Jamanetwork.com

Risk Factors of Vitamin B12 Deficiency
in Patients Receiving Metformin
Rose Zhao-Wei Ting, MBBS; Cheuk Chun Szeto, MD, FRCP; Michael Ho-Ming Chan, FRCPA, FHKCPath;Kwok Kuen Ma, MBBS; Kai Ming Chow, MRCP Background: Identification of risk factors for metformin-
dose and duration of metformin use. Each 1-g/d metfor- related vitamin B12 deficiency has major potential impli- min dose increment conferred an odds ratio of 2.88 (95% cations regarding the management of diabetes mellitus.
confidence interval, 2.15-3.87) for developing vitaminB12 deficiency (PϽ.001). Among those using metformin Methods: We conducted a nested case-control study from
for 3 years or more, the adjusted odds ratio was 2.39 (95% a database in which the source population consisted of confidence interval, 1.46-3.91) (P=.001) compared with subjects who had levels of both serum vitamin B12 and those receiving metformin for less than 3 years. After ex- hemoglobin A1c checked in a central laboratory. We iden- clusion of 113 subjects with borderline vitamin B12 con- tified 155 cases of diabetes mellitus and vitamin B12 de- centration, dose of metformin remained the strongest in- ficiency secondary to metformin treatment. Another 310 controls were selected from the cohort who did not have vitamin B12 deficiency while taking metformin.
Conclusions: Our results indicate an increased risk of
vitamin B
Results: A total of 155 patients with metformin-related
12 deficiency associated with current dose and duration of metformin use despite adjustment for many vitamin B12 deficiency (mean ± SD serum vitamin B12 con-centration, 148.6 ± 40.4 pg/mL [110 ± 30 pmol/L]) were potential confounders. The risk factors identified have compared with 310 matched controls (466.1 ± 330.4 implications for planning screening or prevention strat- pg/mL [344 ± 244 pmol/L]). After adjusting for con- egies in metformin-treated patients.
founders, we found clinically important and statisti-cally significant association of vitamin B12 deficiency with Arch Intern Med. 2006;166:1975-1979 METFORMINHASGREATLY causedustoquestionwhetherthisad-
verse effect is predictable among patients with type 2 diabetes mellitus who receive metformin. To date, knowledge of the risk factors of this adverse event of metformin is still limited. From a clinical standpoint, characterization of risk factors for metfor- min-related vitamin B12 deficiency is the key long-term benefit of metformin in decreas- to better patient care. First, there is likely ing diabetes-related end points, diabetes- to be an improved yield of detecting vita-min B related death, and all-cause mortality in can be identified. Second, subjects identi- fied as having substantial risk for metfor- Author Affiliations:
that metformin decreased serum vitamin B12 We undertook a nested case-control study in the level by 14% to 30%.4-6 The findings of met- New Territories East Cluster region, Hong Kong, formin-related vitamin B12 deficiency have between January 2003 and November 2005. A (REPRINTED) ARCH INTERN MED/ VOL 166, OCT 9, 2006 2006 American Medical Association. All rights reserved.
Downloaded From: https://jamanetwork.com/ on 03/09/2014
min B12 deficiency was assessed by cross-tabulation and the ␹2test or Fisher exact test, whereas continuous variables betweengroups were compared with unpaired t test or the nonparamet- ric Wilcoxon rank sum test as appropriate. To estimate the strength of association, a binary logistic regression model was used for bothunivariate and multivariate analysis to calculate the odds ratio (OR)of metformin-related vitamin B12 deficiency. Our model was con-structed with stepwise selection for relevant variables by previ- ous studies and those associated with metformin-related vita- 12 deficiency at an ␣ level of Ͻ.1 in our univariate analysis.
Because of the lack of a precise standard and uncertainty of defi-nition of vitamin B12 deficiency without supplementary methyl- malonic acid measurement, we repeated the regression analysis Exclusion of SubjectsNot Taking Metformin after excluding cases with borderline vitamin B12 concentration (Ͼ203.3 and Յ298.1 pg/mL [Ͼ150 and Յ220 pmol/L])7,8 to mini-mize misclassification error. Confidence intervals (CIs) re-ported are likelihood based. All P values were 2-sided, and we regarded PϽ.05 as significant. All statistical analyses were per-formed using The Statistical Package for the Social Sciences (Win-dows, version 13.0; SPSS Inc, Chicago, Ill).
Figure 1. Patient profile of the nested case-control study. Cases refer to
subjects with diabetes mellitus and vitamin B12 deficiency during metformin
therapy, whereas the controls had metformin-treated diabetes mellitus but
no vitamin B12 deficiency. Cases and controls were matched according to thedate of blood sampling. To convert vitamin B12 to picograms per milliliter,divide by 0.738.
Between January 2003 and November 2005, a total of 3987
source subjects were evaluated for the study. Figure 1
database was generated by a central chemical pathology labora- summarizes the trial profile. Of 355 cases with a labora- tory that provides vitamin B12 assay service to cover 19% (1.3 mil- lion) of the population in Hong Kong. In other words, we nested concentration Յ203.3 pg/mL [Յ150 pmol/L]), 155 cases the case-control study within a source cohort population, as de-fined by patients who had levels of both hemoglobin A fulfilled the case selection criteria. We therefore identi- fied 155 cases (mean ± SD serum vitamin B12 concentra- 12 being checked for clinical reasons within the study period. To ensure complete data extraction, we also restricted the tion, 148.6 ± 40.4 pg/mL [110 ± 30 pmol/L]; range, 50.1- subjects to have at least 1 year of continuous medical history re- 203.3 pg/mL [37-150 pmol/L]) and 310 matched controls corded on computer. Serum vitamin B12 and folate concentra- (serum vitamin B12 concentration, 466.1±330.4 pg/mL tions were determined by electrochemiluminescent immunoas- [344 ± 244 pmol/L]; range, 204.6-2000.0 pg/mL [151- say (E170 Analytic; Roche Diagnostics, Indianapolis, Ind). The mean±SE lower limit of normal by our vitamin B12 assay was Characteristics of the cases and controls are listed in 253.4±50.1 pg/mL (187 ± 37 pmol/L); therefore, we chose a cut- Table 1. There were no significant differences in se-
off point of 203.3 pg/mL (150 pmol/L) for vitamin B12 defi- rum folate concentration between these 2 groups. Sex dis- ciency. Serum methylmalonic acid concentration evaluation wasnot routinely performed.
tribution, drinking, and smoking habits were similar We defined a case as a Chinese subject with diabetes mellitus among cases and controls. Vegetarians appeared to be more common among cases than controls, with border- 12 deficiency during metformin treatment after ex- cluding subjects who had pernicious anemia (positive Schilling line statistical significance (prevalence 2.6% vs 0.3%; test result or anti-intrinsic factor antibodies), pancreatic exo- P = .04). There was no substantial difference in age, with crine insufficiency, and/or history of gastrectomy or small bowel a mean ± SD age of 72.5 ± 9.3 years among cases vs resection. To avoid potentially confounding causes of vitamin B12, 71.4 ± 11.2 years among controls (P = .24).
patients receiving oral or parenteral vitamin B12 supplementa- In the univariate analysis, vegetarians (OR, 8.19; 95% tion within 3 months prior to study initiation were excluded. Two CI, 0.91-73.9) had a borderline significant association with controls were selected for each case, matched according to the date of blood sampling. Eligible controls for each case were in- dividuals within our database with diabetes mellitus who did not marized in Table 2, the unadjusted analysis showed a sig-
nificant increase in risk associated with the current dose 12 deficiency (serum concentration Ͼ203.3 pg/mL [150 pmol/L]) while taking metformin.
and duration of metformin medication. The metformin- Standardized data collection forms were used to abstract in- treated patients with vitamin B12 deficiency had a mean±SD formation from computerized medical records and pharmacy daily dose of 2.0±0.7 g (interquartile range, 1.5-3.0 g), records. Data collected included demographic information, con- whereas the control group received a daily metformin dose centrations of serum vitamin B12, folate, and blood hemoglo- of 1.4±0.7 g (interquartile range, 1.0-2.0 g). The crude OR bin, white blood cell and platelet counts, mean corpuscular vol- for metformin-related vitamin B12 deficiency was 2.37 (95% ume, documentation of concomitant histamine H2 receptor CI, 1.60-3.52) for a history of metformin use longer than antagonist or proton pump inhibitor therapy, vegetarian diet, 3 years. There was a significant increased risk of metformin- smoking history, and alcohol consumption. Detailed data of met-formin daily dose and duration of metformin therapy at the time related vitamin B12 deficiency associated with each addi- of blood collection were also collected.
tional 1-g/d dose increment of metformin (OR, 2.61; 95% Mean and standard deviation values for continuous variables were calculated to characterize our study population. The asso- Once adjustments were made for potential confound- ciation between discrete variables and metformin-related vita- ing variables (Table 2), the most significant OR was as- (REPRINTED) ARCH INTERN MED/ VOL 166, OCT 9, 2006 2006 American Medical Association. All rights reserved.
Downloaded From: https://jamanetwork.com/ on 03/09/2014
Table 1. Characteristics of Cases of Metformin-Related Vitamin B12 Deficiency and Matched Controls Identified From a
Population-Based Cohort*
Controls
Characteristic
P Value
Moderate to heavy alcohol use (Ͼ1 drink daily, No.)† Use of histamine H2-blocker or proton pump inhibitor therapy, No. (%) Duration of metformin use, median (IQR), y Abbreviation: IQR, interquartile range.
*Unless otherwise indicated, data are reported as mean ± SD.
†Data of smoking and alcohol consumption were missing in 25% and 32% of the study population, respectively.
‡Fisher exact test.
Table 2. Risk Factors Associated With Development of Metformin-Related Vitamin B12 Deficiency, Including Borderline Deficiency
Adjusted OR
Risk Factor
P Value
P Value
Use of histamine H2 receptor antagonist or Daily dose of metformin, per 1-g increment Abbreviations: CI, confidence interval; OR, odds ratio.
sociated with current dose of metformin. This was fol- dose increment; 95% CI, 2.63-5.35) (PϽ.001). Of the 113 lowed by duration of metformin use and patient age. Each subjects excluded, their characteristics, including dose and 1-g/d dose increment conferred a more than 2-fold in- duration of metformin use, were similar to the control sub- creased risk of developing vitamin B12 deficiency with met- jects with serum vitamin B12 concentration exceeding 298.1 formin (adjusted OR, 2.88; 95% CI, 2.15-3.87) (PϽ.001).
pg/mL (220 pmol/L) (details not shown).
Compared with metformin users of less than 3 years, the We also investigated the effects of metformin dose on adjusted OR was 2.39 (95% CI, 1.46-3.91) (P = .001) for the serum vitamin B12 concentration among the cases.
users of metformin for 3 years or more. Increased age ap- Cases were divided into 3 groups according to their cur- peared to be positively related to metformin-related vi- rent daily metformin doses: 1.0 g or less (n = 29), more tamin B12 deficiency albeit with negligible clinical sig- than 1.0 g to 2.0 g (n = 76), or more than 2.0 g to 3.0 g nificance (adjusted OR, 1.36 for each 10-year increment (PϽ.001 by analysis of variance) (Figure 2). Post hoc
in age). Vegetarian diets were also associated with vita- analysis performed by the Bonferroni-adjusted pairwise min B12 deficiency, but the confidence interval was wide comparisons revealed significantly lower vitamin B12 con- (OR, 16.2; 95% CI, 1.69-154.00). We found no signifi- centration in cases receiving more than 2.0 to 3.0 g/d than cantly increased risk for concurrent use of histamine H2 in those receiving more than 1.0 to 2.0 g/d (P = .01) and receptor antagonist or proton pump inhibitor.
those receiving 1.0 g/d or less (PϽ.001). Conversely, when We repeated the analyses after excluding 113 subjects all the cases and controls in this study were included in (24%) with borderline vitamin B12 concentration between the post hoc analysis, there was also a significant differ- 203.3 and 298.1 pg/mL (150 and 220 pmol/L). This did ence in the proportion of subjects with deficient vita- not significantly influence the results (Table 3); current
min B12 concentrations according to the daily metfor- dose of metformin remained the strongest independent pre- min doses. The distribution of subjects (Figure 3) varied
dictor of vitamin B12 deficiency (OR, 3.75 for each 1-g/d with the metformin doses with respect to the vitamin B12 (REPRINTED) ARCH INTERN MED/ VOL 166, OCT 9, 2006 2006 American Medical Association. All rights reserved.
Downloaded From: https://jamanetwork.com/ on 03/09/2014
Table 3. Risk Factors Associated With Development of Metformin-Related Vitamin B12 Deficiency, Excluding Borderline Deficiency
Adjusted OR
Risk Factor
P Value
P Value
Use of histamine H2-receptor antagonist or proton pump inhibitor Daily dose of metformin (per 1-g increment) Abbreviations: CI, confidence interval; OR, odds ratio.
Figure 2. Box-and-whisker plot shows the serum vitamin B12 concentration
Figure 3. Effect of daily metformin dose on the proportion of all subjects
for cases of metformin-related vitamin B12 deficiency according to the with serum vitamin B12 concentration up to 203.3 pg/mL and up to 298.1 different daily doses of metformin received. The lower and upper bounds of pg/mL. Error bars indicate standard error of the mean. To convert vitamin B12 the boxes denote the 25th and 75th percentiles, respectively, and the to picograms per milliliter, divide by 0.738.
horizontal lines in the boxes correspond to the median value. The lower andupper error bars indicate the 10th and 90th percentiles, respectively. Toconvert vitamin B12 to picograms per milliliter, divide by 0.738.
ficiency may result from disorders in intestinal mobilityand/or bacterial overgrowth.7 However, more recent evi-dence has demonstrated that metformin administration cutoff points of either 203.3 or 298.1 pg/mL (PϽ.001 for neither alters the intestinal motility9 nor causes bacte- rial overgrowth.6 On the other hand, metformin may dis-rupt the ileal vitamin B12 absorption.10,11 The vitamin B12- intrinsic factor complex is dependent on the luminalcalcium concentration to facilitate uptake by the ileal cell In our nested case-control study of metformin-related vi- surface receptor, whereas metformin is believed to give tamin B12 deficiency, metformin dose and treatment du- a positive charge to the surface of the membrane, which ration emerged as the most consistent risk factors of vi- would act to displace divalent cations such as calcium.
tamin B12 deficiency within a Chinese population with Impaired calcium availability due to metformin activity diabetes mellitus. Of special interest is that their asso- would therefore interfere with the calcium-dependent pro- ciation remained stable after adjustment for potential con- cess of vitamin B12 absorption.6,10-12 It should be noted founding factors in the multivariate analysis, thus rein- that our study design precluded a full assessment of the forcing our conclusion that higher metformin dose and dietary or supplementary intake of calcium.6 Practi- longer treatment duration are independent risk factors.
cally, our data and the graded relationship of metformin There is also evidence from our post hoc analysis that dose with serum vitamin B12 concentration (Figure 2) sup- serum vitamin B12 concentration showed a dose- port the notion of a causal relation between metformin dependent decrease with increasing dose of metformin.
administration and vitamin B12 deficiency rather than sug- It is impossible to deduce from our case-control study gesting any particular mechanism(s).
the mechanism of metformin-related vitamin B12 defi- Unlike previous studies,13-15 our study demonstrated ciency. The literature has reported that vitamin B12 de- no excess risk of vitamin B12 deficiency among metfor- (REPRINTED) ARCH INTERN MED/ VOL 166, OCT 9, 2006 2006 American Medical Association. All rights reserved.
Downloaded From: https://jamanetwork.com/ on 03/09/2014
min users who concurrently use H2-blockers or proton call attention to the value of vitamin B12 screening,19 par- pump inhibitors. Theoretically, these medications can pro- ticularly among at-risk patients receiving metformin. Our duce malabsorption of dietary protein-bound vitamin B12.
data underscore the need for monitoring subjects un- The lack of association in the current study may stem from dergoing high-dose and/or prolonged-course metfor- imprecise hospital-based medication records. We were unable to track the use of H2-blockers or proton pumpinhibitors that are nowadays commonly prescribed by fam- Accepted for Publication: June 27, 2006.
ily practitioners or purchased over the counter. Such mis- Correspondence: Kai Ming Chow, MRCP, Department
classification of exposure status (to H2-blockers or pro- of Medicine and Therapeutics, Prince of Wales Hospi- ton pump inhibitors), if present and nondifferential, would tal, The Chinese University of Hong Kong, Sha Tin, Hong have had the effect of biasing the OR toward 1.
Kong, China ([email protected]).
To minimize the risk of overmatching, no matching Author Contributions: Study concept and design: Ting and
(other than the date of blood sampling) was used in this Chow. Acquisition of data: Ting, Chan, and Chow. Analy- study, which might have hindered informative results or sis and interpretation of data: Ting, Szeto, Ma, and Chow.
introduced confounding if the matching factor were linked Drafting of the manuscript: Ting and Chow. Critical revi- with the risk of vitamin B12 deficiency. Subjects in our sion of the manuscript for important intellectual content: study, nevertheless, needed to have at least 1 serum vi- Szeto, Chan, and Chow. Statistical analysis: Szeto and tamin B12 sample checked for a clinical purpose before Chow. Study supervision: Chow.
being eligible as controls. Patients with certain charac- Financial Disclosure: None reported.
teristics may therefore have been more likely to be re- Acknowledgment: We thank Clara Law for her contri-
cruited as controls. Clinicians are in general more in- clined to screen for vitamin B12 deficiency amongvegetarians or the geriatric population, for instance, when a hematologic, neuropsychiatric, or cognitive disorder iseither detected or suspected. If this had been the case, 1. Bailey CJ, Turner RC. Metformin. N Engl J Med. 1996;334:574-579.
we would have expected the strength of association be- 2. UK Prospective Diabetes Study (UKPDS) Group. Effect of intensive blood- glucose control with metformin on complications in overweight patients with type derly16-18 or vegetarian subjects, if any, to be negated. This 2 diabetes (UKPDS 34). Lancet. 1998;352:854-865.
3. Tomkin GH, Hadden DR, Weaver JA, Montgomery DA. Vitamin-B is a genuine possibility in the current study because sta- tients on long-term metformin therapy. BMJ. 1971;2:685-687.
tistically significant relationships between increased age, 4. DeFronzo RA, Goodman AM. Efficacy of metformin in patients with non-insulin- vegetarianism, and the development of vitamin B dependent diabetes. N Engl J Med. 1995;333:541-549.
ciency are apparent in the multivariate but not the uni- 5. Wulffele´ MG, Kooy A, Lehert P, et al. Effects of short-term treatment with met- formin on serum concentrations of homocysteine, folate and vitamin B12 in type2 diabetes mellitus: a randomized, placebo-controlled trial. J Intern Med. 2003; Our results have to be interpreted with caution in light of several other limitations. One potential limitation of 6. Bauman WA, Shaw S, Jayatilleke E, Spungen AM, Herbert V. Increased intake of calcium reverses vitamin B12 malabsorption induced by metformin. Diabetes Care.
7. Snow CF. Laboratory diagnosis of vitamin B12 and folate deficiency: a guide for tabolite (plasma homocysteine or methylmalonic acid) the primary care physician. Arch Intern Med. 1999;159:1289-1298.
measurements for confirmation.7,8,16 Controversy exists 8. Klee GG. Cobalamin and folate evaluation: measurement of methylmalonic acid as to whether serum measurement accurately reflects stor- and homocysteine vs vitamin B12 and folate. Clin Chem. 2000;46:1277-1283.
9. Scarpello JH, Hodgson E, Howlett HC. Effect of metformin on bile salt circula- may have been introduced by misclassifying cases and tion and intestinal motility in type 2 diabetes mellitus. Diabet Med. 1998;15:651-656.
controls based on serum vitamin B12 measurement. None- 10. Andrès E, Noel E, Goichot B. Metformin-associated vitamin B12 deficiency. Arch theless, repeated analysis after our careful exclusion of Intern Med. 2002;162:2251-2252.
12 deficiency. Arch Intern Med. 2002;162: qualitatively affect the findings (Table 3), which sug- 12. Schafer G. Some new aspects on the interaction of hypoglycemia-producing bi- gests that any biases thus introduced were likely to be guanides with biological membranes. Biochem Pharmacol. 1976;25:2015- small. Furthermore, the retrospective nature of informa- tion retrieval for our subjects led to predictable limita- 13. Mitchell SL, Rockwood K. The association between antiulcer medication and ini- tions of data completeness. Another potential bias arises tiation of cobalamin replacement in older persons. J Clin Epidemiol. 2001;54: from the unblinded data acquisition method. Finally, the 14. Ruscin JM, Page RL II, Valuck RJ. Vitamin B12 deficiency associated with histamine data from this study cannot be used to assess the inci- (2)-receptor antagonists and a proton-pump inhibitor. Ann Pharmacother. 2002; dence of vitamin B12 deficiency with metformin use.
In conclusion, this nested case-control study in a Chi- 15. Valuck RJ, Ruscin JM. A case-control study on adverse effects: H2 blockers or nese population showed that the risk of vitamin B proton pump inhibitor use and risk of vitamin B12 deficiency in older adults.
J Clin Epidemiol. 2004;57:422-428.
ficiency is magnified in patients who have received both 16. Andrès E, Loukili NH, Noel E, et al. Vitamin B12 (cobalamin) deficiency in elderly a higher dose and longer course of metformin treat- patients. CMAJ. 2004;171:251-259.
ment, independent of other clinical variables. Although 17. Andrès E, Goichot B, Schlienger JL. Food-cobalamin malabsorption: a usual cause this observational study may be subject to residual con- of vitamin B12 deficiency. Arch Intern Med. 2000;160:2061-2062.
founding that cannot be fully corrected for, we believe 18. Liu KW, Dai LK, Woo J. Metformin-related vitamin B12 deficiency. Age Ageing.
our findings should reinforce the heightened vigilance 19. Filioussi K, Bonovas S, Katsaros T. Should we screen diabetic patients using bi- about vitamin B12 deficiency. Enough concerns exist to guanides for megaloblastic anaemia? Aust Fam Physician. 2003;32:383-384.
(REPRINTED) ARCH INTERN MED/ VOL 166, OCT 9, 2006 2006 American Medical Association. All rights reserved.
Downloaded From: https://jamanetwork.com/ on 03/09/2014

Source: https://jamanetwork.com/data/Journals/INTEMED/5558/ioi60103.pdf