2B ), and for CK-2 mass (all vials, Fig. 3B ).

View larger version (26K): [in this window] [in a new window]   Figure 1. Total CK results by analyzer types from (A) survey Z-9606 (all vials), (B) survey Z-9610 (all vials), (C) survey Z-9701 (vial 1), and (D) survey Z-9701 (vials 2–5). Results are shown only for groups of five or more instruments (see Table 2 ). Each point represents the mean value of all results reported for each laboratory.

View larger version (9K): [in this window] [in a new window]   Figure 2. CK-2 results determined by activity from (A) survey Z-9606 (all vials) and (B) survey Z-9701 (vial 1 and vials 2–5). All results are shown for all analytical systems used in each survey (see Table 2 ). Each point represents the mean value of all results reported for each laboratory.

View larger version (12K): [in this window] [in a new window]   Figure 3. CK-2 results determined by mass from (A) survey Z-9606 (all vials) and (B) survey Z-9701 (vial 1 and vials 2–5). All results are shown for all analytical systems used in each survey (see Table 2 ). Each point represents the mean value of all results reported for each laboratory.

Statistical assessments of survey results are tabulated in Table 3 (for total CK measurements), in Table 4 (for CK-2 activity measurements), and in Table 5 (for CK-2 mass assays).

Point plots are shown for the adult male upper reference ranges for total CK activity (Fig. 4 A), CK-2 activity (Fig. 4B ), and CK-2 mass (Fig. 4C ). Statistical assessment of the total CK reference ranges are tabulated in Table 6 .

View larger version (15K): [in this window] [in a new window]   Figure 4. Upper reference limits for adult males, by analyzer type, reported by all Ontario laboratories licensed to perform total CK or CK-2 assays. (A) Total CK activity, (B) CK-2 activity, (C) CK-2 mass.

The relation between the male total CK upper reference limits and the results of each survey for each major analyzer group are shown in Table 7 .

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

 
Total CK activity.
The results obtained for total CK activity for the major analyzer groups in survey Z-9606 are displayed in Fig. 1A . Some instrument groups show a very tight distribution of results—notably the Paramax and RA analyzers—although others do not. This has always been a worrisome feature of these survey results. A statistical analysis of the data is provided in Table 3A. After a preliminary assessment of the results, we found it necessary to separate the results of the Vitros analyzers into three instrumental groups: DT, 250, and the combination of the 500 and 700 analyzers. It is evident that the Paramax and RA systems provide significantly different results; many of the other differences are almost certainly due to the presence of one or two outliers in several instrument groups. What is of great concern is that a standard assay system, such as on the Vitros 250, can produce such a wide range of results.

In survey Z-9610 (Fig. 1B ), the results from the different analyzer groups, at higher CK activities, have changed. There was no statistical difference between the individual Vitros analyzers (the DT instrument was then used by fewer than three laboratories and has been excluded from these results), so all other Vitros analyzer results have been consolidated. The tight distributions previously seen for the Paramax and RA group results are no longer evident, whereas the Dimension and Hitachi groups display a smaller dispersion of results. Still noticeable, however, are the significant differences between the Paramax and all other analyzers (Table 3B).

In survey Z-9701 (Fig. 1C and 1D ), the two total CK activities were chosen, respectively, to be below the upper reference limit and to be midway between the activities used in the previous surveys. Outliers in the vial 1 results (Beckman, Spectrum, and Paramax) caused an increased dispersion of the data, leading to significant differences (Table 3C); the removal of these outliers converted these differences to ones of nonsignificance. The same outliers in the results from vials 2–5 (Table 3D) did not alter the statistical assessments.

We therefore conclude that differences between analyzer results exist across a wide range of total CK activities. Even the use of similar analyzers does not avoid such variability. As Hyltoft Peterson and Hørder illustrate in detail, both analytical imprecision and analytical bias may cause false classification of reported results (9). As noted above, LPTP CK challenges appear to be analyzed with acceptable imprecision, so clearly analytical bias is the major problem. It appears likely, from an inspection of Fig. 1 , that bias could be both negative and positive, thus further complicating the effect on patient classifications. A reviewer suggested that the variability of results could also be due to preanalytical factors such as delay in analyses or differences in storage conditions. Laboratories are asked to process LPTP specimens similar to patient specimens, and indeed each laboratory director is required to certify such handling. Nonetheless, preanalytical processing is clearly a component contributing to the reported variability in results.

How are CK measurements processed by each analyzer? The Vitros systems and the Dimension analyzers use CK calibrators. The remainder use "self-indicating" reactions based on the reduction of NADP by the CK assay and its known molar absorptivity (10). Two possible sources of error suggest themselves: the CK calibrators are not stable at the time they are used or the spectrophotometric accuracies of the analyzers are erroneous. Clearly there is an urgent need for a consensus on the calibration of the CK assay. Bowers and McComb proposed the concept of an International Clinical Enzyme Scale and demonstrated the effectiveness of its application (11)(12). Recent work in the European Commission has further shown the practicability of using standard reference materials to remove the variability of enzyme results in clinical practice (13)(14).

CK-2 activity.
The major techniques used for CK-2 activity determinations are immunoinhibition (Table 2 ). The results obtained in survey Z-9606 (Fig. 2A ) show a very wide variation for non-J&J immunoinhibition determinations compared with those obtained with the Vitros analyzers. An analysis of these results (Table 4 ) shows that although there are no differences between these techniques and the four results obtained by electrophoretic separation of the CK isoenzymes, there is a highly significant difference between the two immunoinhibition techniques. In survey Z-9701, no laboratory used electrophoretic separation (Fig. 2B ) and there continued to be a significant difference between the immunoinhibition methods. A comparison between the mass assay results for vial 1 (see below) with those obtained by the non-J&J analyzers (Fig. 2B ) was made. Serum containing practically no CK-2 (by mass assay) was reported by non-J&J immunoinhibition analyzers to have up to 12 U/L, whereas the J&J analyzers reported values <4 U/L. Users of these non-J&J systems argue that these high values are matched by higher reference ranges so that patient misclassifications do not occur. Nonetheless, from an analytical point of view, the obvious discordance between what CK-2 is actually present and what is measured must raise doubts about their accuracy; in addition, there is an adverse impact on the CK-2 fraction result.

CK-2 mass.
The major technique now used by two-thirds of all provincial laboratories is the mass assay in seven different formulations (Table 2 ). The results for survey Z-9606 are displayed in Fig. 3A . An obvious anomaly are the ACS-180 results. This system is known to give results about twofold greater than other common analyzers; Wians et al. (15) derived a regression equation of ACS 180 (µg/L) = 0.492 2.293 Stratus (µg/L). Results similar to those shown in Fig. 4A have occurred regularly in previous LPTP CK-2 surveys. Of concern is that laboratories using the ACS 180 have identical reference ranges to, for example, the Stratus and IMx analyzers, although their results are twofold greater. Communication with users of the ACS 180 brought this data to their attention. They insisted that their method comparisons justified such reference ranges. It is also evident that the results produced by the IMx and AxSYM are significantly different from each other (Table 5 ), presumably due to differences in calibration. Apart from this aspect, however, users of the IMx analyzer reported results ranging from 25 to 44 µg/L for a sample containing about 35 mg/L CK-2. Several significant differences between other analyzers were also observed (Table 5 ). In survey Z-9701, we used lower values for CK-2 to check if the differences observed in survey Z-9606 occurred across the range of values found in clinical practice (Fig. 3B ). Vial 1 contained very small quantities of CK-2, but significant differences were still observed between several analyzers (Table 5 ). Of interest are the ACS 180 results. In this survey the newly issued recalibrated CKMB II assay (giving results comparable with other mass assay systems) was used by three of the four participants to obvious effect. The differences noted between the IMx and AxSYM results in survey Z-9606 are clearly evident, again, in the results for vials 2–5 in survey Z-9701 (Fig. 3B and Table 5 ). Again, the IMx analyzer reported results ranging from 7 to 28 µg/L on a specimen containing about 20 µg/L. The Stratus, Access, and the ICON also show significant differences (Table 5 ).

The differences between the results of many mass assays are due, to an extent, to differences in calibration. Again, the forthcoming availability of a CK-2 mass standard will reduce this variability (16)(17). That it will likely not remove it entirely is evident from the spread of results shown in Fig. 3A and B.

Reference ranges.
We previously reported on the wide dispersion of total CK and CK-2 reference ranges in Ontario found in 1991 (18). Subsequent to these findings, the Committee encouraged many laboratories to revise or modify their reference values. The considerable effects of preanalytical variables on these analytes has been extensively documented (19). Such effects create considerable difficulties in establishing satisfactory limits. This topic has recently been comprehensively reviewed and detailed advice on analytical goals for generating reference intervals is available (20)(21).

Currently, the reported male upper reference limits for total CK throughout the province form a gaussian curve with a mean value of 199 U/L (SD = 23) with neither kurtosis nor skewing. The individual values, by instrument group, are shown in Fig. 4A . A statistical assessment of these values shows that the upper reference ranges reported for the Paramax analyzer are significantly different from all but three analyzer groups (Table 6 ). Other analyzer groups also show a few significant differences—Vitros, Mira, CX, Dimension, and the RA systems. Is there a relation between the individual values for each analyzer and the results from the three surveys? The Pearson product-moment coefficients of correlation (r) and determination (r) have been tabulated (Table 7 ). An r value less than about 0.4 can be regarded as showing little strength of association between the variables, and this appears to be the case for most analyzers with the notable exceptions of the Vitros DT, Dimension, and RA analyzers. The influence of the upper reference limit on the survey CK results is indicated by the r value, when expressed as a percentage. For example, for the Vitros 250, only 2% of the variation between the results from survey Z-9606 is accounted for by the upper reference limits for these analyzers, whereas for the Dimension analyzers, in the same survey, nearly 70% of the variation is accounted for by the reference limits (Table 7 ).

The male upper reference limits for CK-2 activity and mass are shown in Fig. 4B and C, respectively. Whereas the J&J ranges vary between 5 and 16 U/L, the range for other immunoinhibition assays rise as high as 25 U/L (Fig. 4B ). Mass values for the upper reference range do not exceed 10 µg/L, but in the case of the IMx may be as low as 5 µg/L. In two recent reports investigating troponin T and troponin I values in the acute coronary syndromes, both groups used the Stratus analyzer for CK-2 analyses. One group used 7 µg/L and the other group used 5 µg/L as their upper reference limits (22)(23). In Ontario, the ranges reported for the same analyzer were from 3 to 7 µg/L. One Stratus user reported an upper reference limit of 3 µg/L, whereas values of 4 µg/L were used by participants with the Immuno 1 and Access analyzers; these values appear to be far too low.

The Pearson product-moment coefficients of correlation (r) and determination (r) were also calculated for CK-2 activity and mass assays. Both the J&J and non-J&J immunoinhibition assays and the Stratus and IMx assays showed insignificant r values (and therefore r values), but the AxSYM assay results showed high r (0.65) and r (42%) values in the Z-9606 survey but not in the Z-9701 survey. It is not evident why this should occur except that the former survey had CK-2 values twofold the latter.

Do these variations matter? As noted earlier, the WHO MONICA Project for the diagnosis of myocardial infarction codes an enzyme result as abnormal if it is more than twice the "upper limit of normal" for that test in each laboratory (7). It is evident that the wide range in upper reference limits will undoubtedly alter the diagnostic classification. A report from the Minnesota Heart Survey noted that the incorporation of total CK and CK-2 into the Survey's diagnostic criteria significantly increased the "definite" myocardial infarction rates (24). But such classification will be distorted if either or both of the laboratory's upper reference limits or the analytical value of the determinations are incorrect. These possibilities are very evident in this recent Ontario experience. In this regard, the existence of significant interobserver variation in the interpretation of serum enzyme changes in myocardial infarction may further complicate patient classification (25).

We therefore conclude that currently the routine analyses of total CK and CK-2 in Ontario is unsatisfactory, lacking congruity between laboratories both as regards analytical results and reference ranges.

	Acknowledgments

 
This work was funded entirely by the Laboratory Proficiency Testing Program.

	Footnotes

 
¹ Address reprint requests to this author at the Laboratory Proficiency Testing Program.

	References

Top Abstract Introduction Materials and Methods Results Discussion References

 

1. Lohmann RC, Wood DE, Jacobs WI, Elliot ML. Reliability of white blood cell counting. Arch Pathol Lab Med 1989;13:989-994.
2. McQueen MJ, Patten RL, Henderson AR, Krishnan S, Wood DE, Webb S. Proficiency testing for creatine kinase isoenzyme CK-2 (CK-MB) in Ontario. Clin Chem 1990;36:2102-2105. [Abstract/Free Full Text]
3. Henderson AR, Stark JA, McQueen MJ, Patten RL, Krishnan S, Wood DE, Webb S. Is the determination of creatine kinase-2 after electrophoretic separation accurate?. Clin Chem 1994;40:177-183. [Abstract/Free Full Text]
4. Gerhardt W, Wulff K. Creatine kinase. In: Bergmeyer HU, Bergmeyer J, Grassl M, eds. Methods of enzymatic analysis, Vol. III, 3rd ed.: enzymes 1: oxidoreductases, transferases. Weinheim: Verlag-Chemie, 1983:508–39..
5. Hørder M, Elser RC, Gerhardt W, Mathieu M, Sampson EJ. Approved IFCC recommendation on methods for the measurement of catalytic concentration of enzymes: Part 7. IFCC method for creatine kinase (ATP:creatine N-phosphotransferase, EC 2.7.3.2). Eur J Clin Chem Clin Biochem 1991;29:435-456. [ISI][Medline] [Order article via Infotrieve]
6. Henderson AR, Webb S, Richardson H, Wood DE. A comparison of pooled, fresh-frozen, and lyophilized sera as a matrix for enzyme proficiency testing: the experience of the Ontario Laboratory Proficiency Testing Program. Clin Biochem 1996;29:183-185. [ISI][Medline] [Order article via Infotrieve]
7. . WHO MONICA Project,Tunstall-Pedeo H, Kuulasmaa K, et al. Myocardial infarction and coronary deaths in the World Health Organization MONICA project. Registration procedures, event rates, and case-fatality rates in 38 populations from 21 countries in four continents. Circulation 1994;90:583-612. [Abstract/Free Full Text]
8. Conover WJ. Statistics of the Kolmogorov–Smirnov type. Practical nonparametric statistics 2nd ed. 1980:344-393 John Wiley and Sons New York. .
9. Hyltoft Petersen P H, ørder M. Influence of analytical quality on test results. Scand J Clin Lab Invest 1992;52(Suppl 208):65-88. [Medline] [Order article via Infotrieve]
10. Moss DW, Henderson AR. Enzymes. Burtis CA Ashwood ER eds. Tietz textbook of clinical chemistry 2nd ed. 1994:735-896 WB Saunders Co. Philadelphia. .
11. Bowers GN, Jr, McComb RB. A unifying reference system for clinical enzymology: aspartate aminotransferase and the International Clinical Enzyme Scale [Review]. Clin Chem 1984;30:1128-1136. [Abstract/Free Full Text]
12. McComb RB, Bowers GN, Jr. Alkaline phosphatase and the International Clinical Enzyme Scale. Am J Clin Pathol 1985;84:67-73. [ISI][Medline] [Order article via Infotrieve]
13. Moss DW, Maire I, Calam DH, et al. Reference materials in clinical enzymology: preparation, requirements and practical interests. Ann Biol Clin (Paris) 1994;52:189-198. [Medline] [Order article via Infotrieve]
14. Lessinger J, Férard G, Grafmeyer D, et al. Usefulness of reference materials in calibration of enzyme activities. Eur J Clin Chem Clin Biochem 1995;33:859-864. [ISI][Medline] [Order article via Infotrieve]
15. Wians FH, Baskin LB, Keffer JH, Koch DD. Evaluation and implementation of an immunochemiluminometric assay (ICMA) for quantifying CK-MB concentration. Clin Lab Sci 1995;8:334-338.
16. Vaidya H, Apple F, Boches F, Christensen R, Jones K, Landt Y, et al. Preparation of preliminary standards for creatine kinase-MB (CKMB) mass immunoassays [Abstract]. Clin Chem 1993;39:1256.
17. Green S, Onoroski M, Moore R, Wu A, Lehrer M, Vaidya H. Use of proposed CK-MB standardization material for calibration of mass measurements: clinical correlation of patient sera [Abstract]. Clin Chem 1993;39:1269.
18. Henderson AR, McQueen MJ, Patten RL, Krishnan S, Wood DE, Webb S. Testing for creatine kinase and creatine kinase-2 in Ontario: reference ranges and assay types. Clin Chem 1992;38:1365-1370. [Abstract/Free Full Text]
19. Young DS. Effects of preanalytical variables on clinical laboratory tests, 2nd ed. Washington: AACC Press, 1997:1-1–5-11..
20. Harris EK, Boyd JC. Statistical bases of reference values in laboratory medicine. New York: Marcel Dekker, 1995:361pp..
21. Gowans EM, Hyltoft Petersen P, Blaabjerg O, Horder M. Analytical goals for the acceptance of common reference intervals for laboratories throughout a geographical area. Scand J Clin Lab Invest 1988;48:757-764. [ISI][Medline] [Order article via Infotrieve]
22. Ohman EM, Armstrong PW, Christenson RH, et al. Cardiac troponin T levels for risk stratification in acute myocardial ischemia. N Engl J Med 1996;335:1333-1341. [Abstract/Free Full Text]
23. Antman EM, Tanasijevic MJ, Thompson B, et al. Cardiac-specific troponin I levels to predict the risk of mortality in patients with acute coronary syndromes. N Engl J Med 1996;335:1342-1349. [Abstract/Free Full Text]
24. Burke GL, Edlavitch SA, Crow RS. The effects of diagnostic criteria on trends in coronary heart disease morbidity: the Minnesota Heart Survey. J Clin Epidemiol 1989;42:17-24. [ISI][Medline] [Order article via Infotrieve]
25. Gjørup T, Kelbæk H, Sørensen F, Stenbygård L. Reliability in the interpretation of the course of serum enzyme concentration changes in suspected acute myocardial infarction. Int J Cardiol 1992;35:43-47. [ISI][Medline] [Order article via Infotrieve]