Locally validated biomarker cutoffs should inform micronutrient deficiency burdens

We read the micronutrient deficiencies burden study by Gretchen Stevens and colleagues (November, 2022)¹ with concern, since the stated findings have the potential for unnecessary so-called shockvertising and reactionary policy shifts. For example, the projection made, that 77–92% of preschoolers and non-pregnant women in India have any sentinel micronutrient deficiency, resulting in eroded human potential, contrasts with the steady improvement in Indian human development indicators over time, with attendant improvements in economy, diet, and social welfare programmes. Furthermore, the prevalence of micronutrient deficiency could be inflated if the sociodemographic index-based prediction of missing nutrient deficiencies had low accuracy (not reported in the study). Thus, from these study data from India, we calculate that the prevalence of at least one sentinel deficiency might be considerably inflated (by around 20 percentage points).

Importantly, the biomarker-based cutoffs to diagnose micronutrient deficiency are uncertain, since they have several interdependencies, not least habitual micronutrient intake and body adaptation. These cutoffs are mostly statistically derived from lower limits of their distribution in so-called healthy participants, usually from wealthy white populations with generous animal-based habitual nutrient intakes. The foregoing considerations suggest that, without contextual or local validation of such statistical (instead of functional) cutoffs, extreme caution is warranted when diagnosing deficiency pathologies. A contextual example is where statistical cutoffs for serum zinc from stringently defined healthy participants were substantially lower than usual cutoffs (by 10–18 μg/dL), resulting in markedly lower prevalence of zinc deficiency in preschoolers (6% vs 19%) and in adolescents (6% vs 31%).² Similarly, derived functional cutoffs in Australians indicate that iron deficiency anaemia begins to appear when serum ferritin concentrations reach 10 μg/L, in contrast to the usual cutoffs (12 μg/L for children <5 years and 15 μg/L for individuals aged ≥5 years).³ Finally, the deterministic use of statistical cutoffs to diagnose deficiency, while ignoring its uncertainties, could be problematic; a probabilistic approach might be preferable for estimating the deficiency risk in populations.⁴

The problem is the potential kneejerk policy of universal micronutrient fortification or supplementation in response to shocking deficiencies. Such a policy would need careful deliberation, using all elements of decision making, since temporary fire-fighting measures can eventually achieve permanence, with unknown long-term consequences. Safer and sustainable dietary diversity, which should take centre stage in policy, recedes into oblivion. Nutritional policy must be contextual and break free from a colonial mindset, especially in low-income and middle-income countries, which bear the brunt of industrial, nutrient-product-based policy interventions. For example, priority should be given to contextual validation of the one-size-fits-all cutoffs, from both statistical and biological perspectives, for more realistic, uninflated projections of micronutrient deficiency and disease burden.