Article Text


Prospective evaluation of protein bound vitamin B12(cobalamin) malabsorption in the elderly using trout flesh labelled in vivo with 57Co-cobalamin


Background—The frequency of dietary protein bound vitamin B12 malabsorption in elderly patients remains controversial.

Aims—To evaluate this malabsorption in elderly hospitalised patients using a modified Schilling test.

Patients—Fourteen elderly patients with low B12 blood levels were prospectively selected from 394 hospitalised patients.

Methods—The modified Schilling test was performed with trout labelled in vivo.

Results—The test was normal in five healthy elderly subjects, in 7/8 patients with pancreatic insufficiency, and in nine non-elderly patients with antral gastritis. The low decision limit was established at 3.3% (median 4.8%). From the 14 elderly patients with low B12 prospectively selected from 394 hospitalised patients, seven had a real deficiency with anaemia and an increased homocysteine and/or methylmalonate serum level. The modified Schilling test showed malabsorption in five of these patients, including two in which the standard Schilling test was normal, and three in which the standard Schilling test was partially corrected by an intrinsic factor.

Conclusions—Protein bound vitamin B12malabsorption was detected in at least 0.5% of elderly hospitalised patients, using the labelled trout flesh absorption test.

  • cobalamin deficiency
  • malabsorption
  • gastritis
  • Schilling test

Statistics from

The frequency and aetiologies of cobalamin deficiency due to malabsorption in the elderly are still under discussion, despite the number of studies performed on this subject.1-9 Several groups have observed a low vitamin B12 serum level in as many as 10% of elderly hospitalised patients. Most of these patients showed no clinical or haematological signs of B12deficiency.7 9 It has been suggested that serum B12 is not sensitive enough to diagnose B12deficiency and that serum methylmalonic acid and homocysteine concentration are more sensitive.10–13

The frequency of pernicious anaemia increases with age. A second gastric aetiology of vitamin B12 deficiency corresponds to dietary cobalamin malabsorption.8 This can occur in the absence of intrinsic factor (IF) deficient secretion and is related to a lack of cobalamin release from food proteins, which is the consequence of decreased acid and peptic secretion.14-17It may be assumed that such dietary cobalamin malabsorption is frequent in the elderly population as the prevalence of gastritis increases with age.18 19 The prevalence of vitamin B12deficiency due to this malabsorption has not, however, been prospectively evaluated in an elderly population. Such an evaluation requires the use of an adequate modified Schilling test, with labelled cobalamin bound to food proteins. The protein bound vitamin B12 malabsorption has been studied in chronic gastritis, using modified Schilling tests performed with food proteins labelled either in vitro or in vivo.14 20-26 In our opinion, in vivo incorporation of labelled vitamin B12 to either fish or chicken meat is a more physiological process than adding vitamin B12 to chicken serum in vitro. Fish and chicken meat give an excretion percentage of the tracer in the order of 3–6% whereas chicken serum gives a percentage lower than 2% in healthy subjects.27 This may be explained by the different behaviour of food proteins in releasing vitamin B12 at acidic pH and in being degraded by duodenal juice.28 In addition, the Schilling test performed with chicken serum cannot distinguish malabsorption due to chronic gastritis from that due to chronic pancreatitis.

The aim of the present study was therefore to evaluate the aetiology of vitamin B12 deficiency in elderly patients with a low vitamin B12 blood concentration prospectively selected from 394 elderly hospitalised patients, using a modified Schilling test performed with trout flesh.

Patients and Methods


Controls—Nine healthy, non-elderly subjects (aged 28–66 years), and five healthy, elderly subjects (aged 76–82 years) were included. Inclusion criteria were the absence of anaemia, a normal standard Schilling test, the absence of histological gastritis, and the absence of anti-IF autoantibodies. Informed consent was obtained in accordance with the Declaration of Helsinki.

Elderly patients—A total of 394 elderly patients (aged 70–91 years), hospitalised in the Department of Internal Medicine of the University Hospital Centre, Nancy, were included in the study. None were receiving any treatment or were known to suffer from any intestinal disease impairing cobalamin absorption. Biological parameters measured included serum vitamin B12 and folates, erythrocyte folates, erythrocyte mean corpuscular volume, haemoglobin and ferritin blood levels, and serum anti-IF autoantibodies. Patients who had low vitamin B12 in serum and at least one of the other parameters out of the normal range were investigated for vitamin B12 deficiency and malabsorption, by means of determination of homocysteine and methylmalonic acid in serum, modified and standard Schilling tests, and gastric endoscopic examination of fundic and antral biopsy samples.

Patients with chronic gastritis were aged 51–60 years. The diagnosis was established from endoscopic examination and the study of gastric biopsy specimens. Three patients had fundic gastritis and nine had antral gastritis.

Patients with chronic pancreatitis (n=8) were aged 37–45 years. The diagnosis was established from clinical signs, NBT-PABA test, ultrasonography, and x ray computed tomography.


Plasma pepsinogen I was measured by radioimmunoassay (ORIS, Gif sur Yvette, France). Serum vitamin B12 and folates were measured using a radioisotopic dilution assay (Becton Dickinson Immuno Diagnostics Company, New York, USA). Serum methylmalonate and homocysteine were assayed in serum using capillary gas chromatography mass spectrometry, as recently described.29-31


The gastric biopsy specimens were examined histologically by a single pathologist, who was unaware of the endoscopic findings. Biopsy tissue, fixed in 10% formalin and embedded in paraffin wax, was sectioned at 5 μm and stained with haematoxylin and eosin. Gastritis was graded according to the Sydney system.32 This grades the severity of inflammation, activity (the degree of polymorphonuclear neutrophil infiltration), atrophy, and intestinal metaplasia on a scale from 0 to 3. A subsequent “gastritis score” for each biopsy site was obtained by combining the scores of the four individual characteristics (maximum possible score 12).


The standard Schilling test was performed with the Dicopac test (Amersham, UK). The second stage (ingestion of cobalt-57 labelled cyanocobalamin-IF) was started four days after the first stage (ingestion of cyano-58Co-cobalamin). Urine was collected for 48 hours at each stage. The limit of normal values was established at 10% and 11% of urinary excretion of the tracer for the first and second stages, respectively. The modified Schilling test was performed using trout meat labelled with cyano-57Co-cobalamin. Labelled vitamin B12 (0.02 μCi) was injected into five week old trout after anaesthesia with 0.03% (vol/vol) ethylene glycol monophenyl ester (Merck, FRG). The injection was repeated one week later and the trout were sacrified two weeks later, cooked, wrapped in aluminium foil for 10 minutes at 200°C, then dissected. The meat was liquidised, divided into fractions (one fraction of 54 (10) g per test) corresponding to a radioactivity of 177 000 (5500) cpm and stored at –18°C. The protocol of the test was similar to that of the classic test and included the determination of the meal fraction radioactivity before its ingestion, the intramuscular injection of 1000 μg of non-labelled vitamin B12, and the collection of urine for 48 hours. The determination of excreted radioactivity was estimated by 10 minute γ counting of five 10 ml urine samples. The modified Schilling test was performed at least eight days after the standard test. Correlation of the urinary excretion rate of the tracer with either the Sydney score for the gastric biopsy samples or the pepsinogen blood level was studied using the Spearman rank correlation coefficient.


A low vitamin B12 serum concentration was observed in 40/394 subjects (10.2%); the lowest limit of serum vitamin B12 was 110 pmol/l in our reference population.31 The vitamin B12 serum level of this group was estimated at 83.1 (19.5) pmol/l (range 54.6–109.2 pmol/l). Fourteen of these patients were selected for further studies of vitamin B12 assimilation as they were suspected of having either vitamin B12 deficiency or chronic gastritis, using the criteria defined in the methods section. Nine had a low haemoglobin level and/or an increased erythrocyte mean corpuscular volume, two had a low serum folate level, and three underwent a gastric endoscopic examination for epigastric pain and presented histological signs of chronic gastritis (tables 1 and 2). Anti-IF autoantibodies were detected in the serum of only one patient. The red blood cell folate and the ferritin and iron serum levels were normal in all 14 patients (table 1). None had any neurological symptoms or renal failure.

Table 1

 Analytical data from 14 elderly patients with a low serum vitamin B12concentration

Table 2

 Schilling tests and gastric analytical and histological data of 14 elderly patients with a low serum vitamin B12 concentration

Methylmalonate and homocysteine serum levels were determined in these 14 patients. In most of the anaemic patients, elevation of both parameters was dissociated. Methylmalonate and homocysteine were the only increased parameters in three and five patients, respectively. A least one of these parameters was increased in all patients with macrocytic anaemia (table 1). Plasma pepsinogen I was measured as an index of gastric atrophy.32 33 The lowest normal blood pepsinogen I level limit was established at 20 ng/ml. Pepsinogen I was lower than this limit in only one woman with pernicious anaemia (patient 2, table 1) and one patient with moderate fundic gastritis (patient 7). It was normal in another patient with severe fundic gastritis (patient 6).

The standard Schilling test was performed on the 14 patients (table 2). Stage 1 (without IF) was abnormal (lower than 10%) in four patients with macrocytic anaemia. Stage 2 (with IF) showed a partial correction of the malabsorption in all four patients. One patient (patient 2) had pernicious anaemia with serum anti-IF autoantibodies. All patients with a normal haemoglobin blood level had a normal standard Schilling test, except one (patient 14) whose urinary excretion index of free vitamin B12 was slightly decreased. This patient had macrocytosis.

The normal value of the modified Schilling test was established as the lowest value obtained in the two control groups (n=14). It was estimated to be 3.3%. Values were higher than this limit in the five elderly, healthy subjects, in 7/8 cases of chronic pancreatitis, and in 8/9 non-elderly patients with antral gastritis. In contrast, it was abnormal in all three non-elderly patients with fundic gastritis. The modified Schilling test was performed in the 14 patients and was found to be abnormal in five of them (table 2, fig 1). Two of these patients showed malabsorption of food vitamin B12 as they had both a normal standard Schilling test and an abnormal modified Schilling test (table 2). Only one of these two cases presented histological signs of atrophy of the fundus (table 2). A significant negative correlation was found between the modified Schilling test and the Sydney score for fundic mucosa (r=0.62, p<0.05).

Figure 1

: Schilling tests performed with trout flesh labelled in vivo with 57Co-vitamin B12.


The group of elderly subjects with a low vitamin B12blood level represented 10.1% of the 394 subjects included in our study. This percentage is close to that observed by other authors.1-9 Of the 40 patients with a low vitamin B12 blood level, only seven had a vitamin B12deficiency with macrocytic anaemia and an elevated blood level of methylmalonate and/or homocysteine. Another patient (patient 10) had a haemoglobin blood level and an erythrocyte mean corpuscular volume at the normal value limit and an increased level of homocysteine (table1). The increase in homocysteine corresponded to a vitamin B12 deficiency rather than to a folate deficiency as they all had low B12 and normal red blood cell folate levels. Methylmalonate and homocysteine blood levels have been reported to be sensitive markers for detecting vitamin B12deficiency.10-12 In a study by Stabler et al 34 on patients with a vitamin B12deficiency, 92% of the patients had macrocytosis, 95% an elevated methylmalonate blood level, and 99% an elevated homocysteine level. More recently, Joosten et al showed that 23% and 30% of elderly ambulatory subjects had abnormal methylmalonate and homocysteine blood levels, respectively.13 Our group also observed that methylmalonate and homocysteine blood levels increase in elderly healthy subjects,35 in the absence of haematological or clinical abnormality. The specificity of these parameters for diagnosing a B12 deficiency in the elderly therefore remains to be established.

The present study was the first to use labelled trout meat for performing modified Schilling tests on elderly patients, concurrently with younger adult patients and healthy controls. The value of modified Schilling tests for detecting dietary protein bound vitamin B12 has been discussed in recent literature.17 18 22 24 36 Joosten et alfailed to find an advantage of the protein bound Schilling test performed with chicken serum over the standard Schilling test, in the diagnosis of cobalamin malabsorption in 41 elderly patients with a low vitamin B12 serum level.37 In addition, Scarlett et al recently showed that the diagnostic value of this protein bound Schilling test was limited by the frequent finding of reduced absorption in the healthy elderly.17This was not the case with our test as no abnormal modified Schilling test was observed in elderly patients without vitamin B12deficiency (fig 1).

Our study is the first to evaluate prospectively the frequency of vitamin B12 deficiency due to protein bound vitamin B12 malabsorption in elderly hospitalised patients, using a modified Schilling test with trout flesh labelled in vivo with57Co-vitamin B12. This modified Schilling test was previously described by Dorsherholmen et al.23 It was normal in 7/8 (87%) patients with chronic pancreatitis and therefore enabled the distinction to be made between a lack of dietary vitamin B12 release due to deficient gastric acid secretion, and a lack of haptocorrin degradation due to pancreatic deficiency.38 An abnormal modified Schilling test was found in seven patients (2% of elderly hospitalised patients).

Vitamin B12 malabsorption in chronic gastritis involves two aetiological factors: deficient intrinsic factor, which occurs in pernicious anaemia, and deficient gastric acid and pepsin secretion, which may lead to protein bound malabsorption despite normal or subnormal intrinsic factor secretion.14 20-22 27 We found a significant negative correlation of the modified Schilling test with the Sydney score for fundic mucosa but not with the score for antral mucosa nor with pepsinogen blood levels. It may therefore be suggested that deficient gastric acid secretion was the predominant factor responsible for protein bound vitamin B12malabsorption, assuming that blood pepsinogen reflected pepsin secretion.33 This hypothesis is in agreement with a recent case report from our group which described vitamin B12deficiency with protein bound vitamin malabsorption in a patient receiving long term omeprazole treatment.39 This drug is known selectively to inhibit gastric acid secretion and has no effect on pepsin and intrinsic factor secretion.40

The modified Schilling test is the only Schilling test to be disturbed when the deficient vitamin B12 release from food protein is the only factor responsible for malabsorption.26 This was the case in two patients in our study, representing 0.5% of the elderly hospitalised patients and 28% of the patients with vitamin B12 malabsorption. The trout flesh modified Schilling test may provide a functional test for diagnosing protein bound vitamin B12 malabsorption. It is important to establish this diagnosis, as these patients can be treated by oral administration of vitamin B12.14 20 26 It should be noted that in certain cases, malabsorption is due to both deficient food vitamin B12 release and to bacterial overgrowth related to deficient gastric acid secretion.41 42 In such cases normal vitamin B12 absorption can be restored by oral antibiotic therapy.41 42 More recently it has also been shown that Helicobacter pylori infection of the stomach and food vitamin B12 are intimately associated, H pylori predisposing to a more severe form of malabsorption.43

In conclusion, our study indicated that 2/398 (0.5%) hospitalised elderly patients had vitamin B12 deficiency related exclusively to protein bound vitamin B12 malabsorption. The Schilling test with trout flesh seems to be an efficient and specific model for investigating this type of malabsorption in elderly patients.


View Abstract

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles