ARTICLE IN PRESS

Int. J. Hyg. Environ.-Health 209 (2006) 345–358

1438-4639/$ - se

doi:10.1016/j.ijh

�CorrespondE-mail addr

www.elsevier.de/ijheh

Determinants of aflatoxin levels in Ghanaians: Sociodemographic factors,

knowledge of aflatoxin and food handling and consumption practices

Pauline Jollya,�, Yi Jianga, William Ellisb, Richard Awuahc, Obinna Nnedua,Timothy Phillipsd, Jia-Sheng Wange, Evans Afriyie-Gyawud, Lili Tange, Sharina Personf,Jonathan Williamsg, Curtis Jollyh

aDepartment of Epidemiology, School of Public Health, University of Alabama at Birmingham (UAB), 1665 University Boulevard,

RPHB 217, Birmingham, AL 35294-0022, USAbDepartment of Biochemistry, Kwame Nkrumah University of Science and Technology (KNUST), Kumasi, GhanacDepartment of Crop Science, KNUST, Kumasi, GhanadCollege of Veterinary Medicine, Texas A&M, College Station, TX, USAeDepartment of Environmental Toxicology, Texas Tech, Lubbock, TX, USAfDepartment of Medicine, UAB, Birmingham, AL, USAgCollege of Agricultural and Environmental Sciences, University of Georgia, USAhDepartment of Agricultural Economics, Auburn University, Auburn, AL, USA

Received 26 September 2005; received in revised form 3 February 2006; accepted 8 February 2006

Abstract

Aflatoxins are among the most potent of carcinogens found in staple foods such as groundnuts, maize and other oilseeds. This study was conducted to measure the levels of aflatoxin B1 (AFB1) albumin adducts in blood and aflatoxinM1 (AFM1) metabolite in urine of people in a heavy peanut and maize consuming region of Ghana and to examine theassociation between aflatoxin levels and several socio-demographic factors and food handling and consumptionpractices. A cross-sectional study was conducted in four villages in the Ejura Sekyedumase district of Ghana. A socio-demographic survey was administered to 162 participants. Blood samples were collected from 140 and urine samplesfrom 91 of the participants and AFB1 albumin-adduct levels in blood and AFM1 levels in urine were measured. HighAFB1 albumin-adduct levels were found in the plasma (mean7SD ¼ 0.8970.46 pmol/mg albumin; range ¼0.12–3.00 pmol/mg; median ¼ 0.80 pmol/mg) and high AFM1 levels in the urine (mean7SD ¼ 1,800.1472602.01 pg/mg creatinine; range ¼ non-detectable to 11,562.36 pg/mg; median ¼ 472.67 pg/mg) of most of theparticipants. There was a statistically significant correlation (r ¼ 0:35; p ¼ 0:007) between AFB1–albumin adductlevels in plasma and AFM1 levels in urine. Several socio-demographic factors, namely, educational level, ethnic group,the village in which participants lived, number of individuals in the household, and number of children in thehousehold attending secondary school, were found to be significantly associated with AFB1 albumin-adduct levels bybivariate analysis. By multivariate analyses, ethnic group (p ¼ 0:04), the village in which participants live (p ¼ 0:02),and the number of individuals in the household (p ¼ 0:01), were significant predictors of high AFB1 albumin-adducts.

e front matter r 2006 Elsevier GmbH. All rights reserved.

eh.2006.02.002

ing author. Tel.: +1205 934 1823; fax: +1 205 975 3329.

ess: jollyp@uab.edu (P. Jolly).

ARTICLE IN PRESSP. Jolly et al. / Int. J. Hyg. Environ.-Health 209 (2006) 345–358346

These findings indicate strongly that there is need for specifically targeted post-harvest and food handling andpreparation interventions designed to reduce aflatoxin exposure among the different ethnic groups in this regionof Ghana.r 2006 Elsevier GmbH. All rights reserved.

Keywords: Aflatoxin B1; Aflatoxin M1; Sociodemographic factors; Food consumption; Ghana

Introduction

Aflatoxins are a group of highly toxic metabolitesproduced by the common fungi Aspergillus flavus and A.

parasiticus (Gourama and Bullerman, 1995) and areamong the most potent of carcinogens found in foods.Although the most common aflatoxins (B1, B2, G1 andG2) occur naturally together in various foods indifferent proportions (Park et al., 2002), aflatoxin B1(AFB1) is usually the predominant and most toxic form.AFM1 is a major metabolite of AFB1 and is frequentlyexcreted in milk and urine of humans, dairy cattle andother mammals that have consumed food or feedcontaminated with aflatoxins (Gourama and Bullerman,1995). AFB1 has been shown to cause liver tumors in anumber of animal species (IARC, 1993) and isassociated with hepatocellular carcinoma in humans(Omer et al., 1998; Peers et al., 1976, 1987; Wild et al.,1992a), especially in people with hepatitis B virusinfection (Peers et al., 1976, 1987; Wild et al., 1992a;Henry et al., 1999). Consumption of moderate to highlevels of aflatoxin has also been linked with acute illnessresulting in death in humans (Ngindu et al., 1982;Bourgeois et al., 1971; Krishnamachari et al., 1975; Lyeet al., 1995; CDC, 2004). In the outbreak in Kenya fromJanuary to July 2004, 125 of 317 patients (39.4%) diedfrom acute aflatoxicosis (CDC, 2004). However, chronicaflatoxicosis results from continued consumption of lowto moderate levels of aflatoxins, and the effects areusually subtle.

The highest aflatoxin contamination is found ingroundnuts, maize, and other grains. In Ghana andother West African countries, most rural people dependon these foods for survival. Unfortunately, the basicstaples that are produced and consumed (and the excesstraded in urban and rural markets) are poorly handledand stored prior to marketing and are frequentlycontaminated by toxigenic fungi. Awuah and Kpodo(1996) reported high levels of aflatoxins (5.7–22,168 ppb)in market groundnut samples in Ghana. The sampleswere contaminated by a variety of fungi including A.

flavus. Mintah and Hunter (1978) reported that 50–80%of groundnut samples from the Accra area of Ghana,especially those emanating from the Northern andVolta regions, had aflatoxin levels that exceeded the30 mg/kg hazard level recommended by the FAO/WHO/

UNICEF Protein Advisory Board (Frazier and Westh-off, 1988).

One study on aflatoxin levels in maize conducted inGhana examined only 15 samples collected from majorprocessing sites in Accra and reported the presence ofaflatoxin in 8 (53%) samples (Kpodo et al., 2000). Thelevel of aflatoxin ranged from 2 to 662mg/kg�1. A studyconducted in the neighboring country of Benin showedthat aflatoxin levels in maize varied by agro-ecologicalzones and length of storage time (Hell et al., 2000). Maizesamples with higher aflatoxin levels came from thesouthern Guinea and Sudan savannas of Benin. Storagefor 3–5 months, insect damage and use of local plants forprotection during storage were associated with higheraflatoxin levels (Hell et al., 2000). The Ashanti Region ofGhana falls in the transitional zone between the forest andsavanna and has a bimodal rainfall pattern (US LibraryCongress, 2005) and two maize growing seasons similar tothe forest/savanna mosaic region of Benin (Hell et al.,2000). Dried yam chips that had a moisture content above15% were also found to be contaminated with high levelsof aflatoxin in Benin (Mestres et al., 2004).

In some East and West African countries 75–100% ofindividuals tested have been found to be positive for theAFB1–albumin adduct in their blood with levels rangingfrom 45 to 720 pg AFB1–lysine eq./mg albumin/mlblood (Wild et al., 1992a). Furthermore, unbornchildren are exposed to aflatoxins in utero and theAFB1–albumin adduct has been found in umbilical cordblood (Wild et al., 1991). Gong et al. (2003) showed thatthere is an increase in aflatoxin albumin levels with age;in children 1–3 years it is related to weaning status. Ahigher consumption of maize was correlated with higheraflatoxin albumin adduct level. Ankrah et al. (1994)examined AFB1 and AFG1 in serum, urine and fecalspecimens from 40 volunteers from the Greater Accraregion of Ghana and detected aflatoxins in one or moreof the specimens from 35% of the study participants.However, no studies have been conducted on aflatoxinlevels in people in the Ashanti region of Ghana (apredominantly agricultural and heavy maize producingarea) or on the sociodemographic factors that determinehigh aflatoxin levels. In this study, we examined thelevels of AFB1–albumin adducts in blood and AFM1

metabolite in urine of people living in this region andidentified factors associated with high AFB1 levels.

ARTICLE IN PRESSP. Jolly et al. / Int. J. Hyg. Environ.-Health 209 (2006) 345–358 347

Materials and methods

Participant recruitment, data and sample collection

A cross-sectional field survey of demographic, foodhandling, consumption, health status, health history,and other lifestyle characteristics was conducted fromJune to August 2002 in four villages (Dromankuma,Nkwanta, Hiawoanwu and Kasei) in the Ejura Sekyer-edumase district, Ashanti region of Ghana. This is arural area of Ghana where most individuals aresubsistence farmers using traditional farming practices.The Ejura Sekyedumase district, however, has two rainyor wet seasons which occur from about April throughJune and from September through November, with amean annual rainfall of 1400mm (US Library Congress,2005). The main crops grown in this area are yams,maize, cassava, groundnuts, beans, tobacco, cotton andvegetables (Cudjoe, 2003).

The Ejura District Health Director and other DistrictHealth staff facilitated the study by introducinginvestigators to the opinion community leaders andother people in the villages. The purpose of the studywas explained and after all questions were answered,those present were informed of the time and place wherethe study would be conducted. A total of 162 individualsvolunteered to participate in the study and completedthe survey. A 20ml blood sample was donated by 140 ofthe 162 participants. Following this, participants weregiven sterile covered cups to collect their first urine thenext morning, and the urine samples were collectedpromptly by the study staff. Urine samples wereobtained from 91 participants. Blood and urine sampleswere processed in the laboratories of the Kumasi Centerfor Collaborative Research (KCCR) in Tropical Med-icine, School of Medical Sciences, Kwame NkrumahUniversity of Science and Technology (KNUST). Theblood was separated into plasma and peripheral bloodmononuclear cells (PBMCs) (Jolly, 1997) and thesamples were stored frozen at �80 1C (plasma or urine)or in liquid nitrogen (PBMCs) and shipped to theUniversity of Alabama at Birmingham (UAB) foranalysis. The UAB Institutional Review Board and theMedical School Ethics Committee of the KNUST gaveapproval for the study.

Determination of AFB1–albumin adduct levels in

plasma by radioimmunoassay (RIA)

AFB1–albumin adducts in plasma from study parti-cipants were determined by RIA (Wang et al., 1996).The assay measures aflatoxin that is covalently bound inperipheral blood albumin (Nyathi et al., 1987) andreflects aflatoxin exposure in the previous 2–3 months.Plasma samples were concentrated by high-speed

centrifugal filtration and the concentrated protein re-suspended in phosphate buffered saline (PBS). Plasmaalbumin was determined by a bromocresol purple dyebinding method (Sigma, St. Louis, MO) and the amountof total protein determined using the Bradford proce-dure (Pierce Biotechnology Inc., Rockport, IL). Totalprotein per sample was then digested with Pronase(Calbiochem, La Jolla, CA) and bound aflatoxinextracted with acetone. The RIA procedure (Wang etal., 1996) was used to quantify AFB1–albumin adductsin duplicate plasma protein digests each containing 2mgprotein. Nonspecific inhibition in the assay was deter-mined by processing pooled normal human plasmastandards (Sigma, St. Louis, MO). The standard curvefor the RIA was determined using a nonlinear regressionmethod (Gange et al., 1996). The concentration ofalbumin, total protein, and AFB1–albumin adduct inindividual plasma samples was calculated and the valuesexpressed as the amount of AFB1 per mg albumin(Wang et al., 1996). The detection limit of the assay was0.01 pmol/mg albumin.

Determination of the aflatoxin M1 metabolite in

urine

Metabolic levels of AFM1 in urine were quantitatedby high performance liquid chromatography followingimmunoaffinity cleanup of samples. AFM1 metabolitewas used as a valid indicator of short-term exposure toaflatoxins due to its prevalence in the urine and its dose-dependent relationship with AFB1 intake in the diet(Groopman et al., 1992a).

Affinity chromatography cleanup procedures andHPLC analysis were based on methodologies describedby Groopman et al. (1992b), with modifications of Sarret al. (1995) and Wang et al. (1999). Each of the urinesamples (5.0ml) was adjusted to an acidic pH with0.5ml of 1.0M ammonium formate (pH 4.5), and thevolume was increased to 10ml with water and vortexed.The sample was then applied to a 1ml preparativeAflatest P monoclonal antibody column (VicamLP,Watertown, MA) and aflatoxin eluted at a flow rate ofapproximately 0.3ml/min as described previously(Wang et al., 1999).

For HPLC analysis, a Waters HPLC system (WatersCorporation, Milford, MA) with fluorescence detectioncapabilities was used. A 250mm� 4.6mm LiCrospherRP-18 endcapped column with pore size 100 (A andparticle size 5 mm (Alltech Associates, Deerfield, IL) wasused to resolve aflatoxin metabolites. The mobile phaseconsisted of 22% ethanol in water which was bufferedwith 20mM ammonium formate (pH 3.0). Chromato-graphic separation of aflatoxins was achieved byisocratic elution of the mobile phase for 20min. Sampleswere injected (100 ml) on the column and the elution rate

ARTICLE IN PRESSP. Jolly et al. / Int. J. Hyg. Environ.-Health 209 (2006) 345–358348

was 1.0ml/min. AFM1 peaks were detected at aretention time of approximately 15.4min. The limit ofdetection for this method was 10 pg/ml of urine forAFM1. The identity of AFM1 was confirmed by GC/MSby comparing standard solutions of the metabolite inacetonitrile. Positive identification of AFM1 was basedon retention times and spectra compared with those ofstandards (Sigma, Milford, MO). Urinary concentra-tions of AFM1 metabolites were expressed as pg/mgcreatinine in order to correct for variations in urinedilution among individual samples.

Statistical analysis

Data from the questionnaires were entered intoMicrosoft excel and imported into the SAS softwarepackage version 8.2 (Statistical Analytical System, Gary,NC) for analysis. Basic descriptive statistics wereconducted. Aflatoxin B1 albumin adduct levels weredivided into 2 groups based on the median level of0.8 pmol/mg albumin. Individuals with aflatoxin levelsof 0.8 pmol/mg albumin or above were placed in thehigh aflatoxin group and those with levels less than0.8 pmol/mg in the low aflatoxin group. Crude oddsratio estimates for the relationship between variousfactors and high aflatoxin levels were determined bygenerating 2� 2 contingency tables. In addition, multi-variable logistic regression models were created toexamine the relationship between age, gender, educa-tion, number of individuals in the household, village,ethnicity, sorting of food, number of children insecondary school and aflatoxin levels. To furtherexamine food preference, sorting practices and ethnicgroup within village, Chi square tests of associationwere performed.

Results

Sociodemographic characteristics of study

participants

The average age of the respondents was 40.8 yearswith a median age of 36.5 years (range 19–86 years).Slightly more than half (50.6%) of the participants werefemales and 49.4% were males (Table 1). Majority ofparticipants (43.2%) were from Dromankuma, 24.7%were from Hiawoanwu, 19.8% were from Nkwanta and12.3% were from Kasei. This reflected the size of thevillages with Dromankuma being the largest and Kaseibeing relatively smaller. Most of the participants(36.1%) belonged to the Akan ethnic group, 24.7%belonged to the Dagbani, Basare/Basali, Gonja andKonkomba group, 13.9% to the Grumma, Busanga,Bimoba and Kusasi group, 10.8% to the Moshi, Hausa,Waa, Kokama, Busani, Gungui and Zugu group, and

14.6% to the Ewe, Dagati, Sisala, Wala, Grussi, Frafraand unknown group. These ethnic group categories arethe ones existing and recognized in the region and thecategorization has a cultural basis. Individuals recognizethe ethnic group of their mothers or fathers dependingon whether the group is matriarchal or patriarchal in itsorientation. Therefore, each participant gave only oneethnic group when asked about ethnicity.

There was a significant difference (po0:0001) in theethnic make up of people in the different villages. Themajority of participants from Nkwanta and Kasei(62.5% and 65.0% respectively) were Akan (Table 2).The Ewe, Dagati, Sisala, Wala, Grussi, Frafra andunknown group was the next predominant group inboth districts (21.9% in Nkwanta and 15.0% in Kasei).In contrast, the majority of participants from Droman-kuma (28.8%) and Hiawoanwu (47.5%) belonged to theDagbani, Basare/Basali, Gonja and Konkomba ethnicgroup (Table 2). The second largest group in Droman-kuma (22.7%) was the Gruma, Busanga, Bimoba andKusasi group while the second largest group inHiawoanwu (32.5%) was the Akan.

Less than half (48.7%) of participants reported thatthey had received some type of formal education.Approximately 22.1% of those with formal educationhad only primary school education. A statisticallysignificant difference between males and females withrespect to education was obtained (Table 1). Asignificantly higher proportion of men than womenreceived formal education.

All participants reported that they worked. Most(58.1%) reported farming as their occupation; this wasfollowed by trading (15.6%), farming and trading(5.0%), farming and other 5.0%), and numerous otheroccupations (such as charcoal selling, driving, store-keeping and hair dressing) most with frequencies notexceeding 1% of the sample population. A significantlyhigher proportion of men than women reported thatthey were farmers (Table 1).

The families of majority of the participants (49.0%)had 1–5 individuals in their households. However, a fairproportion of participants (36.3%) had 6–10 householdmembers; 9.6% of participants had 11–15 householdmembers, and 5.1% had 16 or more householdmembers. Approximately 28% of households had threeor more children p 9 years, and 9.7% of householdshad three or more children 10–15 years old. Averagemonthly household income for the families was 299,366cedis (US $ 37.42) (1 US $ ¼ 8000 cedis) and rangedfrom 16,666 to 1,823,334 cedis (US $2.00 to $ 227.92).

Main foods consumed and participants’ awareness of

spoilt food and knowledge of AF

Most participants reported that on a daily basis theyeat groundnuts (63.3%), maize (85.7%) and cassava

ARTICLE IN PRESS

Table 1. Sociodemographic characteristics of the study population by gender

Male n ¼ 80 Female n ¼ 82 Male+Female n ¼ 162 p-value

n (%) n (%) n (%)

Agea

19–29 20 (25.0) 28 (34.1) 48 (29.6)

30–39 21 (26.3) 21 (25.6) 42 (25.9)

40–49 19 (23.8) 13 (15.9) 32 (19.8)

50–69 13 (16.3) 11 (13.4) 24 (14.8)

X70 7 (8.8) 9 (11.0) 16 (9.8) 0.5808

Village

Nkwanta 17 (21.3) 15 (18.3) 32 (19.8)

Kasei 8 (10.0) 12 (14.6) 20 (12.3)

Hiawoanwu 23 (28.8) 17 (20.7) 40 (24.7)

Dromankuma 32 (40.0) 38 (46.3) 70 (43.2) 0.5079

Ethnicity

Group1 24 (30.8) 33 (41.3) 57 (36.1)

Group2 15 (19.2) 7 (8.8) 22 (13.9)

Group3 19 (24.4) 20 (25.0) 39 (24.7)

Group4 9 (11.5) 8 (10.0) 17 (10.8)

Group5 11 (14.1) 12 (15.0) 23 (14.6) 0.5277

Education

No formal 31 (39.2) 51 (63.0) 82 (51.3)

Formal 48 (60.8) 30 (37.0) 78 (48.7) 0.0026*

Formal education level

Primary 11 (22.9) 6 (20.7) 17 (22.1)

Junior secondary 11 (22.9) 10 (34.5) 21 (27.3)

Middle school 16 (33.3) 10 (34.5) 26 (33.8)

Secondary 10 (20.8) 3 (10.3) 13 (16.9)

University 0 (0.0) 0 (0.0) 0 (0.0) 0.5365

Occupation

Farmer 54 (68.4) 39 (48.2) 93 (58.1)

Trader 3 (3.8) 22 (27.2) 25 (15.6)

Farmer/trader 0 (0.0) 8 (9.9) 8 (5.0)

Farmer/other 7 (8.8) 1 (1.2) 8 (5.0)

Other 15 (19.0) 11 (13.6) 26 (16.3) o 0.0001*

Living Situation

Own home 40 (50.6) 41 (51.3) 81 (50.9)

Rent home 17 (21.5) 15 (18.8) 32 (20.1)

Live with family/friends 22 (27.9) 24 (30.0) 46 (28.9) 0.8967

*Significant at po0:05.Group 1: Akan.

Group 2: Gruma, Busanga, Bimoba & Kusasi.

Group 3: Dagbani, Basare/Basali, Gonja & Konkomba.

Group 4: Moshi, Hausa, Waa, Kokama, Busani, Gungui & Zugu.

Group 5: Ewe, Dagati, Sisala, Wala, Grussi, Frafra, and unknown.

Numbers may not always add up to total due to missing responses.aAverage age ¼ 40.8 years (median ¼ 36.5 years; range ¼ 19–86 years).

P. Jolly et al. / Int. J. Hyg. Environ.-Health 209 (2006) 345–358 349

(81.6%) (Table 3). Almost 90% of respondents reportedgrowing their own food and slightly more than halfreported treating maize with pesticide before storage.The pesticides used by the farmers for treating maize isActellic-EC (Pirimiphos methyl) and Actellic super(Pirimiphos methyl with Cypermethrin; Orion Crop

Company, Aukland, New Zealand). Both of thesepesticides are effective against insects during storage ofmaize and are affordable since they are sold to farmersin small measures. Farmers have been educated in theproper use of these pesticides and consider them safe.After treatment of maize, the pesticides breakdown as

ARTICLE IN PRESS

Table 2. Number and percent of the different ethnic groups in each village

Ethnicity Nkwanta N (%) Kasei N (%) Hiawoanwu N (%) Dromankuma N (%) p-value

Group 1 20 (62.5) 13 (65.0) 13 (32.5) 11 (16.6)

Group 2 3 (9.4) 2 (10.0) 2 (5.0) 15 (22.7)

Group 3 1 (3.1) 0 (0.0) 19 (47.5) 19 (28.8) o0.0001

Group 4 1 (3.1) 2 (10.0) 6 (15.0) 8 (12.1)

Group 5 7 (21.9) 3 (15.0) 0 (0.0) 13 (19.7)

Group 1: Akan.

Group 2: Gruma, Busanga, Bimoba and Kusasi.

Group 3: Dagbani, Basare/Basali, Gonja and Konkomba.

Group 4: Moshi, Hausa, Waa, Kokama, Busani, Gungui and Zugu.

Group 5: Ewe, Dagati, Sisala, Wala, Grussi and Frafra and unknown.

Table 3. Food consumed by participants, storage, sorting

and identification of spoiled food and knowledge of aflatoxin

n %

Foods consumed every day

Groundnut 93 63.3

Maize 126 85.7

Cassava 120 81.6

Sorghum 5 3.5

Millet 27 18.9

Grow food consumed

Yes 131 89.7

No 15 10.3

Storage method

With pesticide 69 53.5

Without pesticide 60 46.5

Can you identify spoilt maize

Yes 126 89.4

No 15 10.6

How do you identify spoilage

Rot 63 51.6

Insects and rot 40 32.8

Insects 19 15.6

Do you sort your food

Yes 106 76.8

No 32 23.2

What do you do with spoilt grains

Throw it away 107 84.9

Feed it to animals 10 7.9

Eat it 9 7.1

Knowledge of aflatoxin

Yes 11 7.8

No 131 92.3

Knowledge of foods with which aflatoxin is associated?

Yes 6 4.3

No 135 95.7

Knowledge of health effects of aflatoxin

Yes 7 5.1

No 130 94.9

P. Jolly et al. / Int. J. Hyg. Environ.-Health 209 (2006) 345–358350

the maize is dried in the sun. No pesticides are used forpeanuts in storage.

Most participants (90%) reported that they were ableto identify unwholesome grains. Unwholesomeness ingrains was determined by signs of deterioration such asrot (51.6%), infestation with insects accompanied by rot(32.8%), and presence of insects only (15.6%). Majorityof participants (76.8%) stated that they sort their grains,and 84.9% said that they discard spoilt grains. Smallerpercentages of respondents stated that they either feedspoilt food to animals (7.9%) or eat it (7.1%). Mostrespondents (92.3%) had never heard of aflatoxin, didnot know whether it causes illness (94.9%), nor knewthe foods with which it is associated (95.7%) (Table 3).

Aflatoxin B1–albumin adduct levels (pmol/mg

albumin) of study participants

AFB1–albumin adduct levels were determined for the140 study participants who provided blood samples.Unfortunately, we were unable to collect blood fromparticipants in Kasei and could not determine AFB1

levels for this village. AFB1–albumin adduct levelsranged from 0.12 to 3.00 pmol/mg albumin with a meanof 0.8970.46 and median of 0.80 pmol/mg albumin(Table 4, Fig. 1). All participants were positive forAFB1–albumin adducts and approximately one thirdhad AFB1 levels higher than the mean. The level ofAFB1 in normal human plasma samples (Sigma, St.Louis, MO) used as control in this assay was 0.28 pmol/mg albumin. Also, because the monoclonal antibodyused in the assay targets the whole AFB1 albuminadduct, the AFB1 values that were obtained are higherthan those reported in studies conducted with mono-clonal antibody that target the AFB1–lysine adduct.

Aflatoxin M1 levels in urine of study participants and

correlation between AFB1 and AFM1 levels

AFM1 levels were determined for the 91 studyparticipants who provided urine samples. Importantly,

ARTICLE IN PRESS

Table 4. Descriptive statistics of aflatoxin B1 (AFB1) albumin

adduct and aflatoxin M1 (AFM1) levels in study participants

AFB1 (pmol/mg

albumin)

AFM1 (pg/mg

creatinine)

Number 140 91

Mean7Standard

deviation

0.8970.46 1800.1472602.01

Median 0.80 472.67

Range 0.12–3.00 0–11562.36

Percentiles

25 0.59 186.20

50 0.80 472.67

75 1.10 2595.59

0

10

20

30

40

50

60

70

80

0-0.49 0.5-0.99 1.0-1.49 1.5-1.99 ≥2.0

AFB1 levels in plasma (pmol/mg Albumin)

Nu

mb

er o

f P

arti

cip

ants

Fig. 1. Aflatoxin B1 albumin adduct levels in plasma of study

participants (n ¼ 140). Mean ¼ 0.8970.46 pmol/mg albumin.

Median ¼ 0.80 pmol/mg albumin.

P. Jolly et al. / Int. J. Hyg. Environ.-Health 209 (2006) 345–358 351

83 of the 91 participants (91.2%) were positive forAFM1. AFM1 levels ranged from nondetectable to11,562.36pg/mg creatinine with a mean of 1800.1472602.01 and median of 472.67pg/mg creatinine (Table 4,Fig. 2). Approximately one third of the study participantshad greater than the mean level of AFM1 in urine. Therewas a statistically significant correlation (r ¼ 0:345;p ¼ 0:007) between AFB1–albumin adduct levels inplasma and AFM1 levels in urine for 64 studyparticipants who donated both blood and urine inAugust 2002 (Fig. 3).

Association of sociodemographic characteristics with

AFB1 albumin adduct levels

When sociodemographic characteristics of the studygroup were examined according to AFB1 levels (lowAFB1o0.80 pmol/mg albumin and high AFB1X0.80 p-mol/mg albumin), it was found that level of education,ethnic group and the participants’ villages were sig-

nificantly associated with AFB1 albumin adduct levels(Table 5). Participants who had primary education orless were 2.29 times more likely to have high AFB1 levelscompared with those with a secondary education.Participants from the Dagbani, Basare/Basali, Gonjaand Konkomba and the Gruma, Busanga, Bimoba andKusasi ethnic groups were 3.66 and 3.13 times as likelyto have high AFB1 levels compared to Akans. Also,participants from the Dromankuma and Hiawoanwuvillages were 2.64 and 2.45 times more likely to havehigh levels of AFB1 compared to those from Nkwanta.When the data from Dromankuma and Hiawoanwuwere analyzed separately, education was still a signifi-cant predictor of high aflatoxin level and no significantdifference was observed between ethnic groups. Thislatter finding was not surprising since Dromankuma andHiawoanwu are more similar in ethnic group make-upthan Nkwanta and Kasei. Interestingly, for Dromanku-ma and Hiawoanwu, age was a significant predictor ofhigh aflatoxin levels. For these two villages, there was atrend toward increase in high aflatoxin levels as ageincreased from 19 to 29 years to X70 years. Those X70years were 8.84 times more likely to have high aflatoxinlevels than those 19–29 years (p ¼ 0:049).

Association of family socio-demographic

characteristics with AFB1 albumin adduct levels

When the family socio-demographic characteristicswere examined according to low or high AFB1 levels, thenumber of individuals in the household and the numberof children in secondary school in the household werefound to be significantly associated with high AFB1

levels (Table 6). Study participants from householdswith more than 5 members were 2.39 times more likelyto have high AFB1 levels compared to those fromhousehold with 5 or less family members. Also,participants from households with one or more childrenin secondary school were 4.58 times more likely to havehigh AFB1 levels compared to those from householdswith no children in secondary school.

Multivariate analysis of village, ethnicity, other

socio-demographic characteristics and high AFB1

levels

When multivariate analysis of village and demo-graphic characteristics with high AFB1 level wasconducted, the village in which participants lived andthe number of individuals in a household weresignificant predictors of high AFB1 levels (Table 7).Participants from Dromankuma were 3.27 times aslikely to have high levels of AFB1 compared to thosewho lived in Nkwanta. Although only marginallysignificant, participants from Hiawoanwu were 2.54

ARTICLE IN PRESS

0

5

10

15

20

25

30

35

<100 100-499 500-999 1000-1499 1500-1999 2000-2499 2500-2999 3000-3499 3500-3999 4000-4499 4500-4999 ≥5000

AFM1 Levels in Urine (pg M1/mg Creatinine)

Nu

mb

er o

f P

arti

cip

ants

Fig. 2. Aflatoxin M1 levels in urine of study participants (n ¼ 91). Mean ¼ 1800.1472602.01 pg AFM1/mg creatinine.

Median ¼ 472.67 pg/mg creatinine.

Plasma AFB1 Level (pmol /mg Albumin)

2.52.01.51.00.50.0

Urin

e A

FM

1 Le

vel (

pg /m

g C

reat

inin

e)

12000

10000

8000

6000

4000

2000

0

−2000

r = 0.345 p = 0.007

Fig. 3. Correlation between AFB1 levels in plasma and AFM1

levels in urine of study participants (n ¼ 64).

P. Jolly et al. / Int. J. Hyg. Environ.-Health 209 (2006) 345–358352

times as likely to have high levels of AFB1 compared tothose from Nkwanta. Also, participants from house-holds with more than 5 family members were 3 times aslikely to have high AFB1 levels compared to those fromhouseholds with less than 5 members. There was atendency (p ¼ 0:0747) for those with primary or no

education to be twice as likely to have high AFB1 levelscompared to those with secondary education or above,and for males (p ¼ 0:0880) to have lower AFB1 levelsthan females.

Although non-significant, there was also a tendency forethnicity to be a predictor of high AFB1 levels (Table 7).Participants belonging to the Dagbani, Basare/Basali,Gonja and Konkomba ethnic group tended to be 2.8times as likely (p ¼ 0:0684) and those from the Gruma,Busanga, Bimoba and Kusasi group tended to be 3.1times as likely (p ¼ 0:1088) to have high AFB1 levelscompared to those from the Akan group (Table 7).

Discussion

The responses given by participants in this surveyshow clearly that most participants do not identifyfungal contamination of grains until there are obvioussigns of spoilage such as discoloration, insect infestationor rotting. Further, unwholesome grains may be eaten(boiled, roasted), or processed into other products andconsumed (Table 3). The overwhelming majority of

ARTICLE IN PRESS

Table 5. Bivariate analysis of socio-demographic factors in relation to high Aflatoxin levels

Low aflatoxin N(%) High aflatoxin N(%) OR (95% CI) p-value

Age

19–29 24 (34.3) 21 (30.0) Reference

30–39 19 (27.1) 14 (20.0) 0.84 (0.34–2.08) 0.7099

40–49 14 (20.0) 15 (21.4) 1.22 (0.48–3.12) 0.6710

50–69 6 (8.6) 12 (17.1) 2.28 (0.73–7.16) 0.1559

X70 7 (10.0) 8 (11.4) 1.31 (0.41–4.21) 0.6550

Gender

Female 29 (41.4) 40 (57.1) Reference

Male 41 (58.6) 30 (42.9) 0.53 (0.27–1.04) 0.0630

Level of Education

Secondary 32 (47.8) 20 (28.6) Reference

Primary/none 35 (52.2) 50 (71.4) 2.29 (1.13–4.63) 0.0207*

Ethnicity

Group 1 27 (40.9) 16 (22.9) Reference

Group 2 7 (10.6) 13 (18.6) 3.13 (1.04–9.49) 0.0432*

Group 3 12 (18.2) 26 (37.1) 3.66 (1.45–9.19) 0.0059*

Group 4 10 (15.2) 5 (7.1) 0.84 (0.24–2.91) 0.7882

Group 5 10 (15.2) 10 (14.3) 1.69 (0.58–4.93) 0.3391

Villagea

Nkwanta 21 (30.0) 10 (14.3) Reference

Hiawoanwu 18 (25.7) 21 (30.0) 2.45 (0.92–6.54) 0.0736

Dromankuma 31 (44.3) 39 (55.7) 2.64 (1.09–6.42) 0.0321*

Own Home

No 37 (54.4) 29 (42.0) Reference

Yes 31 (45.6) 40 (58.0) 1.65 (0.84–3.23) 0.1470

*Significant at po0:05.Group 1: Akan.

Group 2: Gruma, Busanga, Bimoba and Kusasi.

Group 3: Dagbani, Basare/Basali, Gonja and Konkomba.

Group 4: Moshi, Hausa, Waa, Kokama, Busani, Gungui and Zugu.

Group 5: Ewe, Dagati, Sisala, Wala, Grussi, Frafra and unknown.aAflatoxin B1 levels were not determined for participants from Kasei.

P. Jolly et al. / Int. J. Hyg. Environ.-Health 209 (2006) 345–358 353

participants did not know of aflatoxin let alone itsharmful health effects. Given the level of poverty andthe lack of awareness of the health effects of consumingaflatoxin-contaminated foods, the amounts of contami-nated groundnuts and maize consumed are likely to bevery high. Although food samples were collected fromparticipants in this study, aflatoxin levels in food werenot determined because of difficulties encountered withequipment and reagents in conducting the analysis inGhana. Wild et al. (1992b) and Egal et al. (2005) found asignificant correlation between dietary intake of aflatox-in and level of AFB1–albumin adduct in plasma samplescollected from the Gambia. Wang et al. (2001) reportedthat average daily intake of AFB1 from food at X14 mgresulted in an average AFB1–albumin adduct level of1.2 pmol/mg albumin in a high-risk population of livercancer patients in Fusui, China. Based on their data, theaverage aflatoxin consumed with food by participants inthis study is approximately 10 mg AFB1 per day. Thus,

the levels of AFB1 biomarkers found in this study arehigher than those reported from the Gambia (Wild etal., 1992b; Egal et al., 2005), Benin (Gong et al., 2003)and the United Kingdom (Turner et al., 1998), but arelower than levels found in areas at high-risk for livercancer in China (Wang et al., 1996, 1999, 2001), wherethe average AFB1–albumin adduct level is usually over1.0 pmol/mg albumin and urinary AFM1 level is over2000 pg/mg creatinine.

The levels of AFM1 found in urine in this study arecomparable to previously reported values for AFM1 inhuman urines with averages of 4.2 ng/ml found inZimbabwe (Nyathi et al., 1987) and a range of0.17–5.2 ng/ml reported in Shanghai (Groopman et al.,1992a). Based on urine volumes, our findings indicate arange from nondetectable to 17.2 ng AFM1/ml urinewith an average level of 1.5 ng/ml urine. Differences inurinary levels of AFM1 are dependent on factors such asvariations in sample size, diet, aflatoxin exposure levels,

ARTICLE IN PRESS

Table 6. Bivariate analysis of family socio-demographic factors in relation to aflatoxin levels

Low aflatoxin n (%) High aflatoxin n (%) OR (95% CI) p-value

Number of individuals in household

1–5 39 (59.1) 26 (37.7) Reference

45 27 (40.9) 43 (62.3) 2.39 (1.20-4.77) 0.0128*

Number of children o9 years in household

0–1 35 (53.0) 28 (41.8) Reference

41 31 (47.0) 39 (58.2) 1.57 (0.79–3.12) 0.1943

Number of children 10–15 years in household

None 37 (56.1) 28 (41.8) Reference

X1 29 (43.9) 39 (58.2) 1.77 (0.89–3.53) 0.0998

Number of children in Primary school

None 34 (51.5) 25 (37.9) Reference

X1 32 (48.5) 41 (62.1) 1.74 (0.87–3.48) 0.1151

Number of children in secondary school

None 63 (95.5) 55 (82.1) Reference

X1 3 (4.6) 12 (17.9) 4.58 (1.14–26.32) 0.0256*

Any family member drink alcohol

No 57 (91.9) 45 (80.4) Reference

Yes 5 (8.1) 11 (19.6) 2.79 (0.81–10.90) 0.1045

Any family member smoke cigarettes

No 57 (91.9) 52 (94.6) Reference

Yes 5 (8.1) 3 (5.5) 0.66 (0.10–3.58) 0.7210

Family Income (cedis)

o100,000 12 (21.4) 7 (14.3) Reference

o200,000 14 (25.0) 9 (18.4) 1.10 (0.32–3.86) 0.8792

o355,000 13 (23.2) 19 (38.8) 2.51 (0.78–8.06) 0.1236

X355,000 17 (30.4) 14 (28.6) 1.41 (0.44–4.55) 0.5636

*Significant at po0:05.

P. Jolly et al. / Int. J. Hyg. Environ.-Health 209 (2006) 345–358354

methods of urine collection, aflatoxin analyses, urineconcentrations, genetic susceptibility to aflatoxins,health and nutritional status of the individual. However,aflatoxin uptake in the diet should be the crucialdeterminant of urinary AFM1 levels.

This study sample was a convenience sample and isprone to all the limitations of convenience sampling.However, the sample may be underestimating theaflatoxin problem since people who self-select toparticipate may be those who want to know about theirhealth. Since the people in the population are similar interms of resources and diet, uniformity of population ismore likely than in a developed country such as theUSA which has great diversity in its population.Concern about generalizability is, therefore, not thesame as in the US population.

Although we did not examine aflatoxin levels duringdifferent seasons, aflatoxin–albumin levels have beenshown to vary by season in the Gambia, withsignificantly higher levels during the dry season whenstudy participants were consuming stored groundnuts

harvested at the end of the wet season of the previousyear and storted under hot, humid conditions thatfavored fungal growth and mycotoxin production (Allenet al., 1992; Wild et al., 2000; Turner et al., 2000).Similar findings were reported from Benin with highestblood contamination in children observed at the end ofthe rainy season in October (Gong et al., 2004).

Four separate seasons occur in South and southwestGhana (US Library Congress, 1994). Heavy rains fallfrom about April through late June followed by a shortdry period (July–August). There is another rainy seasonfrom September through November which is followedby the longer harmattan dry season that lasts fromDecember through March. During the time of samplecollection participants were eating mainly crops har-vested at the end of the September – November rainyseason (usually in December – January) and stored.They were also consuming some fresh crops reaped lateJune to July after the April – June rainy season. Thestored food is likely to be the major source of aflatoxincontamination.

ARTICLE IN PRESS

Table 7. Multivariate analysis of community and demo-

graphic characteristics and high AFB1 levels

OR (95% CI) p-value

Villagea

Nkwanta Reference

Hiawoanwu 2.54 (0.86–7.47) 0.0897

Dromankuma 3.27 (1.20–8.93) 0.0206*

Age

19–29 Reference

30–39 0.85 (0.30–2.40) 0.7534

40–49 0.55 (0.15–1.93) 0.3486

50–69 1.13 (0.25–5.02) 0.8726

X70 0.56 (0.12–2.60) 0.4589

Number of individuals in household

1–5 Reference

45 3.07 (1.19–7.91) 0.0199*

Gender

Female Reference

Male 0.51 (0.24–1.10) 0.0880

Education

Secondary/above Reference

Primary/none 2.14 (0.93–4.94) 0.0747

Ethnicity

Group 1 Reference

Group 2 3.06 (0.78–12.00) 0.1088

Group 3 2.78 (0.93–8.38) 0.0684

Group 4 0.52 (0.12–2.20) 0.3772

Group 5 1.02 (0.27–3.91) 0.9735

*Significant at po0:05.Group 1: Akan.

Group 2: Gruma, Busanga, Bimoba and Kusasi.

Group 3: Dagbani, Basare/Basali, Gonja and Konkomba.

Group 4: Moshi, Hausa, Waa, Kokama, Busani, Gungui and Zugu.

Group 5: Ewe, Dagati, Sisala, Wala, Grussi, Frafra and unknown.aAflatoxin B1 levels were not determined for participants from

Kasei.

P. Jolly et al. / Int. J. Hyg. Environ.-Health 209 (2006) 345–358 355

Bivariate analysis identified several factors associatedwith AFB1 levels in this study. These include level ofeducation, ethnicity, village, number of householdmembers and number of children in secondary school.By multivariate analysis, village, and number of house-hold members were found to be strong predictors ofhigh AFB1 levels. There was also a tendency forethnicity to predict high AFB1 levels. With regard tothe level of education we found that those with asecondary or higher level of education were more likelyto sort their food before preparation. Those withsecondary education may also be able to earn moreincome than those with primary or less education andpurchase food of better quality. However, this is notknown with certainty.

Cultural differences among the different ethnic groupsin the villages that influence food preference, selection,

sorting and preparation practices, as well as generallifestyle patterns, may account for the significantassociation between village of participants and highAFB1 levels. Allen et al. (1992) found that ethnic groupand village were significant predictors of high aflatoxinlevels in the Gambia. In our study, villages with ethnicgroups whose diets/staples are predominantly cereals,dried cassava (konkonte) and groundnuts such as thosein Dromankuma and Hiawoanwu have a high potentialof having AFB1 in their systems since these foods aremost heavily contaminated by aflatoxins in Ghana. Ouranalysis showed that there were significant differences inthe ethnic make-up (po0:0001, Table 2) and frequencyof consumption of maize and groundnuts (p ¼ 0:0004)of people in the different villages. People from Nkwanta(a village with lower AFB1 levels) are predominantlyAkan and reported that they were less likely to eatmaize, groundnuts and dried cassava on a daily basiscompared to people in Dromankuma and Hiawoanwuwho have significantly higher AFB1 levels. The Akanethnic group prefers foods such as Ampesi and Fufuwhich are prepared from fresh yams, cassava andplantain that are generally free of aflatoxin. Also, peoplefrom Dromankuma belong predominantly to theDagbani, Basare/Basali, Gonja, Konkomba and theGruma, Busanga, Bimoba and Kusasi ethnic groups,who consume maize, groundnuts and cassava daily.Further, people from Dromankuma were less likely tosort their food before preparation and consumption(p ¼ 0:002) compared to those from the other villages.From our observations, proper post-harvest practicessuch as sanitary storage and sorting of food were poorlycarried out.

It is also possible that the tendency in associationbetween ethnicity and AFB1 levels may, at least in part,be explainable by differences in genetic characteristicsthat promote or inhibit detoxification of aflatoxin.Enzymes such as glutathione S-transferases (GST) andepoxide hydrolase (EPXH) detoxify aflatoxins (Johnsonet al., 1997; Guengerich et al., 1998). Genotypicdifferences in individuals may result in differences indetoxification and, hence accumulation, of aflatoxin.Certain genetic polymorphisms in the GST and EPXHenzymes may be associated with higher aflatoxin–albu-min adduct levels (McGlynn et al., 1995; Chen et al.,1996). This is currently being examined for this studypopulation.

The association between number of individuals in ahousehold and high AFB1 levels may in part beexplained by the fact that more food is needed to feedlarger households. Therefore, all available food has tobe used up by the household regardless of quality. Whenwe examined sorting of food by household size, wefound that a smaller proportion (73.2%) of householdswith 45 family members sorted their food compared tothe proportion (80.6%) of households with 1–5 family

ARTICLE IN PRESSP. Jolly et al. / Int. J. Hyg. Environ.-Health 209 (2006) 345–358356

members that sorted. However, the difference was notstatistically significant. Without adequate economicresources, larger households may grow, buy andconsume food of poorer quality. The association ofhigh AFB1 with having children in secondary schoolmay also be explained by stretched resources to providefor the needs of the children in secondary schools. Thiscould mean sale of the best food produced in order toobtain money to meet the educational needs, and henceconsumption of food of lower quality.

We understand that this is a relatively small study thatis mainly investigative in nature. However, it providessignificant new information on levels of aflatoxinbiomarkers in people in this region of Ghana andfactors associated with high aflatoxin levels that will beuseful in planning interventions designed to reduceaflatoxin consumption in food by the different ethnicgroups. These interventions are urgently needed con-sidering the high case fatality during outbreaks of acuteaflatoxicosis and the possible more subtle harmfulhealth effects of chronic dietary intake of low tomoderate levels of aflatoxin (Egal et al., 2005; Gong etal., 2003; Oswald et al., 2005; Marin et al., 2002).

Acknowledgements

We thank Jennifer Appawu for help with datacollection and Francis Obuseh for help with data entry.We are grateful to the District Health Director, thePhysician, the Pharmacist, Environmental Field Offi-cers, and other health personnel at the Ejura DistrictHospital for their assistance. Special thanks go to thestudy participants, without whose involvement thisstudy would not have been possible. We thank Dr.Thomas Kruppa, Professor Ohene Adjei and otherlaboratory personnel at the KCCR (KNUST) for use oftheir laboratory facilities and for assistance with cellseparation and shipping. This research was supportedby USAID grant LAG-G-00-96-90013-00 for the PeanutCollaborative Support Research Program and MinorityInternational Research Training grant T37 00077 fromthe Fogarty International Center, National Institutes ofHealth.

References

Allen, S.J., Wild, C.P., Wheeler, J.G., Riley, E.M., Montesa-

no, R., Bennett, S., Whittle, H.C., Hall, A.J., Greenwood,

B.M., 1992. Aflatoxin exposure, malaria and hepatitis B

infection in rural Gambian children. Trans. R. Soc. Trop.

Med. Hyg. 86, 426–430.

Ankrah, N.A., Rikimaru, T., Ekuban, F.A., 1994. Observa-

tions on aflatoxins and the liver status of Ghanaian

subjects. East Afr. Med. J. 71, 739–741.

Awuah, R.T., Kpodo, K.A., 1996. High incidence of

Aspergillus flavus and aflatoxins in stored groundnuts in

Ghana and the use of microbial assay to assess the

inhibitory effects of plant extracts on aflatoxin synthesis.

Mycopathologia 134, 109–114.

Bourgeois, C., Olson, L., Comer, D., Evans, H., Keschamras,

N., Cotton, R., Grossman, R., Smith, T., 1971. Encephalo-

pathy and fatty degeneration of the viscera: a clinico-

pathologic analysis of 40 cases. Am. J. Clin. Path. 56,

558–571.

CDC (Centers for Disease Control and Prevention), 2004.

Outbreak of aflatoxin poisoning – Eastern and Central

Provinces, Kenya, January–June 2004. MMWR 53,

790–793.

Chen, C.J., Yu, M.W., Liaw, Y.F., Wang, L.W., Chiampra-

sert, S., Matin, F., 1996. Chronic hepatitis B carriers with

null genotypes of glutathione S-transferase M1 and T1

polymorphisms who are exposed to aflatoxin are at

increased risk of hepatocellular carcinoma. Am. J. Hum.

Genet. 59, 128–134.

Cudjoe, S.N.A., 2003. Effects of changes in population,

household conditions and farming systems on agricultural

land use in the Volta river basin of Ghana, 1984–2000.

Presented at Deutscher Tropentag 2003, 8–10 October

2003, Georg-August-University, Gottingen. www.tropen-

tag.de/2003/abstracts/full/18.pdf.

Egal, S., Hounsa, A., Gong, Y.Y., Turner, P.C., Wild, C.P.,

Hall, A.J., Hell, K., Cardwell, K.F., 2005. Dietary exposure

to aflatoxin from maize and groundnut in young children

from Benin and Togo, West Africa. Int. J. Food Microbiol.

104, 215–224.

Frazier, W.C., Westhoff, D.C., 1988. Food Microbiology,

fouth ed. McGraw-Hill Book Company, New York, pp.

440–445.

Gange, S.J., Munoz, A., Wang, J.-S., Groopman, J.D., 1996.

Variability of molecular biomarker measurements from

nonlinear calibration curves. Cancer Epidemiol. Biomar-

kers Prev. 5, 57–61.

Gong, Y.Y., Egal, S., Hounsa, A., Turner, P.C., Hall, A.J.,

Cardwell, K.F., Wild, C.P., 2003. Determinants of aflatoxin

exposure in young children from Benin and Togo, West

Africa: the critical role of weaning. Int. J. Epidemiol. 32,

556–562.

Gong, Y., Hounsa, A., Egal, S., Turner, P.C., Sutcliffe, A.E.,

Ha., A.J., Cardwell, K., Wild, C.P., 2004. Postweaning

exposure to aflatoxin results in impaired child growth: a

longitudinal study in Benin, West Africa. Environmental

Health Perspectives 112, 1334–1338.

Gourama, H., Bullerman, L.B., 1995. Aspergillus flavus and

Aspergillus parasiticus, aflatoxigenic fungi of concern in

foods and feed – a review. J. Food Prot. 58, 1395–1404.

Groopman, J.D., Zhu, J., Donahue, P.R., Pikul, A., Zhang,

L.S., Chen, J.S., Wogan, G.N., 1992a. Molecular dosimetry

of urinary aflatoxin DNA adducts in people living in

Guangxi Autonomous Region People’s Republic of China.

Cancer Res. 52, 45–51.

Groopman, J.D., Hasler, J.A., Trudel, L.J., Pikul, A.,

Donahue, P.R., Wogan, G.N., 1992b. Molecular dosimetry

in rat urine of aflatoxin–N7–guanine and other aflatoxin

metabolites by multiple monoclonal antibody affinity

ARTICLE IN PRESSP. Jolly et al. / Int. J. Hyg. Environ.-Health 209 (2006) 345–358 357

chromatography and immunoaffinity/high performance

liquid chromatography. Cancer Res. 52, 267–274.

Guengerich, F.P., Johnson, W.W., Shimada, T., Ueng, Y.F.,

Yamazaki, H., Langouet, S., 1998. Activation and detox-

ication of aflatoxin B1. Mutat. Res. 402, 121–128.

Hell, K., Cardwell, K.F., Setamou, M., Poehling, H.H., 2000.

The influence of storage practices on aflatoxin contamina-

tion in maize in four agroecological zones of Benin, West

Africa. J. Stored Products Res. 36, 365–382.

Henry, S.H., Bosch, F.X., Troxell, T.C., Bolger, P.M., 1999.

Reducing liver cancer – global control of aflatoxin. Science

286, 2453–2454.

IARC, 1993. Aflatoxins. In: Some Naturally Occurring

Substances: Food Items and Constituents, Heterocyclic

Aromatic Amines and Mycotoxins. IARC Monograph on

the Evaluation of Carcinogenic Risk to Humans, vol. 56.

International Agency for Research on Cancer, Lyon,

pp. 245–395.

Johnson, W.W., Yamazaki, H., Shimada, T., Ueng, Y.F.,

Guengerich, F.P., 1997. Aflatoxin B1 8,9-epoxide hydro-

lysis in the presence of rat and human epoxide hydrolase.

Chem. Res. Toxicol. 10, 672–676.

Jolly, P.E., 1997. Replicative characteristics of primary isolates

of the human immunodeficiency virus type 1 in peripheral

blood mononuclear cells, primary macrophages and

CD4+ transformed T-cell lines. Cell. Mol. Biol. 43,

1057–1065.

Kpodo, K., Thrane, U., Hald, B., 2000. Fusaria and

fumonisins in maize from Ghana and their co-occurrence

with aflatoxins. Int. J. Food Microbiol. 61, 147–157.

Krishnamachari, K.A.V.R., Ramesh, V., Bhat, V.,

Nagarajan, V., Tilak, T.B.G., 1975. Hepatitis due to

aflatoxicosis. An outbreak in western India. Lancet i,

1061–1063.

Lye, M.S., Ghazali, A.A., Mohan, J., Alwin, N., Nair, R.C.,

1995. An outbreak of acute hepatic encephalopathy due to

severe aflatoxicosis in Malaysia. Am. J. Trop. Med. Hyg.

53, 68–72.

Marin, D.E., Taranu, I., Bunaciu, R.P., Pascale, F., Tudor,

D.S., Avram, N., Sarca, M., Cureu, I., Criste, R.D., Suta,

V., Oswald, I.P., 2002. Changes in performance, blood

parameters, humoral and cellular immune responses in

weanling piglets exposed to low doses of aflatoxin. J. Anim.

Sci. 80, 1250–1257.

McGlynn, K.A., Rosvold, E.A., Lustbader, E.D., Hu, Y.,

Clapper, M.L., Zhou, T., Wild, C.P., Xia, X.L., Baffoe-

Bonnie, A., Ofori-Adjei, D., Chen, G.C., London, W.T.,

Shen, W.T., Buetow, K.H., 1995. Susceptibility to hepato-

cellular carcinoma is associated with genetic variation in the

enzymatic detoxification of aflatoxin B1. Proc. Natl. Acad.

Sci. USA 92, 2384–2387.

Mestres, C., Bassa, S., Fagbohoun, E., Nago, M., Hell, K.,

Vernier, P., Champiat, D., Hounhouigan, J., Cardwell,

K.F., 2004. Yam chip food sub-sector: hazardous practices

and presence of aflatoxins in Benin. Journal of Stored

Products Research 40, 575–585.

Mintah, S., Hunter, R.B., 1978. The incidence of aflatoxin

found in groundnuts (Arachis hypogea L.) purchased from

markets in and around Accra, Ghana. Plant Science 5,

13–16.

Ngindu, A., Johnson, B.K., Kenya, P., Ngira, J.A., Ocheng,

D.M., Omondi, T.N., Jansen, A.J., Ngare, W., Kaviti, J.N.,

Gatei, D., Siongok, T.A., 1982. Outbreak of acute hepatitis

caused by aflatoxin poisoning in Kenya. Lancet i,

1346–1348.

Nyathi, C.B., Mutiro, C.F., Hasler, J.A., Chetsanga, C.J.,

1987. A survey of urinary aflatoxin in Zimbabwe. Int. J.

Epidemiol. 6, 516–519.

Omer, R.E., Bakker, M.I., Veer, P.V., Hoogenboom, R.L.,

Polman, T.H., Alink, G.M., Idris, M.O., Kadaru, A.M.,

Kok, F.J., 1998. Aflatoxin and liver cancer in Sudan. Nutr.

Cancer 32, 174–180.

Oswald, I.P., Marin, D.E., Bouhet, S., Pinton, P., Taranu, I.,

Accensi, F., 2005. Immunolotoxilogical risk of myco-

toxins for domestic animals. Food Add. Contam. 22,

354–360.

Park, J.W., Kim, E.K., Shon, D.H., Kim, Y.B., 2002. Natural

co-occurrence of aflatoxin B1, fumonsin B1, and ochratox-

in A in barley and maize foods from Korea. Food Addit.

Contam. 19, 1073–1080.

Peers, F.G., Gilman, G.A., Linsell, C.A., 1976. Dietary

aflatoxins and human liver cancer. A study in Swaziland.

Int. J. Cancer 17, 167–176.

Peers, F., Bosch, X., Kaldor, J., Linsell, A., Pluumen, M.,

1987. Aflatoxin exposure, hepatitis B virus infection and

liver cancer in Swaziland. Int. J. Cancer 39, 545–553.

Sarr, A.B., Mayura, K., Kubena, L.F., Harvey, R.B., Phillips,

T.D., 1995. Effects of phyllosilicate clay on the metabolic

profile of aflatoxin B1 in Fisher-344 rats. Toxicol. Lett. 75,

145–151.

Turner, P.C., Dingley, K.H., Coxhead, J., Russell, S., Garner,

C.R., 1998. Detectable levels of serum aflatoxin B1–albu-

min adducts in the United Kingdom population: implica-

tion for aflatoxin B1 exposure in the United Kingdom.

Cancer Epidemiol. Biomarker Prev. 7, 441–447.

Turner, P.C., Mendy, M., Whittle, H., Fortuin, M., Hall, A.J.,

Wild, C.P., 2000. Hepatitis B infection and aflatoxin

biomarker levels in Gambian children. Trop. Med. Int.

Health 5, 837–841.

U.S. Library of Congress, 1994. Ghana: a country study.

Climate. http://countrystudies.us/ghana/31.htm, (accessed

15.12.05).

Wang, J.S., Qian, G.S., Zarba, A., He, X., Zhu, Y.R., Zhang,

B.C., Jacobson, L., Gange, S.J., Munoz, A., Kensler, T.W.,

Groopman, J.D., 1996. Temporal patterns of aflatoxin–al-

bumin adducts in hepatitis B surface antigen-positive and

antigen-negative residents of Daxin, Qidong County,

People’s Republic of China. Cancer Epidemiol. Biomarkers

Prev. 5, 253–261.

Wang, J.S., Shen, X., He, X., Zhou, Y.-R., Zhang, B.-C.,

Wang, J.-B., Qian, G.-S., Kuang, S.-Y., Zarba, A., Egner,

P.A., Jacobson, L.P., Munoz, A., Helzlsouer, K.J., Groop-

man, J.D., Kensler, T.W., 1999. Protective alterations in

phase 1 and 2 metabolism of aflatoxin B1 by oltipraz in

residents of Qidong, People’s Republic of China. J. Natl.

Cancer Instit. 91, 347–354.

Wang, J.S., Huang, T., Su, J., Liang, F., Wei, Z., Liang, Y.,

Luo, H., Kuang, S.Y., Qian, G.S., Sun, G., He, X., Kensler,

T.W., Groopman, J.D., 2001. Hepatocellular carcinoma

and aflatoxin exposure in Zhuqing Village, Fusui County,

ARTICLE IN PRESSP. Jolly et al. / Int. J. Hyg. Environ.-Health 209 (2006) 345–358358

People’s Republic of China. Cancer Epidemiol. Biomarkers

Prev. 10, 143–146.

Wild, C.P., Rasheed, F.N., Jawla, M.F., Hall, A.J., Jansen,

L.A., Montesano, R., 1991. In-utero exposure to aflatoxin

in West Africa. Lancet 337, 1602.

Wild, C.P., Shrestha, S.M., Anwar, W.A., Montesano, R.,

1992a. Field studies of aflatoxin exposure, metabolism and

induction of genetic alterations in relation to HBV infection

and hepatocellular carcinoma in the Gambia and Thailand.

Toxicol. Lett. 64-65, 455–461.

Wild, C.P., Hudson, G.J., Sabbioni, G., Chapot, B., Hall, A.J.,

Wogan, G.N., Whittle, H., Montesano, R., Groopman,

J.D., 1992b. Dietary intake of aflatoxins and the level of

albumin-bound aflatoxin in peripheral blood in the

Gambia, West Africa. Cancer Epidemiol. Biomarkers Prev.

1, 229–234.

Wild, C.P., Yin, F., Turner, P.C., Chemin, I., Chapot, B.,

Mendy, M., Whittle, H., Kirk, G.D., Hall, A.J., 2000.

Environmental and genetic determinants of aflatoxin–albu-

min adducts in the Gambia. Int. J. Cancer 86, 1–7.