Assessment of Laboratory Tests for Plasma Homocysteine -- Selected Laboratories, July-September 1998
Cardiovascular disease, including coronary heart disease and stroke, is the leading cause of death in the United States. Elevated plasma homocysteine (Hcy), generally defined as fasting plasma Hcy levels greater than 15 µmol/L, is an independent risk factor for vascular diseases (1,2). It is unknown whether Hcy is a cause of or a marker for atherosclerosis. A recent statement by the Nutrition Committee of the American Heart Association concluded that until results of clinical trials are available, population-wide Hcy screening is not recommended (3). However, Hcy tests are used in the clinical setting and information on interlaboratory variation, on method variation, is limited. To assess the status of interlaboratory and intralaboratory variation for Hcy analysis, CDC conducted a study of selected laboratories during July-September 1998. This report summarizes findings from the study, which indicates a need to improve analytic precision and to decrease analytic differences among laboratories (4).
Fourteen laboratories participated in the study, including three manufacturers, two government, eight academic, and one clinical research laboratory. Each of three laboratories used two different methods. Selection of laboratories was based on the type of method used for Hcy testing: high-performance liquid chromatography (HPLC), gas chromatography-mass spectrometry (GC-MS), and immunoassay. Laboratories that used HPLC were further subdivided based on the type of detection each laboratory used (electrochemical or fluorescence) and by each of the types of reducing and labeling reagent each used to convert protein-bound and oxidized Hcy into free Hcy and to attach a fluorescent tag to the free Hcy for detection purposes. Laboratories using immunoassay were subdivided into two groups: those using fluorescence polarization immunoassay (FPIA) and those using enzyme immunoassay (EIA). Laboratories participated in a 2-day analysis of 46 blinded plasma samples, four blinded plasma samples with added Hcy, and three plasma quality-control (QC) pools. Interlaboratory and intralaboratory (i.e., between tests run in a laboratory) variation was expressed as a relative standard deviation*. In the absence of target values for the samples analyzed, the GC-MS method was considered arbitrarily as a reference method. Because it used stable-isotopically labeled Hcy as an internal standard, this method is considered to be the most accurate and precise assay available.
For all tests, the mean interlaboratory variation was 9.2% for plasma samples, 8.8% for plasma samples with added Hcy, and 7.6% for the QC pools (Table 1). The mean interlaboratory variation in each method group ranged from 3% to 13%. The group of laboratories performing the FPIA assay had the lowest interlaboratory variation (4.9% for plasma, 3.2% for plasma with added Hcy, and 3.2% for the QC pools). The mean intralaboratory variation was 5.6% for plasma samples, 4.9% for plasma samples with added Hcy, and 4.2% for the QC pools (Table 1). For most laboratories, the intralaboratory variation was less than 10% and the analytical recovery of added Hcy was 85%-115%.
Two of the HPLC methods (HPLC with electrochemical detection and HPLC with fluorescence detection using sodium borohydride as a reducing agent and monobromobimane as a labeling agent) and the EIA method produced results that were, on average, 7.5%, 8.1%, and 7.4% higher than GC-MS results. One HPLC method (HPLC with fluorescent detection using trialkylphosphine as a reducing agent and ABD-F as a labeling agent) produced results that were on average 16.1% lower than GC-MS results. The FPIA method and the two remaining HPLC methods (HPLC with fluorescence detection using either TCEP or TBP as a reducing agent and SBD-F as a labeling agent) showed no deviation compared with results of the GC-MS method.
Analytical quality specifications analysis was performed to test whether the precision** and bias*** of each method were satisfactory (5,6). On the basis of intralaboratory variations, none of the laboratories showed optimum performance for all three types of samples, two laboratories showed desirable performance, and six laboratories exceeded the requirements for minimum performance for at least two types of samples. Three methods performed best regarding analytical precision: GC-MS, FPIA, and HPLC with fluorometric detection using a water-soluble phosphine as reducing agent (TCEP) and SBD-F as fluorescent tag.
With regard to apparent analytical bias, nine laboratories met the requirements for optimum performance with respect to GC-MS, and two laboratories did not meet the requirements for minimum performance. The following three methods performed best regarding apparent analytical bias (versus GC-MS): FPIA, HPLC with fluorometric detection using water-soluble phosphine as reducing agent (TCEP) and SBD-F as fluorescent tag, and HPLC with fluorometric detection using the classical tri-butyl phosphine as reducing agent (TBP) and SBD-F as fluorescent tag.
Reported by: Div of Laboratory Sciences, National Center for Environmental Health, CDC.
The findings in this report indicate that the fully automated FPIA assay performed best with respect to lowest interlaboratory variation, analytical precision, and apparent analytical bias relative to a GC-MS method. Previous findings have shown a good agreement between the FPIA assay and an HPLC assay with internal standardization for approximately 800 serum and plasma samples (7).
Although both the mean interlaboratory variation and the mean intralaboratory variation were less than 10% in this study overall, this variation might be clinically unacceptable because of a graded increase in risk for vascular diseases with increasing plasma Hcy, starting at plasma Hcy concentrations well within the normal range of the population. The findings in this report also indicate that laboratories performing the same method sometimes vary more among themselves than laboratories performing different methods. The analysis also suggests that improvements are needed in the analytical precision to assure that laboratories in an area can use the same reference intervals. To aid this improvement, researchers need to evaluate individual laboratory performance through a program that includes standard reference materials and comparisons with other laboratories. Such an improvement is needed because Hcy has developed from an esoteric test to a clinical test.
The major limitations of this study were the small number of laboratories using each method and the arbitrary selection of the GC-MS method as a reference method, which may itself be biased. As a result, conclusions regarding interlaboratory variation within a method group and with respect to method-specific bias should be interpreted with caution. A high-quality reference method for Hcy is needed to better evaluate a laboratory's or a method's quality.
CDC is promoting efforts to develop a high-order reference method for plasma Hcy by tandem mass spectrometry, the state-of-the-art methodology for accuracy and precision. This method will be a standard for the development and characterization of reference materials. Future studies will assess the variability between laboratories before and after the introduction of standard reference materials.
* Relative standard deviation=standard deviation/mean x 100.
** Analytical precision is less than 0.25 x within-patient variability for optimum performance, less than 0.5 x within-patient variability for desirable performance, and less than 0.75 x within-patient variability for minimum performance.
*** The bias of a method is less than 0.125 x combination of within-patient and between-patient variability for optimum performance, less than 0.25 x combination of within-patient and between-patient variability for desirable performance, and less than 0.375 x combination of within-patient and between-patient variability for minimum performance.
Note: To print large tables and graphs users may have to change their printer settings to landscape and use a small font size.
TABLE 1. Mean interlaboratory and intralaboratory variations of homocysteine (Hcy) plasma samples from 14 laboratories, July-September 1998
* Relative standard deviation.
Return to top.
Disclaimer All MMWR HTML versions of articles are electronic conversions from ASCII text into HTML. This conversion may have resulted in character translation or format errors in the HTML version. Users should not rely on this HTML document, but are referred to the electronic PDF version and/or the original MMWR paper copy for the official text, figures, and tables. An original paper copy of this issue can be obtained from the Superintendent of Documents, U.S. Government Printing Office (GPO), Washington, DC 20402-9371; telephone: (202) 512-1800. Contact GPO for current prices.**Questions or messages regarding errors in formatting should be addressed to email@example.com.
Page converted: 11/10/1999
This page last reviewed 5/2/01