Iron deficiency remains one of the most common micronutrient deficiencies worldwide and is the leading cause of anemia among young children. Although it has declined globally over the past thirty years, children under five in low- and middle-income countries, particularly in sub-Saharan Africa, continue to bear the heaviest burden [1][2]. In children, iron deficiency can hinder cognitive development, slow physical growth, and lead to long-term human capital losses [3][4]. Therefore, international guidelines stress the importance of interventions targeting young children [5][6]. Often linked to other micronutrient deficiencies, it is part of the broader issue of hidden hunger, which affects nearly two billion people globally. This deficiency commonly coexists with stunting and increasingly with overweight within the same populations and households [7][8]. Zinc deficiency, closely related to iron deficiency, further complicates this challenge, particularly in countries like Burundi, with limited progress expected in the next decade [9][10]. Repeated national surveys show high rates of anemia and stunting among children in Burundi [11]-[13], underscoring the urgent need for nutritional interventions to address these issues.

Food-based approaches are important, sustainable alternatives to supplementation and fortification, particularly in rural areas where access to processed foods and health care is limited. Biofortification increases the micronutrient content of staple food crops primarily through two approaches: agronomic biofortification, which applies mineral fertilizers during cultivation to enhance nutrient uptake, and genetic biofortification, which uses conventional plant breeding techniques to improve micronutrient content.

This latter strategy has been widely used for staple crops such as legumes and cereals distributed through international programs such as HarvestPlus, yielding richer and more cost-effective gains in dietary iron, zinc, and vitamin A in many countries.

Nevertheless, the genetic potential observed under controlled breeding conditions does not always translate into comparable nutritional benefits when crops are grown in typical farm soils [14].

The gap between genetic potential and actual crop performance is widening as climate variability reduces the micronutrient content of staple food crops, a trend influenced by both pedological factors and plant-specific physiological processes. This highlights the need for strategies to verify nutrient content under real-world cultivation conditions, rather than relying solely on breeding data [15].

Common bean (Phaseolus vulgaris) is the most widely cultivated and consumed legume worldwide. It plays an essential role in the diet and provides significant nutritional benefits in the Great Lakes region, where per capita consumption is high [14][16][17]. The region features a rich diversity of local varieties and breeding lines, with considerable variation in seed iron and zinc content, making it a prime target for biofortification efforts.

These initiatives have evolved into genetic screening programs in West Africa and Latin America [18][19]. Conventional breeding has produced iron-rich bean varieties now available from the national research institution, ISABU, in Burundi [20][21].

Bean seed iron content is influenced by genetics as well as agronomic and nutritional factors, underscoring the need for local testing. Variety selection can increase seed iron content by 60% - 80% compared with unselected varieties [14]. However, iron content varies considerably by genotype, soil quality, and location, and interactions among these factors can produce variation comparable to genetic differences [22][23].

Soil characteristics such as pH, organic matter, and iron availability at planting determine how much genetic iron potential is expressed in the final seed [24][25].

A variety selected for high iron content under controlled research conditions may perform differently across soil types. In addition, biofortified seeds sold in informal local markets, often several generations removed from certified breeding stock, may not retain the same iron content as the currently recommended ISABU varieties.

Thus, a chain runs from soil to nutrition, in which agricultural practices beyond the simple choice of cultivated varieties influence how a crop’s genetic potential translates into tangible benefits for consumers [26]. The nutritional outcome depends not only on iron content but also on antinutritional factors such as phytic acid and polyphenols, which are naturally present in the seed coat and cotyledons. These substances can markedly reduce iron absorption, even when iron content is accurately measured [16][27][28].

Furthermore, agronomic and compositional characteristics, such as cooking time and drought tolerance [29][30], influence a variety’s adoption and use within households. Thus, the actual nutritional benefit of biofortification depends on a variety’s genetic iron potential, growing conditions, and postharvest bioavailability.

This finding aligns with broader evidence from subsistence farming households in rural Burundi, where child undernutrition is more closely linked to limited dietary diversity and low nonfarm income than to agricultural production alone [31][32]. This underscores the importance of assessing whether available and consumed biofortified beans can improve household iron intake, given local soil and ecological conditions.

This research had two main objectives. First, to compare the initial iron content of four locally available iron-biofortified bean varieties identified by the study population with four biofortified varieties on the official ISABU list that are adapted to the agroecological conditions of the study area. Second, to assess whether iron content remains stable, increases, or decreases after a full cultivation cycle on local soils, taking into account the soil iron status before planting.

This study aimed to assess variation in the iron content of biofortified beans identified as suitable for local cultivation by measuring seed iron content before planting and after harvest, and by analyzing soil samples from each cultivation site. This approach allows the genetic potential of each variety to be distinguished from its actual agronomic performance in Butezi commune, Ruyigi province.

The initial comparison shows that ISABU-recommended varieties had a mean iron content before planting that was more than twice that of popular market-identified varieties. This suggests that biofortified seeds sold in local markets, several generations removed from certified breeding sources, may have lower iron content than the original levels.

This trend aligns with known gaps in seed quality and varietal replacement in Burundi’s bean seed systems [33].

In fifteen of the sixteen variety-by-soil combinations tested, iron content declined. The mean decrease was 51.3% when averaged across the sixteen combinations and 57.9% when computed from the overall group means. This decline was sufficient to reduce the mean postharvest iron content of the ISABU-recommended group (48.4 mg/kg) below the initial mean of 53.1 mg/kg for market-identified varieties.

Mugwiza, which initially had the highest iron content, showed the sharpest decline, losing more than 75% of its initial advantage across all tested soils. Kaneza increased its iron content primarily in clayey soil but lost iron in the other soil types. This suggests that pedological and agricultural conditions may limit iron accumulation during growth, independent of genetic potential [24][25]. However, because soil type in this pilot trial is confounded with hill-specific management and microclimate, this association cannot be fully separated from other site-level factors.

These findings, in which the postharvest iron content of ISABU-recommended varieties approaches or falls below that of market-identified varieties, align with other East African studies showing that the iron content of biofortified and non-biofortified varieties tends to converge under local conditions [34]. This suggests that the iron content of biofortified seeds may decline in markets as seed stock becomes increasingly removed from certified breeding cycles.

Soil-level analysis shows that postharvest iron content was lowest in stony soil, even though stony soil had the highest overall iron and phosphorus content among the four samples. Clayey soil had the highest iron concentrations despite its lower total iron content. This indicates that total soil iron alone does not determine the iron available to the plant. A pattern consistent with published evidence is that high soil phosphorus can reduce iron availability to the plant by forming poorly soluble iron phosphates [27]; however, this mechanism was not directly tested in the present study, since soil pH, organic matter, and extractable iron were not measured. It should therefore be regarded as a plausible explanation rather than a demonstrated cause. Soils are known to differ not only in iron content but also in physicochemical properties that affect iron retention [25].

Consequently, the choice of cultivated varieties should account for these soil-specific characteristics; targeted soil management strategies, such as addressing phosphorus availability, would need to be tested directly in future trials before recommendation.

Iron loss varied by variety regardless of soil type, ranging from 55.1% in Rufutamadeni to 81.2% in Mugwiza. The notable increase observed in Kaneza on clayey soil suggests that genotype-by-environment interactions, already documented for bean seed iron and zinc, can be substantial enough to markedly alter expected outcomes [22][23]. Given the single-hill-per-soil-type design of this pilot study, confirm this observation in trials with multiple hills per soil type before generalizing.

Current understanding of how soil components affect iron retention and bioavailability can explain these soil-related differences. Nitrogen availability is a key driver of protein accumulation in common bean seeds, often independently of iron content [25]. This aligns with the higher crude protein content observed in more fertile soils [35].

In contrast, phosphorus interacts differently with iron: high soil phosphorus levels are reported to promote the formation of poorly soluble iron phosphates in the root zone, thereby reducing the amount of iron available for uptake [27]. This mechanism offers one plausible explanation, among others, for the reduced iron retention observed in phosphorus-rich stony soil [27]; however, it remains a hypothesis rather than a demonstrated cause in this study, since we did not measure soil pH, organic matter, or extractable iron.

Calcium and potassium interact with iron and zinc differently depending on soil type. They compete with zinc and, to a lesser extent, with iron for root uptake [36]. Potassium increases zinc uptake, whereas calcium can decrease it [36].

In addition, calcium influences iron absorption in humans: dietary calcium generally inhibits intestinal absorption of both heme and nonheme iron, likely by disrupting iron transport across the intestinal mucosa [37]. Consequently, a variety’s calcium content may reduce iron availability when both nutrients are consumed in the same meal [37].

Zinc in bean seeds affects bioavailability more than retention. Iron and zinc are thought to compete for similar or overlapping intestinal absorption pathways, including binding sites in the mucus layer that influence how efficiently trace elements reach enterocytes [38]. Consequently, a bean variety rich in both iron and zinc may not deliver proportional nutritional benefits for either mineral.

The importance of crude protein for iron absorption depends on its source and amount. While some animal proteins promote nonheme iron absorption, legume proteins, such as those in common beans, often contain phytic acid, the main antinutrient that limits iron bioavailability in beans, rather than the protein itself [16][28].

These element-specific roles suggest that soil management aimed at strengthening iron biofortification in beans in Ruyigi province may need to consider phosphorus and calcium, among other factors. Furthermore, increasing seed iron content alone does not guarantee increased iron absorption unless zinc, phytic acid, and overall dietary factors are also managed. These soil-nutrient interactions are proposed explanatory mechanisms from the broader literature; the present study did not measure the additional soil parameters (pH, organic matter, extractable iron) needed to test them directly and therefore cannot establish which mechanism, if any, accounts for the differences observed here.

Under optimal conditions, a combination of genetic potential, favorable agronomic characteristics, and market access can benefit farmers [7][39]-[41].

The main advantage of iron-biofortified beans is their higher seed iron content, which can improve iron status and, if iron is preserved through harvest and cooking, may enhance cognitive function. This approach is also highly cost-effective compared with other public health strategies [7][42]-[45]. However, benefits do not depend solely on selecting the right variety. Our results show that, without adequate soil management, genetic iron advantages can be substantially reduced before consumption.

Varieties must be adapted to local soils, where their characteristics are more likely to be preserved [7][39][41]. Multistakeholder partnerships improve access to quality seed and accelerate the adoption of improved varieties [33].

Nutrition-sensitive agricultural strategies, such as biofortification, are most effective when paired with comprehensive policies that strengthen market access, dietary diversity, and health systems [46].

This study compared four ISABU-recommended iron-biofortified bean varieties with four popular market-identified varieties, before and after one cultivation cycle, across four soil types in Butezi. The initial ISABU advantage in iron, more than twice that of the market-identified varieties, decreased by an average of 51.3% across the sixteen variety-by-soil combinations after cultivation (57.9% when computed from the overall group means), falling below market levels.

Soil type was more strongly associated with this pattern than variety identity: clayey soils showed the best retention and the only net gain (Kaneza), while stony soils, despite high total iron and phosphorus content, showed the lowest retention. This pattern is consistent with a possible role of phosphorus in limiting iron availability, though this was not directly tested in the present study and each soil type was represented by only one hill.

No variety proved universally optimal: Kaneza showed the most consistent iron retention, with a net gain on clayey soil; Mugwiza showed the sharpest decline; and Musore and Rufutamadeni showed intermediate, soil-dependent profiles.

These patterns indicate that soil type, rather than variety choice alone, is a factor in whether the genetic iron advantage of ISABU-recommended varieties is expressed at harvest. On clayey soil, such as at Rubaragaza, Kaneza showed the best iron retention in this pilot trial and warrants further, better-replicated testing; on phosphorus-rich stony soil, such as at Sorero, none of the four ISABU varieties retained a majority of their initial iron content. Because this study did not test soil amendments or fertilizer regimens, we do not make specific fertilizer recommendations here; rather, we recommend that future, adequately replicated and randomized trials evaluate whether targeted soil management (for example, addressing phosphorus availability on stony soils) can improve iron retention. A soil-differentiated approach to variety promotion, together with strengthening the formal seed system, should guide further investigation of iron-biofortified beans in Butezi, complemented by additional multiseason, replicated trials.

Belyse, M. conceived and designed the study, participated in field data collection, performed the statistical analysis, interpreted the results, and drafted the manuscript. Mbazi, S., Juma, K., Sayoki, M., Sévérin, S., Caleb S. and Jonathan, N. supervised the study throughout its conduct, including the quality of data collection and analysis, and critically reviewed and revised the manuscript for important intellectual and scientific content. All authors read and approved the final version of the manuscript.