Intake of vitamin B12 in relation to vitamin B12 status in groups susceptible to deficiency: a systematic review

Objective To systematically review the evidence for whether habitual or different levels of experimental intake of vitamin B12 from diet and supplements is sufficient to ensure adequate B12 status in groups most susceptible to vitamin B12 deficiency. Methods We searched MEDLINE, Embase, Cochrane Central Register of Controlled Trials and Scopus up to 21 May 2021, for intervention studies, prospective cohort studies and case-control studies assessing B12 intake from diet and/or supplements in relation to B12 status (s/p-B12, holotranscobalamin, methylmalonic acid, homocysteine or breastmilk B12). Cross-sectional studies were eligible for studies conducted during pregnancy and lactation. Included populations were children (0–18 years), young adults (18–35 years), pregnant or lactating women, older adults (≥65 years) and vegans or vegetarians. Study selection, data extraction and risk of bias assessment were conducted by two assessors independently. The evidence was synthesized qualitatively and classified according to the World Cancer Research Fund. Results The searches yielded 4855 articles of which 89 were assessed in full text and 18 included. Three studies were conducted during pregnancy and three during lactation or infancy – all observational. Eight studies were conducted among older adults; most were interventions among B12-deficient participants. Four studies were eligible for vegetarian and vegans, all interventions. The strength of evidence that habitual B12 intake or an intake in line with the current Nordic recommended intake (RI) is sufficient to ensure adequate status was considered Limited – no conclusion for all included populations. Conclusion Evidence is insufficient to assess if or which level of B12 intake is sufficient to maintain adequate status for all included populations. Population-based cohort studies and low-to-moderate dose interventions that address this question are highly warranted.


Introduction
Vitamin B 12 or cobalamin (B12) is an essential nutrient that is vital for human health, primarily as a coenzyme in one-carbon metabolism (1). B12 contributes to blood cell formation, synthesis of DNA, regeneration of the amino acid methionine and the maintenance of myelin that protects the nerve cells among other functions (2). B12 is absorbed in the form found in foods of animal origin, given adequate secretion of hydrochloric acid and the glycoprotein intrinsic factor excreted from the parietal cells of the stomach. Once absorbed, a low intake will suffice to maintain adequate status as the nutrient is stored in the body and can be partially reabsorbed by the intestine after excretion with bile (3).
B12 deficiency may lead to megaloblastic anaemia, characterized by large and immature red blood cells (2). Other consequences of deficiency include neurological and cognitive impairment (4). While clinical signs of B12 deficiency may take years to develop, several biological markers of B12 status are available, including serum or plasma concentrations of total vitamin B12/ cobalamin (B12), holotranscobalamin (holoTC), methylmalonic acid (MMA) and homocysteine (tHcy) (4). A major risk factor for B12 deficiency is the autoimmune disease pernicious anaemia, which causes destruction of the gastric parietal cells and thereby the intrinsic factor (4). Other population groups are however at increased risk of B12 deficiency, including those who consume limited or no food of animal origin (5), young children with low B12 body stores (6) or older adults with reduced secretion of hydrochloric acid or intrinsic factor (4,7).
Cross-sectional studies show that B12 status stabilizes at intakes of 4-10 μg/d (8). The Nordic Nutrition Recommendation's (NNR) 2012 recommended intake (RI) for vitamin B12 (8) in different population groups can be seen in Table 1. There are uncertainties regarding the vitamin B12 content of breastmilk, and the requirement and intake of infants is understudied (9).
With the current emphasis on lower meat intake and plant-based diets, it is unclear if the intake of vitamin B12 is sufficient to maintain adequate B12 status in people following different dietary patterns, e.g. vegetarians or vegans. In addition, it is unknown if habitual B12 intake reaches RI and if this is sufficient to ensure adequate status in all age groups, including those most susceptible to deficiency. The aim of this systematic review was to summarize the evidence for whether habitual or different levels of experimental intake of vitamin B12 from diet and supplements are sufficient to ensure adequate B12 status in children, pregnant and lactating women, young adults, older adults, vegetarians and vegans.

Methods
This systematic review was conducted according to the guidelines for systematic reviews, developed for the 2022 revision of the NNR (10,11) and preferred reporting for systematic reviews (12). The NNR 2022 is funded by the Nordic Council of Ministers and governmental food and health authorities of Norway, Finland, Sweden, Denmark and Iceland (13). A study protocol was published prior to article selection in database PROSPERO (https://www. crd.york.ac.uk, CRD42021244376).

Eligibility criteria
The research question was specified by the NNR 2022 Committee and the NNR Systematic Review Centre (i.e. the authors) by defining the population, intervention/ exposure, control, timing, study design and setting (PI/ ECOTSS). The PI/ECOTSS ( Table 2) included six healthy populations relevant for the Nordic setting: (1) children (0-18 years), (2) pregnant women, (3) lactating women, (4) young adults (18-35 years), (5) older adults (≥65 years) and (6) vegetarians, including vegans. The intervention/ exposure included both supplemental and dietary intake of vitamin B12, and the comparator was different levels of intake (including placebo). Outcomes were defined as biological markers of vitamin B12 status, either B12, holoTC, MMA, tHcy in plasma or serum, or B12 in breastmilk.
Eligible study designs were randomized control trials (RCTs), cohort studies or case-control studies. Further, cross-sectional studies were included for pregnant and lactating populations, due to the limited time frame of gestation and lactation. Minimum study duration was 4 weeks for RCTs and 12 months for prospective studies. Intervention studies using intravenous vitamin B12 supplementation or toothpaste enriched with vitamin B12 were excluded. Observational studies were limited to European and North American populations.

Search strategy
The literature searches were performed by research librarians from the University of Oslo 21 April 2021 in databases MEDLINE, Embase, Cochrane Central Register

Study selection and data extraction
Screening and selection of studies for inclusion/exclusion was performed independently by two authors (LB and CLA). The screening of titles and abstracts was performed in Rayyan (14). A pilot test was conducted using 10% of the titles and abstracts, in order to harmonize the process. Discrepancies were resolved by discussion with a third author (AÅ). Data from full-text papers included in the systematic review were extracted in standardized extraction forms by authors working independently (EKA, AR, FS).

Risk of bias assessment
Risk of bias in each included study was assessed by two authors (CLA and BT), working independently. The assessment tools used were for intervention studies Cochrane's Risk of bias 2.0 (15) (18). Risk of bias was visualized by using web app Risk-of-bias VISualization (robvis) (19).

Synthesis and strength of evidence
The evidence was synthesized qualitatively, based on study characteristics, context, strengths and limitations, heterogeneity and relevance. In accordance with the guidelines for systematic reviews, meta-analyses were considered if deemed appropriate to combine/pool the different studies, but only when more than three independent RCTs or five cohort studies exist. Strength of evidence was appraised based on risk of bias, consistency/ heterogeneity and precision of the evidence, according to the World Cancer Research Fund's grading: 'Convincing', 'Probable', 'Limited -suggestive', 'Limited -no conclusion', 'Substantial effects unlikely' (13).

Results
The searches yielded 4855 unique articles, of which 89 were read in full text and 18 included (Fig. 1). Articles excluded after full text screening are shown in Supplement 2. For two out of the six populations selected for this systematic review (children other than breastfed infants, and young adults), no eligible studies were identified. Due to heterogeneity in types of interventions, exposures and reported outcomes, no meta-analysis was performed. Results are thus limited to qualitative synthesis.

Study characteristics
Three studies on pregnant women were included (Table 3). In a prospective cohort study from Germany, Koebnick et al. (20) followed (from gestational week 9-12 until the last trimester) 27 lacto-ovo vegetarian pregnant women, 43 low meat eaters (defined as consuming meat <300 g/week and meat products <105 g/week) and 39 omnivore women consuming a Western diet. In another prospective cohort study from Canada, Visentin et al. (21) followed 368 women from gestational week 12-16 until delivery. In a Dutch cross-sectional study, Denissen et al. (22) included 1365 pregnant women in the third trimester. Two studies on lactating women were included (Table 3). In a Danish prospective cohort study, Greibe et al. (23) included 60 mother-child pairs and studied associations both between maternal B12 intake with maternal B12 status and B12 content of breastmilk in addition to associations with infant B12 status. In a cross-sectional study from Norway, Henjum et al. (24) included 193 women 0-6 months postpartum and assessed B12 intake from supplements and diet as well as B12 in breastmilk.
One prospective cohort study from Norway by Hay et al. (25) included infant data on both B12 intake and status at 6, 9 and 12 months.

Intake of B12 in relation to status
Among the studies conducted during pregnancy, Koebnick et al. (20) found that median dietary intake of B12 was 2.5 µg/d, 3.8 µg/d and 5.3 µg/d among lactoovo vegetarians, low meat eaters and omnivore pregnant women, respectively ( Table 4). The corresponding proportion of B12 supplement users was 32, 28 and 21%. The odds of B12 deficiency (<100-130 pmol/L) were almost 4 times higher among ovo-lacto vegetarians and almost 2 times higher among low meat eaters, compared to omnivores. Visentin et al. (21) found that prevalence of B12 deficiency (s-B12 <148 pmol/L) was 17% in early pregnancy and 38% in mid-to-late pregnancy. For every 10-µg increment in maternal total vitamin B12 intake, s-B12 increased by 1.04 pmol/L in both early pregnancy and mid-to-late pregnancy in repeated cross-sectional analyses. Corresponding decreases were observed for tHcy and MMA. In the cross-sectional study by Denissen et al. (22), the mean vitamin B12 intake for all pregnant women was 5.0 µg/day. The corresponding results for subgroups were for omnivores 5.1 µg/day, pescatarians 4.3 µg/ day, vegetarians 3.5 µg/day (self-defined) and lacto-ovo vegetarians 2.3 µg/day (defined by researchers based on FFQ data). The authors found dose-response associations between total dietary vitamin B12 intake with p-B12, holoTC and MMA. The odds of B12 deficiency were lower in the second (5 µg/day) and third tertiles (9.1 µg/ day) of B12 intake, compared with the first tertile (3.2 µg/ day). The analysis showed that a vitamin B12 intake of ≥4.2 μg/day was associated with ~90% lower odds of deficiency compared to lower intake.
Among the studies conducted during lactation, Greibe et al. (23) found that maternal B12 status was significantly correlated with B12 content of breastmilk at 4 months postpartum. Consequently, B12 content of breastmilk at 4 months postpartum correlated with infant p-B12, while there was no correlation between breastmilk B12 and holoTC or MMA in the children at any time point. Exclusively breastfed infants at 4 months had lower p-B12 and holoTC concentrations than infants not exclusively breastfed. Henjum et al. (16) found no significant correlation between maternal B12 intake (from either supplements or diet) and breastmilk B12 content. However, breastmilk B12 content was found to decrease over time during the 6 month period.

Intake of vitamin B12
Hay et al. (17) found that both at 6 and 12 months, all biomarkers of infant B12 status were affected by feeding pattern. At 6 months, B12 status was lower among breastfed infants (with or without complementary feeding) compared to non-breastfed infants. At 12 months, those partially breastfed still had lower B12 status and intake (excluding B12 content of breastmilk) than non-breastfed children.

Risk of bias and strength of evidence
Overall risk of bias in studies in pregnant women was regarded as serious for Koebnick et al. (20) and moderate for Visenitin et al. (21) (Fig. 2). For the cross-sectional study by Denissen et al. (22), we only assessed study quality, which was regarded as moderate. The overall risk of bias in studies of lactating women and their offspring was regarded as low for Greibe et al. (23) and low for Hay et al. (25). The general quality of the cross-sectional studies was regarded as low for Henjum et al. (24).
The strength of evidence that habitual B12 intake or an intake in line with the current Nordic RI is sufficient to ensure adequate status during pregnancy, lactation or infancy is considered Limited -no conclusion. There was Limited -suggestive evidence that breastfed infants (exclusive and/or any breastfeeding) have lower B12 status than non-breastfed infants at 4-6 months of age. Evidence was however Limited -no conclusion for that breastfed infants were at higher risk of B12 deficiency. Overall, the evidence was regarded as limited due to scarcity of data and a lack of eligible studies.

Study characteristics
Eight studies on older adults were included -five RCTs, conducted in Australia (26), the Netherlands (27,28), Switzerland (29) and the United Kingdom (30), one nonrandomized intervention study conducted in the   United States (31), one RCT conducted in the United States (32) and one prospective cohort study conducted in Germany (33) ( Table 5). Sample sizes in intervention studies ranged from 23 to 149 subjects with a mean/median age of 75-82 years. Study duration ranged from 4 to 18 weeks. Supplement doses ranged from 2.5 to 1000 µg/day. One study gave capsules or vitamin B12-fortified milk (28), while the others used oral supplements. The prospective cohort by Jungert et al. (33) studied 332 adults ≥60 years, for 12 years.

Intake of B12 in relation to status
Among studies with deficient or borderline-deficient subjects, Seal et al. (26) found that B12 supplementation of 50 µg/d, but not 10 µg/d, improved B12 status (Table 6). Eussen et al. (27) found that supplemental B12 doses of 2.5, 100, 250, 500 and 1000 µg/d increased s-B12 and holoTC while MMA and tHcy decreased, all in a dose-response manner. Favrat et al. (29) found that B12 supplementation of 1000 µg/d significantly increased s-B12 and decreased MMA and tHcy in comparison to placebo. Hill et al. (30) found that 10 µg/d elicited improvement in B12 status but 500 µg/d was required to normalize p-B12 and

Intake of vitamin B12
holoTC in 90% of participants during the 2-month intervention. Dhonokushe-Rutten et al. (28) found that fortified milk increased s-B12 and decreased MMA and tHcy similar to capsule supplements. Rajan et al. (31) showed that 25 µg/d for 6 weeks was sufficient to normalize MMA in 2 out of 20 subjects while 100 µg/d normalized MMA in an additional 5 out of 20 subjects. Most did not normalize their elevated serum MMA levels until the 1000 µg dose. Among nondeficient older participants, Stabler et al. (32) reported no between group differences in s-B12 concentration (0, 25 or 100 µg/day) in the RCT study. However, MMA increased in the placebo group, while tHcy decreased in the 100 µg group. Stabler et al. (32) reported that 30% of subjects had low B12 status based on elevated MMA (>271 nmol/L). These were not included in the RCT.
In the cohort study by Jungert et al. (33), s-B12 was generally adequate (median 267 pmol/L, Table 6). In addition, 9% of females and 16% of males had B12 deficiency (s-B12 ≤ 148 pmol/L). Median B12 intake was 5.3 µg/d and 28% of females and 15% of males had an intake below 4.0 µg/d. In longitudinal analysis, s-B12 was associated with supplemental but not dietary B12 intake. Supplement users had ~97 pmol/L higher s-B12 than nonusers.  (Fig. 3). In the nonrandomized intervention by Rajan et al. (31), risk of bias was considered low overall and for all included domains. Risk of bias in the cohort by Jungert et al. (33) was considered moderate (Fig. 2).

Risk of bias and strength of evidence
The strength of evidence that habitual B12 intake or an intake in line with the current Nordic RI is sufficient to ensure adequate status among older adults is considered Limited -no conclusion due to a lack of eligible population-based prospective studies and low-dose intervention studies. The strength of evidence that higher experimental doses of B12 result in higher biomarkers of status in older adults with B12 deficiency is considered Convincing but no optimal intake level could be defined.

Study characteristics
Four RCTs conducted among vegetarians and vegans were included (34-37) ( Table 7). The studies were conducted in Hong Kong (34), the United Kingdom (35), Germany (36) and Italy (37). All studies were conducted among adults, three among practising vegetarians or vegans (34,35,37), while one studied effects of a vegan diet (36).  Kwok et al. (34) found that 500 µg/d of B12 raised s-B12 to 380 pmol/L and lowered tHcy to 11.3 (Table 8). Obersby et al. (35) found that, in intention-to-treat (ITT) analysis, tHcy decreased from 14.7 to 9.1 µmol/L in the group supplemented with 500 µg/d. Del Bo et al. (37) found that both 2000 µg/week and 50 µg/day (350 µg/ week) improved B12 status, reflected by higher s-B12 and holoTC and lower MMA and tHcy. Only s-B12 differed significantly between the two doses at 90 days follow-up. Lederer et al. (36) found that after 4 weeks of vegan diet, s-B12 and holoTC had decreased significantly compared to the meat-rich control diet. MMA and tHcy were not significantly different.

Risk of bias and strength of evidence
Overall risk of bias was considered low for Del Bo et al. (37) and Kwok et al. (34), some concerns for Lederer et al. (36) and high for Obersby et al. (35) (Fig. 3).
The strength of evidence that habitual B12 intake or an intake in line with the current Nordic RI is sufficient to ensure adequate status among vegetarians and/or vegans is considered Limited -no conclusion as there were no eligible prospective cohort studies that investigated B12 intake in relation to status among vegetarians or vegans.

Discussion
The results of this systematic review show that there is a scarcity of prospective studies into B12 intake in relation to B12 status for most of the included populations. In addition, most intervention studies used high supplemental doses and short study durations, making it difficult to conclude on the long-term effect of low-dose B12 supplementation. There are some indications that breastfed infants have lower B12 status than non-breastfed infants. Data are however not sufficient to assess the relevance of this finding. There were not enough data to assess B12 intake in relation to B12 status in other groups.
In pregnancy, the strength of evidence that habitual B12 intake or an intake in line with the current Nordic RI is sufficient to ensure adequate status was considered Limited -no conclusion. Assessing dietary intake and   (32) status in the pregnant state is associated with some difficulties. A dietary assessment in early pregnancy can be obscured by pregnancy nausea and/or vomiting (38) and might not be an accurate reflection of habitual pre-pregnancy intake or intake as the pregnancy advances. In addition, plasma volume expansion can make changes in nutritional biomarkers over pregnancy difficult to assess (39). Future studies, preferably prospective-and population-based, are required to investigate how B12 intake relates to both maternal and infant B12 status. During lactation, the strength of evidence that habitual B12 intake or an intake in line with the current Nordic RI is sufficient to ensure adequate status was considered Limited -no conclusion. There was however Limitedsuggestive evidence that breastfed infants had lower B12 status than non-breastfed infants at 4-6 months of age. This is similar with findings from a 2017 systematic review that found insufficient evidence to evaluate the timing of introducing of complementary food or beverage on infant B12 status (40). Further, a 2018 systematic review found evidence for associations between breastmilk B12 content and maternal intake of B12 (9) but methodological disparities obscured any firm conclusions. Since infant B12 stores can be low if maternal intake during pregnancy was low, ensuring adequate B12 intake during lactation is important to prevent deficiency (41). The relevance of the observed lower B12 status among breastfed infants cannot be determined based on the findings included in the current review.
Among older adults, the strength of evidence that habitual B12 intake or an intake in line with the current Nordic RI is sufficient to ensure adequate status among older adults is considered Limited -no conclusion. The interventional studies overall found higher B12 status with higher B12 intake, but the definitions of what constituted optimal B12 status differed and thus the interpretations of the findings. The RCTs mostly included older adults with B12 deficiency, likely caused by reduced absorption (42), and it is difficult to transfer the results to maintenance of vitamin B12 status. Thus, the results should be interpreted with that in mind. In addition, almost all studies were short term (<6 months) and may have not shown the full potential of low-dose vitamin B12 supplements to normalize B12 status. It is estimated that B12 deficiency affects approximately 2%-35% of the older population, depending on age range and region (4). These estimates may not be representative of the general older adult population in the Nordic region. Population-based studies that assess the prevalence of B12 deficiency in relation to intake among older adults are warranted.

Intake of vitamin B12
The strength of evidence that habitual B12 intake is sufficient to ensure adequate status among vegetarians and/or vegans is considered Limited -no conclusion. It is known that a vegan diet does not provide enough B12 and eventually requires supplementation (43). However, there was an unexpected lack of prospective cohort studies to show the effect of lacto-ovo vegetarian diet on B12 status. Thus, more prospective studies are needed to clarify if B12 supplementation should be recommended also to lacto-ovo vegetarians.

Limitations and strengths
Strengths of this review include a highly standardized process of literature searches, article selection, data extraction, risk of bias assessment and evidence grading. Limitations of this work are mainly related to the scarcity of data and relevant studies. Since B12 deficiency takes years to develop; cross-sectional studies were only included for pregnant and lactating women. This is both a limitation and a strength, since it not only greatly reduced the number of eligible studies but also ensured higher quality data. Lastly, all included RCTs were short term and will thus not show the long-term effects of B12 supplementation.

Conclusion
In conclusion, evidence is insufficient to assess if habitual B12 intake or an intake in line with the current Nordic RI is sufficient to maintain adequate status for all included populations. Population-based cohort studies and low-to-moderate dose interventions that address this question are highly warranted. assistance with the literature searches, and the university librarians at Karolinska Institutet for peer reviewing the search strategy.