Validation of the clinical use of any diagnostic test focuses on three main principles:
- Technical feasibility of the test;
- Diagnostic performance of the test, such as sensitivity, specificity, and positive and negative predictive value both in different populations of patients and compared to the gold standard;
- Clinical utility of the test, i.e., how the results of the diagnostic test will be used to improve management of the patient.
The serum measurements of holo-TC involve the use of standard laboratory immunoassay techniques. In the first step, holo-TC in the serum sample is separated by magnetic microspheres coated with monoclonal anti-human transcobalamin antibodies. The cobalamin bound to the holo-TC is then released and measured by a competitive binding radioimmunoassay.
The diagnostic performance must be compared to the established gold standard for measuring cobalamin deficiency. This is particularly problematic, since there is currently no established gold standard. As noted in the Description section, serum levels of total cobalamin are considered poorly sensitive and specific, and there have been several reports of the serum measurements proposing serum measures of methylmalonic acid (MMA) and homocysteine as an alternative gold standard. Two general categories of patients may be considered; for example, those with symptoms suggestive of cobalamin deficiency, for whom the diagnosis may be confirmed by a treatment response to cobalamin supplementation. One possible strategy would be to develop diagnostic parameters for holo-TC (i.e., the establishment of cut-off points for normal versus low values) based on this known population, followed by remeasuring holo-TC after treatment. In a second step, the established diagnostic parameters could be applied to an independent population (representative of the United States’ population and diet) with suggestive symptoms. One population of interest is composed of asymptomatic patients who are considered at risk for cobalamin deficiency, such as those with high-risk nutritional factors (i.e., elderly patients or those with restrictive diets), or those with a predisposing disease or condition, such as gastrectomy or autoimmune disease. It is thought that identification of subclinical disease can prompt early treatment such that clinical symptoms do not develop. Given the absence of a definitive gold standard, confirmation of a diagnosis of subclinical disease is problematic.
According to the FDA decision summary, the cut-off values for holo-TC were based on a normal population instead of a population of those with known cobalamin deficiency. For example, the low value of holo-TC, 37 pmol/L, was based on a study of 303 normal Finnish individuals. This study has also been published in the peer-reviewed literature. Participants included 226 normal elderly subjects and 80 normal, non-elderly adult subjects. Patients were excluded from the trial if they had hyperhomocysteinemia, evidence of a possible cobalamin deficiency. In addition, patients in the lowest one third of holo-TC results underwent additional testing with MMA; those with elevated MMA levels were also excluded. In the normal reference population, the holo-TC range was 25–254 pmol/L with a 95% central reference interval of 37–171 pmol/L. Therefore, the cut-off value for a low result was established at 37 pmol/L. This cut-off value was then applied to the results of 107 patients with presumed cobalamin deficiency, as evidenced by different combinations of an increased plasma homocysteine or MMA level, or a low total serum cobalamin level, defining patients with potential, possible, or probable cobalamin deficiency. A total of 48% of those with presumed deficiency had a holo-TC below 37 pmol/L. The frequencies of low holo-TC levels increased with increasing pretest probability of cobalamin deficiency. For example, among the 16 patients thought to have the highest pretest probability of cobalamin deficiency, based on elevated levels of homocysteine and MMA, 100% had low levels of holo-TC. Therefore, this study used levels of homocysteine and MMA as the gold standard. Based on this standard, the sensitivity of the test was only 48% among those with potential, possible, or probable cobalamin deficiency. The authors conclude that further studies are needed to confirm the clinical utility and specificity of holo-TC in diagnosis of subclinical cobalamin deficiency.
Hvas and Nexo reported on a study of 143 subjects who were divided into four groups, those with a confirmed diagnosis of cobalamin deficiency based on a decreased total serum cobalamin (<200 pmol/L) and increased MMA (>0.70 ųmol/L), a second group thought to be normal based on normal values of total serum cobalamin and MMA, and finally two additional groups with an uncertain diagnosis due to conflicting values of total cobalamin and MMA. Although these authors used the reference interval established in the above study (i.e., 24–157 pmol/L), the cut-off for a low result was set at 50 pmol/L. Using this cut-off point, measurements of holo-TC had a sensitivity of 1.00 and specificity of 0.89 in classifying patients very likely to be, or not be, cobalamin deficient. Among the 73 patients with conflicting levels of MMA and total cobalamin, 39 had low holo-TC levels. Without a gold standard, it is difficult to interpret the results in this group with an uncertain diagnosis. As noted by the authors, it is not possible to determine whether or not holo-TC correctly classified the individual as deficient or not.
Hermann and colleagues reported on another series of patients using the same 37 pmol/L cut-off established by Loikas. This study included 93 omnivorous German controls, and several other groups of patients considered at risk for cobalamin deficiency: 111 German and Dutch vegetarians, 122 apparently health Syrians, 127 elderly Germans, and 92 patients with renal failure. In addition to holo-TC, MMA, total serum cobalamin, and homocysteine were measured. A total of 72%, 50%, and 21% of vegetarians, Syrians, and the elderly had holo-TC levels of less than 35 pmol/L. Similar to the study above, these low levels of holo-TC were associated with either normal or high levels of MMA. Conversely, high levels of MMA were associated with normal holo-TC levels in other patients. Again, it is difficult to interpret the clinical significance of these conflicting laboratory values.
There are inadequate data to establish holo-TC testing as an alternative to either total serum cobalamin or levels of MMA or homocysteine. Cut-off points for low levels of cobalamin were based on a study of a homogeneous population of 303 Finnish subjects. These values for a homogeneous population of Finnish subjects with a diet high in fish might not be able to be extrapolated to the heterogeneous American population and diet. Furthermore, these cut-off points require confirmation in a larger population of patients whose cobalamin status is unknown.
Advocates of holo-TC testing posit that this laboratory test can identify early subclinical stages of cobalamin deficiency, permitting prompt initiation of treatment, specifically supplementary cobalamin dietary supplementation. This hypothesis was not directly tested in any of the identified published literature. In the absence of a gold standard, the clinical significance of subclinical cobalamin deficiency must be further studied by understanding the natural history of this condition. Does subclinical deficiency inevitably progress to clinical deficiency? Does cobalamin supplementation normalize the values? How variable are cobalamin levels within patients? These clinical issues have not been well addressed in the literature. Finally, for all patients at risk, i.e., vegetarians, the elderly, and post-gastrectomy patients, the recommended treatment of subclinical disease is further dietary supplementation of cobalamin. This recommendation is appropriate, regardless of the level of measured cobalamin.
A MEDLINE database literature search through July 2007 did not identify any new published literature that would prompt reconsideration of the policy statement, which remains unchanged.
Serum holo-transcobalamin as a marker of vitamin B12 deficiency is not supported by evidence in the peer-reviewed medical literature that:
- permits conclusions on the effect on health outcomes;
- demonstrates an improvement in net health outcome;
- demonstrates that use of this test is as beneficial as established alternatives.