Articles
Reduction in exposure to carcinogenic aflatoxins by postharvest intervention measures in west Africa: a community-based intervention study P C Turner, A Sylla, Y Y Gong, M S Diallo, A E Sutcliffe, A J Hall, C P Wild
Summary Lancet 2005; 365: 1950–56 Molecular Epidemiology Unit, Centre for Epidemiology and Biostatistics, University of Leeds, Leeds UK (P C Turner PhD, Y Y Gong PhD, A E Sutcliffe BSc, Prof C P Wild PhD); Infectious Disease Epidemiology Unit, London School of Hygiene & Tropical Medicine, University of London, London UK (Prof A J Hall PhD); and Institut Pasteur de Guinée, Kindia, Republic of Guinea (A Sylla BSc, M S Diallo DES) Correspondence to: Prof C P Wild, Molecular Epidemiology Unit, Centre for Epidemiology and Biostatistics, Leeds Institute of Genetics, Health, and Therapeutics, University of Leeds, Leeds LS2 9JT, UK
[email protected]
Background Aflatoxins are fungal metabolites that frequently contaminate staple foods in much of sub-Saharan Africa, and are associated with increased risk of liver cancer and impaired growth in young children. We aimed to assess whether postharvest measures to restrict aflatoxin contamination of groundnut crops could reduce exposure in west African villages. Methods We undertook an intervention study at subsistence farms in the lower Kindia region of Guinea. Farms from 20 villages were included, ten of which implemented a package of postharvest measures to restrict aflatoxin contamination of the groundnut crop; ten controls followed usual postharvest practices. We measured the concentrations of blood aflatoxin–albumin adducts from 600 people immediately after harvest and at 3 months and 5 months postharvest to monitor the effectiveness of the intervention. Findings In control villages mean aflatoxin–albumin concentration increased postharvest (from 5·5 pg/mg [95% CI 4·7–6·1] immediately after harvest to 18·7 pg/mg [17·0–20·6] 5 months later). By contrast, mean aflatoxin–albumin concentration in intervention villages after 5 months of groundnut storage was much the same as that immediately postharvest (7·2 pg/mg [6·2–8·4] vs 8·0 pg/mg [7·0–9·2]). At 5 months, mean adduct concentration in intervention villages was less than 50% of that in control villages (8·0 pg/mg [7·2–9·2] vs 18·7 pg/mg [17·0–20·6], p0·0001). About a third of the number of people had non-detectable aflatoxin–albumin concentrations at harvest. At 5 months, five (2%) people in the control villages had non-detectable adduct concentrations compared with 47 (20%) of those in the intervention group (p0·0001). Mean concentrations of aflatoxin B1 in groundnuts in household stores in intervention and control villages were consistent with measurements of aflatoxin–albumin adducts. Interpretation Use of low-technology approaches at the subsistence-farm level in sub-Saharan Africa could substantially reduce the disease burden caused by aflatoxin exposure.
Introduction Staple foods in west Africa and other parts of the developing world are frequently contaminated with aflatoxins, metabolites of aspergillus species.1 Aflatoxins have a carcinogenic action on the liver and act synergistically with chronic hepatitis B virus.1 A joint Food and Agriculture Organisation of the United Nations/WHO committee thus concluded that reduced intake of aflatoxins in places where infection with hepatitis B virus is endemic would have a substantial effect on the incidence of hepatocellular carcinoma,2 which accounts for about a fifth of all cancers in men in west Africa.3 Furthermore, aflatoxins can cause growth inhibition and immune suppression in animals.4 Consistent with these observations, aflatoxin exposure in west Africa at the time of weaning has been associated with impaired child growth5–7 and with a decreased immune response.8 These health effects further emphasise the potential benefit to public health of reducing aflatoxin exposure in developing countries. Aflatoxin contamination of foods is especially severe after long-term crop storage because excessive heat, humidity, and insect and rodent damage result in proliferation and spread of fungal spores. The main 1950
crops affected include maize (corn) and groundnuts (peanuts), which together are the main dietary staples in many parts of sub-Saharan Africa. The reliance on subsistence farming, the limited food diversity, and toxin contamination result in high aflatoxin exposures throughout life.9 By use of a blood-based biomarker, aflatoxin–albumin adducts, data for several west African countries show that more than 98% of children and adults have detectable amounts in their blood.9,10 Exposures are orders of magnitude higher than those allowed by regulation in Europe, the USA, and other parts of the developed world. Several approaches can prevent aflatoxin exposure in developing countries.11 Because much food contamination occurs during postharvest storage, methods to remove nuts or kernels damaged by fungus before storage and to restrict humidity during storage could reduce fungal growth and toxin production. However, the effect of this strategy on exposure in subsistence-farming communities in Africa has not been assessed. Our aim was to assess the use of improved methods of groundnut storage for subsistence farmers in a rural area of Guinea. Aflatoxin and infection with hepatitis B www.thelancet.com Vol 365 June 4, 2005
Articles
virus are endemic in this region, and groundnuts are the main source of aflatoxin exposure.12–14 We introduced a package of postharvest intervention measures to farmers in ten villages at the time of groundnut harvest and made comparisons with farmers in ten neighbouring villages that used usual storage practices.
Methods Participants The study was undertaken in 20 villages in the Kindia prefecture of lower Guinea, which were close to each other (within about 60 km of Kindia) to ensure that climate—as well as practices of groundnut cultivation, harvesting, and storage—were much the same in all villages. The ten intervention villages were to the north of Kindia and ten control villages were to the southeast to keep to a minimum the risk of prevention strategies being used in control villages. In every village we recruited, by a simple random sampling procedure, 15 families who were subsistence farmers who grew groundnuts. In all families the head of the household and his spouse, or if unavailable another man and woman who were older than 16 years, were included. Eligible families were those who had communal family meals, were permanently resident in the village, and had no signs of illness at the time of recruitment. We identified 600 individuals and monitored their aflatoxin–albumin concentrations. The study was approved by the Comité National d’Ethique pour la Recherche en Santé in Guinea and by the ethics committee of the London School of Hygiene and Tropical Medicine, UK. The aim of the study was explained to community leaders in the villages and then to eligible individuals in their own language. All participants gave informed consent.
Intervention strategies We introduced a package of intervention measures to improve the drying and storage of groundnuts. Local government agricultural advisers, who were employed to provide guidance to subsistence farmers, explained the intervention strategy to the farmers and demonstrated the different techniques. Families included in the study stored between ten and 25 bags of groundnuts (each bag weighed about 50 kg). The panel shows the intervention methods. In the control villages, farmers were left to follow their usual postharvest practices, which sometimes included one or more of the parts of the intervention package.
Survey data The study consisted of three main surveys: the first at harvest time in September or October, 1999 (survey 1); the second about 3 months postharvest in December or January, 2000 (survey 2); and the third in February or March, 2000, 5 months after harvest (survey 3, figure 1). At the first survey we recorded the age, sex, ethnic group, www.thelancet.com Vol 365 June 4, 2005
duration of residence in village, and profession of the 600 individuals. The date of groundnut harvest, time of transfer from the field to the village, and the rainfall between harvest and storage were also noted. A simple dietary questionnaire was administered to every person at all surveys to assess recent consumption of groundnuts and other foods, including maize. Additional surveys in October or November, 1999 (intermediate survey 1), and in January or February, 2000 (intermediate survey 2), were done between the main surveys. At these visits fieldworkers and agricultural workers were able to verify compliance with the intervention measures and administer the dietary questionnaire. The temperature and humidity in the storage facility were determined by use of a handheld thermohygrometer (HANNA HI-8564, Philip Harris, Leicester, UK) at the time of harvest (survey 1) and at intermediate survey 2.
Panel: Intervention measures Hand sorting Farmers were shown how to identify groundnuts that were visibly mouldy or had damaged shells. Damaged kernels were removed and discarded before storage. Drying on mats Groundnuts are commonly spread on the ground for sun drying, making them susceptible to humidity and difficult to gather in the event of unexpected rain. Therefore, locally produced natural-fibre mats for the sun-drying process were provided. Sun drying Incomplete sun drying leaves residual humidity in the groundnuts during storage. Farmers were shown how to judge the completeness of sun drying by shaking the kernels to listen for the free movement of the dried nuts. Storage in natural-fibre bags Farmers most frequently use plastic or other synthetic bags for storage, which promote humidity. We therefore provided natural-fibre jute bags. Wooden pallets In the storage facilities, bags of groundnuts are often stored on the floor or on stones leading to the risk of humidity from the earthen floors. We provided locally made wooden pallets on which to store the bags. Insecticide One of the main factors affecting aflatoxin formation is the presence of insects in storage facilities and bags, which produce humidity via metabolic activity and spread fungal spores. We gave about 10 kg of a locally available insecticide (acetilite) to every family to be sprinkled in small quantities on the floor of the storage facility under the wooden pallets at the start of storage and intermittently afterwards.
1951
Articles
Blood-sample collection
Sept/Oct
Groundnut-sample collection
Dec/Jan
Intermediate survey 1 Survey 1 Harvest
Feb/Mar
Intermediate survey 2 Survey 2
Survey 3
Figure 1: Outline of study design 30 people from 20 villages (ten intervention and ten control) participated.
villages to identify variables that significantly correlated with aflatoxin–albumin adducts at each of the two follow-up surveys, independently of the intervention strategies. Variables identified by these models were used in the final multivariable model in all villages together with baseline (ie, survey 1 at harvest) adduct concentrations to analyse the adjusted effects of the intervention strategies. The final statistical analyses in multivariable models were done with individual and village mean concentrations of aflatoxin–albumin adducts; specific intervention activities (eg, hand sorting) were not retained in the model. For mean concentrations per village, individuals were assigned the mean adduct concentration for their village of residence. All analyses were done with STATA software version 7.0.
Role of the funding source Aflatoxin exposure To monitor the effectiveness of the intervention we took blood samples from participants at the three main survey points; sera were separated and stored at –20oC. Concentrations of aflatoxin–albumin adducts were measured by ELISA.15 In a subset of three families per village a sample of groundnuts from the household store was collected by incremental sampling at each of the three main surveys to measure the concentration of aflatoxin B1 by use of thin-layer chromatography after extraction in methanol/chloroform.16
Statistical analysis Because our study recruited a man and woman from every family, the survey variables were regarded as individual (eg, food consumption) or as household (eg, harvest practices). In the control villages we used questionnaire data from the women for harvest and storage practices because they usually take more responsibility for farming in these communities than do men. Frequencies or means of individual and household variables were calculated at the group level (intervention or control) and village level for comparison with aflatoxin–albumin concentrations. Data for aflatoxin– albumin adducts were not normally distributed, and were therefore natural-log transformed (all aflatoxin–albumin adduct data are presented as unadjusted geometric means with 95% CIs unless otherwise stated). Concentrations of aflatoxin–albumin adducts in control participants were first compared with every variable in the control villages by univariable analysis. We included data for adducts only from women to avoid clustering; however, men and women did not differ in aflatoxin–albumin adduct concentration in a paired analysis. We tested the difference between means for a specific variable by the t test or by ANOVA. Variables with a tendency towards significance (p0·2) were then entered into a multivariable model in the control 1952
The sponsor of the study had no role in the collection, analysis, and interpretation of the data; in the writing of the report; or in the decision to submit the article for publication. The corresponding author had full access to the data and was responsible for the final decision to submit the paper for publication.
Results 150 pairs of men and women were recruited from ten villages for both the intervention and control groups. Participants in the intervention group were younger than those in the control group (table 1). Three main ethnic groups (Soussou, Peulh, and Kissi) were represented, and intervention villages had an overrepresentation of Soussou compared with the control group. Neither age nor ethnic group correlated with concentrations of aflatoxin–albumin adducts. Data were obtained for 535 (89%) of 600 individuals in survey 1, for 529 (88%) of 600 in survey 2, and for 514 (86%) of 600 in survey 3. Men and women did not differ in adduct concentration at any of the three individual surveys (p=0·4, 0·4, and 0·8 at the three surveys, respectively; paired t test). There was substantial variation in aflatoxin–albumin adduct concentrations between villages in both groups (figure 2). Overall, the mean adduct concentration was lowest at harvest in both groups, but was higher in the intervention group than in the control group at this time (7·2 pg/mg [95% CI 6·2–8·4] vs 5·5 pg/mg [4·7–6·1]; p=0·02). The fairly low concentrations at harvest were as expected, since aspergillus and aflatoxin contamination were already present in the field but then increase during storage. In control villages the mean concentration of aflatoxin–albumin adducts increased during the 5 months from 5·5 pg/mg (95% CI 4·7–6·1) immediately postharvest to 18·7 pg/mg (17·0–20·6) after 5 months. By contrast, the mean concentration in the intervention villages after 5 months of groundnut www.thelancet.com Vol 365 June 4, 2005
Articles
Intervention (n=300)
Control (n=300)
Male sex Age (years)
150 28·6 (16–59)
150 33·7 (16–60)
Ethnic origin Soussou Peulh Kissi Other
245 (82%) 23 (8%) 30 (10%) 2 (1%)
204 (68%) 41 (14%) 47 (16%) 8 (3%)
150 (100%) ·· ·· 150 (100%)
10 (7%) 79 (53%) 61 (41%) 52 (35%)
150 (100%) ·· ··
4 (3%) 107 (71%) 39 (26%)
150 (100%) ·· ·· 150 (100%)
6 (4%) 70 (47%) 74 (49%) 19 (13%)
50 40 30 20
Temperature (ºC) Survey 1 Intermediate survey 2
36·8 (36·5–37 ·0) 36·2 (35·9–36 ·6) 27·8 (27·7–28 ·0) 27·3 (27·1–27 ·5)
Humidity (%) Survey 1 Intermediate survey 2
71·4 (71·2–71 ·6) 71·0 (70·7–71 ·3) 68·3 (68·1–68 ·4) 68·1 (67·9–68 ·2)
Aflatoxin–albumin (pg/mg)
Harvest and storage Groundnut drying surface Mat Ground Mixed Hand sorting (yes) Type of bag Natural fibre Plastic Mixed Pallet Wooden Other None Insecticide use (yes)
Control villages
60
10 0 Intervention villages
60 50 40 30
Data are number (%) or mean (95% CI). Percentages have been rounded to the nearest whole number.
Table 1: Demographic data and harvest practices of intervention and control groups
storage (survey 3) was much the same as that immediately postharvest (survey 1; 7·2 pg/mg [6·2–8·4] vs 8·0 pg/mg [7·0–9·2]), but with a moderate increase at survey 2 (11·7 pg/mg [10·1–13·7], figure 3). At the end of the study period, mean concentration of aflatoxin–albumin adducts in the intervention villages was less than 50% of that in control villages (8·0 pg/mg [7·0–9·2] vs 18·7 pg/mg [17·0–20·6]; difference between groups 10·7 pg/mg [10·1–11·1], p0·0001, figure 3).We also assessed the number of individuals with non-detectable concentrations of adducts (defined as 3 pg/mg) as an indicator of especially low exposure. At the time of harvest, adducts were not detectable in about a third of individuals in both intervention and control groups (figure 3). However, at 5 months postharvest (survey 3), only five (2%) people had nondetectable concentrations in the control villages compared with 47 (20%) of those in the intervention group (figure 3; difference between groups was 42 [32–60], p0·0001). Concentrations of aflatoxin B1 in groundnuts in household stores in intervention and control villages were consistent with the findings for aflatoxin–albumin adducts—ie, concentrations did not differ between groups at harvest, but then rose greatly in control villages compared with intervention villages (table 2). www.thelancet.com Vol 365 June 4, 2005
20 10 0 1
2 Survey points
3
Figure 2: Geometric mean aflatoxin–albumin adduct concentration in individual control (upper) and intervention (lower) villages at survey points Lines represent individual villages; lines of the same colour in both graphs do not represent the same village.
In the control villages the practices introduced in the intervention package were practised infrequently (table 1), apart from hand sorting. 52 (35%) families reported some attempt to remove visibly mouldy nuts before storage, although adduct concentrations were higher in those who practised this technique in control villages. Insecticide use was associated with a 50% lower mean adduct concentration at survey 2 (9·2 pg/mg [95% CI 6·3–13·5] vs 18·1 pg/mg [15·4–21·3], p=0·03) but not at survey 3. Other intervention measures were not associated with adduct concentrations in control villages, but all of which were also infrequently practised (table 1). Climate conditions (ie, rainfall, temperature, and humidity) did not differ substantially between intervention and control villages, and there was no consistent relation between these factors and aflatoxin–albumin adduct concentrations at the three surveys. Control farmers who delayed moving groundnuts from fields to the village immediately postharvest had higher mean concentrations of aflatoxin–albumin adducts at survey 2 than those who did not (20·0 pg/mg [95% CI 14·8–27·1 vs 13·0 pg/mg 1953
Articles
[9·4–17·9]; difference 7·0 pg/mg [5·4–8·6], p=0·05), an effect that was not significant at survey 3 (20·3 pg/mg vs 17·5 pg/mg). Groundnut consumption was frequent (ie, 85% in the preceding 48 h) at all surveys in both intervention and control villages. Maize was consumed less frequently than were groundnuts (between 5% and 45% reported consumption in the week before the five surveys). Overall, maize consumption was higher in the intervention group than in the control group but was not associated with aflatoxin–albumin adduct concentration (data not shown). Table 3 shows the effect of the intervention on aflatoxin–albumin adduct concentrations at both the individual level and the village mean level; adjusted mean concentrations were much the same for the two types of analyses. The adjusted mean adduct concentration in the intervention group was lower than that of the control group at survey 2, and less than half that in the control group at the end of the study (survey 3).
24
Mean aflatoxin–albumin (pg/mg)
Intervention Control 20 16 p0·0001 12 8 4 0
Non-detectable aflatoxin–albumin (%)
40
30
Discussion 20
p0·0001
10
0 1
2 Survey points
3
Figure 3: Concentration of aflatoxin–albumin adducts over study period Geometric mean and 95% CI of adducts in the intervention and control groups at the three main surveys (upper). Percentage and 95% CI of individuals with non-detectable adducts in the intervention and control groups at the three main surveys (lower). Aflatoxin B1 (parts per billion)
Survey 1 Survey 2 Survey 3
Control (n=30)
Intervention (n=30)
p*
11 (9–17) 22 (17–33) 55 (39–72)
9 (5–11) 9 (8–11) 17 (11–22)
·· 0·0001 0·0001
*Compared with corresponding control group by non-parametric test of medians. Data are median (IQR).
Table 2: Aflatoxin B1 in groundnuts at main survey points
This community-based intervention has shown a striking reduction in aflatoxin exposure by use of simple, low-technology postharvest practices in a rural subsistence-farming community. Exposure was more than halved 5 months after harvest in individuals from the intervention villages. Moreover, only 2% of individuals in the control villages had undetectable concentrations of aflatoxin–albumin adducts compared with about 20% in the intervention villages. In west Africa more than 98% of people test positive for this biomarker9 and thus a substantial number of individuals had their exposure reduced to very low levels by the intervention. Our observations might be an underestimate of the true benefit because exposure was monitored only until 5 months postharvest. Because aflatoxins continue to accumulate with long-term storage the effect of the intervention may have been greater later in the year, although interpretation of subsequent findings might have been complicated by the diminished availability of groundnuts at later times postharvest. The intervention strategy was based on restricting the spread and proliferation of aspergillus spores in storage. Methods such as those used in our intervention have
Aflatoxin–albumin adduct concentration (pg/mg) Survey 2
Individual level Village mean level
Survey 3
Control
Intervention
p*
Control
Intervention
p*
15·5 (9·5–25·0) 17·6 (16·1–19·3)
12·7 (7·8–20·7) 12·3 (11·2–13·5)
0·568 0·0001
18·4 (14·2–24·0) 18·7 (17·8–19·5)
8·0 (6·1–10·5) 8·2 (7·8–8·5)
0·0001 0·0001
*Compared with control group after adjustment for baseline adduct concentration and significant factors from analysis in control villages. Data are geometric mean (95% CI).
Table 3: Concentration of aflatoxin–albumin adducts at follow-up surveys
1954
www.thelancet.com Vol 365 June 4, 2005
Articles
been suggested for postharvest control in developing countries at the level of the small-scale local producers11,17 but have not been previously assessed at the individual level—which has been made possible by the availability of a blood-based biomarker of aflatoxin exposure.10 This scale of farming is still the most common practice in west Africa and improvements are thus of potential benefit to large populations across many countries. The intervention villages were chosen to be as similar as possible to the controls, although some degree of geographical separation was needed to avoid crossover of intervention practices. Villages in Kindia are dispersed settlements with houses scattered among fields, and control and intervention villages are indistinguishable in terms of culture and agricultural practice. Differences between the areas could have affected our findings, but we do not think that these effects would negate the strong results we obtained. Concentrations of aflatoxin–albumin adducts and the frequency of groundnut consumption were much the same between the intervention and control villages at the time of harvest. Other variables that differed between groups (eg, age, ethnic group, and maize consumption) did not correlate with adduct concentrations. This intervention was targeted only at the groundnut crop, because these nuts are the main cause of aflatoxin exposure in this region of Guinea.13,14 We were therefore able to test whether intervention against a single crop would have a substantial effect on overall aflatoxin exposure. The effect of the intervention on aflatoxin concentrations in the stored groundnuts (table 2) was consistent with those seen on adduct concentrations in individuals. Aflatoxin contamination can occur in the field, and therefore postharvest measures will only reduce, rather than eliminate, exposure. Nevertheless, even small reductions may be of major benefit in such high-exposure regions. The findings from our community-based intervention can be compared with those from chemoprevention studies18 on aflatoxin in China, in which use of chlorophyllin led to a 55% reduction in urinary aflatoxin biomarkers. However, such a chemoprevention approach would be difficult to implement in a sustainable way in affected communities. The reduction in aflatoxin–albumin adducts would be expected in the longer-term to be associated with a decrease in the burden of aflatoxin-related disease. These adducts and other aflatoxin biomarkers are associated with increased risk of hepatocellular carcinoma in adults19,20 and with impaired growth in young children.5–7 Our results should encourage other workers to assess the benefits of intervention on child health and cancer risk. Many of the 360 million chronic carriers of hepatitis B virus worldwide are living in countries with high aflatoxin exposure, and the www.thelancet.com Vol 365 June 4, 2005
implementation of simple postharvest measures is thus clearly of potential benefit in the reduction of hepatocellular carcinoma incidence.21 The simple intervention measures used were readily accepted by farmers, who recognise the problem of food spoilage. The measures were introduced as a package and consequently we did not test whether specific parts were more effective than others. However, it is noteworthy that the good practices of the intervention package were not widely used in control villages. Because the measures can be easily introduced as a package and each component will probably be beneficial, we think that there is little need to study the effectiveness of individual components. The large variations in aflatoxin–albumin adduct concentrations between villages are consistent with previous observations from west Africa.22 Data for the Kindia region formed the basis for our study design.13,14 Some of the variation may be a result of the known heterogeneity of aflatoxin contamination in a crop or of factors that we did not account for, including foodpreparation methods, quantities of groundnuts consumed (instead of frequency of consumption), and other dietary factors that could modulate aflatoxin metabolism. Key questions in the use of any intervention strategy in the developing world are those of cost and sustainability. The cost of these locally available materials was fairly low for each farmer—about US$50 per household, most of which was for the wooden pallets (about $10 each). Gross national product per head in Guinea is about $1100. These costs should be set against the benefits in terms of avoiding food spoilage and the improved market value of any surplus crop. Future studies need to address the sustainability of such intervention measures in the light of both economic and health benefits in these communities. In particular, the re-use of wooden pallets needs assessment, given their relative expense; so too do cheaper alternatives such as use of large stones to lift storage bags off the earthen floor. Our intervention study encourages the use of primaryprevention strategies at postharvest in rural west African communities. This intervention was on groundnuts, but similar postharvest principles apply to maize, and studies on this crop should be a priority for the region. There are many other potential intervention procedures against aflatoxin contamination, including chemoprevention, genetic modification of resistant crops, and biocompetition with non-aflatoxigenic strains of aspergillus.11 However, the substantial reductions in aflatoxin exposure recorded here by working with local farmers and by use of readily available materials and local agricultural expertise could be a rapid and inexpensive approach to reducing the burden of aflatoxin-associated disease in many parts of subSaharan Africa. 1955
Articles
12
Conflict of interest statement We declare that we have no conflict of interest.
13
Acknowledgments This work was supported by a grant from the NIEHS, USA ES06052. We thank Paola Pisani for comments on the original design of the study and Jean-Jacques Castegnaro for assistance in establishing the analysis of aflatoxins in groundnuts. Lisa Worrilow and Angela Rowley were involved in some of the laboratory analysis. A Sylla and M S Diallo acknowledge the support of ICRETT fellowships during the development of this project. References 1 World Health Organization International Agency for Research on Cancer. Some traditional herbal medicines, some mycotoxins, naphthalene and styrene. IARC Monogr Eval Carcinog Risks Hum 2002; 82: 1–556. 2 World Health Organization. Food additives series 40: safety evaluations of certain food additives and contaminants. Aflatoxins. WHO: Geneva, 1998: 359–468. 3 World Health Organization International Agency for Research on Cancer. Cancer in Africa: epidemiology and prevention. Liver cancer. IARC Sci Publ 2003; 153: 299–314. 4 Raisuddin S, Singh KP, Zaidi SI, Paul BN, Ray PK. Immunosuppressive effects of aflatoxin in growing rats. Mycopathologia 1993; 124: 189–94. 5 Gong YY, Cardwell K, Hounsa A, et al. Dietary aflatoxin exposure and impaired growth in young children from Benin and Togo: cross sectional study. BMJ 2002; 325: 20–21. 6 Gong YY, Egal S, Hounsa A, et al. Determinants of aflatoxin exposure in young children from Benin and Togo, West Africa: the critical role of weaning. Int J Epidemiol 2003; 32: 556–62. 7 Gong YY, Hounsa A, Egal S, et al. Postweaning exposure to aflatoxin results in impaired child growth: a longitudinal study in Benin, West Africa. Environ Health Perspect 2004; 112: 1334–38.
1956
8
Contributors P C Turner contributed to the design of the study, was responsible for supervision of the laboratory analysis, and was involved in writing the paper. A Sylla supervised and did the fieldwork, and did part of the aflatoxin analysis in food. Y Y Gong did part of the laboratory work and statistical analysis, and contributed to writing the paper. M S Diallo did the fieldwork and analysis of aflatoxins in food samples. A E Sutcliffe did analysis of aflatoxin-exposure biomarkers. A J Hall contributed to the design of the study, did statistical analysis, and contributed to writing the paper. C P Wild was responsible for the overall design and planning of the study, as well as for the preparation of the paper.
9 10 11
14
15
16 17
18
19 20
21
22
Turner PC, Moore SE, Hall AJ, Prentice AM, Wild CP. Modification of immune function through exposure to dietary aflatoxin in Gambian children. Environ Health Perspect 2003; 111: 217–20. Montesano R, Hainaut P, Wild CP. Hepatocellular carcinoma: from gene to public health. J Natl Cancer Inst 1997; 89: 1844–51. Wild CP, Turner PC. The toxicology of aflatoxin as a basis for public health decisions. Mutagenesis 2002; 17: 471–81. Wild CP, Hall AJ. Primary prevention of hepatocellular carcinoma in developing countries. Mutat Res 2000; 462: 381–93. Diallo MS, Sylla A, Sidibe K, Sylla BS, Trepo CR, Wild CP. Prevalence of exposure to aflatoxin and hepatitis B and C viruses in Guinea, West Africa. Nat Toxins 1995; 3: 6–9. Sylla A, Diallo MS, Castegnaro J, Wild CP. Interactions between hepatitis B virus infection and exposure to aflatoxins in the developmentof hepatocellular carcinoma: a molecular epidemiological approach. Mutat Res 1999; 428: 187–96. Turner PC, Sylla A, Diallo MS, Castegnaro JJ, Hall AJ, Wild CP. The role of aflatoxins and hepatitis viruses in the etiopathogenesis of hepatocellular carcinoma: a basis for primary prevention in Guinea-Conakry, West Africa. J Gastroenterol Hepatol 2002; 17: S441–S448. Chapot B, Wild CP. ELISA for quantification of aflatoxin–albumin adducts and their application to human exposure assessment. In: Warhol M, van Velzen D, Bullock GR, eds. Techniques in diagnostic pathology. San Diego: Academic Press, 1991: 135–55. Association of Official Analytical Chemists. Official methods of analysis. 15th edn. Arlington: AOAC, 1990: 1184–85. Crompton JAF, Tyler PS, Hindmarsh PS, Golob P, Boxall RA, Haines CP. Reducing losses in small farm grain storage in the tropics. Trop Sci 1993; 33: 283–318. Egner PA, Wang JB, Zhu YR, et al. Chlorophyllin intervention reduces aflatoxin–DNA adducts in individuals at high risk for liver cancer. Proc Natl Acad Sci USA 2001; 98: 14601–06. Wang LY, Hatch M, Chen CJ, et al. Aflatoxin exposure and risk of hepatocellular carcinoma in Taiwan. Int J Cancer 1996; 67: 620–25. Qian GS, Ross RK, Yu MC, et al. A follow-up study of urinary markers of aflatoxin exposure and liver cancer risk in Shanghai, People’s Republic of China. Cancer Epidemiol Biomarkers Prev 1994; 3: 3–10. Hall AJ, Wild CP. Liver cancer in low and middle income countries—prevention should target vaccination, contaminated needles, and aflatoxins. BMJ 2003; 326: 994–95. Allen SJ, Wild CP, Wheeler JG, et al. Aflatoxin exposure malaria and hepatitis B infection in rural Gambian children. Trans R Soc Trop Med Hyg 1992; 86: 426–30.
www.thelancet.com Vol 365 June 4, 2005
