Evaluation of an enzymatic method for the rapid and specific determination of urea in raw milk. Pierre J. Broutin Bentley Instruments, Inc., 4004 Peavey road, Chaska, MN55318, USA. Summary : A new enzymatic method with an analytical speed of 150 samples per hour has been evaluated against the NF AFNOR V04-217 French standard for the routine determination of urea in raw milk . This method is based upon an enzymatic reaction and formation of a colored complex which intensity is measured spectrophotometrically and correlated to the urea concentration after instrument calibration. The evaluation was performed with 135 individual milk samples collected from five different herds, 50 bulk milk samples and 15 spiked samples. The repeatability expressed as the Standard Deviation of Repeatability (Sr) was respectively 0.46 mg/dl and 0.60 mg/dl for individual and bulk milk samples. The accuracy expressed as the Standard Deviation of Accuracy (Sy,x) was respectively 2.11 mg/dl and 1.59 mg/dl for individual and bulk milk samples. The method is considered simple, accurate and relatively inexpensive and could be used to obtain an improved urea determination in routine analysis compared to the infrared methods. 1.Introduction A large number of dairy herd control laboratories are now proposing urea testing to their members. The interest in measuring urea has considerably increased during the last five years with the advent of new technologies and a better understanding of the positive and multiple impacts of its monitoring. Various studies are suggesting that an appropriate monitoring of the urea level in dairy cattle can help to precisely meet nutritional requirements, increase the milk yield, improve reproductive performance and reduce nitrogen excretion into the environment(1,2,7). Several methods are already available for the routine analysis of urea such as the enzymatic Flow Injection Analysis (FIA) and colorimetric (e.g. DMAB, DMA...) direct methods or the indirect infrared methods [e.g. Anadis FT MI600, Bentley B2000, Foss MilkoScan...]. The chemical methods have the advantage to be more specific to the urea determination and accurate. However they can involve hazardous chemicals and be time consuming and as such are not always suitable for the routine analysis of large quantity of samples. The infrared methods have the advantage to be rapid and relatively inexpensive but significantly less accurate. Prior studies have reported an accuracy of 4.83 mg/dl and 5.3 mg/dl for the infrared filter analyzers and of 3.7 mg/dl for the Fourier Transform IR milk analyzers(3,4). Urea contained in a milk matrix doesn’t have any specific absorption wavelengths in the mid infrared region. The two major urea absorption wavelengths (1600 and 1690 cm-1) are not available in an aqueous medium because of the strong absorption of water (3,5) and the other two urea absorption wavelengths located at 1161 cm-1 and 1472 cm-1 overlap with the absorption wavelengths of the main milk components (i.e. fat, protein, lactose and citric acid...). This lack of specificity explains the lower accuracy of these methods and implies that many calibration samples are necessary to calibrate the instruments. A new simple and inexpensive automated method has been evaluated (6) which provides the specificity and accuracy of an enzymatic reaction and is suitable for the rapid routine determination of urea in raw milk with an analytical speed of 150 samples per hour . 2. Principle and Instrumentation 2.1. Principle The method is based on a modified Berthelot reaction. A small amount of milk is automatically sampled and dispensed in a well thermostated at 40/C and split enzymatically with urease into ammonia and carbon dioxide. After an incubation time, a dye containing a catalyst and an activator are added to form a green complex which intensity is correlated to the ammonia concentration. The color intensity is measured at 700 nm in a trans-reflectance flow cell and translated into an urea value after instrument calibration. Two measurements were performed per sample, one without enzyme, and one with enzyme in order to account for the natural ammonia due to the bacteria proteolytic activity and remove the interferences from the sample matrix. The differential reading allowed to precisely determine the true urea concentration. 2.2. Instrumentation and Reagents The study was carried out on a standard ChemSpec 150 ( Bentley Instruments Inc., USA) equipped with a 17-well thermostated carrousel, an autosampler, an internal computer, a pump assbly. to sample the milk and deliver the chemicals, a touch screen monitor, a flow cell and a spectrometer to monitor the absorbance at 700 nm. The ChemSpec speed was set to 150 samples per hour. The reagents made of a dye, activator and an enzyme buffered solution were supplied in a kit for 2500 samples and prepared according to the manufacturer recommendations. The reagents were stored between 2-8/C and brought up to room temperature before analyzing the samples. The same reagents were used throughout the evaluation. 3. Materials and Methods 3.1. Milk Samples 150 individual and 50 bulk milk samples were collected from five different herds from the Jura region (France) and preserved with bronopol at 0.02%. Fifteen individual milk samples were spiked with 30 mg/dl urea in order to obtain a higher variability in the data set. The evaluation took place between February and April 1999 at the French Center for Research and Control of Analysis in the Dairy Industry. 3.2. Reference Method The reference analysis was performed in duplicates using the French AFNOR V04-217 (8) manual enzymatic reference method for the determination of ammoniac and urea in milk. The method was controlled and adjusted with two aqueous solutions containing urea at 20 and 60 mg/dl. 3.3. Instrumental Measurements All the samples were analyzed in duplicates on the ChemSpec 150 at an analytical speed of 150 samples per hour in the differential mode. The ChemSpec 150 was calibrated with a single milk standard provided by the manufacturer. The reference value of the standard was 31.75 mg/dl and was measured using the pH Differential method. The calibration was performed automatically by a simple linear regression forced through zero. 4. Repeatability Evaluation 4.1. Results The method repeatability was evaluated with 150 individual ( 15 were spiked with 30 mg/dl urea) and 50 bulk milk samples analyzed in duplicates in normal routine operating conditions. The samples were analyzed by batch of 20 samples as follows : Series 1 rep1, series 2, rep 2,..., series n, rep 1, series n, rep 2 . The repeatability was calculated according to the International IDF Standard 128: 1985 (9). Table 1. Repeatability Results Individual Milk Ind.Milk + spiked samples Bulk Milk n 135 150 (148) 50(49) Mean 16.16 19.44(19.05) 22.16 Sx 8.39 12.83(12.46) 7.04 Sr 0.46 0.69(0.58) 0.64(0.60) r 1.27 1.90(1.61) 1.78(1.51) With n : number of samples, Sx : standard deviation of the concentration range; Sr: standard deviation of repeatability, r : maximum deviation between duplicates, () results after removing outliers. 4.2. Discussion Two spiked samples were considered as outliers and removed from the data set. The Sr on the spiked samples was higher than the Sr obtained on the other samples. This may indicate that the enzyme was losing some sensitivity. The method gave a Sr between 0.46 and 0.60 mg/dl (Table.1) depending on the type of milk analyzed. The Sr on the reference method was slightly lower with a Sr of 0.49 mg/dl. 5. Accuracy Evaluation 5.1 Results The method accuracy was evaluated with 102 individual milk samples among the 135 collected and 50 bulk milk samples. The analysis were done using the ChemSpec standard configuration. Tab 2. Accuracy results summary. Fig 2. Reference versus Predicted values on 98 individual milk samples Ind. Milk Bulk Milk 70 60 n 102 (98) 49 Predicted ChemSpec Values [mg/dl] 50 Y=ax +b 1.060X+0.925 1.024X+1.389 40 (1.067X+0.728 ) 30 20 Sx-y 2.68(2.26) 1.585 10 Sy,x 2.59(2.11) 1.595 0 With n : number of samples, Sx-y : standard deviation of 0 10 20 30 40 AFNOR Reference [mg/dl] 50 60 70 accuracy before regression; Sy,x : standard deviation of accuracy after regression; () results after removing outliers. 5.2. Discussion Four samples were considered as outliers and were removed from the data set although the determination on these four samples was precise with a Sr of 0.46 mg./dl . No clear reasons have been found to explain these outliers. A mean bias after regression of 0.728 and 1.389 mg/dl [Tab. 2] was respectively observed on the individual milk and bulk milk samples between the AFNOR and ChemSpec methods. The reference value for the Bentley standard used to calibrate the ChemSpec was obtained with the pH Differential method selected as reference method by the International Dairy Federation. This could explain the bias between the AFNOR and ChemSpec methods. This bias could be easily accounted for by analyzing the Bentley standard with the method routinely used by the laboratory and calibrate with this new reference. The standard deviation of accuracy (Sy,x), was between 2.11 and 1.59 mg/dl respectively for the individual and bulk milk samples. Conclusions This evaluation shows that the ChemSpec method compared to the AFNOR NF V04-217 standard has a standard deviation of accuracy (Sy,x) between 1.59 and 2.11 mg/dl and a repeatability (Sr) between 0.46 and 0.60 mg/dl depending of the type of samples analyzed. This represents respectively an accuracy improvement of 58% and 43% over the filter and FTIR infrared analyzers. Therefore the ChemSpec enzymatic method, with an analytical speed of 150 samples per hour, could be used to obtain an improved urea determination in routine analysis compared to the infrared determination. Acknowledgements The author would like to thank the Center for Research and Control of Analysis in the Dairy Industry [CecaLait] for organizing this study and Mr. Philippe Trossat for carrying out this evaluation and supplying the data. References (1) J.SA. Jonker, RE. Kohn, and R. A. Erdman. - Using Milk Urea Nitrogen to predict Nitrogen Excretion and Utilization Efficiency in Lactating Dairy Cows. 1998 J. Dairy Science 81:2681-2692. (2) Gustafsson & Carlsson, 1993. Effect of silage quality, protein evaluation systems and milk urea content on milk yield and reproduction in dairy cows. (3) P.W. Hansen, Urea determination in milk using Fourier Transform infrared spectroscopy and multivariate calibration. (4) O. Leray , Ph. Trossat , 1996. La lettre de CecaLait N/21 Evaluation : Le MilkoScan 4000. (5) D. Lefier SRTAL-INRA, Poligny France, analytical methods for determination of urea in milk. (6) Ph. Trossat , 2000. La lettre de CecaLait N/32 Evaluation : Le ChemSpec 150. (7) Intérêt des dosages d’urée dans le lait pour l’élevage et la transformation. La Lettre de CecaLait, 1995, N/15. (8) AFNOR Standard NF V04-217 : Determination of ammoniac and urea content in milk. (9) International IDF Standard 128: 1985 : Definition and evaluation of the overall accuracy of indirect methods of milk analysis.
Pages to are hidden for
"ICAR_02"Please download to view full document