Diagnosis, Treatment and Long-Term Management of Vitamin B12 Deficiency in Adults: A Delphi Expert Consensus

1. Introduction

Vitamin B12 (cobalamin) deficiency can affect several organs, such as the bone marrow and the peripheral and central nervous systems [1]. The signs and symptoms of deficiency are variable and mostly nonspecific. Patients often seek help in primary medical care.

Dietary sources of vitamin B12 are foods of animal origin. Insufficient dietary intake of vitamin B12 (i.e., less than 4-5 µg/day) can cause vitamin B12 deficiency. In addition, intestinal absorption of vitamin B12 requires the release of vitamins from food proteins, normal secretion and function of intrinsic factor, and appropriate gastrointestinal acidity. Malabsorption of vitamin B12 due to pernicious anemia (lack of intrinsic factor or antibodies against intrinsic factor), atrophic gastritis, or intestinal or gastric resection is a main cause of clinically manifested vitamin B12 deficiency among adults. Pernicious anemia usually manifests as gastrointestinal symptoms (episodes of abdominal distress, distension, nausea, and diarrhea) in addition to vitamin B12 deficiency-related symptoms.

Manifestations related to the hematopoietic system (i.e., megaloblastic anemia) can be diagnosed using widely available laboratory markers. Approximately 30-50% of patients with vitamin B12 deficiency have some degree of neurological involvement [2]. Psychiatric symptoms such as mental and mood disorders and cognitive dysfunction are common in people with B12 deficiency [2,3,4]. A delay in the diagnosis of the neuropsychological manifestations of vitamin B12 deficiency can cause symptoms to become irreversible [5,6,7].

The serum concentration of vitamin B12 is a commonly used marker of vitamin B12 status. The majority of vitamin B12 in blood is bound to haptocorrin, which is not available for B12-dependent enzymatic reactions in cells. Holotranscobalamin (i.e., vitamin B12 bound to transcobalamin) constitutes a small fraction of total serum B12, which is biologically active. Adenosylcobalamin and methylcobalamin deficiency results in increased plasma concentrations of methylmalonic acid and homocysteine, respectively. Circulating vitamin B12 may not be lowered in all patients with B12 deficiency [8]. Using methylmalonic acid and homocysteine to aid in the diagnosis of clinically manifested vitamin B12 deficiency has advantages and limitations [8].

The severity of vitamin B12 deficiency symptoms at first admission affects the choice of treatment modality. For example, initial parenteral treatment has been used when the clinical symptoms are severe [9]. More recent studies have shown that the effectiveness of parenteral and high-dose oral B12 (i.e., 1 mg/d) is similar in correcting serum concentrations of vitamin B12, suggesting that high-dose oral B12 can be used for maintenance treatment [10,11,12].

The UK National Institute for Health and Care Excellence (NICE) released guidelines intended to facilitate the medical care of people with vitamin B12 deficiency specifically in the UK [13]. However, due to uncertainties and lack of evidence on multiple aspects related to the diagnosis and treatment of vitamin B12 deficiency, widely accepted guidelines are still needed due to between-country differences in medical practices, health care systems and generally limited resources. The aim of this study was to establish an internationally acceptable consensus based on a review of the literature published over the past 20 years. First, we summarized consistent results related to aspects of the diagnosis and treatment of vitamin B12 deficiency in the literature. Second, in areas with insufficient evidence, we used a modified Delphi method to establish a practice-oriented expert consensus that can guide the medical care of people with vitamin B12 deficiency and those at risk for deficiency.

2. Materials and Methods

2.1. Methods of the Scoping Review

A literature search was conducted in PubMed on 10 February 2023 according to the search terms shown in Table S1 and by applying language (only English) and time filters (from 01 January 2003). The search was performed according to a detailed protocol (unpublished) that was prepared before starting the study. The articles were selected according to predefined inclusion and exclusion criteria (Table S2). The initial search identified 432 potentially relevant articles (Figure S1). After screening the titles and abstracts, 170 papers qualified for full-text screening. Thereafter, 19 articles were excluded for various reasons (Table S3), and the remaining 151 articles were classified into the following categories according to their primary topics: 1) clinical manifestations of B12 deficiency (neurological and hematological endpoints), 2) risk groups, 3) causes of B12 deficiency, 4) laboratory diagnosis of B12 deficiency, and 5) treatment and long-term management of the deficiency.

Key information was extracted from each publication (whenever relevant) on the study question, characteristics of the participants, the intervention, the biomarkers used to diagnose B12 deficiency, the clinical tests, the outcomes and the key findings. In addition, we defined key open questions raised by each article. The data extraction table was used to 1) identify topics with sufficient between-study agreement and 2) for a stepwise prioritization of open questions to be used in a Delphi survey. The Delphi method was used to determine experts’ consensus on the open questions [14].

2.2. Recruitment of Scientific Board Members and Panelists

The scientific board members were invited by the leading authors (RO and KR) based on medical expertise, practical experience in the field of B12 deficiency, publication records, seniority, and willingness to participate. Nine out of 12 experts accepted the invitation to serve as scientific board members. The remaining 3 experts declined due to conflicts of interest. The scientific board consisted of 1 pharmacist and 9 medical doctors (4 from neurology/neuropsychiatry and 5 from internal medicine). The scientific board was responsible for planning the study, setting and refining the survey questions and discussing and interpreting the results of the surveys. The board met once prior to the first survey round (June 2023) and once after the second survey round (November 2023). To ensure neutrality of the results, the scientific board members were not eligible to participate in the surveys.

The panelists (i.e., the survey participants) were recruited through different channels. First, all corresponding authors of the 151 original articles identified through the literature search were invited to participate. Second, experts were identified through professional networks of board members, in addition to key opinion leaders in the field who had not published original studies in the last 20 years. Forty-six panelists agreed to participate in the survey. The first survey round was completed by 46 panelists, while the second survey was completed by 42 panelists (4 non-responders in the second round). Two online surveys were conducted using the SurveyMonkey Genius® program. For both surveys, the responses of the panelists were anonymous.

2.3. Conduction of the Delphi Survey

A two-round web-based survey was conducted between June 2023 and October 2023. The survey included general questions related to the professional and geographical background of the participants, experience in the vitamin B12 field, and questions related only to the topics of vitamin B12 diagnosis and treatment/management (Table S4). The first survey consisted of questions offering single or multiple answers, including numerical estimates when appropriate. Several open questions were included where the panelists were able to provide their individual input as free text (Table S4).

The second survey aimed to test the consistency of the response of the panelists on the topics that achieved consensus in the first survey (Table S5). In the second survey, the panelists rated each of the questions on a 5-point Likert scale (strongly agree, agree, neutral, disagree, and strongly disagree). We included the option “I do not have the expertise to answer this question” as a possible answer in both rounds of the survey. The data analysis included only the responses of people who considered themselves qualified to answer a specific question.

2.4. Data Analysis and Consensus Definition

After returning the answers of the first survey round, we calculated the mean and 80% confidence intervals (CIs) of the percentage of panelists who agreed with each of the answers. Only answers with a level of agreement of at least 50% (as a lower bound of the 80% CI) were revised and used in the second round of the survey. In the second round of the survey, the upper two categories on a 5-point Likert scale (i.e., the “agree” and “strongly agree” categories) were combined. A question was considered to reach consensus if the combination of the “agree” and “strongly agree” categories constituted at least 50% of the total responses (50% as the lower bound of the 95% CI of the mean response). Questions that did not achieve consensus were considered topics with uncertainty or candidate topics for future research. The mean percentage and the 80% or 95% CIs of the panelists who agreed with the answers of the survey rounds are shown in Tables S4 and S5, respectively.

3. Results

3.1. Vitamin B12 Deficiency in the Medical Literature

Several topics showed consistent results between the studies (Table S6). Macrocytic anemia is not related to the presence or absence of neurological symptoms or their severity [15,16,17,18]. The serum concentration of vitamin B12 is commonly used as a primary marker of vitamin B12 status. This is because 30-40% of people with neurological or hematological symptoms related to B12 deficiency may have normal vitamin B12 concentrations [19,20,21,22]. Serum vitamin B12 concentrations below 148 pmol/L (or 156 pmol/L in certain studies) are commonly considered to suggest a frank deficiency (examples [23,24,25,26,27]). Serum B12 concentrations between 148 and 220 pmol/L or 260 pmol/L are often considered to indicate mild deficiency [28,29,30]. Some studies measured plasma homocysteine concentrations [5,31] or both homocysteine and methylmalonic acid [32,33,34] along with serum B12. For homocysteine and methylmalonic acid concentrations, there is a large variability in the cutoff values and combinations with other markers. Low serum B12 concentrations and elevated methylmalonic acid [35] are not consistently correlated with the presence or severity of polyneuropathy or the clinical response to B12 treatment [36]. Vitamin B12 treatment increases serum B12 concentrations and decreases methylmalonic acid and homocysteine levels [15], but normalization of these markers does not always correspond to clinical improvement. In general, determining the severity of B12 deficiency and the response to treatment primarily relies on evaluating clinical symptoms.

The use of the glucose-lowering drug metformin is causally related to lowered serum concentrations of vitamin B12 [37,38,39,40,41,42,43]. The effect has been shown as early as 3 months after starting the drug [44]. Low serum vitamin B12 concentrations in people using metformin are associated with hyperhomocysteinemia [45], a higher incidence of neuropathy [46,47], and worse neuropathy scores [48]. A recently published longitudinal study showed that people using metformin had a greater risk of neuropathy [HR = 1.84 (95% CI, 1.62, 2.10)] than did those not using metformin [46]. There was a dose‒response association between the daily dose of metformin and a greater risk of neuropathy [46], which could suggest that low vitamin B12 concentrations are a direct cause of neuropathy in those patients. Vitamin B12 supplementation can increase serum B12 concentrations in users of metformin [44,49], but it remains unclear whether it may reduce neuropathy risk.

Vitamin B12 deficiency is common in elderly people [50,51] and is often explained by food-cobalamin malabsorption [52,53]. There is currently no clear evidence that vitamin B12 deficiency has an etiological role in frailty [51] or sarcopenia [54] in elderly people. The presence of some gastrointestinal disorders constitutes a major risk factor for vitamin B12 deficiency. Pernicious anemia often occurs parallel to other autoimmune disorders [24] and in people who are first-degree relatives of patients with this condition [55]. Positive antibodies against intrinsic factor or parietal cell antibodies are found in 30-50% of patients with neurological manifestations related to B12 deficiency [2,56]. Pernicious anemia may be associated with deficiency of multiple nutrients, such as folate and iron [57]. Lowered serum vitamin B12 has been reported in people with atrophic gastritis [58,59], Crohn’s disease [60,61,62], celiac disease [63], infection with H. pylori [64], gastric surgeries [65,66,67,68,69], and chronic treatment with proton pump inhibitors [70]. Up to 70% of patients who undergo gastrectomy may develop vitamin B12 deficiency 12-24 months after surgery if not supplemented [71,72].

In people with vitamin B12 malabsorption, 1000 - 1500 µg/day of oral vitamin B12 [73,74,75,76] or 1000 µg i.m. B12 every 1-3 months is adequate as a maintenance treatment that can maintain serum B12 concentrations within the reference range [15,77]. Lowered serum concentrations of vitamin B12 are unlikely to be corrected by using 3-6 µg/day B12 from food supplements or high oral B12, i.e., 1000 µg provided once weekly [27,72,78]. Oral and i.m. protocols are likely to be equivalent in correcting serum B12 concentrations [12,79]. Available vitamin B12 forms such as cyanocobalamin and methylcobalamin are safe and beneficial [23,80].

Neuropsychiatric conditions are common in patients with B12 deficiency, although the available studies are heterogeneous regarding patient characteristics and the diagnostic tools employed [2,3,26,30,56,81,82,83,84,85,86]. This variability in particular concerns the differentiation between peripheral sensory neuropathy and sensory neuronopathy (gangliopathy) alone or as part of the subacute combined spinal cord degeneration (SCD) syndrome. At least 50% of patients have both manifestations concurrently [21]. Therefore, in patients with sensory symptoms, differential diagnoses for both disorders need to be considered.

The time needed for hematological and neurological symptoms to recover after starting vitamin B12 treatment varies among subjects [11,87], but high-quality follow-up studies are not available. In general, subjective improvement occurs earlier and is more impressive than objective recovery of neurological function [26]. The earliest subjective improvements after B12 treatment occurred for paresthesia and balance, leading to full recovery [86]. Up to 25% of patients have been reported to retain severe neurological symptoms despite normalization of blood marker levels [6,7]. In approximately 20% of patients with neurological signs and symptoms, recovery may become evident after 3 months of starting the therapy and remain partial [56]. Symptoms of neuropathy mostly improve within several months but may take up to one year for sensory symptoms to resolve after the start of B12 therapy [2]. In patients with SCD, sensory neuronopathy may improve after 7 weeks of B12 therapy (range: 2–32 weeks) [21]. In general, B12 treatment for a minimum of 2 months improves signs and symptoms in all patients with SCD, while Romberg's sign and mild sensory disturbances in the toes and fingers may persist in some patients after several weeks or months of B12 therapy [5,88]. In a group of deficient patients who were treated and followed up for 2-24 months, the patients with stomatitis showed a complete recovery [2]. In another study, neuropsychological tests were corrected after 6 weeks of multiple i.m. injections of 1000 µg B12 [89]. Functional recovery measured by an activity of daily living score may take as long as 12 months, although single cognitive tests may already show some improvement within 3 months [81].

The duration and severity of pretreatment deficiency symptoms have a clear impact on the time course and degree of recovery after starting B12 therapy [1,85], thus emphasizing the importance of early identification and prevention.

3.2. Results of the Delphi Survey

3.2.1. Characteristics of the Survey Panelists

Medical doctors constituted 67% of the 46 panelists in the first round of the survey and 64.4% of the 42 panelists in the second round. The geographical distribution of the panelists in the two survey rounds was as follows: 26.2% were from Asia and the Middle East, 47.6% were from central and western Europe, 14.3% were from eastern Europe, and 11.9% were from North America. The specialties of the 42 panelists in survey 2 were as follows: n = 16 internal medicine, n = 7 neurology, n = 6 other medical specialists, n = 11 researchers in the B12 field, n = 2 people with medical backgrounds working in B12-related societies or groups (Figure S2).

3.2.2. Delphi Survey Rounds

The questions that received consensus in the first survey (Table S4) were revised and used in the second survey (Table S5). The final consensus statements are summarized in Table 1 and Table 2.

3.2.3. Consensus on the Clinical Practice of Diagnosing Vitamin B12 Deficiency

The panelists agreed that the delay in the diagnosis of vitamin B12 deficiency may be explained by barriers such as the variability of symptoms, low doctor awareness of the disease, not paying attention to the patient’s symptoms, and limited access to modern laboratory markers (Table 1). There is agreement that some of these barriers can be overcome by educating both doctors and patients. The panelists agreed that symptoms related to the peripheral and central nervous systems are the most common symptoms in patients with clinically manifested vitamin B12 deficiency (Figure S3). At the same time, it was agreed that symptoms affecting the central nervous system (mental function and neurological function), followed by those affecting the peripheral nervous system, are the most difficult symptoms to link to B12 deficiency (Figure S4), suggesting the need for more specific awareness in a primary medical care setting and improvement in examination options. Moreover, there is agreement that metabolic biomarkers such as methylmalonic acid and homocysteine can be useful in guiding the diagnosis and optimizing and monitoring of treatment with vitamin B12 (Table 1). The panelists agreed that periodic measurement of serum concentrations of vitamin B12 can be helpful in identifying people at risk of deficiency, such as people receiving long-term metformin treatment.

If concentrations of serum vitamin B12 are above the reference range and people have no specific medical conditions, the panelists agreed that inquiring about usage of supplemental vitamin B12, repeating the measurements of serum B12 after a given time interval and measuring routine blood markers of renal or liver function or malignancies could provide a clue toward the cause of high serum B12 unexplained by excess intake (Table 1).

In the context of clarifying the cause of vitamin B12 deficiency, the experts agreed that food-cobalamin malabsorption may cause vitamin B12 deficiency in elderly people even if the dietary intake of B12 is adequate and the person has no gastrointestinal disorders. To support identifying the cause of vitamin B12 deficiency, doctors need to inquire whether the person adheres to a diet that is low in foods from animal sources and should ask the person about disease history, regular use of certain medications and recreational use of nitric oxide. There is agreement that malabsorption of vitamin B12 can be suspected in people with autoimmune disorders, gastric surgery, a family history of pernicious anemia, and people with positive test results for antibodies against intrinsic factors or parietal cells in the serum. Moreover, assessing folate and iron statuses can aid in the differential diagnosis of vitamin B12 deficiency (Table 1). The algorithm for diagnosing vitamin B12 deficiency and its causes agreed among the panelists (Figure 1). In people with symptoms suggesting vitamin B12 deficiency, the general practitioner can collect information on medical history, perform physical examinations, and perform basic laboratory tests such as blood count and serum B12. Advanced clinical and laboratory markers can be ordered if there are indications (e.g., intrinsic factor antibodies in cases of gastrointestinal symptoms). After the preliminary workup, treatment with B12 can start. Depending on the clinical symptoms, some patients may need referral to other specialties.

3.2.4. Consensus for Clinical Practices of Treatment, Prophylaxis, and Long-Term Management of Vitamin B12 Deficiency

Due to the lack of evidence on the comparative effectiveness and safety of different chemical forms of vitamin B12 (e.g., cyanocobalamin and methylcobalamin), the panelists agreed that evidence from randomized controlled trials is needed. The prophylactic use of vitamin B12 was regarded as necessary for people diagnosed with atrophic gastritis, those with previous bariatric surgery, people at risk of B12 deficiency due to illnesses or medications, people with low or no consumption of animal source foods and people ever diagnosed with vitamin B12 deficiency when they decided to become pregnant (Table 2).

Regarding the route of administration, it was agreed that acuity and severity of the symptoms of vitamin B12 deficiency necessitate prioritizing treatment with parenteral vitamin B12. Contraindications to parenteral use, such as in people receiving anticoagulant medication, need to be taken into consideration, and oral supplementation may be preferred in those cases. The decision on the route of administration during long-term treatment needs to balance patient preference and treatment goals. If vitamin B12 treatment failed to alleviate the patient’s symptoms, the panelists agreed that doctors should consider alternative conditions that may explain the symptoms and should reevaluate the appropriateness of the dose of B12 therapy. In nonresponsive cases, doctors can consider using parenteral B12, especially if oral B12 treatment has been used in the past or when serum vitamin B12 is not normalized (Table 2).

The detection and treatment of vitamin B12 deficiency during pregnancy, lactation and infancy should receive high priority due to the otherwise serious impact on fetal and infant development. Moreover, prophylactic B12 supplementation should be used from prepregnancy until the end of the lactation period in women with previously diagnosed B12 deficiency or who are at risk of vitamin B12 deficiency due to low dietary intake of B12.

Aspects related to the diagnosis and treatment of vitamin B12 deficiency that did not achieve agreement among the panelists (Table S7) are briefly discussed below.

4. Discussion

4.1. Delphi Consensus

The present work was undertaken with the aim of reviewing the current knowledge related to the diagnosis and treatment of vitamin B12 deficiency. In addition, we sought expert opinion to guide the diagnosis and treatment of vitamin B12 deficiency, especially in areas where there are still heterogeneous views and practices. This study identified controversially discussed topics that need to be investigated in future studies.

Our results suggest that introducing educational elements for health care professionals and patients could shorten the path for diagnosing vitamin B12 deficiency and accelerate the initiation of treatment. The need to use different laboratory markers of vitamin B12 to aid in diagnosis stresses the importance of making these markers widely available and affordable, which is not the case in many countries. On the other hand, to avoid costs for the health care system, measuring vitamin B12 biomarkers needs to be tailored to the patient’s health condition. Furthermore, regular measurement of serum vitamin B12 concentrations in all patients at risk (e.g., people using metformin) may cause unnecessary costs in the long term, which needs to be weighed against the costs of vitamin B12 treatment itself. The high concentrations of vitamin B12 encountered while suspecting B12 deficiency may be due to underreporting supplement use, autoimmune disorders affecting the B12 assay, or renal or liver diseases.

The most valuable source for identifying the cause of vitamin B12 deficiency is listening to patients’ complaints, disease history, and clinical examination. Symptoms affecting the nervous system or mental function are more difficult to diagnose and may escape detection due to the low specificity of the symptoms and the limited time available for each patient in primary care settings.

Identifying the cause of vitamin B12 deficiency would enable tailoring the treatment with vitamin B12 (dose, frequency, route of administration, and duration). In people with severe signs and symptoms, the use of injectable forms of vitamin B12 for several weeks is acknowledged, at least in the initial treatment phase. The therapeutic regimen can be revised later when symptoms improve or when patients need to receive vitamin B12 for a longer duration. Prophylactic use of vitamins should receive special attention in at-risk groups and during critical life stages, such as pregnancy and lactation, or in elderly individuals.

4.2. Additional Points Raised in the Board Discussion

The board members agreed that in patients with clinically suspected severe B12 deficiency, there must not be any delay in initiating appropriate B12 treatment. Serum specimens for measuring vitamin B12 marker(s) should be collected prior to treatment. Laboratory confirmation of this deficiency can help cover B12 treatment costs by health insurance in many countries.

The board discussed that measuring serum B12 concentrations to monitor B12 treatment should be decided on a case-by-case basis (for example, when people show new symptoms, to check compliance with oral treatment and to check whether B12 was absorbed after oral administration). Although one-third of patients with vitamin B12 deficiency can have normal serum B12 concentrations, measuring serum vitamin B12 is still a convenient screening test when the clinical picture is suggestive of vitamin B12 deficiency. The use of holotranscobalamin as a screening marker of B12 deficiency may increase the cost of the measurements.

The use of vitamin B12 as a prophylactic measure in patients with atrophic gastritis, after bariatric surgery, in vegans, and due to illnesses or medications has reached the consensus of the panelists. The board members discussed the importance of considering prophylactic vitamin B12 in people using metformin [90] or L-DOPA [91,92] to reduce the possible increased risk of neuropathy due to vitamin B12 deficiency, although this needs to be tested in randomized controlled trials.

There is currently no gold standard for choosing the dose and route of B12 administration. Various protocols have been used successfully. In general, whenever B12 deficiency is clinically manifested or the patient cannot absorb B12, therapeutic doses between 1000 µg/d and 2000 µg/day orally or 1000 µg i.m. (e.g., daily, then weekly, then monthly for maintenance treatment) can be used. The severity of the deficiency symptoms may necessitate choosing injectable forms of B12 at least in the starting phase. The clinical response to B12 treatment in SCD patients is expected to be better when large doses of B12 are administered more frequently than when B12 is administered less frequently [93]. The treatment regimen can be revised later to fulfill both the treatment goals and patient convenience.

The scientific board members noted that the duration needed to resolve symptoms of vitamin B12 deficiency can vary. The lack of clinical improvement after 4-8 weeks for anemia and after 6-12 months for neurological signs could suggest that the symptoms are not due to B12 deficiency or that the dose of vitamin B12 or the route of administration needs to be adjusted to the severity of symptoms. Since serum B12 concentrations do not always mirror the clinical picture, this marker is not optimally suited to monitor the effect of treatment, although it can help uncover a lack of patient compliance and malabsorption of oral B12 treatment. An insufficient clinical response to B12 treatment during the first weeks of therapy should not routinely lead to the discontinuation of treatment.

4.3. Strengths and Limitations

This study is based on a scoping review that evaluated existing consistent data on practice-relevant topics while identifying areas with an urgent need for consensus where there is still no evidence. The invited scientific board members were not eligible for the surveys, which can ensure survey neutrality. The professionals who participated in the surveys had sufficient diversity regarding their disciplines and countries of practice. Therefore, these consensus points may be generalized to many countries. This study has several limitations, such as the use of only PubMed and the fact that the search was limited to the last 20 years.

4.4. Conclusions

The present study provides an overview of consistent data on the diagnosis and treatment of vitamin B12 deficiency. In addition, the Delphi study established a robust consensus on various aspects of diagnosing and treating B12 deficiency to support medical practice in areas where there is insufficient evidence. The experts agreed on the need for educational and organizational changes in the current practices. Although clinical symptoms should receive more weight in diagnosing B12 deficiency, the panelists recognized the need to use serum vitamin B12 concentrations as a cost-effective screening marker and the need to additionally measure a metabolic marker to support the diagnosis. The B12 dose and treatment regimens needed to be adjusted according to the severity of the symptoms and the cause of the deficiency. Several topics require future in-depth investigations, such as the clinical benefit of using prophylactic B12 supplementation in some groups who are at risk of B12 deficiency due to accompanying diseases or medications.