Laboratory Evaluation for Vitamin B₁₂ Deficiency: The Case for Cascade Testing

Abstract

Objective

Potential vitamin B₁₂ deficiency is a common clinical diagnostic problem, and many providers have a low threshold for initiating therapy. The goal of this study was to systematically evaluate current practice patterns regarding the laboratory evaluation of suspected vitamin B₁₂ deficiency.

Methods

This retrospective study reviewed the electronic medical records of 192 patients initiated on intramuscular vitamin B₁₂ injections.

Results

Only 12 patients had objectively documented hematologic responses: decrease of mean corpuscular volume by ≥5 fL with stable or improved hemoglobin. Another 5 patients had equivocal hematologic responses. There was one plausible neurologic response. Thus, only 18 (9.4%) of 192 patients had data supportive of a clinical response. In these 18 patients, the baseline serum B₁₂ level was ≤107 pg/mL; only 3 patients also had a baseline serum methylmalonic acid level, which was ≥1.29 μmol/L in all 3 patients.

Conclusions

Currently, only a small minority of patients initiated on intramuscular vitamin B₁₂ supplementation derive any meaningful clinical benefit. Furthermore, current testing recommendations for vitamin B₁₂ deficiency are usually not followed. Up-front ordering of a diagnostic testing cascade is recommended to improve compliance; an example is presented with decision points chosen to improve specificity for clinically evident vitamin B₁₂ deficiency without loss of sensitivity. Ultimately, a better understanding of vitamin B₁₂ physiology is needed to develop and evaluate laboratory tests that more accurately reflect true intracellular vitamin B₁₂ status.

The laboratory evaluation of possible vitamin B₁₂ deficiency is often prompted by the presence of macrocytic anemia.¹ In clinically evident vitamin B₁₂ deficiency, macrocytosis (mean corpuscular volume [MCV] >99 fL) usually precedes the development of anemia; less commonly neurologic manifestations can occur when both values are normal.² However, macrocytosis is a common nonspecific finding in adults undergoing an automated complete blood count (CBC). Even less specific are neurologic abnormalities that may include paresthesias, ataxia, dementia, and depression. Rarely do patients derive any meaningful neurologic response to vitamin B₁₂ supplementation (personal communication with author, GRS).

Published estimates of the sensitivity and specificity of serum B₁₂ measurement vary widely, largely due to the lack of a gold standard for diagnosis.^3,4 Serum B₁₂ levels below 100 pg/mL are reported to have specificity approaching 90% for diagnosing clinically-manifest deficiency.³ However, in clinical practice, with more chronically ill, elderly patients, the practical specificity may be lower.⁵ Using elevated metabolite levels (ie, serum methylmalonic acid [MMA] or homocysteine [Hcy]) to define deficiency, the “screening” serum B₁₂ cutoff used in many diagnostic algorithms has been raised to 250 pg/mL to 300 pg/mL to provide greater sensitivity.⁶ With this approach, the prevalence of vitamin B₁₂ deficiency may be as high as 15% to 20% in the elderly.⁷ Of course, most of these patients will have no hematologic or neurologic manifestations of B₁₂ deficiency (ie, subclinical deficiency).⁸ Holotranscobalamin (holoTC) has been reported to offer slightly better performance than total serum B₁₂ levels in diagnosing deficiency, but these study designs have been criticized, and there is currently no consensus that holoTC should replace conventional serum B₁₂ testing.^8,9

Given the limitations of current testing for vitamin B₁₂ deficiency, and because its administration is inexpensive and fairly innocuous, many clinicians have a low threshold for prescribing vitamin B₁₂ supplementation (usually by intramuscular injection to bypass issues with absorption). However, this approach may give a false sense of effective medical intervention for both the patient and provider, thereby diverting attention away from further evaluation and appropriate diagnosis.

In recent years, a number of diagnostic algorithms for evaluating vitamin B₁₂ deficiency have been promulgated, but adherence to these recommendations in clinical practice is unknown. Also, how well these algorithms actually perform in terms of predicting meaningful response to vitamin B₁₂ supplementation is unclear. In this study, we sought to systematically evaluate current practice patterns regarding the laboratory evaluation of suspected vitamin B₁₂ deficiency.

Methods

This study was approved with waiver of written informed consent by the Institutional Review Board of Marshfield Clinic, a large, multi-specialty, multi-site group practice with regional centers throughout central and northern Wisconsin. Medical records of 250 potential subjects were electronically identified utilizing the Healthcare Common Procedure Coding System (HCPCS) code for vitamin B₁₂ injection (J3420) and International Classification of Diseases (ICD-9-CM) codes related to vitamin B₁₂ deficiency (266.2, 281.0-1), with 125 randomly selected for manual review from each of the calendar years 2000 and 2005. Fifty-eight patients were excluded, leaving 192 cases (93 from year 2000 and 99 from year 2005) that formed the basis for this study. Reasons for exclusion were previous vitamin B12 injections within 12 months before the reference date (26 cases), documented alcohol abuse (16 cases), undergoing chemotherapy or hormonal therapy for a malignancy (9 cases), and no documentation that vitamin B₁₂ therapy was initiated (7 cases). Very few patients had well-documented neurologic exams, which generally precluded objective evaluation of neurologic improvement.

Vitamin B₁₂ was measured on the Bayer Immunol analyzer from June 1999 to August 2004 with a reference range of 185-1000 pg/mL (137-740 pmol/L). Starting in June 2002, the comment “Borderline low B₁₂ level (100-250 pg/mL); suggest serum methylmalonic acid if clinically indicated” was addended to all laboratory reports with B₁₂ levels in that range. In August 2004, B₁₂ testing was moved onto the Beckman Access/DXI instrument which uses a competitive immunoassay incorporating intrinsic factor. The correlation between the two methods was: Access = 0.966 x (Immuno1) + 12; (n = 42 lab employees). Based on complaints from Marshfield Clinic neurologists regarding a longstanding history of referrals for low B₁₂ results, the lower limit of the reference range was changed at that time to 160 pg/mL.

Methylmalonic acid in serum was determined by liquid chromatography-tandem mass spectrometry using butanol derivatization with deuterated MMA as the internal standard on an Applied Biosystems API 3000 with Analyst 1.4 software. The reference range was 0.0-0.40 μmol/L.

Plasma Hcy was assayed by an enzymatic method measuring S-adenosylmethionine catalyzed by homocysteine-S-methyltransferase. Our laboratory utilized age-adjusted reference ranges (eg, <16.5 for males 61-80 years and <14.8 for females 61-80 years).

Baseline patient data had to be within 6 months before initiation of intramuscular B₁₂ therapy. Follow-up data had to be at least 2 months after this date, but not more than 12 months. Hematologic response was defined as a decrease in MCV by ≥5 fL while the hemoglobin (Hgb) remained stable or increased. Based on baseline serum B12 and MMA testing alone, patients were placed into one of four categories derived from current testing recommendations. Minimal or no evidence of B₁₂ deficiency (Group 1) was defined as (a) no documented testing for vitamin B₁₂ status, (b) normal serum vitamin B₁₂ (>250 pg/mL) regardless of MMA value, or (c) low serum vitamin B₁₂ (any level ≤250 pg/mL) with normal MMA. Possible vitamin B₁₂ deficiency (Group 2) was defined as moderate to borderline low serum vitamin B₁₂ (100-250 pg/ mL) with no documented testing for MMA. Probable vitamin B₁₂ deficiency, unconfirmed (Group 3) was defined as very low serum vitamin B₁₂ (<100 pg/mL), but no documented testing for MMA. Lastly, vitamin B₁₂ deficiency, confirmed by laboratory criteria (Group 4) was defined as low serum vitamin B₁₂ (≤250 pg/mL) with elevated MMA.

Analyses for this study were primarily descriptive, the principle goal being to present the clinical experience at a single institution with respect to the adequacy of laboratory evaluation for patients on vitamin B₁₂ supplementation. Associations among laboratory tests were examined graphically with scatter plots and were measured using the Spearman rank correlation coefficient. Results in this report were deemed statistically significant at the 5% level (P<0.05) without adjustment for multiple comparisons.

Results

Patients started on intramuscular vitamin B₁₂ were mostly female (64%) and elderly (median age 75 years). Laboratory data were measured before or at the time of diagnosis (ie, at baseline) in the majority of patients (table 1). Figures 1, 2, and 3 show the frequency distribution of baseline values for serum vitamin B₁₂, Hgb, and MCV, respectively. Baseline Hgb was inversely correlated with serum creatinine (Cr): correlation coefficient r = -0.21 (P=0.01). Forty patients had baseline MMA levels; of these, 35 also had baseline Hgb and MCV data. Serum MMA was directly correlated with Cr: r = 0.33 (P=0.04). Closer inspection of the relationship between MMA and Cr showed that 11 of 12 patients with a baseline serum Cr >1.6 mg/dL had an elevated serum MMA (>0.40 μmol/L), compared with only 7 of 26 having a Cr <1.5. Although MCV was significantly correlated with MMA (r =0.43, P=0.01), no cutoff value for MCV was apparent that could usefully identify patients having an increased MMA. The inverse correlation between MCV and B₁₂ did not quite reach statistical significance (r = -0.15, P=0.07). No significant relationships were identified between MMA and B₁₂ or MMA and Hgb.

Table 1.

Demographics and baseline laboratory results.

Patients (n)	192
Female (%)	64.1
Age
median (years)	75.2
range	18–93
B12
# tested	171
median (pg/mL)	168
range (pg/mL)	26–2000
Hgb
# tested	165
median (g/dL)	13.1
range (g/dL)	7.5–17.2
MCV
# tested	165
median (fL)	91.7
range (fL)	63–131
MMA
# tested	40
median (μmol/L)	0.4
range (μmol/L)	0.1–16
Serum creatinine
# tested	179
median (mg/dL)	1.0
range (mg/dL)	0.5–5.4

Hgb, hemoglobin; MCV, mean corpuscular volume; MMA, methylmalonic acid

Figure 1.

Frequency distribution of baseline values for serum vitamin B₁₂.

Figure 2.

Frequency distribution of baseline values for serum hemoglobin (Hgb).

Figure 3.

Frequency distribution of baseline values for mean corpuscular volume (MCV).

Of the 192 patients included in the study, B therapy was initiated in calendar years 2000 and 2005 for 93 and 99 patients, respectively. Their stratification by diagnostic category (Groups 1-4) is shown in table 2. The lab report addendum suggesting testing of serum MMA for B₁₂ levels in the 100-250 pg/mL range did result in 29 of 99 patients (29%) having a baseline serum MMA in 2005, compared with 11 of 93 (12%) in 2000. However, this had little effect on patient classification, and surprisingly, relatively fewer patients with probable or confirmed deficiency (Group 3 or 4) were identified in 2005 (13/99, 13%) than in 2000 (22/93, 24%). No plausible explanation (other than chance) was identified, and therefore, data from 2000 and 2005 were combined in subsequent analyses reported below.

Table 2.

Diagnostic category by year.

	Year
	2000	2005
Group*	n (%)	n (%)
1: Minimal/no evidence	33 (35.5)	35 (35.4)
2: Possible	38 (40.9)	51 (51.5)
3: Probable	18 (19.4)	6 (6.1)
4: Confirmed	4 (4.3)	7 (7.1)
Total patients (N)	93	99

*Groups were defined as follows: Group 1 – (a) no documented testing for vitamin B₁₂ status, (b) normal serum vitamin B₁₂ (>250 pg/mL) regardless of MMA value, or (c) low serum vitamin B12 (any level ≤ 250 pg/mL) with normal MMA. Group 2 – moderate to borderline low serum vitamin B₁₂ (100–250 pg/mL) with no documented testing for MMA. Group 3 (unconfirmed) – very low serum vitamin B₁₂ (<100 pg/mL), but no documented testing for MMA. Group 4 (confirmed by laboratory criteria) – low serum vitamin B₁₂ (≤250 pg/mL) with elevated MMA.

Only 88 patients had both baseline and follow-up B levels assessed. As illustrated in figure 4, B₁₂ levels rose for the vast majority of the patients in Groups 2, 3, and 4, providing laboratory evidence that supplementation was occurring. Follow-up B₁₂ levels were ≥100 pg/mL for all but one individual. A significantly greater number of patients had both baseline and follow-up Hgb and MCV assessed: 136 and 131, respectively. There were no significant correlations between the changes in these variables and the B12 increment (data not shown). In contrast, the change in MCV was significantly correlated with the baseline B₁₂ level (r = 0.27, P=0.003); this relationship is illustrated in figure 5. Vertical lines at 100 pg/mL and 250 pg/mL indicate the cutoffs used for group assignment. Data points in the lower left quadrant represent patients in whom the baseline B₁₂ was <100 pg/mL, and MCV decreased as would be expected with a hematologic response to B₁₂ supplementation.

Figure 4.

Study of vitamin B₁₂ supplementation: B₁₂ response group and time.

Figure 5.

Study of vitamin B₁₂ supplementation: mean corpuscular volume (MCV) change versus baseline B₁₂.

Remarkably, 35.4% of patients had no laboratory evidence to justify therapeutic intervention (Group 1), and another 46.3% had only possible B₁₂ deficiency (Group 2). Only two of these patients had a hematologic response. Notably, these two patients had the lowest baseline B₁₂ values of all the patients in either Group 1 or 2 (103 and 107 pg/mL).

Twenty-four patients had probable B₁₂ deficiency defined as a baseline B₁₂ <100 pg/mL without measuring MMA (Group 3). Eighteen of these had both baseline and follow-up Hgb and MCV data and are included in figure 5. Eight patients met the criteria for hematologic response, and another four had equivocal hematologic responses (eg, incomplete data or changes not quite fulfilling those criteria.) The one patient with a plausible neurologic response was in this group. She was a woman, age 68 years, in whom paresthesias of both feet had resolved when she was examined 6 weeks after initiation of B₁₂ therapy; criteria for hematologic response were not met.

Eleven patients had “confirmed” B₁₂ deficiency defined as a serum B ≤250 and serum MMA >0.40 μmol/L (Group 4). All but one had baseline Hgb and MCV data, but just seven (shown in figure 5) had follow-up hematologic data. Only two of these seven patients had well-documented hematologic responses. Three of the other five patients (all age 75 years or greater) had a serum Cr ≥1.6 that could explain the slightly elevated serum MMA. Four patients had inadequate data to assess for a hematologic response; only one had baseline values suggestive of clinical B₁₂ deficiency. In summary, only three of the eleven patients with confirmed B₁₂ deficiency had hematologic data to corroborate clinical deficiency.

Of the 40 patients (21%) who had baseline serum MMA levels, five had a serum B₁₂ <100 pg/mL, where one could arguably initiate vitamin B₁₂ without further testing. One of these patients did not have corroborative evidence of B₁₂ deficiency (normal MMA, Hgb, and MCV). Only 11 (5.7%) patients had Hcy testing before initiation of B₁₂ therapy. Two of these (with levels of 74.3 μmol/L and 27.2 μmol/L, respectively) appeared to have clinical B₁₂ deficiency. Notably, three patients with levels between 19.1 and 28.8 μmol/L had no corresponding laboratory evidence or clinical response to support B₁₂ deficiency; one of these three was clearly iron deficient (serum ferritin = 3 ng/mL). Too few patients had follow-up MMA or Hcy levels to evaluate for so-called metabolic responses.

Baseline intrinsic factor (IF) antibodies, which would support a more specific diagnosis of pernicious anemia, were evaluated in 15 patients and were positive in 5 patients. In one patient (DAT-positive hemolytic anemia responsive to prednisone), this was a false positive. None of the ten patients with a negative IF antibody test had laboratory evidence or a clinical picture of B₁₂ deficiency. The Schilling test, which has become largely unavailable in recent years, was performed on only four patients. It was reportedly positive in three; two of these three had evidence of a hematologic response to B₁₂ supplementation.

An interesting observation was that 37 patients (27 females and 10 males) had a history of gastrointestinal surgery that may have predisposed them to malabsorption of certain nutrients including vitamin B₁₂. Of these surgical procedures, 24 were gastric bypass or gastrectomy (the vast majority for morbid obesity), and 20 of these patients were women. Of the thirty-five patients that had baseline serum B₁₂ levels, five were <100 pg/mL, 27 were between 100 pg/mL and 250 pg/mL, and three were >250 pg/mL. Only three patients had evidence of a hematologic response.

Table 3 summarizes the findings in 27 patients who were started on B₁₂ therapy but found to have another cause of anemia. None of the 27 patients in table 3 had clinical data to suggest a meaningful response to B₁₂ therapy.

Table 3.

Other causes of anemia and their B₁₂ levels.

Cause	Number of patients	Number with B12 ≤250 pg/ml	Range of B12levels ≤250 pg/ml
Iron deficiency	16	13/15*	57-238
Renal insufficiency	3	3/3	143-238
Autoimmune hemolytic anemia	3	3/3	143-177
Myelodysplastic syndrome	2	1/2	145
Plasma cell myeloma	2	2/2	152, 166
Folate deficiency	1	1/1	86

*One iron deficiency patient without a reliable baseline serum B₁₂ level.

Surprisingly, iron deficiency was nearly as common as clinical B₁₂ deficiency in this cohort of 192 patients started on B₁₂ injections. In 10 patients this was well-documented, and it was strongly suspected in another six in whom there was incomplete laboratory documentation. In 13 of 15 patients with baseline B₁₂ levels, they were 250 pg/mL or below (ranging from 57 pg/ mL to 238 pg/mL). The highest MCV in an iron deficient patient was 90.5 fL; whereas, the lowest MCV in a patient with documented or equivocal clinical response to B₁₂ supplementation was 91.8 fL.

Three patients had renal insufficiency as the likely cause for their anemia, with serum Cr values between 3.0 mg/dL and 5.4 mg/dL. Three patients had well-documented autoimmune hemolytic anemia with a positive direct antibody test; B₁₂ levels ranged between 143 pg/mL and 177 pg/mL, and all three had an elevated MCV due to reticulocytosis. Another patient had a spuriously high MCV of 122.8 fL due to cold agglutinin disease; his MCV data were deleted from the statistical analysis and table 3. Two patients likely had a myelodysplastic syndrome (MDS) explaining their macrocytic anemia. Two patients were diagnosed with plasma cell myeloma within 2 months after starting B₁₂ therapy.

Folate testing, either serum folate or red cell folate, was done in a small minority of patients in this study, and only one patient was identified with possible deficiency. This woman, age 91 years, had a slightly low serum folate level of 3.6 ng/ mL (reference range 4-20 ng/mL), and baseline values were B 86 pg/mL, Hgb 11.6 g/dL, and MCV 98.2 fL with no follow-up labs. Her decreased sensation below the knees bilaterally was unchanged with B12 and folate supplementation.

In five patients, drugs may have been the cause of borderline to mild macrocytosis (MCV ≥96.1 fL) without anemia. The implicated drugs were chemotherapeutic agents in two, phenytoin in one, both phenytoin and methotrexate in one, and hydroxyurea in one. No patients were identified in which hypersegmented neutrophils were reported in the electronic medical record within 6 months before initiating intramuscular vitamin B₁₂ supplementation.

Discussion

The elderly age distribution and female predominance were expected given the demographics of vitamin B₁₂ deficiency. Also, some of the female predominance may be attributable to the higher frequency of gastric bypass surgery in women; 20 of 24 such patients were women.

The positive correlation between serum MMA and renal function is well known.¹⁰ Caution is advised in interpreting serum MMA levels in patients with renal insufficiency (eg, Cr >1.6 mg/dL). In one study from the United Kingdom, average MMA levels increased from 0.25 mmol/L in people age 65 years to 74 years, to 0.38 mmol/L in people >85 years.¹¹ Declining renal function and slightly lower B₁₂ levels only partially accounted for this change. Elevated serum Hcy is another metabolite indicator of B₁₂ deficiency but is less specific than MMA, and there are not well established diagnostic cutoffs. Like B₁₂, Hcy is largely cleared by the kidneys, and levels are higher in patients with renal insufficiency.¹²

How to manage patients with a history of gastrointestinal surgery is unclear. This study could not evaluate the frequency with which low vitamin B₁₂ levels are encountered in this patient cohort, but it seems fairly common. Notably, these patients rarely have any meaningful clinical response to parenteral vitamin B₁₂ supplementation. Some might argue that a low threshold for B₁₂ therapy is reasonable as a preventative measure, rather than waiting for symptomatic deficiency to develop.¹ This issue deserves further study, particularly with the recent increase in gastric procedures for morbid obesity.

The quagmire surrounding vitamin B₁₂ deficiency contrasts sharply with iron deficiency, which is much more common, readily confirmed by laboratory testing, and predictably responsive to supplementation. Notably, iron deficiency was nearly as common as B₁₂ deficiency in this cohort of patients started on B₁₂ injections. The rarity of folate deficiency is consistent with recent data from the United States and Canada, pointing to the introduction of widespread folic acid fortification of many foods.^13,14

Hematologists and hematopathologists frequently encounter patients with low B₁₂ levels that are found to have another cause for their macrocytosis with or without anemia (eg, hemolysis with reticulocytosis, drug-related macrocytosis, MDS, or other neoplasms). Incorrectly attributing a macrocytic anemia to B₁₂ deficiency delays further evaluation that could uncover the true underlying cause, and of course, appropriate treatment is delayed as well.

Myelodysplastic syndromes, particularly 5q syndrome, often present with a macrocytic anemia. The medical records of 13 patients (not included in the 192 patient cohort) with MDS diagnosed at Marshfield Clinic in which cytogenetics revealed a 5q deletion (accompanied by noncomplex cytogenetic abnormalities in two) were reviewed. All 11 of the patients tested at baseline had a normal serum B₁₂ level >250 pg/mL. Although these data suggest that low serum B₁₂ levels are uncommon in patients with MDS, one author (GRS) has anecdotally encountered several MDS patients with low B₁₂ levels that were unresponsive to supplementation. Further testing for MMA could clarify the situation in most cases.

The dramatic presentation and response of patients with fullblown vitamin B₁₂ deficiency provides some insight into recognizing diagnostic pitfalls. Two recent cases of pernicious anemia encountered by one author (GRS) during the preparation of this manuscript serve as examples. Both patients had baseline B₁₂ levels <50 pg/mL, positive intrinsic factor (IF) antibodies, and severe macrocytic anemia (Hgb <5.8 g/dL and MCV >134.5 fL) with striking poikilocytosis that morphologically raised some concern for an MDS (possibly with myelofibrosis). In the one patient tested, MMA was 32.8 μmol/L and Hcy 94.9 μmol/L. Both patients had tremendous elevations in their lactate dehydrogenase level (>2600 U/L) and slight to moderately increased indirect bilirubin, which raised consideration of a hemolytic anemia. In fact, one patient had a positive direct antiglobulin test (albeit weak reactivity) and an increased reticulocyte count (but not nearly enough to explain the severe macrocytosis). The other patient had a low haptoglobin level. Within 3 weeks of starting parenteral B₁₂ therapy, Hgb and MCV of both patients had normalized (>12.2 g/dL and <97.3 fL, respectively).

In the electronic medical record, the reporting of hypersegmented neutrophils was of no help in diagnosing vitamin B₁₂ deficiency. This concurs with most reports in the literature stating that the presence of hypersegmented neutrophils on the peripheral blood smear is neither sensitive nor specific for this purpose.¹⁵ However, careful review of a peripheral blood smear by a hematologist or hematopathologist may provide helpful clues regarding the underlying cause of a macrocytic anemia.

Approximately 75% of serum B₁₂ is bound to haptocorrin (formerly called transcobalamin I) whose function is unknown.⁹ Therefore, total serum B₁₂ levels largely reflect B₁₂ that is not bio-available. Total transcobalamin, consisting of holoTC and apo-transcobalamin, is the major carrier protein in the plasma/serum that delivers B₁₂ to the tissues. Conceptually, holoTC should more accurately reflect intracellular B₁₂ levels. Sequencing of the TCN1 gene for haptocorrin has identified a 999G>T polymorphism with heterozygosity in about 12% of individuals of European ancestry.¹⁶ Based on limited familial studies, heterozygosity for this polymorphism is associated with slightly lower total haptocorrin and, hence, slightly lower total serum B₁₂ levels. This polymorphism may help explain why whites, on average, have lower total haptocorrin and B₁₂ levels than blacks. A 776C>G polymorphism in the transcobalamin gene (TCN2) that affects holoTC has been described as commonly occurring in the Portuguese population.¹⁷ Extremely rare mutations in the TCN2 gene may result in severe intracellular cobalamin depletion.¹⁸ More work needs to be done to sort out the genetics and pathophysiology related to cobalamin deficiency.¹⁹

Some, but not all, studies have found marginally better performance (as assessed by receiver operator curves) by holoTC when compared with serum B₁₂.^9,20–22 However, the “gold standards” for classifying patients as B₁₂ deficient have been dubious (ie, elevated MMA or Hcy, with minimal or no clinical correlation). One study used red blood cell (RBC) cobalamin as the “gold standard” and found that holoTC slightly outperformed B₁₂. Curiously, techniques to measure RBC cobalamin were reported four decades ago and yet are still not widely available. Rare patients with pernicious anemia have spurious elevations of vitamin B₁₂ when using methods based on competitive binding of serum B₁₂ with reagent intrinsic factor.²³

A recent National Health and Nutrition Examination Survey (NHANES) roundtable recommended conventional serum B₁₂ over holoTC at this time, since holoTC methods are relatively new and would benefit from additional performance studies.²⁴ Furthermore, the same roundtable summary expressed caution regarding the common assumption that detecting subclinical deficiency early is important to enable public health interventions that prevent its progression to clinical deficiency. These experts questioned the value of further hypothesis generating epidemiologic studies and stated that prospective clinical trials are urgently needed.^8,25,26

From a cost-effective public health standpoint, some might advocate widespread oral vitamin B₁₂ supplementation without laboratory testing to provide insurance against subclinical B₁₂ deficiency.²⁷ Although the general public and commercial health food suppliers have a seemingly insatiable appetite for “fountain of youth” dietary supplements, it seems premature to recommend mandatory fortification of the food supply without well-conducted, randomized clinical trials. Several patients in the current study were given vitamin B₁₂ injections solely at their request, and sometimes for psychological benefit; patient education would seem more appropriate.

In terms of future studies, a relatively small, well designed, prospective clinical trial may be enlightening. Patients could initially be identified by laboratory testing (eg, a cohort of patients with serum B₁₂ levels of <200 pg/mL). A comprehensive battery of laboratory tests could be obtained both at baseline and, with the exception of genetic tests, at defined intervals once treatment had been initiated (eg, MMA, Hcy, holoTC, a full CBC, IF antibodies, ferritin, serum iron/total iron binding capacity, serum folate, RBC folate, and sequencing of the TCN1 and TCN2 genes). For study purposes, perhaps the Schilling test could even be resurrected to better understand the absorption issues. A careful neurologic exam would be documented by a neurologist and repeated at prescribed intervals. Treatment would be standardized.

This retrospective study has several limitations. Patients were selected on the basis of clinical initiation of intramuscular B treatment. Thus, this was not a random or representative sample of the general population. Laboratory testing before initiation of therapy was inconsistent, and monitoring even more inconsistent. Also, there was no standard treatment protocol, and numerous patients were incompletely treated; either they or their provider discontinued the therapy. In patients deemed to have had a laboratory response to B supplementation, there was no control group of untreated or placebo-treated patients, so the possibility that high MCV values decreased because of regression toward the mean cannot be ruled out. It was difficult during chart review to capture all confounding variables (eg, not having stable baseline or follow-up CBC data). This could be due to blood loss, medication changes, or other alterations in clinical status. Comprehensive neurologic exams were rarely conducted, so there may have been some patients with small objective improvements that were overlooked. This data may be considered somewhat dated; however, there have been no significant changes in B₁₂ testing over the past decade.

In summary, of 192 patients started on intramuscular B injections, only 35 (18.3%) had laboratory evidence supporting “probable” or “confirmed” deficiency. Furthermore, only 18 patients (9.4%) had data supportive of a clinical response. In these 18 patients, the baseline serum B level was ≤107 pg/mL; only three had a baseline MMA level, which was ≥1.29 μmol/L for all three. In terms of diagnosing clinically evident vitamin B deficiency with a reasonable expectation of clinical response, these limited data suggest that current diagnostic algorithms could be modified to increase specificity without sacrificing sensitivity. Additional studies are needed to corroborate these findings and further explore the benefit (if any) in detecting and treating “subclinical” vitamin B₁₂ deficiency.

The following are offered as practical suggestions that providers and laboratories may find useful.

1. Do not test patients for B deficiency without a clear clinical indication; ie, macrocytosis (usually with anemia) or (less commonly) neurologic signs or symptoms potentially referable to vitamin B₁₂ deficiency. Patients with microcytosis (or even an MCV <90 fL) should not undergo B₁₂ testing solely for anemia.

2. No single test establishes a diagnosis of vitamin B₁₂ deficiency. Laboratory cutoffs vary. In this study, all patients with plausible hematologic or neurologic responses had baseline B₁₂ levels of ≤107 pg/mL. Yet, even very low serum B levels (<100 pg/mL) and B₁₂ deficiency “confirmed” by an elevated MMA should be viewed cautiously; the complete clinical picture needs to be considered. Keep in mind that patients with serum Cr values ≥1.6 mg/dL frequently have slightly elevated serum MMA and Hcy. IF antibodies have good specificity with a sensitivity of about 50% by most accounts,³ and this assay is suggested when the initial laboratory data suggests B₁₂ deficiency.

3. Laboratories should recommend a B₁₂ testing cascade that includes MMA and IF antibodies as indicated to follow-up low B₁₂ levels without requiring a separate order by the provider or a second blood draw. An example of such a cascade is shown in figure 6. To increase specificity, the serum B₁₂ cutoff has been conservatively adjusted to our laboratory’s lower reference range cutoff (160 pg/mL), and the serum MMA cutoff has been increased to 0.80 μmol/L. Sign-out of low serum B₁₂ results by a hemato-pathologist or clinical chemist (much like serum protein electrophoresis) can provide individualized interpretative comments enhancing communication to providers. In our laboratory, which averages between 50 and 60 serum B₁₂ results per day, only about 4% of samples have values <160 pg/mL, so this could be accomplished with relatively little additional professional effort.

4. Whatever the exact therapeutic regimen (not addressed in this study), laboratory reevaluation at about 2 to 3 months by repeating the CBC, B₁₂ and MMA testing is recommended. For patients with macrocytic anemia, if the CBC has remained the same and the patient has been compliant in receiving B₁₂ therapy, then it may be time to reassess. Have the B₁₂ and MMA (or Hcy) levels corrected? What is being gained by the patient in continuing parenteral therapy only to “treat” the B₁₂ and MMA numbers? Other diagnoses (eg, MDS, hemolytic anemia) should be considered.

Figure 6.

Testing cascade for vitamin B₁₂ deficiency.