Table of Contents Author Guidelines Submit a Manuscript
Journal of Spectroscopy
Volume 2018, Article ID 5152832, 6 pages
https://doi.org/10.1155/2018/5152832
Research Article

Detection of Veterinary Antimicrobial Residues in Milk through Near-Infrared Absorption Spectroscopy

1Grupo de Engenharia e Espectroscopia de Materiais, Departamento de Física, Universidade Federal de Juiz de Fora, 36036-900 Juiz de Fora, MG, Brazil
2Departamento de Ciências de Alimentos, Universidade Federal de Lavras, 37200-000 Lavras, MG, Brazil

Correspondence should be addressed to Leandro da Conceição Luiz; moc.liamg@ordnaelorvil

Received 25 February 2018; Accepted 6 May 2018; Published 3 June 2018

Academic Editor: Maria Carmen Yebra-Biurrun

Copyright © 2018 Leandro da Conceição Luiz et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract

This study focuses on detection of antimicrobial residues in milk through Fourier transform near-infrared spectroscopy. Simulated and real samples were considered. The simulated ones take into account veterinary drugs added in milk samples in the following concentrations: enrofloxacin 100 μg/L, terramycin 100 μg/L, and penicillin 4 μg/L. The statistical tool used to discriminate the samples was the principal component analysis (PCA). Our results show that, with this experimental procedure, it is possible to discriminate different types of antimicrobials dissolved in milk. Moreover, the methodology was able to detect real sample milked on different days after the injection of ceftiofur hydrochloride which is in principle a zero grace period antimicrobial. The methodology proved to be fast and accurate within the maximum residue limits allowed by European Agency for Medicinal Products and Ministry of Agriculture Livestock and Food Supply from Brazil.

1. Introduction

According to the Food and Agriculture Organization of the United Nations, milk is one of the most consumed foods in the world. It not only has its importance in the nutritional level, but also plays an important role in the economy. Global consumption of milk and its derivatives exceeds 6 billion of consumers [1].

The milk contains protein, carbohydrates, lipids, minerals, and vitamins which accomplish important biochemical and nutritional functions, particularly to children and elderly people. The bovine milk contains about 87.1% of water, 4.0% of fat, 3.3% of protein, 4.6% of lactose, and 0.7% of ash. The basic protein content in milk is casein in 78.3%, whey protein 19%, and others totaling 2.7% [2, 3]. Concerning carbohydrates, lactose is the main one. Constituted by two monosaccharides, glucose and galactose, they carry nutritional important functions, such as providing 16.8 kJ/g of energy to people [4, 5]. Milk and dairy products are considered a good source of calcium due its high bioavailability. The latter can be understood as the fraction of ingested nutrient and food that is absorbed and used in physiological and normal metabolic functions and storage [6, 7]. Although the potentiality of milk as food is unquestionable, restrictions on its intake exist in people allergic to lactose and casein.

Adulterations in milk have been highly reported in developing countries, such as Pakistan, Brazil, India, and China [8]. Mostly the aim is to increase volume with the addition of water [9]. However, there are other problems, such as the contamination of milk by residues of veterinary drugs that may be present when the cow is milked in the grace period. The most common drugs are antimicrobials and anti-inflammatories. They are widely used in the treatment of dairy cattle diseases, such as mastitis, diarrhea, and lung diseases, also in prevention and control, or to increase the production and growth of animal [1013]. A study performed by Van Boeckel et al. showed that, between 2000 and 2010, the consumption of antibiotics by the world population has increased by 36% and is related to the appearance of drug-resistant bacteria. From this total, 76% is mainly due to the countries that composes the BRICS (Brazil, Russia, India, China, and South Africa). Fraction of this increase is due to ingestion of animals or their food derivatives contaminated with antibiotics [14].

The overuse of drugs in dairy herd results in detectable traces in the milk. When their concentrations are over the maximum residue limit (MRL), they can cause health damage to the consumer ranging from allergic reactions to bacterial resistance [13] Contaminated milk is harmful to health and should be discarded [11, 15, 16]. In 2014, Van Boeckel et al. related the excessive use of antibiotics in animals to the appearance of super bacteria in humans. In order to monitor the milk content, the regulatory agencies use a variety of analytical methods to detect antibiotics traces in the milk, such as high-performance liquid chromatography (HPLC), gas chromatography coupled to mass spectroscopy (CG-MS), and antimicrobial detection kits. Nevertheless, there are drawbacks in their use such as sample preparation, qualified manpower, complex procedures, time consuming, and high cost. Moreover, normally the tests of antibiotics are specific to a class of antimicrobials.

The search for high sensitive techniques that allow the detection of residues of antibiotics in milk has been carried out for decades. In 1996, Verdon and Couedor used a reverse-phase HPLC technique to determine ampicillin residues which is able to detect 3 μg/L and to quantify up to 10 μg/L of the drug [17]. In 2002, Sivakesava and Irudayaraj carried out a study showing the feasibility of measuring tetracycline at μg/L levels with Fourier transform near-infrared (FT-NIR) spectroscopy and Fourier transform medium-infrared (FT-MIR) spectroscopy. Nevertheless, they reported high prediction errors and suggested that the methodology should be confirmed with naturally contaminated samples and other drug residues [18]. In 2003, Jankovská and Sustová used FT-NIR to analyse cow milks. However, the technique was used to describe physical-chemical characteristics of milk. In addition, partial least squares (PLS) regression was used to develop calibration models for the examined milk components. Through these results, they suggest that the NIR spectroscopy is applicable for a rapid analysis of milk composition [19]. In 2010, Brandão et al. studied fat concentration in milk samples by means of noninvasive techniques, FT-IR and FT-NIR absorption. They concluded that the wavelength of 2308 nm can be used to determine the fat concentration of milk without other components’ influence [20]. In 2014, Zhang et al. [21] examined UHT and pasteurized milks to verify the presence of residues of tetracyclines, sulfonamides, sulfamethazine, and quinolone. The milk samples were collected in highly populated cities of China and analysed by the enzyme-linked immunosorbent assay (ELISA). No residue of veterinary drugs has exceeded MRL established by China, European Union, and CAC (Codex Alimentarius Commission). Nevertheless, because of the high number of residues present in milk, they recommended that the control mechanisms should be rigorously applied in order to keep these residues at safe levels. In 2015, Moharana et al. [22] analysed the veterinary drug enrofloxacin in cow milk samples obtained from two cities of India. According to the authors, the enrofloxacin is the most rampantly used drug in veterinary practice. To analyse the samples, they used reverse-phase HPLC. With a limit of detection of 100 μg/L, they have verified that 8% of the samples had values above the MRL. From the brief historical review described above, it is clear that there is a need for more in-depth studies exploring detection limits, embracing bigger classes of antibiotics, and analysing real samples.

This work deals with Fourier transform near-infrared (FT-NIR) spectroscopy associated with principal components analysis (PCA) to detect traces of veterinary antimicrobials (enrofloxacin, terramycin, penicillin, and ceftiofur hydrochloride) bellow the MRL allowed by the legislation of European Agency for Medicinal Products (EMEA) which is adopted in Brazil by Ministry of Agriculture Livestock and Food Supply (MAPA) [23]. The detection of the ceftiofur hydrochloride in milk was performed for two consecutive days after the drug has been administered to the animal.

2. Materials and Methods

The analyses were performed in the Process and Products Laboratory (LPP) and in the Materials Spectroscopy Laboratory (LEM), located in the Physics Department of Federal University of Juiz de Fora, Brazil.

2.1. Milk Samples

Some samples used in this work were raw milk from nonmedicated cows. They were collected at a farm located in the city of Rio Pomba, MG, Brazil, near the laboratory. After milking, the raw milk samples were immediately stored and kept refrigerated at 5°C until analysis which was performed after approximately one hour. The samples were submitted to chemical-physical analysis to verify their conformity with the established standards [24], that is, cryoscopy, density, pH, acidity (Dornic and Alizarol tests), fat, protein, lactose, and solids [2527]. Each measurement was performed in triplicate. The results are shown in Table 1.

Table 1: Result of physical-chemical characterization of milk samples. Average values and their standard deviations.
2.2. Contamination Simulation

Initially, two portions were separated, one as a control sample and the other as a self-controlled one with a veterinary drug. The drugs used were the antimicrobials: enrofloxacin Baytril® injective 10% that has 10 g of enrofloxacin plus 100 ml of vehicle in its composition, terramycin/LA Zoetis/Pfizer® injective that has 20 g of oxytetracycline plus 100 g of vehicle in its composition, and reinforced pentabiotic Pfizer with penicillin.

The simulation was done according to the active principle of each drug and not in its volume. More explicitly, in the analysis with PCA, the discrimination of contaminated milk in μg/L is given through the active principle, MRL as provided by legislation, and not through the drug (excipients + active principle). For this purpose, each veterinary drug was first diluted in distilled water, and finally, part of this dilution was added in the genuine milk, in order to reach the concentration of the active principle in the milk. The values equivalent to the MRL allowed by EMEA and MAPA for the antimicrobials and its metabolites are as follows: 100 μg/L to enrofloxacin, 100 μg/L to terramycin, and 4 μg/L to penicillin [22, 28, 29]. This methodology was applied previously to the anti-inflammatory sodium diclofenac [30].

2.3. Real Contaminated Samples

First, the milk free from drugs was collected from a cow used as a control. The CeF-50 Ceftiofur Agner Union injective drug, which has 50 g of ceftiofur hydrochloride plus 1 mL of vehicle in its composition, was then administered to the cow. This drug does not have a grace period. Milk is then collected for two consecutive days. All the samples were immediately stored and kept refrigerated at 5°C until analysis which was performed after approximately one hour.

2.4. Analysis Using FT-NIR Method

Analyses of the samples were carried out with the Multi Purpose FT-NIR Analyser from Bruker operating in the reflectance mode in the range of 13.500 to 3.700 cm−1 wavenumbers with a Te-InGaAs detector and 4 cm−1 of resolution. The OPUS® software version 5.5 was used for data acquisition. The samples were placed in borosilicate cuvettes with 8 mm thickness. Each analysis was performed in triplicate with 32 scans for simulated and control samples.

2.5. Statistical Analysis

The reflectance spectra and their first derivatives were made and analysed with the software OriginPro®8 SR2 v.8.0891(B891). The eigenvalues were calculated with the software BioStat version 5.3. The principal component analysis was conducted with the software The Unscrambler® X version 10.3.

3. Results

Table 1 shows the average results obtained for the milk quality parameters of the samples and the reference value established by the Brazilian Normative Instruction 62 from MAPA and the FAO/TCP/KEN/6611 [24]. The results are in the range accepted by the legislation.

Figure 1 shows the reflectance spectra (top) and their first derivatives (bottom) of genuine milk and pure antimicrobials. From the figure, one can see that the spectra are very different from each other above all, in the first derivatives.

Figure 1: Reflectance spectra of pure samples of genuine milk and veterinary drugs with their respective first derivatives.

Figure 2 shows the reflectance spectra (top) and their first derivatives (bottom) of genuine and contaminated milk samples within the MRL. On the contrary of Figure 1, now the differences between the spectral profiles are not apparent due to the very low concentration of the antibiotics. Therefore, we have to rely on a statistical model. This will be accomplished by means of PCA to discriminate control and tampered milk samples.

Figure 2: Reflectance spectra and their first derivatives of genuine and contaminated milk samples within the MRL.

The principal component of a set of data (dimensionality) is obtained by means of an analysis that consists in finding the eigenvalues of the covariance matrix [31]. Each eigenvector has a corresponding eigenvalue. The eigenvectors with higher eigenvalues are the principal components and are ordered from the higher to the smaller ones furnishing the components in significance degree [32]. Table 2 shows the explained and cumulative variances of the principal components of genuine and contaminated milk samples within the MRL.

Table 2: Explained and cumulative variances of genuine and contaminated milk samples within the MRL.

Figure 3 shows the score plot with the clustering of milk samples: control (GEN1, GEN2, GEN3), contaminated with 100 μg/L of enrofloxacin (EN100.1, EN100.2, EN100.3), contaminated with 100 μg/L of terramycin (TE100.1, TE100.2, TE100.3), and contaminated with 4 μg/L of penicillin (PE4.1, PE4.2, PE4.3).

Figure 3: Hotelling score plot of statistical analysis (PCA) showing clustering data for control milk samples GEN1, GEN2, and GEN3 (blue square); milk with 100 μg/L of enrofloxacin, EN100.1, EN100.2, and EN100.3 (green triangles); 100 μg/L of terramycin, TE100.1, TE100.2, and TE100.3 (black rhombuses); and milk with 4 μg/L of penicillin, PE4.1, PE4.2, and PE4.3 (red circles).

From Figure 3, it is clear the formation of clusters resulting from the high degree of similarity between groups of samples. Four groups are present, one in each quadrant. Group 1 (squares) refers to genuine milk (control samples) and is located in the first quadrant. Group 2 is located in the second quadrant (circles). This cluster represents a group of contaminated milk samples with 4 μg/L of penicillin. The third quadrant contains elements of group 3 which is related to contaminated milk samples with 100 μg/L of enrofloxacin (triangles), while the fourth quadrant is occupied by group 4 (rhombuses) containing contaminated milk samples with 100 μg/L of terramycin. Among the elements of the groups, none is far apart from each other which discard the presence of outliers. The PCA accurately discriminated the samples in groups despite the very low concentration of the antimicrobials. From PC1, one can observe that the penicillin and enrofloxacin have the same score in contraposition to terramycin and genuine milk. This trend may be depicted in the absorption spectra (Figure 2). Therefore, PC1 represents the degree of milk contamination with antibiotics. Note that the group of data containing genuine milk is close to the centre of the axis. The contaminated clusters have different distances and different positions from the centre. This is related to the fact that the concentrations of the drugs are different. For example, samples with 100 μg/L of medication (groups 3 and 4) have similar position. From the above reasoning, we are led to infer that PC2 is related to the milk similitude.

All samples analysed until now were raw milk. Concerning pure antimicrobial data one may verify that the PCA is also able to discriminate between them. Figure 4 deals with such fact. It shows the PCA of average values of reflectance spectra of genuine milk, pure enrofloxacin, pure terramycin, and pure penicillin, and milk-enrofloxacin in 100 μg/L, milk-terramycin in 100 μg/L, and milk-penicillin in 4 μg/L, respectively. It can be seen that the milk-like samples (GEN, EN100, TE100, and PE4) are located very close in relation to PC1 which indicates the degree of similarity between them (cluster formation). For this reason, the samples of enrofloxacin antimicrobial (ENR), penicillin antimicrobial (PEN), and terramycin antimicrobial (TER) are so far apart from the group. Pure penicillin sample is detached from the other clusters because it was the only solid matrix (powder), while the others were liquid.

Figure 4: Score plot of statistical analysis (PCA) for average values of samples: pure enrofloxacin (ENR), pure terramycin (TER), pure penicillin (PEN), genuine milk (GEN), milk-enrofloxacin and milk-terramycin in 100 μg/L each (EN100 and TE100), and milk-penicillin in 4 μg/L.

Table 3 shows the explained and cumulative variances for each of the principal components of the average values of pure enrofloxacin, pure terramycin, pure penicillin, genuine milk, milk-enrofloxacin and milk-terramycin in 100 μg/L each, and milk-penicillin in 4 μg/L.

Table 3: Explained and cumulative variances of average values of samples: pure enrofloxacin, pure terramycin, pure penicillin, genuine milk, milk-enrofloxacin and milk-terramycin in 100 μg/L each, and milk-penicillin in 4 μg/L.

Figure 5 shows the methodology applied in a real situation. It shows data of a control milk, without medication, and contaminated milk in two consecutive days after the administration of the drug ceftiofur hydrochloride in a cow. Table 4 has the respective explained and cumulative variances from the data of Figure 5.

Figure 5: Score plot of statistical analysis (PCA) for samples of genuine milk (GEN) and milk collected for 2 days after the cow took the ceftiofur hydrochloride.
Table 4: Explained and cumulative variances of samples of genuine milk and milk collected for 2 days after the cow took the ceftiofur hydrochloride.

Cluster formation was observed for the genuine milks (GEN1, GEN 2, and GEN3) and one-day and two-day drug administered milk (1DAY and 2DAY). It turns out that the GEN and 2DAY groups are along the same PC1, showing the similarity between the groups. This is due to the metabolization of the antibiotic in the milked cow after two days of the drug administration. Reinforcement of this supposition is apparent when compared with one day milked sample with genuine milk. The PC2 is connected with time of milking as the spectroscopic measurements were performed after the last day of milking (2DAY).

4. Conclusions

This article dealt with the identification of the antimicrobials enrofloxacin, terramycin, penicillin, and ceftiofur in milk samples in the MRL permitted by regulatory agencies. The methodology developed is based on the combination of Fourier transform near-infrared (FT-NIR) spectroscopy jointly with principal component analysis (PCA). The experimental procedure was able to detect antibiotics traces in a fast and accurate way. The methodology was applied also to detect ceftiofur hydrochloride in milk of a cow where the drug was administered. Beyond trace detection, we are also able to follow the metabolization process in the animal. Our results clearly demonstrate the potentiality of the method for the development of a portable prototype.

Data Availability

The data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest

The authors declare that there are no conflicts of interest regarding the publication of this paper.

Acknowledgments

The authors thank the Brazilian funding agencies CAPES (PNPD 2871/2011), CNPq (309100/2016-0), and FAPEMIG (MPR 00004-13 and MPR 01068/16) for financial funding. Leandro da Conceição Luiz would like to thank the Chemical Department of Rural Federal University of Rio de Janeiro, Brazil, for the access to the software The Unscrambler.

References

  1. FAO/WHO, Dairy Production and Products: Milk and Milk Products, 2017.
  2. P. Walstra, J. T. M. Wouters, and T. J. Geurts, Dairy Science and Technology, CRC Press, Boca Raton, FL, USA, 2nd edition, 2006.
  3. V. K. Tavanti, L. G. Bassi, G. C. C. Ferreira et al., “Composição e capacidade de coagulação de leites de vacas holandesas e girolandas,” Revista do Instituto de Laticínios Cândido Tostes, vol. 370, no. 64, pp. 5–9, 2009. View at Google Scholar
  4. Food and Agriculture Organization of United Nations (FAO), Food Energy—Methods of Analysis and Conversion Factors, Chapter 3, Food and Agriculture Organization of United Nations, Rome, Italy, 2003.
  5. M. C. S. Pereira, L. P. Brumano, C. M. Kamiyama, J. P. F. Pereira, M. P. Rodarte, and M. A. O. Pinto, “Lácteos com baixo teor de lactose: uma necessidade para portadores de má digestão da lactose e um nicho de mercado,” Revista do Instituto de Laticínios Cândido Tostes, vol. 67, no. 389, pp. 57–65, 2012. View at Publisher · View at Google Scholar
  6. R. S. Gibson, L. Perlas, and C. Hotz, “Improving the bioavailability of nutrients in plant foods at the household level,” Proceedings of the Nutrition Society, vol. 65, no. 2, pp. 160–168, 2006. View at Publisher · View at Google Scholar · View at Scopus
  7. J. M. Carbonell-Capella, M. Buniowska, F. J. Barba, M. J. Esteve, and A. Frígola, “Analytical methods for determining bioavailability and bioaccessibility of bioactive compounds from fruits and vegetables: a review,” Comprehensive Reviews in Food Science and Food Safety, vol. 13, no. 2, pp. 155–171, 2014. View at Publisher · View at Google Scholar · View at Scopus
  8. C. E. Handford, K. Campbell, and C. T. Elliott, “Impacts of milk fraud on food safety and nutrition with special emphasis on developing countries,” Comprehensive Reviews in Food Science and Food Safety, vol. 15, no. 1, pp. 130–142, 2016. View at Publisher · View at Google Scholar · View at Scopus
  9. W. W. G. Nascimento, M. P. F. Souza, A. C. M. M. Valente, V. Anjos, M. A. M. Furtado, and M. J. V. Bell, “Results from portable and of low cost equipment developed for detection of milk adulterations,” Food Science and Technology, vol. 37, pp. 38–41, 2017. View at Publisher · View at Google Scholar · View at Scopus
  10. M. A. V. P. Brito and C. C. Lange, Resíduos de Antibióticos no Leite, Comunicado Técnico 44, Embrapa, Brasília, Federal District, Brazil, 2005.
  11. D. Pontes Netto, M. O. Lopes, M. C. S. Oliveira et al., “Levantamento dos principais fármacos utilizados no rebanho leiteiro do Estado do Paraná,” Acta Scientiarum. Animal Sciences, vol. 27, no. 1, pp. 145–151, 2005. View at Publisher · View at Google Scholar
  12. T. F. Landers, B. Cohen, T. E. Wittum, and E. L. Larson, “A review of antibiotic use in food animals: perspective, policy, and potential,” Public Health Reports, vol. 127, no. 1, pp. 4–22, 2012. View at Publisher · View at Google Scholar
  13. R. W. Han, N. Zheng, Z. N. Yu et al., “Simultaneous determination of 38 veterinary antibiotic residues in raw milk by UPLC-MS/MS,” Food Chemistry, vol. 181, pp. 119–126, 2015. View at Publisher · View at Google Scholar · View at Scopus
  14. T. P. Van Boeckel, S. Gandra, A. Ashok et al., “Global antibiotic consumption 2000 to 2010: an analysis of national pharmaceutical sales data,” The Lancet Infectious Diseases, vol. 14, no. 8, pp. 742–750, 2014. View at Publisher · View at Google Scholar · View at Scopus
  15. S. J. Forsythe, Microbiologia da Segurança Alimentar, ArtMed, Porto Alegre, RS, Brazil, 2002.
  16. L. Redding, Understanding the Use of Antibiotics on Small Dairy Farms in Rural Peru, University of Pennsylvania, Philadelphia, PA, USA, 2014, Publicly Accessible Penn Dissertations Paper 1414.
  17. E. Verdon and P. Couedor, “Determination of ampicillin residues in milk by ion-pair reversed phase high performance liquid chromatography after precolumn derivatization,” Journal of Pharmaceutical and Biomedical Analysis, vol. 14, no. 8–10, pp. 1201–1207, 1996. View at Publisher · View at Google Scholar · View at Scopus
  18. S. Sivakesava and J. Irudayaraj, “Rapid determination of tetracycline in milk by FT-MIR and FT-NIR spectroscopy,” Journal of Dairy Science, vol. 85, no. 3, pp. 487–449, 2002. View at Publisher · View at Google Scholar · View at Scopus
  19. R. Jankovská and K. Sustová, “Analysis of cow milk by near-infrared spectroscopy,” Czech Journal of Food Sciences, vol. 21, no. 4, pp. 123–128, 2011. View at Publisher · View at Google Scholar
  20. M. C. M. P. Brandão, A. P. Carmo, M. J. V. Bell, and V. C. Anjos, “Characterization of milk by infrared spectroscopy,” Revista do Instituto de Laticínios Cândido Tostes, vol. 373, no. 65, pp. 30–33, 2010. View at Google Scholar
  21. Y. D. Zhang, N. Zheng, R. W. Han et al., “Occurrence of tetracyclines, sulfonamides, sulfamethazine and quinolones in pasteurized milk and UHT milk in China’s market,” Food Control, vol. 36, no. 1, pp. 238–242, 2014. View at Publisher · View at Google Scholar · View at Scopus
  22. B. Moharana, P. K. Venkatesh, S. P. Preetha, and S. Selvasubramanian, “Quantification of enrofloxacin residues in milk sample using RP-HPLC,” World Journal of Pharmacy and Pharmaceutical Sciences, vol. 4, no. 10, pp. 1443–1450, 2015. View at Google Scholar
  23. EMEA, The European Agency for the Evaluation of Medicinal Products: Veterinary Medicines and Inspections, Committee for Veterinary Medicinal Products: Enrofloxacin, EMEA, London, UK, 2002.
  24. FAO/TCP/KEN/6611 Project, Milk Testing and Quality Control, Milk Processing Guide Series V2.
  25. Food and Agriculture Organization of United Nations (FAO), Milk Analysis.
  26. A. F. Bodenmüller, J. C. Damasceno, I. T. S. Previdelli, R. G. Santana, C. E. C. O. Ramos, and G. T. Santos, “Tipologia de sistemas de produção baseada nas características do leite,” Revista Brasileira de Zootecnia, vol. 39, no. 8, pp. 1832–1839, 2010. View at Publisher · View at Google Scholar · View at Scopus
  27. M. S. Robim, Avaliação de Diferentes Marcas de Leite UAT Comercializadas no Estado do Rio de Janeiro e o Efeito da Fraude por Aguagem na Fabricação, Composição e Análise Sensorial de Iogurte, Universidade Federal Fluminense, Niterói, Brazil, 2011, Dissertação.
  28. E. Pacheco-Silva, J. R. Souza, and E. D. Caldas, “Resíduos de medicamentos veterinários em leite e ovos,” Química Nova, vol. 37, no. 1, pp. 111–122, 2014. View at Publisher · View at Google Scholar · View at Scopus
  29. A. M. Mangerona, O. D. Mozzer, R. F. Vieira, C. R. Dias, and D. R. Cappelini, “Comparação do período de carência em tecidos e leite bovino após administração de enrofloxacino e diclofenaco sódico (Quinotril Plus®) em diferentes vias parenterais,” A Hora Veterinária, vol. 33, no. 196, 2013. View at Google Scholar
  30. L. C. Luiz, M. J. V. Bell, R. A. Rocha, T. O. Mendes, and V. C. Anjos, “Análise de resíduos de diclofenaco sódico veterinário em leites por espectroscopia no infravermelho próximo,” Revista Brasileira de Ciências da Saúde, vol. 18, no. 3, pp. 219–224, 2014. View at Publisher · View at Google Scholar
  31. R. E. Santos, “Principal component analysis applied to digital image compression,” Einstein, vol. 10, no. 2, pp. 135–139, 2012. View at Publisher · View at Google Scholar · View at Scopus
  32. S. Sayad, Principal Component Analysis, University of Toronto, Toronto, ON, Canada, 2010.