|
|
||||||||
Review |
1 Department of Pharmacology, University of Oxford, Oxford OX1 3QT, UK.
2 Department of Pharmacology and 3
Locus for Homocysteine and Related Vitamins, University of Bergen, 5021 Bergen, Norway.
4 Department of Clinical Biochemistry AKH, Aarhus University Hospital, DK-8000 Aarhus C, Denmark.
5 Clinical Trial Service Unit, Radcliffe Infirmary, University of Oxford, Oxford OX2 6HE, UK.
6 Departments of Biochemistry and Clinical Medicine, Trinity College, Dublin 2, Ireland
aAddress correspondence to this author at: Department of Pharmacology, University of Oxford, Mansfield Rd., Oxford OX1 3QT, UK. Fax 44-1865-27-1882; e-mail helga.refsum{at}pharmacology.oxford.ac.uk.
| Abstract |
|---|
|
|
|---|
Methods: Published data available on Medline were used as the basis for the recommendations. Drafts of the recommendations were critically discussed at meetings over a period of 3 years.
Outcome: This review is divided into two sections: (a) determination of homocysteine (methods and their performance, sample collection and handling, biological determinants, reference intervals, within-person variability, and methionine loading test); and (b) risk assessment and disease diagnosis (homocystinuria, folate and cobalamin deficiencies, cardiovascular disease, renal failure, psychiatric disorders and cognitive impairment, pregnancy complications and birth defects, and screening of elderly and newborns). Each of these subsections concludes with a separate series of recommendations to assist the clinician and the research scientist in making informed decisions. The review concludes with a list of unresolved questions.
| Introduction |
|---|
|
|
|---|
The introduction of tHcy assays in the mid-1980s (10)(11) started a new era of research on Hcy. However, it was the advent of immunoassays in the latter half of the 1990s (12)(13) that changed tHcy determinations from research tools to widely used clinical chemistry tests. As a result, interest in this field has increased exponentially in both routine diagnostics and research.
Although several reviews on tHcy determinations have been published (14)(15)(16)(17)(18), few have provided recommendations for their use in clinical practice (19)(20)(21)(22)(23). Our aim was to review the practical aspects of tHcy determinations in clinical practice as well as in the research setting and to survey the data on tHcy in diagnostics or as screening tests in several target populations.
Ideally, guidelines should be established by a multidisciplinary team including all relevant stakeholders, and the recommendations should be according to evidence-based medicine (24). There are not sufficient data in the Hcy field, however, to use such an approach. In particular, data from controlled clinical trials are sparse. Nevertheless, both clinicians and scientists need guidelines on how to use the tHcy assay and how to interpret the results. Accordingly, we have adopted an approach in which we base our recommendations on an expert opinion, using the evidence available and our long-term experience with tHcy measurements in research and in routine laboratory medicine. Each section provides a summary of the evidence for use of tHcy determinations in a particular setting and concludes with our personal recommendations. Wherever possible, the quality of the evidence available (Table 1
) is provided in the associated tables.
|
| Methods |
|---|
|
|
|---|
| Facts and Recommendations in tHcy Determinations and Assessment |
|---|
|
|
|---|
|
|
|
methods and their performance
Methods for tHcy were introduced in the mid-1980s (10)(11) and overcame the problems related to the presence of multiple unstable Hcy species in plasma. In all available assays, plasma or serum is initially treated with a reducing agent that converts all Hcy species into the reduced form, HcyH, which is measured either directly or after derivatization (Fig. 2
) (14).
For a thorough evaluation of assay design and performance, readers are referred to published reports (14)(15)(16)(17)(18). Briefly, the tHcy methods can be classified into two groups: chromatographic methods and enzyme and immunoassays (16)(17). The latter group measure only tHcy but are usually simple to perform. The fluorescence polarization immunoassay (FPIA), run on Abbotts IMx and AxSYM platforms (12)(26), is widely used in both research and routine laboratory settings. Chromatographic assays include amino acid analysis; HPLC with ultraviolet, fluorescence, or electrochemical detection; capillary electrophoresis with fluorescence detection; gas chromatographymass spectrometry (GC-MS); and liquid chromatography with tandem MS (MS-MS) (14)(17)(18). The chromatographic assays usually require skilled staff and are labor-intensive, and throughput may be low (14). The advantages of chromatographic assays include wide analytical range, simultaneous determination of other compounds (e.g., other sulfur amino acids and MMA), and sometimes lower cost than commercial reagent-based assays (14)(17).
The different tHcy methods give comparable results (16)(27), but the variations among methods and among laboratories are considerable (16)(28)(29)(30)(31). Ideally, from the known biological inter- and intraindividual variation in tHcy, the bias should be <10% (0.375 x CVbetween-person), and the imprecision no higher than 5% (0.75 x CVwithin-person), but many methods do not fulfill these criteria (29).
There is a need for standardization of tHcy assays (16). Certified reference material is lacking, which is a problem because different types of calibrator materials often yield different values (32). Hcy calibrators in water or plasma often have greater imprecision than plasma-based calibrators (29)(30). The inclusion of an internal calibrator improves performance for some (29)(33) but not all methods(34). The quality of tHcy measurements has improved in recent years, but the problem of standardization remains unresolved.
recommendations in the routine clinical setting
340 µmol/L).
recommendations in the research setting
sample collection and handling
Food intake and diurnal and seasonal variations.
A small meal will not influence tHcy concentrations in healthy people (35)(36), whereas intake of a large, protein-rich meal may increase the plasma tHcy concentration by
1015% after 68 h (36)(37). This may explain the diurnal variation, with tHcy concentrations being lowest in the first part of the day and highest in the evening (38). Plasma tHcy is probably not subject to seasonal variation (39)(40).
Posture during blood collection/venous stasis.
The tHcy concentration is not altered by the duration of venous stasis (15). Blood samples collected in the supine position have
10% lower mean tHcy concentrations than those collected in the sitting position (15), possibly because plasma albumin (which binds Hcy) is reduced in the supine position. This phenomenon may contribute to the lower tHcy concentrations observed in the acute phase after myocardial infarction compared with samples collected 23 months later (41)(42). Likewise, comparison of tHcy concentrations in patients confined to bed with those in healthy (sitting) controls may bias results of casecontrol studies.
Plasma vs serum.
Traditionally, it has been recommended that tHcy should be measured in plasma because the use of anticoagulant allows immediate sample processing. Serum, even if optimally prepared, yields slightly higher values than plasma (see below). Optimally collected EDTA or heparin plasma gives identical results, whereas citrated plasma yields 515% lower tHcy values (43). The influence of coagulant also depends on the method used. For example, EDTA improves the fluorescence yield in the monobromobimane assay (32) but is not compatible with some of the GC-MS methods (11).
tHcy export from blood cells and use of stabilizers.
After blood collection, but before removal of the blood cells, there is a time- and temperature-dependent increase in tHcy (Fig. 3
). This is attributable to an ongoing release of Hcy from erythrocytes (44). At room temperature, the increase in tHcy is
1 µmol · L-1 · h-1, but it is not dependent on the initial tHcy concentration (43). Hence, this corresponds to an
10% increase per hour in a typical sample with 10 µmol/L tHcy, but only 3% in a 30 µmol/L sample (43). Because serum is prepared from a blood sample left at room temperature for 3060 min to allow coagulation, serum concentrations will usually be
510% (
0.51 µmol/L) higher than those obtained in optimally prepared plasma (15).
|
The increase in tHcy is prevented by immediate centrifugation and removal of the blood cells or by keeping samples cooled on ice until centrifugation (15). The use of gel separator tubes that are rapidly centrifuged also prevents the increase of tHcy in serum for at least 48 h (45).
Several stabilizers may prevent the formation or release of Hcy from the blood cells (15)(46). Sodium fluoride appears to reduce the increase, but this is probably attributable to an osmotic effect on the red cells, diluting the plasma (46). Adenosine analogs, such as 3-deazaadenosine, are effective but not compatible with assays based on S-adenosylhomocysteine hydrolase, including the FPIA (47). Acidic citrate stabilizes the tHcy concentrations for a few hours, but its use usually leads to small but consistent changes in tHcy (46). Determination of tHcy in whole-blood lysates has been suggested as an alternative to plasma (48), but this will require a new set of reference intervals.
Overall, the problem with stabilization of tHcy in whole blood is only partly solved, and use of stabilizers usually leads to small but systematic deviations at baseline (46). Most experts therefore still recommend that plasma is prepared optimally with cooling or immediate centrifugation.
Stability of tHcy in stored plasma/serum.
After removal of the blood cells, tHcy in plasma or serum is stable. No changes are observed for at least 4 days at room temperature (43), for several weeks in the refrigerator, or for several years at -20 °C (15). Freezethaw cycles are usually tolerated well (15); however, after freezing, our experience is that inhomogeneity of the sample matrix is a common problem. Hence, thorough mixing of the samples is required after thawing.
Hemolysis.
Hemolysis may interfere with some tHcy assays but will usually not change the plasma tHcy concentrations per se. Although the tHcy concentration in erythrocytes is lower than in plasma (48), hemolysis has to be extensive to change the tHcy concentration even by a small percentage (43).
recommendations in the routine clinical setting
recommendations in the research setting
biological determinants
Determinants of plasma tHcy include genetic, physiologic, and lifestyle factors; various diseases (Table 3
); and drugs (Table 4
). Many of these factors cause a change in tHcy concentrations by altering the function or blood concentrations of the B vitamins, in particular folate and cobalamin, and/or by influencing renal function or, more rarely, by influencing enzyme activities. The causes of increased tHcy concentrations vary according to the age of the person and the degree of tHcy increase (Table 5
).
|
|
|
Low folate or cobalamin status or renal impairment account for the majority of cases with increased tHcy (49)(50)(51)(52)(53). In populations eating food fortified with folic acid, renal impairment and cobalamin deficiency are the most important determinants (54). Homozygosity for the methylenetetrahydrofolate reductase (MTHFR) 677C
T polymorphism is the most common genetic determinant (55)(56). Individuals with the MTHFR 677TT genotype usually have
2.5 µmol/L higher tHcy than those with the 677CC variant (57)(58), but it depends on the folate (59)(60) and riboflavin status (61)(62). Most other genetic polymorphisms in enzymes related to Hcy have little effect on tHcy concentrations (63).
recommendations in the routine clinical setting
recommendations in the research setting
T polymorphism), physiologic (age, sex, pregnancy, menopausal state, renal function), lifestyle (smoking, coffee intake, diet, drugs), and blood (folate, cobalamin, creatinine) determinants of tHcy.
reference intervals
In general, the reference intervals are calculated as the 2.5th97.5th percentile interval (or 95% reference interval) for presumed healthy individuals. The blood samples should be collected and handled under conditions reflecting usual clinical practice. Reference intervals may be established for different populations to account for important differences, including those related to nonmodifiable factors such as age, gender, or ethnicity, as well as modifiable factors, such as nutritional status, lifestyle, and disease. Some of these will be briefly mentioned below.
Age, gender, and pregnancy.
tHcy concentrations increase throughout life (Fig. 4
) and approximately double from childhood to old age. After puberty, males have higher mean tHcy concentrations than females. The gender difference in mean tHcy is
2 µmol/L, but it becomes less with increasing age (64)(65), and the proportion with tHcy above a given upper reference limit is similar in adult men and women. During pregnancy, both the mean concentrations and upper reference limits for tHcy are markedly lower, a finding that is only partly explained by the hemodilution and reduced plasma albumin during pregnancy (66)(67).
|
Renal function and creatinine.
tHcy is dependent on renal function and creatinine synthesis (68)(69)(70)(71). Usually, the reference limits for tHcy are calculated after excluding persons with increased creatinine or impaired renal function. Another possibility is to establish different reference limits for different creatinine concentrations, for example, by use of a nomogram. Such data are currently not available.
Nutritional status and lifestyle.
The marked effect of vitamin status on the reference intervals highlights the problem of defining "presumed healthy individuals". In most adults who do not eat food fortified with folic acid, the upper reference limit is 1520 µmol/L or even higher (64)(72)(73). However, in adults with good vitamin status or a healthy lifestyle, the upper reference limit is
12 µmol/L (Fig. 5
) (72)(74)(75). The introduction of folic acid fortification in the US has markedly reduced the prevalence of hyperhomocysteinemia (76). In adults eating a typical but nonfortified diet, the use of folic acid supplements or a healthy lifestyle will lower the mean tHcy concentrations by 1030%, and the upper reference limits will be reduced by a similar proportion (72)(77)(78).
|
Ethnicity.
Plasma tHcy concentrations differ among ethnic groups (64)(79), but the effect on the upper reference limit is relatively small between groups living in the same area and eating a similar diet (64). In some regions of the world, in particular in developing countries, tHcy concentrations may be very high in the general population. For example, in a group of presumed healthy Asian Indians, the 95th percentile was >50 µmol/L, a finding that was only partly explained by their low cobalamin status (80).
Evaluation.
There are two opposing needs: on the one hand, simplicity with few thresholds; on the other hand, the ability to identify those individuals with a truly abnormal value, taking all relevant factors into account. Age, pregnancy, and renal function are important (Table 6
). The intake of folic acid as either supplements or through fortification of foods must also be considered (Table 6
).
|
The statistically defined reference interval as discussed here may be different from the desirable tHcy concentrations (81). Because nearly everyone can lower their tHcy by increasing their vitamin intake, it has been suggested that the tHcy reference limits should be based on individuals who are vitamin replete (74)(75). However, it is still not known whether such low tHcy concentrations will have any beneficial effect on health. Hence, until more evidence is available, the conventional approach to distinguishing abnormal from normal, i.e., the statistical upper reference limit, should be used.
recommendations in the routine clinical setting
recommendations in the research setting
within-person variability
The within-person variability refers to the relationship between repeated testing on the same person at different time points. In the clinical routine, the within-person variability reflects the reliability of a single measurement, and it determines the magnitude of change between two measurements that is significant, for example, in response to treatment or disease. In research studies, the within-person variability has important implications for the power of the study.
The total within-person variability is the sum of within-person biological and analytical variability and is usually given as the within-person SD or CV. The within-person CV for tHcy is
8% in healthy adults over a 1-year period (29). Hence, in an individual with a true mean tHcy of 10 µmol/L, retesting would yield a value between 9.2 and 10.8 µmol/L in
70% and 8.4 and 11.6 µmol/L in 95% of the tests. The within-person variability for tHcy may seem substantial but is comparable to the variability for total cholesterol and systolic blood pressure (40).
Hyperhomocysteinemia is associated with higher within-person variability. In individuals with tHcy >40 µmol/L (56), the within-person variability is
25% after 48 months and
35% after
2 years. A similarly high variability has been observed for persons with cobalamin deficiency (82). Changes in factors known to influence tHcy (Tables 3
and 4
) are likely to increase variability.
To decide whether there is a significant change in the tHcy concentration between two measurements from the same individual, both the analytical CV and the biological within-person variability must be considered. This so-called critical difference [2.77 x (CVanalytical2 + CVbiological2)1/2] will be at least 20% (83) or, for most laboratories, substantially higher. In relation to tHcy measurements, the size of the analytical CV relative to the biological within-person CV is quite large (
50%). Hence, the magnitude of the critical difference can be reduced by simply testing more replicates of each sample (84).
In cross-sectional or retrospective studies, a single measurement of tHcy in each individual will typically underestimate the true strength of association of tHcy with disease by
15% (40). In prospective studies, a single tHcy measurement at enrollment may not reflect the individuals long-term mean tHcy concentration, a phenomenon referred to as the regression dilution bias. The magnitude of this bias is greater in studies with longer follow-up intervals. Failure to correct for regression dilution may underestimate the relative risks of disease by
20% after 2 years and 50% after 10 years (85).
recommendations in the routine clinical setting
recommendations in the research setting
the methionine loading test
The methionine loading test was originally developed to diagnose heterozygosity for cystathionine ß-synthase (CBS) deficiency (86). Currently, the test is used to identify individuals with mild impairment of Hcy metabolism, in particular in CVD patients, in whom fasting tHcy may be normal but the postload tHcy concentration is increased (87)(88).
Methionine loading involves intake of 100 mg methionine/kg of body weight and measurement of tHcy usually 46 h after methionine ingestion (3). A 2-h postmethionine tHcy has been validated and may be more practical in the clinical setting (89), but it probably has greater within-person variability than the 4- or 6-h test (90). There are few side effects of methionine loading (91), but one possible death from severe overdose was reported recently (92).
In adults, the mean post-methionine-load tHcy measured at 4 or 6 h is
30 µmol/L, i.e., a 20 µmol/L increase above the fasting value, or 3 times the fasting value (3)(83)(93). The upper reference limit is
5 times the fasting tHcy concentration (Table 6
). Several factors are associated with increased post-methionine-load tHcy concentrations, including higher age; male sex; impaired renal function; low concentrations of folate, cobalamin, and vitamin B6; the MTHFR 677TT genotype; and heterozygosity for CBS deficiency (90)(94)(95)(96)(97)(98). However, with the exception of impaired B6 function and heterozygosity for CBS deficiency (95)(97), most of these factors increase both the pre- and postload tHcy concentrations. For example, in the European COMAC cohort, (3)(88), among the controls with postload tHcy >50 µmol/L (
95th percentile), only 5% had a fasting tHcy <10 µmol/L (approximately the median). Hence, the majority of individuals with increased post-methionine-load tHcy have fasting tHcy concentrations either above normal or in the high normal range.
recommendations in the routine clinical setting
recommendations in the research setting
| Facts and Recommendations for the Use of tHcy in Diagnosis and Risk Assessment |
|---|
|
|
|---|
|
Each section below will have the following plan. The facts will first be reviewed under the headings "target populations", "background for using tHcy", and "use of tHcy in risk assessment/diagnosis". We will then give recommendations separately for the routine clinical and research setting. For all sections, the relevant information related to tHcy determinations and assessment in both the clinical and the research setting is summarized in Table 2
.
homocystinuria
Target populations.
The target populations include patients, in particular children and young adults, with symptoms of homocystinuria, including thromboembolism, lens dislocation, progressive myopia, osteoporosis, marfan-like appearance, unexplained mental retardation, psychiatric disorders, or megaloblastic anemia, as well as siblings or children of patients with homocystinuria.
Background.
Homocystinuria refers to the rare inborn errors of metabolism leading to urinary excretion of large amounts of homocystine combined with severely increased plasma tHcy concentrations, usually >100 µmol/L (25). The most common cause is CBS deficiency, but impaired Hcy remethylation attributable to defects in methionine synthase, in MTHFR, or in factors or enzymes involved in the transport or metabolism of cobalamin may also occur (99). Independent of site of defect, these patients have a high risk of premature, frequently fatal, thromboembolic events (70).
CBS deficiency is an autosomal recessive disease. The reported worldwide birth prevalence is
1 in 300 000, but the prevalence is higher in Ireland and New South Wales (70)(100). Data from Scandinavia based on genetic analyses in newborns (101) and extensive tHcy testing of the population show that CBS deficiency may be much more common, i.e., at least 1 in 20 000 live-born children.
CBS deficiency is divided into pyridoxine-responsive and -nonresponsive variants, depending on the effect of pyridoxine on the tHcy concentration (70). Patients with pyridoxine-nonresponsive CBS deficiency often have a severe clinical phenotype with early symptom debut, but both the responsive and nonresponsive groups may suffer from serious complications (70). Lens dislocation and myopia are usually the earliest signs, but some patients experience few typical symptoms and are first diagnosed after a CVD event (102). Early diagnosis and treatment with pyridoxine and/or folic acid and betaine, preferably from infancy, can prevent CVD events and most of the clinical symptoms (100)(103). The beneficial effect of tHcy-reducing therapy is independent of pyridoxine responsiveness (100).
Most individuals heterozygous for CBS deficiency have normal fasting tHcy, but their urinary tHcy concentrations may be increased (97), and some, but not all, respond to a methionine loading with an abnormal increase in tHcy (97)(104). Increased postload tHcy concentrations are a risk factor for vascular disease (88) and neural tube defects (NTDs) (105), but heterozygosity for CBS deficiency is not a frequent finding in these conditions (106)(107), and heterozygous individuals do not seem to be at increased CVD risk (70).
Compared with CBS deficiency, less is known about the inborn errors leading to impaired Hcy remethylation. These defects usually, but not always, become clinically apparent early in life and are associated with developmental delay, failure to thrive, megaloblastic anemia, and myelopathy (99). In MTHFR deficiency, megaloblastic anemia does not occur, and failure to thrive is the most common sign (99). The phenotypic expression of all variants of homocystinuria may vary substantially, even within the same family (99). The effect of treatment with cobalamin, folic acid, and/or betaine is variable, but may be successful if the treatment is initiated early (99).
tHcy in diagnosis of homocystinuria.
Except in close family members of a homocystinuria patient, hyperhomocysteinemia is rarely explained by the genetic defects causing homocystinuria. However, in individuals with increased tHcy >100 µmol/L, homocystinuria should be suspected. Milder variants may produce lower tHcy concentrations. High methionine (1050 times normal) combined with low cystathionine is indicative of CBS deficiency; low or normal methionine combined with increased cystathionine suggests a Hcy remethylation defect (70). Those with MTHFR deficiency have red cell folate concentrations within reference values, whereas those with methionine synthase deficiency or defects in intracellular cobalamin metabolism have low red cell folate and a megaloblastic anemia (99). Marked increases in both tHcy and MMA are found with some defects in cobalamin transport and metabolism (99).
recommendations in the routine clinical setting
recommendations in the research setting
folate and cobalamin deficiencies
Target populations.
The target population includes individuals with clinical symptoms suggestive of folate or cobalamin deficiency and individuals who are at risk of developing a deficiency (Table 7
and Fig. 6
).
|
Background.
Folate deficiency occurs at all ages and is usually a result of poor diet, malabsorption, alcoholism, or use of certain drugs (108). It is common during pregnancy (108). The prevalence in US adults was
20% before folic acid fortification, but it is now <2% (76). Cobalamin deficiency is most often observed in the elderly (prevalence
1015%), where it is nearly always attributable to malabsorption caused by lack of intrinsic factor (pernicious anemia), gastric atrophy, or ileal disease (109). Newborns frequently have low cobalamin (see below) (110). Age-independent causes include inadequate intake (e.g., vegetarians) or use of certain drugs (Table 4
).
Diagnosis of folate or cobalamin deficiency is not straightforward. The definition of deficiency is debated, and there are no gold standards for diagnosis (2)(108)(109). Megaloblastic changes are a late event in the disease process (111), and neuropsychiatric symptoms may be present without hematologic findings (109)(112). Furthermore, chronic, low-normal vitamin concentrations are related to serious conditions such as CVD, cancer, pregnancy complications, birth defects, psychiatric disorders, and cognitive impairment (113)(114)(115). Hence, the aim is increasingly for early diagnosis in a preclinical state.
The basis for use of tHcy and MMA measurements in assessment of vitamin status is because these metabolites begin to increase sharply at low-normal concentrations of cobalamin (MMA and tHcy) and folate (tHcy only; Fig. 7
) (50)(116)(117)(118)(119). Increased MMA or tHcy most frequently results from impaired vitamin function, and the metabolite concentrations will usually return to normal within 1 or 2 weeks after appropriate vitamin treatment (82)(120)(121).
|
Comparison of blood tests.
No single blood test used for assessment of vitamin status provides certain diagnosis, but each has advantages and limitations (2)(121). Below we discuss folate, cobalamin, MMA, and tHcy measurements. For information on other tests used in the diagnosis of vitamin deficiency (e.g., blood and bone marrow examinations, serum gastrin, autoantibodies, Schilling test, and deoxyuridine suppression test), readers are referred to specialized reviews (2)(121).
Blood vitamin measurements are established and simple to perform, but are increasingly considered to be neither sufficiently sensitive nor specific (2)(122)(123)(124)(125). There is limited consensus on reference limits and desirable concentrations (108)(109). In addition, folate measurements, in particular red cell folate, are hampered by low analytical performance (2). These factors may explain why up to 50% of the individuals with low vitamin concentrations have no other biochemical or clinical finding and why patients with low-normal vitamin concentrations may have symptoms that respond to vitamin treatment (108)(109)(121)(126)(127). New assays for holotranscobalamin (holoTC) have now become available (128)(129). Low holoTC concentrations could be an early marker of cobalamin deficiency (111); recent studies show promising results (130)(131)(132)(133), but the clinical utility of this test remains to be documented.
Among the metabolites, MMA is often deemed superior to tHcy in relation to cobalamin deficiency (134)(135) because it is more specific and less susceptible to preanalytical errors than tHcy. In a person with normal renal function and low cobalamin, MMA concentrations above the threshold are often seen as proof of disturbed cobalamin function (1)(118). A disadvantage, however, is that MMA measurements are expensive and not widely available (2). Some data suggest that increased MMA predicts neither symptoms of deficiency nor response to cobalamin treatment (136)(137)(138).
The argument against the use of tHcy measurements is that it is not a specific test because the concentration is influenced by numerous factors (Tables 3
and 4
). However, most of these factors change tHcy concentrations through their effect on folate or cobalamin status. In addition, a large proportion of folate and cobalamin measurements are used as screening tests or in relation to nonspecific symptoms that can be explained by a deficiency of either vitamin (118)(138). Thus, as a screening test, tHcy is probably superior to MMA because tHcy measurements are less expensive, widely available, and reflect both cobalamin and folate status.
The diagnostic utility of tHcy and MMA in vitamin assessment is well documented, and interested readers are referred to published reports (1)(49)(82)(118)(120)(135) and reviews (108)(109)(121). Unfortunately, few studies have actually compared the sensitivity or specificity of these metabolites with folate and cobalamin measurements in relation to clinical outcome. With the lack of gold standards for diagnosis, optimal design of such studies requires that the inclusion criterion should be a finding, symptom, or disease suggesting vitamin deficiency but identified by means other than vitamin or metabolite measurements. An alternative approach is to use a large cohort with sufficient clinical or hematologic data (see Fig. 8
).
|
tHcy in assessment of vitamin status.
Despite the limitations of the published studies, the evidence suggests that an increased tHcy or MMA combined with a low vitamin concentration is better than either alone to identify individuals with symptoms (Fig. 8
) (1)(53)(54)(125) and those who would benefit from treatment (Fig. 9
) (49)(82)(127)(139). High tHcy or MMA concentrations may also help to identify individuals at risk of developing adverse effects from drugs that interfere with folate or cobalamin status (140)(141)(142).
|
tHcy measurements are increasingly being used to screen for vitamin deficiency in the general or high-risk populations (53)(118)(121)(126)(138). Such studies have used tHcy thresholds consistent with the upper reference limits listed in Table 6
, i.e., corresponding to the 95th97.5th percentiles in a presumed healthy population. The vitamin thresholds used for screening are usually 200250 pmol/L for cobalamin (2)(118)(126) and 710 nmol/L for serum folate (50)(53)(108). These thresholds fit with the concentrations at which tHcy begins to increase steeply (Fig. 7
), but are substantially higher than the statistical lower reference limits for cobalamin and for folate in nonfortified populations (108)(109).
It is usually suggested that vitamin measurements should precede tHcy (or MMA) determination (2)(118). However, in patients with no or only diffuse symptoms, tHcy measurements are probably superior primary screening tests (Figs. 8
and 9
) (53)(126)(138).
recommendations in the routine clinical setting
recommendations in the research setting
cvd
Target populations.
The target populations include individuals with established, or at high-risk of developing, arterial or venous occlusive vascular disease.
Background.
Homocystinuria, irrespective of the enzymatic defect, leads to severely increased tHcy and a high incidence of arterial and venous thromboembolic events (70)(99). These patients also have vascular changes, and this observation led McCully in 1969 (143) to propose his Hcy theory of atherosclerosis in the general population. Since then, the association between tHcy and CVD has been documented in many epidemiologic studies (for details see Refs. (4)(144)(145). Moderately increased tHcy is related to both venous and arterial occlusive disease (3). The relationship between tHcy and CVD is dose dependent and independent of other risk factors (3)(4). Data from knock-out mice support the view that increased Hcy causes vascular disease (146)(147). In humans, the deleterious effects of Hcy on endothelial and vascular function and blood coagulation provide pathophysiologic explanations for the increased CVD risk in hyperhomocysteinemia (3)(148)(149).
Nevertheless, for three reasons there has been debate about the causal role of tHcy, and, hence, its clinical significance in CVD (150): (a) the MTHFR 677C
T polymorphism, which is a strong risk factor for increased tHcy but not for CVD; (b) the apparent discrepancy between prospective and retrospective casecontrol studies; and (c) the lack of data from controlled clinical trials.
The apparent lack of MTHFR 677C
T effect (57) is related to the low power of most studies (150). In a recent metaanalysis including 11 000 CVD cases and 13 000 controls, the TT genotype was associated with a 16% increase in CVD risk (58). This risk enhancement corresponds to the extent of tHcy increase observed in individuals with the TT genotype (4)(58)(150). Thus, the results for individuals with the MTHFR 677TT genotype support the hypothesis that Hcy is causally related to CVD (58).
The discrepancy between prospective and retrospective casecontrol studies (4)(144)(151) arises in part because most prospective studies start with a healthy population. In such studies, tHcy is a modest predictor of CVD events. In two recent large-scale metaanalyses, a 25% (3 µmol/L) reduction in tHcy in population-based cohort studies was associated with a 1116% decrease in the risk of ischemic heart disease, 1922% decrease in the risk of stroke and 25% decrease in the risk of deep venous thrombosis (4)(145). In contrast, increased tHcy is a strong risk factor for CVD events and mortality in patients with coronary artery disease (Fig. 10
), diabetes, renal failure, and systemic lupus erythematosus (152)(153)(154)(155)(156). Hence, irrespective of causality, increased tHcy is a prognostic factor in certain groups with a high CVD risk profile.
|
Results from intervention trials with B vitamins are sparse. Some studies show that tHcy-lowering therapy slows the progression of coronary and peripheral atherosclerosis (157)(158)(159). However, a recent study found that patients receiving folic acid, cobalamin, and vitamin B6 in combination had increased risk of in-stent restenosis (160). Many of the ongoing trials will lack power as a result of the folic acid fortification program (161). Thus, we may only know the public health implications of B-vitamin intervention by performing a metaanalysis of results from ongoing CVD trials.
tHcy in CVD risk assessment.
Epidemiologic data suggest that tHcy screening in the general population is not justified (4)(144)(162). However, in high-risk populations, increased tHcy is a prognostic marker for an increased risk of new CVD events or mortality (150). Whether this reflects a causal role of tHcy may be answered by the ongoing clinical trials (163). The outcome of the trials will certainly influence recommendations for the use of tHcy in CVD risk assessment.
The question about desirable tHcy concentrations is being debated. There is a doseresponse relationship between tHcy and CVD risk from low concentrations up to at least 20 µmol/L (3)(152)(164). Some experts suggest that tHcy should be <10 µmol/L (20)(81). With such a cutoff, 3050% of the general population would be defined as "hyperhomocysteinemic"; this proportion would be higher among CVD patients and the elderly. A more realistic target can be proposed from prospective studies; a significant and substantial increase in CVD risk is observed above 1315 µmol/L (144)(152)(156)(164)(165)(166)(167). Firm recommendations about desirable tHcy concentrations will be available only after the results of clinical trials are reported and may depend on whether clear thresholds for risk reduction after vitamin intervention are demonstrated.
Overall, the use of tHcy in CVD risk assessment is controversial in relation to both how and when to use the test and what to do about the result. Within our group, we were unable to reach a consensus, and the recommendations listed below for the clinical setting reflect the majority view.
recommendations in the routine clinical setting