Feature Story

Getting to the heart of homocysteine testing

February 2000
Karen Sandrick

In 1991, the Mayo Clinic performed a respectable number of homocysteine analyses-about 800. In 1998, the Rochester, Minn., facility completed an extraordinary number of homocysteine tests, upwards of 45,000, and in 2001, Mayo Clinic expects the total to increase another 18 percent.

The volume of homocysteine testing has grown steadily since the early ’90s when prospective trials and case-controlled studies started showing that this amino acid product of methionine demethylation was an independent risk factor for cardiovascular disease. "It’s pretty clear from most cross-sectional and case-controlled studies and some prospective studies that hyperhomocysteinemia increases the risk for not only myocardial infarction, but also coronary artery disease, peripheral vascular disease, stroke, cerebrovascular disease, restenosis of coronary arteries that have undergone balloon angioplasty, and even death from coronary artery disease," says Herbert Naito, PhD, chief, ancillary testing and satellite facilities, Louis Stokes Cleveland VA Medical Center (Arch Intern Med. 2000;160:422-434).

Now that laboratory methods make it comparatively easy to measure homocysteine, physicians are jumping on the homocysteine bandwagon-even though there has been no clear clinical proof that a decrease in homocysteine will protect an individual from a future cardiovascular event.

But clinicians are faced with many unanswered questions: Should homocysteine be used to screen general patient populations or only those predisposed to cardiovascular risk? Will homocysteine become another cholesterol? How should clinicians use the results of homocysteine testing to counsel patients?

Laboratory professionals, meanwhile, are wrestling with questions of their own: What is a normal homocysteine level? How do homocysteine tests compare in their analytical precision and bias? Is total homocysteine equal to reduced and oxidized homocysteine?

At the American Association for Clinical Chemistry meeting last July, Joseph McConnell, PhD, explained that in vitro studies suggest a causal relationship between elevated homocysteine and cardiovascular disease. Homocysteine oxidizes low-density lipoprotein, which may precipitate atherosclerosis and is toxic to the endothelium. Homocysteine also decreases the expression of thrombomodulin on the surface of endothelial cells. Thrombomodulin interacts with the protein C activation pathway, and its decrease results in inhibition of this anticoagulant pathway, potentially promoting thrombosis. Homocysteine activates platelets and increases platelet aggregation, says Dr. McConnell, director of the cardiovascular risk assessment laboratory and codirector of the biochemistry genetics laboratory at Mayo. Dr. McConnell adds that most in vitro studies have been performed at much higher than physiologic homocysteine concentrations, so the suggested causal relationship between homocysteine and cardiovascular disease implied from these studies should be viewed with a grain of salt.

In addition to causing cardiovascular disease by increasing thrombogenic events, hyperhomocysteinemia may trigger atherosclerotic events by enhancing smooth muscle cell proliferation, intimal-medial wall thickness, endothelial cell injury, thromboxane A2 formation, lipid abnormalities, LDL oxidation, and lipoprotein (a) binding to fibrin, says Dr. Naito (Ann Intern Med. 1999;131:363-375).

Clinical trials are beginning to suggest that lowering the homocysteine level is beneficial. A mild to moderate elevation in homocysteine can increase the risk of coronary artery disease by 2.5 times. Projections from a number of small studies suggest that lowering homocysteine by only 5 µmol/L may reduce the death rate from coronary artery disease by 10 percent, Dr. Naito says (JAMA. 1995;274:1049-1057). A drop in the plasma homocysteine concentration also may cut the rate of recurrence of restenosis of blood vessels. "So there’s enough evidence that hyperhomocysteinemia is an independent and dose-related risk factor, and it may even be a causal factor. It may be like glucose in that if it gets too high, it becomes a real toxin," Dr. Naito adds.

But how much homocysteine is too much? Most laboratories use 15 µmol/L as the cutoff point between normal and abnormal. A study conducted from 1991 to 1994 and published in the Annals of Internal Medicine pointed out, however, that the reference range may be much lower. This study is unique, says Dr. Naito, because the authors carefully selected their reference population to exclude those with established risk factors for hyperhomocysteinemia-that is, those with low vitamin B12 or folate intake, persons with renal dysfunction, and pregnant or postmenopausal women. According to the data, the reference range for men appears to be between 5 and 11.4 µmol/L; in women it ranges from 4 to 10.4 µmol/L. A finding above 11.4 µmol/L in men and 10.4 µmol/L in women can be considered abnormal (Ann Intern Med. 1999;131:331-339).

Determining the normal range for homocysteine has been complicated in the last few years by the addition of folic acid to bread and cereal products. Because folate fortification changes the normal reference values for homocysteine, it interferes with prospective studies of risk reduction and cardiovascular disease event outcomes, said Dr. McConnell in an interview with CAPTODAY. In the middle of these studies, he points out, the homocysteine level changed for the entire population-not just the patients taking vitamins.

Larry Brace, PhD, associate professor of pathology and director of coagulation services, University of Illinois, Chicago, adds that a direct relationship may exist between the risk of atherosclerosis and incrementally rising homocysteine levels. "If you could reduce your serum homocysteine level to near zero, that would present the lowest risk for atherosclerosis," he says. "But any incremental rise above that may carry some risk, even for someone with homocysteine in the normal range."

To think that homocysteine may be similar to cholesterol is tantalizing. In addition to two major clinical trials with hard cardiovascular endpoints published by 1988, as well as other trials examining angiographic evidence of coronary arteries, an explosion of more recent evidence confirms the relationship between cholesterol and coronary artery disease. "That is just not the case with homocysteine," says Andrew G. Bostom, MD, laboratory director, Memorial Hospital of Rhode Island, Pawtucket. "I’m not saying we won’t have those data, it’s just that we don’t have the data now, and it will probably be another five years to have comparable evidence to what we used to base the 1988 cholesterol education panel guidelines on."

As was done for cholesterol, a panel of experts would have to examine an abundance of data to decide what level represents a risk factor for CAD, Dr. Naito says. "How do we get a reliable figure that we can depend on?" he asks. "What are the variables, besides aging, gender, adequate vitamin B6, B12, folate intake, disease states, and medication interference, that can influence homocysteine concentration in the blood? In addition, what is the biological variation of homocysteine in the blood from day to day? What is the best method of treatment? And how do we know that if someone is below a certain level, they are at minimal risk?

"Another question is, should we screen the general public for hyperhomocysteinemia?" asks Dr. Naito. "Perhaps clinicians should agree that homocysteine should be analyzed in only target patient groups. Patients who have none of the conventional risk factors for cardiovascular disease but have clinical evidence of atherosclerosis are ideal candidates," he says. Other good candidates are patients who have a family history of premature coronary artery disease or early death due to heart disease and patients who have balloon angioplasty or carotid artery bypass grafts. Homocysteine levels are higher in older patients, particularly men, though women catch up after menopause (Atherosclerosis. 2000;149:163-168).

Diabetics are at greater risk for elevated homocysteine and cardiovascular consequences; just a 5-µmol/L increase in homocysteine has been associated with a 60 percent increase in cardiovascular disease mortality among diabetics, as compared to a 17 percent increase in those without diabetes mellitus (Circulation. 2000;101: 1506-1511). Homocysteine also is higher in patients with hyperthyroidism, says Dr. Naito (Ann Intern Med. 1999;131:348-351).

A recent study in Circulation (2000;102:605-610) suggested that if cardiac bypass patients with high baseline homocysteine levels are not treated within 2.5 years, they have a 2.6 times greater risk of recurring cardiac events, adds Dr. Naito. However, other evidence of a direct connection between a reduction in homocysteine and a drop in cardiovascular events is lacking, Dr. McConnell says.

So what should laboratory professionals tell clinicians who are ordering homocysteine tests for their patients? "Physicians in local hospitals who are potentially going to be using this methodology should be notified that homocysteine is a marker for cardiovascular disease, but efforts to reduce homocysteine may not affect patients’ risk for cardiovascular events," Dr. McConnell says.

Patients who already have other cardiovascular disease risk factors may benefit from knowing they have elevated homocysteine so they can be more aggressive about reducing their LDL cholesterol or maintaining a heart healthy diet, he adds.

For laboratory professionals, the question is which homocysteine test to use. Until recently, high-performance liquid chromatography was the most common method, and two HPLC methods came to the forefront, the primary difference being the fluorescent derivatizing agent used. Each method has its advantages and disadvantages. Monobromobimane derivatization occurs fairly rapidly, so it lends itself to automated analysis, Dr. McConnell says. Sodium borohydride typically is used to reduce homocysteine bound to protein or other thiols to produce free homocysteine. Derivatization with monobromobimane then occurs and takes only a few minutes. Monobromobimane obtained from commercial sources often has fluorescent impurities, which necessitates gradient elution to separate all other peaks from homocysteine, the peak of interest.

HPLC methods that use SBD-F (ammonium-7-fluorobenzo-2-oxa-1,3-diazole-4-sulphonate) or ABD-F (4 aminosulfonyl-7-fluoro-benzo-2-oxa -1,3-diazole) are not as easily adapted to automation, Dr. McConnell says. Reduction typically is done with tris-carboxethyl-phosphine (TCEP) and takes about 30 minutes. Proteins are precipitated with trichloroacetic acid. Derivatization takes a little longer with SBD-F-about 60 minutes at 60°C-than with monobromobimane, but SBD-F has fewer fluorescent impurities, "which allows for efficient separation of homocysteine from other compounds using isocratic HPLC elution methods," Dr. McConnell says. "Chromatographic times are typically less than five minutes," he adds.

Subject to debate is that some HPLC methods use internal standards, but others do not. "When choosing an internal standard, you would ideally want a substance that is not present in the body but that has characteristics similar to homocysteine," Dr. McConnell says. "There is always a question as to whether a molecule that is not homocysteine behaves exactly like homocysteine during both the reduction and derivatization processes." Of all the internal standards that have been used, he adds, cystamine appears to be becoming the internal standard of choice.

Also a problem for many laboratories is that they do not have the equipment, technical expertise, or time to adopt HPLC technology, says Christine Pfeiffer, PhD, chief, global micronutrient section, National Center for Environmental Health at the Centers for Disease Control and Prevention, Atlanta. (Drew Scientific, Portsmouth, RI, however, recently introduced an HPLC system, DS30, that does not require special technical expertise.)

Providing an alternative to HPLC methods are enzyme immunoassay kits. EIA techniques, which are commercially available in microtiter plate or automated formats, are based on the same principle: reduction of homocysteine and mixed disulfides, conversion of free homocysteine to S-adenosyl-L-homocysteine (SAH), and a competitive immunoassay-enzyme-linked immunosorbent assay or fluorescence polarization immunoassay for SAH.

The fully automated FPIA from Abbott Laboratories, Abbott Park, Ill., reduces disulfide forms of homocysteine with dithiothreitol, enzymatically converts homocysteine to SAH, and adds an anti-SAH antibody and a fluorescein tracer.

Use of such immunoassays is growing. In a CAP Survey of 289 laboratories that are conducting homocysteine analyses, the first such Survey conducted by the College, 80 percent were using the technique. FPIA compares favorably with HPLC.

Good agreement was noted among all laboratories in their analysis of three samples with varying concentrations of homocysteine, which were sent out by the College. The coefficient of variation was under 10 percent in most cases, reports Anthony A. Killeen, MD, associate director for clinical laboratories, University of Michigan Medical Center, Ann Arbor, and a member of the CAP Clinical Chemistry Resource Committee. Although the information comes from only one mailing, Dr. Killeen says, "The initial data look quite promising in terms of analytical performance of the methods that are in use by laboratories." Data from a second Survey will be available and analyzed soon, he adds.

Dr. Pfeiffer also found good comparability among HPLC, gas chromatography-mass spectrometry, and immunoassays. The mean interlaboratory variation and mean intralaboratory variation was less than 10 percent for 14 government, academic, and clinical laboratories (Clin Chem. 1999;45(8):1261-1268).

Her analysis and the results of an external quality assessment program from Denmark published a month later, nevertheless, saw room for improvement, she says. Although the interlaboratory variation was low, it might not be acceptable clinically because of the graded rise in risk for cardiovascular disease with increasing homocysteine, even within the normal range. Laboratories using the same method also sometimes varied more among themselves than laboratories that used different methods.

Dr. Pfeiffer concluded that a need existed to improve analytical precision so laboratories could use the same reference intervals. She also determined that it was necessary to evaluate the performance of individual laboratories with standard reference materials characterized by a reference method.

"Having standardized reference materials and a reference method will give the last necessary push to come to closer agreement among the methods and among laboratories," Dr. Pfeiffer says. "Having proficiency testing out there helps, because everyone can evaluate themselves over time and in comparison with other laboratories. But it’s not the same as having a reference material with a certified value and a reference method that establishes that value," she adds.

The CDC is optimizing a liquid chromatography-mass spectrometry technique as a potential reference method, but its development must be completed and it must be thoroughly evaluated, which will take several months, Dr. Pfeiffer says.

Scientists from the Mayo Clinic and elsewhere were to meet in December with representatives from the National Institute of Standards and Technology to begin discussions regarding preparation of standard reference materials for homocysteine, Dr. McConnell says. It may take at least a year before a standard reference material is approved, he predicts.

But all the proficiency testing in the world isn’t going to amount to a hill of beans if laboratories don’t pay attention to preanalytical variables. "The samples that are sent out by the College have stable levels of homocysteine and thus do not control for preanalytical variables," says Dr. McConnell. "If you are not taking care of preanalytical variables, you may be reporting inaccurate results despite acceptable proficiency testing results."

For laboratories that are beginning to develop homocysteine testing, control of preanalytical variables is crucial, Dr. McConnell adds, because homocysteine in whole blood is highly unstable. If a sample of blood is collected and left at room temperature, within an hour the homocysteine level can increase by 10 percent or more, he says. If left at room temperature for four to 24 hours, homocysteine levels can rise in whole blood samples by 35 to 75 percent. Samples, therefore, should be processed within an hour if they aren’t placed on ice, and they should be centrifuged within 15 minutes if at all possible, Dr. McConnell advises.

Standard operating procedure in many facilities, however, is for phlebotomists to collect eight or 10 samples at once. If a sample for homocysteine testing is first on the list, it may take an hour or more before it reaches the laboratory. Then it may sit on a bench for 30 to 45 minutes before a technologist places it in the centrifuge and spins it down.

Dr. McConnell recommends that laboratory professionals ask phlebotomists to place samples for homocysteine testing in ice immediately after they are drawn. If kept on ice at 2°C, homocysteine will not change significantly for about four hours.

Reports about the need to obtain fasting samples vary. One study involving only 13 patients found no difference in homocysteine levels among fasting and nonfasting patients. A second study showed that homocysteine levels were lower if samples were drawn two to four hours after breakfast; a third indicated that homocysteine levels increased six hours after a meal, Dr. McConnell says.

In the experience of Robert H. Williams, PhD, fasting is essential. Changes in homocysteine levels admittedly are transient after meals; eating initially will raise the methionine level and eventually alter homocysteine. The nature of a meal also influences homocysteine levels. "If you’re going to have eggs and bacon for breakfast, your methionine levels are going to be sky high. So if you’re trying to get a true baseline homocysteine level in a patient, you need to know they’ve been fasting," says Dr. Williams, director of general chemistry and toxicology and assistant professor of pathology, University of Illinois, Chicago.

New tests for homocysteine may soon be available to laboratories. Already at the Mayo Clinic, tandem mass spectrometry is used to measure homocysteine. The technique, which was developed at Mayo by Piero Rinaldo, MD, PhD, does not require derivatization. It takes a stable isotope of homocystine (homocysteine-homocysteine disulfide) and runs it as the internal standard. "The internal standard thus controls for differences in reduction as well as any other potential loss of homocysteine during sample preparation," Dr. McConnell says.

Other, more direct enzymatic-type methods are becoming available, but Dr. McConnell cautions laboratory professionals to make sure the methods have been fully validated and compared to established methods before bringing them onboard.

Also on the horizon is measurement of the multiple forms of homocysteine. The medical community doesn’t know precisely what each of the various forms of homocysteine does, says Dr. Williams, who, along with Dr. Brace, has been studying homocysteine in stroke patients. Drs. Brace and Williams have been examining the multiple forms of homocysteine because, "Most compounds, when they are bound to protein, are not biologically active, so the protein-bound portion of homocysteine probably is not responsible for any pathological effects," Dr. Williams says.

More than likely, the reduced form of homocysteine is the most active form of the compound because the sulfhydryl group is participating in the reactions, Dr. Williams explains. Still, trying to target the multiple forms of homocysteine has been difficult because these substances change quickly in blood after it has been removed from a patient. Measuring the multiple forms of homocysteine consequently is time-consuming and ineffective. It requires collecting blood in a heparinized tube that contains a derivatization agent or a specific trapping reagent to separate the different forms of homocysteine. This method traps and measures the reduced or oxidized free fractions of homocysteine, like homocystine and the mixed disulfide cysteine-homocysteine.

Alternatively, it involves collecting the sample in EDTA, chilling it immediately, rapidly spinning it down, and separating and acidifying the plasma-all within 15 minutes or less. "Most laboratories can’t do that realistically; it just doesn’t happen in the real world," Dr. Williams says.

But Dr. Williams has devised another way to test homocysteine fractions that is less labor intensive. His method is described in a paper that he has submitted to Clinical Chemistry. For now, therefore, Dr. Williams cannot elaborate. "Stay tuned for further developments," he says.

Karen Sandrick is a freelance writer in Chicago.