Changing the bar for B12 deficiency: possible clinical impact of NICE’s thresholds revisions

Introduction

Vitamin B12 (B12) deficiency can impact health at any age if timely supplementation is not provided (1,2). Severe B12 deficiency in pregnancy has been associated with increased risks of gestational diabetes, preeclampsia, and preterm birth during pregnancy, as well as impaired growth and neurological damage in newborns (1). In older adults, it has been linked to cognitive decline, with even borderline-low levels associated with conditions such as Alzheimer’s disease, vascular dementia, and Parkinson’s disease (2). The UK’s National Institute for Health and Care Excellence (NICE) acknowledges that there is currently no universally accepted gold standard for diagnosing B12 deficiency and total serum B12 measurement remains the first line and most commonly used diagnostic tool to clinically guide supplementation decisions (3). NICE has issued new recommendations for the diagnosis and management of B12 deficiency in individuals aged 16 years and older. The clinical practice guideline clearly identifies the conditions associated with an increased risk of B12 deficiency that warrant testing. It also emphasizes that total and active B12 can generally be used interchangeably as initial diagnostic tests, with the exception of pregnant women, for whom active B12 is preferred. Finally, the guideline introduces new diagnostic thresholds, supported by a systematic review (SR) of the literature aimed at identifying the most accurate methods for diagnosing B12 deficiency and achieving optimal patient outcomes (4).

NICE currently recommends using a total serum B12 threshold of 180 ng/L to identify absolute B12 deficiency, and a range of 180–350 ng/L to indicate possible deficiency. In the latter case, confirmation via additional biomarkers such as serum methylmalonic acid or plasma homocysteine are advised although often not performed, due to cost implications for services (3). NICE has emphasized that these thresholds are intended to be broadly applicable across most commonly used assays, in the case that their lower reference limits (LRLs) tend to be similar. Alternatively, laboratories may use their own validated thresholds for total B12 (3). However, NICE guidance remains vague on this point. It is likely that the LRL of the reference interval (RI), either defined by the assay manufacturer or established through published independent clinical studies, may be considered acceptable for meeting this indication. Retrospective case series based on total B12 measurements from laboratory records are often used to estimate local reference ranges. However, since this test is frequently (and inappropriately) performed as a screening tool or to monitor supplementation, the resulting threshold levels may be overestimated (4).

Recent studies indicate that the LRLs adopted in some assays might be considered aligned with NICE’s thresholds, for example, 181 ng/L for the Siemens Centaur XP and 191 ng/L for the Roche Cobas (5). Manufacturer-reported LRLs for commonly used assays indicate LRLs of 187 ng/L for Abbott, 145 ng/L for Beckman, 211 ng/L for Siemens, and 197 ng/L for Roche. Except for Beckman, which uses a lower LRL, the other assays might be roughly considered “to be similar” and quite aligned with NICE’s threshold.

Two main quandaries should, however, be approached.

• Firstly, are the aforementioned thresholds, including the ones recommended by NICE, considered reliable when applied to rule out possible B12 deficiency?
• Secondly, could harmonized thresholds be considered for assays declaring comparable LRLs?

Under focus on the thresholds recommended by NICE

The clinical validity of the LRLs retrieved from clinical studies or defined by the manufacturers has to be critically appraised in relation to the study design (statistical approach, sample size, and selection criteria). The main critical issue lies in the fact that the age of the investigated reference populations is generally well lower (median age ~45 years) than those typically undergoing B12 testing in our laboratories (median age 65 years) (5,6). Furthermore, the NICE threshold of 180 ng/L appears to have been defined more on the basis of expert consensus than on rigorous scientific assessment without experimental data. Although the guideline is supported by a SR of the literature, this review primarily evaluates the diagnostic accuracy of different B12 thresholds and reports the assays used, rather than justifying the specific cut-off adopted by NICE (3).

The threshold of 180 ng/L for detecting deficiency quoted by NICE appears to originate from a single study included in the SR, but this threshold was derived using outdated methods with no traceability to current assay standards (3,7). As such, it may not be applicable to modern immunoassays (4). Furthermore, this threshold appears to be lower than those used in recent clinical trials, and adherence to the NICE cut-off could result in underdiagnosis of B12 deficiency in individuals who might still benefit from supplementation (8).

An alternative threshold of 244 ng/L (180 pmol/L) for identifying B12 deficiency appears to be better supported in the SR, as it demonstrated higher diagnostic specificity than the 180 ng/L cut-off (9). This value was determined using the Abbott AxSYM analyser, a relatively modern assay that can be regarded as a precursor to current-generation Abbott methods, with comparisons available to the more recent replacement methods.

Meanwhile, the upper reference limit of 350 ng/L reported by NICE for possible deficiency is based on a study using the Siemens Advia Centaur analyser (10), and also in this case comparisons are available by the manufacturer to the more recent Siemens replacement methods.

Under focus on harmonization between methods

Relying on the manufacturers’ defined LRLs to determine whether methods are sufficiently harmonized, as done by NICE, is undoubtedly a very rough assessment. Assessing inter-assay variability of B12 measurements, especially around the thresholds proposed by NICE, is essential to determine whether these cut-offs can be universally applied or need adjustment based on the assay used. A recent study has investigated the comparison between the most commonly used total B12 methods, demonstrating that B12 values, obtained from different immunoassays, are not interchangeable, even after the implementation of the World Health Organization (WHO) International Standard (IS 03/178), which aimed to improve standardization (11). In particular, Abbott and Beckman methods were shown to overestimate and underestimate, respectively, B12 concentrations vs. Roche and the other methods. Furthermore, Beckman method was affected by both proportional and systematic errors. Although the inter-method bias was well below the desirable goal of 9.4% for Abbott Architect, and to some extent also lower for Siemens ADVIA Centaur, the wide limits of agreement prevented any firm conclusion regarding the agreement between Roche and the other methods (11). In other words, the thresholds for serum total B12 recommended by NICE should not be assumed applicable to all assays, and this variability underscores the need for laboratories to use assay-specific or assay-adjusted thresholds when interpreting B12 concentrations.

Therefore, it is crucial to take into consideration the analytical platforms used to derive the thresholds recommended by NICE (180 ng/L as threshold of absolute deficiency and 350 ng/L as threshold of possible deficiency), or reported in the SR (244 ng/L as threshold of deficiency) (3), to convert them into values detectable by the most commonly used methods to measure total B12.

Under conversion of the threshold levels

With reference to the equations published in the recent paper by Cesana et al. (11), inverse regressions may be used for obtaining the “X calibrated value” together with its 95% CI according to Draper and Smith (12). This procedure allows to convert the value determined by Roche (used as reference method since the one traced to the IS 03/178) into those derived by other assays. Notably the ordinary least square (OLS) were performed on a serum set of 97 samples and on a set of 11 selected specimens used for External Quality Assessment (EQA). The former were used to obtain the cut-off for Roche Cobas, Beckman DxI, Advia Centaur Siemens, and the latter for deriving thresholds for Atellica Siemens and Alinity Abbott (Table 1) (11).

Concerning the value of 180 ng/L originally suggested by NICE as the threshold for absolute deficiency, a tentative conversion into a Roche value could be considered, taking account of the original study used Quantaphase (QP) kit (Bio-Rad Laboratories, Richmond, CA, USA) (7). Notably, the analytical conversion has a questionable validity according to the published data on the relationship between QP and Roche E-170 (13). In addition, the standard error is not available in the original paper (13). Furthermore, an overestimate of Roche by 48 units on the median value of 438 pmol/L (593 ng/L) has been obtained from an analysis affected by small proportional and constant systematic error between the two methods, on log transformed data.

The threshold for B12 deficiency of 244 ng/L (180 pmol/L), reported in the SR to be associated to an acceptable specificity (3) was obtained in the original study by AxSYM analyser Abbott (9). This value may tentatively be converted into Abbott Architect analyser, by accounting for the regression (OLS) between the two analytical platforms declared by the manufacturer (slope =1.09; intercept =3.7) on serum [the 95% confidence interval (CI) was not reported] (14). The obtained threshold for Architect is 230 ng/L. This cut-off may be converted into Roche by using the inverse regression applied to OLS calibrations of the published paper performed on the serum set (11). As a result, for Roche, the estimated value of 204 ng/L is practically equal to the threshold of 203 ng/L recently used in some clinical trials (15).

Similarly, the cut-off of 350 ng/L [originally determined by ADVIA Centaur (10)] may be converted into Roche, obtaining a value of 379 ng/L by using the OLS calibrations reported in the original paper on the serum set (11).

Among Abbott analysers, the thresholds are similar, although derived from different sample sets (see Table 1). Of note, the 95% CIs of the thresholds obtained for Roche and Abbott analysers overlap and this agrees with the evidence that the % bias between these two methods is well within the desirable goal of bias, although the Limit of Agreement are quite wide (see Table 1).

For Siemens too, the 95% CIs of the thresholds obtained for Roche and Atellica or Advia Centaur analysers overlap; however, the wide limits of Siemens 95% CI impacts on the clinical applicability of these thresholds (see Table 1).

For Beckman, we tentatively estimated the threshold at 148 ng/L, but a more conservative 95% URL of 163 ng/L should be preferred, although these levels should be cautiously considered as it seems that a recalibration of the analytical system to the WHO IS 03/178 before estimating the thresholds is required (11).

Conclusions

Referrals to dementia screening clinics continue to increase across high income countries. Across low- and middle-income countries, many cases remain undiagnosed (16). Every year nearly 10 million new cases are identified (17). In many people (85%), neurological, neuropsychiatric and cognitive symptoms are the main presenting symptoms of low B12, compared to less than 20% presenting with anaemia (18). It is estimated that 10–20% of over 60s in Europe and North America have low B12 levels, and the prevalence in low- and middle-income countries much higher, where diets are low in animal products (19). With such large numbers of people potentially affected, it is imperative to use B12 assay limits which best capture potentially low B12 in a timely manner. Treatment, whether administered intramuscularly or orally, has traditionally been regarded as safe and relatively inexpensive when compared to the more tangible costs of laboratory investigations. Notably, the high uncertainty or low precision of diagnostic thresholds often leads to further second-level testing, increasing the overall burden. As a result, many clinicians advocate for empirical B12 supplementation in individuals with apparently low circulating B12 levels, offering a trial period before pursuing more extensive assessments of neuropsychiatric and/or cognitive symptoms, given the possibility that these may be reversible.

However, considering supplementation as relatively inexpensive remains somewhat speculative, as it does not take into account indirect costs. These include opportunity costs for both Regional or National Health Services (e.g., physician time) and caregivers (e.g., loss of working hours).

A detailed cost analysis of B12 deficiency management across various clinical scenarios could demonstrate that initiating treatment, including clinical evaluations, specialist consultations (with dietary and nutrition assessments), neurological evaluations, and laboratory monitoring, incurs costs approximately 2.5 times higher than those associated with the diagnostic phase, which largely relies on laboratory testing alone (1).

The tentative conversion of the thresholds shows that, even for assays demonstrating acceptable agreement, the derived thresholds are affected by wide CIs.

In other words, their precision may represent a critical limitation in clinical practice. To minimize the risk of failing to identify patients with true cobalamin deficiency, it may be reasonable to adopt the upper limit of the CI.

This approach is clinically tolerable, given that B12 supplementation is inexpensive and associated with few, if any, documented adverse effects.

It is also noteworthy that the threshold identified for Roche (204 ng/L) is practically identical to that used in some current clinical trials (15).

Similarly, the thresholds derived for Abbott Alinity, very close to those of Architect, could be considered in future clinical validation studies.

By contrast, for methods not traceable to the WHO IS (e.g., Beckman, Siemens), threshold conversion should be interpreted with caution. In these cases, prioritizing the upper limit of the 95% CI may be advisable, even though this would likely increase both the number of subjects identified with deficiency and the demand for second-level confirmatory tests, with an inevitable impact on healthcare costs.

This situation illustrates how improvements in traceability and harmonization of laboratory methods can directly influence patient outcomes as well as the economic sustainability of healthcare systems.

Acknowledgments

None.

Footnote

Provenance and Peer Review: This article was commissioned by the editorial office, Journal of Laboratory and Precision Medicine. The article has undergone external peer review.

Peer Review File: Available at https://jlpm.amegroups.com/article/view/10.21037/jlpm-25-49/prf

Funding: None.

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://jlpm.amegroups.com/article/view/10.21037/jlpm-25-49/coif). S.F. serves as an unpaid editorial board member of Journal of Laboratory and Precision Medicine from May 2025 to April 2027. The other authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Ferraro S, Lucchi S, Montanari C, et al. Vitamin B12 deficiency in newborns: impact on individual's health status and healthcare costs. Clin Chem Lab Med 2025;63:559-71. [Crossref] [PubMed]
2. Moore E, Mander A, Ames D, et al. Cognitive impairment and vitamin B12: a review. Int Psychogeriatr 2012;24:541-56. [Crossref] [PubMed]
3. Vitamin B12 deficiency in over 16s: diagnosis and management. London: National Institute for Health and Care Excellence (NICE); March 6, 2024.
4. Ferraro S, Da Molin S, Cereda C, et al. Vitamin B12 Deficiency and New Recommendations by the National Institute for Health and Care Excellence: A Challenging Clinical and Laboratory Topic. Clin Chem 2025;hvaf037. [Crossref] [PubMed]
5. Jassam N, Bancroft T, Luvai A, et al. Vitamin B12 reference intervals on Beckman, Roche and Siemens analytical platforms. Ann Clin Biochem 2023;60:417-22. [Crossref] [PubMed]
6. Solé-Enrech G, San-José P, Aliste-Fernández M, et al. Vitamin B12 and folate levels in a healthy population: establishing reference intervals. Clin Chem Lab Med 2019;57:e173-5. [Crossref] [PubMed]
7. Matchar DB, McCrory DC, Millington DS, et al. Performance of the serum cobalamin assay for diagnosis of cobalamin deficiency. Am J Med Sci 1994;308:276-83. [Crossref] [PubMed]
8. Ferraro S, Tosi M, Verduci E, et al. Biomarkers to measure the need and the effectiveness of therapeutic supplementation: a critical issue. Clin Chem Lab Med 2025;63:e218-20. [Crossref] [PubMed]
9. Heil SG, de Jonge R, de Rotte MC, et al. Screening for metabolic vitamin B12 deficiency by holotranscobalamin in patients suspected of vitamin B12 deficiency: a multicentre study. Ann Clin Biochem 2012;49:184-9. [Crossref] [PubMed]
10. Bondu JD, Nellickal AJ, Jeyaseelan L, et al. Assessing Diagnostic Accuracy of Serum Holotranscobalamin (Active-B12) in Comparison with Other Markers of Vitamin B12 Deficiency. Indian J Clin Biochem 2020;35:367-72. [Crossref] [PubMed]
11. Cesana BM, da Molin S, Jassam N, et al. Assessing the harmonization of current total vitamin B12 measurement methods: relevance and implications. Clin Chem Lab Med 2025; Epub ahead of print. [Crossref]
12. Draper NR, Smith H. Applied regression analysis 3rd. Edition. NY: John Wiley & Sons, Inc.; 1998:83-6.
13. Yetley EA, Pfeiffer CM, Phinney KW, et al. Biomarkers of vitamin B-12 status in NHANES: a roundtable summary. Am J Clin Nutr 2011;94:313S-21S. [Crossref] [PubMed]
14. 510(k) Summary Abbott AxSYM ® B12. Summary of Safety and Effectiveness Information Supporting a Substantially Equivalent Determination. 2005.
15. Upadhyay RP, Chowdhury R, Mundra S, et al. Comparative effectiveness of daily therapeutic supplementation with multiple micronutrients and iron-folic acid versus iron-folic acid alone in children with mild-to-moderate anaemia in rural India: an open-label, randomised controlled trial. Lancet Glob Health 2025;13:e497-507. [Crossref] [PubMed]
16. Li X, Feng X, Sun X, et al. Global, regional, and national burden of Alzheimer's disease and other dementias, 1990-2019. Front Aging Neurosci 2022;14:937486. [Crossref] [PubMed]
17. WHO Dementia factsheet. March 2025. Available online: https://www.who.int/news-room/fact-sheets/detail/dementia (accessed 29 August 2025).
18. Wolffenbuttel BH, Owen PJ, Ward M, et al. Vitamin B(12). BMJ 2023;383:e071725. [Crossref] [PubMed]
19. Vitamin B12. Available online: https://ods.od.nih.gov/factsheets/VitaminB12-HealthProfessional/. Accessed September 2025.