Aflatoxins in Midwestern Corn

Charles R. Hurburgh, Jr.

Aflatoxin has been a problem in midwestern corn in years when there were drought conditions. Significant incidence occurred in 1983 and, most recently, in 1988. Although aflatoxin is a chronic problem in southwestern states, the effect on the corn market is much more pronounced when midwestern corn is affected. In 1988, nine corn-belt states (SD, NE, MN, IA, MO, WI, IL, IN, OH) produced 3.8 billion bushels of corn, 78% of total U.S. production (Iowa Crop and Livestock Reporting Service, 1989). Aflatoxin in these areas obviously creates major difficulties in finding clean stocks for export and sensitive domestic food uses.

Midwestern grain handlers receive corn very rapidly, especially at country elevators during harvest. Because aflatoxin test procedures are relatively slow and variable, identification of contaminated lots is nearly impossible on a large scale. Country elevators typically receive producer-deliveries at a rate about 1 load per minute.

The objectives of this article are: (1) to summarize available data on aflatoxin outbreaks in the Midwest and (2) to relate reported aflatoxin levels to the needs and constraints of midwestern grain handlers.

Aflatoxin data for midwest corn

Incidence data is available for the two most recent aflatoxin outbreaks--1983 and 1988. Any comparisons of data between states, however, or even within states, must be used with caution because sampling and analytical errors were large (coefficients of variation on the order of 25-50%) (Dickens and Whitaker, 1983) and protocols for collecting samples were not uniform.

Figure 1 shows the results of testing 99 samples of 1983 Iowa corn for aflatoxin as reported by Schmitt and Hurburgh (1989). Each sample was a composite of six 4-lb samples collected from farm bins in the 99 Iowa counties. Average July-August temperatures over the entire state were 2-3° higher than normal, but the southern half of the state also received about 6 inches less rainfall than normal. The aflatoxin contamination was concentrated in areas with high temperature and low rainfall, but there was not a high mathematical correlation, as shown in Table 1.

Figure 2 has the same type of data collected in Indiana (Tuite et al., 1984). The samples were hand-harvested from fields just before harvest. Again, the concentration of aflatoxin in localized areas is evident.

The usual method for presenting aflatoxin data is to tabulate the percentage of samples with aflatoxin contents in excess of certain levels (e.g. 20 ppb or 100 ppb). Iowa data from 1983 (Schmitt and Hurburgh, 1989) and 1988 (Hurburgh, 1989) are presented in this manner in Table 2.

These data illustrate the difficulty in generalizing about statewide or areawide averages. One might conclude that the average for all 1983 Iowa corn was 21 ppb, more than the 20 ppb action level. Yet the map in Figure 1 clearly shows that large areas of Iowa were relatively free of aflatoxin. Aflatoxin results must be interpreted on a localized, case-by-case basis. Average aflatoxin figures are meaningful only to indicate the risk of creating feed mixes that have aflatoxin levels in excess of action levels.

The local variability of aflatoxin is apparent in Table 3 from Schmitt and Hurburgh (1989). These are the individual-subsample results from the ten counties in which tests of the composite samples indicated more than 100 ppb (1983 crop). The composites weighed about 11-12 lbs; the individual subsamples weighed, 1.5-2.0 lbs. Even in these clearly high-risk areas, a significant portion of the samples tested less than 20 ppb. It is inappropriate to apply generalizations about average levels to individual lots. The grain handler receiving many lots may experience the local average but individual grain sellers will be treated unfairly by averages. Unfortunately, summary reports to the public tend to accentuate such generalizations, to the detriment of the grower.

Extensive publicity forced most midwestern states to conduct some sort of survey program in 1988. A summary of results in given in Table 4. Although Iowa and Illinois had greater incidences of aflatoxin, the diversity of survey protocols and sampling frequencies greatly weakens any scientific conclusions that might be drawn. Given the variability within a local area, the average level of aflatoxins in corn harvested in a state can not be used to estimate or forecast the extent of toxin occurring in that area. Unfortunately, the general public and media view such figures as indicative of the amount of contaminated grain in a state.

Other sources of public information are summaries of samples voluntarily submitted to either state laboratories or veterinary diagnostic laboratories. Both of these sources are completely biased because the samples were submitted because they were suspect, either because of a positive black-light test or from adverse animal-health symptoms. In 1988, 75% of the samples submitted to the Iowa Veterinary Diagnostic Laboratory had more than 20 ppb, with 30% having more than 100 ppb (ISU Veterinary Diagnostic Laboratory, 1989, unpublished). This was a totally unrealistic portrayal of incidence in Iowa, yet, lacking more reliable data, this type of information has been used to estimate state-wide levels.

The incidence problem is further compounded by the natural incentive for each state to downplay its situation. State government agencies are usually the agents for the Food and Drug Administration in their respective states, and, thus, are called upon to do the testing. No state wants to portray itself as a center of contamination, thereby risking market isolation. Clearly, is incidence surveys are to be done, they should be designed and managed by an agency with no local conflict of interest. Developments in testing technology may make marketplace screening so routine that there will be no need for government monitoring data.

Aflatoxin development is weather-related. Laboratory studies and field studies in the southeastern United States have provided information on critical weather parameters (Lillehoj, 1983; Diener et al., 1987). High temperature (day-time high and nighttime low) and lack of rain are necessary conditions for crop stresses that magnify aflatoxin development. But average weather conditions for a state or substate area have not been effective in predicting specific aflatoxin levels. Localized weather information is needed, as well as more knowledge of compounding agronomic factors. The extreme variation in high-risk areas (Table 3) supports the contention that factors in addition to weather conditions have great influence over the presence or absence of aflatoxin (Schmitt and Hurburgh, 1989).

Weather data can play an important role in dealing with the sporadic incidence in midwestern corn. Aflatoxin has been reported in years when the Iowa July-August mean temperature has exceeded 76-77°F. These also were dry years; low rainfall gives little evaporative cooling, thus contributing to temperature rise (Thompson, 1986). Macro-weather data may be very useful in estimating the risk of aflatoxin, even though it may not predict specific levels or incidence percentages. The Iowa Aflatoxin Task Force, in its August 1989 report, suggested that national weather data could be used to forecast high-risk areas before harvest (ICGA, 1989).

Midwestern corn compared to southeastern corn

Aflatoxin contamination occurs relatively frequently in southeastern corn production. But organized survey data is relatively scarce; individual handlers monitor inbound corn at their discretion (Hagler, 1990). Table 5 published survey data for the southeastern states (Nichols, 1987).

It seems that southeastern corn is routinely affected by aflatoxin to the same extent that Corn-Belt corn is in a stress year. But the southeastern states contribute only a small percentage of total U.S. production. Furthermore, if levels of aflatoxin in contaminated southeastern corn equal those of stress-year midwestern corn, random commingling will produce blends that usually have less than 20 ppb of aflatoxin.

Aflatoxin and grain handlers

The country grain handler faces a decision as to whether to accept or reject aflatoxin-contaminated corn. At country elevators, that decision must be made in 1-2 minutes, the time normally available for grading. The more complete inspection for grade factors, done at river grain terminals, processing plants, and export elevators, allows more time for aflatoxin testing at those points.

Aflatoxin levels also are the most variable in farmer-deliveries, which often originate from an individual field. The variations shown to Table 3 probably are typical of what a country elevator in a high-risk area would receive. Lot-to-lot variations are compounded by the well-documented random error in sampling and testing, on the order of ±50%.

The only truly rapid screening method available to country elevators is the black-light. The ELISA-based test require 8-10 minutes per sample, which is too long for accept/reject decisions. And moreover, it requires 15-30 minutes to prepare a sample suitable for the test. Even if used properly, the black light gives 40-60% false positives (fluorescence without aflatoxin) (Schmitt and Hurburgh, 1989; Shotwell et al., 1975). Attempts to relate the number or weight of fluorescing particles to aflatoxin concentration have been unsuccessful, with coefficients of variation in excess of 100% (Shotwell, 1983). The black-light test can establish risk, but not definitive presence or absence of aflatoxin at any level. Therefore, the country grain handler’s problem is one of assessing risk and of estimating the average concentration of large lots consolidated from smaller farmer deliveries. If more then 20-30% of individual lots are truly in excess of 20 ppb, then the risk of having a consolidated shipment test greater than 20 ppb is relatively great.

Judicious use of the black light can reduce this risk. For 1983 and 1988 Iowa corn, the average concentration of samples with fewer than 5 BGYF particles (whole kernel) per kg was 10 ppb, with 70% false positives, whereas samples with 5 or more BGYF particles averaged 65 ppb with only 20% false positives. Shotwell et al. (1985) reported 35% false positives for samples of North Carolina corn containing 4 or more BGYF particles per kilogram. Individual samples with low glower counts did have much greater levels (and vice-versa, but, overall, samples with low glower counts had less aflatoxin. Using glower count to assess risk is not the same as using glower count to predict aflatoxin level.

The point is that country elevators in high-volume grain production areas need sampling and analysis techniques compatible with their handling constraints. Their decision process must necessarily be based on risk probabilities. Their legal options for handling grain must be flexible enough to accommodate the variability in sampling, testing and lot-to-lot aflatoxin levels. Accurate identification of all contaminated lots is not possible at this point.

The county grain elevator also acts as a warehouse for growers without sufficient on-farm storage to meet their marketing needs. Warehouses are licensed under either state or federal statutes, but, in either instance, warehouses must maintain corn of quality equal or better than that certified on warehouse receipts. Generally warehouse receipts are only issued for the standard U.S. Grades, which by implication means less than 20 ppb aflatoxin. Clearly, warehouses in high-volume areas must know the aflatoxin status of grain received. The cost of finding aflatoxin in corn or peanuts stored in a warehouse can be great. The sampling error is yet further magnified in a storage structure compared to a bulk carrier. In 1988, there were reports of warehouses with suspect grain receiving unequal treatment from various state and federal jurisdictions.

Analysis and Commentary

The incidence data demonstrate the difficulty of assessing aflatoxin levels in a growing area. So far, aflatoxin contamination in midwestern corn has been limited enough that the natural consolidation of grain lots into larger shipments eventually diluted aflatoxin concentrations to less than 20 ppb most of the time. This would not apply to users drawing from localized high-risk areas. A much better early-warning system is needed to identify potential high-risk areas.

The surveys at harvest do respond to public pressure for information, but their methodology should be standardized. The estimated average aflatoxin levels and distribution are very uncertain, which means that the data should be explained carefully. Because each state has a vested interest in not finding aflatoxin, surveys should be administered at the Federal level.

The current FDA policy of considering aflatoxin as an adulterant creates a strong incentive not to know aflatoxin levels. The U.S. high-volume grain industry is designed around the ability to adjust quality by mixing, blending, and dilution. Labeling of grain as unmarketable, particularly on the basis of tests with great uncertainty, is a major problem for handlers. Therefore, it is sometimes perceived as advantageous not to know, or not to suspect. The public would be better protected by a more open, flexible system that provides disposition options at the raw commodity level, rather than outright rejections. The burden, in the present system, is placed mainly on the end-user preparing to convert corn into food, meat, or dairy products. When aflatoxin occurs in corn in the Midwest, FDA simply does not have the resources to police an adulterant-oriented approach to aflatoxin. The grain industry has to have enough realistic disposal options to create an incentive for self-regulation at the local elevator level and for monitoring by the Federal Grain Inspection Service at the export and processor levels. No handler will be able to make perfect identification of aflatoxin lots. Therefore, grain receival procedures should be designed to reduce risk and remove with some certainty the lots with the greatest aflatoxin levels. Successive application of such methods will relieve the burden on the final users and protect the consuming public to a greater degree than is now done.

Aflatoxin should no longer be regarded as a rare event in the Midwest. Aflatoxin risk assessment, monitoring, and test procedures should become part of the normal course of grain business. There is no reason why each new outbreak should be treated as a new and surprising event, with its associated public concern about safety. The problems of aflatoxin in midwestern corn are no different than are associated with aflatoxin in other areas, but the volume of grain handled in the Midwest greatly accentuates both the public-health and grain-handling difficulties. Fortunately, technology exists, or could exist with appropriate research support, to deal with aflatoxin in a practical manner that will better protect the public.

References

1. Dickens, J.W., and T. Whitaker. 1983. Aflatoxin testing procedures for corn. In Aflatoxin and Aspergillus flavus in Corn. Southern Cooperative Series Bulletin 279. Alabama Agr. Exp. Station, Auburn University, Auburn, AL.
    
2. Diener, U.L. R.J. Cole, T.H. Sanders, G.A. Payne, L.S. Lee, and M.A. Klich. 1987. Epidemiology of aflatoxin formation by Apergillus flavus. Ann. Rev. Phytopathology 25:249-270.
    
3. Hagler, W.M. 1990. Aflatoxin in field crops in the southeastern United States. Presented at ARS Aflatoxin Symposium, Agricultural Research Service, USDA, Beltsville, MD.
    
4. Hurburgh, C.R., Jr. 1989. Aflatoxin in 1989 Iowa Corn. Agri. Engr. Dept. Staff Papers Series FPE 89-10. Agr. Engr. Dept., Iowa State Univ., Ames, IA.
    
5. ICGA. 1989. Aflatoxin: Strategies for the Future. Iowa Corn Growers Association, West Des Moines, IA.
    
6. Iowa Crop and Livestock Reporting Service. 1989. Iowa Crop Report 89(12), December 1989.
    
7. Lillehoj, E.B. 1983. Effect of environmental and cultural factors on aflatoxin contamination of developing corn kernels. In Aflatoxin and Aspergillus flavus in Corn. Southern Cooperative Series Bulletin 279. Alabama Agr. Exp. Station, Auburn University, Auburn, AL.
    
8. Nichols, T.E. 1987. Aflatoxin in the Southeastern USA. In Aflatoxin in Maize: A Proceedings of the Workshop. CIMMYT, Mexico.
    
9. Schmitt, S.G., and C.R. Hurburgh, Jr. 1989. Distribution and measurement of aflatoxin in 1983 Iowa corn. Cereal Chem 66(3):165-168.
    
10. Shotwell, O.L., M.L. Goulden, A.M. Jepson, W.F. Kwolek, and C.W. Hesseltine. 1975. Aflatoxin occurrence in some white corn under loan, 1971. III. Association with bright greenish-yellow fluorescence in corn. Cereal Chem. 52:670-677.
     
11. Shotwell, Odette. 1983. Aflatoxin detection and determination in corn. In Southern Cooperative Series Bulletin 279. Alabama Agr. Exp. Station, Auburn University, Auburn, AL.
     
12. Thompson, L.M. 1986. Cliamtic change, weather variability and corn production. Agron. J. 78(4):649-653.
     
13. Tuite, J., R. Sensmeir, C. Koh-Knox, and R. Noel. 1984. Preharvest afaltoxin contamination of dent corn in Indiana in 1983. Plant Dis. 68:893-895.

Table 1. Temperature and Cumulative Precipitation Departures During July and August 1983 in Nine Iowa Crop Reporting Districts (Schmitt and Hurburgh, 1989).

^a Parts ber billion, average of county composites.

Table 2. Aflatoxin data for 1983 and 1988 Iowa Corn^a (after Schmitt and Hurburgh, 1989; Hurburgh, 1989).

^a Samples collected from farm bins in 1983, at country elevator dryers in 1988.

Table 3. Distribution of aflatoxin concentrations in county composite samples that contained more than 100 ppb. 1983 Iowa corn (Schmitt and Hurburgh, 1989).

Table 4. A summary of aflatoxin survey results for the 1988 corn crop, midwestern states.

^a Percentage of false positives from black light: IA-57%, IN-45%, NE-more than 50%.
^b Weighted by relative state crop production, excluding Missouri.
Date supplied by respective state regulatory laboratories.

Table 5. Aflatoxin in North Carolina corn, elevator surveys 1976-1980 (after Nichols, 1987).