Cobalamin (vitamin B₁₂) is a water-soluble vitamin obtained by humans from animal sources, including meat, dairy products and eggs.¹ This vitamin is vital in deoxyribonucleic acid (DNA) synthesis, red blood cell formation and neurological functioning.² It is also linked to the synthesis of various neurotransmitters, and thus, it has been implicated in the pathogenesis of various psychiatric disorders.³

Deficiency presents with broad, non-specific clinical features, including symptoms related to anaemia, gastrointestinal issues, neurological impairments and psychological disturbances.⁴ Often, neuropsychiatric symptoms are the first clinical manifestation and precede the haematologic and gastrointestinal symptoms.^5,6,7 Psychiatric manifestations of vitamin B₁₂ deficiency can include depression, apathy, irritability, dementia, mania, catatonia, delirium, anxiety and psychosis.^8,9,10

The prevalence of low serum or plasma vitamin B₁₂ concentrations among adults varies, with rates of 2.7% in the United States, 8.3% in the United Kingdom and 14.7% in Germany.¹¹ In Africa, available data on prevalence indicate that 35% and 15% were deficient in Kenya and Ethiopia, respectively; however, in Ghana, the prevalence was relatively lower (7%).¹² Epidemiological data for cobalamin deficiency in South Africa are limited.

Psychiatric patients are frequently screened for vitamin B₁₂ deficiency in the absence of haematologic or neurologic findings,¹³ with some studies going as far as recommending that routine screening for serum vitamin B₁₂ levels should be adopted by all hospitals for psychiatric patients.^14,15 Based on the authors’ clinical experience, routine screening for vitamin B₁₂ deficiency is common practice across psychiatric departments in KwaZulu-Natal, including at Townhill Hospital. At this facility, admitted patients are typically screened for vitamin B₁₂ deficiency, irrespective of the presence of specific risk factors.

Certain groups are at heightened risk for vitamin B₁₂ deficiency because of factors that impair intake or absorption. These include individuals with restricted diets, pregnant women, the elderly, people with gastritis and certain autoimmune disorders (e.g., pernicious anaemia), people with intestinal diseases (e.g., Crohn’s, celiac disease), those with surgeries involving the gastrointestinal tract and people on medications that inhibit B₁₂ absorption, such as metformin, proton pump inhibitors (PPIs), H₂-receptor antagonists and oral contraceptives.^{4,16,17,18,19}

A diagnosis of vitamin B ₁₂ deficiency is challenging, as confirmation relies on laboratory findings, which remain controversial. In terms of laboratory investigations for B ₁₂ deficiency, there are four biological markers of B ₁₂ : total serum vitamin B ₁₂ , holotranscobalamin (holoTC) and measures of B ₁₂ metabolites homocysteine (Hcy) and methylmalonic acid (MMA). Each of these tests have limitations. ⁴ The serum B ₁₂ is the most frequently used test and it measures the circulating concentration of vitamin B ₁₂ bound to both vitamin binding proteins, transcobalamin and haptocorrin. ²⁰ A critical problem with these assays is that they can be influenced by the presence of interfering anti-intrinsic antibodies (particularly in patients with pernicious anaemia), thereby giving elevated vitamin B ₁₂ concentrations in cobalamin-deficient patients. ²¹ There is also a need for scientific consensus defining cut-off values to diagnose deficiency states. ²² Measuring Hcy and MMA is based on the rationale that cobalamin deficiency leads to the accumulation of these metabolites. ²³ While these tests may have greater sensitivity for cobalamin deficiency, they remain problematic in diagnosing B ₁₂ deficiency. Differential causes need to be considered if MMA levels are elevated. Other possible causes of elevated MMA include renal impairment or dehydration, ²⁴ and if Hcy is elevated, it may reflect folate deficiency; therefore, this would have to be excluded. ²⁵ Therefore, vitamin B ₁₂ deficiency can be suspected only when serum levels are low, and MMA and Hcy are elevated without renal disease, volume depletion and folate deficiency. A variety of conditions may impact serum cobalamin levels, but MMA or Hcy level within the normal range suggests that vitamin B ₁₂ deficiency can be excluded. ²⁶ Measuring holoTC, the bioactive form of vitamin B ₁₂ , remains debatable, with one large study concluding that holoTC has the highest diagnostic accuracy in predicting deficiency ²⁷ and another extensive study showing that the holoTC immunoassay cannot be used to reliably determine the vitamin B ₁₂ status. ²⁸ In our setting, the serum vitamin B ₁₂ is the most cost-effective test and widely used test to assess for cobalamin deficiency.

Clinical utility is increasingly utilised in health care, and critical measures of clinical utility include clinical effectiveness and cost-effectiveness. However, as Smart highlights, this narrow focus overlooks other essential factors that influence clinical decision-making, including appropriateness, accessibility, practicality and acceptability, all of which should be considered when assessing clinical utility.²⁹ Cost-effectiveness analysis (CEA) assesses the outcomes and costs of various strategies or interventions.³⁰

While diagnosing and treating vitamin B₁₂ deficiency in the psychiatric population is essential, the prevalence of this condition among patients admitted to Townhill Hospital remains unclear. Consequently, it is uncertain whether routine screening is justified in non-vulnerable patients who lack haematological or neurological findings. Furthermore, current testing practices may need to be more reliable, potentially contributing to unnecessary costs in an already resource-constrained environment. This study investigated the clinical utility and cost-effectiveness of routine screening for vitamin B₁₂ deficiency in adult psychiatric patients admitted to Townhill Hospital.

A total of 496 adult patients (18 years and older) were admitted to Townhill Hospital during the study period. Out of the total sample, 81 participants (16.3%) were excluded from the study because of the absence of documented serum vitamin B₁₂ levels in their medical records. In addition, 49 participants (9.9%) were excluded because their files were unavailable at the hospital registry department. The final sample thus consisted of 366 (73.8%) cases.

Key sociodemographic and clinical characteristics are summarised in Table 1.

The mean age was 35.95 years (SD ± 13.44). Most patients were female (n = 198, 54.1%), and 45.9% (n = 168) were male. The study population predominantly comprised African patients, accounting for 75.1% of the cohort (n = 275).

Four of the patients (1.1%) were pregnant.

Most of the participants (n = 290, 79.2%) reported following a regular omnivorous diet, while 12 patients (3.3%) were identified as vegetarian. The dietary preference information was missing for 64 patients (17.5%).

The majority of the participants, totalling 78.4% (n = 287), were established mental health care users, while for 79 individuals (21.6%), it was their first contact with mental health services.

In this study population, 76 patients (20.8%) had a diagnosis of human immunodeficiency virus (HIV), with other common comorbidities including hypertension (n = 60, 16.4%), dyslipidaemia (n = 35, 9.6%), diabetes (7.4%) and epilepsy (n = 26, 7.1%). A total of 193 patients (52.7%) reported alcohol consumption, of whom 12.8% (n = 12.8%) were diagnosed with alcohol use disorder. The primary psychiatric diagnoses included schizophrenia (n = 118, 32.2%), schizoaffective disorder (n = 67, 18.3%) and bipolar 1 disorder (n = 65, 17.8%), with a notable prevalence of cannabis use disorder (n = 47, 17.8%) and multiple substance use disorders (n = 47, 12.8%). Antipsychotics were the most prescribed medication (n = 342, 93.8%), followed by mood stabilisers (n = 191, 52.2%), antidepressants (n = 86, 23.5%) and antiretrovirals (n = 76, 20.8%). Other medications included metformin (n = 26, 7.1%) and PPIs (n = 11, 3.0%), with no use of oral contraceptives or H2-receptor antagonists in this population. A total of 81 patients (22.1%) had a history of prior non-abdominal surgery, while 17 (4.6%) had a history of prior abdominal surgery.

Key variables related to the testing of vitamin B₁₂ are outlined in Table 2.

The mean serum vitamin B₁₂ level was 423.86 mmol/L (SD ± 233.37). The interquartile range (IQR) was 243 mmol/L, with values ranging from 118 mmol/L to 1476 mmol/L. The median serum vitamin B₁₂ concentration was 359 mmol/L. In this study, eight (2.2%) patients had low vitamin B₁₂ levels. Only one (0.3%) of these patients had a Hcy level test performed, and none received an MMA test. Most of the study sample (n = 358, 97.8%) had normal serum vitamin B₁₂ levels.

There was a statistically significant association between low vitamin B₁₂ levels and both pregnancy and dietary status, with more of these patients having a low vitamin B₁₂ level (p < 0.05). However, although these were statistically significant, the effect sizes were small (phi = 0.18 and 0.15, respectively), and the assumption of expected cell counts exceeding five was violated in the pregnancy group.

There was no statistically significant association between mean age and vitamin B₁₂ deficiency, including no statistically significant difference in the percentage of those with low vitamin B₁₂ levels and advanced age of > 60 years (p > 0.05).

Associations of low vitamin B₁₂ with sociodemographic variables are outlined in Table 3.

A statistically significant lower mean serum vitamin B₁₂ level was observed in individuals with schizophrenia (t = 2.57, p = 0.01). However, no significant association was found between a diagnosis of schizophrenia and vitamin B₁₂ deficiency (χ² = 0.14, p > 0.05).

A general trend of a lower mean serum vitamin B₁₂ level in patients with comorbidities was seen, but this was not statistically significant (p-values > 0.05). However, there was a statistically significant difference regarding patients with diabetes, with more of these patients having a low vitamin B₁₂ level (χ² = 3.72, p = 0.05).

There was no statistical difference between patients on chronic medication and mean serum vitamin B₁₂ levels (p-values > 0.05). However, there was a statistically significant difference with regard to patients using metformin, with more of these patients having a deficient vitamin B₁₂ level (χ² = 3.97, p = 0.05).

The results showed no statistical differences between patients with surgical history, alcohol use or alcohol use disorder and mean serum vitamin B₁₂ levels (p-values > 0.05). However, there was a statistically significant difference regarding patients with a history of abdominal surgery, with more of these patients having a low vitamin B₁₂ level (χ² = 6.19, p = 0.05).

Associations of low vitamin B₁₂ with important clinical variables are summarised in Table 4a and Table 4b.

Overall, three of the patients with low vitamin B₁₂ had no identified risk factors (38%), while the remainder had at least one risk factor (n = 5, 62%). Two of the deficient patients had two or more risk factors.

The frequency of risk factors in patients with vitamin B₁₂ deficiency is summarised in Table 5.

This study investigated the clinical utility and cost-effectiveness of screening for vitamin B₁₂ deficiency in adult psychiatric patients admitted to Townhill Hospital. This is the first documented analysis of the clinical utility and cost-effectiveness of routine vitamin B₁₂ screening in a South African inpatient psychiatric population. Routine laboratory screening of psychiatric patients is widely practiced, primarily to detect occult medical conditions. It also facilitates the identification of underlying medical illnesses that may exacerbate or mimic psychiatric symptoms. Adequate screening tests should meet criteria for acceptability, suitability for the population and cost-effectiveness and should positively impact patient outcomes.³¹

In our study, only 2.2% of individuals who underwent routine screening for vitamin B₁₂ deficiency were found to have low cobalamin levels. The total cost per patient with vitamin B₁₂ deficiency was R5780.73, a notably high figure. Additionally, less than half of the patients diagnosed with vitamin B₁₂ deficiency received appropriate treatment, contributing to an increased cost-utility ratio. Consequently, our findings do not support routine screening for vitamin B₁₂ deficiency, as it is not cost-effective.

The study identified a statistically significant difference in the prevalence of low vitamin B₁₂ levels with established risk factors for cobalamin deficiency. Screening for vitamin B₁₂ deficiency should be guided by clinical history (including inquiries about well-known risk factors) and physical findings. We recommend limiting ‘routine testing’ for deficiency to individuals exhibiting physical signs suggestive of deficiency, mental health care users not responding to treatment, and specific high-risk groups. These groups include pregnant women, the elderly, individuals on medications associated with vitamin B₁₂ deficiency (such as metformin, PPIs, H₂-receptor antagonists and oral contraceptives), those with conditions strongly linked to deficiency (such as Crohn’s disease, celiac disease and pernicious anaemia), and individuals with a history of abdominal surgery. In a resource-constrained healthcare environment, better patient outcomes and management may be achieved by directing resources towards higher-yield interventions.