Elevated Methylmalonic Acid and Total Homocysteine Levels Show High Prevalence of Vitamin B₁₂ Deficiency after Gastric Surgery

Methods

Between September 1991 and March 1993, 65 patients who had had gastric surgery were identified at the Philadelphia Veterans Affairs Medical Center. These patients were identified either by review of gastrointestinal radiographs, surveys of the house-staff assigned to the medicine and surgery inpatient services, or referral of outpatients from physicians in the medical clinic. Four of the 65 patients were excluded: Three were receiving vitamin B₁₂ therapy, and one had a hepatoma. Hepatoma can produce increased levels of vitamin B₁₂-binding protein, which may complicate interpretation of serum vitamin B₁₂ levels.

Patients who had not had gastric surgery (controls) were drawn from 127 consecutive patients attending one author's Philadelphia Veterans Affairs Medical Center clinic between November 1992 and March 1993. One hundred seven controls participated, and 20 either declined to participate or did not complete the required blood tests.

We determined the type of gastric surgery that had been done either from patient reporting or by reviewing radiologic, endoscopic, or surgical records. In most patients (51 of 61), we determined the year surgery had been done from patient report or chart review. For patients who could not provide the year of surgery but could specify the decade, we used the mid-decade year. For example, if the patient said that the surgery had been done in the 1950s, we recorded the year as 1955.

Serum vitamin B₁₂ and folate levels were determined at the Philadelphia Veterans Affairs Medical Center using a commercially available radioligand kit (Bio-Rad, Diagnostics Group, Hercules, California). In the hospital's laboratory, normal values for vitamin B₁₂ and folate levels were 171 to 840 pmol/L and 5 to 39 nmol/L, respectively. The remaining serum samples were frozen at − 20 °C and were shipped to Denver so that serum methylmalonic acid and total homocysteine levels could be analyzed by the stable isotope dilution gas chromatography-mass spectrometry method [27-30]. The normal range for serum methylmalonic acid levels (determined in 50 normal blood donors 18 to 65 years of age) is 73 to 271 nmol/L, and the normal range for serum total homocysteine levels is 5.4 to 16.2 µmol/L [22].

Vitamin B₁₂ deficiency was defined as one of the following: 1) a serum vitamin B₁₂ level less than 221 pmol/L and an elevated methylmalonic acid level; 2) a serum vitamin B₁₂ level less than 221 pmol/L and a total homocysteine level that decreased after vitamin B₁₂ therapy; or 3) in patients unavailable for treatment, a serum vitamin B₁₂ level less than 221 pmol/L, a folate level greater than 9 nmol/L, and an elevated total homocysteine level.

Hemoglobin level, hematocrit, and mean corpuscular volume were measured by automatic devices. Macrocytosis was defined as a mean corpuscular volume of 95 fL or less. The peripheral smears of 71% of patients (43 of 61) and 88% of controls (94 of 107) were reviewed by one hematologist who was blinded to each participant's vitamin B₁₂ level, hemoglobin level, hematocrit, and gastric surgery status. Hypersegmentation was defined as five neutrophils with five or more lobes or one neutrophil with six lobes per 100 cells counted.

Treatment

Vitamin B₁₂ treatment generally consisted of daily intramuscular injections of 1000 µg of vitamin B₁₂ for 5 days, followed by monthly injections. Folic acid was given orally, 1 mg/d. Serum vitamin B₁₂, folate, methylmalonic acid, and total homocysteine levels were measured 1 to 6 weeks after treatment.

Statistical Analysis

Data were examined to determine whether the variables were suitable for parametric analyses. Although relatively modest, the skew for the numeric variables necessitated that several variables be transformed to logs for entry into two-way analysis of variance or be subjected to nonparametric analyses.

The comparison between patients and controls for levels of vitamin B₁₂, folate, methylmalonic acid, and total homocysteine was done by two-factor analysis of variance on log-transformed variables. In each analysis of variance, race was included as a factor (along with study group) to control for possible race-by-group interactions.

We used unpaired t-tests or Mann-Whitney U tests to do comparisons of other numeric variables, such as hemoglobin and mean corpuscular volume; comparisons between other groups, such as patients with a positive and patients with a negative peripheral blood smear; and comparisons between deficient and nondeficient patients. We used chi-square tests to compare groups on dichotomous variables (such as white patients compared with black patients). Spearman correlations were used to assess the association between the time since surgery and other variables.

The Human Studies Subcommittee and the Research and Development Committee of the Philadelphia Veterans Affairs Medical Center approved the study.

Results

Clinical Characteristics

The 61 patients (who had had gastric surgery) and 107 controls (who had not) were similar in the ratio of men to women (60:1 compared with 104:3), age, and race (Table 1). The indications for surgery included peptic ulcer disease (56 patients), obesity (3 patients), gastric cancer (3 patients [2 of whom had previously had surgery for peptic ulcer disease]), and gastric lymphoma (1 patient). The type of gastric surgery could be determined in 36 of 61 patients (59%). The types of surgery were Billroth II (23 patients), repair of perforated ulcer (6 patients), vagotomy and pyloroplasty (2 patients), gastric bypass or gastric banding for obesity (3 patients), Billroth I (1 patient), total gastrectomy (1 patient), and unknown (25 patients). Surgery had been done a median of 20 years before the study began (range, 1 month to 50 years). Four patients had had surgery after 1990; 10, between 1980 and 1989; 16, between 1970 and 1979; 12, between 1960 and 1969; 4, between 1950 and 1959; and 5, between 1940 and 1949.

Prevalence of vitamin B₁₂ Deficiency

On the basis of transformed geometric means, patients were determined to have had substantially lower serum vitamin B₁₂ levels (162 pmol/L compared with 300 pmol/L; P = 0.004) and folate levels (17.2 nmol/L compared with 22.2 nmol/L; P = 0.04) than controls. The patients also had substantially higher methylmalonic acid levels (234 nmol/L compared with 117 nmol/L; P < 0.001) and total homocysteine levels (12.3 µmol/L compared with 10.9 µmol/L; P = 0.02). Almost half of the patients (48%) and 21% of the controls had vitamin B₁₂ levels less than 221 pmol/L (P < 0.001); 31% of the patients and 9% of the controls had vitamin B₁₂ levels less than 148 pmol/L (P < 0.001) (Table 2). Substantially more patients ([19] [31%]) than controls ([2] [2%]) met one of the definitions of metabolically confirmed vitamin B₁₂ deficiency (P < 0.001) (Table 2).

Of the 21 vitamin B₁₂-deficient participants, 8 (38.1%) had both increased methylmalonic acid and total homocysteine levels, 9 (42.8%) had increased methylmalonic acid levels alone, and 4 (19.1%) had increased total homocysteine levels alone. Both deficient controls had elevated methylmalonic acid and normal total homocysteine levels. In 12 of the 19 patients with vitamin B₁₂ deficiency (63%), total homocysteine levels were 2 standard deviations above the mean (> 16.2 µmol/L); in 5 (26%), values were 3 standard deviations higher than the mean (> 21.3 µmol/L).

Measurements of methylmalonic acid and total homocysteine levels were essential to the accurate diagnosis of true vitamin B₁₂ deficiency in the 51 patients with serum vitamin B₁₂ levels less than 221 pmol/L and were especially important in the controls. Sixty-five percent of the 29 patients with serum vitamin B₁₂ levels less than 221 pmol/L were actually vitamin B₁₂-deficient, as indicated by elevated metabolite levels (6 of 10 patients with vitamin B₁₂ levels less than 221 pmol/L and 13 of 19 patients with vitamin B₁₂ levels less than 148 pmol/L). Controls had nonspecific serum vitamin B₁₂ levels. Metabolite levels were elevated in only 9% of controls (2 of 22) (one control had a vitamin B₁₂ level less than 148 pmol/L, and the other had a vitamin B₁₂ level between 148 and 221 pmol/L). If serum vitamin B₁₂ levels had been the only criterion on which the decision to institute lifelong therapy was based, 22 controls would have been treated, but only 2 would have benefited.

The median year of surgery was 1971 for the vitamin B₁₂-deficient patients and 1972 for those without deficiency (P > 0.2). We saw no correlation between the years since surgery and serum vitamin B₁₂, methylmalonic acid, or total homocysteine levels: All P values were greater than 0.1, and all Spearman correlation coefficients were small (absolute values less than 0.17). Two patients developed evidence of vitamin B₁₂ deficiency during the study (that is, serial determinations indicated that their vitamin B₁₂ levels decreased and their serum metabolite levels increased). These patients' surgeries had been done 4 and 30 years, respectively, before the patients developed vitamin B₁₂ deficiency.

Most (14 of 19) of the vitamin B₁₂-deficient patients were elderly (65 to 83 years of age), but this was because most study patients (patients who had had gastric surgery) were elderly (39 of 61). The difference in the prevalence of vitamin B₁₂ deficiency between patients who were older than 65 years of age (36%) and those who were younger than 65 years of age (23%) was not statistically significant (P = 0.14). However, 36% (14 of 39) of the patients aged 65 to 83 years had vitamin B₁₂ deficiency compared with 2% (1 of 58) of the older controls (P < 0.001).

Prevalence of Folic Acid Deficiency

One patient (1.6%) and one control (1%) had folic acid deficiency. In both of these participants, the elevated total homocysteine levels returned to normal with oral folate therapy.

Hematologic Evaluations

Patients had lower mean hemoglobin levels than did controls (mean ±SD, 127 ± 26 g/L compared with 147 ± 14 g/L; P < 0.001) (Table 3). In addition, vitamin B₁₂ deficiency was associated with substantial anemia; the mean hemoglobin level was almost 20 g lower in the vitamin B₁₂-deficient patients (114 ± 30 g/L) than in the patients without deficiency (133 ± 23 g/L). Patients and controls had similar mean corpuscular volumes (89 ± 9.0 fL and 90 ± 5.7 fL, respectively), but 20% of patients and 4.3% of controls had a mean corpuscular volume less than 80 fL (P = 0.001). Although we did not measure iron status, the lower mean corpuscular volume suggests that the patients developed iron deficiency after gastric surgery. Patients with vitamin B₁₂ deficiency and patients without deficiency had similar mean corpuscular volumes and ranges (90 fL [range, 68 to 113 fL] compared with 88 fL [range, 71 to 106 fL]). A trend toward macrocytosis was seen in the vitamin B₁₂-deficient patients compared with patients who had no deficiency (6 of 19 [32%] compared with 7 of 42 [17%]) (P = 0.09).

Hypersegmentation was noted in 7 of 43 patients, 6 of 94 controls, and only 2 of the 11 vitamin B₁₂-deficient patients who had peripheral smears. Neither vitamin B₁₂-deficient control had hypersegmentation, but the folate-deficient patient and control did have hypersegmentation.

Response to Replacement Therapy

The initial methylmalonic acid and total homocysteine levels in the 21 deficient participants (19 patients and 2 controls) are shown in Figure 2. Serum methylmalonic acid and total homocysteine levels markedly decreased in each of the 15 participants who were available for treatment. Vitamin B₁₂ therapy decreased the elevated total homocysteine levels, even in participants whose levels were already in the top part of the normal range. We could not treat 6 of the 21 participants because 3 died and 3 were lost to follow-up.

Elevated Metabolite Levels in Participants with Normal vitamin B₁₂ Levels

Eight patients (13%) and 8 controls (8%) had elevated metabolite levels and serum vitamin B₁₂ levels of at least 221 pmol/L. Two of these 16 participants (1 patient and 1 control) had folate deficiency. Of the remaining 14 participants, 3 died and 3 were lost to follow-up. Of the 8 remaining patients and controls, 5 (3 patients and 2 controls) participated in a therapeutic trial. In 4 of these 5 participants, the elevated metabolite levels decreased substantially with vitamin B₁₂ treatment. The fifth participant, who only had an elevated total homocysteine level, was treated sequentially with folic acid and then vitamin B₁₂. Despite this treatment, the participant's total homocysteine level remained elevated.

Discussion

Our results indicate that the prevalence of biochemically confirmed vitamin B₁₂ deficiency in patients who have had gastric surgery is much higher than previously appreciated. We detected vitamin B₁₂ deficiency in 31% of patients who had had gastric surgery compared with 2% of controls of similar age and race. In patients 65 to 83 years of age, the prevalence increased to 36%; in controls, however, the prevalence remained unchanged (2%). We identified four additional participants with normal vitamin B₁₂ levels whose elevated metabolite levels decreased substantially after vitamin B₁₂ treatment. These participants may have had early vitamin B₁₂ deficiency. They are not included in our prevalence data, however, because their vitamin B₁₂ levels at this time were normal.

For the diagnosis of vitamin B₁₂ deficiency, we required a low serum vitamin B₁₂ level and elevated levels of specific metabolites (serum methylmalonic acid and total homocysteine). It was useful to measure levels of both metabolites. Although 80% of the deficient patients had elevated serum methylmalonic acid levels, 20% of the patients only had elevated serum total homocysteine levels. Determining metabolite levels was especially important in diagnosing vitamin B₁₂ deficiency and guiding therapy in elderly persons who had not had gastric surgery. We found that a low serum vitamin B₁₂ level in persons who had not had gastric surgery was unlikely to indicate vitamin B₁₂ deficiency. Measurement of metabolite levels showed substantial vitamin B₁₂ deficiency in only 2 of these 22 participants. However, in patients who had had gastric surgery, low vitamin B₁₂ levels were much more likely to predict true B₁₂ deficiency (19 of 29 patients).

The prevalence of vitamin B₁₂ deficiency that we found in patients who had had gastric surgery is higher than the 1% to 20% found in most previous studies of such patients [4, 5]. This higher prevalence is probably caused by three factors: 1) the use of more metabolic variables to detect tissue vitamin B₁₂ deficiency in our study; 2) the older median age of our patient sample [3, 4, 31]; and 3) the greater median and absolute number of years since surgery in our study.

Previous studies probably underestimated the true prevalence of vitamin B₁₂ deficiency in patients who had had gastric surgery, because the diagnosis was based on the presence of megaloblastic anemia rather than low serum vitamin B₁₂ levels or elevated metabolite levels. Rygvold [7] found that half of his vitamin B₁₂-deficient patients (deficiency defined as serum vitamin B₁₂ levels less than 148 pmol/L) had normal hemoglobin levels. It is now well recognized that hematologic abnormalities may be absent in patients with documented vitamin B₁₂ deficiency who present with neurologic disease [12, 23, 32]. Although we found that the vitamin B₁₂-deficient patients were anemic, vitamin B₁₂ deficiency was not associated with increased mean corpuscular volume or hypersegmented neutrophils. Further, because iron deficiency is commonly found in patients who have had gastric surgery [5, 24], the expected macrocytosis may be masked. However, because we did not measure iron status, we cannot assess the contribution of iron deficiency to the absence of macrocytosis.

The older median age of our patients may also have contributed to the higher prevalence of vitamin B₁₂ deficiency we found compared with the prevalence found in other series. The median age of our patients (67 years) is higher than that found in earlier studies (53, 57, and 46 years) [3, 4, 31].

In our study, the median interval between surgery and evaluation for vitamin B₁₂ deficiency was 20 years (range, 1 month to 50 years). In previous reports [3-57, 8], the median was not given, but the interval between surgery and evaluation for vitamin B₁₂ deficiency ranged from months to only 23 years. The longer median time after surgery in our patients may have allowed for the appearance of patients with late-onset vitamin B₁₂ deficiency in tissue who would not have been identified in the earlier studies.

We found that the prevalence of vitamin B₁₂ deficiency in the controls was lower than the 14.5% prevalence noted in an earlier study of elderly patients [2]. Part of this difference may be secondary to differences in patient selection for the two studies. In the earlier study, patients who had had gastric surgery were not excluded; in our study, however, we rigorously excluded persons who had had gastric surgery from the control group. The patients in this previous study, ranging in age from 65 to 99 years (mean, 80 years), were older than the controls in our study (mean age, 63 years; range, 33 to 81 years). However, age alone is unlikely to explain the difference found in the prevalence of vitamin B₁₂ deficiency. Fifty-eight (54%) of the controls in our study were at least 65 years of age, and only one (who was 66 years of age) had a vitamin B₁₂ level less than 221 pmol/L and had elevated methylmalonic acid or total homocysteine levels.

The clinical significance of our study is that vitamin B₁₂ deficiency is often unrecognized in persons who have had gastric surgery. At the beginning of our study, only 3 of 65 patients who had had gastric surgery were receiving vitamin B₁₂ replacement therapy, although an additional 19 were found to be deficient. Clinicians may be monitoring their patients only for the development of neurologic findings or macrocytic anemia because previous studies have emphasized the hematologic manifestations [3-510, 11] and the neurologic [12] and neuropsychiatric symptoms and signs [26]. These signs, however, help detect vitamin B₁₂ deficiency at a later stage than do the metabolic measures we used. Clinicians also may underestimate the need for continued monitoring of serum vitamin B₁₂ levels in patients who had had gastric surgery many years ago. Clinicians may assume that persons who have not developed low serum vitamin B₁₂ deficiency 5 to 6 years after surgery will not develop it in the future. Although body stores of vitamin B₁₂ are depleted in this period if no vitamin B₁₂ is being absorbed, many patients who have had gastric surgery have enough acid [10], pepsin [33], and intrinsic factor to continue to absorb diminished but adequate levels of protein (food)-bound vitamin B₁₂ for many years. As patients age, absorption of food-bound vitamin B₁₂ may be impaired by the age-associated loss of acid and pepsin [9, 34] or therapy with histamine blockers [35, 36]. These patients may therefore develop vitamin B₁₂ deficiency, as we have shown, as long as 30 years after surgery.

Clinicians must recognize that in patients who have had gastric surgery, continued monitoring of serum vitamin B₁₂ levels is necessary because, even when neurologic and hematologic findings are absent, a low or low-to-normal serum vitamin B₁₂ level is a strong predictor of vitamin B₁₂ deficiency. Moreover, neurologic dysfunction responds best to vitamin B₁₂ when treated early [12, 13].

Evidence showing that hyperhomocysteinemia is an independent risk factor for vascular disease is growing. Therefore, delaying the diagnosis of vitamin B₁₂ deficiency may have cardiovascular consequences [15-18]. Because most vitamin B₁₂-deficient patients in our study had elevated total homocysteine levels, premature coronary artery, peripheral, and cerebrovascular disease may also be preventable if vitamin B₁₂ deficiency is recognized promptly. The total homocysteine levels found in our vitamin B₁₂-deficient patients were, in fact, higher than the levels found to be associated with carotid stenosis in one recent study [15]. However, because of our study design, we could not determine the difference in atherosclerotic cardiovascular disease between patients who were found to be vitamin B₁₂-deficient and those who were not.

Because of the absence of overt symptoms during the early phases of vitamin B₁₂ deficiency and the serious and often irreversible neurologic and, possibly, cardiovascular consequences of late diagnosis, we endorse periodic determinations of serum vitamin B₁₂ levels in all patients who have had gastric surgery, including those having gastric surgery for obesity. We recommend that lifelong parenteral vitamin B₁₂ therapy be given to patients with a vitamin B₁₂ level less than 221 pmol/L. Once patients are receiving therapy, vitamin B₁₂ levels do not need to be measured. In contrast, in patients with low serum vitamin B₁₂ levels who have not had gastric surgery, we recommend measuring folate, methylmalonic acid, and total homocysteine levels before therapy is begun. Because of the low prevalence of vitamin B₁₂ deficiency in this population, we would begin therapy only if folate levels were normal and metabolite levels were elevated.

Acknowledgments: The authors thank John R. Hansell, MD, Elizabeth Quinlain, PA-C, and Michael J. Palmieri, PhD, for administrative assistance; Linda Albert, Sandra Moore, Barbara Whiteman, Linda Farb, and Bev Raab for expert technical assistance; and Jonsi Jones for manuscript preparation.

Dr. Chin: Philadelphia College of Pharmacy and Science, 600 South 43rd Street, Philadelphia, PA 19104-4494.

Dr. Abrahm: Medical Service (111), Philadelphia Veterans Affairs Medical Center, University and Woodland Avenues, Philadelphia, PA 19104.

Dr. Berry: Division of Biochemical Development and Molecular Diseases, the Children's Hospital of Philadelphia, 34th Street and Civic Center Boulevard, Philadelphia, PA 19104.

Drs. Allen and Stabler: Division of Hematology, University of Colorado Health Sciences Center, 4200 East 9th Avenue, Box B170, Denver, CO 80262.