Association of vitamin B12 deficiency in people with type 2 diabetes on metformin and without metformin: a multicenter study, Karachi, Pakistan.

OBJECTIVE: To assess the prevalence of vitamin B12 deficiency in people with type 2 diabetes mellitus (T2DM) on metformin and without metformin.
METHODOLOGY: Between May 2018 and January 2019, this prospective multicenter observational study recruited participants from seven centers in four provinces of Pakistan (Sindh, Punjab, Baluchistan and Khyber Pakhtunkhwa). Participants with T2DM treated with metformin for >2 years and those not on metformin underwent assessment of hemoglobin, vitamin B12, homocysteine and diabetic neuropathy (vibration perception threshold (VPT) >15V) and painful diabetic neuropathy (Douleur Neuropathique 4 (DN4) ≥4) and Diabetic Neuropathy Symptom (DNS) score ≥1.
RESULTS: Of 932 subjects, 645 (69.2%) were treated with metformin, while 287 (30.8%) were not on metformin. Overall, B12 deficiency (<200 pg/mL) was significantly higher in metformin users of 25 (3.9%), compared with non-metformin users of 6 (2.1%), while B12 insufficiency (200-300 pg/mL) was significantly lower in metformin users of 117 (18.4%) compared with non-metformin users of 80 (27.9%). Subjects with B12 deficiency and insufficiency with hyperhomocysteinemia (≥15) were found in 19 (76%) µmol/L and 69 (60.5%) µmol/L in metformin users compared with 6 (100%) µmol/L and 57 (73.1%) μmol/L in non-metformin users, respectively. VPT>25 and DN4 score ≥4 were significantly higher in B12-deficient metformin users compared with non-metformin users. Similarly, DNS score ≥1 was non-significantly higher in B12-deficient metformin users compared with non-metformin users.
CONCLUSION: This study shows that vitamin B12 insufficiency was frequently found in our population and may progress into B12 deficiency. It is also associated with neuropathy in subjects on metformin. Further interventional studies to assess the benefit of B12 treatment on painful neuropathy in patients on metformin may be warranted. B12 levels may be checked in people with T2DM using metformin for >2 years. © Author(s) (or their employer(s)) 2020. Re-use permitted under CC BY-NC. No commercial re-use. See rights and permissions. Published by BMJ.

Accumulating evidence suggests that metformin as the first choice of therapy for glycemic control may lead to low levels of serum vitamin B12.

Vitamin B12 insufficiency was frequently found in our population and may progress into B12 deficiency that is also associated with neuropathy in subjects on metformin.

In clinical practice, B12 levels may be checked especially in people with type 2 diabetes mellitus using metformin for >2 years to confirm B12 deficiency and/or B12 insufficiency.

Introduction

Worldwide, type 2 diabetes mellitus (T2DM) has affected an estimated 463 million people in 2019 and projected to reach 700 million by 2045, reported by the International Diabetes Federation (IDF).1 In the recent second National Diabetes Survey of Pakistan 2016–2017, the prevalence of diabetes was 26.3%.2 The American Diabetes Association (ADA), the European Association for the Study of Diabetes and the IDF recommend metformin as the first choice of therapy for glycemic control.1 3 Accumulating evidence from both observational and interventional studies has revealed that vitamin B12 deficiency may occur with metformin treatment.4 It has also been reported that vitamin B12 deficiency ranges from 9% to 52% in people with T2DM and has been partially attributed to long-term use of metformin.5–7

Vitamin B12 is essential for remethylation of homocysteine (Hcy) to methionine and B12 deficiency could lead to hyperhomocysteinemia, which has been associated with macrovascular complications in people with T2DM.8 B12 deficiency may also increase the severity of peripheral neuropathy in T2DM.9 However, reports are contradictory on the association between metformin-induced vitamin B12 deficiency and peripheral neuropathy.9–11 Furthermore, there are limited studies assessing metformin-induced vitamin B12 deficiency in people with T2DM12–14 and no such study assessing the relationship to diabetic neuropathy in Pakistan.

This study was undertaken to establish the prevalence of B12 deficiency in people with T2DM treated with metformin and its relationship to diabetic peripheral neuropathy (DPN) in Pakistan.

Methodology

This prospective multicenter observational study was conducted by Baqai Institute of Diabetology and Endocrinology (BIDE), Baqai Medical University (BMU), Karachi, Pakistan. Duration of study was between May 2018 and January 2019. An estimated sample size of 1000 subjects of which 750 have T2DM treated with metformin for >2 years and 250 non-diabetics without metformin was calculated. Subjects were selected from seven tertiary care centers across four provinces of Pakistan (Sindh, Punjab, Baluchistan and Khyber Pakhtunkhwa).

Subjects with a history of pernicious anemia, iron deficiency anemia, malabsorption (celiac disease, inflammatory bowel disease, gastrointestinal surgery), malnutrition (pure vegans, anorexia nervosa), history of thyroid disease and thyroxine treatment and/or a history of other organ-specific autoimmune conditions (vitiligo, Addison’s, primary ovarian failure, hypoparathyroidism), peripheral arterial disease and history of another cause of neuropathy were excluded. Subjects with previous gastric resection or bariatric surgery or on a vegetarian diet, who had received oral or intramuscular vitamin B12 supplementation, vitamin D supplementation, multivitamins, calcium supplements and proton-pump inhibitors (PPI) within the last 3 months, pregnancy and hearing or visual impairment or dementia were also excluded.

Baseline demographic and anthropometric details including age, gender, duration of metformin use, daily dose of metformin, blood pressure and body mass index (BMI) were noted using predesigned questionnaire. Blood samples were collected into a dedicated evacuated tube for biochemical parameters including hemoglobin (Hb), serum vitamin B12, and Hcy levels. From all centers, blood samples were transported to the laboratory of BIDE-BMU. Equipment used throughout the study were standardized with measure of quality assurance.

Vitamin B12 was analyzed using the Roche Diagnostic cobas e411 Immunoassay System—a fully automated, random access, software-controlled system for immunoassay analysis. The e411 vitamin B12 assay employs a competitive test principle using intrinsic factor specific for vitamin B12. In the sample, vitamin B12 competes with the added vitamin B12 labeled with biotin for the binding sites on the ruthenium-labeled intrinsic factor complex. Serum vitamin B12 >300 pg/mL was defined as normal, 200–300 pg/mL insufficient and <200 pg/mL as deficient.15

Subjects with vitamin B12 deficiency and insufficiency underwent assessment of Hcy levels (<15 µmol/L normal, ≥15 µmol/L hyperhomocysteinemia).16 17 Subjects underwent assessment of vibration perception threshold (VPT), Douleur Neuropathique 4 (DN4) score and Diabetic Neuropathy Symptom (DNS) score. VPT was measured on the pulp of the large toe on both right and left legs with a neurothesiometer.18 VPT was considered normal (<15V), intermediate (16–25V), and abnormal (>25V).19 The DN4 comprised 10 questions and a score ≥4 was used to define neuropathic pain.20 A DNS score ≥1 was considered to be indicative of neuropathy.21

Statistical analysis

Data analysis was performed in Statistical Package for Social Sciences (SPSS V.20). Student’s t-test, analysis of variance, χ2 test, and Fisher’s exact test were applied to check the significant difference between groups. Pearson’s correlation analysis was used to examine the relationship between vitamin B12 and other parameters. A two-tailed p value <0.05 was considered statistically significant.

Results

Out of 1000 sample size, 932 subjects were recruited of whom 287 (30.8%) were not on metformin supplementation and 645 (69.2%) were on metformin supplementation. The mean age of non-metformin users was 39.77±14.95 years and metformin users were 51.16±14.64 years. Metformin users had a higher BMI (27.91±5.12 vs 26.03±5.42, p<0.0001), systolic blood pressure (134.41±18.58 vs 126.94±18.06, p<0.0001) and diastolic blood pressure (81.61±13.62 vs 78±14.6, p=0.001). Hb was significantly lower in metformin users (13.41±2.32) compared with non-metformin users (14.05±2.3) (table 1).

Table 1

Comparison of demographic, anthropometric and clinical characteristics between non-metformin users and metformin users

Variable	Non-metformin users	Metformin users	P value	Overall
n	287	645	–	932
Age (years)	39.77±14.95	51.16±14.64	<0.0001	47.66±15.64
Gender
Male	157 (54.7%)	280 (43.4%)	0.001	437 (46.9%)
Female	130 (45.3%)	365 (56.6%)		495 (53.1%)
Marital status
Single	68 (23.7%)	12 (1.9%)	<0.0001	80 (8.6%)
Married	219 (76.3%)	633 (98.1%)		852 (91.4%)
Duration of DM (years)	–	8.03±5.4	–	8.03±5.4
BMI (kg/m2)	26.03±5.42	27.91±5.12	<0.0001	27.36±5.28
Systolic blood pressure (mm Hg)	126.94±18.06	134.41±18.58	<0.0001	132.2±18.73
Diastolic blood pressure (mm Hg)	78±14.6	81.61±13.62	0.001	80.54±14.01
Sulfonylureas	–	119 (18.44%)		119 (18.44%)
Thiazide diuretics	–	39 (6.06%)		39 (6.06%)
Dipeptidyl peptidase-4 (DPP4) inhibitor	–	164 (25.4%)		164 (25.4%)
Hb (g/dL)	14.05±2.3	13.41±2.32	<0.0001	13.61±2.33

Data presented as n (%) or mean±SD.

Student’s t-test and χ2 test were applied.

P<0.05 considered to be statistically significant.

BMI, body mass index; DM, diabetes mellitus.

Overall, B12 deficiency (<200 pg/mL) was significantly higher in metformin users of 25 (3.9%) compared with non-metformin users of 6 (2.1%), while B12 insufficiency (200–300 pg/mL) was significantly lower in metformin users of 117 (18.4%) compared with non-metformin users of 80 (27.9%). Subjects with B12 deficiency and insufficiency with hyperhomocysteinemia (≥15) were found in 19 (76%) µmol/L and 69 (60.5%) µmol/L in metformin users compared with 6 (100%) µmol/L and 57 (73.1%) μmol/L in non-metformin users, respectively (table 2).

Table 2

Vitamin B12 and homocysteine in non-metformin and metformin users

Parameters	Metformin use		P value
	No	Yes
Vitamin B12
Deficiency (<200)	6 (2.1%)	25 (3.9%)	0.002
Insufficiency (200–300)	80 (27.9%)	117 (18.4%)
Normal (>300)	200 (70%)	494 (77.7%)
B12 levels and homocysteine
B12 deficient and homocysteine
Normal	0 (0%)	6 (24%)	0.197
Hyper	6 (100%)	19 (76%)
B12 insufficient and homocysteine
Normal	21 (26.9%)	45 (39.4%)	0.05
Hyper	57 (73.1%)	69 (60.5%)
Overall homocysteine
Normal	21 (25%)	51 (36.7%)	0.047
Hyper	63 (75%)	88 (63.3%)

Data presented as n (%).

χ2 test was applied.

P<0.05 considered to be statistically significant.

Either the VPT>25 or DN4 score ≥4 was significantly higher in B12-deficient metformin users compared with non-metformin users. Similarly, DNS score ≥1 was non-significantly higher in B12-deficient metformin users compared with non-metformin users (table 3).

Table 3

Association of VPT, DNS and DN4 scores with B12 deficiency

Parameters	Non-metformin users					Metformin users
	B12					B12
	<200	200–300	>300	P value	Overall	<200	200–300	>300	P value	Overall
VPT
<15	4 (66.6%)	33 (44.6%)	65 (33.5%)	<0.0001	102 (37.2%)	7 (29.2%)	69 (68.3%)	53 (16.8%)	<0.0001	129 (29.3%)
15–25	1 (16.7%)	12 (16.2%)	21 (10.8%)		34 (12.4%)	0 (0%)	14 (13.9%)	45 (14.3%)		59 (13.4%)
>25	1 (16.7%)	29 (39.2%)	108 (55.7%)		138 (50.4%)	17 (70.8%)	18 (17.8%)	217 (68.9%)		252 (57.3%)
DNS
<1	5 (83.3%)	61 (79.2%)	116 (62.7%)	0.001	182 (64.2%)	4 (16%)	31 (27%)	76 (18.8%)	0.137	111 (20.4%)
≥1	1 (16.7%)	16 (20.8%)	69 (37.3%)		86 (35.8%)	21 (84%)	84 (73%)	328 (81.2%)		433 (79.6%)
DN4
<4	6 (100%)	70 (92.1%)	146 (80.2%)	0.01	222 (84.1%)	7 (29.2%)	70 (60.9%)	175 (43.3%)	<0.0001	252 (46.4%)
≥4	0 (0%)	6 (7.9%)	36 (19.8%)		42 (15.9%)	17 (70.8%)	45 (39.1%)	229 (56.7%)		291 (53.6+%)

Data presented as n (%).

χ2 test was applied.

P<0.05 considered to be statistically significant.

DN4, Douleur Neuropathique 4; DNS, Diabetic Neuropathy Symptom; VPT, vibration perception threshold.

B12 levels were not associated with age (r=0.172, p<0.0001), BMI (r=−0.089, p=0.013), duration of diabetes (r=0.017, p=0.706), VPT (r=0.262, p<0.0001), DNS score (r=0.128, p<0.0001) and DN4 score (r=0.318, p<0.0001), while B12 levels were negatively correlated to duration of metformin use (r=−0.24; p=0.0001) (figure 1A), dose of metformin use (r=−0.21; p=0.0001) (figure 1B), HbA1c (r=−0.09, p=0.378) and Hcy levels (r=−0.147, p=0.038) (table 4).

Figure 1

Correlation of vitamin B12 level with (A) duration and (B) dose of metformin in metformin users.

Table 4

Correlation between B12 and various parameters

Parameters	Correlation	P value
Age	0.172	<0.0001
BMI	−0.089	0.013
Duration of DM	0.017	0.706
Duration of metformin	−0.236	<0.0001
Daily dose of metformin	−0.21	<0.0001
VPT	0.262	<0.0001
DNS score	0.128	<0.0001
DN4 score	0.318	<0.0001
HbA1c	−0.09	0.378
Homocysteine	−0.147	0.038

P<0.05 considered to be statistically significant.

BMI, body mass index; DM, diabetes mellitus; DN4, Douleur Neuropathique 4; DNS, Diabetic Neuropathy Symptom; VPT, vibration perception threshold.

Discussion

This is the largest multicenter study to date assessing the relationship between metformin use and B12 deficiency, and its association with diabetic neuropathy. In this study, significantly increased prevalence of B12 deficiency was observed in people with T2DM treated with metformin as compared with non-metformin users.9 14 22 On contrary, B12 insufficiency was significantly higher in non-metformin users compared with metformin users. It indicates that B12 insufficiency was generally found in our population, and after initiation of metformin in people with diabetes, the B12 insufficiency may develop into B12 deficiency. Moreover, we observed that subjects with B12 deficiency have high VPT (>25), DNS score (≥1) and DN4 score (≥4) as compared with non-metformin users, similar to Algeffari and Singh et al’s studies.23 24 Indeed, Zalaket et al showed reversal of neuropathy after B12 supplementation.25

Regarding the clinical significance of biochemical vitamin B12 deficiency versus true tissue deficiency, a significant debate already exists. Up to now, the most commonly surrogate markers used for detection of vitamin B12 deficiency are plasma Hcy and methylmalonic acid.26 In our population, concurrently elevated Hcy levels were also observed in people with B12 insufficiency and B12 deficiency.27 However, measurement of additional biomarkers for more comprehensive assessment of B12 deficiency, such as holotranscobalamin, methylmalonic acid, red cell-B12, and plasma concentrations of methylation indices, is beyond the scope of our study.

Vitamin B12 deficiency is a multifactorial condition caused by insufficient intake (nutritional deficiency) as well as acquired or inherited defects that disrupt B12 absorption and processing pathways. Similarly, metformin-induced B12 deficiency is also thought to occur due to vitamin B12 malabsorption such as alteration of bile acid metabolism, small intestinal bacterial overgrowth, or effects on intrinsic factor secretion, but a more currently accepted explanation is the interference by metformin on calcium-dependent membrane action responsible for vitamin B12 intrinsic factor absorption in the terminal ileum.28 The use of PPIs is also thought to contribute to B12 deficiency, although this does not appear to be a factor in our study. Both observational and interventional studies have shown that the duration and dose of metformin are also associated with B12 deficiency and neuropathy.11 29 30 A recent study from Qatar, however, showed no association between metformin use and B12 deficiency or diabetic neuropathy.20 de Groot-Kamphuis et al have shown a lower prevalence of DPN in people with T2DM on metformin compared with those not on metformin.31 Our study confirms a weak but significant correlation between B12 levels and duration and dose of metformin. A significant association has also been found with age, gender, married individuals, BMI and blood pressure with B12 levels in metformin users.29 In the present study, the metformin users were significantly older, but no such association between age and B12 levels exists in related studies.11 32

In current study, significantly increased but low Hb levels were observed in metformin users compared with non-metformin users. In prior studies, the significant association between B12deficiency and low Hb concentration was also noted.33 34 Metformin-induced B12 deficiency has been attributed to alterations in small bowel motility and enhanced bacterial overgrowth or interference of metformin with calcium-dependent intrinsic factor release.35 To date, there are no guidelines recommending routine screening for B12 deficiency in T2DM subjects on metformin, although the recent ADA-ADA consensus guidelines recommended the assessment of B12 in subjects with DPN being treated with metformin.36

Strengths and limitations

This is a cross-sectional multicenter study and therefore a true cause effect between metformin use and B12 deficiency cannot be established. We have attempted to exclude other confounding factors, although the patients on metformin were older. We lack complete data regarding VPT, DNS score and DN4 score from all centers is our limitation. Glycemic control not being assessed is also a limitation of this study. All laboratory assessments were undertaken in a central lab and exactly the same protocols were used to assess for diabetic neuropathy and painful diabetic neuropathy.

Conclusion

This study shows that vitamin B12 insufficiency was frequently found in our population and may progress into B12 deficiency. It is also associated with neuropathy in subjects on metformin. Further interventional studies to assess the benefit of B12 treatment on painful neuropathy in patients on metformin may be warranted. B12 levels may be checked in people with T2DM using metformin for >2 years.