Review Reports - Drug-Induced Vitamin B12 Deficiency: Current Evidence, Diagnostic Challenges, and Therapeutic Strategies

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The topic is clinically relevant, and the manuscript attempts to connect drug exposure, cobalamin metabolism, diagnosis, monitoring, and treatment. However, the evidence-identification process is not reproducible, the synthesis does not grade study quality, several observational associations are presented too causally, and clinically directive thresholds and treatment regimens are offered without adequate guideline-level support. The citation system is seriously corrupted, including shifted references and a figure citing a nonexistent reference.

1 L35-70 The epidemiological framing relies on loosely connected and sometimes extrapolated prevalence estimates. The text moves from 5–15% prevalence to values of 26% in the general population and 40% in older adults without defining deficiency, biomarker thresholds, or geographic context. It then estimates more than 150 million metformin users worldwide and up to 10% of the Spanish population using metformin, although the cited sources do not clearly establish these calculations. Similar broad estimates are used for chronic PPI exposure. Please replace extrapolations with directly reported, population-specific estimates and explicitly account for differences in assays and cutoffs. These unsupported estimates currently exaggerate the magnitude of the problem and weaken the rationale.

2 L72-84 The literature-search method is not reproducible and is insufficient for a review claiming to summarize current clinical evidence. The authors provide database names and isolated keywords but omit the search dates, full Boolean strategies, language restrictions, inclusion and exclusion criteria, duplicate removal, screening process, number of retrieved and included records, and reasons for excluding evidence. Terms such as ‘clinical relevance,’ ‘methodological quality,’ and ‘contribution’ are not operationally defined, and no formal assessment of study quality or certainty is reported. A narrative format does not justify an opaque evidence-selection process, particularly when diagnostic thresholds and treatment regimens are proposed.

3 L85-129 The physiology section is largely textbook description and does not establish a distinctive conceptual framework appropriate for Metabolites. Several statements need greater precision, including the assertion that the intestinal microbiota cannot synthesize enough cobalamin to meet human needs and the implication that hepatic stores allow deficiency to remain merely paucisymptomatic for a fixed 3–5 years. The review should distinguish dietary cobalamin, microbial corrinoids, circulating total B12, holotranscobalamin, intracellular cofactor availability, and functional metabolites. Figure 1 is described as an original figure adapted from references 1 and 10, but the degree of adaptation and permission status are unclear. The current section adds length without delivering the metabolic depth promised by the journal and title.

4 L133-195 The metformin section does not consistently separate mechanistic hypotheses, biochemical changes, and clinically meaningful deficiency. Reduced calcium-dependent uptake, altered motility, bacterial overgrowth, and microbiome changes are discussed in a manner that suggests comparable evidential support, although the underlying studies differ greatly in design and directness. Risk factors such as male sex, smoking, obesity, dyslipidemia, and coronary disease are listed without effect estimates or an explanation of whether they are independent predictors. The statement that the association is ‘clear’ and that decline begins early should be supported by a structured comparison of randomized trials, longitudinal cohorts, and cross-sectional studies, including definitions of deficiency, baseline B12 status, dose, duration, confounding, and certainty of evidence.

5 L196–244 and Table 1 The evidence concerning PPIs and combined PPI–metformin exposure is overstated. The manuscript acknowledges that the main evidence is observational and cannot establish causality, yet subsequently presents a dose- and duration-dependent relationship and potentially additive effects as if they were established clinical facts. Table 1 assigns ‘moderate to high’ evidence to the PPI association without a defined grading framework and gives fixed onset periods and treatment thresholds that are not derived transparently. It also conflates food-bound B12 malabsorption with clinically verified deficiency. The authors should systematically address confounding by age, comorbidity, diet, indication, polypharmacy, atrophic gastritis, and healthcare utilization. Without this analysis, the central comparison in the manuscript is unreliable.

6 L245–278 and Figure 2 The section on other drugs contains unsupported causal language, uneven evidence, and citation mismatches. Epidemiological studies are said to demonstrate a ‘causal relationship’ for H2 antagonists, which is not justified by observational association. Phenytoin, colchicine, cholestyramine, contraceptives, nitrous oxide, and GLP-1 receptor agonists are grouped together despite fundamentally different mechanisms and evidence levels. The reported annual incidence of 2.6% with GLP-1 receptor agonists is not linked to a clearly described primary study, population, diagnostic definition, or exposure duration. In addition, the citations are shifted: the anticonvulsant statement cites reference 20 although the relevant reference is 21, and the same displacement continues for contraceptives, nitrous oxide, and GLP-1 therapy. This section cannot support the broad title ‘Drug-Induced Vitamin B12 Deficiency.’

7 L279–418 and Figures 3 The clinical section combines manifestations of severe classical B12 deficiency with outcomes potentially associated with modest drug-related reductions in serum B12, without establishing that the same risk magnitude applies. The neuropathy discussion appropriately acknowledges heterogeneous findings, but the text nevertheless recommends periodic monitoring and implies preventability without defining which patients benefit, the monitoring interval, or whether screening improves neurological outcomes. The comparison of metformin-associated B12 deficiency with diabetic neuropathy requires a more rigorous account of reverse causation, diabetes duration, glycemic control, renal function, and ascertainment methods. Figure 3 is said to derive from references 20–22, but those references concern different drug exposures because of the citation displacement, making the figure’s evidential basis unclear.

8 L435–479 and Table 2 The diagnostic section presents heterogeneous biomarker thresholds as a quasi-universal algorithm. The proposed grey zone of 148–287 pmol/L, the statement that values above 287 pmol/L generally require no further work-up, and fixed cutoffs for MMA, homocysteine, and holotranscobalamin are highly dependent on assay, renal function, age, folate and vitamin B6 status, pregnancy, liver disease, and clinical context. Table 2 labels diagnostic performance merely as ‘low,’ ‘good,’ or ‘very good’ without sensitivity, specificity, confidence intervals, reference standards, or population. MMA is described as the most sensitive and specific marker while renal impairment is only briefly noted. These simplifications are unsuitable for a clinical review and could lead to misclassification.

9 L483–536 and Figure 4 The monitoring recommendations and author-developed diagnostic–therapeutic algorithm are not adequately supported. The manuscript alternates between risk-based monitoring, annual testing, baseline screening, and undefined ‘periodic’ measurement without reconciling the recommendations or grading their evidence. The ADA 2026 guidance is cited as reference 35, but reference 35 is an observational monitoring study and the ADA document is reference 36. Most seriously, Figure 4 states that it was adapted from references 45 and 35 even though the bibliography contains only 43 references. The algorithm was not developed through a stated consensus method, evidence-to-decision framework, or external validation. A clinically directive algorithm cannot be accepted when its source citations are incorrect or nonexistent.

10 L537–699 The treatment section is excessively prescriptive relative to the evidence reviewed. It provides fixed intramuscular loading and maintenance schedules, a universal 8-week oral regimen, and weekly oral maintenance for metformin-associated deficiency without adequately distinguishing cyanocobalamin from hydroxocobalamin, national practice standards, etiology, adherence, severity, neurological involvement, or the need for rapid correction. Evidence that high-dose oral therapy normalizes serum B12 is treated as equivalent to evidence for neurological recovery, despite the manuscript acknowledging that clinical-outcome data are limited. Statements that oral treatment can be used in ‘almost all scenarios’ and that methylcobalamin may be more effective for neuropathic symptoms require stronger evidence. These recommendations could influence patient care and therefore require guideline-level substantiation.

11 L700–744 and 745–833 The conclusions are more definitive than the underlying evidence and the reference system is not sufficiently reliable for publication. The review concludes that metformin and PPIs act through potentially additive mechanisms and advocates proactive detection, although the evidence-selection process is opaque and many recommendations rest on observational data. Only 43 references are used for an exceptionally broad review, several central areas rely on secondary sources, and references 17 and 18 appear to be two fragments of one article, shifting subsequent numbering. Numerous in-text citations are malformed, including ‘functions10,’ ‘neuropathy2,’ ‘deficiency29,’ ‘patients1,37,’ and ‘disease1.’ The work was funded by ITF Research Pharma, one author is employed by the sponsor, and the other reports consultancy; although disclosed, this context makes transparent selection and grading of treatment evidence especially important.

Comments on the Quality of English Language

The manuscript is generally readable but contains overly long sentences, awkward grammatical structures, misuse of articles and prepositions, inconsistencies in singular and plural forms, and nonuniform punctuation and terminology.

Author Response

Reply to the Review Report

REVIEWER 1

“The topic is clinically relevant, and the manuscript attempts to connect drug exposure, cobalamin metabolism, diagnosis, monitoring, and treatment. However, the evidence-identification process is not reproducible, the synthesis does not grade study quality, several observational associations are presented too causally, and clinically directive thresholds and treatment regimens are offered without adequate guideline-level support. The citation system is seriously corrupted, including shifted references and a figure citing a nonexistent reference”.

Answer: We sincerely thank the Reviewer for the thorough and constructive evaluation of our manuscript. We have carefully addressed each of the concerns raised and have revised the manuscript accordingly. Below, we provide a detailed, point-by-point response to all comments, together with a description of the corresponding changes made in the revised version. In particular, we have strengthened the description of the evidence-identification process, clarified the interpretation and limitations of observational evidence, revised clinically directive statements to ensure appropriate support from guidelines and primary literature, and carefully reviewed and corrected the reference and citation system throughout the manuscript, including the references associated with the figures. We believe that these changes have substantially strengthened the manuscript.

The manuscript was also carefully revised for English language and style, with improvements in grammar, sentence structure, terminology consistency and overall readability.

Answer: We thank the Reviewer for this important comment. We have carefully reviewed references [1–9] and revised the epidemiological section accordingly. Specifically, we have removed prevalence estimates and extrapolations that were not adequately supported by the cited sources, clarified the influence of diagnostic criteria and biomarker thresholds on reported estimates of vitamin B12 deficiency, and replaced generalized estimates of medication use with population-specific data where available. We have also revised the wording to distinguish observed epidemiological findings from estimates and to ensure that the conclusions are proportionate to the available evidence. These changes improve the accuracy, transparency, and consistency of the epidemiological background and strengthen the rationale for the clinical relevance of drug-associated vitamin B12 deficiency.

Answer: We thank the Reviewer for this constructive comment. Although this manuscript was designed as a narrative review rather than a systematic review or meta-analysis, we agree that greater methodological transparency is essential. Accordingly, we have revised the Methods section to provide further details on the literature-search strategy, including the databases searched, the search period and date of the final search, the search terms used, the eligibility criteria, and the manual screening of reference lists from key reviews, meta-analyses, and clinical guidelines. We have also clarified the criteria guiding the selection of publications, including their clinical relevance, methodological quality, recency, and contribution to the practical management of drug-associated vitamin B12 deficiency. Furthermore, we now explicitly acknowledge that no formal risk-of-bias assessment was performed, in keeping with the narrative design of the review. These revisions improve the transparency of the evidence-identification and selection process and clarify the methodological scope and limitations of the review.

Answer: We thank the Reviewer for this insightful comment. We have revised the physiology section to strengthen its biochemical and metabolic focus and improve the precision of several statements. In particular, we have removed the specific estimate suggesting that vitamin B12 deficiency remains paucisymptomatic for 3–5 years, replacing it with a more cautious description of the role of hepatic stores in delaying clinical manifestations. We have also expanded the discussion of the two principal intracellular coenzyme forms of cobalamin, methylcobalamin and adenosylcobalamin, and their respective roles in one-carbon metabolism and mitochondrial metabolism, linking these pathways to the accumulation of homocysteine and methylmalonic acid as functional biomarkers of deficiency. In addition, we have revised the description of the absorption pathway, including the role of intrinsic factor, ileal uptake, and transport to peripheral tissues. We acknowledge the importance of distinguishing dietary cobalamin from microbial corrinoids and circulating vitamin B12 measurements from intracellular functional status, and have considered these distinctions in refining the section. Finally, we have reviewed the description and attribution of Figure 1 to ensure that its origin and relationship to the cited sources are clearly stated. These changes aim to provide a more mechanistically grounded account of cobalamin physiology and better align the section with the metabolic focus of the review.

Answer: We thank the reviewer for this valuable and constructive comment. We agree that several potential risk factors mentioned in the original manuscript required more specific supporting evidence and that the distinction between reductions in circulating vitamin B12 concentrations and clinically relevant deficiency following prolonged metformin exposure needed to be clarified.

Accordingly, we have revised this section to better differentiate proposed mechanisms according to the strength and directness of the supporting evidence. We have also added specific references and effect estimates for risk factors supported by the available literature, while removing or qualifying associations that are not consistently supported by adjusted analyses.

Furthermore, we have clarified the complementary contributions of randomized controlled trials, longitudinal cohort studies, and cross-sectional investigations, taking into account differences in baseline vitamin B12 status, metformin dose and duration, definitions of deficiency, and potential confounding. The revised text distinguishes reductions in serum vitamin B12 concentrations from biochemical deficiency and clinically significant neurological or hematological manifestations. Although biochemical changes may occur relatively early during metformin treatment, the timing and clinical significance of these changes vary among individuals and depend on baseline vitamin B12 status, cumulative exposure, and other patient-related risk factors.

Finally, we rechecked the findings reported in the Chapman and Yang meta-analyses against the original publications and revised the manuscript to ensure that the corresponding effect estimates and 95% confidence intervals are accurately reported and interpreted in the appropriate clinical context.

Answer: We thank the reviewer for this important and constructive comment. We agree that the previous version of the manuscript did not sufficiently distinguish the observational evidence linking proton pump inhibitor (PPI) use to vitamin B12 deficiency from evidence establishing causality. We also acknowledge that some statements regarding dose- and duration-related effects, potential additive effects of concomitant PPI and metformin treatment, and the strength of the available evidence required greater qualification.

In response, we have substantially revised Section 4.2 and Table 1. First, we have removed the duplicated description of the study by Lam et al. and retained its principal findings while explicitly emphasizing that the observed associations do not establish causality. We have also revised the discussion of the proposed pathophysiological mechanisms to distinguish impaired absorption of food-bound vitamin B12 from low circulating vitamin B12 concentrations, biochemically confirmed deficiency, and clinically significant neurological or hematological manifestations.

Second, we have reassessed the epidemiological estimates and risk-related factors discussed in this section. We have clarified the limitations of interpreting studies that use serum vitamin B12 thresholds of 300 pg/mL, as such thresholds may include individuals with borderline concentrations and do not necessarily indicate confirmed deficiency. We have also reviewed the interpretation of the reported association between PPI exposure, treatment duration, and dose. Statements implying universal clinical thresholds for treatment duration or daily dose have been removed or qualified where the evidence does not support their general applicability.

Third, we have expanded the discussion of potential confounding factors, explicitly addressing age, comorbidities, dietary vitamin B12 intake, the indication for PPI treatment, polypharmacy, atrophic gastritis and other gastrointestinal conditions, as well as differences in healthcare utilization that may influence testing and diagnosis. We now emphasize that, although these factors may be addressed in statistical analyses, residual confounding cannot be excluded and may affect the interpretation of observational associations.

Finally, we have revised the discussion of concomitant PPI and metformin exposure. Although these medications may affect vitamin B12 status through different proposed mechanisms, the existence of distinct mechanisms does not in itself demonstrate an additive or synergistic effect. We have therefore qualified our previous statements and clarified that the magnitude of any additional risk remains uncertain in the absence of direct evidence from studies evaluating combined exposure and treatment interactions. Table 1 has also been revised to avoid unsupported qualitative grading of the evidence and to distinguish proposed mechanisms, observational associations, biochemical findings, and clinically relevant outcomes.

We believe that these revisions provide a more balanced and methodologically cautious interpretation of the available evidence, address the concerns raised by the reviewer, and improve the reliability of the comparison between metformin- and PPI-associated vitamin B12 deficiency.

Answer: We thank the reviewer for this important observation. We agree that the previous version of this section did not sufficiently distinguish between the mechanisms involved, the strength of the available evidence, and the degree to which the reported associations support a diagnosis of vitamin B12 deficiency.

We have substantially revised Section 4.3 to adopt a more cautious and evidence-based approach. Specifically, we have replaced causal language regarding H2 receptor antagonists with wording that reflects the observational nature of the available evidence. We have also clarified the proposed mechanisms and clinical relevance of the other drugs discussed, distinguishing between potential effects on vitamin absorption or serum concentrations and functional inactivation of vitamin B12, as occurs with nitrous oxide exposure.

Regarding GLP-1 receptor agonists, we acknowledge that the evidence specifically linking these agents to vitamin B12 deficiency remains limited. We have therefore removed the previously reported annual incidence of 2.6%, which was not sufficiently supported by a clearly characterized primary study, and revised the text to emphasize the uncertainty surrounding the magnitude and clinical significance of any potential association.

In addition, we have reviewed the relevant references to ensure that the citations correspond to the statements they support. Figure 2 has also been reviewed to ensure consistency with the revised text and to avoid implying equivalent evidence or established causality across drugs with different mechanisms and levels of supporting evidence (for example colestiramine).

Finally, we have clarified the distinction between drugs supported by more substantial clinical evidence and other agents for which the evidence remains limited or context-dependent. The review's central conclusions are based primarily on the better-characterized associations involving metformin and proton pump inhibitors, while the other drugs are discussed as additional or context-specific considerations rather than as equivalent contributors to vitamin B12 deficiency.

We believe these changes improve the scientific accuracy, balance, and clinical relevance of this section and address the reviewer's concerns.

Answer: We thank the reviewer for this insightful comment. We agree that the clinical manifestations of established vitamin B12 deficiency should be distinguished from the potential consequences of more modest reductions in serum vitamin B12 associated with drug exposure. We have revised the clinical section to clarify this distinction and to emphasize that a reduction in serum vitamin B12 does not necessarily indicate functional deficiency or imply an equivalent risk of clinical complications.

We have also expanded the discussion of the relationship between metformin-associated vitamin B12 deficiency and peripheral neuropathy. The revised text acknowledges the heterogeneity of the available evidence and further discusses relevant methodological limitations, including diabetes duration, cumulative glycemic exposure, renal function, differences in diagnostic criteria and ascertainment methods, and the limitations of cross-sectional observational designs. We have emphasized that the available evidence does not establish a consistent causal relationship between metformin-associated vitamin B12 reductions and diabetic neuropathy, while recognizing that clinically significant vitamin B12 deficiency may contribute to neurological symptoms in selected patients.

Furthermore, we have refined the discussion of vitamin B12 assessment and monitoring to emphasize an individualized approach based on treatment exposure, additional risk factors and compatible clinical manifestations, rather than implying that a universal monitoring interval or a clinical benefit from systematic screening has been established. We have also clarified that metformin should not routinely be discontinued solely because of concern about vitamin B12 deficiency and that confirmed deficiency should be appropriately treated.

Finally, we have reviewed Figure 3 and its supporting references to ensure that the cited evidence corresponds to the statements presented and that the figure is consistent with the revised text.

We believe these changes provide a more balanced interpretation of the clinical evidence and more accurately distinguish established complications of vitamin B12 deficiency from outcomes that remain uncertain in patients receiving drug therapy.

Answer: We thank the reviewer for this important comment. We agree that the previous version of the diagnostic section presented several biomarker thresholds too categorically and did not sufficiently emphasize the limitations associated with their interpretation.

We have revised Sections 6.1–6.3 to clarify that serum vitamin B12 thresholds and indeterminate ranges are dependent on the analytical method, population and clinical context, rather than representing universally applicable diagnostic boundaries. We have also removed the categorical implication that serum concentrations above 287 pmol/L generally exclude the need for further investigation, emphasizing instead the importance of clinical suspicion, predisposing conditions and complementary biochemical assessment.

Furthermore, we have revised the discussion of functional biomarkers to avoid presenting methylmalonic acid (MMA) as an infallible or universally superior diagnostic marker. The revised text explicitly addresses the influence of renal impairment on both MMA and homocysteine concentrations and acknowledges the additional effects of folate and vitamin B6 status, age, pregnancy, liver disease and other relevant clinical and analytical factors.

Finally, we are revising Table 2 to provide a more clinically informative account of the principal biomarkers, their potential applications and their limitations, avoiding unsupported qualitative classifications of diagnostic performance. Where quantitative diagnostic estimates are reported, these should be supported by appropriate references and interpreted in relation to the population and diagnostic reference standard used.

We believe these changes provide a more balanced and clinically appropriate presentation of vitamin B12 deficiency assessment and reduce the risk of misclassification arising from isolated biomarker thresholds.

Answer: We thank the reviewer for highlighting the inconsistencies in the monitoring recommendations and the need to clarify the methodological status of the proposed algorithm.

We have revised Section 6.3 to distinguish baseline risk assessment from subsequent monitoring during metformin treatment. The revised text emphasizes that baseline vitamin B12 measurement may be appropriate in patients with predisposing factors or compatible clinical findings, while acknowledging that universal testing before treatment initiation has not been established as a requirement for every patient.

We have also reconciled the previous references to periodic and annual testing. The revised section clarifies that current guideline recommendations support consideration of periodic vitamin B12 assessment, particularly in patients with anemia, peripheral neuropathy or additional risk factors, without establishing a universally applicable monitoring interval. We have therefore qualified the statements concerning annual testing and distinguished proposals made in some reviews from formal guideline recommendations. In addition, we have clarified that the observational findings on testing practices identify a potential gap in clinical implementation but do not establish the optimal monitoring frequency or demonstrate the benefits of a particular screening schedule.

Regarding Figure 4, we have corrected the source attribution in the figure legend and reviewed the consistency of the relevant citations. We have also revised Section 6.4 to clarify that the diagnostic–therapeutic algorithm is an author-developed conceptual proposal informed by the literature reviewed. We acknowledge that it was not developed through a formal consensus process or structured evidence-to-decision methodology and has not undergone external validation. Accordingly, it is presented as a pragmatic framework to support individualized clinical reasoning rather than as a validated diagnostic instrument or a formal guideline recommendation.

We believe these revisions improve the transparency, consistency and clinical interpretability of the manuscript while preserving the practical value of the proposed framework without overstating the available evidence.

Answer: We thank the reviewer for this important comment regarding the clinical applicability and strength of the treatment recommendations. We agree that the original version of this section did not sufficiently distinguish between evidence supporting biochemical correction and evidence demonstrating clinical benefit, particularly in patients with neurological manifestations.

We have substantially revised Section 7 to ensure that therapeutic recommendations are proportionate to the available evidence and appropriately qualified according to the underlying etiology, clinical severity, neurological involvement, treatment adherence, and applicable national guidelines.

First, we have revised the discussion of oral vitamin B12 replacement to clarify that high-dose oral treatment is an effective option for correcting biochemical deficiency in selected patients, but that the available comparative evidence does not establish equivalent neurological recovery or long-term clinical outcomes compared with intramuscular treatment. We have also moderated statements regarding the suitability of oral treatment in patients with significant malabsorption or severe clinical manifestations.

Second, we have revised the discussion of intramuscular treatment to avoid presenting a single loading and maintenance schedule as universally applicable. The revised text acknowledges that dosing regimens vary according to the cobalamin formulation, national clinical recommendations, the cause and severity of deficiency, and the presence of neurological or hematological manifestations. We have also clarified that cyanocobalamin and hydroxocobalamin should not be assumed to share identical dosing intervals.

Third, we have removed the suggestion that methylcobalamin has established superiority for neuropathic symptoms in medication-associated vitamin B12 deficiency. The revised text emphasizes that the available evidence does not support a definitive recommendation favoring one cobalamin formulation over another on the basis of superior neurological outcomes in this setting.

Fourth, we have clarified the distinction between initial treatment, long-term replacement for persistent or irreversible causes, and preventive supplementation in patients receiving metformin and/or proton pump inhibitors. We have removed the implication that a universal oral maintenance schedule is appropriate and emphasized that treatment duration, dose, and route should be individualized.

Finally, we have refined the discussion of response monitoring to prioritize clinical evolution and appropriate hematological assessment, with additional biochemical or functional markers used according to the initial diagnostic findings and subsequent clinical course. We have also emphasized the need for prompt treatment when clinically significant neurological manifestations are suspected and for reassessment when the response is inadequate.

We believe that these revisions improve the clinical accuracy and balance of the treatment section, distinguish established recommendations from areas of uncertainty, and address the reviewer's concerns regarding the strength and applicability of the therapeutic guidance.

Answer: We thank the reviewer for this comprehensive and important assessment. We agree that the conclusions should more clearly reflect the limitations of the available evidence and that the reliability and transparency of the reference system are essential to the scientific validity of this narrative review.

First, we have revised the Conclusions section to reduce the strength of statements that could be interpreted as establishing causality or demonstrating clinical benefits that have not been adequately confirmed. In particular, we now clarify that metformin and proton pump inhibitors have been associated with impaired vitamin B12 status through distinct proposed mechanisms, while the potential additive effects of combined exposure remain uncertain. We have also moderated our recommendations regarding proactive detection, explicitly acknowledging that the available evidence does not establish the benefit of universal screening and supporting an individualized, risk-based approach.

Second, we have clarified the distinction between biochemical correction and clinically meaningful outcomes. The revised conclusions acknowledge that vitamin B12 replacement can correct biochemical deficiency, whereas evidence concerning comparative neurological recovery and the clinical benefits of systematic screening remains limited. Treatment recommendations are therefore framed in terms of individualized clinical assessment and applicable guidelines rather than as universal regimens.

Third, we have revised the conclusions to emphasize the uncertainties that remain regarding risk stratification, screening strategies, and patient-important outcomes. This approach preserves the practical relevance of the review while ensuring that its recommendations are proportionate to the evidence discussed.

Finally, we are addressing these issues through a systematic cross-check of the in-text citations against the reference list, verification of the relevant bibliographic records, and correction of citation formatting and numbering throughout the manuscript. We are also reviewing the support for central clinical claims to ensure that primary studies, systematic reviews, and relevant clinical guidelines are cited appropriately

Reviewer 2 Report

Comments and Suggestions for Authors

Reviewer Comments to the Authors

The manuscript addresses an important and clinically relevant topic. Drug-associated vitamin B12 deficiency is frequently encountered in clinical practice, particularly in patients receiving long-term metformin or proton pump inhibitors, and bringng the available evidence together in a practical review could be useful for clinicians.I found the manuscript interesting and generally well structured. The sections on metformin, PPIs, diagnostic biomarkers, and treatment provide a useful framework for the reader. Several aspects would benefit from further clarification and revision before the manuscript is ready for publication. My main concerns relate to the literature-search methodology, accuracy and consistency of the references, interpretation of some clinical evidence, and the recommendations regarding vitamin B12 replacement therapy.

Major Comments

Scope and relevance to Metabolites

The review is currently more clinically oriented than metabolically focused. Given the scope of Metabolites, I think the manuscript would benefit from a somewhat stronger discussion of the metabolic aspects of vitamin B12 deficiency. In particular, the roles of methylmalonic acid, homocysteine, holotranscobalamin, functional vitamin B12 deficiency, and one-carbon metabolism could be discussed in greater depth. It would also be useful to mention the influence of renal function on the interpretation of MMA and related biomarkers.

This addition would, in my view, strengthen the scientific focus of the review and improve its fit with the journal.

Lines 72–84

The authors provide the databases and general search terms used, which is helpful. However, a few additional methodological details would make the review more transparent and reproducible.

Please consider specifying the period covered by the search, the date of the final literature search, and clearer inclusion and exclusion criteria. It would also be helpful to explain briefly how the authors assessed the relevance and methodological quality of the selected publications.

References and citation consistency

I noticed thst substantial problem with the reference number from approximately Reference 17 onward. Several citations in the text do not seem to correspond to the sources listed under the same numbers.

This appears to affect, among others, the sections dealing with phenytoin, hormonal contraceptives, nitrous oxide, GLP-1 receptor agonists, Gao et al., Farooq et al., ADA recommendations, Chappell et al., and some of the treatment studies.

I would strongly suggest checking the reference list systematically and then verifying every in-text citation after the numbering has been corrected. Figure 4 also cites Reference 45, although I could not identify this reference in the bibliography, while Reference 43 appears not to be cited in the main text.

Lines 93–97

The statement that hepatic tissue stores approximately 3–10 mg of vitamin B12 deserves another check. My understanding is that commonly cited values generally with only a proportion located in the liver.

Please verify this value against an authoritative source and revise the wording if necessary.

Lines 132–195

This is an important and generally informative section. However, some of the listed risk factors for vitamin B12 deficiency—particularly sex, smoking, obesity, dyslipidemia, and coronary artery disease—would benefit from specific supporting references. I would also recommend the authors to distinguish more clearly between an early biochemical decrease in serum vitamin B12 and the development of clinically relevant deficiency after prolonged metformin exposure. This distinction may help reconcile the relatively early changes reported in some intervention studies with the longer treatment durations associated with clinically significant deficiency.

The results of the Chapman and Yang meta-analyses should also be checked against the original publications and reported with the relevant effect estimates and confidence intervals.

Lines 196–241

The discussion of PPIs is useful, but the description of the study by Lam et al. should be revised. This was a case-control study rather than a cohort study. It would be preferable to describe the study design and exposure categories exactly as reported in the original publication.

Table 1

Table 1 is useful for comparing metformin and PPIs, but I have two concerns.

First, the evidence categories “High” and “Moderate to high” are provided without explaining how the evidence was graded. It may be preferable either to remove these categories or to use a clearly defined grading system.

Second, the table lists the usual onset for metformin-associated changes as “months,” whereas elsewhere the manuscript emphasizes several years of exposure. These statements are not necessarily contradictory, but the distinction between an early biochemical reduction in B12 and later clinically significant deficiency should be made clearer.

The statement that metformin has a “limited impact” on free vitamin B12 from supplements would also benefit from further explanation or supporting evidence.

Lines 245–275

For H2 receptor antagonists, I suggest using the term “association” rather than “causal relationship,” unless stronger causal evidence can be provided.The statements concerning colchicine and cholestyramine would benefit from references. The section on nitrous oxide could also be updated, particularly in view of contemporary evidence concerning recreational exposure.

The discussion of GLP-1 receptor agonists is interesting and timely. However, given that the evidence regarding vitamin B12 deficiency in this context is still developing, the reported incidence should be presented somewhat cautiously and the limitations of the available evidence should be acknowledged.

Section 5

A few details in this section should be checked for consistency.

The manuscript defines macrocytosis using an MCV threshold >98 fL, whereas >100 fL is more commonly used. Please consider using a consistent definition throughout the manuscript.

The definition of neutrophil hypersegmentation should also be verified against standard hematological criteria.

I would additionally suggest checking the numerical findings attributed to Gao et al. and Farooq et al. against the original articles. The statement that up to 20% of patients with B12 deficiency may have no hematological abnormalities would benefit from a specific supporting reference.

Finally, I suggest replacing the term “impotence” with the more appropriate clinical term “erectile dysfunction.”

Lines 427–434

Please verify the stated adequate intakes of 4.5 and 5 μg/day against the cited source.

Since metformin is a major focus of the review, a brief discussion of its use during pregnancy, gestational diabetes might also add value to this section, particularly in relation to maternal and fetal vitamin B12 status.

Section 6

This section is particularly relevant to Metabolites and could perhaps be developed further.

I recommend rechecking the proposed cut-offs for total vitamin B12, MMA, homocysteine, and holotranscobalamin against the original sources. Since these thresholds may vary according to the analytical method and population, this limitation should be made clear to the reader.

The Danish immunoassay data should also be checked to ensure that the reported reference intervals correspond to the correct analyte.

It would additionally be useful to emphasize that renal impairment can increase MMA independently of vitamin B12 deficiency and therefore complicate interpretation.

Lines 504–529

The wording “monitoring serum cyanocobalamin levels” should be reconsidered. Routine laboratory testing generally measures total serum vitamin B12/cobalamin rather than specifically cyanocobalamin.

Lines 530–536

The proposed clinical algorithm is potentially useful. However, the caption states that the figure was adapted from References 45 and 35, while Reference 45 does not appear to be included in the reference list.

Please check this citation.

Section 7

I believe this section deserves particularly careful revision because it provides recommendations that readers may apply in clinical practice.

The evidence supporting the proposed oral maintenance schedules should be clarified. In particular, the manuscript recommends an initial period of daily high-dose oral vitamin B12 followed by once-weekly maintenance, including a specific weekly maintenance regimen for patients receiving metformin. I am not convinced that the evidence presented in the manuscript is sufficient to support this schedule as a general recommendation.

Minor Comments

There are also several relatively straightforward editorial issues that can be addressed during revision: “Vitamin b12” should be changed to “Vitamin B12”; “Vitamina B12” in Table 2 should be corrected; “intrinsic factors” should be changed to “intrinsic factor” where appropriate; and “impotence” should preferably be replaced by “erectile dysfunction.” Units such as mcg/µg and pg/mL/pmol/L should be standardized throughout. The citation style should also be checked carefully, as several stray numbers and inconsistent citation formats are visible in the manuscript.

Overall Recommendation: Major Revision

This is a useful review addressing an important clinical problem, and I believe the manuscript has the potential to become a valuable contribution after revision. Most of the concerns raised above appear addressable. However, the reference-list inconsistencies, several clinical and numerical issues, and particularly the evidence supporting the proposed vitamin B12 maintenance regimen should be resolved before the manuscript is considered further.

Author Response

Reviewer 2

Scope and relevance to Metabolites

This addition would, in my view, strengthen the scientific focus of the review and improve its fit with the journal.

We thank the reviewer for this valuable suggestion. We agree that the original version was primarily focused on the clinical implications of drug-induced vitamin B12 deficiency. To improve the fit of the manuscript with the scope of Metabolites, we have expanded the discussion of the metabolic aspects of vitamin B12 deficiency.

Specifically, we have incorporated additional information regarding the role of vitamin B12 in one-carbon metabolism and methylmalonyl-CoA metabolism, the pathophysiological basis of functional vitamin B12 deficiency, and the clinical interpretation of metabolic biomarkers including methylmalonic acid (MMA), homocysteine, and holotranscobalamin. We have also added a discussion on the influence of renal dysfunction on MMA concentrations and the limitations this may introduce when interpreting biochemical evidence of vitamin B12 deficiency.

These additions have been incorporated into Sections 3, 5 and 6 and we believe they substantially strengthen the mechanistic and metabolic perspective of the review.

Lines 72–84

The authors provide the databases and general search terms used, which is helpful. However, a few additional methodological details would make the review more transparent and reproducible.

Thanks for the comment. Although this manuscript was designed as a narrative review rather than a systematic review, we agree that additional methodological information improves transparency and reproducibility. Accordingly, we have expanded the Methods section to specify the databases searched, the time period covered, the date of the final literature search, the inclusion and exclusion criteria, and the approach used to assess the relevance and methodological quality of the selected publications. These modifications do not alter the scope of the review but provide greater clarity regarding the identification and selection of the supporting literature.

References and citation consistency

I noticed that substantial problem with the reference number from approximately Reference 17 onward. Several citations in the text do not seem to correspond to the sources listed under the same numbers.

We thank the reviewer for identifying these inconsistencies. Following this comment, the entire reference list and all in-text citations were systematically reviewed and corrected. Errors in reference numbering were identified and resolved, particularly in the section beginning around Reference 17, where a formatting issue led to misalignment between several references and their corresponding citations.

All in-text citations were subsequently rechecked against the corrected bibliography, including those related to phenytoin, hormonal contraceptives, nitrous oxide, GLP-1 receptor agonists, Gao et al., Farooq et al., ADA recommendations, Chappell et al., and treatment studies. In addition, all figure citations were verified and corrected where necessary. The incorrect citation to Reference 45 in Figure 4 has been amended, and all references have now been cross-checked to ensure consistency throughout the manuscript.

We are grateful to the reviewer for bringing this issue to our attention.

Lines 93–97

Please verify this value against an authoritative source and revise the wording if necessary.

We thank the reviewer for this observation. We have re-evaluated the statement regarding hepatic vitamin B12 stores and agree that the original wording could incorrectly imply that the entire estimated amount corresponded specifically to hepatic stores. We have therefore removed the 3–10 mg estimate and revised the text to distinguish more clearly between body vitamin B12 reserves and the liver as a major storage site. The corresponding reference has also been revised.

Lines 132–195

The results of the Chapman and Yang meta-analyses should also be checked against the original publications and reported with the relevant effect estimates and confidence intervals.

We thank the reviewer for this valuable comment. We agree that several potential risk factors mentioned in the original manuscript required more specific supporting evidence and that the distinction between an early biochemical reduction in serum vitamin B12 and clinically relevant vitamin B12 deficiency after prolonged metformin exposure should be made clearer.

Accordingly, we have revised this section by adding specific references to the risk factors for which sufficient evidence is available and by removing or qualifying associations that are not consistently supported. We have also clarified that reductions in serum vitamin B12 may occur relatively early after initiation of metformin therapy, whereas the development of clinically relevant or symptomatic deficiency is more commonly associated with longer exposure and depends on additional patient-related risk factors and baseline vitamin B12 status.

Finally, we rechecked the Chapman and Yang meta-analyses against the original publications and revised the manuscript to report the corresponding effect estimates and 95% confidence intervals.

Lines 196–241

We thank the reviewer for identifying this inaccuracy. We have corrected the description of the study by Lam et al. to accurately identify it as a nested case-control study rather than a cohort study. We have also revised the text to report the exposure categories as defined in the original publication, including ≥2 years of PPI or H2-receptor antagonist exposure, the corresponding reference group, and the reported dose categories. The associated odds ratios and 95% confidence intervals have also been included. Given the observational case-control design, we have retained appropriate caution regarding causal interpretation.

Table 1

Table 1 is useful for comparing metformin and PPIs, but I have two concerns.

The statement that metformin has a “limited impact” on free vitamin B12 from supplements would also benefit from further explanation or supporting evidence.

We thank the reviewer for these important observations. The qualitative evidence categories (“High” and “Moderate to high”) have been removed from Table 1, as no formal evidence-grading framework was applied in this narrative review. We have also revised the description of the time course associated with metformin to distinguish between the early biochemical reduction in serum vitamin B12, which has been observed within weeks to months, and clinically relevant deficiency associated with prolonged exposure. Finally, the statement that metformin has a “limited impact” on free vitamin B12 from supplements has been removed because the available evidence more directly supports an effect on vitamin B12 absorption than a specific conclusion regarding the absorption of free supplemental vitamin B12.

Lines 245–275

We thank the reviewer for these important observations. We have revised the terminology used for H2 receptor antagonists and now refer to an association, rather than a causal relationship, between their use and vitamin B12 deficiency.

The statements concerning colchicine and cholestyramine have been removed because these drugs are peripheral to the main scope of the review and we considered it preferable not to retain claims without sufficiently specific supporting references.

The section on nitrous oxide has been updated to better reflect contemporary evidence regarding recreational exposure. The revised text now emphasizes the concept of functional vitamin B12 deficiency, highlights the neurological manifestations associated with repeated recreational use, and clarifies that serum vitamin B12 concentrations may not always accurately reflect intracellular vitamin B12 status, whereas MMA and homocysteine may provide additional diagnostic information.

Finally, the section on GLP-1 receptor agonists has been revised to reflect the emerging and predominantly observational nature of the available evidence. The previously reported incidence estimate has been removed because it did not specifically establish the incidence of vitamin B12 deficiency. The revised text now acknowledges that evidence specifically linking GLP-1 receptor agonist therapy to vitamin B12 deficiency remains limited and that neither causality nor the magnitude of any specific risk can currently be established.

Section 5

A few details in this section should be checked for consistency.

The manuscript defines macrocytosis using an MCV threshold >98 fL, whereas >100 fL is more commonly used. Please consider using a consistent definition throughout the manuscript.

The definition of neutrophil hypersegmentation should also be verified against standard hematological criteria.

Finally, I suggest replacing the term “impotence” with the more appropriate clinical term “erectile dysfunction.”

We thank the reviewer for these careful observations. Section 5 has been revised for consistency. Macrocytosis is now defined using an MCV threshold >100 fL throughout the manuscript, and the definition of neutrophil hypersegmentation has been revised according to standard hematological criteria. The numerical findings attributed to Gao et al. and Farooq et al. were rechecked against the original publications and corrected where necessary. In addition, the unsupported numerical statement regarding the proportion of patients without hematological abnormalities has been removed and replaced with a more conservative statement acknowledging that vitamin B12 deficiency may occur without anemia or macrocytosis. Finally, the term “impotence” has been replaced with “erectile dysfunction.”

Lines 427–434

Please verify the stated adequate intakes of 4.5 and 5 μg/day against the cited source.

We thank the reviewer for this helpful comment. We verified the vitamin B12 Adequate Intake values against the cited EFSA source. EFSA proposes an AI of 4.5 μg/day during pregnancy and 5 μg/day during lactation; the terminology in the manuscript has been revised accordingly [36]. We have also added a brief discussion of metformin use during pregnancy and gestational diabetes, with particular consideration of maternal vitamin B12 status, one-carbon metabolism and fetal exposure to metformin. A specific reference addressing the interaction between metformin, folate and vitamin B12 during diabetic pregnancy has been added [37].

Section 6

This section is particularly relevant to Metabolites and could perhaps be developed further.

The Danish immunoassay data should also be checked to ensure that the reported reference intervals correspond to the correct analyte.

It would additionally be useful to emphasize that renal impairment can increase MMA independently of vitamin B12 deficiency and therefore complicate interpretation.

We thank the reviewer for these valuable suggestions. We have revised Section 6 to clarify that the proposed cut-offs for total vitamin B12, methylmalonic acid (MMA), homocysteine and holotranscobalamin may vary according to the analytical method, laboratory, and population studied, and therefore should be interpreted within the appropriate clinical context. We also rechecked the cited Danish immunoassay data to ensure that the reported reference intervals correspond to the correct analyte. Finally, we have expanded the discussion of MMA to emphasize that renal impairment may increase MMA concentrations independently of vitamin B12 deficiency and may therefore complicate diagnostic interpretation.

Lines 504–529

We thank the reviewer for this observation. We agree that routine laboratory testing measures total serum vitamin B12 (cobalamin) rather than specifically cyanocobalamin. The terminology has therefore been corrected throughout the manuscript to refer to serum vitamin B12 levels.

Lines 530–536

Please check this citation.

We thank the Editor for identifying this issue. The figure legend has been revised and the reference numbering has been corrected. The previous citation to reference 45 resulted from an earlier version of the manuscript. The legend now correctly cites references 39 and 40, which correspond to the sources used for the development of the algorithm.

Section 7

I believe this section deserves particularly careful revision because it provides recommendations that readers may apply in clinical practice.

We thank the reviewer for this important observation. We agree that the available evidence does not support a universal once-weekly oral maintenance schedule, including a specific weekly regimen for patients receiving metformin. These prescriptive recommendations have therefore been removed. The revised section now distinguishes the use of high-dose daily oral vitamin B12 for initial correction from the more limited evidence regarding long-term maintenance and clarifies that maintenance treatment should be individualized according to the etiology and severity of deficiency, clinical response, biomarker evolution and persistence of underlying risk factors.

Minor Comments

We thank the reviewer for these careful editorial observations. The manuscript has been revised throughout to ensure consistent terminology and formatting. Specifically, “Vitamin b12” has been corrected to “Vitamin B12”, “Vitamina B12” in Table 2 has been corrected, “intrinsic factors” has been changed to “intrinsic factor” where appropriate, and “impotence” has been replaced by “erectile dysfunction”. In addition, units and laboratory reporting formats have been standardized throughout the manuscript, and the citation format and reference numbering have been carefully reviewed and corrected to eliminate inconsistencies and stray citations.

Reviewer 3 Report

Comments and Suggestions for Authors

This narrative review covers mechanisms, diagnosis, monitoring and treatment of B12 deficiency associated with metformin, PPIs and other drugs. The topic is clinically relevant and the writing is generally clear. However, several issues need to be addressed before it can be published.

Major comments

Reference integrity. Ref 17 contains only authors; its title appears as ref 18, so from entry 18 onward, in-text citation [n] corresponds to list entry n+1 (e.g., [20] phenytoin = entry 21; [29] Gao = entry 30; [41] Cochrane = entry 42). Further errors not explained by this shift: [34] (L511, L512, L540) matches no entry; [36] at L525 (13.5% monitoring audit) corresponds to Chappell (entry 35), whereas other uses of [36] correspond to Pratama (entry 37); Figure 4 cites a nonexistent ref 45; Figure 3 cites refs 20–22, none of which addresses metformin and anemia; the ADA/EASD consensus is not in the reference list; All citations should be re-verified against the claims they support.
Missing key evidence. The long-term placebo-controlled HOME trial (de Jager et al., BMJ 2010) and the DPP/DPPOS data, both cited in the ADA Standards, are not discussed, whereas the small single-centre study by Kiran et al. is described in detail (L169–182). Please add these trials and more recent evidence. A summary table of key studies (design, n, population, B12 cutoff, outcome) is recommended.
Functional deficiency and renal function. Please discuss whether low serum B12 in metformin users reflects true tissue deficiency, including the MMA and homocysteine data from the HOME trial and alternative explanations such as altered B12 distribution. L468–469 notes that MMA rises in renal impairment, but homocysteine is also renally dependent; this caveat is absent from Table 2, and its implications for diabetic patients with CKD, in whom both markers may be falsely elevated, should be discussed.
Figure 4. The algorithm addresses metformin only, does not specify how "true or functional deficiency" is established (thresholds, grey zone, MMA), and omits PPI deprescribing despite this being recommended in the Conclusions (L717). The "≥2 risk factors" threshold for pre-treatment testing and the "every 2–3 years" monitoring interval are not supported by the text or references.

Author Response

REVIEWER 3

1.- Reference integrity. Ref 17 contains only authors; its title appears as ref 18, so from entry 18 onward, in-text citation [n] corresponds to list entry n+1 (e.g., [20] phenytoin = entry 21; [29] Gao = entry 30; [41] Cochrane = entry 42). Further errors not explained by this shift: [34] (L511, L512, L540) matches no entry; [36] at L525 (13.5% monitoring audit) corresponds to Chappell (entry 35), whereas other uses of [36] correspond to Pratama (entry 37); Figure 4 cites a nonexistent ref 45; Figure 3 cites refs 20–22, none of which addresses metformin and anemia; the ADA/EASD consensus is not in the reference list; All citations should be re-verified against the claims they support.

Answer: We thank the reviewer for carefully identifying the inconsistencies in the reference list and in-text citations. We have thoroughly reviewed and verified the references throughout the manuscript, corrected the numbering and bibliographic inconsistencies, and checked that each citation appropriately supports the corresponding statement. We have also reviewed the references associated with Figures 3 and 4, corrected the relevant citations, and verified the inclusion and attribution of the supporting sources, including the applicable diabetes guidelines and consensus documents.

These corrections were made to improve the accuracy, consistency, and traceability of the evidence cited throughout the manuscript. We are grateful to the reviewer for highlighting these issues, which have helped us improve the scientific rigor and editorial quality of the revised version.

2.- Missing key evidence. The long-term placebo-controlled HOME trial (de Jager et al., BMJ 2010) and the DPP/DPPOS data, both cited in the ADA Standards, are not discussed, whereas the small single-centre study by Kiran et al. is described in detail (L169–182). Please add these trials and more recent evidence. A summary table of key studies (design, n, population, B12 cutoff, outcome) is recommended.

Answer: We thank the reviewer for this valuable suggestion. We agree that the previous version gave disproportionate emphasis to the small single-center study by Kiran and Naik while insufficiently discussing the more informative long-term evidence on metformin-associated vitamin B12 deficiency.

We have substantially revised Section 4.1 to incorporate the randomized, placebo-controlled HOME trial by de Jager et al. and the long-term DPP/DPPOS analysis by Aroda et al. The revised section describes their study designs, populations, treatment regimens, vitamin B12 thresholds, and principal findings, including the effects of prolonged metformin exposure on biochemical vitamin B12 deficiency and the reported hematological and neurological outcomes. We have also incorporated a more recent meta-analysis of observational studies while acknowledging the heterogeneity of diagnostic thresholds and the limitations inherent to observational evidence.

In addition, we have reduced the emphasis on the study by Kiran and Naik. The revised discussion explicitly distinguishes reductions in serum vitamin B12 concentrations from confirmed deficiency and from clinically established hematological or neurological complications. This approach provides a more balanced assessment of the evidence and avoids overstating causal relationships between biochemical changes and clinical outcomes.

We are grateful to the reviewer for this recommendation, which has helped us strengthen the evidence base and improve the clinical relevance of this section.

3.- Functional deficiency and renal function. Please discuss whether low serum B12 in metformin users reflects true tissue deficiency, including the MMA and homocysteine data from the HOME trial and alternative explanations such as altered B12 distribution. L468–469 notes that MMA rises in renal impairment, but homocysteine is also renally dependent; this caveat is absent from Table 2, and its implications for diabetic patients with CKD, in whom both markers may be falsely elevated, should be discussed.

Answer: We thank the reviewer for highlighting the need to distinguish reductions in circulating vitamin B12 from evidence of functional deficiency and to address the limitations of metabolic biomarkers in patients with impaired renal function.

We have revised Section 6.2 to incorporate evidence from the HOME trial and its subsequent post hoc analysis. The original randomized trial demonstrated a reduction in serum vitamin B12 concentrations with long-term metformin treatment, whereas the increase in homocysteine was not statistically significant in the primary analysis. The subsequent analysis reported a significant increase in MMA, providing additional evidence consistent with impaired intracellular vitamin B12-dependent metabolism in at least some patients. We have also clarified that these findings do not establish that every reduction in serum vitamin B12 necessarily represents clinically meaningful tissue deficiency, and we have discussed altered tissue distribution as a proposed, but insufficiently established, alternative explanation.

Furthermore, we have revised Section 6.3 to explicitly state that renal impairment can increase both MMA and homocysteine independently of vitamin B12 status. We now emphasize the importance of considering eGFR and other relevant clinical and biochemical factors when interpreting these markers, particularly in patients with diabetes and CKD, in whom reliance on either marker alone may lead to diagnostic misclassification.

These changes provide a more balanced interpretation of the available evidence and clarify the limitations of functional biomarkers in routine clinical practice.

4.- Figure 4. The algorithm addresses metformin only, does not specify how "true or functional deficiency" is established (thresholds, grey zone, MMA), and omits PPI deprescribing despite this being recommended in the Conclusions (L717). The "≥2 risk factors" threshold for pre-treatment testing and the "every 2–3 years" monitoring interval are not supported by the text or references.

Answer: We thank the reviewer for these important observations regarding Figure 4. We would like to clarify that the algorithm was intentionally designed to address metformin-associated vitamin B12 deficiency, as metformin is the medication supported by the most extensive clinical evidence concerning reductions in circulating vitamin B12 concentrations and the risk of deficiency during long-term treatment. Other medications, including proton pump inhibitors (PPIs), are discussed separately in the manuscript because their evidence base and clinical implications differ.

Regarding the distinction between biochemical and functional deficiency, we acknowledge that this distinction requires an integrated diagnostic approach. Serum vitamin B12 should be interpreted in conjunction with clinical findings and relevant risk factors, with MMA considered when vitamin B12 results are indeterminate and additional clarification is clinically warranted. We also recognize that MMA, and homocysteine when measured, must be interpreted cautiously in patients with impaired renal function.

Concerning PPIs, the intention of Figure 4 is not to provide a comprehensive management algorithm for every medication potentially associated with vitamin B12 deficiency. The decision to discontinue or reduce a PPI should be individualized according to its indication, ongoing clinical benefit and the patient's overall risk profile, rather than being considered an automatic consequence of a low vitamin B12 concentration.

Finally, we acknowledge that the specific thresholds of at least two risk factors and a fixed reassessment interval of every two to three years should not be interpreted as universally established recommendations in the absence of supporting evidence. The clinical approach should instead be guided by individual risk factors, symptoms, previous laboratory findings and the overall clinical context. We have considered these limitations to ensure that the proposed algorithm is interpreted as a pragmatic, evidence-informed framework rather than a validated clinical prediction rule.

Reviewer 4 Report

Comments and Suggestions for Authors

Authors presented a clinically important topic on vit. B12 deficiency and its relation to concomitant pharmacotherapy. However, this review does not add anythyng useful to the current practice. It includes informations relatively well-known to medical students. Therefere, I recommend rejection of the manuscript.

Author Response

Reviewer 4

We thank the reviewer for the time invested in the evaluation of our manuscript and for the candid comments. We acknowledge that several aspects of vitamin B12 physiology and deficiency are well established. However, the primary objective of this review was not to provide a general overview of vitamin B12 deficiency, but rather to integrate and critically discuss the evidence regarding drug-associated vitamin B12 deficiency, particularly in relation to metformin and acid-suppressive therapies, which are highly prevalent in routine clinical practice.

In response to the reviewers’ comments, we have substantially revised the manuscript to strengthen its clinical relevance and practical applicability. The revised version includes a more detailed discussion of the metabolic basis of vitamin B12 deficiency, updated evidence regarding metformin- and PPI-associated deficiency, clarification of the distinction between biochemical alterations and clinically relevant deficiency, a revised interpretation of diagnostic biomarkers and their limitations, and a more cautious treatment section reflecting the strength and limitations of the available evidence.

We hope that these revisions have improved the manuscript and enhanced its value as a focused clinical review on the diagnosis, monitoring and management of medication-associated vitamin B12 deficiency.

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors The authors have comprehensively addressed all major and minor concerns raised in the previous review. Point‑by‑point revisions have been implemented across the manuscript, covering literature search transparency, epidemiological data interpretation, mechanistic description, assessment of evidence strength for observational findings, biomarker diagnostic limitations, monitoring strategies, therapeutic recommendations, figure attributions, and reference formatting. The revised narrative review now properly distinguishes observational associations from causal conclusions, clearly separates biochemical changes from clinically meaningful patient‑relevant outcomes, and avoids over‑interpretation of drug‑nutrient interactions related to metformin and proton‑pump inhibitors. Methodological limitations of narrative synthesis are explicitly stated. Diagnostic cut‑offs, biomarker performance, and confounding factors (including renal function affecting methylmalonic acid and homocysteine) are adequately discussed. The author‑proposed diagnostic‑therapeutic algorithm is appropriately framed as a conceptual pragmatic framework rather than validated guideline recommendation. Treatment suggestions are suitably qualified and individualized, without rigid universal dosing regimens. Potential conflicts of interest and funding sources are fully disclosed. Figures and tables have been updated with corrected citations and necessary caveats. Although Table 2 retains descriptive qualitative terms for biomarker performance, sufficient footnotes and main‑text discussion highlight their limitations and context‑dependent cut‑off values. The overall scientific rigor, balance, and clinical readability of the manuscript have been substantially improved. I recommend acceptance for publication.

Reviewer 2 Report

Comments and Suggestions for Authors

I thank the authors for their detailed responses to my previous comments and for the substantial revisions made to the manuscript. The revised version has improved. Most of my previous concerns have been adequately addressed.

Few minor issues remain:

Line 391 - The authors state in their response that the discussion of colchicine was removed. However, colchicine is still discussed in the revised manuscript.

Reviewer 3 Report

Comments and Suggestions for Authors

No more comments

Reviewer 4 Report

Comments and Suggestions for Authors

The manuscript has beem elongated and includes much more fonts amd is scientifically sound. However, it still in my opinion a mere generic review regarding vitamin B12 deficiency taht does not add anything new to the scientific field.