Aflatoxin Regulations in a Network of Global Maize Trade

Felicia Wu1*, Hasan Guclu2

1 Department of Environmental and Occupational Health, University of Pittsburgh, Pittsburgh, Pennsylvania, United States of America, 2 Department of Biostatistics and
Public Health Dynamics Laboratory, University of Pittsburgh, Pittsburgh, Pennsylvania, United States of America

Abstract
Worldwide, food supplies often contain unavoidable contaminants, many of which adversely affect health and hence are
subject to regulations of maximum tolerable levels in food. These regulations differ from nation to nation, and may affect
patterns of food trade. We soughtto determine whether there is an association between nations’ food safety regulations
and global food trade patterns, with implications for public health and policymaking. We developed a network model of
maize trade around the world. From maize import/export data for 217 nations from 2000–2009, we calculated basic
statistics on volumes of trade; then examined how regulations of aflatoxin, a common contaminant of maize, are similar or
different between pairs of nations engaging in significant amounts of maize trade. Globally, market segregation appears to
occur among clusters of nations. The United States is at the center of one cluster; European countries make up another
cluster with hardly any maize trade with the US; and Argentina, Brazil, and China export maize all over the world. Pairs of
nations trading large amounts of maize have very similar aflatoxin regulations: nations with strict standards tend to trade
maize with each other, while nations with more relaxed standards tend to trade maize with each other. Rarely among the
top pairs of maize-trading nations do total aflatoxin standards (standards based on the sum of the levels of aflatoxins B1, B2,
G1, and G2) differ by more than 5 mg/kg. These results suggest that, globally, separate maize trading communities emerge;
and nations tend to trade with other nations that have very similar food safety standards.

Citation: Wu F, Guclu H (2012) Aflatoxin Regulations in a Network of Global Maize Trade. PLoS ONE 7(9): e45151. doi:10.1371/journal.pone.0045151
Editor: Alejandro Raul Hernandez Montoya, Universidad Veracruzana, Mexico
Received March 29, 2012; Accepted August 15, 2012; Published September 25, 2012
Copyright: ß 2012 Wu, Guclu. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits
unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This work was funded by the National Cancer Institute (NCI) of the National Institutes of Health (NIH), Grant No. 5R01CA153073-2. The content is solely
the responsibility of the authors and does not necessarily represent the official views of NCI or NIH. The funders had no role in study design, data collection and
analysis, decision to publish, or preparation of the manuscript.
Competing Interests: The authors have declared that no competing interests exist.
* E-mail: [email protected]

Introduction peanuts, tree nuts (almonds, pistachios, hazelnuts, pecans, and

Brazil nuts), and milk.
Multiple nations worldwide have set food safety standards for Aflatoxin B1, the most toxic aflatoxin, is the most potent
maximum tolerable levels of certain contaminants in food, for the naturally occurring chemical liver carcinogen known. The risk of
purpose of protecting public health. These standards, in turn, have liver cancer in individuals exposed to chronic hepatitis B virus
important impacts on world food trade. Of interest is whether (HBV) infection and aflatoxin may be up to 30 times greater than
there is an association between nations’ food safety regulations and the risk in individuals exposed to HBV of aflatoxin alone [1].
global food trade patterns. For example, do nations with strict food Acute aflatoxicosis, causing severe gastrointestinal symptoms and
safety standards tend to trade more with each other, while nations often death, results from high aflatoxin doses. In recent years,
with more relaxed food safety standards also tend to trade with hundreds of aflatoxicosis cases in Africa have resulted from
each other? Do the main food exporters tend to have stricter or
consumption of contaminated maize [2]. Aflatoxin exposure may
more relaxed food safety standards themselves?
also be associated with stunting in children [3,4] and immuno-
We are interested in these questions from the perspective of how
suppression [5]. Recently, aflatoxin exposure has been associated
food safety regulations that national governments impose may or
with liver cirrhosis; aflatoxin and HBV exposure may synergize to
may not have impacts on a global level, and what the implications
substantially increase cirrhosis risk [6].
might be for food safety and policy decision-making. Hence, we
Currently, over 5 billion people worldwide are at risk of chronic
developed a network model to represent the global trade patterns
exposure to aflatoxin in food [2]. Maize is one of the main sources
of maize from nation to nation. We examined these maize trade
of human exposure to aflatoxin, because it is highly consumed
patterns as function ofaflatoxin regulations in maize, with a focus
worldwide and unfortunately is also one of the most susceptible
on regulations for maize intended for human consumption, for
crops to aflatoxin contamination [2,7].
each nation.

Aflatoxin: Background Aflatoxin regulations worldwide and their potential

Aflatoxins are secondary metabolites of the common foodborne impacts
fungi Aspergillusflavus and A. parasiticus, which colonize crops in Over 100 nations have set regulatory limits on allowable
tropical and subtropical regions worldwide. These fungi can also aflatoxin levels in human food or animal feed. Most of these
produce aflatoxin in storage, transportation, and food processing. nations regulate the sum of the levels of the four most prominent
Aflatoxin contamination primarily occurs in maize, spices, types of aflatoxins in food: B1, B2, G1, and G2. Hence, in our

PLOS ONE | www.plosone.org 1 September 2012 | Volume 7 | Issue 9 | e45151

 Aflatoxin Regulations in a Network of Food Trade

paper, ‘‘aflatoxin’’ is meant to refer to the sum of these aflatoxins encourage exportation of their best crops to preserve export
unless otherwise specified. markets. The poor-quality crops would be left for domestic
In industrial nations, aflatoxin contamination in food primarily consumption, inadvertently increasing liver cancer risk among
inflicts economic rather than health burdens. It reduces the price HBV-infected populations [9].
paid for crops, and can cause disposal of large amounts of food. The other aspect to former UN Secretary-General Annan’s
Losses from aflatoxin in the US – in the hundreds of millions USD statementconcerns the purported adverse economic impacts to
annually – are associated with market loss rather than health Africa of attempting to trade with the European Union. This
effects [9], as enforcement of aflatoxin standards and aflatox- statement about economic loss was based on estimates in Otsuki et
incontrol methods have largely eliminated harmful exposures in al. [12], who developed an economic model of expected aflatoxin
food. In low-income nations, however, health impacts of aflatoxin contamination in African crops and how much of their export
are more severe. Many individuals are not only malnourished but market would thus be lost. But are African nations in fact trading
also chronically exposed to high aflatoxin levels primarily through much food with the EU at all? Or do these nations as well as other
the staple foods of maize and peanuts, resulting in deaths from food-producing nations worldwide tend to export more of their
aflatoxicosis and liver cancer. Low-income nations often lack the food to nations that have more relaxed food safety standards?
resources, technology, and infrastructure necessary for routine According to a 2005 World Bank report [13], the losses suffered by
food monitoring and aflatoxin control. Aflatoxin exposuresare African nations attempting to export foods to the EU was not
typically highest in sub-Saharan African and Asian nations [7]. nearly as severe as had been predicted in [12], and the African
Further complicating the problem is that for a given level of shares for certain foods (dried fruits) actually increased. Wu [9]
aflatoxin exposure, cancer risk is more severe in low-income also estimated a much lower loss to African nations from the EU
nations than in the industrial world because of higher HBV aflatoxin regulations, and hypothesized that this was because food
prevalence [1]. trade patterns between Africa and the EU were not on such a large
Globalization of food trade has exacerbated aflatoxin-related scale as estimated by [12].
losses in three unfortunate ways: These food trade patterns will have a very important impact on
the sustainability of nations that rely upon food exports for their
1. Strict aflatoxin standards mean that many nations will export market economies, and may also have impacts for global public
their best-quality foods and keep contaminated foods domes- health. It is for the purpose of answering these questions, and
tically, resulting in higher aflatoxin exposure in low- or middle- exploring the nature of food safety regulations and potential
income nations where hepatitis prevalence is high. impacts on food trade and global health, that we have developed a
2. Even the best-quality foods produced in some nations may be social network model of global maize trade to facilitate under-
rejected for export because of aflatoxin levels exceeding the standing of this association.
tolerable limit, resulting in millions of dollars in losses.
3. The cost of a rejected food shipment is substantial (about Social networkmodels and their applications
$10,000 per lot in demurrage fees, [8]), even if the lot can be Social network models have been used in public health research
returned to the country attempting to export. to explain and predict a variety of phenomena, includingthe
spread of infectious disease and how to control that spread [14],
These dilemmas led former United Nations(UN) Secretary- how to control disease spread [15], and patterns of obesity and
General Kofi Annan to recognize the magnitude of the problem of smoking prevalence in social circles of friends, family members, co-
setting appropriate aflatoxin standards worldwide. He comment- workers, and acquaintances ([16,17], see also [18]). In the field of
ed, ‘‘The EU [European Union] regulation on aflatoxins costs mycotoxins, a network analysis has been done on contaminants
Africa $670 million each year in exports. And what does it reported through the EU Rapid Alert System for Food and Feed
achieve? It may possibly save the life of one citizen of the EU every (RASFF) by nation [19]. However, the history of social network
two years. Surely a more reasonable balance can be found’’ [9]. modeling extends as far back as the early 1900s, when
Annan had based his statement upon the report of the Joint FAO mathematical models described malaria transmission and a
(Food and Agriculture Organization)/World Health Organization threshold level for the Anopheles mosquitoes that transmit the
(WHO) Expert Committee on Food Additives’ 49th meeting on disease [20]. Is it pointed out in [21] that social network models
aflatoxin, which assessed the effect of aflatoxin regulations on liver enable us to understand behaviors at both individual and
cancer depending on HBV prevalence [10,11]. Joint FAO/WHO population/global levels in a way that simple random sampling
Expert Committee on Food Additives (JECFA) developed two cannot do, because random sampling removes individuals (in the
scenarios, to determine the effect of moving from an enforced case of this work, individual nations) from the social context that
aflatoxin standard of 20 mg/kg(or 20 ng/g) to 10 mg/kg, in two may influence their behavior.
hypothetical nations: one with HBV prevalence of only 1%, and
another with HBV prevalence of 25%. In the first nation, Methods
tightening the aflatoxin standard yielded a drop in the estimated
population risk of 2 additional cancers per year per billion people. We developed a social network model of world food trade,
In the second nation, tightening the aflatoxin standard for this focusing on maize because of its importance to populations’ diets
population yielded a drop in the estimated population risk of 300 worldwide, large volume of trade worldwide, and propensity to be
additional cancers per year per billion people. contaminated with aflatoxin. Each nation is represented as an
Hence, in rich food-importing nations with low HBV preva- individual node or ‘‘actor’’ in the model, connected to each other
lence, tightening the aflatoxin standard would reduce cancer risk in pairs by flows of imported and exported maize. To quantify the
by an amount so small as to be undetectable by epidemiological relative importance of maize trade patterns between nations, we
methods. But food-exporting regions with high HBV incidence – used the United Nations Commodity Trade Statistics Database
China, Southeast Asia, and Africa – could have greater health risk (UN Comtrade, comtrade.un.org), gathering country codes and
due to stringent aflatoxin standards. Until aflatoxin control analyzing total maize exports and imports from and to each of
methods become available and affordable, strict standards would these nations (to and from every other nation) for each year from

PLOS ONE | www.plosone.org 2 September 2012 | Volume 7 | Issue 9 | e45151

 Aflatoxin Regulations in a Network of Food Trade

2000 to 2009. We summedmaize trade data (exports and imports) maize worldwide, with over half a billion metric tons of maize
on a nation-by-nation basis for the years 2000 to 2009. exported to other nations in the past decade. This amount exceeds
Then we converted these global maize trade data into a the next largest exporter’s total trade amount by over four-fold.
weighted and directed network model in the software program Argentina and Brazil, two neighboring South American nations,
PajekTM [22], in which each nation is represented as a node, and are also large exporters. China is the largest Asian maize-exporting
the edges (or lines) are export/import connections between the nation, while France and Hungary, despite their relatively smaller
countries with weights (represented by thickness of the arrows in geographic size compared with other main exporters, also export
the network diagram) equal to amount of maize traded. The size of large quantities of maize. South Africa is the largest African maize-
each node is proportional to the square root of the total amount of exporting nation. Among countries that import maize, Japan is the
maize exportedby that nation from 2000 to 2009, for ease of largest maize-importing nation. The Republic of Korea, Mexico,
visualization. The direction of the edges, or arrows, denotes the Egypt, Taiwan, and Spain have also imported large quantities of
direction of maize trade: each arrow between two nations emerges maize over the last decade.
from the nation exporting the maize, and points to the nation An interesting facet of these trade statistics is that several
importing the maize. The ‘‘distance’’ between all trading pairs of countries that are among the top 20 maize exporters are also
nations is then minimized in the program by using a force-based among the top 20 maize importers. These include the United
layout so that the network representation reveals clusters: groups of States, Canada, The Netherlands, Mexico, and Germany. This
nations that tend to trade large amounts of maize amongst each may be for a variety of reasons, including that imported and
other, which appear to form a sub-network within the larger exported maize may serve different destinations (e.g., food or feed),
network. may be needed for different purposes at different times of the year,
We also calculated the degree (of connectedness) of each node and may reflect policies of individual agreements among grain
(e.g., of each nation). The degree is the number of edges connected companies in different nations. In fact, the United States and
to the node. In the network model of maize exports, the degree Canada have established a trade relationship in which they each
(rather out-degree) represents the number of other nations with have exported and imported large amounts of maize to each other
which one nation has exported any maize from 2000–2009. Degrees over the last decade.
and clustering patterns, taken together, are important in maize Table 2 lists the 20 exporting nations with the highest out-
trade networks to understand how vulnerable (or resilient) nations degree (i.e., the number of nations to which it has exported at least
would be in the event of food shortage elsewhere in the world. If a one consignment of maize from 2000–2009) and the 20 importing
nation that exports maize has a high degree, then any issues that nations with the highest in-degree (i.e., the number of nations from
alter maize availability in that nation would potentially affect which it has imported maize). Not surprisingly, as Tables 1 and 2
many other nations. These other nations, in turn, are even more show, many of the top maize exporters worldwide are also the
vulnerable if they do not regularly import maize from many, if nations that export maize to the largest total number of nations.
any, other nations worldwide. On the other hand, if an importing Again, the United States has the largest degree, exporting maize to
nation has a high degree (rather in-degree), this means that it is 181 different nations worldwide from 2000 to 2009. Likewise,
importing maize from many different nations. Thus, any failure to Argentina, South Africa, France, Canada, and Brazil also export
produce maize at expected levels in one of those nations would not maize to a large number of nations.
necessarily jeopardize maize supplies in the importer. However, although Japan, Korea, and Mexico are the largest
In addition to analyzing which nations export and import maize-importing nations, they are not included in the 20 nations
maize, and with whom, we compiled information about aflatoxin that import maize from the largest number of nations as indicated
regulations in nations worldwide. These regulations were taken by degree (Table 2, right columns). The implication is that they are
from the Food and Agriculture Organization [23] report on importing large amounts of maize from a small number of
mycotoxin regulations worldwide. Because the network model is countries. Also of interest is that many European nations are
specific to maize trade, we only includedaflatoxin standards included in both columns, such as France, Germany, Italy, The
relevant to maize in our database. For example, if a nation such as Netherlands, Spain, and the United Kingdom. They conduct a
Kenya has an aflatoxin standard listed for peanuts but not for substantial amount of maize trade amongst themselves within
maize, then the nation is coded as not having set an aflatoxin Europe.
standard for maize. The social network model of maize trade between and amongst
We examined whether nations trading maize had similar or nations is shown in Figure 1 (export volume represented roughly
dissimilar aflatoxin regulations, to understand the ease or difficulty by size of nodesfor each nation [Wu and Guclu, unpublished
of exporting nations to provide maize that had sufficiently low data]). An edge is drawn between two nations if they had traded
aflatoxin levels. The aflatoxin regulations are the basis for another more than one million MTs total in the years 2000 to 2009. The
network model representation, in which the maize trade patterns direction of the arrows indicates the direction of the maize trade:
are recreated, but thenode sizes are made equal and are color- from an exporting nation to an importing nation.
coded based on the relative strictness of aflatoxin regulations. Figure 1 shows that at least two distinct clusters emerge when
Again, when distances between trading pairs of nations are distances are minimized among all trade partnerships: European
minimized, the proximity of nodes reveals whether nations that nations in one cluster of inter-trade (lower left portion of Figure 1),
have similar aflatoxin standards are clustered together in maize and the United States and other American nations in another such
trading patterns. cluster (upper right portion of Figure 1). Meanwhile, Argentina,
China, and Brazil are at the center of the network: China
Results exporting maize to Asian nations, and Argentina and Brazil
trading with multiple different nations across the world.
Table 1 contains data on the top 20 maize-exporting nations In the maize trading network, the United States is at the center
and the top 20 maize-importing nations, as well as the total of a star-shaped topology. It exports large quantities of maize to a
amount of maize traded in metric tons from 2000 to 2009. As large number of nations, many of which do not import significant
Table 1 shows, the United States is by far the largest exporter of amounts of maize from any other nations. Hence, if any

PLOS ONE | www.plosone.org 3 September 2012 | Volume 7 | Issue 9 | e45151

 Aflatoxin Regulations in a Network of Food Trade

Table 1. Top maize exporting and maize importing nations worldwide, based on volume of trade from 2000–2009.

Top maize exporting nations and total amount exported

Rank 2000–2009, MTs Top maize importers and total amount imported 2000–2009, MTs

1 USA 526,670,541 Japan 170,279,244

2 Argentina 123,527,253 Republic of Korea 90,841,881
3 France 71,269,591 Mexico 69,857,045
4 China 65,558,093 Egypt 51,446,403
5 Brazil 54,473,911 Taiwan 47,282,122
6 Hungary 28,557,159 Spain 45,302,592
7 Canada 23,311,927 USA 33,978,967
8 Ukraine 19,568,172 Netherlands 28,629,716
9 South Africa 15,021,879 Malaysia 27,703,058
10 Paraguay 12,051,097 Iran 27,178,624
11 Mexico 11,923,079 Colombia 26,821,972
12 India 11,738,537 Canada 26,012,453
13 Germany 10,400,097 Algeria 20,230,143
14 Serbia 6,797,441 Italy 16,678,997
15 Thailand 5,366,268 Germany 16,548,899
16 Romania 4,859,320 Israel 15,658,213
17 Switzerland 3,620,319 Saudi Arabia 15,613,125
18 Netherlands 3,601,194 Portugal 14,245,589
19 Austria 3,394,665 Morocco 14,083,900
20 Bulgaria 2,962,606 United Kingdom 13,815,724

Source: United Nations Commodity Trade Statistics Database (UN Comtrade, comtrade.un.org). MTs = metric tonnes.
doi:10.1371/journal.pone.0045151.t001

Table 2. Nations with highest degrees of maize exports and imports: number of other nations with which it trades.

Maize-exporting nations and total number of nations to Maize-importing nations and total number of nations from
Rank, by degree which they export which they import

1 USA 181 France 69

2 Argentina 150 Germany 66
3 South Africa 128 USA 66
4 France 122 Netherlands 62
5 Canada 108 Canada 58
6 Brazil 101 Italy 57
7 China 95 United Kingdom 56
8 Italy 94 Spain 53
9 Netherlands 86 Egypt 53
10 India 85 Switzerland 51
11 Australia 78 Turkey 47
12 Ukraine 72 Austria 46
13 Hungary 72 Saudi Arabia 46
14 Thailand 70 Russian Federation 44
15 Spain 70 South Africa 44
16 Germany 69 United Arab Emirates 44
17 Turkey 69 Bulgaria 43
18 United Arab Emirates 66 Israel 41
19 United Kingdom 63 Romania 40
20 Chile 60 Belgium 40

Source: UN Comtrade.
doi:10.1371/journal.pone.0045151.t002

PLOS ONE | www.plosone.org 4 September 2012 | Volume 7 | Issue 9 | e45151

 Aflatoxin Regulations in a Network of Food Trade

Figure 1. Global maize trade network emphasizing top exporters. The circle sizes are loosely proportional to the amount of maize exported.
Each line represents export/import amount greater than 1 million metric tons from 2000–2009 [Wu and Guclu, unpublished data].
doi:10.1371/journal.pone.0045151.g001

circumstance jeopardized the amount of maize that the US could Figure 2b shows that among the maize-trading clusters
afford to export in any given season, some countries could identified previously, aflatoxin regulations look very similar. It is
experience a substantial loss in maize supply – particularly nations not surprising that the European maize trading community in the
in Latin America. This was particularly relevant in early years of lower left is homogenously colored, as the EU has set aflatoxin
large amounts of US maize being directed to ethanol production, standards that apply across all member states [25]. This EU
and impacts on other nations [24]. However, other countries that standard is relatively strict compared with other parts of the world.
import large amounts from the US (in the Middle East and Asia) However, it is interesting that the United States, which has a
are importing maize from other parts of the world as well. relatively relaxed total aflatoxin standard of 20 mg/kg (‘‘total
European nations trade much of their maize amongst each aflatoxin’’ refers to the sum of the levels of aflatoxins B1, B2, G1,
other; from the outside, several of these nations also import maize and G2: the four major types of aflatoxins), primarily exports
from Brazil and Argentina. However, none of these European maize to other nations that also allow relatively larger amounts of
nations imported more than one million MTs of maize over ten aflatoxin in maize. There are, however, several exceptions: several
years from the United States, despite the extremely large volume Latin American and Middle Eastern nations that have strict
of maize exports from the US. On the whole, this portion of the aflatoxin standards (Honduras, Cuba, Chile, Turkey, Tunisia, and
maize trading network appears more stable; as a relatively smaller Syria) import large amounts of maize from the US. Notably,
number of nations are receiving most of their maize imports from African nations do not export substantial amounts of maize to the
just one nation. EU. Several sub-Saharan African nations are in a maize trading
Many of the largest maize importers source their maize from cluster in the upper left of Figure 2b, while northern African
multiple different parts of the world, such as Japan and Korea nations such as Egypt, Morocco, and Tunisia are more in the
(importing from the United States, Brazil, Argentina, China, center of the global trading pattern.
Israel, and South Africa) and several European nations. However, These trends are highlighted as well in Table 3, which lists the
Mexico, though a large maize importer, purchases almost all of its 20 pairs of nations that have engaged in the greatest volume of
maize from just one nation: the United States. This is also the case maize trade in the last ten years, the aflatoxin standards in these
with many other nations in the Americas. nations (compiled from [23]), and the total amount of maize
Figures 2a and 2b depict the aflatoxin standards for maize set by traded from 2000–2009. Nations are not included in this table if
each individual nation around the world. In Figure 2a, darker hues they have set aflatoxin standards for other foodstuffs such as
represent nations with stricter aflatoxin standards, while nations in peanuts, but not for maize. In some nations, the aflatoxin standard
gray have not yet set maximum allowable standards for aflatoxin is set specifically for aflatoxin B1, rather than for total aflatoxins.
in maize. Figure 2b is the maize network representation with To extrapolate to an estimate of total allowable aflatoxins, we
colors indicative of the relative stringency of the aflatoxin standard multiplied the maximum allowable concentration of aflatoxin B1
for maize. by two.

PLOS ONE | www.plosone.org 5 September 2012 | Volume 7 | Issue 9 | e45151

 Aflatoxin Regulations in a Network of Food Trade

Figure 2. Color-coded maximum aflatoxin levels in maize by country: a) On the world map, and b) on the trade network. Each edge in
2b represents an export/import amount greater than 1 million metric tons from 2000–2009.
doi:10.1371/journal.pone.0045151.g002

Among these major maize trade relationships, there are but a with each other do not differ by more than 5 mg/kg. In fact, in the
few instances in which an exporting nation trades maize to an majority of these top 20 trading relationships, the importing and
importing nation with a significantly different aflatoxin standard. exporting nations have the same aflatoxin standard for maize.
In general, aflatoxin regulations in two nations that trade maize

PLOS ONE | www.plosone.org 6 September 2012 | Volume 7 | Issue 9 | e45151

 Aflatoxin Regulations in a Network of Food Trade

Table 3. Top volumes of maize trade worldwide from 2000–2009.

Rank Top exporter-importer pairs and their total aflatoxin (AF) standards in mg/kg maize Total amount (MT)
Exporter AF standard Importer AF standard

1 USA 20 Japan 20 159,377,000

2 USA 20 Mexico 20 69,764,700
3 USA 20 Taiwan 15 44,212,000
4 USA 20 Korea 20 41,657,300
5 China 40 Korea 20 36,446,400
6 USA 20 Egypt 20 35,540,100
7 USA 20 Canada 15 25,933,000
8 USA 20 Colombia 20 21,726,900
9 Canada 15 USA 20 21,161,900
10 France 4 Spain 4 18,682,400
11 France 4 Netherlands 4 14,901,600
12 Brazil 30 Iran 30 12,588,000
13 Mexico 20 USA 20 10,947,000
14 Argentina 20 Chile 5 10,625,700
15 USA 20 Algeria 20 10,457,700
16 USA 20 Dominican Rep. 20 10,325,300
17 Argentina 20 Spain 4 10,311,600
18 China 40 Malaysia 35 10,119,800
19 France 4 UK 4 9,899,890
20 Argentina 20 Egypt 20 9,734,360

Source: UN Comtrade and FAO [22]. MT = metric tonnes.

doi:10.1371/journal.pone.0045151.t003

Discussion African maize exports would be adversely affected by the strict EU

aflatoxin standard. Our findings here are in agreement with those
Earlier social network analyses in public health have demon- of [9] and [13]. Other crop exports may be, however; which was
strated that ‘‘like attracts like.’’ Controlling for other socioeco- not analyzed in this study.
nomic and demographic factors, smokers tend to be in closer social In between these two distinct clusters are nations that export
bonds with other smokers [17], and obese individuals form closer maize to multiple different parts of the world: Argentina, Brazil,
bonds with other obese individuals [16]. The same appears to be and China. These three nations supply maize to Africa, the
true of nations that trade maize with each other and the relative Americas, Asia, Europe, and the Middle East. Though they
strictness of food policies that pertain to maize in individual themselves have relatively relaxed aflatoxin standards, Argentina
nations. Namely, nations that share strong food trade ties tend to and Brazil export maize to multiple European nations with much
have similar regulations on allowable levels of aflatoxin in maize. stricter aflatoxin standards. However, all three nations trade more
Certain trade clusters for maize clearly emerge. The United with other nations that have relaxed standards or no standards at
States is the largest exporter of maize in the world, and occupies a all for aflatoxin in maize.
corner of the maize trade network that places it in the center of a The largest importer in the world of maize isJapan; which, in an
cluster of nations that import maize almost solely from it. Most US average year, imports nearly twice as much as the next largest
maize trade is done in with Canada, Latin American nations, and maize-importing nation, Korea. Mexico, Egypt, and Taiwan are
Middle Eastern nations; it does not export much maize at all to also large maize importers. Japan, Korea, and Egypt all import
Europe. Notably, the aflatoxin regulations in the nations to which maize from at least three different continents; which makes them
the US exports maize are roughly the same as the US Food and less vulnerable to maize scarcity if maize supplies for export
Drug Administration aflatoxinregulations. become limited in one part of the world. Taiwan and Mexico, on
European nations form another cluster of maize trade, the other hand, import almost all of their maize from the United
completely separated from the US maize cluster (i.e., no direct States. Hence, their maize supply is heavily dependent upon
links between the US and any European nation at the level of 1 continued availability of maize for export from the US. Possibly
million MTs maize traded from 2000 to 2009). In this European because they rely heavily on maize imports, all five of these nations
cluster, it is not surprising that most nations have the same have relatively relaxed standards for aflatoxin in maize.
aflatoxin regulation, as many of these nations are EU member These clustering patterns and directionality of maize trade have
states that share mycotoxin regulations [23]. Within this cluster, important implications for food security. If the members in a
France and Hungary are the main maize exporters, while Spain cluster are not well-connected, then the cluster may be more
and the Netherlands are main importers. In contrast to former UN vulnerable to any adverse consequences that may affect the central
Secretary General Annan’s statement, because EU member states nation in the cluster. Because the United States is at the center of a
do not import much maize at all from Africa, it is unlikely that large cluster of North and Latin American nations, which

PLOS ONE | www.plosone.org 7 September 2012 | Volume 7 | Issue 9 | e45151

 Aflatoxin Regulations in a Network of Food Trade

themselves are hardly connected with each other or with other enactment at different points in history, to examine whether the
nations in the network, then this portion of the network of maize trade communities preceded the enactment of regulations, or vice
trade would become extremely vulnerable to reduced food supply versa.
and increased prices if anything should affect the quality or If the specific trade community determines food safety
quantity of US maize available for export. This was relevant in the regulations in multiple nations, then the implication is that certain
case of US maize ethanol production and its attendant effects on central nations in the network have a large amount of power in
maize-importing nations in recent years. Because European determining regulations elsewhere worldwide. However, if regu-
nations are more densely clustered in maize trade and also source lations determine trade patterns, then nations should be aware of
maize from Brazil and Argentina, they may be more stable to the implications their standard setting will have on who their
fluctuations in maize supply in any one part of the world. future food trading partners worldwide will be. Because the
One limitation of static network models is the inability to prove implications are important for food security in ways that extend
causality. We have shown that nations tend to cluster into maize beyond aflatoxin exposure, this is an area of research that deserves
trading communities that share similar aflatoxin regulations. further attention for the purposes of policy decision-making.
However, do the regulations cause the trade patterns to emerge
as they are, or do the trading patterns influence the regulations
Author Contributions
that nations set? Or might both be possible? Although causality for
one of these cannot be proven, one future research direction Conceived and designed the experiments: FW. Performed the experiments:
involves temporal modeling of food trade and regulation HG. Analyzed the data: FW HG. Wrote the paper: FW HG.

References
1. Groopman JD, Kensler TW, Wild CP (2008) Protective Interventions to Prevent 13. World Bank (2005) Food Safety and Agricultural Health Standards: Challenges
Aflatoxin-Induced Carcinogenesis in Developing Countries. Annu Rev Public and Opportunities for Developing Country Exports. Report No. 31207, http://
Health 29: 187–203. siteresources.worldbank.org/INTRANETTRADE/Resources/Topics/
2. Strosnider H, Azziz-Baumgartner E, Banziger M, Bhat RV, Breiman R, et al. Standards/standards_challenges_synthesisreport.pdf, last accessed July 2012.
(2006) Workgroup Report: Public Health Strategies for Reducing Aflatoxin 14. Ferguson NM, Cummings DAT, Fraser C, Cajka JC, Cooley PC, et al. (2006)
Exposure in Developing Countries. Environ Health Perspect 114: 1898–1903. Strategies for mitigating an influenza pandemic. Nature 442: 448–452.
3. Gong YY, Cardwell K, Hounsa A, Egal S, Turner PC, et al. (2002) Dietary 15. Epstein JM, Goedecke DM, Yu F, Morris RJ, Wagener DK, et al. (2007)
aflatoxin exposure and impaired growth in young children from Benin and Controlling pandemic flu: the value of international air travel restrictions. PLoS
Togo: cross sectional study. BMJ 325: 20–21. ONE 2: e401.
4. Khlangwiset P, Shephard GS, Wu F (2011)Aflatoxins and Growth Impairment: 16. Christakis NA, Fowler JH (2007)The spread of obesity in a large social network
A Review. Crit Rev Toxicol 41: 740–755. over 32 years. NEJM 357: 370–9.
5. Jolly PE, Jiang Y, Ellis WO, Wang JS, Afriyie-Gyawu E, et al. (2008) Modulation 17. Christakis NA, Fowler JH (2008)The collective dynamics of smoking in a large
of the human immune system by aflatoxin. In: Leslie JF, Bandyopadhyay R, social network. NEJM 358: 2249–58.
Visconti A, editors. Mycotoxins: detection methods, management, public health
18. Lyons R (2011)The Spread of Evidence-Poor Medicine via Flawed Social-
and agricultural trade. Oxfordshire, UK: CAB International. 41–52.
Network Analysis. Statistics, Politics, and Policy 2: A2.
6. Kuniholm MH, Lesi OA, Mendy M, Akano AO, Sam O, et al. (2008) Aflatoxin
19. Petroczi A, NepuszT, Taylor G, Naughton DP (2011) Network analysis of the
exposure and viral hepatitis in the etiology of liver cirrhosis in The Gambia,
West Africa. Environ Health Perspect 116: 1553–7. RASFF database: a mycotoxin perspective. World Mycotoxin J 4: 329–38.
7. Liu Y, Wu F (2010) Global Burden of Aflatoxin-Induced Hepatocellular 20. McKenzie FE, Samba ES (2004)The role of mathematical modeling in evidence-
Carcinoma: A Risk Assessment. Environ Health Perspect 118: 818–24. based malaria control. Am J Trop Med Hyg71: 94–6.
8. Wu F (2008) A Tale of Two Commodities: How EU Mycotoxin Regulations 21. Valente TW (2010) Social Networks and Health: Models, Methods, and
Have Affected Food Industries. World Mycotoxin J 1: 71–8. Applications. New York: Oxford University Press.
9. Wu F (2004)Mycotoxin Risk Assessment for the Purpose of Setting International 22. deNooy W, MrvarA, Batagelj V (2005) Exploratory Social Network Analysis
Regulatory Standards. Environ SciTechnol 38: 4049–55. with Pajek. New York, Cambridge University Press.
10. JECFA (1998) Safety evaluation of certain food additives and contaminants, 23. Food and Agriculture Organization (2004) Worldwide regulations for mycotox-
WHO Food Additives Series 40, Aflatoxins. In: The forty-ninth meeting of the ins in food and feed in 2003. Rome, FAO Press, ISBN 92-5-105162-3.
Joint FAO/WHO Expert Committee on Food Additives. Geneva: World Health 24. Boddiger D(2007)Boosting biofuel crops could threaten food security. Lancet
Organization. 370: 923–4.
11. Henry SH, Bosch FX, Troxell TC, Bolger PM (1999) Public health: Reducing 25. Van Egmond HP, Schothorst RC, Jonker MA(2007) Regulations relating to
liver cancer – global control of aflatoxin. Science 286: 2453–2454. mycotoxins in food: perspectives in a global and European context. Anal
12. Otsuki T, Wilson JS, Sewadeh M (2001) Saving Two in a Billion: Quantifying BioanalChem 389: 147–57.
the Trade Effect of European Food Safety Standards on African Exports. Food
Policy26: 495–514.

PLOS ONE | www.plosone.org 8 September 2012 | Volume 7 | Issue 9 | e45151