Diagnostics & Imaging Week Washington Editor
WASHINGTON — This year's annual meeting of the American Association for Clinical Chemistry (AACC; Washington) demonstrates that the clinical lab professional is no more wedded to the tried and true than any other healthcare professional. One of Sunday's presenters discussed several sources of inaccuracy in cholesterol measurements, which he blamed on the reagents used in assays that are commonly employed to determine a patient's risk of cardiovascular disease. Another presenter chipped away at the role of C-reactive protein (CRP) in prediction of heart disease and stroke risk by presenting data in support of lipoprotein-associated phospholipase A2 (Lp-PLA2) as a superior biomarker.
In a discussion titled "New Tests for Traditional Markers of Cardiovascular Disease," Greg Miller, MD, a professor of pathology at Virginia Commonwealth University (VCU; Richmond, Virginia), said that among the measurement issues facing clinical lab professionals is that of how to evaluate the performance of the various assays and their reagents for homogeneous high-density lipoprotein-C (HDL-C) and its low-density lipoprotein counterpart, LDL-C. "Even though these have been around for a decade, there are still important issues," he said.
Miller reminded his audience that the standards published in the Adult Treatment Panel III (ATP III) by the National Heart, Lung and Blood Institute "are the current clinical practice guidelines" but said ATP IV is in development. In ATP III, "the emphasis is on multiple risk factors," including lipoproteins and trigylcerides, and ATP III addresses "treatment beyond LDL-C lowering for persons with trigylcerides beyond 200 milligrams per deciliter (mg/dl)" in addition to high cholesterol counts.
He described the relationship between cardiovascular risk and lipoprotein counts as anything but linear, and "as research progresses we continue to learn more and more" about that relationship.
"The clinical goal of a standard measurement program is to get comparable results in all labs," Miller said. The current gold standard for measurement of laboratory references is beta quantification, a method that starts blood samples off with an 18-hour trip through a centrifuge (ultracentrifugation), which isolates very low-density lipoproteins (VLDL) from LDLs and HDLs.
He remarked that the cut-off weight for VLDLs is 1.006 kilograms per liter, but cautioned attendees that "we need to remember that each of these so-called lipid fractions contain a lot of sub-particles." Among the LDL fractions he referred to are remnants of chylomicron LDL cells, which are fairly large LDLs that have secreted their triglyceride contents.
Another complication is the fact that a lipoprotein "that is normally a minor component may be present in relatively high concentrations" in some patients, Miller stated, making benchmarks for these assays an all-the-more urgent need. Hence, laboratory professionals need to scrutinize carefully the results of a beta quantification effort to validate an assay's reliability.
Unfortunately for lab operators, reagents and analyzers do not offer a mix-n-match situation. "You cannot mix calibrators and reagents from different manufacturers. It has to be used as a package deal," he said, so any benchmarking of an analyzer and reagent pairing has to be individually traced back to the standards set by the network reference labs at the Centers for Disease Control and Prevention (CDC; Atlanta). Any lab operator that attempts to calibrate equipment by using the published literature faces a possible headache imposed by the fact that "the literature often uses modifications to beta quantification."
Add in interferences such as bilirubin and ascorbate, and the task becomes more onerous. Miller said that because bilirubin is a marker for liver disease, the fact that it "can cause a spectrophotometric effect" in the presence of hydrogen peroxide creates the possibility that liver disease may go undetected.
He noted that the next generation of reagents is under development, referring to abstracts that address reagents made by Diazyme (Poway, California) and Roche Diagnostics (Basel, Switzerland), among others. These, he said, should "reduce interferences from other proteins, particularly hemoglobins."
In a side note regarding payment policy, Miller noted that the Friedewald measurement for LDL, which sets the threshold for reimbursement for trigylcerides at 400 mg/dl, "may be too high based on scientific evidence."
He referred to a 2004 paper by Andre Tremblay, PhD, and five others at the University of Ottawa (Ottawa, Ontario) that compared the Friedewald method with ultracentrifugation, concluding that the overestimation of the standard deviation for cholesterol runs to almost 8% when trigylcerides come in at 197 mg/dl.
This figure jumps to 18.4% when trigylcerides register at 302 mg/dl, and the effect is even more pronounced at trigylcerides measurements of 426 mg/dl, which generates a standard deviation bias of almost 28%.
Gary Myers, PhD, chief of clinical chemistry in the Division of Laboratory Sciences at the CDC, gave attendees a closer look at the contrast between C-reactive protein (CRP) and lipoprotein-associated phospholipase-2 (Lp-PLA2) as predictors of stroke and heart disease. Myers' presentation made clear that he, at least, sees the latter as closing the gap with the former.
He reminded attendees that CDC and the American Heart Association (AHA; Dallas) published a 2003 statement on the public health implications of CRP as a measure of risk for stroke, but he also presented data from more than 25 studies that indicated a disturbingly wide range of relative risk, from just more than 1.0 to almost 4.5 in patients who had elevated levels of CRP, but who were free of any diagnosed disease.
Myers also alluded to the larger risk calculation dilemma by stating that "there are probably more than 100 biomarkers" associated with coronary heart disease, but noting that a retrospective analysis of more than 120,000 patients in a number of nations indicated that "about 62% of those who had coronary heart disease had only one or none of these risk factors," referring to smoking, hypertension, diabetes and high cholesterol.
This information came from an article appearing in the Aug. 20, 2003, edition of the Journal of the American Medical Association, authored by a team of nine led by Umesh Khot, MD, of Indiana Heart Physicians (Indianapolis).
Myers characterized Lp-PLA2 as "sort of the new kid on the block," but said, "it has started to catch people's attention." This molecule is a calcium independent member of the phospholipase family, and is produced by monocytes, macrophages and t-lymphocytes, and "is finding more applications as a risk factor for ischemic stroke."
He presented an animated video that showed a molecule with Lp-PLA2 attached slipping from inside a blood vessel into the wall of the artery. Once the Lp-PLA2 penetrated the wall, a macrophage invades and absorbs the Lp-PLA2-bearing cells. The used-up macrophage then converts to a foam cell, which triggers inflammation.
"We know cholesterol is not a very good predictor of stroke," Myers said, citing six studies that demonstrated no statistically significant correlation, but he said that among others, the Rotterdam Study and the Malmo Diet & Cancer Study "demonstrate substantial risk" associated with Lp-PLA2.
He also cited the Atherosclerosis Risk in Communities (ARiC) study, which he said showed that high levels of Lp-PLA2 combined with sufficiently high levels of hypertension "increased your risk [of ischemic stroke] by almost seven-fold." Still, screening with this test is not yet recommended.
As for the performance of current assays for Lp-PLA2, Myers discussed the PLAC, an enzyme-linked immunosorbent assay (ELISA) that detects the mass of the analyte rather than its activity, and does not cross-react with other phospholipidases. However, this assay is classified as highly complex under the Clinical Labs Improvement Amendments of 1988 (CLIA), whereas the AutoPLAC, designed to run on automated clinical chemistry systems, falls into the moderate complexity category under CLIA.
The AutoPLAC "has been validated on Roche, Hitachi and Olympus systems," he said, hinting that this technology is ready to go. The maker of the PLAC, diaDexus (South San Francisco, California), obtained a 510(k) for the assay earlier this year.