Comparison of Various Methods for Estimating Body Fat in Dogs
Obesity is considered one of the most common forms of malnutrition occurring in dogs. Laboratory methods of evaluation of body composition in live dogs have included dual-energy X-ray absorptiometry (DEXA) and deuterium oxide (D2O) dilution. Clinical methods of evaluation include assigning a body condition score (BCS) based on visual observation, palpation, and morphometric measurements. This study used these four methods to evaluate 23 healthy, adult, client-owned dogs. Good correlation (coefficient of determination [r2]=0.78) was found between measurements of percent body fat (%BF) determined by the D2O dilution method and the DEXA scan. Percent body fat can also be estimated using BCS (r2=0.92 comparison with DEXA) or by using morphometric measurements with simple calculations (r2=0.92 comparison with DEXA).
Introduction
Obesity is the most common form of malnutrition occurring in dogs and is defined as body weight exceeding optimum by 15% to 20% owing to excessive adipose tissue.1 Dogs having 20% body fat above optimum are by definition obese, but often they are not described as such during routine physical examination. It has been estimated that the prevalence of overweight and obese dogs is 24% to 30%.1 Health risks associated with or exacerbated by obesity include traumatic and degenerative orthopedic disorders; cardiovascular disease; decreased tolerance to heat and exercise; predisposition to diabetes mellitus, hypertension, and hyperlipidemia; and compromised immune function.1
Physical examination, visual observation, and palpation may be used to assign a body condition score (BCS).2 A BCS is a semiquantitative assessment of body composition with a range of categories from cachectic to severely obese. A nine-point system has been previously correlated with assessment of body composition and percent body fat (%BF) as determined by dual-energy X-ray absorptiometry (DEXA) in dogs.3
Anthropomorphic measurements such as skin-fold thickness over the triceps area have been used in humans to estimate body composition and %BF. Morphometric measurements have also been used in dogs and cats, but little has been published comparing objective body measurements with body composition. Dogs deposit and store fat subcutaneously in various locations, including the thoracic, lumbar, and coccygeal areas as well as intra-abdominally.1 With weight gain, the dimension that changes the most is pelvic circumference (PC),1 and specific measurements of PC can be used to estimate %BF.
Dual-energy X-ray absorptiometry has been used to estimate body composition and is a noninvasive method for estimating body fat content. It utilizes X rays at two different energy levels (70 and 140 kilovolts peak [kVp]) to differentiate the type and amount of each tissue in the part of the body being scanned. Based on the exponential attenuation of the different energy levels of X rays emitted, it differentiates body tissues into bone (mineral), lean, and soft tissue mass. The accuracy of DEXA has been validated in pigs, poultry, cats, dogs, and rodents by cadaver analysis.4–10 Dual-energy X-ray absorptiometry has been shown to correlate with indirect estimates of body fat using deuterium oxide (D2O) dilution in healthy laboratory dogs.11
The D2O dilution technique is based on the fact that body water is predominantly associated with nonfat tissue; therefore, a measure of total body water provides an indirect measure of fat-free (i.e., lean body) mass. Deuterium oxide is a stable, nontoxic tracer that is freely exchangeable with water. It is given intravenously (IV) and can be used as a noninvasive method to estimate body composition. Fat-free mass, as determined by D2O dilution technique, has been correlated with body fat content estimated by carcass analysis in dogs.12
The objectives of this study were to compare estimates of body fat determined by DEXA and by D2O dilution, and to compare estimates of body fat determined by DEXA with assigned BCS and with objective body measurements in otherwise healthy, client-owned dogs.
Materials and Methods
All animal procedures were approved by the University of Tennessee Institutional Animal Care and Use Committee. Twelve male and 11 female, adult, client-owned dogs of various breeds were evaluated. Age ranged from 2 to 14 years (mean, 6.2 years±standard deviation [SD] 3.1); body weight ranged between 3.82 and 39.77 kg (mean, 18.51 kg±SD 10.64) [Table 1].
Subjective Body Measurements
One investigator (Bartges), using a previously described nine-point scale,2 assigned a BCS to each dog using visual inspection and palpation. The scoring system ranges from 1 (i.e., emaciated ) to 9 (i.e., grossly obese).
Objective Body Measurements
Anthropometric measurements in cm using a simple tape measure were performed at the following sites:1 height at shoulder; length from midpoint between the cranial scapulae to the base of the tail; length from the occipital protuberance to the base of the tail; PC measured around the level of the flank; and length from tuber calcaneus to midpatellar ligament (i.e., hock-to-stifle length [HS]). This allowed the following anthropometric values to be calculated:1
-
Body mass index (BMI) = Body weight (BW)kg/(height at shouldercm × length from occipital protuberance to base of tailcm)
-
Body fat was calculated according to gender-specific formulas:1
\(Males\ (\%BF)\ =\ {-}1.4\ (HS_{cm})\ +\ 0.77\ (PC_{cm})\ +\ 4\)
\(Females\ (\%BF)\ =\ {-}1.7\ (HS_{cm})\ +\ 0.93\ (PC_{cm})\ +\ 5\)
\(Either\ gender\ (\%BF)\ =\ ({-}0.0034\ [HS_{cm}]^{2}\ +\ 0.0027\ [PC_{cm}]^{2}\ {-}\ 1.9)/BW_{kg}\)
Total body water content was estimated by measuring D2O isotope dilution at equilibrium after IV administration of D2O. Dogs were weighed, and blood was collected from a jugular vein into a heparinized vacuum tube. Sterilized D2O was administered into a cephalic vein via a 19-gauge catheter at a dosage of 0.275 g/BWkg. Syringes were weighed before and after administration of D2O to determine exact dose administered. The IV catheter was flushed with 10 mL of sterile 0.9% saline solution to ensure that the entire dose was delivered. Two hours after D2O administration, dogs were reweighed, and blood was collected into a heparinized vacuum tube from a jugular vein. Plasma was separated and stored at −30°C until analysis by nuclear magnetic resonance (NMR). No side effects or complications occurred with this procedure.
The following method was performed to determine D2O content.11 Plasma samples were thawed to room temperature (20° to 22°C) and analyzed for D2O at 25°C using a NMR spectrometer operating at field strength of 7.05 Tesla as previously described.11
Total body water was calculated as:
\(Total\ body\ H_{2}O\ (g)\ =\ (\mathit{g\ D}_{2}\mathit{O\ injected})\ {-}\ (\mathit{m}_{\mathit{0}}{-}\mathit{m}_{\mathit{1}})(\mathit{D}_{\mathit{1}}{-}\mathit{D}_{\mathit{0}}\mathit{/100))(0.985)(18/20)}\ {\div}\ (D_{1}{-}D_{0})/100\)where m0 is the dog’s body weight immediately before D2O; m1 is the dog’s weight at the time of sample collection; D1 is atom % D2O in the plasma sample obtained after D2O administration and equilibrium of D2O; D0 is atom % D2O in blood plasma before the dose was administered (D1 and D0 were obtained from NMR measurements in actual plasma samples and subsequent comparison with standard curve constructed from reference D2O plasma samples); 0.985 is a correction for incorporation of deuterium into nonexchangeable organic constituents; and 18/20 is the correction factor for the difference in molecular weight between D2O and H2O. Assuming fat-free mass to contain 73.2% moisture, %BF was calculated from measured total body water according to the following formula:
\(\%BF\ =\ 100\ {-}\ \%total\ body\ H_{2}O/0.732\)Dual-energy X-ray absorptiometrya scans were performed immediately following blood collection from dogs. General anesthesia was induced using acepromazine (0.02 mg/kg IV) and torbugesic (0.2 mg/kg IV) followed by inhalation anesthetic (i.e., isoflurane and oxygen) using an anesthetic mask. Dogs were maintained in dorsal recumbency. Scanning was done in the whole-body mode, which required approximately 20 minutes. Individual scans were analyzed using pediatric software designed for use with the DEXA instrument. Reproducibility of this technique in the authors’ laboratory has yielded a coefficient of variance of 2.35% for %BF assessment.
Statistical Analysis
Simple linear regression analysis was used to determine correlation between estimates of %BF determined by DEXA and %BF determined by D2O, BCS, and morphometric analysis.
Results
Dogs ranged in BCS from 4 to 8 and ranged in %BF, as determined by DEXA, from 4.00% to 40.80% [Table 1]. No correlation was found between body weight and age or between body weight and BCS [Table 2]. Poor correlation was found between body weight and %BF as determined by DEXA or by D2O [Table 2].
Percent Body Fat Determined by D2O Versus DEXA
Regression analysis of %BF by D2O or by DEXA indicated close relative agreement, with a slope of 0.823 and a coefficient of determination (r2) of 0.78 [Figure 1]. However, the intercept differed significantly (P<0.01) from 0, indicating a lack of absolute agreement.
Body Condition Score Versus %BF by DEXA
Correlation between BCS and %BF by DEXA was significant (r2=0.92, P<0.001) [Figure 2]. In this study, ideal body condition (BCS=5) was associated with 11%±2%BF. Each unit increase in BCS was associated with an increase in fat of approximately 8.7%.
Morphometric Measurements Versus %BF by DEXA
The correlation between BMI and %BF by DEXA was significant (r2=0.54, P<0.001); however, the intercept (43) differed significantly from 0 (P<0.001), indicating lack of absolute agreement [Table 2; Figure 3A]. Correlation was significant between %BF determined using morphometric measurements with either-gender or sex-specific formulae when compared with %BF by DEXA (r2=0.92, P<0.001; r2=0.90, P<0.001, respectively) [Table 2; Figures 3B, 3C]. Intercepts for these regression equations did not differ statistically from 0, indicating agreement.
Discussion
Accurate analysis of canine body composition is a challenge clinically. The gold standard for body composition analysis is chemical analyses of tissues from euthanized dogs, which is not feasible in the clinical setting. Both DEXA and D2O dilution techniques have been validated in dogs using chemical carcass analyses.912 Lauten, et al. confirmed a high degree of precision with DEXA in dogs and that no significant differences exist between results obtained via DEXA and chemical analysis for %BF and percent lean mass.9
The theory behind the D2O dilution technique is that body water content is estimated using D2O as a tracer that will equilibrate with body water. A known quantity of D2O is administered and distributes uniformly throughout all pools of body water. A sample of one of these pools can then be analyzed to determine the total volume of water in which the tracer has been dispersed. Time for equilibrium and tracer lost through respiration and urine has been questioned. Burkholder, et al. showed that the stable isotope concentration was reached at 90 minutes and that there was negligible loss of tracer in urine.12 Unmeasured isotope loss occurs in most studies, as shown by overestimation of absolute body water. One explanation is nonaqueous hydrogen exchange, which can contribute 2% to 5% of the overestimated water; however, this does not account for all of the overestimation.12 Another loss may be through pulmonary respiration. Given IV, the isotope passes through the lungs immediately, and loss of water with associated isotope can occur. This loss has not been measured in dogs. Other unknown losses may also occur.
In this study, a good relative agreement was seen between %BF by DEXA and %BF by D2O dilution. Estimation of %BF by DEXA averaged 13% higher than by D2O dilution. These results are similar to a previous study.11 The D2O dilution method provided estimates for %BF that were <0, suggesting that total body fat was underestimated. Son, et al. suggest using a different equation based on results from estimation of %BF by D2O dilution and carcass analysis; however, this would not change the results by >1%.11 In this study, the use of this alternative equation may have changed the estimated %BF by D2O dilution to a value closer to that obtained by DEXA. This equation was not used in the current study, because correlation between the D2O dilution method and DEXA method would not have changed; however, the intercept for the equation for D2O dilution would have been closer to 0.
Use of D2O analysis for body fat estimation is a difficult and time-consuming procedure, and it is not a practical clinical test. Correlation between %BF by D2O dilution and %BF by DEXA allows the clinician to use DEXA as a practical clinical or research tool with low radiation exposure and without patient discomfort. Dual-energy X-ray absorptiometry has been used traditionally to assess body composition with regard to bone and soft tissue. Work has been done recently to further divide analysis of soft tissue into fat and bone-free lean tissue, which allows estimation of %BF. With the increased use of DEXA in human medicine, more veterinary institutions will likely have access to this technique in the future.
Since most veterinary clinics do not own or have access to DEXA at present, more practical methods of determining %BF are to assign a BCS and to estimate %BF from morphometric measurements. In the current study, correlation between BCS and %BF by DEXA was good. The ideal BCS of 5 was associated with 11%BF in the current study. Each unit increase in BCS was associated with an 8.7% increase in %BF. These results are similar to a study done by Laflamme, et al., in which correlation of BCS to %BF by DEXA was 0.9 (P<0.01), and the ideal BCS was associated with 19%±8%BF.3 Each unit increase in BCS was associated with a 5%BF increase.
The present study used one person to subjectively measure BCS at the time of DEXA scan and D2O dilution. It might have been useful to average multiple measurements or to have multiple people scoring each animal, which may have resulted in a more rigorous subjective examination. In a clinical setting, often only one doctor is examining an animal and assessing the BCS; therefore, this study may be a truer reflection of clinical judgment and application.
Morphometric measurements are easy to perform with the dog standing squarely and looking straight ahead. Good correlation existed between measurements made from the “sex-specific” or “either-gender” %BF calculations and the %BF determined by DEXA. These calculations used height at shoulder and PC as the major determinants. Correlation with %BF by DEXA was less when the BMI was used; therefore, the BMI calculation does not appear to be as useful. Since PC is the measurement that changes most with weight gain in the dog, using this site in the calculation is recommended; however, because of the variety of body proportions in different dog breeds, these measurements may not always correlate with %BF. Further studies comparing dogs of differing conformation, such as the greyhound and English bulldog, are warranted.
Conclusion
Good correlation has been found between measurements of %BF obtained by D2O dilution and by DEXA scan. Percent body fat can be estimated clinically using morphometric measurements based on simple calculations or using a nine-point BCS system. These measurements correlate well with the %BF determined by the DEXA method. Further clinical studies are required to compare %BF calculated from morphometric measurements among different dog breeds.
Lunar model DPX; Lunar Corp., Madison, WI
Citation: Journal of the American Animal Hospital Association 40, 2; 10.5326/0400109
Citation: Journal of the American Animal Hospital Association 40, 2; 10.5326/0400109
Citation: Journal of the American Animal Hospital Association 40, 2; 10.5326/0400109
Correlation between estimated percentage of body fat (%BF) determined by deuterium oxide (D2O) and estimated %BF determined by dual-energy X-ray absorptiometry (DEXA).
Correlation between body condition score (BCS) and estimated percentage of body fat (%BF) determined by dual-energy X-ray absorptiometry (DEXA).
Correlation of estimated percentage of body fat (%BF) determined by dual-energy X-ray absorptiometry (DEXA) and (A) body mass index (BMI); (B) estimated %BF using either-sex formula (Either); and (C) estimated %BF using sex-specific formula (Sex specific).
