Rising atmospheric CO₂ increases global threat of zinc deficiency

Summary

Background

Increasing concentrations of atmospheric carbon dioxide (CO₂) lower the content of zinc and other nutrients in important food crops. Zinc deficiency is currently responsible for large burdens of disease globally, and those populations who are at highest risk for zinc deficiency also receive most of their dietary zinc from crops. By modeling dietary intake of bioavailable zinc for the populations of 188 countries under both an ambient CO₂ and elevated CO₂ scenario, we sought to estimate the impact of anthropogenic CO₂ emissions on the global risk of zinc deficiency.

Methods

We used established methods of estimating per capita per day bioavailable intake of zinc for the populations of 188 countries at ambient CO₂ concentrations. We then modeled zinc intake at elevated and ambient CO₂ concentrations and estimated the risk for inadequate zinc intake among the populations of different nations under the two scenarios (ambient and elevated CO₂) by calculating population-weighted estimated average requirements (EARs), and using the EAR cut-point based method. We then compared the size of the populations at risk under the two scenarios in each country.

Findings

Anthropogenic emissions of CO₂ are likely to place between 132 million—180 million people at new risk of zinc deficiency by around 2050. The people likely to be most affected live in Africa and South Asia, with nearly 48 million residing in India alone. Global maps of increased risk show significant heterogeneity and can be used to help guide interventions aimed at reducing this vulnerability.

Interpretation

Our results indicate that one heretofore unquantified human health impact associated with anthropogenic CO₂ emissions will be a significant increase in the human population at risk for zinc deficiency.

Introduction

Adequate zinc nutrition is a cornerstone of global maternal and child health, and roughly 17% of the global population was estimated to be at risk of zinc deficiency in 2011^1,2. Zinc deficiency increases the risk of premature delivery and reduces growth and weight gain in infants and young children³. Adequate zinc intake is also important in proper immune function.⁴ Preventive zinc supplementation in zinc deficient populations decreases morbidity from childhood diarrhea, acute lower respiratory infections, and all-cause mortality^5–8 The global burden of disease attributed to zinc deficiency is high with greater than 100,000 deaths per year from diarrhea and pneumonia in children under 5 years of age attributable to zinc deficiency.⁹

Food crops, which serve as an important source of dietary zinc for billions of people around the world, have recently been shown to contain lower concentrations of zinc and other nutrients when grown at elevated concentrations of carbon dioxide (heretofor referred to as [CO₂])¹⁰. When grown under open field conditions at [CO₂] the world is expected to experience by 2050¹¹ (roughly 550 p.p.m.), wheat (−9.1%), rice (−3.1%), barley (−13.6%), soy (−5.0%) and field peas (−6.8%) have significantly reduced zinc content.

In order to understand the global health implications of these changes, we modeled the per capita availability of dietary zinc and phytate (a phosphate storage molecule that inhibits absorption of zinc in the diet) for 188 countries under both ambient and elevated [CO₂] scenarios. For each country, under each scenario, we calculated the percent of the population at risk for inadequate zinc intake. By comparing the results for each scenario, we are able to estimate the percent of each national population that would be placed at new risk of inadequate zinc intake as a result of rising [CO₂] in the atmosphere. We are also able to identify geographic regions where populations are particularly vulnerable to the nutritional impacts of increasing [CO₂] as a result of their diets and their overall zinc and phytate intake.

Methods

Study Design

Estimation of the effects of [CO₂] on zinc and phytate content of plant source foods

To estimate the size of the effect of elevated [CO₂] on zinc and phytate concentrations, we performed a meta-analysis using previously published data, pooled by crop, from free air carbon dioxide enrichment (FACE) and open-top chamber (OTC) experiments in which crops were grown at ambient and elevated [CO₂] and the edible portion of the food crop was tested for zinc and/or phytate¹⁰. Phytate concentrations were only found to change significantly in wheat (P < 0.05) in response to elevated [CO₂] so only wheat phytate concentrations were adjusted in our scenarios. Estimates of the effect of elevated [CO₂] on nutrient content were used to adjust per capita nutrient intake from each food commodity as described below.

Estimation of the adequacy of zinc in national food supplies based on national food balance data

National food balance sheet (FBS) data from 2003–2007, which are available from the Food and Agriculture Organization of the United Nations (FAO)¹², were analyzed to estimate country-specific, per capita zinc and phytate intake under both ambient (375–384 p.p.m. during this time period)¹³ and elevated (roughly 550 p.p.m.) [CO₂] scenarios. We estimated the prevalence of inadequate zinc intake under each scenario by comparing the estimated absorbable zinc content of the national food supply with the population’s estimated physiological requirements for absorbed zinc. Detailed methodological and model assumptions, as well as results based on original data, have been described previously^{2, 14, 15}. We take risk of inadequate zinc intake based on FBS analysis as a proxy for risk of zinc deficiency, and, for clarity,we use the term “risk of zinc deficiency” instead of “risk of inadequate zinc intake” in our results and discussion.

The FAO FBS provide estimated country-specific data for 210 countries/areas on the average daily per capita consumption of 95 ‘standardized’ food commodities (kcal/capita/d). Of these 210 countries/areas, we could obtain demographic data for 188 countries which became the subjects of our analysis. The remaining 22 territories and states, mostly small island entities, were excluded from further analyses as we restricted our analyses to national populations. Plant-source food commodities reported in the food balance sheets were initially categorized as C₃ legumes, C₃ tubers, C₃ other plants, or C₄ grasses. If crop-specific data on the effects of elevated [CO₂] on changes in the zinc and phytate contents were available from FACE and/or OTC experiments, commodities were subsequently assigned to these “primary” groups (maize, peas, rice, sorghum, soy, wheat, barley and potatoes). When crop-specific data were not available, commodities were assigned to one of three “composite” groups, which were composed of the weighted means of crop-specific data from the FACE and/or OTC experiments (C₃ plants: wheat and barley, with and without rice; C₃ legumes: soy and peas; C₄ grasses: corn and sorghum). Tubers other than potatoes were assumed to be closest to potatoes and the values for potatoes were used for those crops. Because rice is grown under very different (immersion) conditions than other C₃ plants, it was not clear whether it should be included in our composite estimates of the mean effect of elevated [CO₂] on C₃ plants. To address this uncertainty, we generated two different models for adjusting zinc and phytate contents of the FAO FBS food commodities, one which included rice in the C₃ plant composite estimates and one which did not. The assignment of each food commodity is listed in Table S1. The “best-estimate” model was considered to be that which did not include rice in the C₃ weighted mean.

The FAO FBS country-specific data on food availability (kcal/capita/d) were used to calculate the per capita zinc and phytate contents of the daily food supply (mg/capita/d), prior to accounting for the effects of elevated [CO₂]. The estimated zinc and phytate content (mg/100kcal) for each food commodity at ambient [CO₂] was obtained from a composite nutrient composition database created specifically for the analysis of the zinc and phytate content of national food supplies as reported¹⁴. Mean estimated zinc and phytate content (mg/100kcal) for each food commodity was calculated, adjusting for the effects of food processing methods (e.g. decortication, milling, fermentation and nixtamalization), according to previous regional assumptions^2,14. The mean per capita zinc and phytate intakes for each country were calculated as the sum of the zinc and phytate contribution from each food commodity. To model physiological zinc intake at elevated [CO₂], we used Monte Carlo simulations to account for uncertainties. In the simulation, 1000 random draws were made from the established range of altered nutrient concentrations for each food crop/commodity to generate confidence intervals (CI) which reflect the precision of the estimates free of distributional assumptions.

The fractional absorption of zinc and the absorbable zinc content of the daily food supply for each country under both ambient and elevated [CO₂] scenarios were predicted using a saturation response model of zinc absorption as a function of dietary zinc and phytate¹⁶. The age and sex distribution of country populations (IHME University of Washington; 2010 Revision of the World Population Prospects¹⁷) were used to calculate the country-specific theoretical mean daily per capita physiological requirement for zinc, as developed by the International Zinc Nutrition Consultative Group (IZiNCG)¹. Because there is a lack of consensus on physiological requirements for zinc, we generated an additional model using the effect estimate of elevated [CO₂] on nutrient levels from the best-estimate model, but with country-specific theoretical mean daily per capita physiological requirements for zinc based on the Food and Nutrition Board of the United States Institute of Medicine (FNB/IOM) recommendations¹⁸. For each scenario, we calculated the estimated percentage of the mean physiological requirement for zinc available in the national food supply by dividing the estimated absorbable zinc content of the national food supply by the calculated national physiological requirement (IZiNCG or FNB/IOM). We then applied an estimated average requirement (EAR) cut-point based method to estimate the percentage of national populations at risk for inadequate zinc intake, assuming a normal population distribution with a 25% inter-individual variation as has been done in previous analyses of global risk of zinc deficiency.^19,20 Variation in dietary zinc requirements takes into account both variation in requirements for absorbed zinc (i.e. variations in metabolism and rate of zinc turnover) as well as variation in the fractional absorption of zinc. We took the difference between the population at risk at elevated [CO₂] and the population at risk under ambient [CO₂] as our measure of impact.

Statistical analyses

Regional classifications are based on the reporting regions of the Global Burden of Diseases, Injuries, and Risk Factors 2010 Study²¹ with the exception that we broke out India and China separately because of their large population sizes. Regional and global data were weighted by national population sizes. All statistical analyses were completed using SAS System for Windows release 9.3 (SAS Institute, Cary, North Carolina) and the R statistical package V 3.0. Data are presented as means (95% CI), unless otherwise noted.

Results

We found that rising [CO₂] threatens populations around the world with a significant increase in the risk of zinc deficiency. Specifically, we found that concentrations of CO₂ anticipated by 2050 would be likely to place roughly 138 (± 18) million people newly at risk for zinc deficiency and that those in Africa and parts of Asia will be most affected (Figure 1). Globally, rising [CO2] threatens to increase the percent of the human population at risk for zinc deficiency by 2.1% while increasing risk in sub-Saharan Africa, South Asia, and India by 3.9%, 2.9% and 4.2% respectively (Table 1). We found the population of India particularly vulnerable to the impact of elevated [CO₂] on crop nutrients with nearly 48 million additional people estimated to be newly at risk for zinc deficiency.

Figure 1.

Global map showing the absolute percentage increase in the risk of zinc deficiency in response to elevated atmospheric [CO₂].

Table 1.

Altered risk of zinc deficiency ± CI under elevated [CO₂] (550 p.p.m.) scenario, calculated regionally.

VARIABLE	High- income	Southern and Tropical L. America	Central and Andean L. America and Carib.	Central and Eastern Europe	Central Asia, North Africa and Middle East	Sub- Saharan Africa	South Asia	India	East and Southeast Asia and Pacific	China	Global
Number of Countries	30	5	27	20	28	48	5	1	21	3	188
Population (millions)	937.2	249.4	301.4	330.1	481.4	757.8	355.0	1140.5	606.8	1337.7	6497.5
Energy (kcal/day)	3423.9	3031.4	2835.7	3285.8	3089.2	2350.5	2233.7	2295.2	2585.2	2905.0	2776.2
Zinc (mg/d)	12.8 (12.7, 12.9)	11.9 (11.8, 12.0)	10.3 (10.2, 10.4)	11.2 (11.1, 11.3)	13.1 (13.0, 13.2)	8.0 (7.9, 8.1)	9.0 (8.9, 9.1)	9.3 (9.1, 9.5)	8.6 (8.5, 8.7)	13.2 (13.0, 13.4)	10.9 (10.8, 11.1)
Phytate (mg/d)	1162.2 (1158.8, 1165.6)	1162.0 (1157.4, 1166.6)	1881.7 (1880.1, 1883.3)	1183.0 (1176.1, 1189.9)	2702.7 (2680.3, 2725.1)	1777.3 (1776.3, 1778.3)	1981.2 (1953.4, 2009.0)	2286.4 (2248.5, 2324.3)	1436.0 (1434.6, 1437.4)	1440.0 (1427.0, 1453.0)	1707.3 (1699.8, 1714.8)
Absorbable zinc (mg/d)	3.27 (3.26, 3.28)	3.17 (3.16, 3.18)	2.47 (2.45, 2.49)	3.04 (3.03, 3.05)	2.48 (2.47, 2.49)	2.10 (2.09, 2.11)	2.18 (2.17, 2.19)	2.11 (2.09, 2.13)	2.41 (2.40, 2.42)	3.12 (3.10, 3.14)	2.65 (2.64, 2.66)
% mean physiological requirement	158.6 (158.3, 158.9)	162.2 (161.5, 162.9)	129.9 (129.0, 130.8)	146.2 (145.7, 146.8)	129.4 (129.1, 129.7)	118.9 (118.1, 119.8)	118.2 (117.6, 118.8)	110.3 (109.2, 111.4)	124.4 (123.7, 125.1)	153.2 (152.4, 154.0)	135 (134.7, 135.3)
Estimated % of pop. with inadequate zinc intake	8.1(8.0, 8.2)	6.9 (6.7, 7.1)	19 (18.4, 19.6)	10.6 (10.4, 10.8)	19.6 (19.4, 19.8)	29.5 (28.6, 30.4)	27.3 (26.6, 28.0)	35.4 (34.0, 36.8)	24.6 (23.8. 25.4)	8.4 (8.2, 8.6)	19.5 (19.2, 19.8)
Absolute % increase in population with inadequate zinc intake	0.6 (0.5, 0.7)	0.5 (0.3, 0.6)	2.0 (1.4, 2.6)	1.0 (.8, 1.2)	2.5 (2.3, 2.7)	3.9 (3.0, 4.8)	2.9 (2.2, 3.6)	4.2 (2.8, 5.6)	2.4 (1.6, 3.2)	0.6 (0.4, 0.8)	2.1 (1.8, 2.3)
Relative % increase in population with inadequate zinc intake	6.9 (6.0, 7.8)	6.5 (4.1, 8.9)	11.1 (7.8, 14.4)	10.2 (8.2, 12.2)	14.6 (13.2, 16.0)	14.9 (11.6, 18.2)	12.2 (9.8, 14.6)	13.4 (9.0, 17.8)	10.4 (7.8, 13.0)	7.3 (4.6, 10.0)	10.6 (9.5, 11.7)
Population newly at risk of inadequate zinc intake (millions)	5.4 (4.5, 6.2)	1.1 (0.7, 1.6)	5.9 (4.2, 7.7)	3.3 (2.8, 3.9)	12.2 (11.1, 13.2)	29.7 (22.6, 36.8)	10.4 (8.0, 12.8)	47.8 (32.1, 63.5)	14.8 (10.1, 19.5)	7.8 (5.1, 10.5)	138.4 (120.0, 156.8)

Data are presented first for high-income countries, and then by region with China and India broken out separately because of their population size. Data are weighted by national population sizes and are for 188 countries. Regional classifications are based on the reporting regions of the Global Burden of Diseases, Injuries, and Risk Factors 2010 Study, and are grouped according to geographical location and dietary patterns. Absorbable zinc is mathematically related to dietary phytate and zinc according to the Miller equation.

Our results are robust regardless of which model we used to estimate the change in prevalence of risk for zinc deficiency. While the choice of FNB/IOM versus the IZiNCG estimates of the physiological requirements for zinc intake has very large implications for the percentage of the global population at risk for zinc deficiency at ambient [CO₂] (66% using the FNB/IOM model, and 17% using the IZiNCG model¹⁴), (Table S2), the estimated impact of elevated [CO₂] on global risk of zinc deficiency using these two models varied much less. The total number of people estimated to be placed at new risk for zinc deficiency using the IOM model was 180 (± 16) million, while the IZiNCG based model estimated 138 (± 18) million. Both models estimated the absolute increase in the percent of the global population at risk under elevated [CO₂] as 2–3%.

In addition, the decision of whether or not to include rice in the composite index for C₃ grains did not have a large effect on our results. With rice included in the composite index, we estimate 133 (± 18) million people at new risk for zinc deficiency globally (Table S3) compared with 138 (± 18) million when rice is not included in the composite index. Nor does the inclusion or exclusion of rice from the composite index or the choice of which physiological requirements to use alter appreciably the order of countries most impacted by rising [CO₂] (Table S4, Figure S1).

We chose to follow IZiNCG recommendations with respect to the best model for estimating risk of zinc deficiency because these estimates of physiological zinc requirements are based on a larger number of studies conducted among both men and women and including studies in both developed and developing countries; they are intended to be generalizable internationally. We also chose, for our “best estimate” model, to use the model that does not include rice in calculating the effect size of elevated [CO₂] on C₃ grains. Because rice is grown primarily under immersion conditions that are substantially different from other grains, we believed its inclusion could be misleading. As discussed above, however, these decisions did not affect our conclusions in a meaningful way.

Discussion

The global [CO₂] in the atmosphere is expected to reach 550 p.p.m. in the next 40–60 years, even if further actions are taken to decrease emissions¹¹. These concentrations of CO₂ have been shown to reduce the nutritional value of important food crops¹⁰, and here we show that such nutrient reductions threaten an additional 138 million people concentrated in Africa and South Asia with an increased risk of zinc deficiency. Zinc deficiency has been consistently shown to be associated with compromised immune function and increased susceptibility to morbidity and mortality from infectious diseases.^6,22

Our analysis does not include the changes in the global diet that will almost certainly take place over the next few decades while global [CO₂] rises to 550 p.p.m. We have tried to isolate the CO₂ effect by simply modeling a world in which food availability is the same as in 2010 but the nutrient contents of those foods have changed in response to elevated [CO₂]. Anticipating how the global diet is likely to change over the coming decades is murky business. Economic growth may allow populations to consume more calories or receive a higher percentage of their calories from animal source foods. However, it is also estimated that agricultural production globally will have to roughly double by 2050 in order to keep up with increased demand²³, and the combination of water scarcity, arable land degradation, and climate change represent very significant headwinds to such increases in production²⁴. Because of this complexity, we believe the simplest approach is to model diets that are unchanged with respect to calories and composition, an achievement that many would consider optimistic in the fact of rapidly changing environmental conditions.

We have also made no attempt to account for population growth in our analysis. The human population is expected to rise to between 9 and 10 billion by 2050²⁵, but we have used 2010 estimates of population size for our analysis. This means that our numbers of individuals likely to be placed at risk for zinc deficiency are almost certainly a considerable underestimate. A simple scaling of population growth to the effect we have measured would lead us to conclude that, in fact, 187 million people (using 9.5 billion as an estimate of the 2050 global population) are likely to become newly zinc deficient as a result of increased [CO₂]. And the fact that most of this population growth is expected to occur in the regions that are disproportionately impacted by the nutritional consequences of rising [CO₂] suggests that even this number is likely an underestimate. However, in order to maintain the most transparent analysis possible with the fewest assumptions, we have not attempted to project these demographic changes but believe that our results are a conservative estimate.

One assumption we do make for this study is that impacts of elevated [CO₂] on crop nutrients that have been quantified in developed country settings for a subset of crops and cultivars consumed globally can be generalized to estimate nutrient intakes around the world. Of course, to be certain of the nutritional impacts of elevated [CO₂] on the global population we would need to conduct FACE experiments for every consumed cultivar of every food crop in every country—an undertaking that is not feasible. But we are reassured that having found a very similar pattern of effects across 41 different cultivars of six different food crops grown on three continents in seven locations over ten years under vastly different growing conditions,¹⁰ that these nutrient changes are a robust finding and are likely to be similar across the different growing conditions around the world. This assumption was also recently supported by a broad meta-analysis showing similar changes in the nutrient content of a diverse number of plants across many plant tissues and many locations.²⁶

An additional conservative assumption embedded in this analysis is that food availability in populations around the world is in proportion to physiological requirements. Children under 5 and woman (especially during pregnancy) are likely to be at an increased risk of zinc deficiency due to increased nutrient requirements. The assumed optimal distribution of foods is unlikely to be met in most settings, but in the absence of global data on food distributions, an assumption must be made, and this assumption is the most conservative approach. Less optimal food distributions would lead to increased impacts of elevated [CO₂] on risk of zinc deficiency.

Finally, we have assumed that there is no change in the zinc content of animal source foods. There is clear evidence that most plants, not just food crops, have lower concentrations of zinc when grown at elevated [CO₂]. Meta-analyses of plants that include many different tissues from a variety of grasses, trees, and shrubs show consistent reductions in zinc content²⁶, making it likely that animal forage would have reduced zinc content in a world experiencing higher atmospheric [CO₂]. However, there are no data available on how these changes in the nutrient content of forage might alter the concentrations of zinc in animal source foods like meat, milk, or eggs. Until such data are available, we can only assume no change in nutrient concentrations, but this, too, is likely to lead to underestimates of the impact of rising [CO₂] on risk of zinc deficiency.

The effect we have identified highlights an issue of social justice. Wealthier people are associated with higher CO₂ emissions²⁷ while the people who are most vulnerable to the nutritional impacts of rising [CO₂] are those who receive the smallest percentage of their dietary zinc from animal source foods (Figure 2). These tend to be the poorest people within a country or region. The wealthy world’s CO₂ emissions are putting the poor in harm’s way.

Figure 2.

As percentage of dietary zinc received from animal-source foods (ASF) rises, the risk of new zinc deficiency in response to elevated [CO₂] falls. These data are derived from analysis of the food balance sheets. The percentage of dietary zinc availability from animal source foods (ASF; e.g. meat, fish, milk, etc.) was calculated by dividing the amount of per capita daily bioavailable zinc from ASF by total per capita daily bioavailable zinc estimates. The % increased risk of zinc deficiency was calculated as described in the methods section of this manuscript. Zinc content in animal source foods is assumed to remain unchanged.

By modeling national data for 188 countries, we identify populations who are at highest risk of increased zinc deficiency as a consequence of rising [CO₂]. These populations could be the target of interventions designed to address this risk. Such interventions might include zinc supplementation, fortification of staple foods with additional zinc, the application of zinc containing fertilizers to crops, and the development and introduction of bio-fortified crop strains such as rice and wheat. Earlier work has also shown that, at least for rice, different cultivars of a crop show different levels of sensitivity to the [CO₂] impact on zinc content which may provide an opportunity for breeding crop cultivars with lower nutritional sensitivity to rising [CO₂]¹⁰.

Anthropogenic change to Earth’s natural systems will impact human health in multiple ways through pathways that are often quite complex²⁸. Here we describe one such pathway that would have been challenging to anticipate in advance of the experimental data. We suspect that there will be others as human transformation of natural systems becomes increasingly profound and pervasive.

Research in Context.

Evidence before this study

Prior to this study, there was strong evidence that zinc deficiency was a significant global health problem affecting at least 17% of the global population^1,2 and responsible for large burdens of disease around the world.^5–9 More recently, strong evidence has emerged from free air carbon dioxide enrichment (FACE) experiments that the edible portion of food crops grown at elevated atmospheric [CO₂] have lower zinc, iron, and protein contents than identical cultivars of the same crops grown under identical growing conditions at ambient [CO₂] concentrations.¹⁰ These same experiments showed that phytate concentrations were lower in wheat cultivars grown at elevated [CO₂] but phytate content in other food crops was unaffected.¹⁰

Added value of this study

This is the first study to combine data on nutrient changes in food crops anticipated at higher levels of atmospheric CO₂ with estimates of dietary intake of most of the world’s population in order to model the effect of rising [CO₂] on the global risk of zinc deficiency.

Implications of all the available evidence

The study indicates that, in addition to disrupting the global climate system, anthropogenic CO₂ emissions are also threatening millions of people with increased risk of zinc deficiency. Future studies will evaluate the impact of rising [CO₂] on the adequacy of global iron and protein intake. This study also indicates that the distribution of the populations at risk for increased zinc deficiency is heterogeneous and concentrated in Africa and South Asia. From a policy perspective, these findings suggest that interventions including biofortification of staple food crops, supplementation, and fortification should be targeted at those populations identified as most vulnerable. In addition, these finding provide additional support for the urgent need to mitigate global CO₂ emissions.