PROCEDURAL AND MATHEMATICAL CONSIDERATIONS IN UREA DILUTION ESTIMATION OF BODY COMPOSITION IN LAMBS 1

Transcription

1 PROCEDURAL AND MATHEMATICAL CONSIDERATIONS IN UREA DILUTION ESTIMATION OF BODY COMPOSITION IN LAMBS 1 S. J. Bartle 2, O. A. Turgeon, Jr. 3, R. L. Preston 2 and D. R. Brink 4 Texas Tech University, Lubbock and University of Nebraska, Lincoln ABSTRACT Two experiments (n 46 and 56, respectively) were conducted to evaluate urea dilution as an estimator of body composition in lambs and to address certain procedural and mathematical considerations in this technique. In Exp. 1, 14 blood samples were taken over 240 rain after urea infusion. The equation describing the urea clearance curve was:apun 9.7e -'1727(min) e -'O021(min), pools 1 and 2, respectively (r 2.99, P <.001;individual lamb effects removed). In the combined experiments, urea space (US) was related to percentage of empty body water (PEBH20) by the equation US (empty body weight basis; r 2.56, P <.001). The regression equation indicates that the US-PEBH20 relationship in lambs is different from that reported in cattle, even though urea clearance kinetics are similar. Although the prediction equations appeared to be biologically valid, considerable error was associated with the composition estimates. The PEBHz O was predicted as well by live weight (r 2.69; SEy 9 x 3.0) as by US in these experiments. The twosample method (T12 minus TO) to determine the change in marker concentration was shown to be related more closely (r 2.56) to PEBH20 than the standard muhisample extrapolation to TO method (r z.0 and.38 for pools 1 and 2, respectively). An equilibration time of 9 to 12 rain provided the best estimate of body composition in lambs. The results demonstrate the importance of accurate sample timing and precise plasma urea-n determination and indicate that the US procedure may require modification for use in lambs. (Key Words: Urea, Dilution, Lambs, Body Composition, Techniques.) Introduction Urea space is proportional to empty body water and thereby is a useful estimator of body composition in cattle (Hammond et al., 1984; Rule et al., 1986; Bartle et al., 1987). The regression equation relating urea space to empty body fat is similar to the known relationship between empty body water and fat after the underestimation of empty body water by 1 Approved jointly for publication by the Dean of Agric. Sci., Texas Tech Univ., as publication No. T and Agric. Res. Div., Univ. of Nebraska, as Journal Series No Funds were provided in part by the Thornton Endowment, Texas Tecb Univ. 2Dept. of Anirn. Sci., Texas Tech Univ. 3Current address: Koers Consulting Service, Inc., Salina, KS Dept. of Anim. Sci., Univ. of Nebraska. Received October 2, Accepted February 11, urea dilution is considered (Bartle et al., 1987). Urea dilution, however, accounts for only 60 to 70% of the variation in body components in most reports. Because the regression equations appear to be biologically valid, based on the relationships among body components, the lack of a higher coefficient of determination suggests that factors other than the urea space-body composition relationship contribute to the variation. Several physiological and procedural factors can introduce variation into this relationship. Individuals may vary in rate of urea transfer to the rumen, though this is not significant in fasted cattle (Bartle and Preston, 1986), in renal urea clearance rate or in percentage of gut fill. Inaccurate determination of body composition or live weight also can contribute to variation, as can procedural errors such as inaccurate blood sampling times or imprecise laboratory determination of plasma urea-n. The objectives of these experiments were to evaluate urea dilution as an estimator 1920 J. Anita. Sci :

2 UREA DILUTION IN LAMBS 1921 of body composition in lambs, to establish the urea clearance curve in lambs and to address certain procedural questions and possible errors in urea dilution estimation of body composition. Procedures Experiment I. Twenty ewe and 30 wether Rambouillet x Suffolk lambs (initial wt kg) were blocked by sex, stratified by weight and randomly assigned to four slaughter groups (initial, 30, 38 and 45 kg live wt). The initial slaughter group contained five lambs; later slaughter groups consisted of 15 lambs. Lambs were grown from 20 to 30 kg at three rates (slow, medium and rapid; see Turgeon et al. [1986] for a complete description of the growing conditions). Diets for lambs at the three growth rates contained 70, 50 and 30% roughage, respectively. Lambs remaining after the 30-kg slaughter were grown on the 30% roughage diet to their respective slaughter weights. Slaughter weight was the average of four weights (not shrunk) taken on consecutive days before slaughter. Urea dilution was determined 1 d before slaughter by infusing a saline solution containing 10% urea (.13 g urea/kg LW) as described by Preston and Kock (1973). Urea was infused and samples were collected (after thorough flushing) through a jugular catheter. Blood samples were collected before infusion (TO) and at 3, 6, 9, 12, 15, 18, 21, 25, 30, 35, 40, 60, 120 and 240 min after infusion. Plasma was collected and stored at -20 ~ C for plasma urea-n (PUN) determination using an autoanalyzer (Marsh et al., 1965; Fingerhut et al. 1966). Urea space (US) was calculated on both a live weight (LW) and an empty body weight basis (EBW) as described by Bartle et al. (1987): urea space mg urea-n infused/(change in PUN x LW x 10), where PUN is expressed as mg urea-n/100 ml. Change in PUN was calculated as T12 minus TO. Different equilibration times were calculated as T6, T9 or T15 minus TO. Empty body chemical composition (water, protein, fat and ash) was determined after whole-body grinding (digesta removed). Gut fill was calculated as the weight difference between the full and emptied gastrointestinal tract. Slaughter and chemical analysis procedures were presented by Turgeon et al. (1986). Experiment 2. Fifty-six lambs from a variety of sources were slaughtered at the Texas Tech Meats Laboratory. All lambs were held off feed and water overnight. Urea dilution was determined in a manner similar to that in Exp. 1, except that a 20% urea solution was infused (.13 g urea/kg LW), only TO and T12 samples were taken, and PUN was determined as described by Chaney and Marbach (1962) and Searle (1984). Lambs were slaughtered immediately after urea dilution. Live weight, weight of the full and emptied gastrointestinal tract and carcass weight were determined. After chilling overnight, carcass specific gravity was determined, and empty body composition was calculated as described by Garrett et al. (1959). The regression equation describing urea clearance in Exp. 1 was determined using the nonlinear regression (NLIN) procedure (SAS, 1985). Variable means within experiments and regression equations relating urea space to body components for Exp. 1 and 2 and the combined results were computed and evaluated using MEANS and regression (REG)procedures, respectively (SAS, 1985). Urea space measurements were the independent variable in the REG procedure. Results and Discussion Means and ranges of LW, EBW, percentage gut fill, US and percentage empty body fat (PEBF) and water (PEBH20) are shown in Table 1. Four lambs in Exp. 1 were not included in the analysis because of incomplete infusion or missed blood samples. In Exp. 1, lambs (n 46) averaged 36 kg LW and ranged from 8 to 25.4 PEBF. Lambs in Exp. 2 averaged 50.9 kg and ranged from 7.5 to 32.5 PEBF. Body composition and weight of experimental lambs were representative of the industry. A 12-min equilibration time was achieved in all lambs. The maximum coefficient of variation of replicate PUN determinations was 2.0% in both experiments. Body Composition Estimation. Equations relating US to PEBF and PEBH20 are shown in Table 2. Similar relationships were found in the two experiments, although coefficients of determination tended to be greater (LW basis) when lambs were shrunk (Exp. 2) than when lambs were fed diverse diets and not shrunk prior to US determination (Exp. 1). The low coefficient of determination in the LW-based relationship (Exp. 1) may have been due to the large range in percentage gut fill, which influenced LW. Lambs in the 30-kg slaughter group fed 70, 50 and 30% roughage diets averaged 29.5, 23.0 and 19.8% gut fill, respec-

3 1922 BARTLE ET AL. TABLE 1. MEANS AND RANGES OF LIVE WEIGHT (LW), EMPTY BODY WEIGHT (EBW), PERCENTAGE GUT FILL, UREA SPACE, EMPTY BODY FAT AND EMPTY BODY WATER Exp. 1 Exp. 2 Range Range Item Mean a SD b Low High Mean c SD b Low High Live wt, kg Empty body wt, kg Gut fill, % of LW Urea space (LW), % Urea space (EBW), % Empty body fat, % , Empty body water, % an 46. bsd standard deviation. CN 56. TABLE 2. EQUATIONS RELATING UREA SPACE TO PERCENTAGE EMPTY BODY WATER (PEBH20) AND EMPTY BODY FAT (PEBF) Empty body component Equation r 2 Sy.x Water, % Exp. 1 a PEBH20 Exp. 2 f All lambs Cattleg Fat, % Exp. 1 a PEBF Exp. 2 f All lambs Cattleg US b,lw c US. EBW e US LW US. EBW US LW US. EBW US. LW US. EBW.30 d US, LW US, EBW US, LW US, EBW US, LW US, EBW US, LW I3S, EBW.75 an 46; composition determined chemically. bus urea space. clw live weight. dp <.001 for all r 2. eebw empty body weight. fn 56; composition calculated from carcass specific gravity (Garrett et al., 1959). gbartle et al. (1987).

4 UREA DILUTION IN LAMBS 1923 tively. Urea space tended to overestimate the PEBH20 in lambs fed 70 and 50% roughage diets. Omitting these lambs from the relationship improved the coefficient of determination for estimation of PEBH20 from.30 to.52 (LW basis). These differences illustrate the problem in estimating body composition when EBW is not a constant proportion of LW. Currently, no means are available to deal with differences in gut fill in the US procedure other than managing the animals to achieve uniform fill. Holding ruminants off feed before urea dilution may minimize area transfer to the rumen (Bartle and Preston, 1986). Mean US (EBW) and PEBH20 values were similar (Table 1), but the range in US was 12.9 and 14.9 percentage units greater than the range in PEBH20 for Exp. 1 and 2, respectively. The greater range in US resulted in less than one percentage unit change in PEBH20/percentage unit change in US (slope.47; Table 2), which is considerably different from the regression equation in cattle (slope.81; Table 2). The intercepts of the regression equations in lambs and cattle reflect the less than 1:1 relationship between US and PEBH20 (31.7 vs 13.0%, respectively). The intercepts can be interpreted to indicate an underestimation of PEBH20 by US. The reason for the substantial difference between species in the relationship between US and PEBH20 is not clear. The equations, however, yield similar results within the physiological range of PEBH20. For example, solving both equations for 50.4% US (average US in this study) results in PEBH20 estimates of 53.8 and 55.4% for cattle and lamb equations, respectively. The percentage relationship between PEBH20 and PEBF in Exp. 1 lambs (Table 3) was similar to that observed by Garrett et al. (1959) in lambs and Arnold et al. (1985) in cattle. As in cattle, regression equations relating US and PEBF appear to be valid, if the intercept of the PEBH2 O-US equation is considered. The US-PEBH20 intercept (Table 2)indicates that US underestimates PEBH20 by 31.7%, which is consistent with the underestimation of PEBF at zero US (34.2%; %; 87.9% the intercept of the PEBF-PEBHzO equation, Table 3). Table 4 shows the relationships between empty body water and LW in this study and similar regression equations in cattle. The PEBH20 was predicted as closely by US as by LW in these experiments. The similar body component-lw equations between experiments indicate that uniform lambs were used. Diverse populations of cattle have been shown to have different body component-lw relationships but similar body component-us relationships (Bartle et al., 1987). Mathematical Considerations. Figure 1 shows the urea clearance curve (individual lamb effects removed) in Exp, 1. Clearance was best described by a two-pool model (Shipley and Clark, 1972): APUN 9.7e -'1727(min) e --O021(min) (n 46, r 2.99, Sy 9 x.09) which is similar to the equation for cattle (S. J. Bartle, unpublished data): APUN 15.2e -'1572(min) e -.-O020(min) (n 4, r 2.99, Sy 9 x 1.25) The intercepts indicate that an equivalent quantity of urea (LW basis) causes a greater TABLE 3. EQUATIONS RELATING EMPTY BODY FAT AND EMPTY BODY WATER IN EXP. 1 AND SIMILAR LITERATURE EQUATIONS Author Equation r 2 This paper, Exp. 1 Lambs (Garrett et al., 1959) Cattle (Arnold et al., 1985) apebf percentage empty body fat. bpebh20 percentage empty body water. Cp <.001 for all r 2. PEBF a PEBH20 b PEBF PEBH 20 PEBF PEBH20.96 c.99.98

5 1924 BARTLE ET AL. TABLE 4. EQUATIONS RELATING LIVE WEIGHT (LW) AND EMPTY BODY WATER Basis Equation r 2 Sy.x Mass Exp. 1 EBH20 a LW.95 b.8 Exp LW All lambs LW Cattle c LW Percentage Exp. 1 PEBH20 d LW Exp LW All lambs LW Cattle c LW aebh20 empty body water, kg. bp <.001 for all r 2. CAverage equation, calculated from Table 4, Bartle et al. (1987). dpebh20 percentage empty body water. 16. Urea Clearance in Lambs PUN 9.7e ~ n 46, r.99, Sy.x e "'0021(min) min ,. I ~ 8- t 10.4e "'O021(min), pool 2 t Z t a. t t 4- I 4~---9.7e "'1727(min), pool 1 % % %t 1'2 2'0 4'0 do Time after infusion, min Figure 1. Change in plasma urea-n (PUN) after infusion with.13 g urea/kg live weight.

6 UREA DILUTION IN LAMBS 1925 PUN increase in cattle than in lambs. This difference may be one reason for the variation between species in prediction equations and suggests that more urea (per kg LW) should be infused in lambs. The concentration of the infusion solution, however, should not exceed 20%, because infusion solutions containing greater than 22% urea cause red blood cell hemolysis in vitro (R. L. Preston, unpublished data). Urea dilution differs from most other dilution techniques in that only two samples (TO and T12; Preston and Kock, 1973) vs a series of samples and extrapolation to TO (concentration with instantaneous mixing) are used to determine the change in marker concentration. To compare the two-sample method with the muhisample method, components of the clearance curve (Figure 1) were extrapolated to TO. Observed intercepts for each lamb (I1 and I2, for pools 1 and 2, respectively) were used to calculate urea space (EBW basis) and then related to PEBH~O. The intercept of pool 1 (I1, rapid equilibration pool) was not related to PEBH20 (r 2.0). The intercept of pool 2 (I2; slow equilibration pool) was related to PEBH20 by the equation PEBH (r 2.38, Sy. x-- 3.2, P <.001). Using the extrapolated I2 intercept to determine the change in marker concentration is a standard approach (Shipley and Clark, 1972). In this study, the extrapolated intercept was not as closely related to body composition as the change in marker concentration determined by the two-sample method (r 2.56; Table 2). The slope of the second pool appears to be largely renal clearance of urea (Bartle and Preston, 1986), which is not known to be related to body composition. Table 5 shows the relationships of PEBH20 and US (Exp. 1) calculated after 6, 9, 12 and 15 min equilibration time. The equations show that best estimates are found 9 to 12 min after infusion, similar to cattle results (Preston and Kock, 1973; Kock and Preston, 1979; Meissner et al., 1980). The consistent equilibration time is supported by the uniform rate constants in the lamb and cattle urea clearance curves. Procedural Errors. One procedural error that often is unavoidable is inaccurate timing of the second (12-min) blood sample. The urea clearance curve indicates that a 1-min error in the 12-rain sampling time would result in an approximate 2.0% error in the US estimate. Because urea clearance is not linear, this error will not be consistent over greater timing differences. Expected deviation in PUN for equilibration times other than 12 min can be calculated from the urea clearance equation. A second procedural error is imprecise determination of PUN. Because the change in marker concentration is the difference between two PUN values, any lack of precision may be compounded and can result in substantial variation in the body composition estimate. As shown in the example presented in Table 6, a 5% coefficient of variation (CV) in replicate PUN determinations can cause US (EBW basis) to vary from 60 to 81%. Even a 2% CV can cause US to vary from 65 to 74%. These examples demonstrate the importance of precise PUN determination. If PUN is 10 mg urea-n/100 ml or greater, a goal of < 1.5% CV is reasonable. Also, if it is necessary to TABLE 5. EFFECT OF EQUILIBRATION TIME ON THE ESTIMATION OF PERCENTAGE EMPTY BODY WATER (PEBH20) IN EXP. 1 (N 46) a Equilibration, time, min Equation r 2 Sy,x 6 PEBH US b.46 c US US US aempty body weight basis. bus urea space calculated after various equilibration times. Cp <.001 for all r 2.

7 1926 BARTLE ET AL. TABLE 6. EXAMPLES OF THE EFFECT OF PLASMA UREA NITROGEN (PUN) DETERMINATION ERRORS I. Variables used in examples: TO 10 mg urea-n/100 ml T12 20rag Urea-N dose 1,966 mg Empty body wt 28.4 kg II. Effect of a 5% coefficient of variation (CV) in PUN determination: TO range 9.5 to 10.5 mg urea-n/100 ml T12 range 19.0 to 21.0 Minimum APUN 8.5; EBUS a -- 81% Maximum APUN 11.5; EBUS 60% Ill. Effect of a 2% CV in PUN determination: TO range 9.8 to 10.2 mg urea-n/100 ml T12 range 19.6 to 20.4 Minimum APUN 9.4; EBUS 74% Maximum APUN 10.6; EBUS 65% aebus urea space, empty body weight basis. repeat an individual PUN determination, both TO and T12 should be repeated to eliminate run-to-run variation. In conclusion, urea dilution was related to body composition in lambs. Considerable variation was noted, however, and the proportion of variation in body composition accounted for by urea space was not satisfactory. The US-PEBH 20 regression equation was found to be different from that reported in cattle, suggesting that the urea dilution procedure may require modification for use in species other than cattle. The urea clearance curves indicate that infusing a greater amount of urea/kg LW may improve the relationship in lambs. The two-sample method was more closely related to body composition than the multisample extrap- olation to zero time method. Several possible sources of error were discussed and suggestions were made to minimize their effect. Literature Cited Arnold, R. N., E. J. Hentges and A. Trenkle Evaluation of the use of deuterium oxide dilution for determination of body composition in steers. J. Anita. Sol. 60:1188. Bartle, S. J., S. W. Kock, R. L. Preston, T. L. Wheeler and G. W. Davis Validation of urea dilution to estimate in vivo body composition in cattle. J. Anita. Sci. 64:1024. Bartle, S. J. and R. L. Preston Plasma, rumen and urine pools in urea dilution estimation of body composition. J. Anim. Sei. 63:77. Chaney, A. L. and E. P. Marbach Modified reagents for determination of urea and ammonia. Clin. Chem. 8:130. Fingerhut, B., R. Ferzola, W. H. Marsh and A. B. Miller Automated methods for blood glucose and urea with adaptation for simultaneous determination. Clin. Chem. 12: 570. Garrett, W. N., J. H. Meyer and G. P. Lofgreen The comparative energy requirements of sheep and cattle for maintenance and growth. J. Anita. Sci. 18:528. Hammond, A. C., T. S. Rumsey and G. L. Haaland Estimation of empty body water in steers by urea dilution. Growth 48:29. Kock, S. W. and R. L. Preston Estimation of bovine carcass composition by the urea dilution technique. J. Anita. Sci. 48:319. Marsh, W. H., B. Fingerhut and H. Miller Automated and manual direct methods for the determination of blood urea. Clin. Chem. 11:624. Meissner, H. H., J. H. van Staden and E. Pretorius In vivo estimation of body composition in cattle with tritium and urea dilution. 1. Accuracy of prediction equations for the whole body. S. Aft. J. Anim. Sci. 10:165. Preston, R. L. and S. W. Kock In vivo prediction of body composition in cattle from urea space measurements. Proc. Soc. Exp. Biol. Med. 143: Rule, D. C., R. N. Arnold, E. J. Hentges and D. C. Beitz Evaluation of urea dilution as a technique for estimating body composition of beef steers in vivo: Validation of published equations and comparison with chemical composition. J. Anim. Sci. 63:1935. SAS SAS User's Guide: Statistics. SAS Inst., Inc., Cary, NC. Searle, P. L The Berthelot or indophenol

8 UREA DILUTION IN LAMBS 1927 reaction and its use in the analytical chemistry of nitrogen: A review. Analyst 109: 549. Shipley, R. A. and R. E. Clark Compartment analysis: Two-pool open systems. In: Tracer Methods for In Vivo Kinetics. pp Aca- demic Press, New York. Turgeon, O. A., Jr., D. R. Brink, S. J. Bartle, T. J. Klopfenstein and C. L. Ferrell Effect of growth rate and compensatory growth on body composition in lambs. J. Anita. Sci. 63:770.