Cobalamin in the human diet is a bacterial product ingested and stored by animals and strict vegans are therefore liable to deficiency. The prevalence of cobalamin deficiency, as defined by serum vitamin B₁₂ <200 ng/l and methylmalonic acid >0.27 μmol/l, is 1.6% of subjects over 51 years of age in the 2001–2004 National Health and Nutrition Examination Survey ¹ in the USA. Of this age group, 3.2% have a low serum B₁₂ <200 ng/l. Ingested cobalamin is released from food proteins by pepsin and acid and bound initially by transcobalamin I or haptocorrin (R binder). This binder also binds other cobinamides in the diet which are metabolically inert. Pancreatic enzymes release the cobalamin from transcobalamin I and permit binding by intrinsic factor. The intrinsic factor–cobalamin complex is attached to cubam, a multiligand receptor, which is a combination of cubilin² and amnionless and is taken up by endocytosis, into the ileal cell. The cobalamin is then released from the endosome and bound to transcobalamin II (holotranscobalamin) in the ileal cell and exported into the portal circulation. Cobalamin undergoes enterohepatic circulation via the liver and bile ducts with 1.4 μg/day excreted in the bile, of which 1 μg/day is reabsorbed in the ileum. Holotranscobalamin is the active form of cobalamin and is taken up by holotranscobalamin receptors (TCII receptors) on cells throughout the body, particularly liver, kidney and bone marrow cells. At the cellular level in the target tissue the holotranscobalamin undergoes endocytosis via the transmembrane TCII receptor. Holotranscobalamin then undergoes lysosomal degradation, releasing cobalamin for metabolic reactions.

Cobalamin is a cofactor in two important biochemical reactions. In the first, methylcobalamin acts as a cofactor for methionine synthase in the production of methionine from homocysteine. The remethylation of cobalamin requires the donation of the methyl group from methyltetrahydrofolate as it is converted to tetrahydrofolate, thus linking cobalamin to folate and 1-carbon metabolism. The second cobalamin reaction occurs in the mitochondrion. Cobalamin is converted to adenosylcobalamin, a cofactor for the enzyme methylmalonyl-CoA mutase, which converts methylmalonyl-CoA (the product of propionate metabolism) to succinyl-CoA. Methionine produced in the first reaction is converted to adenosylmethionine and is a vital source of methyl groups critical for a series of methylation reactions involving proteins, phospholipids, neurotransmitters, RNA and DNA. In cobalamin deficiency, methylmalonic acid and homocysteine levels are therefore elevated. Reduced methionine synthesis is thought to result in a decrease in methylation of myelin basic protein, resulting in the neuropathies associated with cobalamin deficiency which are irreversible once subacute combined degeneration ³ of the cord has occurred.

Folate absorption and metabolism

Dietary folate polyglutamates are thermolabile, water-soluble vitamins found in leafy green vegetables. Folate deficiency arises from dietary deficiency, impaired absorption or increased requirements. Folate polyglutamates in the diet must be hydrolysed to monoglutamates by hydrolases, operating maximally at pH 5.5 in the presence of zinc and further converted to pteroylglutamate, before absorption can take place. Folate carriers transport pteroylpolyglutamate rapidly into the luminal cells where it is methylated using methyl cobalamin as cofactor and reduced to 5-methyltetrahydrofolate (5-methyl THF) in the enterocyte before entering the portal venous system. Unconverted pteroylglutamate remains in the luminal cells.

Just like cobalamin, there is significant enterohepatic recirculation of folate, amounting to 90 μg/day, and biliary drainage results in rapid fall in serum folate levels, whereas deprivation of dietary folate takes up to 3 weeks to cause serum levels to fall. Two-thirds of plasma folate is non-specifically bound to plasma folate binding proteins including albumen and one-third circulates as free folate.

There is sufficient retention of folate by the renal tubules to prevent urinary vitamin loss; this is achieved by megalin uptake of filtered folate-binding protein ⁴ and bound folate. Cubam,² which binds intrinsic factor-cobalamin complex, is also important in the uptake of albumen from the renal tubules, which may also contribute to folate retention.

Folate transport into cells is dependent upon two mechanisms, reduced folate carrier (58 kDa), which is a low-affinity high-capacity system, and folate receptors (44 kDa), of which there are three isoforms – alpha and beta are attached to the cell surface through a glycosyl-phosphatidylinositol anchor and gamma is secreted by enteric mucosal cells. Methyltetrahydrofolate bound to the folate receptor undergoes endocytosis by clathrin-coated pits or caveolae. Antimetabolites such as methotrexate also bind to the reduced folate carrier and folate receptor. Passive diffusion is an alternative mechanism by which folate can enter cells. The relative contributions of the different mechanisms are not known. Folate receptors may be expressed on malignant cells and have become potential targets for delivery of cytotoxic agents linked to folate.

Folates participate in 1-C metabolism and thereby facilitate the essential cellular metabolism of methionine, serine, glycine, choline and histidine in the biosynthesis of purine and deoxythymidine monophosphate (dTMP) in the synthesis of pyrimidines and thus DNA (Fig. 10.1).

Figure 10.1 Metabolic diagram. Note: Cobalamin bound to transcobalamin II is in its most oxidized 3+ state and enters the cell by TCII receptor, it undergoes one reductase step to cobalamin (I) in the cytoplasm and a second reductase step to cobalamin (II) in the mitochondria. The methionine cycle takes place in the cytoplasm and the succinyl-CoA synthesis in the mitochondria.

Intracellular folates are compartmentalized between the cytosol and mitochondria and the major forms are tetrahydrofolate (THF), 5-methyl THF and 10-formyltetrahydrofolate (10-formyl THF). Homocysteine is converted to methionine by methionine synthetase using cobalamin as a cofactor and methyl THF as the methyl group donor. Cobalamin deficiency therefore results in inactivation of methionine synthase, resulting in accumulation of 5-methyl THF, which cannot be converted back to 5,10-methylene THF. Folate is then unavailable for pyrimidine and purine synthesis – the methyl-trap hypothesis, which was advanced to explain why cobalamin deficiency often results in a functional folate deficiency. Furthermore, methyl THF is a very poor substrate for the enzyme responsible for folate polyglutamation, folylpoly-γ-glutamate synthetase, which prefers THF and 10-formyl THF. Folate deficiency is thought to cause megaloblastic anaemia by inhibiting the production of 5,10-methylene THF polyglutamate form, which acts as a cofactor in the rate-limiting step in the production of DNA, the synthesis of deoxythymidine monophosphate (dTMP). Thus, in the absence of cobalamin, polyglutamate synthesis ceases and monoglutamate forms are not retained by cells. This explains why in cobalamin deficiency serum folate levels may be found to be elevated and red cell folate levels normal or low. These metabolic pathways explain the interrelationship of serum B₁₂, serum folate and red cell folate results seen in the clinical laboratory.

Folate deficiency is therefore associated with elevation of homocysteine levels, reduced methionine synthesis and defective purine and pyrimidine synthesis, the latter resulting in morphologically visible megaloblastic change in bone marrow cells. Cobalamin deficiency also results in a functional folate deficiency with indistinguishable megaloblastic change and/or the more insidious and dangerous potential for neurological damage.

Rationale for investigation of cobalamin or folate status

Investigation of the vitamin B₁₂ and folate status of individuals is not restricted to investigation of individuals with classical features of megaloblastic anaemia alone because neuropathy and neuro-psychiatric changes may occur in B₁₂ deficiency in the absence of macrocytosis or anaemia.⁵^–⁹ The finding that folate supplementation reduced the incidence of neural tube defects¹⁰ by 25–46% in the USA and Canada highlights the importance of defining optimum population folate levels. Increased plasma homocysteine and serum methylmalonic acid (MMA) levels have been advocated¹¹^–¹⁴ as sensitive indicators of folate and cobalamin deficiency that may be subclinical. Introduction of metabolite testing to routine laboratory practice was limited in the past as a result of technical difficulty in measurement but is now becoming more available, though the clinical benefit of this approach has been questioned by some authors.^15,¹⁶ Elevated homocysteine levels are an independent vascular disease risk factor¹⁷ and are also associated with risk of idiopathic venous thrombosis.¹⁸ In the USA, dietary supplementation with folate was introduced in 1998 to reduce neural tube defects and by lowering homocysteine levels may achieve a further health gain in reduced myocardial infarction and stroke. Food supplementation with folate remains a controversial step in European countries because of a possible increased risk of cancers^19,²⁰ (see Scientific Advisory Committee on Nutrition, at: www.sacn.gov.uk, for update). Increased plasma homocysteine levels occur in both B₁₂ and folate deficiency as a result of reduction in methionine synthesis (Fig. 10.1). Some laboratories have utilized homocysteine as an initial screening test for abnormalities of cobalamin and folate metabolism which may be particularly appropriate for investigation of suspected inherited cobalamin or folate disorders in children.^21,²² MMA measurement though not widely available in the UK, requiring gas chromatography-mass spectrometry (GC-MS), is available at a limited number of university departments. In contrast, plasma homocysteine measurement by high-performance liquid chromatography (HPLC) and commercial enzyme immunoassays is widely available. The limitations of total serum B₁₂ measurement have been highlighted by studies that showed poor positive predictive value (i.e. healthy persons with a low level or low cobalamin levels with no evidence of deficiency) and <100% negative predictive value of 95% (i.e. 5% clinically deficient with normal level).^15,²³ In addition, some publications have highlighted the presence of severe cobalamin deficiency concurrent with a normal cobalamin level using some of the current commercial assays.^24,²⁵ The introduction of holotranscobalamin assays,²⁶^–²⁸ now available on an Abbott automated immunoassay platform, provides a readily accessible method of assessing the physiologically active form of cobalamin rather than the less relevant total B₁₂. The advent of a commercial holotranscobalamin assay occurred over 20 years after depletion of serum holotranscobalamin II was identified as an early sign of cobalamin deficiency.²⁹ Adoption of new methodology in laboratory medicine in the UK requires careful evaluation in a clinical setting and this research appears inhibited by lack of training in research methodology and financial constraints in service laboratories. Serum methylmalonic acid is increased in renal impairment and 7.9% of subjects over 51 have levels in excess of 0.27 μmol/l.¹ MMA must therefore be used in conjunction with serum cobalamin and holotranscobalamin and must be interpreted in the light of renal function. Urinary MMA measurement can be used to compensate for the impact of renal impairment.

Haematological features of megaloblastic anaemia

Megaloblastic anaemia resulting from impaired DNA synthesis is characterized by the presence of megaloblastic red cell precursors in the bone marrow and occasionally also in the blood. Megaloblasts have a characteristic chromatin pattern (Fig. 10.2) and increased cytoplasm as a result of asynchrony of nuclear and cytoplasmic maturation with a relatively immature nucleus for the degree of cytoplasmic haemoglobinization. The delay in nuclear maturation caused by delay in DNA synthesis resulting from lack of vitamin B₁₂ or folate is also seen in all lineages, particularly granulocytic marrow precursors with giant metamyelocytes (Fig. 10.3) and hyperlobated neutrophils with increased lobe size as well as number of nuclear segments (see Chapter 5, Fig. 5.10). In severe pernicious anaemia, a progressive increase in mean red cell volume (MCV) up to 130 fl occurs, with oval macrocytes, poikilocytes and hypersegmentation of neutrophils (>5% with more than five nuclear lobes).³¹ The neutrophil hypersegmentation index is an equivalent automated parameter on some cell counters, although hypersegmentation does not always respond to a therapeutic trial. The mean platelet volume is decreased and there is increased platelet anisocytosis, as detected by the platelet distribution width (PDW). The MCV falls to 110–120 fl as megaloblastic change advances. Howell–Jolly bodies and basophilic stippling are seen in the red cells.

Figure 10.2 Photomicrographs of bone marrow films stained by May–Grünwald–Giemsa. (A) Megaloblasts. (B) Normoblasts for comparison.

Figure 10.3 Photomicrograph of bone marrow film stained by May–Grünwald–Giemsa showing giant metamyelocytes.

Differential Diagnosis of Macrocytic Anaemia

Macrocytic red cells are also seen in myelodysplastic syndromes, which can be suspected from the presence of hypogranular neutrophils (see Chapter 5, Fig. 5.76) or monocytosis. Excess alcohol consumption results in an increased MCV as a result of round macrocytes, although rarely does it go higher than 110 fl unless coexisting folate deficiency is present. Hypothyroidism, liver disease, aplastic anaemia, rare inherited orotic aciduria or Lesch–Nyhan syndrome also have a high MCV. Automated reticulocyte counts facilitate detection of increased red cell turnover and high MCV as a result of haemolysis or bleeding. Coexisting iron deficiency or thalassaemia trait may mask macrocytic changes, although a high red cell distribution width indicates anisocytosis and the need for blood film review. Congenital dyserythropoietic anaemias types I and III and erythroleukaemia exhibit some features of megaloblastic erythropoiesis that are unrelated to B₁₂ and folate. Drugs interfering with DNA synthesis (e.g. azathioprine, zidovudine or hydroxycarbamide) result in macrocytosis and megaloblastic erythropoiesis. Anticonvulsant therapy interferes with folate metabolism,³² whereas the impact of oral contraceptives on folate absorption and metabolism is controversial.³³ Prolonged nitrous oxide anaesthesia destroys methylcobalamin and causes acute megaloblastic change.³⁴ Methotrexate inhibits dihydrofolate reductase and toxicity can be reversed with folinic acid, which is already in the tetrahydrofolate form and effectively reverses the metabolic block, which is not achieved solely with folic acid.

Testing strategy for suspected cobalamin or folate deficiency

Microbiological cobalamin and folate assays and competitive radiodilution binding assays for measurement of cobalamin and folate, which were often performed together, have largely been replaced by separate analysis by automated binding assays. The application of a suitable testing strategy for patients suspected of having cobalamin or folate deficiency is shown in Tables 10.1–10.3.

Table 10.1 Significance of clinical details

Symptoms or signs		Possible significance
Tiredness, palpitations, pallor		Anaemia
Slight jaundice		Ineffective erythropoiesis
Neurological
	Cognitive impairment, optic atrophy, loss of vibration sense, joint position sense; plantar responses normal or abnormal; tendon reflexes depressed or increased	Cobalamin deficiency, subacute combined degeneration of the spinal cord and sensori/motor peripheral neuropathies
Dietary and gastrointestinal history
	Vegetarian or vegan; poor nutrition (e.g. tea and toast diet in elderly or students); dietary fads	Low iron stores and iron deficiency
		Cobalamin deficiency in babies born to mothers who are vegans
		Folate deficiency (often with iron deficiency)
	Weight loss, bloating and steatorrhoea, particularly nocturnal bowel movements	Features of malabsorption and folate deficiency, e.g. due to coeliac disease, tropical sprue
	Mouth ulcers, abdominal pain, perianal ulcers, fistulae	Terminal ileal Crohn’s disease – cobalamin deficiency
	Glossitis, angular cheilosis and koilonychia	Cobalamin and combined iron deficiency
	Alcohol history	Poor diet and interference with folate metabolism
History of autoimmune disease in patient or family
	Hypothyroidism, pernicious anaemia or coeliac disease	Increased likelihood of pernicious anaemia or coeliac disease
Surgery
	Gastrectomy/bowel resection	Cobalamin deficiency usually 2 years post-gastrectomy
		Ileal disease resulting in cobalamin deficiency
		Blind loop syndromes
Physical appearance
	Grey hair, blue eyes, vitiligo	Association with pernicious anaemia
Pregnancy		Increased iron and folate requirements.
		Cobalamin levels fall by 30% in the 3rd trimester
		Holotranscobalamin levels are unaltered in late pregnancy
Malabsorptive syndrome
	Tropical sprue, bacterial overgrowth, fish tape worm in Scandinavian countries	Combined folate and iron deficiency
		Cobalamin deficiency
Drug history		See text
Other haematological disorders
	Myeloproliferative neoplasms, haemolytic anaemias, leukaemias	Increased folate utilization may result in folate deficiency
	Myeloma	Paraprotein interference with cobalamin assays resulting in falsely low cobalamin levels, which normalize on treatment of myeloma

Table 10.2 Laboratory tests in suspected cobalamin or folate deficiency

Table 10.3 Clinical and laboratory checklist for diagnosis of pernicious anaemia

	Laboratory criteria	Clinical criteria
Minor criteria	Macrocytosis	Parasthesiae, numbness or ataxia
	Anaemia	Hypothyroidism
	Raised plasma homocysteine	Vitiligo
	Gastric pH above 6 Raised serum gastrin	Family history of pernicious anaemia or hypothyroidism
	Positive gastric parietal cell antibody
Major criteria	Low serum B₁₂ (<180 ng/l) or raised serum methylmalonic acid (>0.75 μmol/l) in presence of normal renal function
	Megaloblastic anaemia not resulting from folate deficiency
	Positive intrinsic factor antibodies using high-specificity test.
	Holotranscobalamin level <23 μmol/l
Reference standard criteria	Schilling test^a shows malabsorption of oral cyanocobalamin corrected by coadministration of intrinsic factor

a Reagents for Schilling tests currently unavailable. A non-isotopic B₁₂ absorption test utilizing recombinant intrinsic factor and holotranscobalamin measurement is under development.

Table 10.1 highlights the important clinical details that should be elicited by the clinician and submitted with the request to assist the laboratory in interpretation of the numeric results of cobalamin and folate assays. Ideally, test requests should not be accepted without this information, which could be incorporated into electronic order communication from user to laboratory.

Table 10.2 lists the important laboratory investigations that should be performed – results must not be reported in isolation from other laboratory results and clinical details. If investigations are performed in different laboratories, authorization and release of results requires access to all laboratory data on the individual patient. Laboratory information systems should facilitate this cross-disciplinary access. For example, intrinsic factor antibody results should be available to haematology or clinical chemistry laboratories undertaking cobalamin, MMA or homocysteine assays.

Table 10.3 provides a list of clinical and laboratory features for diagnosis of pernicious anaemia. These criteria avoid undue reliance on a single B₁₂ assay and should help clinicians to make a diagnosis even when some critical tests, e.g. Schilling tests, are not available.

In view of the lack of specificity and sensitivity of serum cobalamin assays and frequent lack of availability of other diagnostic tests, basing the diagnosis of pernicious anaemia, which requires lifelong parenteral B₁₂ therapy, solely on laboratory results, is not straightforward. A checklist of laboratory and clinical diagnostic criteria, as shown in Table 10.2, helps to achieve a greater degree of diagnostic certainty than any single diagnostic test and permits the diagnosis to be made even when a single diagnostic test is anomalous or unavailable. Clinical and other laboratory criteria thus provide additional supportive evidence of an autoimmune aetiology, even if the more demanding diagnostic laboratory criteria are not met.

Limitations of Cobalamin Assays

Sensitivity and Specificity of Cobalamin and Holotranscobalamin Assays

Utility of receiver operator characteristic curves

There has been little data on sensitivity and specificity of current B₁₂ assays, due to the difficulty in defining a truly deficient study population. Some authors have suggested B₁₂ assays and measurement of methylmalonic acid are no better than tossing a coin, to determine the presence or absence of deficiency. The study by Clarke et al.²⁸ provides data which permits calculation of the specificity and sensitivity of a current B₁₂ immunoassay in the detection of cobalamin deficiency in a community study of 1621 subjects over age 65 with normal renal function. Subjects were defined as cobalamin deficient if the methylmalonic acid was elevated above 0.75 μmol/l. Deficiency was found in 4.3% of subjects over 65 years of age with normal renal function. The mean B₁₂ level of these subjects was 151 pmol/l (202 ng/l) by Siemens Centaur assay (range 110–199 pmol/l). Table 10.4 illustrates the calculation to derive specificity and sensitivity for the Siemens Centaur B₁₂ assay using a cut-off point of 200 pmol/l (270 ng/l).

Table 10.4 The calculation to derive specificity and sensitivity for the Siemens Centaur B₁₂ assay

This study demonstrates that, while values over 270 ng/l have a high (98.4%) negative predictive value for the presence of disease, values below 270 ng/l include a high percentage (28.2%) of individuals with normal methylmalonic acid levels and presumably no evidence of cobalamin deficiency, resulting in a large grey area. This is reflected in the poor specificity (71.8%) of the assay using this cut-off point. The choice of the appropriateness of the cut-off point can be explored further using receiver operator characteristic (ROC) curves. Laboratories able to construct ROC curves will be able to select an appropriate cut-off point to identify those patients with possible or definite disease.

Selection of a cut-off point for cobalamin of below 150 ng/l, will identify subjects with higher probability of presence of cobalamin deficiency. Some authors ¹³ have advocated choosing a cut-off point 25% below the reference range lower limit for any particular assay.

If, for example, a cut-off point of 125 pmol/l (168 ng/l) is chosen, the specificity of a value below this level, i.e. the number of normal individuals who fall below the cut-off point, will be markedly reduced (FP) and therefore the specificity of the test (TN/TN+FP) improves to 95%, although the detection of individuals with true deficiency who lie above the cut-off point, i.e. false negatives will be markedly increased, resulting in a sensitivity (TP/TP+FN) of 35%. The ROC curve therefore allows a laboratory to select a cut-off point that meets the objective of the laboratory – to have either a highly specific but low sensitivity test or to have a test of poor specificity but high sensitivity. Clearly cobalamin assays do not meet the criteria for an ideal test of high sensitivity and high specificity which would lie on the coordinates 0.1 to the left of the graph. The Axis-Shield/Abbott holotranscobalamin assay in this pivotal study is seen to have slightly superior ROC curves (see Fig. 10.4).

Figure 10.4 Receiver operator characteristic (ROC) curves (showing holotranscobalamin levels in μmol/l (abnormal <44 μmol/l) and B₁₂ levels in pmol/l (1 pmol/l = 1.344 ng/l). (A) The ROC curve for methylmalonic acid levels >0.75 μmol/l (definite metabolic B₁₂ deficiency). (B) Methylmalonic acid levels of >0.45 μmol/l (probable B₁₂ deficiency).

(Redrawn from Clarke et al. 2007).²⁸

Utility of holotranscobalamin, methylmalonic acid and homocysteine assays

Holotranscobalamin assays gave a greater area under the curve, 0.85 versus 0.76 in the above study, and superior sensitivity and specificity. Holotranscobalamin is the physiologically active fraction of total cobalamin.²⁷^–²⁹ An isolated abnormal result of serum B₁₂ should not be the sole criterion on which treatment decisions are based and a repeat assay and other confirmatory and clinical evaluation are necessary prior to a diagnostic conclusion. Additional secondary testing with metabolite levels, or holotranscobalamin and monitoring of treatment response is recommended. High B₁₂ levels have been described in subjects with no myeloproliferative neoplasm and who are not on cobalamin therapy or vitamin supplementation.³⁵ This is thought to be due to immunoglobulin-complexed B₁₂ resulting in assay interference. False normal B₁₂ levels^24,²⁵ have been described in subjects with high titre intrinsic factor antibody²⁵ and may also occur due to presence of heterophile antibody interference.

Holotranscobalamin assays ^27,²⁸ may challenge total B₁₂ assays as a first-line test in cobalamin assessment. Holotranscobalamin has been shown to be unaffected by assay interference from high-titre intrinsic factor antibody levels.³⁶ In addition, holotranscobalamin is not subject to the 30% fall in total B₁₂ levels seen in normal pregnancy, which makes low total B₁₂ levels uninterpretable during pregnancy (Fig. 10.5).³⁰

Figure 10.5 Comparison of HoloTC, cobalamins and HoloHC in pregnant women at 18th, 32nd, 39th gestational week (gw) and at 8 weeks’ postpartum (8 pp) (n = 141). Means and 95% CI (calculated from the normal distributed log transformed data) for HoloTC (white), cobalamins (white + grey shaded) and measured HoloHC (grey + black shaded) are shown. The grey shaded area indicates HoloHC saturated with true cobalamins (cobalamins – HoloTC) and the black shaded area indicates HoloHC saturated with analogues (measured HoloHC + HoloTC – cobalamins). CI, confidence interval.

(Redrawn from Morkbak et al. (2007), Figure 1.³⁰)

Access to homocysteine, methylmalonic acid and holotranscobalamin assays facilitates more precise definition of cobalamin and folate status in patients in whom prolonged B₁₂ therapy may have been initiated inappropriately and continued unnecessarily or, conversely, discontinued inappropriately because of lack of confidence in the original assessment. The study by Gorringe et al.¹⁶ showed that only 27/49 patients were anaemic or macrocytic with a low total B₁₂ of <170 ng/l and elevated MMA. All of these patients treated with B₁₂ therapy corrected elevated MMA levels, suggesting the presence of a metabolic deficiency. However, only 15/27 showed a haematological response.

Homocysteine levels also fell by >25% in 47/49 patients after B₁₂ treatment. Many of these patients had no clinical evidence or symptoms of cobalamin deficiency and may reflect subjects with subclinical deficiency, which could have subtle cognitive impairment, or may just represent a compensated metabolic state of no clinical consequence. Elevation of homocysteine levels is seen in folate deficiency and is therefore less specific than MMA measurement. The causes of cobalamin deficiency are shown in Table 10.5.

Table 10.5 The causes of cobalamin deficiency

		Supportive information/diagnostic tests
Reduced intake
	Strict vegetarian/vegan	Dietary history
	Dietary fad that excludes dairy products and meat	Ethnic origin/culture
	Breastfed babies of mothers who are vegetarian or cobalamin deficient
	Poor dietary intake in elderly
Malabsorption as a result of loss or inactivity of intrinsic factor
	Addisonian pernicious anaemia	Diagnostic criteria for pernicious anaemia (see Table 10.3)
	Gastrectomy (partial or total)	History of gastric surgery
	Bacterial overgrowth or parasitic infestation of small bowel	Radiolabelled lactose breath tests for bacterial overgrowth Repeat Schilling test post-antibiotic therapy
	Pancreatic dysfunction: failure of trypsin release of B₁₂ from R binding proteins	Pancreatic function tests; exocrine pancreatic dysfunction results in abnormal Schilling test but clinical deficiency is rare
	Malabsorption as a result of failure of B₁₂-intrinsic factor complex uptake in ileum – ileal resection	Radiological, enteroscopic or capsule camera study of small bowel for Crohn’s disease of terminal ileum or tuberculous ileitis
	Congenital Imerslund–Gräsbeck syndrome	Subjects of Scandinavian origin
	Tropical sprue	Small bowel biopsy
	Zollinger–Ellison syndrome	Multiple gastric and duodenal ulcers
		Pancreatic adenoma on imaging
Food cobalamin malabsorption
	Atrophic gastritis with achlorhydria	Endoscopic and gastric biopsy findings
	Gastric surgery
Abnormal transport proteins
	Transcobalamin II deficiency	Megaloblastic anaemia in presence of normal cobalamin levels; transcobalamin II and holotranscobalamin levels reduced
	Transcobalamin I deficiency	No evidence of clinical deficiency but low serum cobalamin levels
		Possible fall in holotranscobalamin levels in elderly
Inborn errors of cobalamin metabolism (see reviews^21,²²)		Serum and urinary methylmalonic acid and metabolite measurement
Acquired drug effects
	Nitrous oxide: chronic repeated exposure	Drug history
	Colchicine: chronic usage impairs B₁₂ uptake resulting from diarrhoea
	Metformin reduces B₁₂ levels in diabetics.

Clinical and Diagnostic Pitfalls of Folate Assays

Serum folate is altered by acute dietary change and interruption of enterohepatic recycling; it can therefore be low without significant tissue deficiency. This may be a particular problem in hospital inpatients. Red cell folate was originally advocated as correlating better with megaloblastic change ^37,³⁸ reflecting the folate status over the lifespan of the red cells (2–3 months), but a subsequent study suggested that little was to be gained by the addition of red cell folate analysis because only 14% of patients with low serum folate also have low red cell folate levels.³⁹ Minor haemolysis in vitro may cause spurious elevation of serum folate levels because the red cell folate may be 10–20 times the serum value. More than half of the patients with severe cobalamin deficiency have a low red cell folate because impaired methionine synthesis results in accumulation of methyltetrahydrofolate (MTHF) monoglutamate, which diffuses out of cells resulting in a high serum folate.^40,⁴¹ Treatment with cobalamin alone will correct the low red cell folate and high serum folate levels. Concern over the inter-method variability of red cell folate assays, and questions about the additional benefits³⁹ of measurement of both serum and red cell folate, have reduced the use of red cell folate assays. However, satisfactory results are possible if appropriate care is taken with preanalytical sample preparation and analysis. 5-methyl THF is a very labile substance and the addition of sodium ascorbate has reduced the coefficient of variation by half in serum folate assays in external quality assurance surveys.⁴² The interplay between serum B₁₂, serum folate and red cell folate and plasma homocysteine and MMA is shown in Table 10.7. In view of the limitations of both serum and red cell folate assays, it is prudent to measure both. An international reference method^43,⁴⁴ for serum folate has been recognized by the Joint Committee on Traceability in Laboratory Medicine (JCTLM) and is now used to verify the target values in UK National External Quality Assessment Scheme (NEQAS) Haematinics surveys (www.ukneqas-haematinics.org.uk). This should improve the accuracy of serum folate assays, which have been bedevilled by inter-method differences and variation in recommended reference ranges between manufacturers. Red cell folate assays still suffer from a lack of a standardized method for haemolysate preparation and matrix effects, which result in large inter-method differences.

Clinicians and patients using different laboratories require information about derivation of reference ranges, which will now be included in the standards for laboratory accreditation in the UK. Definition of population reference ranges for both serum and red cell folate have been difficult to achieve due to changing diet and food supplementation and assay variability. Elevated levels of homocysteine have been found in elderly subjects,^45,⁴⁶ indicating possible subclinical folate deficiency in the elderly. Folate reference ranges provided by kit manufacturers also show large inter-method differences. In the USA, some authors have advocated cessation of folate testing since, following dietary supplementation of flour, folate deficiency is very unusual. The causes of clinical deficiency and supportive information or diagnostic tests are shown in Table 10.6 and the interactions between cobalamin and folate are shown in Table 10.7.

Table 10.6 The causes of folate deficiency

		Supportive information/diagnostic tests
Reduced intake
	Poor diet particularly alcoholics (wine and spirits because beer contains folate)	Dietary and alcohol history
	Elderly or students ‘tea and toast diet’
	Dietary fads
	Premature babies
	Unsupplemented parenteral nutrition
Malabsorption
	Coeliac disease (often with coexisting iron deficiency) Tropical sprue	Antiendomysial, antigliadin tests, antitissue transglutaminase
	Small bowel resection, malabsorption syndromes	Small bowel biopsy
Drug effects
	Sulphasalazine, methotrexate, trimethoprim-sulphamethoxazole, pyrimethamine, phenytoin, sodium valproate, oral contraceptives	Drug history
		Bone marrow aspiration
Hereditary hyperhomocysteinaemia
	Homozygotes for C677T MTHFR have lower folate levels
Increased folate turnover
	Pregnancy: progressive fall in 3rd trimester
	Increased requirements for breastfeeding
	Skin disease – severe psoriasis or exfoliation
	Haemodialysis
	Haemolysis: haemoglobinopathy, paroxysmal nocturnal haemoglobinuria, autoimmune haemolytic anaemia (see Chapters 11–13)

Table 10.7 Interaction between serum cobalamin, serum folate, red cell folate, plasma homocysteine, serum methylmalonic acid, urinary methylmalonic acid and holotranscobalamin

Standards, Accuracy and Precision of Cobalamin and Folate Assays

There are currently no internationally recognized reference methods for serum cobalamin measurement, but isotope dilution liquid chromatography tandem mass spectrometric methods have recently been accepted as international reference methods for the quantification of folate species in serum.^43,⁴⁴ As a result, international reference materials have been developed (by the World Health Organization, WHO 03/178, and by the National Institute of Standards and Technology, NIST SRM), with values assigned for folate species by the tandem mass spectrometric methods.⁴⁷ Although reference methods have not been verified for the whole blood matrix as yet, there is a WHO whole blood international standard (95/528) with consensus values for total folate.⁴⁸

Evaluation of commercial automated binding assays by recovery experiments has shown under-recovery of added 5-methyl THF and over-recovery of pteroylglutamic acid (PGA),⁴⁹ whereas a suitably calibrated microbiological assay recovers closer to 100%.⁵⁰ Differential sensitivity of assays to pteroylglutamic acid and genetic variability in the proportion of in vivo formyl folates may be a factor in inter-method variability.

A microbiological assay was the method used to assign a potency value to the British Standard for human serum B₁₂,⁵¹ and this was later re-classified as the 1st WHO International Standard (IS) (81/563). The 2nd WHO IS for serum B₁₂, 03/178, was ratified in 2007,⁴⁷ the values adopted being a consensus of the contemporary B₁₂ protein-binding assays.

External quality assessment schemes have shown serum B₁₂ intra-method coefficients of variation (CV) of 6–10% and as much as 20% at clinically relevant levels; there is thus a substantial ‘grey’ indeterminate range between normal and low values. Serum folate intra-method CVs are between 6% and 12% and higher CVs of up to 20% are seen for red cell folate assays. Overall between-method CVs may be as high as 35% for the serum methods and can reach 50% for the whole blood assays, suggesting considerable method differences. The causes of this variability include patient factors as well as pre-analytical, analytical and post-analytical factors, as discussed later.

Genetic Factors

A number of methylenetetrahydrofolate reductase polymorphisms that alter the proportion of formylfolate in serum have been described ⁵² and this could be a potential source of disparity in the response of some sera to different assays. Individuals homozygous for C677T genetic polymorphism have 25% higher plasma homocysteine levels than controls.⁵³ A genetic–nutrient interactive effect is noted in that the polymorphism confers a greater effect on homocysteine levels in those individuals with low folate levels. Cigarette smoking, age, renal disease, drugs including levodopa and folate supplements all affect homocysteine levels.

Pre-analytical Sample Preparation

Serum B₁₂ is stable at room temperature and is not affected by sample handling, unless the sample is haemolysed. Holotranscobalamin is a sensitive marker of recent cobalamin intake and day-to-day variation is 10%. A change in value of twice the day-to-day variation may be taken as a significant change. Folate is affected by recent dietary intake and ideally fasting samples should be taken. However, this is difficult in practice and assumes that the reference range was also based on fasting samples. Marked loss of folate activity is observed as a result of light and temperature instability. Because red cells contain 30–50 times more folic acid than serum, even slight haemolysis will affect serum folate analysis. Thus, rapid transportation and separation prior to analysis, avoidance of storage at room temperature and the storage of samples at 2–8°C for a maximum of 48 h, or at –20°C for no longer than 28 days are all critical factors in the accuracy and precision of serum folate assays. Presence of haemoglobin as a result of lysis in a plasma or serum sample can be readily determined and may be quantified by haemoglobinometry.

The addition of sodium ascorbate 5 mg/ml will stabilize folate in serum, extending sample storage times. Stabilization of serum folate with sodium ascorbate added to the primary blood collection tube would improve the reproducibility of routine serum folate assays as shown in external quality control surveys,⁴² but would necessitate introduction of separate B₁₂ and folate sample tubes since ascorbate interferes with cobalamin analysis. EDTA plasma is unsuitable and heparinized plasma may result in higher values. Samples must be fibrin free and without bubbles.

Analytical Factors

Analytical sensitivity or limit of detection (LOD) varies between methods. It is defined as the concentration of analyte at 2SD of 20 replicates above the zero standard and for B₁₂ assays is normally in the region of 22 pmol/l (30 ng/l) and for folate is 0.68 nmol/l (0.3 μg/l). This is sometimes confused with the functional sensitivity of an assay, a term that defines the analyte concentration at which the CV of the assay is 20%. It is preferable that the functional sensitivity limit of serum B₁₂ assays is closer to 37 pmol/l (50 ng/l) than the 111 pmol/l (150 ng/l) quoted by some kits because this provides increased sensitivity at the clinically important lower end of the reference range.

For many folate assays, functional sensitivity is in the region of 2.26 nmo1/l (1.0 μg/l) or less, although the Roche Elecsys assay quotes 4.5 nmol/l (2.0 μg/l).

Limitations and Interference

Ascorbic acid destroys vitamin B₁₂; therefore ascorbate-treated serum cannot be used for B₁₂ assays. Methotrexate and folinic acid interfere with folate measurement because these drugs cross-react with folate-binding proteins. Minor degrees of haemolysis significantly increase serum folate values as a result of high red cell folate levels. Lipaemia with >2.25 mmol/l (2 g/l) of triglycerides and bilirubin >340 µmol/l (200 mg/l) may affect assays.

High cobalamin levels were thought to be due to B₁₂ therapy, vitamin supplementation, myeloproliferative neoplasms or liver disease. In the absence of these factors, assay interference may result from immunoglobulin–B₁₂–transcobalamin complexes.³⁵

Post-analytical Factors

The clinical interpretation of laboratory data should take account of the positive or negative predictive value of a result. The report should include a reference range, the derivation of which should be indicated. Food supplementation with folate or voluntary additional vitamin intake has resulted in bimodal distributions of vitamin levels in some populations, further complicating the definition of normality.

Methods for cobalamin and folate analysis

Microbiological bioassays and radiodilution assays for serum B₁₂ and folate ⁵⁴ are still used, albeit by a decreasing minority of laboratories and continue to play an important role in the evaluation of new automated methods. They are also used in population studies where they are useful in providing information on the long-term comparability of results. (They are detailed in the 9th edition of this book.)

Modern methods are highly automated, heterogeneous, competitive protein-binding assays with chemiluminescence or fluorescence detection systems.

General principles of competitive protein binding assays

The majority of automated single-platform, multianalyte, random-access analysers offer assays for serum B₁₂, folate and homocysteine by non-isotopic competitive protein binding or immunoassays. Second antibodies may be utilized as part of the separation procedures. These assays have been designed for serum assays and the assay platforms may not be optimized for analysis of haemolysates.

Serum B₁₂ assays

Release from Endogenous Binders and Conversion of Analyte to Appropriate Form

Nearly 99% of serum B₁₂ is bound to endogenous binding proteins (transcobalamin I, II and III) and must be released from these before measurement. The release step utilizes alkaline hydrolysis (NaOH at pH 12–13) in the presence of potassium cyanide (KCN), which converts cobalamin to the more stable cyanocobalamin and dithiothreitol (DTT) to prevent rebinding of released B₁₂. Alkaline hydrolysis requires subsequent adjustment of pH to be optimal for the binding agent.

Binding of B₁₂ to Kit Binder

The binding of B₁₂ to kit binder is the competitive step of the assay. Serum-derived cyanocobalamin competes with labelled cobalamin, which is usually complexed to a chemiluminescent or fluorescent substrate or enzyme, for limited binding sites on porcine intrinsic factor. Specificity for cobalamin is ensured by purification of the intrinsic factor (IF) or by blocking contaminating corrinoid binders (R binders) by addition of excess blocking cobinamide. Specificity of pure and blocked IF can be demonstrated by the addition of cobinamide to sera. There should be no increase in assay value. Some assays use only the alkaline denaturation step to inactivate the endogenous binders. It is important that assays are not be affected by the presence of high-titre anti-intrinsic factor antibody in patient sera. The Siemens Centaur, Beckman Access, Siemens Immulite 2000/2500 and Tosoh assays have all been vulnerable to this false normal B₁₂ effect and carry product literature warnings to this effect. Roche and Abbott assays have not demonstrated this problem with such sera and the holotranscobalamin assay by Axis-Shield is also unaffected.^25,^26,³⁶

Separation of Bound and Unbound B₁₂

Following competitive binding, the separation of bound and unbound B₁₂ is achieved by a number of electro- or physico-chemical and immunological methods. The Roche Elecsys utilizes an electrochemiluminescence measuring cell in which the bound B₁₂–ruthenium–IF complex, attached to paramagnetic particles by biotin–streptavidin, is magnetically captured onto the surface of an electrode. The Abbott Axsym uses polymer microparticles (beads) coated with porcine IF to bind B₁₂ and the bound B₁₂ is then immobilized by irreversible binding to a glass fibre matrix. These methods are designed to improve separation of bound and unbound B₁₂; they may utilize murine anti-intrinsic factor antibody–enzyme conjugates as part of the signal generation.

Signal Generation

The bound fraction is then detected by addition of a chemiluminescent, fluorescent or colorimetric enzyme substrate, which results in generation of fluorescence or light emission. There are two types of signal: flash, which is pH or electrically induced, and plateau, which is sustained. The initial rate of reaction or the area under the curve is used to calculate the result.

Electrochemiluminescence Immunoassay

In the Roche Elecsys platforms a voltage is applied to the electrode on which the bound B₁₂–ruthenium–IF complexes have been immobilized by magnetic attraction. This generates electrochemical luminescence that is measured by the photomultiplier, the relative light units being inversely proportional to the sample B₁₂ concentration. The signal that is produced is timed and integrated by the instrument’s software.

Enzyme-linked Fluorescence Generation

In the Abbott Axsym substrates such as 4-methyl-umbelliferyl phosphate are cleaved by an enzyme (i.e. alkaline phosphatase–ligand complex) and the resulting reaction generates fluorescence, the intensity of which is inversely proportional to the B₁₂ concentration in the sample.

The Siemens Centaur and the Abbott Architect use acridinium esters bound to B₁₂–IF complex coupled to paramagnetic particles; photons are emitted in response to pH change. The Siemens Immulite 2000/2500 (formerly DPC) uses adamantyl dioxatane phosphate as an alternative substrate, which is cleaved by alkaline phosphatase-labelled B₁₂–IF complex, resulting in generation of a plateau chemiluminescent signal. Alkaline phosphatase/4 methyl-umbelliferyl phosphate is utilized by the Tosoh Eurogenetics method and Beckman Coulter Access employs alkaline phosphatase/dioxatane phosphate (Lumi-Phos) for signal generation. Precise descriptions of the assay methods are given in the product literature.

Holotranscobalamin assays

Principle

About 6–20% of B₁₂ is bound to TC II forming the physiologically active complex HoloTC and in this form is taken up by cells. Levels of holotranscobalamin (HoloTC) B₁₂ are 30–160 pmol/l. The remainder of the serum B₁₂ is bound to transcobalamin I (haptocorrin), which is involved in the transport of B₁₂ to the liver and enterohepatic circulation thereof. Haptocorrin also binds B₁₂ in the gastric contents and B₁₂ is released from haptocorrin by pancreatic proteases prior to capture by IF. HoloTC is thought to be the first metabolite to decrease following reduced intake or absorption of B₁₂.^28,²⁹ A commercial assay is available from Axis-Shield and Abbott.²⁷ The clinical utility of this assay is being confirmed particularly in pregnancy sera where changes in transcobalamin I make total B₁₂ levels uninterpretable. The relative merits of HoloTC compared with metabolite measurement remain to be clarified. As a sensitive marker of cobalamin malabsorption, holotranscobalamin levels that correct with small oral doses of B₁₂ and the use of recombinant intrinsic factor⁵⁵^–⁵⁸ could provide the basis for a non-isotopic B₁₂ absorption test to replace the Schilling test, currently unavailable in the UK.

Holotranscobalamin ‘Active B₁₂’ Immunoassay

Measurement of the ‘active form of B₁₂’²⁷ is based on a monoclonal antibody to cobalamin bound to TCII known as HoloTC. This ‘active’ form of B₁₂ amounts to 20% of the total serum B₁₂. Reference range values are 19–134 pmol/l; pernicious anaemia values are <5 pmol/l. HoloTC is thought to be more sensitive to early deficiency than the measurement of total B₁₂ and may be a better indicator of vulnerability to impaired cognitive function.²⁸ Levels are elevated by renal failure, but are unaffected by pregnancy.³⁰

HoloTC levels below 23 pmol/l show an increase in macrocytosis and red cell distribution width. The HoloTC assay appears to be unaffected by assay interference due to high intrinsic factor antibody levels which has caused false normal B₁₂ levels in severely deficient patients, as reported to the Medicines and Healthcare products Regulatory Agency (MHRA) by UK NEQAS for three of five major manufacturers of serum B₁₂ assays. As expected, HoloTC levels do not correlate with total serum B₁₂ and there is an indeterminate zone with borderline levels which may represent early or subclinical deficiency. HoloTC levels below 20 pmol/l occur in approximately 10% of laboratory requests for B₁₂.

HoloTC Radioimmunoassay

The HoloTC radioimmunoassay ⁵⁹ uses magnetic microspheres coated with monoclonal antibody to holotranscobalamin and achieves separation from haptocorrin by magnetic separator. A⁵⁷Co B₁₂ tracer together with a reducing and a denaturing agent are then added to destroy the HoloTC linkage. When the B₁₂ binder containing IF is added, the free B₁₂ and tracer compete for binding. The unbound tracer is removed by centrifugation and the bound fraction is measured using a gamma counter. The measured radioactivity reflects the competition between tracer and vitamin B₁₂ bound to transcobalamin (i.e. HoloTC). The concentration of vitamin B₁₂ in the sample is calculated from a calibration curve using recombinant human HoloTC. The assay only requires a 0.4 ml sample volume, the coefficient of variation is <10%, the limit of detection is 10 pmol/l and the assay time is 4 h.

Quantitation of Transcobalamin Saturation

Nexo and colleagues ^60,⁶¹ described a method permitting measurement of total TC and HoloTC. The method uses B₁₂ modified by acid treatment and bound to magnetic beads, which can then be used to remove unsaturated TC or apoTC from serum. The remaining HoloTC is then measured by an enzyme-lined immunosorbent assay (ELISA). Thus the total and HoloTC can be measured and the TC saturation (HoloTC/Total TC) can be quantitated.

In a study of 137 healthy blood donors the reference range for HoloTC was 40–150 pmol/l. Some 10% of circulating TC is saturated with a reference range of 5–20%; 15–50% of B₁₂ is bound to TC. In subjects who were B₁₂ deficient, HoloTC was 2–34 pmol/l and the TC saturation was 0.4–3%, well below the reference interval, providing a clear cut-off from normal sera. Nexo’s method combines a sensitive ELISA ⁶⁰ for HoloTC with a simple procedure for removal of the unsaturated TC or apoTC.

Serum folate methods

The first methods used for measurement of serum folate were microbiological assays. Radioisotope dilution (RID) assays were subsequently developed and the newer commercial, automated, competitive-binding assays are based on similar principles. As with B₁₂, the use of the original microbiological and RID procedures for serum and red cell folate measurements has diminished. Definition of assay response to different forms of folate is crucial for inter-assay comparisons, particularly in view of the effect of dietary supplementation with folate.

Release from Endogenous Binders

Serum folate present as 5-methyl THF is released from endogenous binders (two-thirds weakly bound to serum proteins and a minority to membrane-derived soluble folate receptors) by alkaline denaturation. DTT or monothioglycerol (MTG) is used in most folate assays to prevent reattachment of folate to the endogenous binders and to keep the folate in its reduced form.

Binding of Folate to Folate-Binding Protein

β-Lactoglobulin, isolated from cow’s milk, is commonly used as a binding agent in folate assays, i.e. as the folate-binding protein (FBP). Unfortunately, these lactoglobulins are not specific for the attachment of 5-methyl THF and will bind many forms of folate, dependent on pH. These properties have been utilized for the standardization of folate assay. The physiologically active form (5-methyl THF) is very unstable. Pteroylglutamic acid (PGA), present in vitamin supplements, is more stable, binds to FBP and has been used as an alternative standard for the folate assays. Its use as a standard depends on the observation that the binding affinities for 5-methyl THF and PGA are equivalent ⁶² at pH 9.3 ± 0.1. It is therefore essential that the pH is strictly maintained at the binding stage of the procedure. In most of these assays, serum-derived folate competes with chemiluminescent or enzyme-labelled folate for limited sites on FBP. In the Roche Elecsys methods the FBP is labelled with ruthenium and competition exists between endogenous and biotin-labelled folate.

Separation of Bound and Unbound Folate

Following competitive binding the separation of bound and unbound folate is achieved by a number of electro- or physico-chemical and immunological methods.

In the Abbott Axsym ion capture assay, murine anti-FBP immunoglobulin G (IgG) results in negatively charged polyanion–folate complexes. These are captured electrostatically using a positively charged glass fibre matrix (charge imparted by high molecular-weight quaternary ammonium compound) that removes the bound fraction. The Beckman Access uses murine anti-FBP goat antimouse IgG coated on paramagnetic particles. The Siemens Immulite platforms use murine anti-FBP coated on polystyrene beads with separation by centrifugation.

Signal Generation

Generation of electro-chemiluminescence, chemiluminescent, fluorescent or light emission is similar to that for B₁₂ methods. In assays where the bound B₁₂ complex is labelled with chemiluminescent esters, signal is generated in response to pH change or is electrically induced.

Red cell folate methods

Whereas Lactobacillus casei responds equally to both triglutamates and monoglutamates, the affinity of the FBP varies with the number of glutamate residues. Reproducible protein-binding assays for red cell folate can only be achieved by release and conversion of the protein-bound folate polyglutamates, mainly 5-methyl THF with four or five additional glutamate moieties, to a monoglutamate form. There must be adequate dilution of the red cells in hypotonic solution, a pH between 3.0 and 6.0 (ideally pH 4.5–5.2) for optimal conditions for plasma folate deconjugase (polyglutamate hydrolase) and ascorbic acid to stabilize the reduced forms.^63,⁶⁴

Haemolysate preparations for the newer assay platforms vary widely. The concentration of ascorbic acid varies from 0.09% to 1%, dilution factors from 1:5 to 1:31 and the duration of haemolysate preparation from 40 to 180 min. Some assays require the addition of protein to lysates before analysis and use bovine serum or human serum albumin, whereas others need only aqueous solutions. The pH of lysing diluent varies from 3.0 to 4.0 and that of the deconjugase step from 4.0 to 6.8 (the final pH of the lysate after protein addition varies from 4.4 to 7.5). These various factors may help to explain the large intermethod differences detected in external quality assessment surveys. Inadequate lysis and deconjugation will give falsely low results.^65,⁶⁶

If advice is not given by the manufacturer a suitable haemolysate method described below may be used. However, when the haemolysate is analysed using the serum folate methodology, this preparation method may not be suitable for all instruments.

Haemolysate Preparation

Red cell folate samples are usually collected in EDTA tubes and may be stored at 4°C for up to 48 h prior to lysate preparation. The lysate is prepared by adding 0.1 ml of whole blood of known haematocrit (Hct) to 1.9 ml of 1 g/dl freshly prepared aqueous ascorbic acid, incubated at room temperature for 60 min, in the dark. The ascorbic acid stabilizes the folate and the pH of 4.6 allows plasma folate deconjugase to convert the polyglutamate forms to the monoglutamate form. Folate activity should be assayed straight away although the lysate may be stored at 4°C for no more than 24 h prior to analysis. Storage at –20°C for longer periods is permissible, but approximately 10% decrease in activity is noted following a single freeze/thaw cycle. Whole blood in heparin, EDTA, or citrate–phosphate–dextrose (CPD) can be used in some, but not all, assays.

Calculation of Red Blood Cell Folate from Haemolysate Folate Result

1. Multiply the haemolysate folate value by the dilution factor (e.g. 21 if a 1:21 dilution) to obtain the folate concentration of whole blood in µg/l.

2. For an uncorrected red cell folate value in µg/l, divide the whole blood folate result in (1) by the Hct (l/l).

3. The result obtained in (2) should, if possible, be corrected for the serum folate value, which for patient populations taking supplemented dietary folate may now be quite significant, as is illustrated in the following worked example:

Corrected red cell folate (μg/l) =

Red cell folate in μg/l − [serum folate in μg/l × (1 − Hct)/Hct]

e.g. when Hct is 0.45, red cell folate is 180 μg/l and serum folate is 28 μg/l, then

Corrected red cell folate (μg/l) = 180 − [28 × (1 − 0.45)/0.45] = 145.8.

Many serum assays have an upper limit of 22–28 μg/l and the serum sample must therefore also be diluted and retested to obtain accurate results. Serum folate is traditionally quoted as µg/l or ng/ml; to convert to SI units (Système International d’unités); µg/l × 2.265 = nmol/l and nmol/l × 0.44 = μg/l.

Serum B₁₂ and Folate and Red Cell Folate Assay Calibration

Cobalamin in serum is protein bound and therefore standards for total B₁₂ assays should be gravimetrically prepared cyanocobalamin in either a lyophilized serum or protein adjusted matrix. A WHO lyophilized serum standard with consensus values for serum B₁₂ is available from the National Institute for Biological Standards and Control (NIBSC) (www.nibsc.ac.uk);⁴⁷ however, the protein binding B₁₂ assays are not usually calibrated with this material.

For serum folate analysis, 5-methyl THF is the physiologically active folate form and therefore should be used as the standard. However, 5-methyl THF is highly unstable and historically PGA has been used as the primary calibrator either gravimetrically added to aqueous standards or used to assign values to secondary serum matrix standards. The principle underlying the use of PGA is dependent on the equimolar binding of 5-methyl THF or PGA at pH 9.3.⁶² More recent work suggests the pH of equimolar binding may be nearer 8.9 than 9.3.

The recent development of isotope dilution, liquid chromatography, tandem mass spectrometry (ID-LC/MS/MS) reference methods for folate derivatives in serum ^43,⁴⁴ has permitted the introduction of higher-order serum and plasma reference materials with accurate values for folates. A lyophilized serum standard (WHO 03/178)⁴⁷ is available from NIBSC and three frozen plasma preparations (FPP) with assigned values for 5-methyl THF and folic acid (FA) are available from the NIST in the USA (www.nist.gov/index.html).

Whole Blood Folate Standards

An international reference preparation for whole blood folates ⁴⁸ is available from NIBSC although the haemolysate preparations are generally analysed using the serum method protocols and calibration curves.

Primary Instrument Calibration

The majority of manufacturers still use gravimetric PGA calibrators or PGA values assigned to serum matrix calibrators referenced to older RID methods as their primary instrument calibration. Siemens now utilizes 5-methyl THF as its instrument calibrator, although this is referenced to aqueous gravimetric factory standards. Methods calibrated in this way should be optimized for the detection of 5-methyl THF and should not require the adjustment of the binding pH to 9.3. Currently only one major manufacturer has opted to harmonize the calibration of its serum folate assay with the WHO 1st International Standard for Total Folate in Serum. Results from recent surveys of the UK NEQAS External Quality Assessment (EQA) Haematinics Scheme suggest that this method (Abbott) is now accurate when compared to values obtained by the ID-LC/MS/MS reference method (UKNEQAS Haematinics steering committee report October 2009).

Internal Adjustment Calibration

Most automated assay systems use calibration curves that are stored on barcode systems with each reagent lot. The barcode also contains the mathematical formulae for shifting or adjusting the observed responses to the master curve when the instrument requires routine calibration. The calibration interval within reagent lots usually varies from 7 to 28 days. Most instruments will also require at least a 2-point calibration when changing lot numbers of reagent packs, when replacing system components and when internal quality control results are out of range. Reagents should be discarded at the end of the stability intervals and should not be used beyond the expiry dates. Particular care should be taken to ensure that calibrators and reagents are correctly mixed and homogeneous, especially those containing particulates. Aliquots of reagents may be stored at 2–8°C or at –20°C if permitted by the manufacturer.

Internal Quality Control

As a minimum requirement, two levels of quality control material should be assayed daily when samples are being analysed. A choice of batch analysis, or true random access, may be preferred by the operator, although in view of folate instability prolonged on-board times are to be avoided by consideration of the test repertoire for a given set of samples. Suitable controls are available commercially (Bio-Rad Laboratories).

Methylmalonic acid measurement

Principle

Methylmalonic acid (MMA) is one of several dicarboxylic acids present in urine and plasma. Before separating MMA from other interfering substances it is necessary to obtain derivatives of these compounds. Urine concentration of MMA is higher than plasma concentration and needs to be normalized for urine creatinine concentration and corrected for the effects of renal impairment or dehydration.

Serum measurement offers added convenience, is unaffected by diet and can use the same sample as that obtained for B₁₂.

Methods

Plasma or serum MMA is extracted, purified and, using tert-butyldimethylsilyl derivatives of MMA, measured by GC-MS. A deuterated stable isotope of MMA is used as an internal standard. The use of dicyclohexyl, another derivative of MMA, is described by Rasmussen.⁶⁷

A method for screening for inborn errors of metabolism using O-(2,3,4,5,6-pentafluorobenzyl) hydroxylamine HCl to derive oxoacids, followed by liquid partition chromatography, was described by Hoffman et al.⁶⁸ In more recent HPLC or capillary electrophoresis methods,^69,⁷⁰ MMA can be derivatized using the fluorescent compound 1-pyrenyldiazomethane to yield a fluorescent monoester adduct. After separation by HPLC or capillary electrophoresis, short-chain dicarboxylic acids are quantified following laser-induced fluorescence activation.

Reference ranges for serum MMA measured by GC-MS and using mean ± 3SD of log transformed data are 53–376 nmol/l ⁷¹ and 50–440 nmol/l (Rasmussen et al.⁷²).

Homocysteine measurement

Principle

Homocysteine is a disulphide amino acid present at low concentrations in cells (<1 µmol/l) and in plasma at 5–15 µmol/l.⁷³ Homocysteine has a reactive sulphydryl group that forms disulphide bonds with homocysteine or cysteine or protein sulphydryl groups to form the oxidized form of homocysteine: homocystine, homocysteine-cysteine mixed disulphide and protein-bound homocysteine. The free homocysteine is <2% of plasma homocysteine and 80% is as protein-bound homocysteine; the remainder is homocystine or mixed disulphides. Reducing conditions convert all these species to homocysteine by reduction of disulphides; therefore total homocysteine is measured.

For quantification, plasma homocysteine requires protein precipitation and reduction of disulphide bonds. The S-H group of homocysteine is derivatized using a thiol-specific reagent and the resulting adduct is detected. A variety of methods have evolved ⁷³ from ion-exchange amino acid analysers,⁷⁴ radioenzymatic determination, capillary gas chromatography⁷⁵ stable isotope dilution combined with capillary GC-MS,⁷⁶ liquid chromatography electrospray tandem mass spectrometry⁷⁷ and HPLC methods using fluorochromophore detection.

HPLC methods that use fluorochromophore detection systems are still available but are not as widely used in clinical laboratories as previously. These methods first reduce the disulphide bonds followed by the removal of protein and the derivatization of the thiol groups. Separation requires reverse-phase HPLC and high organic mobile phase such as heptane sulphonic acid/methanol mobile phase at pH 1.9–2.0 for example. In the method from Drew Scientific, detection is by fluorescence of the fluorophore at λ_ex 385 nm and λ_em 515 nm.

Development of enzymatic and immunoassays for homocysteine has dramatically changed the availability of its measurement in routine laboratories. These are discussed further in the following section.

Immunoassay for Homocysteine Measurement

Automated enzyme immunoassays ⁷⁸ for homocysteine have now been developed by a number of manufacturers and have reduced the dependence of clinical laboratories on highly specialized instrumentation such as gas chromatography, mass spectrometry and HPLC. The general principle of these methods is the release of protein-bound homocysteine by buffered DTT and the reduction of homocystine and mixed disulphides to homocysteine. The enzymatic conversion of free homocysteine to s-adenosyl-L-homocysteine (sAH) is then achieved with s-adenosyl homocysteine hydrolase in the presence of excess adenosine. Competitive reactions are used to quantify total homocysteine and methods are further sub-classified based on differences in the separation and detection of sAH. These include fluorescence polarization immunoassay (FPIA), chemiluminescent immunoassay (CIA), enzyme-linked immunosorbent assay (ELISA) and other enzyme immunoassays (EIA). More recently latex enhanced agglutination immunoassays have also been developed for homocysteine quantitation.

In the Axis-Shield, Siemens Centaur and Siemens Immulite 2000/2500 (formerly Diagnostic Products) methods, synthetically derived sAH is bound to the separator system (coated wells, paramagnetic particles, or polystyrene beads). Labelled murine anti-sAH is added and in the presence of sample-derived sAH competes for binding to immobilized sAH. The concentration of labelled anti-sAH bound to the separation phase sAH is inversely proportional to the concentration of sAH derived from the original sample. An appropriate substrate and suitable conditions for colour, chemiluminescence, or fluorescence generation permit analyte quantitation.

The enzyme immunoassays methods utilize enzyme systems in the detection phase of the assay. For example, in the Axis-Shield EIA method a second rabbit anti-mouse antibody labelled with enzyme horseradish peroxidase is added to the bound fraction. Colour is generated by the addition of the substrate, tetramethylbenzidine and read at 450 nm.

Immunoturbidimetric assays, suitable for use on a coagulation analyser, have also been developed. One such method has been produced by Instrumentation Laboratory (IL). This assay converts all homocysteine forms enzymatically to sAH in the same way as described above. Plasma-derived sAH and a homocysteine conjugate with multiple binding sites compete for limited binding sites on microlatex particles coated with anti-SAH antibodies. The resultant agglutination is measured at 405 nm.

Other Enzyme Assays for Homocysteine

One of the more novel methods for homocysteine measurement is described by its manufacturer (Diazyme Europe GmbH) as an enzyme cycling system with amplification of the detection signal. The sample total homocysteine is converted to methionine and sAH by homocysteine-s-methyltransferase in the presence of s-adenosylmethionine (sAM). sAH is hydrolysed to adenosine and homocysteine which is cycled back into the conversion reaction. The adenosine formed is hydrolysed to inosine and ammonia that reacts with glutamate dehydrogenase in the presence of NADH. The conversion to NAD⁺ is detected spectrophotometrically at 340 nm. The kit is transferable to a number of platforms, including the Roche Diagnostics Hitachi, Siemens (formerly Dade) Vista Dimension and the Beckman Synchron CX, LX and DXC.

Standardization of Homocysteine Methods

Although the exact process differs between techniques, the chromatographic methods are calibrated, using gravimetrically prepared standards of homocysteine and other closely related thiol compounds, as a ratio to a stable internal standard such as 2-mercaptoethylamine. The ratio of the sample signal to internal standard is used to quantify the unknown concentration. Where this is described in any detail, most immunoassays are standardized with synthetically derived sAH that may be dispersed in buffered aqueous solution or added in modified human serum. This means that the standard material is not subject to the release, reduction and enzymatic conversion of the endogenous homocysteine but has the advantage of stability. Some enzymatic methods employ gravimetrically prepared homocysteine.

Reference Methods for Homocysteine

There are three procedures quoted on the Joint Committee for Traceability in Laboratory Medicine (JCTLM) database as high-order reference procedures for homocysteine in human serum.^79,⁸⁰ Methods include isotope dilution, gas chromatography, mass spectrometry (ID/GC/MS) and two methods based on isotope dilution, liquid chromatography, tandem mass spectrometry (ID/LC/MS/MS). These methods are used to assign homocysteine (and folate) values to frozen serum reference material (SRM 1955) produced by NIST in the USA. Lyophilized materials with values assigned by the reference procedures are not available yet.

Relative Performance of Homocysteine Methods

Proficiency testing shows that there are small but significant differences in the values produced by the enzymatic methods when compared with HPLC and immunoassays.⁸¹^–⁸⁵ While the within-run precision is acceptable for all three types of technology the between-laboratory data reveals a greater imprecision in the HPLC method group.⁸⁵ When evaluated in the light of intra-individual variability in homocysteine levels, FPIA meets the performance goal suggested by Fraser and Petersen.⁸⁶

Pre-analytical Variables in Homocysteine Testing

Plasma homocysteine is elevated in both B₁₂ and folate deficiency and is elevated in individuals with a common genetic polymorphism of methylenetetrahydrofolate reductase, C677T polymorphism, the homozygous form of which is present in 10% of Western populations and results in homocysteine levels 25% above the normal upper limit. A raised homocysteine level is an independent risk factor for vascular disease, including myocardial infarction and stroke.¹⁸ Homocysteine levels are affected by age, renal function, smoking, excessive coffee consumption and drugs, including levodopa. Correction of elevated homocysteine levels in patients with cobalamin or folate deficiency by administration of one or both vitamins provides proof of a deficiency state, which may be subclinical.

Factors affecting albumin levels will alter homocysteine because it is protein bound and venepuncture should not be performed after venous stasis or following the subject resting in supine position. Plasma is preferred because cells leak homocysteine, resulting in an increase of 10% during clot formation from uncoagulated samples. Even plasma will show an increase of 10% per hour at room temperature as a result of leakage from red cells.⁸⁷ Samples should be kept on ice and centrifuged as soon as possible, at least within 1 h.

Dynamic testing of cobalamin folate metabolism

Measurement of metabolite response at 1 week following cobalamin or folate treatment provides the opportunity to perform an in vivo dynamic function test of the cobalamin–folate metabolic pathways and confirm a diagnosis of tissue deficiency of one vitamin or the other.

Another dynamic functional test of cobalamin–folate metabolism, utilized primarily in research laboratories, is the deoxyuridine suppression test,⁸⁸ in which failure of suppression by deoxyuridine of tritiated thymidine ³H-TdR incorporation into DNA in cobalamin or folate deficient is corrected by treatment with the appropriate haematinic. (The method was described in the 9th edition of this book.)

Investigation of the cause of cobalamin deficiency

Once cobalamin deficiency has been confirmed by the finding of an unequivocally low serum B₁₂ result (with or without confirmatory raised metabolite levels and response to therapy), the aetiology of the low cobalamin should be elucidated as in Table 10.5. Gastric parietal cell antibodies are present in 90% of patients with pernicious anaemia, but this is of low specificity, being found in 15% of elderly subjects, and is therefore of little discriminatory use. Achlorhydria as a cause of cobalamin malabsorption may be suspected by the presence of raised gastrin levels.⁸⁹

Intrinsic factor antibody measurement

Principle

Two types of antibody to IF have been detected in the sera of patients with pernicious anaemia. Type I blocks the binding of B₁₂ to IF, whereas type II prevents the attachment of IF or the IF–B₁₂ complex to ileal receptors. Type II antibodies (precipitating antibodies) may be precipitated by IF–B₁₂ complex and sodium sulphate at pH 8.3 in barbitone buffer. More than 60% of patients with pernicious anaemia are reported to have type I or type II antibodies.^90,⁹¹ Automated competitive binding assays are now available and have largely superseded radiodilution assays. A number of manufacturers are developing IF antibody assays for their multianalyte immunoassay platforms.

Assay methods have been reviewed ⁹⁰ and one method for detection of types I and II IF antibody based on radiodilution competitive binding was described in the 9th edition of this book.

Intrinsic Factor Antibody Kits

Radioisotopic competitive binding assay kits that only detect blocking (type I) antibodies are being superseded by automated competitive binding assays for inclusion in the test repertoire of these multianalyte platforms. ELISA (enzyme-linked immunosorbent assay) kits that detect both type I and type II antibodies are available from a number of companies.

Principle of Binding Assay for Type I Intrinsic Factor Antibodies

Current methods for serum type I intrinsic factor antibodies (IFAb) are adaptations of the B₁₂ competitive binding assays. Patient serum is incubated with an intrinsic factor alkaline phosphatase conjugate and intrinsic factor antibody in the patient sample binds to the intrinsic factor conjugate. Paramagnetic particles coated with murine monoclonal antibody specific for the vitamin B₁₂ binding site on intrinsic factor are added to the reaction vessel and intrinsic factor conjugate that has not been blocked by sample intrinsic factor antibody binds to the mouse monoclonal antibody on the solid phase (paramagnetic particles). Separation is achieved by the application of a magnetic field to hold the particles and the excess unbound material (IF conjugate) is washed away. The chemiluminescent substrate is added to the reaction vessel and the light generated read at 530 nm in a luminometer.

This type of method is potentially subject to interference from free B₁₂ in the patient’s serum because it binds to the IF conjugate, which is then removed in the washing step. Under normal circumstances 99% of B₁₂ in serum is bound but patients with high levels of circulating free B₁₂ (>444 ng/l), as might be the case following treatment with intramuscular B₁₂, may record disproportionately elevated levels of intrinsic factor antibody.

ELISA Methods for Type I and Type II Intrinsic Factor Antibodies

Serum is incubated in the presence of IF bound to a solid phase in such a way that both the type I and type II binding sites are available for binding IFAb.⁹¹ Excess unbound serum is removed and the solid phase is further incubated with conjugate-labelled (e.g. horseradish peroxidase) antihuman IgG. Unbound conjugate is removed and substrate is added to develop the signal, which is proportional to the amount of IFAb in the original serum. The specificity of the IF antibody assay will depend on the purity of the solid-phase IF. Purified porcine intrinsic factor or recombinant intrinsic factor are used in different ELISAs and the UK NEQAS intrinsic factor antibody quality control surveys have shown variable positive rates for different types of ELISA,⁹² perhaps reflecting different sensitivities and specificity of patient sera.

Interpretation

International reference and calibration material for IFAb is not available and results of both types of method are expressed in arbitrary units or as a ratio of a cut-off deemed as positive by each manufacturer. As a consequence of the use of arbitrary units and the requirement to detect both antibodies concurrently, independent quality control material is not available.

Investigation of absorption of B₁₂

In patients who are B₁₂ deficient and who are IFAb negative, it is important to establish whether the capacity to absorb the vitamin is normal. Absorption tests should be reserved for those individuals in whom low B₁₂ levels result in genuine tissue deficiency, confirmed by supportive laboratory or clinical findings (e.g. macrocytosis, hypersegmented neutrophils, megaloblastic anaemia, neuropathy, neuropsychiatric features or elevation of cobalamin dependent metabolites) to avoid excessive investigation of ‘falsely low or indeterminate’ serum B₁₂ levels.

The withdrawal of reagents for the traditional radiolabelled Schilling tests has hampered the full investigation of patients with cobalamin deficiency.

Development of Non-Isotopic B₁₂ Absorption Tests Using Holotranscobalamin Levels and Recombinant Intrinsic Factor

Holotranscobalamin levels increase 3–4 h after oral cobalamin intake and, after 3 doses of 9 μg cyanocobalamin given 6-hourly on day 1, peak serum HoloTC was detected at 24 h in healthy adults maintained on a normal diet and given bread and juice to aid absorption of each cobalamin dose. Cellular uptake of HoloTC will be greater in deficient individuals and therefore the time course of a rise in HoloTC and the dose of cyanocobalamin may vary. Further work to optimize this approach to develop a standardized B₁₂ absorption test is ongoing.⁵⁵^–⁵⁷ Bhat et al.⁵⁸ gave 10 μg of cyanocobalamin at 6-hourly intervals ×3 and defined a normal response as a rise in plasma HoloTC of >15% and >15 pmol above baseline measurement; the majority of Asian women studied corrected low holotranscobalamin levels, suggesting dietary deficiency as the cause.

Non-Isotopic B₁₂ Absorption Test Utilizing Recombinant Intrinsic Factor in Combination with Holotranscobalamin Levels

Native (human) or bovine intrinsic factor is no longer commercially available. Hvas et al.⁵⁷ have therefore utilized recombinant intrinsic factor from 1 g of the plant leaves of transgenic Arabidopsis thaliana (Pharma Skan, Skanderborg, Denmark) to show enhanced cobalamin absorption and holotranscobalamin levels in subjects with B₁₂ deficiency. Three patterns were observed: one group showed a large increase in HoloTC (mean rise 13 pmol/l) from baseline levels with cyanocobalamin, indicating dietary deficiency; clinically deficient individuals showed only a small rise (mean 5 pmol/l) in HoloTC after cyanocobalamin, which could be enhanced (>10 pmol by day 3) by addition of intrinsic factor, indicating a lack of intrinsic factor (e.g. gastrectomy or autoimmune pernicious anaemia); and a third group with no increase in HoloTC, with or without intrinsic factor, a finding consistent with malabsorption.

Urinary Excretion of Radiolabelled B₁₂ With and Without Intrinsic Factor (Schilling Test)

Radiolabelled cyanocobalamin and bovine intrinsic factor reagents are no longer available. (Details of the principles of the Schilling tests can be found from the references ^93,⁹⁴ and in the 10th edition of this book.)

B₁₂ binding capacity of serum or plasma: transcobalamin measurement

Principle

Transcobalamin I (TCI) binds 80% or more of the total serum B₁₂ and the B₁₂–TCI complex is known as holohaptocorrin. TCII is the minor but important transport protein that delivers B₁₂ to the tissues. The B₁₂–TCII complex is known as holotranscobalamin (HoloTC) and 6–25% of total serum B₁₂ is carried in this complex. HoloTC is believed to be the first B₁₂ metabolite that decreases on inadequate B₁₂ absorption. TCII and III are normally virtually unsaturated unless an individual is undergoing B₁₂ treatment. TCII and III should therefore be measured prior to B₁₂ treatment. TCI and III (R binders) are glycosylated proteins differing in their sugar moiety. Chronic myeloid leukaemia, myelofibrosis and other myeloproliferative neoplasms are characterized by increased levels of TCI and therefore total serum B₁₂. Primary liver cancer (fibrolamellar hepatoma) is also associated with synthesis of large quantities of an abnormal form of TCI. It has been suggested that some low B₁₂ levels without evidence of B₁₂ deficiency may result from a decrease in R binder concentration.⁹⁵ Congenital absence of R binders TCI and III results in very low serum B₁₂ but no evidence of B₁₂ deficiency.⁹⁶ In congenital absence of TCII,^21,²² which results in fulminating pancytopenia and megaloblastosis within 2 months of birth, the serum B₁₂ is normal, unsaturated B₁₂ binding capacity is decreased as a result of absent TCII and B₁₂ absorption is reduced; the deoxyuridine suppression test is abnormal and corrected by B₁₂.

Unsaturated B₁₂ binding capacity and transcobalamin identification and quantitation

Measurement of unsaturated B₁₂ binding capacity is rarely undertaken and is only available in reference laboratories. (Details of the method were given in the 9th edition of this book.) It is mainly of use for detection of transcobalamin I deficiency as a cause for very low serum B₁₂ with no clinical features of cobalamin deficiency. It is also required for diagnosing the rare TCII deficiency. In addition, the assay has been used as a tumour marker for primary liver cancer.

Reference Ranges for Transcobalamins

The normal range ⁹⁷ for serum unsaturated B₁₂ binding capacity is 670–1200 ng/l; for plasma collected into EDTA-sodium fluoride it is 505–1208 ng/l,⁹⁸ for TCI, 49–132 ng/l, TCII, 402–930 ng/l and for TCIII, 80–280 ng/l.

Acknowledgements

The description of metabolic pathways for cobalamin, folate and homocysteine were assisted by Hematology Basic Principles and Practice⁹⁹ and Homocysteine in Health and Disease.¹⁰⁰ We wish to thank Annie Lee, Julie Bonser, staff at UK NEQAS Haematinics Department of Haematology and Good Hope NHS Trust for collaboration on production of international reference standards and experimental work on folate and B₁₂ assays, Dr Christine Pfeiffer at Centers for Disease Control and Prevention (CDC) Atlanta for collaborative work on the international reference method for serum folate and inter-method comparisons and Dr Susan Thorpe at NIBSC for collaborative work on international reference standards. Thanks also to family, friends and colleagues for ever-present support and encouragement.

References

1 Centers for Disease Control and Prevention. Available at: www.cdc.gov/ncbddd/B₁₂/table4.html.

2 Fyle J.C., Madsen M., Hojrup P., et al. The functional cobalamin (vitamin B₁₂)-intrinsic factor receptor is a novel complex of cubilin and amnionless. Blood. 2004;103:1573-1579.

3 Pearce J. Subacute combined degeneration of the cord. Putnam–Dana Syndrome European Neurology. 2008;60:53-56.

4 Birn H. The kidney in vitamin B₁₂ and folate homeostasis: characterization of receptors for tubular uptake of vitamins and carrier proteins. Am J Physiol Renal Physiol. 2006;291:F22-F36.

5 Waters A.H., Mollin D.L. Observations on the metabolism of folic acid in pernicious anaemia. Br J Haematol. 1963;9:319-327.

6 Jewsbury E.C. Subacute combined degeneration of the cord and achlorhydric peripheral neuropathies without anaemia. Lancet. 1954;3:307-312.

7 Healton E.H., Savage D.G., Brust J.C.M., et al. Neurologic aspects of cobalamin deficiency. Medicine. 1991;70:229-245.

8 Lindenbaum J., Healton E.B., Savage D.G., et al. Neuropsychiatric disorders caused by cobalamin deficiency in the absence of anaemia or macrocytosis. N Engl J Med. 1988;318:1720-1728.

9 Smith D., Refsum H. Vitamin B₁₂ and cognition in the elderly. Am J Clin Nutr. 2009;89:707S-711S.

10 Rush D. Folate supplements prevent recurrence of neural tube defects. FDA Dietary Supplement Task Force. Nutr Rev. 1992;50:22-28.

11 Savage D.G., Lindenbaum J., Stabler S.P., et al. Sensitivity of serum methylmalonic acid and total homocysteine for diagnosing cobalamin and folate deficiencies. Am J Med. 1994;96:239-246.