Small Ruminant Research 106S (2012) S49S53

Contents lists available at SciVerse ScienceDirect

Small Ruminant Research

journal homepage: www.elsevier.com/locate/smallrumres

Sheep milk as a potential indicator of environmental exposure to dioxin-like polychlorinated biphenyls (dl-PCBs)
a , M. Perugini a, , M. Esposito b , L. Baldi c , M. Amorena a E.G. Herrera Nunez
a b c

Dipartimento di Scienze degli Alimenti, University of Teramo, Via Crispi 212, 64100 Teramo, Italy IZS del Mezzogiorno, via Salute 2, 80055 Portici (Na), Italy ORSA Osservatorio Regionale Sicurezza Alimentare, via Salute 2, 80055 Portici (Na), Italy

a r t i c l e
Keywords: Dioxin-like PCBs Sheep milk GIS

i n f o

a b s t r a c t
Dioxin-like polychlorinated biphenyls are lipophilic persistent organic pollutants which accumulate in the fat tissue of animals. Moreover, in mammals, these contaminants are secreted into milk during lactation periods. The contamination level of milk is closely related to the animal dl-PCBs exposure through diet, inhalation and dermal contact. The study of the relationship between the environment in which the animals live and the milk contamination level is a key component in assessing the presence of pollutants in dairy products. In this paper, soil was selected as a reliable matrix of environmental pollution with dlPCBs as it reects as well the possible contamination of vegetation and air in the areas where sheep graze and live. However, dl-PCBs analysis could be problematic due to the high cost. To overcome this problem, the approach of this study was focused on using GIS and geostatistics to integrate data from different research institutions, aiming to assess the relationship between milk and soil contaminated with dl-PCBs. Therefore, samples of soil and sheep milk collected in Campania (Italy) during 2008 monitoring plans were introduced in a GIS. The relationship between the two variables showed heteroscedasticity, and the global regression technique OLS indicated a R2 of 0.24. To improve the OLS results, the altitude was integrated in the GIS considering that this variable can inuence PCBs contamination. The results showed that altitude is strongly related with soil contamination but was not strongly related with milk contamination. Thus, analyzing the relationship between milk and soil contamination dividing data in three different groups of altitude ranges, the R2 values were higher. Furthermore, this correlation changes with the altitude, becoming weaker with higher altitude. These results show the possibility to use sheep milk as indicator of PCBs contamination at least from sheep bred under 300 m of altitude. 2012 Elsevier B.V. All rights reserved.

1. Introduction The dioxin-like PCBs (dl-PCBs) are a group of 12 polychlorinated biphenyls, showing chemical and toxicological

This paper is part of a supplement titled SIPAOC Congress 2010 guest edited by Antonello Carta, Maria Teresa Manfredi, Giuseppe Moniello and Massimo Trabalza-Marinucci. Corresponding author. Tel.: +39 0861 266988; fax: +39 0861266989. E-mail address: mperugini@unite.it (M. Perugini). 0921-4488/$ see front matter 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.smallrumres.2012.04.033

properties similar to those of dioxins. They are persistent organic pollutants (POPs) (Colles et al., 2008; Liu and Liu, 2009) and, due to their carcinogenic and mutagenic effects and their endocrine disrupting action (Salihoglu and Tasdemir, 2009; Vives et al., 2008) represent a threat to human health. Since they are widespread environmental contaminants, they are found in environmental samples, in food for human consumption, in animal feed and in grassland (De Mul et al., 2008). In the environment, these compounds decompose very slowly and due to their liposolubility they accumulate in the food chain and

S50

E.G. Herrera Nu nez et al. / Small Ruminant Research 106S (2012) S49S53 Table 1 dl-PCBs concentrations in sheep milk and soil samples (WHO-TEQ-ng/kg). Number of observations 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 dl-PCBs concentrations in milk WHO-TEQ ng/kg 0.84 1.885 0.567 0.71 0.66 1.163 0.582 0.605 0.934 0.08 0.599 0.41 0.623 1.4 2.65 0.859 0.597 0.578 6.57 0.58 1.31 0.768 4.8 4.74 1.21 4.75 dl-PCBs concentrations in soil WHO-TEQ ng/kg 0.03 0.26 0.29 0.08 0.03 0.38 0.17 0.25 0.21 0.14 0.07 0.22 0.09 0.12 0.31 0.10 0.08 0.16 0.45 0.05 0.43 0.37 0.38 0.37 0.14 0.26

represent a serious hazard for human and animal health. Due to atmospheric deposition, these compounds accumulate in the leafy vegetation (such as grass) and in the soil, that acts as a natural sink. Once settled into soils, dl-PCBs concentration is directly related with the environmental pollution (Zhang et al., 2009; Vives et al., 2008). Grazing animals as sheep, ingest contaminated grass and these contaminants are rapidly absorbed from the gastrointestinal tract (Faye and Sinyavskiy, 2008; Esposito et al., 2010). Dl-PCBs once absorbed by animals, due to their lipophilic nature, accumulate in the liver and adipose tissues and during lactation periods, milk represent the main excretion pathway for these contaminants (Schmid et al., 2003; Esposito et al., 2009). Campania (Italy) is strongly inclined toward agriculture, and it is a region where sheep extensive breeding is largely used. Until the 1990s the shepherd system was the most important way of breeding: it was based on pastures located in places far from each other; now the situation has changed, with the permanent breeding. Sheep introduce dl-PCBs both eating forage which has been contaminated by atmospheric deposition and ingesting direct contaminated soil from pasture. In assessing the impact of soil contamination on plants, it appears that for PCBs there is an increasing concentration with decreasing soil particle size (Smith and Jones, 2000). Considering that the sheep breeding system is permanently in the open air, milk could be an ideal matrix for monitoring the PCBs presence in the environment providing exposure information through a non-invasive collection method. Since 2007, the Campania Region Health Authority launched a dioxin surveillance plan in order to control the pollution degree and examine contamination extent and sources. This plan was focused on milk (sheep, cow and buffalo) samples collection to determine the PCDD/F and dl-PCB levels. At the same time the Agenzia Regionale Protezione Ambientale della Campania (ARPAC) monitored the levels of PCBs-dioxins in the soil. In this context, Geographic Information Systems (GIS) rise as a powerful tool for data analysis. GIS is a computer-based technology and methodology for collecting, managing, analyzing, modeling and presenting geographic data for a wide range of applications (Davis, 2001). Features on maps have spatial relationships which can be very important in many applications, and GIS is an excellent tool for determining such characteristics (Davis, 2001). One of the chief virtues of GIS is that it allows the visualization of spatial data as well as providing a means of utilizing fuzzy data. While the quantitative science prefers clear and precise facts, GIS provides a way to include data that is not so pristine (Schuurman, 2004). GIS offer a variety of geo-processing features, all dealing with the management of data (e.g. dissolving, clipping, merging and intersecting) allowing to aggregate information from one layer to another. Transformation, conversion or the translation of data from one unit or level to another is the next step of coupling. Most important for the process of integration is the interpolation of single data-points to area-wide information. Interpolation creates a surface, a regionalization, from point-data by geostatistical operations (Jopp et al., 2011).

This paper aims to study the relationship between sheeps milk and soil contaminated with dl-PCBs as a rst stage of further studies to assess a valid bioindicator of environmental pollution. Data from different research institutes were integrated in a GIS as it allows to manage different information layers and perform spatial analyses.
2. Materials and methods Sheep milk samples were collected from 26 farm located in Campania. The sheep breeding was extensive livestock farming and no feed were given to the animals. Total dl-PCBs concentration in sheeps milk samples is expressed as the sum of 12 congeners in WHO-TEQ pg/g fat in agreement with the Regulation EC 1881/2006 (Table 1). All data derive from the EU monitoring plan, Residues National Plan (PNR) and Surveillance Regional Plan for PCBs and dioxins carried out in 2008 by Campania Region. These data were obtained in collaboration with the Istituto Zooprolattico Sperimentale del Mezzogiorno-Portici, Italy, where dl-PCBs analyses in sheeps milk were carried out, according to the EPA method 1613 (USEPA, 1994). Total dl-PCB concentrations in soil samples are expressed as the sum of 12 congeners (WHO-TEQ ng/kg) and all data derived from the ARPA Campania web site (http://www.arpacampania.itat cont area2.asp?id=10). 2.1. Geostatistical analyses Data was analyzed using ESRI ArcGis software (Environmental Systems Research Institute, Redlands, CA) version 9.3. The Ordinary Kriging method was used to interpolate soil contamination values. The parameters of lag size, number of neighbors and shape of semivariogram were combined to obtain the best result. The different maps obtained were compared using cross-validation and the best combination was the one with the less root mean square error. The interpolation of soil contamination values was performed to overcome the problem of different sampling locations between the variables (milk and soil), obtaining a continuous map. This permitted to estimate soil contamination values for each point where milk was sampled. The relationship between soil and milk contaminated with dl-PCBs was analyzed using Ordinary Least Squares (OLS). Furthermore, Koenker and Global Morans Index statistical tests were applied. The Koenker test was performed to determine

E.G. Herrera Nu nez et al. / Small Ruminant Research 106S (2012) S49S53

S51

Fig. 1. Interpolation of soil contamination values using Ordinary Kriging with a circular semivariogram. The map shows the interpolation between dl-PCBs values found in the soil and milk sampling points. heteroscedasticity. In case of statistically signicance, the linear regression model is heteroscedastic, and the robust probabilities are the only reliable results (Watson and Teelucksingh, 2002; Seddighi et al., 2000). The Global Morans Index test was performed to the residuals of the regression to determine the existence of a spatial autocorrelation. The null hypothesis states that there is no spatial clustering of the values associated with the geographic features in the study area. When the P-value is small and the absolute value of the Z score is large enough that it falls outside of the desired condence level, the null hypothesis can be rejected. If the index value is greater than 0, the set of features exhibits a clustered pattern. If the value is less than 0, the set of features exhibits a dispersed pattern (Sullivan and Unwin, 2003).

3. Results The best interpolation of soil contamination values was obtained using a circular semivariogram, a lag size of 4619.94 m and 15 neighbors (Fig. 1). Sheeps milk data were integrated to the map obtained with Ordinary Kriging to perform regression analysis

between milk and soil contaminated with dl-PCBs. The relationship between the variables presented a R2 of 0.24. Furthermore, for a 95 percent condence level, the relationship between soil and milk contamination was statistically signicant heteroscedastic (P < 0.05). Therefore, the signicance of the soil contamination was determined by the robust probabilities, which was statistically signicant (P < 0.05). The Global Morans Index test showed that for a 95 percent condence level, the residuals were clustered. Considering the results of the Koenker test and that heteroscedasticity could be caused from the omission of relevant regressors, we integrated the Digital Elevation Model (DEM) to calculate the altitude for each soil and milk sampling point in order to improve the OLS performance. The altitude and milk contamination were not strongly related (R2 = 0.18) but the altitude was strongly related to soil contamination (R2 = 0.53).

S52

E.G. Herrera Nu nez et al. / Small Ruminant Research 106S (2012) S49S53

Fig. 2. Separation of data in three groups of different altitude ranges ( = 0100 m; = 100300 m; altitude on the relationship between milk and soil contamination in the studied area.

= 3001000 m) in order to evaluate the inuence of

The visual analysis of all data integrated in GIS showed that sampling points were distributed geographically in three groups (Fig. 2). Calculating the altitude from DEM for each point, data was divided in the following altitude ranges: 0100 m, 100300 m and 3001000 m. The regressions performed for each group showed R2 of 0.91, 0.81 and 0.35 for the groups of 0100 m, 100300 m and 3001000 m respectively. The Global Morans Index indicated random distribution of the residuals for the three groups. 4. Discussion The geostatistic approach allowed the description and modeling of spatial patterns of soil contamination with dl-PCBs, resulting in a prediction of soil pollution at unsampled locations. Kriging is the preferred method in many spatial statistical analyses and is considered the best linear unbiased estimator, therefore, the resulting map was a reliable information layer to estimate pollution concentration in the locations where milk was sampled. Performing the OLS, the coefcient of determination was low and the residuals were clustered (spatial autocorrelation). The Koenker test showed that heteroscedasticity was present indicating that the regression was missing a

key explanatory variable. Both, the heteroskedasticity and autocorrelation lead to biased and inefcient estimates of parameters with considerably standard errors. Therefore, it was necessary to improve the OLS. Thus, the Digital Elevation Model of the Campania Region was introduced to analyze the relationship between the variables considering altitude, as the latter can inuence dl-PCBs concentrations (Wang et al., 2009). Although, altitude and milk contamination were not strongly related, the altitude inuences the soil contamination. Considering that PCBs milk contamination is inuenced by soil pollution and the latter is inuenced by altitude, we have hypothesized that milk contamination was inuenced by soil contamination in different order of magnitudes depending on altitude. In addition, exploiting the visual advantages of GIS, which allowed the visualization on data in a geographical context, the separation in three groups of different altitude ranges was performed to test the hypothesis of the inuence of soil in milk contamination. The separation of data in three groups with different altitude ranges leaded to better R2 values, conrming that the relationship between milk and soil contamination was variable within the study area. Moreover, the relationship between milk and soil contaminated with dl-PCBs

E.G. Herrera Nu nez et al. / Small Ruminant Research 106S (2012) S49S53

S53

apparently weakens with altitude considering the R2 values of the three groups (0.91 > 0.81 > 0.35). Nonetheless, more data is necessary to conrm such pattern. Since soil reects environmental pollution very well (Zhang et al., 2009; Vives et al., 2008) and sheeps milk showed to be strongly related with soil contamination, the results obtained in this paper, could be the base for future research in assessing bioindicators of environmental contamination with dl-PCB. 5. Conclusions The approach proposed in this paper is advantageous as it allows to integrate and analyze data from different sources. Furthermore, the study of environmental variables seems to be highly related to geographic location, thus, GIS and geostatistics are essential in cutting edge research. The results obtained in this study evidence that milk and soil contaminated with dl-PCBs are strongly correlated suggesting that the sheep milk reects environmental PCBs contamination. However, the relationship between these two variables is inuenced by altitude and the R2 values decrease with higher altitude. This research paper evidences the potential use of sheep milk as indicator of PCBs environmental pollution and represents an innovative method to create a monitoring contamination system. Moreover, if integrated with more data and validated with more temporal series, a simple predictive model could be constructed. Conict of interest statement None of the authors have any nancial or personal interest that would inappropriately inuence or bias the contents of this paper. References
Colles, A., Koppen, G., Hanot, V., Nelen, V., Dewolf, M., Noel, E., Malisch, R., Kotz, A.K.K., Biot, P., Vinkx, C., Schoeters, G., 2008. Fourth WHOcoordinated survey of human milk for persistent organic pollutants (POPs): Belgian results. Chemosphere 73, 907914. Commission Regulation (EC) No. 1881/2006 (19 December 2006) setting maximum levels for certain contaminants in foodstuffs. Off. J. Eur. Union, 364, 524. Davis, B., 2001. GIS: A Visual Approach, second edition. OnWord Press, New York.

De Mul, A., Bakker, M.I., Zeilmaker, M.J., Traag, W.A., van Leeuwen, S.P.J., Hoogenboom, R.L.A.P., Boon, P.E., van Klaveren, J.D., 2008. Dietary exposure to dioxins and dioxin like PCBs in the Netherlands anno 2004. Regul. Toxicol. Pharmacol. 51, 278287. Esposito, M., Cavallo, S., Serpe, F., DAmbrosio, R., Gallo, P., Colarusso, G., Pellican, R., Baldi, L., Guarino, A., Serpe, L., 2009. Levels and congener proles of polychlorinated dibenzo-p-dioxins, polychlorinated dibenzofurans and diodi-like polychlorinated biphenyls in cows milk collected in Campania, Italy. Chemosphere 77, 12121216. Esposito, M., Serpe, F., Neugebauer, F., Cavallo, S., Gallo, P., Colarusso, G., Baldi, L., Iovane, G., Serpe, L., 2010. Contamination levels and congener distribution of PCDDs, PCDFs and diodi-like PCBs in buffalos milk from Caserta province (Italy). Chemosphere 79, 341348. Faye, B., Sinyavskiy, Y., 2008. Impact of pollution on animal products. In: Proceedings of the NATO Advanced Research Workshop on Impact of Pollution on Animal Products. Jopp, F., Reuter, H., Breckling, B., 2011. Modelling Complex Ecological Dynamics, An Introduction into Ecological Modeling. Springer, New York. Liu, J., Liu, W., 2009. Distribution of polychlorinated dibenzo-p-dioxins and dibenzofurans (PCDDs/Fs) and dioxin-like polychlorinated biphenyls (dioxin-like PCBs) in the soil in a typical area of eastern China. J. Hazard. Mater. 163, 959966. Salihoglu, G., Tasdemir, Y., 2009. Prediction of the PCB pollution in the soils of Bursa, and industrial city in Turkey. J. Hazard. Mater. 164, 15231531. Schmid, P., Gujer, E., Zenneg, M., Studer, C., 2003. Temporal and local trends of PCDD/F levels in cows milk in Switzerland. Chemosphere 53, 129136. Schuurman, N., 2004. GIS A Short Introduction. Blackwell Publishing, Oxford. Seddighi, H.R., Lawler, K.A., Katos, A.V., 2000. Econometrics: A Practical Approach. Taylor & Francis Group, Routledge. Smith, K.E.C., Jones, K.C., 2000. Particles and vegetation: implications for the transfer of particle-bound organic contaminants to vegetation. Sci. Total Environ. 246, 207236. Sullivan, D.O., Unwin, D.J., 2003. Geographic Information Analysis. John Wiley &Sons, Hoboken. US Environmental Protection Agency (USEPA), October 1994. USEPA Method 1613. Tetra-through Octa-chlorinated Dioxins and Furans by Isotope Dilution HRGC/HRMS. USEPA Ofce of Water Engineering and Analysis Division, Washington, DC. Vives, I., Muller, A., Umlauf, G., Christoph, E., Mariani, G., Skejo, H., Cenci, R., Sena, F., Beone, G., 2008. Levels of PCDD/Fs and trace elements in supercial soils of Pavia Province (Italy). Environ. Int. 34 (7), 9941000. Wang, P., Zhang, Q., Wang, Y., Wang, T., Li, X., Li, Y., Ding, L., Jiang, G., 2009. Altitude dependence of polychlorinated biphenyls (PCBs) and polybrominated diphenyls ethers (PBDEs) in surface soil from Tibetan Plateau, China. Chemosphere 76, 14981504. Watson, P., Teelucksingh, S., 2002. A Practical Introduction to Econometric Methods, Classical and Modern. The University of the West Indies Press. Zhang, S., Peng, P., Huang, W., Li, X., Zhang, G., 2009. PCDD/PCDF pollution in soils and sediments from the Pearl River Delta of China. Chemosphere 75, 11861195.