PRESENCE OF ORGANOPHOSPHATE INSECTICIDES IN ...

325Presence of organophosphate...

Nicolas SOUMIS1* , Marc LUCOTTE1, Delaine SAMPAIO2, Diane CRUZALMEIDA2, Dalie GIROUX3, Silmara MORAIS2, Pierre PICHET4

PRESENCE OF ORGANOPHOSPHATE INSECTICIDES IN FISHOF THE AMAZON RIVER

1Author to whom correspondence should be addressed. Tel. : (514) 987-4080 Fax: (514) 987-3635 E-mail: [email protected] of Environmental Sciences/GEOTOP, University of Québec in Montréal (UQAM), PO Box 8888, succ.

Centre-Ville, Montréal, Québec, H3C 3P8, Canada.3Laboratory of Environmental Biology, Federal University of Pará, Campus of Santarém, Santarém, PO Box

241, CEP 68040-060, Pará, Brazil.4Department of Political Science, UQAM, PO Box 8888, succ. Centre-Ville, Montréal, Québec, H3C 3P8, Canada.5Department of Chemistry, UQAM, PO Box 8888, succ. Centre-Ville, Montréal, Québec, H3C 3P8, Canada.

ABSTRACT – Trace levels of three organophosphate insecticides (OPI) were detected in eightfish species from the region of Santarém, State of Pará, Brazil. Individual concentrations of OPIin fish ranged from less than the detection limit to 2,1 ppb. Mean concentrations of chlorpyrifos,malathion, and methyl-parathion were 0,3 ± 0,3, 0,1 ± 0,1, and 0,3 ± 0,3 ppb, respectively.Pellona flavipinnis, the largest and fattest piscivorous species analyzed, was the most contami-nated. Since an inhabitant of this Amazonian region consumes 220 g of fish per day on average,ingested doses of chlorpyrifos, malathion, and methyl-parathion may reach up to 308, 220, and462 ng·d-1, respectively. Compared to acceptable daily intakes (ADI), quantities of OPI absor-bed via fish consumption on a daily basis are far below deleterious levels. We estimated thateven considering the highest OPI contents detected, the average daily fish consumption of anadult of 60 kg would have to increase by ca. 1 950, 5 450, and 2 600 times to reach ADI ofchlorpyrifos, malathion, and methyl-parathion, respectively. Neither fish diet nor fish lipid contentenabled us to completely explain the interspecific differences observed.Key-Words: Brazilian Amazon, fish contamination, organophosphate insecticides, acceptable dailyintake (ADI), human alimentation.

Presença de inseticidas organofosforados nos peixes do RioAmazonas

RESUMO – As concentrações de três inseticidas organofosforados (IOF) foram determinadasem oito espécies de peixes da região de Santarém, Estado do Pará, Brasil. As concentraçõesindividuais de IOF dos peixes situam-se entre abaixo da limite de detecção e 2,1 ppb. Asconcentrações médias de clorpirifós, malathion e metil parathion foram respectivamente de 0,3 ±0,3, 0,1 ± 0,1 e 0,3 ± 0,3 ppb. Pellom Flavipinnis, a espécie piscívora de maior tamanho e maisgordurosa, apresentou as concentrações mais elevadas. Considerando que um morador dessa regiãoda Amazônia brasileira consome em média 220 g de peixe por dia, as doses de clorpirifós, malathione metil parathion podem atingir respectivamente 308, 220 e 462 ng·d-1. Tendo em conta as "dosesdiárias aceitáveis" (acceptable daily intakes – ADI), as quantidades diárias de IOF absorvidas viaconsumo de peixe estão muito abaixo dos níveis perigosos para a saúde humana. Nós estimamosque mesmo com o teor mais elevado de IOF detectado, o consumo diário médio de peixe de umadulto de 60 kg teria que aumentar cerca de 1 950, 5 450 e 2 600 vezes para atingir as respectivasADI de clorpirifós, malathion e metil parathion. Nem o hábito alimentar dos peixes e nem o seuteor de gorduras podem explicar completamente as diferenças interespecíficas observadas.Palavras-Chave: Baixo Amazonas, contaminação em peixes, inseticidas organofosforados, dosesdiárias aceitáveis (ADI), alimentação humana.

ACTA AMAZONICA 33(2): 325-338

326 Soumis et al.

INTRODUCTION

We previously performed a study ofpesticide consumption in the region of Santa-rém, an important urban center of the LowerAmazon, Brazil (Soumis et al., 2000). Basedon surveys among several local crop produc-ers and agricultural suppliers, this study esta-blished that besides other pesticides such asfungicides and herbicides, organophosphate in-secticides (OPI) were the most common com-pounds into the pesticide arsenal of this region,their consumption reaching ca. 2 000 kg·year-1

(active ingredients). The importance of theseinsecticides is also reflected by pesticide con-sumption figures for Brazil gathered by the Fo-od and Agriculture Organization (FAO) thatindicate an upward trend in the use of OPI bet-ween 1991 (1 850 tons) and 1995 (5 384tons)1 . These insecticides are currently usedon crops established on floodplains (várzeas)of the Amazon River. Várzea designate theportion of land along rivers which is floodedduring the rainy season. Annual floods carrylarge amounts of nutrient-rich Andean se-diments which deposite on várzeas, a fertili-zing phenomenon that ensures an enhanced agri-cultural potential for these environments(Furch, 1997; Ohly and Junk, 1999; Zarin,1999).

Apart from supporting agriculture,várzeas also represent important fishing sitesin view of their abundant and diverseichtyofauna (Ruffino, 1996; Cerdeira et al.,1997; Ruffino et al., 1998). Now the multipleuse of this environment may be problematicsince the contamination of fish by OPI is likelyto occur for several reasons. First, várzea cropsare located in the vicinity of aquatic ecosys-tems. Hence, wind drift, volatilization, andfrequent heavy rainfalls in the Amazon Basin(IBGE, 1977, in Salati, 1985; Ohly, 1987)may cause OPI to reach streams and lakesnearby crops (Racke, 1992). Second, OPI re-

main in their original form long enough(from several days to a few months) to allowthem to move away from their application si-te. Third, many OPI are hydrophobic com-pounds; in aqueous media, they have muchmore affinity for the biological compartments(Sabharwal and Belsare, 1986; Cowgill et al.,1991).

Basic information on the pesticide bur-dens in fish remains of first concern in theLower Amazon as fisheries represent one ofthe principal economic activities in this region,and as fish is the main source of animal pro-teins in the diet of local populations (Saint-Pauland Bayley, 1979; Ferreira et al., 1996; Isaacet al., 1996). According to our knowledge, veryfew studies has yet been conducted in order toassess the concentration of agricultural pesti-cide in the Amazonian ichtyofauna (Torres etal., 2002). Considering a situation which mayjeopardize the health of the local populationsand alter an important pillar of the regionaleconomy, the present study constitutes a pre-liminary assessment of the ichtyofauna con-tamination by three OPI commonly used onvárzea crops : chlorpyrifos (O, O-diethyl O-(3, 5, 6-trichloro-2-pyridyl) phosphorothioate;CAS : 2921-88-2; Kow : 158 489), malathion(O, O-dimethyl S-(1, 2-dicarbethoxyethyl)phosphorodithioate; CAS : 121-75-5; Kow:776), and methyl-parathion (O, O-dimethyl O-(4-nitrophenyl) phosphorothioate; CAS : 298-00-0; Kow : 1 288). The following objectiveswere pursued : 1) to obtain a portrait of thefish contamination by OPI in the region ofSantarém; 2) to compare the results obtainedwith acceptable daily intakes2 (ADI) in orderto determine if the daily doses absorbed bylocal populations reach a theoreticallycritical level; 3) to verify the existence of si-gnificant interspecific differences in contami-nation levels and to assess some biological fac-tors that may be responsible for suchdifferences.

1According to FAOSTAT, a FAO’s database available on the Web (http://apps.fao.org/page/collection). Data from FAO areonly available from 1991 to 1995.

2Acceptable daily intake was defined by the World Health Organization (WHO) as the daily intake of a chemical which,during an entire lifetime, appears to be without appreciable risk on the basis of all known facts at the time (WHO, 1962,in Klaassen and Eaton, 1991). ADI is expressed in mg of chemical by kg of the consumer body.


STUDY AREA

Located at the confluent of the Tapajósand the Amazon Rivers in the State of Pará,Santarém (2° 25’ S, 54° 43’ W) is the princi-pal urban center of the Lower Amazon.This region was selected for many reasons.First, an important concentration of agriculturalactivities take place around Santarém(Scatena et al., 1996). Second, this region alsoencompasses one of the greatest concentrationsof várzeas of the Brazilian Amazon.Third, as the largest fishing port of the LowerAmazon, Santarém receives the catches from14 municipalities along the Amazon River(Isaac et al., 1996; Ruffino et al., 1998). It isthus possible to collect at Santarém’s fishmarkets specimens coming from the wholeregion.

MATERIAL AND METHODS

The sampling campaign was conductedby the end of September of 1998, during thelow water season. At that time of year, as ri-ver recedes to its lowest level (Fig. 1), thewhole várzeas are free of water and agriculturalactivities attain its optimum. Consequently,the highest pesticide consumption is expectedto be found during this season. One undredand twenty fresh specimens, equally distributed between eight species (Tab. 1), were bou-ght from a fish market located in the port ofSantarém during a five-day sampling.According to salemen and fishermen, specimenscollected were known to belong from theregion.

Figure 1 - Fluctuations of water level for 1998. Measurements were made twice a day at thePortobrás’ wharf in Santarém, located at the confluence of the Amazon and Tapajós Rivers(Data from Delegacia da Capitania dos Portos dos Estados do Pará e Amapá em Santarém).

emancifitneicS )nailizarB(emannommoC teiD tnetnocdipiL

suluconomalhciC )xipS( éranucut suorovicsip )wol(%5,0mumoporcamamossoloC )reivuC( iuqabmat suorovibreh )wol(%2,0

(siladrapsucrasopiL )uanletsaC ódobiraca suorovitirted )wol(%1,0ertnevirudamossolyM )reivuC( mumocucap suorovibreh )hgih(%0,7

sinnipivalfanolleP )senneicnelaV( ocnarbápapa suorovicsip )hgih(%4,6sumissisomauqsnoicsoigalP )lekceH( acnarbadacsep suorovicsip )etaidemretni(%0,2

irerettansurtnecogyP )renK( ujacahnarip suorovicsip )etaidemretni(%2,1mutaicsafnodozihcS )zissagA( mumocucara suorovibreh )etaidemretni(%0,2

Table 1 - Names, diet and lipid content of fish species sampled from the market of Santarém.

J F M A M J J A S O N D

m onth

water level (m

)

sam p ling period

0

1

2

3

4

5

6

7

8


m onth

water level (m

)

sam p ling period

0

1

2

3

4

5

6

7

8


m onth

water level (m

)

sam p ling period


m onth

water level (m

)

sam p ling period

0

1

2

3

4

5

6

7

8


m onth

water level (m

)

sam p ling period

0

1

2

3

4

5

6

7

8


m onth

water level (m

)

sam p ling period

0

1

2

3

4

5

6

7

8


m onth

water level (m

)

sam p ling period


m onth

water level (m

)

sam p ling period

0

1

2

3

4

5

6

7

8


m onth

water level (m

)

sam p ling period

0

1

2

3

4

5

6

7

8


m onth

water level (m

)

sam p ling period


m onth

water level (m

)

sam p ling period

0

1

2

3

4

5

6

7

8

328 Soumis et al.

In the laboratory, filets free of scales,skin, and bones were drawn from fish so asto be frozen (-20 °C) in amber-glass bottleswith teflon-septum caps. Extraction of OPIresidues from the filets was carried out accord-ing to methods 304-E1 (extraction) and 304-C2 (cleanup) of the American Food and DrugAdministration (FDA) suitable for fatty matri-ces3 . This solid/liquid multiresidue extractionprocedure allows for the simultaneousextraction of several nonpolar pesticides. Asa result, only parental OPI were quantitativelyrecovered since it is impossible to extractpolar metabolites of OPI with thisprocedure.

We slightly modified the FDA methodsto better fit our needs. Hereafter are the fewmodifications we brought. For the extractionstep, 30 g (wet weight) of fish filet were grou-nded with 60 g of anhydrous sodium sulfate(Na2SO4). We determined the quantity of fishaccording to fat content (Junk, 1985) to avoidoverwhelming the retention capacity of the cle-anup column. For the concentration step, weused a rotary evaporation system instead ofthe Kuderna-Danish concentrator recommendedby the FDA since we achieved better resultswith the former device. For the cleanup step,we followed instructions for the methylenechloride elution system, although we only usedeluants "2" and "3" (50% methylene chloride,0,35% acetonitrile, 49,65% hexane (v/v/v) and50% methylene chloride, 1,5% acetonitrile,48,5% hexane (v/v/v), respectively). As wecollected all of our analytes through these twoeluants, the eluant "1" proposed by FDA wasuseless in our case. All reagents were ACSgrade; they were concentrated and tested forcontamination beforehand.

Cleaned extracts were concentrated to 2ml prior to analysis. OPI analyses were con-ducted with a Varian 3800 gas chromatograph(GC). Our system was equipped with an au-tosampler (Varian 8200), a splitless glass in-sert (0.5 mm ID), a SPB-5 fused silica capilla-ry column (Supelco, 30 m lenght, 0,25 mm IDand 0,25 µm film) protected by a deactivatedand uncoated fused silica guard column (Siltek,

5 m length, 0,25 mm ID), and provided with athermoionic specific detector (TSD). TSD re-sponse was monitored during routine analysesby external standards. Injector and TSD tem-peratures were set at 230 and 300 °C, res-pectively. The temperature programming of theGC oven was as follow : initial temperature,80 °C, initial hold time, 120 s; ramp rate of 40°C·min.-1 to 150 °C, hold time, 60 s; ramp rateof 4 °C·min.-1 to 200 °C, hold time, 300 s; andramp rate of 40 °C·min.-1 to 290 °C, final holdtime, 360 s; total run time, 30,5 min. Heliumwas the carrier gas (1,5 ml·min.-1) while nitrogenwas the make-up gas (28,5 ml·min.-1). Eachanalysis was made in triplicate, with an in-jection volume of 2 µl.

Recovery rate (± standard deviation) foreach OPI was 110% ± 10%. Detection limitsfor chlorpyrifos, malathion, and methyl-para-thion were 2,45, 1,65, and 3,20 ppm (con-centrations of the extract injected in the GC),respectively. Concentrations of insecticides infish (wet weight) were calculated separately inorder to compare each one to its own ADI.Undetected compounds were assumedto be present at half the detection limit (seeabove) of the GC rather than absent fromspecimens.

In order to further our understanding ofthe ichtyofauna contamination, we tested so-me biological factors that may influence OPIlevels in specimens. We first checked for theexistence of interspecific differences. Then, weexplored two factors (diet and lipid content)which might explain these interspecific differ-ences. Determination of diet (Tab. 1) was madeaccording to literature (Goulding and Carvalho,1982; Soares et al., 1986; Ruffino and Isaac,1995; Junk et al., 1997; Ferreira et al., 1998).Lipid contents were estimated according toJunk (1985), by using figures for the low wa-ter season (Tab. 1). For all parameter tested,we used Kruskal-Wallis’ tests and a set ofWilcoxon-Mann-Whitney’s tests every time aKruskal-Wallis test showed a significant differ-ence. In addition, Kendall’s rank correlationtests were used to verify the existence of arelation between the lipid content and the OPIlevel in filets.

3These methods are found in the Pesticide Analytical Manual, volume 1. This document is available on the Web at thefollowing URL : http://vm.cfsan.fda.gov/~frf/pami3.html (see chapter 3).


RESULTS

Description of the fish sample

Our results indicate that OPI foundin the specimens analyzed are at trace levels.As depicted by the histograms ofFigure 2, which illustrate distributions of the

120 specimens according to their OPIconcentrations, most specimens are veryslightly contaminated albeit a restricted fewshow much more higher concentrations. More-over, the exponential pattern of distribu-tions clearly justifies the use of nonpara-metric tests for further statistical dataanalysis.

Figure 2 - Statistical distributions of the 120 specimens according to their concentrations inchlorpyrifos, malathion, and methyl-parathion. sk = skewness, sx

2 = variance, θ = 1/µ (pa-rameter of the exponential law, where µ is the mean).

Table 2 lists some statistical parametersfor the whole sample and for each species.Considering the whole sample, the majorityof specimens are contaminated to a detectablelevel by chlorpyrifos and methyl-parathion(59,2% and 68,3%, respectively), although thisis not true in the case of malathion (only 25,8%detected). Furthermore, 20% of the sampleshow no detectable contamination by any ofthe three OPI. When specimens are dividedaccording to their species (Tab. 2), it appearsthat the occurrence and the difference betweenthe particular mean and median vary fromone species to another. These observationssuggest specific trends, which becomemore obvious (2 words) from the boxplots ofFigure 3.

Daily doses absorbed by popula-tions and comparison toacceptable daily intakes

According to Cerdeira et al. (1997), theaverage fish meat consumption of a riverine po-pulation living near Santarém is about 220g·d-1 inhabitant (a point is missing). Althoughit is likely that urban populations consume asmaller quantity of fish per day (Shrimptonand Giugliano, 1979; Smith, 1979; Amoroso,1981), we extrapolated this figure to the wholeregion of Santarém. Table 3 estimates OPIdaily doses ingested by local populations viatheir fish consumption.

Acceptable daily intakes (ADI) forchlorpyrifos, malathion, and methyl-parathion

330 Soumis et al.

are fixed at 0,01, 0,02, and 0,02 mg·kg-1, respec-tively (Lu, 1995). Assuming a hypotheticalbody weight of an adult consumer of 60 kg4 ,we calculated that critical daily doses for chlo-rpyrifos, malathion, and methyl-parathion are0,6, 1,2, and 1,2 mg·d-1, respectively. Consi-dering the latter figures, it is obvious that eventhe maximum daily doses to which local po-pulations are exposed (Tab. 3) are far smallerthan those required to reach ADI. In fact, dailyfish consumption would have to increase byca. 1 950, 5 450, and 2 600 times to reach ADIof chlorpyrifos, malathion, and methyl-para-thion, respectively. In other words, with such"security margins", probabilities of attainingthe critical doses stated above with fish fromthe sampled market are infinitesimal.

Biological factors influencing Thecontamination (2 words)

level of fish by OPI

According to Kruskal-Wallis’ tests, in-terspecific differences that we already suspec-ted from boxplots of Figure 3 proved to bevery highly significant for all three OPI. Fur-ther analysis using Wilcoxon-Mann-Whitney’stests indicated that these differences are causedby 1) P. flavipinnis (higher than all other spe-cies) and P. squamosissimus (higher than L.pardalis, P. nattereri, and S. fasciatum, butlower than P. flavipinnis) in the case of chlor-pyrifos; 2) P. flavipinnis (higher than all otherspecies, except C. macropomum and S. fascia-tum) and L. pardalis (lower than all other spe-

4Based on a study made in a small community along the Tapajós River (Brasilía Legal), the average Amazonian adult femaleweight is 58 kg and the average adult male weight is 63 kg (D. Mergler, personal communication).

Table 2 - Statistical parameters for the whole sample and for each species.puorG eziS edicitcesnI egnaR naideM naeM .veddtS ecnerruccO a

n )bpp( )bpp( )bpp( )bpp±( )%(sofiryprolhc 4,1-1,0 1,0 3,0 3,0 2,95

elpmaselohw 021 noihtalam 0,1-0,0 1,0 1,0 1,0 8,52noihtarap-lyhtem 1,2-1,0 2,0 3,0 3,0 3,86

sofiryprolhc 0,1-1,0 1,0 2,0 3,0 0,06suluconom.C 51 noihtalam 2,0-0,0 1,0 1,0 1,0 3,31

noihtarap-lyhtem 4,0-1,0 2,0 2,0 1,0 3,35sofiryprolhc 1,1-1,0 2,0 3,0 3,0 7,66

mumoporcam.C 51 noihtalam 7,0-0,0 1,0 1,0 2,0 0,04noihtarap-lyhtem 5,0-1,0 3,0 3,0 1,0 7,68

sofiryprolhc 9,0-1,0 1,0 2,0 2,0 7,64siladrap.L 51 noihtalam 4,0-LD< LD< LD< 1,0 3,31


ertnevirud.M 51 noihtalam 2,0-LD< 1,0 1,0 1,0 7,6noihtarap-lyhtem 7,0-1,0 3,0 3,0 2,0 0,06

sofiryprolhc 4,1-2,0 4,0 5,0 3,0 0,08sinnipivalf.P 51 noihtalam 0,1-1,0 1,0 2,0 2,0 0,04


sumissisomauqs.P 51 noihtalam 2,0-LD< 1,0 1,0 1,0 7,33noihtarap-lyhtem 4,0-1,0 2,0 2,0 1,0 3,35

sofiryprolhc 3,0-1,0 1,0 1,0 1,0 7,64irerettan.P 51 noihtalam 3,0-LD< 1,0 1,0 1,0 7,6


mutaicsaf.S 51 noihtalam 3,0-LD< 1,0 1,0 1,0 3,35noihtarap-lyhtem 4,0-1,0 2,0 2,0 1,0 3,35

a: Determines the ratio of specimens in which the compound has been positively detected.< DL : Smaller than the detection limit


cies, except C. monoculus) in the case of mala-thion; 3) P. flavipinnis (higher than all otherspecies) and C. macropomum (higher than C.monoculus, P. nattereri, and S. fasciatum, butlower than P. flavipinnis) in the case of meth-yl-parathion. Although these observationsshow that some species may be marked outfrom the batch sample, none of them is as re-peatedly distinct as P. flavipinnis. The lattermay hence be seen as greatly responsible forthe interspecific differences registered.

Influence of both biological factorstested (diet and lipid content) in order to ex-plain interspecific differences varies accordingto the insecticide at stake. According toKruskal-Wallis’ tests, diet only seems to exertan influence in the case of malathion, differ-ences being very highly significant. Results ofWilcoxon-Mann-Whitney’s tests clearly pointout that the detritivorous diet distinguishes it-self (lower) from the two other dietary groups,the latter remaining statistically undiscernible.

Regarding the lipid content, differencesbetween groups are at least highly significant(Kruskal-Wallis) for all three OPI. Results ofWilcoxon-Mann-Whitney’s tests indicate thatthese differences are caused by the high lipidcontent group, which distinguishes itself(higher) from other groups, chlorpyrifos andmethyl-parathion respectively showing as hi-ghly and very highly significant. The situa-tion with malathion is less obvious since thehigh lipid content group is different (higher)from the low lipid content group, without therebeing any significant difference between thesegroups and the intermediate one. Furthermore,Kendall’s rank correlation tests show that ifthere is a relation of an at least significant na-ture between the lipid content and the contami-nation level, this relation remains weak accor-ding to Kendall’s tau-b coefficients (t).

DISCUSSION

Choice of the study period andrepresentativeness of the sample

Being tightly linked to the seasonalwater cycle of the Amazon River, agricultureon várzeas is subjected to wide fluctuations inthe course of a year (Fig. 1). During the high

Figure 3 - Distribution boxplots for each spe-cies analyzed. O = outlier value (> 1,5 inter-quartile range); * = extreme value (> 3interquartile range).

malathion

(ppb

)

S. fa

sciatum

P. nattere

ri

P. squ

amos

issim

us

P. flav

ipinnis

M. d

uriven

tre

L. par

dalis

C. m

acropo

mum

C. m

onoc

ulus

methy

l-p

arat

hion

(ppb

)

2 ,0

1 ,6

1 ,2

0 ,8

0 ,4

0 ,0

1 ,0

0 ,8

0 ,6

0 ,4

0 ,2

0 ,0

chlorp

yrifo

s (pp

b)

1 ,4

1 ,2

1 ,0

0 ,8

0 ,6

0 ,4

0 ,2

0 ,0

malathion

(ppb

)

S. fa

sciatum

P. nattere

ri

P. squ

amos

issim

us

P. flav

ipinnis

M. d

uriven

tre

L. par

dalis

C. m

acropo

mum

C. m

onoc

ulus

methy

l-p

arat

hion

(ppb

)

2 ,0

1 ,6

1 ,2

0 ,8

0 ,4

0 ,0

1 ,0

0 ,8

0 ,6

0 ,4

0 ,2

0 ,0

chlorp

yrifo

s (pp

b)

1 ,4

1 ,2

1 ,0

0 ,8

0 ,6

0 ,4

0 ,2

0 ,0

met

hyl-

para

thio

n (p

pb)

mal

athi

on (

ppb)

chlo

rpyr

ifos

(pp

b)

332 Soumis et al.

tilization during that time, insecticides runoffis also likely to occur. Third, as fish popula-tions are more confined due to the drying-upof some aquatic habitats, fishing activities in-crease during the low water season (Ohly andJunk, 1999). Probabilities are thus higher toconsume a contaminated specimen during thatperiod of the year.

The eight species collected cannot givean exhaustive picture of the Lower Amazon’sichtyofauna biodiversity. Indeed, more thansixty fish species are found in the several mar-kets of Santarém (Isaac et al., 1996; Ruffino etal., 1998). Nevertheless, only ten species outof those sixty account for 80-86% of thecatches landed in Santarém (Isaac et al., 1996;Ruffino et al., 1998). Since the eight speciescollected during this study are among the lat-ter, we can reasonably assume that our sampleis representative of the commercial speciescurrently found in Santarém during the lowwater season.

Sanitary hazards linked to fishconsumption

As far as OPI are concerned, the dailyfish consumption does not appear to representa health hazard for the Amazonian populationssince detected levels of these compounds arefar below their ADI. Several considerations

water season, most várzea soils are flooded andagricultural activities are scarce, being oftenconfined to small gardens on stilts (Padoch andPinedo-Vasquez, 1999). However, as heavyrainfalls are frequent during this period of theyear, runoff with the few pesticides applied isprone to occur. On the contrary, during thelow water season, agriculture and pesticideapplications reach their optimum as plenty ofsoil is available for crops. However, becauseprecipitations are less abundant during thisseason, pesticide runoff may be lower in com-parison with the high water season.

In a comprehensive study of pesticidecontamination in fish, it would have been re-levant to conduct several sampling campaignsalong the year in order to assess each set ofenvironmental and agricultural conditions.However, keeping in mind that we conducteda preliminary study on this issue, we ratherfocused on a single period we considered themost critical. For several reasons, we estimatedthat the low water season could be that criticalperiod during which fish contamination may beat its highest. First, and as explained above, itis expected to find the highest pesticide con-sumption at that time of the year. Second, theaveraged daily precipitations in September areabout 3 mm (Xie and Arkin, 1996, in Zeng,1999)5 . Hence, although fish contamination ismore likely to come from wind drift and vola-

Table 3 - Estimation of daily doses of OPI ingested by populations of the region of Santarémwhen consuming fish from the marketa.

a: Based on a daily fish consumption of 220 g (Cerdeira et al., 1997).< DL : Smaller than the detection limit

sofiryprolhc 0,22d·gn(sesodlaminiM 1- ) noihtalam LD<

)nemicepsdetanimatnocsselehtnodesab( noihtarap-lyhtem 0,22sofiryprolhc 0,22

d·gn(sesodnaideM 1- ) noihtalam 0,22)elpmaselohwehtfonaidemehtnodesab( noihtarap-lyhtem 0,44

sofiryprolhc 0,66d·gn(sesodegarevA 1- ) noihtalam 0,22

)elpmaselohwehtfonaemehtnodesab( noihtarap-lyhtem 0,66sofiryprolhc 0,803

d·gn(sesodlamixaM 1- ) noihtalam 0,022)nemicepsdetanimatnoctsomehtnodesab( noihtarap-lyhtem 0,264

5Averaged data for September, from 1985 to 1993.


Determination of biologicalfactors affecting the contamina-

tion levels in fish

Although we found that the contamina-tion level varies from a species to another, theexact causes of this situation still remain un-known. On the one hand, statistical tests haveshown that no difference was induced by thediet in the case of chlorpyrifos and methyl-par-athion. However, the same biological factorseems to induce a very highly significant dif-ference in the case of malathion. While thisdifference is due to the detritivorous grouprepresented by L. pardalis alone, it is yet du-bious to draw any conclusion from a singlespecies as differences could arise from otherspecific factors. On the other hand, statisti-cally significant differences caused by the lipidcontent also remain ambiguous. By correlat-ing the lipid contents of fish with their con-tamination levels, we expected to find a con-comitant augmentation of both variables, aphenomenon which was assessed by Kendall’srank correlation tests. However, these testshave shown but a weak relation between thosevariables. Moreover, only the high lipid con-tent group distinguished itself from the othergroups. It appears that the differences aremainly due to P. flavipinnis since M. duriventre– the other species comprised in the high-lipidgroup – presents relatively low OPI levels inspite of its higher lipid content. Thus, althoughthe lipid content may be partially responsiblefor the interspecific differences, other influentfactors which were not assessed in this studymay intervene to a larger extent. The specificmetabolism of fish could represent a relevantanswer to the differences observed, but thereis yet no study dealing with toxicokinetic as-pects of Amazonian fish. One may also notethat the spatial distribution of specimens – aparameter that was not assessed in this study– can account for the differences observed.

The absence of any significant differencecaused by the diet in the case of chlorpyrifosand methyl-parathion may suggest some indi-cations about the phase and the way these OPIare introduced in the aquatic environment, andas to how they contaminate the ichtyofauna.

may however temper this reassuring conclusi-on. Only three OPI among the most com-monly used pesticides in Santarém were ana-lyzed, thus neglecting other potentially harm-ful products. Moreover, only parental com-pounds were quantitatively recovered by theFDA procedures, letting metabolite concentra-tions of these insecticides unknown. This situ-ation constitutes an important lack of thisstudy when considering the high vulnerabilityof OPI to in vivo transformations (IPCS, 1986;Gallo and Lawryk, 1991) and the great toxic-ity of some of these metabolites. Although itis a widely accepted assertion that metabolitesare rapidly excreted by living organisms be-cause of their low affinity for lipid matrices, itwas demonstrated that some fish species maybioaccumulate malaoxon (a metabolite ofmalathion) because they lack the enzymes re-quired for its hydrolysis (Areechon and Plum,1990). However, it is unlikely that analyzedfish contain a sufficiently high concentrationof metabolites to reach critical doses. Since thequantity of OPI found in specimens is manytimes lower than their respective ADI, the fishwould have to accumulate much greater quan-tities of metabolites than those detected fortheir parental compounds, a situation notlikely to occur (Hall and Kolbe, 1980; Takaseand Oyama, 1985; Cunha Bastos et al.,1999).

A last limit still deserves attention. Eachinsecticide was individually compared to itsparticular ADI. Now, considering that mostspecimens are contaminated with more thanone compound and given the existence of syn-ergistic or additive effects between some OPI(Gallo and Lawryk, 1991), we must be carefulwhen drawing conclusions from norms estab-lished for a single insecticide. However, onemay argue that thanks to the wide securitymargins we defined, it is unlikely that additiveor synergistic effects of OPI present in fishreach a deleterious level. Nevertheless, theconcomitant absorption of pesticides fromother alimentary sources such as fruits or veg-etables may remain dangerous. In a similarway, the possible presence of other neurotoxicchemicals in fish – particularly organomercuriccompounds (Lebel et al., 1997) – may causeunsuspected synergistic effects which are notcovered by this study.

334 Soumis et al.

Despite the interest of these prospective leads,we remain careful regarding the validity of thefollowing hypotheses mainly because ofthe low contamination levels and the relativelyhigh number of specimens with undetectedOPI.

Two phenomena may explain the simi-lar contamination levels of the different dieta-ry groups : 1) an even distribution of contami-nants among all food sources for fish, whichwould allows equal body burdens viabiomagnification; 2) the presence of OPI in thewater column which may be readily absorbedby bioconcentration. Both phenomena implythe presence of dissolved chlorpyrifosand methyl-parathion, forms unlikely tobe introduced in water by runoff or leachingfrom treated soils since these two insecticidesare known to be well adsorbed onto soil par-ticles (Wauchope, 1978; Agnihotri et al., 1981).Now there are only two possible ways to finddissolved OPI in the water column. One me-ans is through volatilization, atmospheric trans-port, and precipitations (Racke, 1992;Bidleman, 1999; Van Dijk and Guicherit, 1999).These routes play a major role in OPI trans-portation, especially during a period of lowrainfall. We can also suspect some bioconcen-tration arising from a slow desorption processof OPI from the sediments, which constitutean important sink for these contaminantsassocied with soil particles (Hughes et al.,1980; Sabharwal and Belsare, 1986; Bhushanet al., 1997).

Between both hypotheses onbioaccumulation stated above, bioconcentrationseems to be the principal route through whichOPI are absorbed by fish. Indeed, OPI are nottransfered along the food chain to a great ex-tent because they are easily decayed by the invivo digestive processes of animals (IPCS,1986), plant metabolism (Sabharwal andBelsare, 1986; Racke, 1992), and even chemi-cal/microbial processes in sediments and floo-ded soils (Sabharwal and Belsare, 1986;Adhya et al., 1987), thus tempering biomagni-fication processes. Moreover, there is someevidence that bioconcentration is the most rele-vant route of contaminant exposure in theaquatic habitat (Chiou et al., 1977; Bruggemanet al., 1981).

CONCLUSION

Focusing on the agricultural pollution ofthe ichtyofauna by OPI, this study has shownthat detected concentrations of some of theseinsecticides in fish are several times lower thantheir respective ADI. It thus appears that thecontamination of the ichtyofauna by OPI isunlikely to occur to a problematical extent inthe region of Santarém. Pellona flavipinnis, alarge and high lipid-content piscivorous spe-cies, was the most contaminated fish of oursample. Still, the exploratory and partial na-ture of this study should not be overlooked.Concerning the assessment of the contamina-tion by other classes of pesticides or by OPImetabolites, we allowed ourselves to specula-te on the wide security margins we determined.Nevertheless, we insist on the fact that eitherthis statement or the one on the plausible ab-sence of deleterious synergistic or additive ef-fects must imperatively be validated bycomplementary studies. Besides, our conclu-sions are only applicable to the present OPIamounts in use and the consequences of agreater pesticide consumption on várzeas re-main unknown.

To further address the issues raised bythis study we strongly encourage local scien-tific authorities to develop a monitoring pro-gram aiming at the assessment of contamina-tion levels of a wide array of pesticides in theichtyofauna. Particularly, it would be of firstinterest to characterize the seasonal fluctuationsof the contamination, considering the huge hy-drological modifications taking place through-out the year and their impact on várzea’s agri-culture. It would also be interesting to deter-mine the actual effect of floods on the fate ofpesticide residues in the soils of várzeas : afurther degradation (hydrolysis) or a desorp-tion toward the water column.

ACKNOWLEDGMENTS

This study was supported by the Natu-ral Sciences and Engineering Research Councilof Canada (NSERC) and the International De-velopment Research Center of Canada (IDRC).The authors would like to thank Dr. MauroLuis Ruffino from Projeto IARA at Santarém


for his helpful advices, the students of theLaboratório de Biologia Ambiental at the Fed-eral University of Pará (UFPa) in Santarém fortheir implication on the field, and Pierre Cayerand Sophie Tran from the Institute of Envi-ronmental Sciences at the University of Québecat Montréal (ISE-UQAM) for their help in thelaboratory. Revision of the manuscript wasmade by Emmanuel Roy.

LITERATURE CITED

Adhya, T.K.; Wahid, P.A.; Sethunathan, N.1987. Persistence and biodegradation ofselected organophosphorus insecticides inflooded versus non-flooded soils. Biologyand Fertility of Soils, 4(1):36-40.

Agnihotri, N.P.; Pandey, S.Y.; Jain, H.K.;Srivastava, K.P. 1981. Persistence, leach-ing and movement of chlorfenvinphos,chlorpyrifos, disulfoton, fensulfothion,monocrotophos and tetrachlorvinphos insoil. Indian Journal of Agricultural Chem-istry, 14(1-2):27-31.

Amoroso, M.C.M. 1981. Alimentação em umbairro pobre de Manaus. Acta Amazônica,11(3):supplement of 43 pp.

Areechon, N.; Plum, J.A. 1990. Sublethal ef-fects of malathion on channel catfish,Ictalurus punctatus. Bulletin of Environ-mental Contamination and Toxicology,44:435-442.

Bhushan, R.; Thapar, S.; Mathur, R.P. 1997.Accumulation pattern of pesticides intropical fresh waters. Biomedical Chroma-tography, 11(3):143-150.

Bidleman, T.F. 1999. Atmospheric transportand air-surface exchange of pesticides.Water, Air, and Soil Pollution, 115(1-4):115-166.

Bruggeman, W.A.; Martron, L.B.J.M.;Koolman, D.; Hutzinger, O. 1981. Accu-mulation and elimination kinetics of di-,tri-, and tetrachlorobiphenyls by goldfishafter dietary and aqueous exposure.Chemosphere, 10:811-815.

Cerdeira, R.G.P.; Ruffino, M.L.; Isaac, V.J.1997. Consumo de pescado e outrosalimentos pela população ribeirihha do

Lago Grande de Monte Alegre, PA –Brasil. Acta Amazônica, 27(3):213-228.

Chiou, C.T.; Freed, V.H.; Schmedding, D.W.;Kohnert, R.L. 1977. Partition coefficientand bioaccumulation of selected organicchemicals. Environmental Science andTechnology, 11(5):475-478.

Cowgill, U.M.; Gowland, R.T.; Ramirez, C.A.;Fernandez, V. 1991. The history of achlorpyrifos spill : Cartagena, Colombia.Environment International, 17:61-71.

Cunha Bastos, V.L.F.; Silva Filho, M.V.;Rossini, A.; Cunha Bastos, J. 1999. Theactivation of parathion by brain and liverof a Brazilian suckermouth benthic fishshows comparable in vitro kinetics. Pesti-cide Biochemistry and Physiology,64(3):149-156.

Ecobichon, D.J. 1991. Toxic effects of pesti-cides. In : Amdur, M.O.; Doull, J.;Klaassen, C.D. (Ed). Casarett and Doull’stoxicology. The basic science of poisons.4e édition. Mc Graw-Hill, New York. pp.565-622.

Ferreira, E.J.G.; Zuanon, J.A.S.; dos Santos,G.M. 1996. List of commercial fish spe-cies from Santarém, State of Pará, Brazil.NAGA, The ICLARM Quarterly, 19(3):41-44.

Ferreira, E.J.G.; Zuanon, J.A.S.; dos Santos,G.M. 1998. Peixes comerciais do MédioAmazonas, região de Santarém, Pará.Cooleção Meio Ambiente: Série EstudosPesca. 18. Instituto Brasileiro do MeioAmbiente e dos Recursos NaturaisRenováveis (IBAMA), Brasília. 212 pp.

Furch, K. 1997. Chemistry of várzea and igapósoils and nutrient inventory of their flood-plain forests. In : Junk, W.J. (Ed). TheCentral Amazon Floodplain. EcologicalStudies. 126. Springer-Verlag, Berlin. pp.47-67.

Gallo, M.A.; Lawryk, N.J. 1991. Organic phos-phorus pesticides. In : Hayes, W.J.; Laws,E.R. (Ed). Handbook of pesticide toxicol-ogy. 2: Classes of pesticides. AcademicPress Inc., San Diego. pp. 917-1123.

Goulding, M.; Carvalho, M.L. 1982. Life his-tory and management of the tambaqui

336 Soumis et al.

(Colossoma macropomum, Characidae):An important Amazonian food fish.Revista Brasileira de Zoologia, 1:107-133.

Hall, R.J.; Kolbe, E. 1980. Bioconcentration oforganophosphorus pesticides to hazardouslevels by amphibians. Journal of Toxicol-ogy and Environmental Health, 6:853-860.

Hughes, D.N.; Boyer, M.G.; Papst, M.H.;Fowle, C.D.; Rees, G.A.V.; Baulu, P.1980. Persistence of three organophospho-rus insecticides in artificial ponds andsome biological implications. Archives ofEnvironmental Contamination and Toxi-cology, 9:269-279.

IPCS. 1986. Organophosphorus insecticides:A general introduction. World Health Or-ganization (WHO), Geneva.

Isaac, V.J.; Milstein, A.; Ruffino, M.L. 1996.A pesca artesanal no Baixo Amazonas:análise multivariada da captura porespécie. Acta Amazônica, 26(3):185-208.

Junk, W.J.; Soares, M.G.M.; Saint-Paul, U.1997. The fish. In : Junk, W.J. (Ed). Thecentral Amazon floodplain. Ecological Stud-ies. 126. Springer-Verlag, Berlin. pp. 385-408.

Klaassen, C.D.; Eaton, D.L. 1991. Principlesof toxicology. In : Amdur, M.O.; Doull,J.; Klaassen, C.D. (Ed). Casarett andDoull’s toxicology. The basic science ofpoisons. 4th edition. McGraw-Hill, NewYork. pp. 12-49.

Lebel, J.; Roulet, M.; Mergler, D.; Lucotte, M.;Larribe, F. 1997. Fish diet and mercury ex-posure in a riparian Amazonian popula-tion. Water, Air and Soil Pollution, 97:31-44.

Lu, F.C. 1995. A review of the acceptable dailyintakes of pesticides assessed by WHO.Regulatory Toxicology and Pharmacology,21:352-364.

Ohly, J.J. 1987. Untersuchungen über dieEignung der natürlichen Pflanzenbeständeauf den Überschwemmungsgebieten(várzea) am mittleren Amazonas,Brasilien, als Weide für den Wasserbüffel(Bubalus bubalis) während derterrestrischen Phase des Ökosystems.

Göttinger Beiträge zur Land- undForstwirtschaft in den Tropen undSubtropen, 24.

Ohly, J.J.; Junk, W.J. 1999. Multiple use ofCentral Amazon floodplains: reconcilingecological conditions, requirements for en-vironmental protection, and socioeco-nomic needs. Várzea, 13:283-299.

Padoch, C.; Pinedo-Vasquez, M. 1999. Farm-ing above the flood in the várzea ofAmapá : Some preliminary results of theProjeto Várzea. Várzea, 13:345-354.

Racke, K.D. 1992. Degradation of organophos-phorus insecticides in environmental ma-trices. In : Chambers, J.E.; Levi, P.E. (Ed).Organophosphates. Chemistry, fate and ef-fects. Academic Press, San Diego. pp. 47-78.

Ruffino, M.L. 1996. Potencialidades dasvárzeas para os recursos pesqueiros : umavisão sócio-econômica e ecológica.Empresa Brasileira de PesquisaAgropecuária (EMBRAPA) – ConselhoAssessor Regional Norte, Boa Vista,Roraima, Brazil.

Ruffino, M.L.; Isaac, V.J. 1995. Life cycle andbiological parameters of several BrazilianAmazon fish species. NAGA, TheICLARM Quarterly, 18(4):41-45.

Ruffino, M.L.; Isaac, V.J.; Milstein, A. 1998.Fisheries ecology in the lower Amazon: atypical artisanal practice in the tropics.Ecotropica, 4:99-114.

Sabharwal, A.K.; Belsare, D.K. 1986. Persis-tence of methyl parathion in a carp rear-ing pond. Bulletin of EnvironmentalCotamination and Toxicology, 37:705-709.

Saint-Paul, U.; Bayley, P.B. 1979. A situaçãoda pesca na Amazônia Central. ActaAmazônica, 9(4):109-114.

Salati, E. 1985. The climatology and hydrol-ogy of Amazonia. In : Prance, G.T.;Lovejoy, T.E. (Ed). Key environment :Amazonia. Pergamon Press, New York.pp. 18-48.

Scatena, F.N.; Walker, R.T.; Homma, A.K.O.;De Conto, A.J.; Ferreira, C.A.P.;Carvalho, R.A.; Neves da Rocha, A.C.P.;


Don Santos, A.I.M.; De Oliveira, P.M.1996. Cropping and fallowing sequencesof small farms in the "terra firme" land-scape of the Brazilian Amazon : a casestudy from Santarém, Pará. EcologicalEconomics, 18:29-40.

Shrimpton, R.; Giugliano, R. 1979. Consumode alimentos e alguns nutrientes emManaus. 1973-4. Acta Amazônica,9(1):117-141.

Smith, N.J.H. 1979. Pesca no Rio Amazonas.INPA/CNPq, Manaus. 154 pp.

Soares, M.G.M.; Almeida, R.G.; Junk, W.J.1986. The trophic status of the fish faunain Lago Camaleão, a macrophyte domi-nated floodplain lake in the middle Ama-zon. Amazoniana, 9(4):511-526.

Soumis, N.; Roulet, M.; Lucotte, M. 2000.Characterization of pesticide consumptionin the county of Santarém, Pará, Brazil.Acta Amazônica, 30(4):615-628.

Takase, I.; Oyama, H. 1985. Uptake andbioconcentration of disulfoton and its oxi-dation compounds in carp, Cyprinus

carpio L. Journal of Pesticide Science,10(1):47-53.

Torres, J.P.M.; Pfeiffer, W.C.; Markewitz, S.;Panse, R.; Malm, O.; Japenga, J. 2002.Dichlorodiphenyl trichlorocthane in soil,river sediment, and fish in the Amazon inBrazil Environmental Reseaech, 88(2):134-139.

Van Dijk, H.F.G.; Guicherit, R. 1999. Atmo-spheric dispersion of current-use pesti-cides : A review of the evidence frommonitoring studies. Water, Air, and SoilPollution, 115(1-4):21-70.

Wauchope, R.D. 1978. The pesticide contentof surface water draining from agriculturalfields – A review. Journal of Environmen-tal Quality, 7:459-472.0

Zarin, D.J. 1999. Spatial heterogeneity andtemporal variability of some Amazonianfloodplain soils. Várzea, 13:313-321.

Zeng, N. 1999. Seasonal cycle and interannualvariability in the Amazon hydrologiccycle. Journal of Geophysical Research,104(D8):9097-9106.

Recebido: 20/03/2002

Aceito: 10/02/2003

338 Soumis et al.

PRESENCE OF ORGANOPHOSPHATE INSECTICIDES IN ...

Documents

Transcript of PRESENCE OF ORGANOPHOSPHATE INSECTICIDES IN ...