Evaluating methods to assess the body condition of female ...

BioOne sees sustainable scholarly publishing as an inherently collaborative enterprise connecting authors, nonprofit publishers, academicinstitutions, research libraries, and research funders in the common goal of maximizing access to critical research.

Evaluating methods to assess the body condition of female polar bearsAuthor(s): Anthony M. Pagano, Karyn D. Rode, and Stephen N. AtkinsonSource: Ursus, 28(2):171-181.Published By: International Association for Bear Research and Managementhttps://doi.org/10.2192/URSU-D-16-00029.1URL: http://www.bioone.org/doi/full/10.2192/URSU-D-16-00029.1

BioOne (www.bioone.org) is a nonprofit, online aggregation of core research in the biological,ecological, and environmental sciences. BioOne provides a sustainable online platform for over 170journals and books published by nonprofit societies, associations, museums, institutions, and presses.

Your use of this PDF, the BioOne Web site, and all posted and associated content indicates youracceptance of BioOne’s Terms of Use, available at www.bioone.org/page/terms_of_use.

Usage of BioOne content is strictly limited to personal, educational, and non-commercial use.Commercial inquiries or rights and permissions requests should be directed to the individual publisheras copyright holder.

https://doi.org/10.2192/URSU-D-16-00029.1

http://www.bioone.org/doi/full/10.2192/URSU-D-16-00029.1

http://www.bioone.org

http://www.bioone.org/page/terms_of_use

Evaluating methods to assess the body conditionof female polar bears

Anthony M. Pagano1,2,4, Karyn D. Rode1, and Stephen N. Atkinson3

1U.S. Geological Survey, Alaska Science Center, 4210 University Drive, Anchorage, AK 99508, USA2Department of Ecology & Evolutionary Biology, University of California, Santa Cruz, CA 95060, USA

3Box 19, Group 7, RR2, Dugald, Manitoba, R0E 0K0, Canada

Abstract: An animal’s body condition provides insight into its health, foraging success, and overallfitness. Measures of body composition including proportional fat content are useful indicators ofcondition. Isotopic dilution is a reliable non-destructive method for estimating the body compositionof live mammals, but can require prolonged handling times. Alternatively, bioelectrical impedanceanalysis (BIA) has promise as a quick method for estimating the body composition of live mammals,but measurements can potentially be affected by field conditions. Body condition indices (BCI) andenergy density models can also be used to assess body condition based on morphological measurements,but may not reliably reflect an animal’s energy stores. Here we evaluate BIA, BCI, and an energydensity model in measuring the energy stores of female polar bears (Ursus maritimus). We examine therelationship between total body fat (TBF) derived from isotopic dilution to these alternative methods for9 female polar bears from 14 captures on the sea ice of the southern Beaufort Sea in April 2014–2016.An energy density model, BCI, and BIA-derived measures of TBF were poor predictors of TBF derivedfrom isotopic dilution. We suggest energy density, BCI, and BIA may not be predictive of an animal’sbody fat at fine scales (e.g., among individuals within the same sex, reproductive status, and season).In particular, BIA should provide similar measures of body composition as isotopic dilution, but itfailed to reliably measure TBF of individual bears. These limitations in the precision of body conditionmeasures should be considered when planning future studies.

Key words: bioelectrical impedance analysis, body composition, body condition index, body fat, energy density,isotopic dilution, polar bear, Ursus maritimus

DOI: 10.2192/URSUS-D-16-00029.1 Ursus 28(2):171–181 (2017)

Body condition provides an indication of an animal’senergy stores (Speakman 2001, Schulte-Hostedde et al.2005). Measures of condition are useful to evaluate in-dividual health, foraging success, reproductive potential,and overall fitness as well as to evaluate environmen-tal conditions experienced by the individual (Virgl andMessier 1992, Atkinson and Ramsay 1995, Atkinsonet al. 1996, Winstanley et al. 1999, Speakman 2001,Stevenson and Woods 2006, Molnar et al. 2011). In pop-ulation studies, condition may be one of the first param-eters to exhibit measureable change before reproductionand survival (vital rates) effects are detectable (Derocheret al. 2004, Williams et al. 2013, Obbard et al. 2016).Measures of an animal’s energy stores are of particularimportance for species that store (and use) large amountsof energy in the form of body fat–adipose tissue as part of

4email: [email protected]

a life-history adaptation for dealing with seasonal varia-tion in food availability.

Polar bears (Ursus maritimus) undergo dramaticchanges in storage energy as an adaptation to living in anenvironment where food availability is highly seasonal(Ramsay and Stirling 1988, Stirling and Øritsland 1995).In the late spring and early summer, when access to lipid-rich marine mammal prey such as ringed seals (Pusahispida) increases, polar bears become hyperphagic andtheir adipose tissue deposits undergo extensive hypertro-phy (Ramsay and Stirling 1988). The percent body fatof most individuals peaks by late summer or autumn andvaries according to sex, age class, and reproductive status,with adipose tissue among pregnant females comprisingup to half of their total body mass (Atkinson and Ramsay1995, Atkinson et al. 1996). During other seasons, foodavailability is less predictable and hence storage of bodyfat is more variable (Atkinson and Ramsay 1995, Atkin-son et al. 1996). Superimposed on this annual cycle of

171

172 METHODS TO ASSESS POLAR BEAR BODY CONDITION � Pagano et al.

body fat are the potential effects of declines in Arctic seaice, which in some areas appear to be reducing polar bearforaging success (Regehr et al. 2007, 2010; Cherry et al.2009; Pilfold et al. 2015). Given this changing ecosystem,reliable measures of body condition (in particular, energystores) are needed to evaluate the effects of declines insea ice on polar bear condition, health, and populationviability (e.g., Stirling and Derocher 1993; Stirling et al.1999; Obbard et al. 2006, 2016; Rode et al. 2010, 2012,2014b; Patyk et al. 2015).

Various methods have been used to assess the bodycondition of polar bears, including body condition in-dices (Stirling et al. 1999, Cattet et al. 2002), lipid con-tent of adipose tissue (Thiemann et al. 2006, McKinneyet al. 2014, Sciullo et al. 2016), energy density models(Molnar et al. 2009), and measures of body composi-tion (Arnould and Ramsay 1994, Atkinson and Ramsay1995, Atkinson et al. 1996, Sciullo et al. 2016). Thesemethods differ from one another in the way body condi-tion is conceptualized and in the assumptions associatedwith measuring it (e.g., Cattet 1990, Green 2001, Speak-man 2001, Barthelmess et al. 2006, Schamber et al. 2009,Hwang et al. 2015). In comparison with other methods ofcondition assessment, body composition analyses facili-tate a more direct and absolute evaluation of the amountof stored energy in an animal, especially in species suchas polar bears that are known to store energy in large de-posits of adipose tissue. Whole-body chemical extractionprovides the most accurate measure of body composition(Speakman 2001), but requires sacrificing the animal.

Two methods for empirically measuring the body com-position of live animals are isotopic dilution and bioelec-trical impedance analysis (BIA). These methods assumean animal’s mass consists of 2 compartments: fat massand fat-free (lean body) mass with water restricted tothe animal’s fat-free mass (Speakman 2001). Based onthis assumption, isotopic dilution and BIA can be usedto estimate an animal’s total body water (TBW) content,which, in combination with the animal’s body mass andregression equations derived from chemical extraction(e.g., Farley and Robbins 1994), can be used to predicttotal body fat (TBF) and protein content.

Isotopic dilution has been used for decades to mea-sure the body composition of mammals (Richmond et al.1962, Nagy and Costa 1980), including wild brown bears(U. arctos) and polar bears (Atkinson and Ramsay 1995;Atkinson et al. 1996; Hilderbrand et al. 1999, 2000; Polis-chuk et al. 2002). Dilution involves dosing chemicallyimmobilized animals with a known amount of either tri-tium (HTO) or deuterium oxide (D2O) and then subse-quently measuring the enrichment levels of these isotopes

within the animal’s water pool upon equilibration. Iso-topic dilution is unaffected by environmental conditions,body position, or anesthesia levels (Farley and Robbins1994, Speakman 1997). However, for adult bears it re-quires immobilization for ≥90 minutes to allow equili-bration (Farley and Robbins 1994), skill in performinginjections of isotopes and blood collection, and time andcost associated with laboratory analyses. Given theseconstraints, isotopic dilution is likely impractical as amethod for monitoring body condition of polar bears inany wide-ranging applications.

Bioelectrical impedance analysis has been developedmore recently to measure the body composition of mam-mals (Gales et al. 1994, Hundertmark and Schwartz 2002,Barthelmess et al. 2006, Pitt et al. 2006), including wildbrown, American black (U. americanus), and polar bears(Atkinson and Ramsay 1995; Hilderbrand et al. 1999,2000; Gau and Case 2002; Harlow et al. 2002; McKinneyet al. 2014; Sciullo et al. 2016). Bioelectrical impedanceanalysis involves measuring the electrical conductance ofthe body using a known alternating current across elec-trodes positioned at distal ends of the body (Kushner1992, Marken Lichtenbelt 2001). This provides a mea-sure of electrical resistance across the body, which incombination with a measure of the distance between theelectrodes and a measure of body mass, can be relatedto TBW, which is often derived from isotopic dilution(Kushner 1992, Farley and Robbins 1994, Marken Licht-enbelt 2001). Bioelectrical impedance analysis can berapidly performed, but can be prone to errors dependingon the animal’s body position, depth of anesthesia, in-juries, ambient and body temperature, and moisture onor in contact with the animal’s body (Farley and Rob-bins 1994, Marken Lichtenbelt 2001). The last 3 fac-tors are of particular concern for using BIA to measurethe body composition of polar bears on the sea ice be-cause of the combination of low and variable ambienttemperatures and the potential for water, snow, or iceto affect resistance measures. For example, Farley andRobbins (1994) reported erroneous results in using BIAwhen bears were resting in standing water. These po-tential sources of error led Vongraven et al. (2012) toconclude that BIA was unsuitable as a standard moni-toring tool for polar bears. Yet, BIA has been used forpolar bears on land (Atkinson and Ramsay 1995, Sciulloet al. 2016) and on the sea ice (McKinney et al. 2014).Sciullo et al. (2016) found weak but significant correla-tions between TBF in polar bears derived from BIA andlipid content in adipose tissue and energy density. Nev-ertheless, Sciullo et al. (2016) concluded that potentialsources of error in the field make BIA impractical for

Ursus 28(2):171–181 (2017)

METHODS TO ASSESS POLAR BEAR BODY CONDITION � Pagano et al. 173

monitoring body condition of polar bears. However, un-like lipid content in adipose tissue or energy density mod-els, BIA measures an animal’s body composition follow-ing a 2-compartment model and, hence, weak correlationsmay reflect differences in how these methods assess con-dition rather than measurement errors. In fact, body com-position measures from BIA conducted on polar bears onthe sea ice have yet to be evaluated in comparison withbody composition measures using techniques that alsoassume a 2-compartment model (e.g., isotopic dilution).

Given the potential value of BIA as a rapid measureof polar bear body composition, we conducted isotopicdilutions and BIA measurements of female polar bears ofthe Southern Beaufort Sea subpopulation captured on thesea ice. We used percent TBF as an indicator of bears’energy stores and evaluated the reliability of body com-position measures from BIA in comparison with similarmetrics derived from isotopic dilution. We also comparedestimates of percent TBF from these 2 methods with es-timates of condition derived from a body condition index(Cattet et al. 2002) and an energy density model (Molnaret al. 2009). Both of these alternative condition metricshave been used to monitor changes in body conditionwithin polar bear subpopulations (Obbard et al. 2006,2016; Rode et al. 2010, 2012, 2014a, b).

MethodsWe captured adult (≥4 yr) and subadult (2–3 yr) fe-

male polar bears without dependent young on the seaice north of Prudhoe Bay, Alaska, USA, during April2014–2016. We located polar bears from a helicopterand immobilized them with a rapid-injection dart (PalmerCap-Chur Equipment, Douglasville, Georgia, USA) con-taining zolazepam–tiletamine (Telazol R©; Zoetis, Parsip-pany, New Jersey, USA [Stirling et al. 1989]). In 2015and 2016, bears were recaptured 8–11 days later as partof a separate study. We measured total body mass (TBM)at each capture using an electronic load cell suspendedfrom an aluminum tripod. We measured straight-linebody length (SLBL) as the straight-line distance fromthe tip of the nose to the end of the last tail vertebra whilebears were sternally recumbent. We measured snout-to-vent length (SVL) following the dorsal contours (Farleyand Robbins 1994). Bears that had not been previouslycaptured were aged based on counts of cementum an-nuli from an extracted vestigial premolar (Calvert andRamsay 1998). Procedures were approved by the AnimalCare and Use Committees of the U.S. Geological Survey,Alaska Science Center and the University of California,

Santa Cruz. Research was approved under U.S. Fish andWildlife Service Marine Mammal Permit MA690038.

Bioelectrical impedance analysisWith immobilized bears positioned in sternal recum-

bency, we attached anterior electrodes to the upper lip atthe level of the canines while we attached posterior elec-trodes to needles positioned approximately 3 cm on ei-ther side of the base of the tail (Farley and Robbins 1994).We measured resistance (R) using a Quantum X analyzer(RJL Systems, Clinton Township, Michigan, USA) usingan alternating current of 425 μA at 50 kHz. The analyzerhas a resistance range of 0–1,000 ohms and a resolutionof 0.1 ohms. We took 3 separate resistance measurementsand used the average. In 2015–2016, we took BIA, SVL,and TBM measurements at both the initial capture andrecapture. Bears were completely immobile and relaxedat the time of measurement.

We calculated TBW from resistance and SVL mea-surements using the regression equation for polar bearsderived from Farley and Robbins (1994), where TBW =−1.86 + 0.231 × (SVL2/R) + 0.074 × (TBM). Weused measures of TBW to determine measures of TBFbased on Farley and Robbins (1994), where %TBF =98.01 − 1.28 × (%TBW).

Isotopic dilutionFollowing BIA measurements, we inserted a jugular

catheter to facilitate blood sampling and administrationof isotopes. We collected an initial blood sample to serveas a baseline measure of D2O. We then injected the bearintravenously with a precisely weighed ( ± 0.01 g; OhausScout Pro, Parsippany, New Jersey, USA) dose of 0.06–0.14 g/kg of 99.9% enriched D2O (Isotec, Inc., Miamis-burg, Ohio, USA or Cambridge Isotope Laboratories,Inc., Tewksbury, Massachusetts, USA) with NaCl addedto make it 0.9% isotonic and sterilized using a 0.2-μ Mil-lipore filter (Corning, Inc., Corning, New York, USA). Oninjection, we back-washed the syringe with blood 3 timesto ensure all D2O had been injected into the bear. Wekept the bear immobilized for 90–120 minutes after theinjection of D2O to allow isotope equilibration (Arnould1990; Farley and Robbins 1994; D.P. Costa, Universityof California, Santa Cruz, personal communication). Wecollected serial blood samples 60 minutes, 90 minutes,and 120 minutes after dosing to evaluate equilibrationcurves (Speakman 1997). Bears recaptured in 2015 and2016 were again subject to isotopic dilution as describedabove to measure potential changes in composition. Atrecapture, we assessed the disappearance rates of D2Obetween capture and recapture by collection of an initial

Ursus 28(2):171–181 (2017)


blood sample and serial blood samples collected 60 min-utes and 90 minutes after dosing of D2O. We collectedblood in evacuated tubes without anticoagulants and cen-trifuged tubes to separate serum from red blood cells. Westored serum frozen in cryogenic vials at −20◦C untilanalysis.

We analyzed serum samples for D2O concentration(Metabolic Solutions, Inc., Nashua, New Hampshire,USA). We calculated TBW as the ratio of the amountof D2O injected to the concentration of D2O in the bodyat equilibration per the plateau method (Speakman 1997,eq. 17.11). We used measures of TBW to determine mea-sures of TBF based on Farley and Robbins (1994), where%TBF = 96.85 − 1.33 × (%TBW).

Body condition metricsUsing data on TBM and SLBL, we used 2 metrics to

estimate the body condition of female polar bears cap-tured in this study. The first was an energy density modeldeveloped by Molnar et al. (2009). This model estimatesstructural mass from SLBL and storage mass from SLBLand TBM (Molnar et al. 2009). The model uses bodycomposition measures from previous studies and energydensity estimates of fats and proteins to convert storagemass to storage energy (MJ; Molnar et al. 2009). Weconverted measures of TBM and SLBL to storage en-ergy (E) using the equation for adult female polar bears,where E = 26.14 × TBM − 390.53 × SLBL3 (Molnaret al. 2009). We converted storage energy to energy den-sity (MJ/kg) by dividing by lean body mass based on theequations of Molnar et al. (2009). We excluded measuresfrom subadult females because storage energy equationshave not been developed for this group (Molnar et al.2009).

The second metric was a residual body condition index(BCI: Cattet et al. 2002). This index uses the standard-ized residuals from the regression of TBM and SLBL asa measure of condition, with a potential range of −3 to+ 3 (Cattet et al. 2002). Rode et al. (2014a) used mea-sures of TBM and SLBL from polar bears captured in theSouthern Beaufort Sea subpopulation between 1982 and2013 to develop a BCI for polar bears in the SouthernBeaufort Sea subpopulation using the methods of Cattetet al. (2002). We applied the equation from Rode et al.(2014a) to the TBM and SLBL measurements from thebears captured in this study, where BCI = (lnTBM −2.29 × lnSLBL + 6.7) / (2.58 − 0.45 × lnSLBL).

Statistical analysesLarge amounts of food in the stomach (i.e., gut fill)

can cause erroneous measures of body composition us-

ing either BIA or isotopic dilution (Farley and Robbins1994, Hilderbrand et al. 1999); therefore, we excludedmeasurements from bears that were identified to havefull stomachs at capture based on stomach palpation. Weused paired t-tests to compare estimates of TBW andTBF derived using BIA versus isotopic dilution. We thenused linear regressions to evaluate the relationships be-tween (1) BIA and isotopic dilution in their estimatesof TBW and TBF; (2) resistance and SVL2/ resistance(i.e., conductor volume) with isotopic dilution-derivedTBW (Farley and Robbins 1994) to assess whether mea-surements of SVL or equations derived from Farley andRobbins (1994) might influence our TBW measurementsderived using BIA; (3) BIA-derived TBW and TBF andisotopic dilution-derived TBW and TBF with TBM toassess whether body mass alone is a suitable predictor ofbody composition; and (4) energy density and BCI com-pared with TBF derived from isotopic dilution and fromBIA. We conducted all analyses in Program R (R CoreTeam 2014), and considered differences of P ≤ 0.05 to besignificant. For regressions, we report the coefficient ofdetermination (r2) and the standard error of the estimate(SEE).

ResultsWe captured 4 adult female polar bears in 2014, 3

adult female and 1 subadult female polar bears in 2015,and 2 adult female polar bears in 2016. In 2015, we re-captured 3 of these bears 9–11 days later. In 2016, werecaptured both bears 8–9 days later. One adult female in2014 was identified to have a full stomach at capture andwas excluded from analyses. Therefore, we used pairedBIA and isotopic dilution measurements from 14 cap-tures of 9 individuals. One individual had only a singleBIA resistance measurement; all others received 3 resis-tance measures. Bioelectrical impedance analysis resis-tance measures within individuals differed by ≤5.8 ohmswith a maximum standard error in BIA-derived percentTBF of 1.6% (Fig. 1). Snout-to-vent length measures dif-fered between capture and recapture of individuals by anaverage of 6.5 cm (SE = 2.1, n = 5). D2O equilibrationoccurred within 90 minutes for all measurements (Fig.2). Measures of TBM, TBW, and TBF were all greaterin adult bears than in the subadult (Table 1). Total bodywater derived from BIA differed significantly from esti-mates made by isotopic dilution (t13 = −3.6, P = 0.003).Percent TBF derived from BIA did not differ significantlyfrom estimates made by isotopic dilution (TBF: t13 = 1.5,P = 0.15).

Ursus 28(2):171–181 (2017)


Fig. 1. Mean ( ± SE) percent total body fat (TBF) de-termined by bioelectrical impedance analysis basedon 3 separate resistance measurements from femalepolar bears (Ursus maritimus) captured on the seaice of the southern Beaufort Sea in April 2014–2016.

Kilograms of TBW derived from BIA accounted for70% of the variation in kg TBW as measured by isotopicdilution (SEE = 8.70, P < 0.001; Fig. 3a), but there wasno relationship in percent TBW between the 2 methods(r2 = 0.12, SEE = 2.55, P = 0.22; Fig. 3b). Similarly,kg TBF were weakly related between the 2 methods (r2

= 0.28, SEE = 8.39, P = 0.05; Fig. 3c), but there was norelationship for percent TBF (r2 = 0.12; SEE = 3.39, P= 0.22; Fig. 3d).

Bioelectrical impedance analysis resistance andSVL2/resistance accounted for 63% and 65%, respec-tively, of the variation in kg TBW derived from iso-topic dilution (BIA resistance: SEE = 9.77, P < 0.001;SVL2/resistance: SEE = 9.45, P < 0.001). Total bodymass accounted for 91% of the variation in kg TBWderived from isotopic dilution (SEE = 4.80, P < 0.001;Fig. 4a), but only 66% of the variation in kg TBW derivedfrom BIA (SEE = 11.73, P < 0.001; Fig. 4b). Total bodymass accounted for 58% of the variation in kg TBF asmeasured by isotopic dilution (SEE = 6.38, P = 0.002;

Fig. 2. Equilibration of deuterium oxide (D2O) en-richment levels (mean ± SE; parts per million) in 9 fe-male polar bears (Ursus maritimus), captured on thesea ice of the southern Beaufort Sea in April 2014–2016, prior to and following an intravenous injectionof D2O.

Fig. 4c), but was not related to kg TBF derived from BIA(r2 = 0.18, SEE = 15.02, P = 0.13; Fig. 4d). Neitherpercent TBF derived from isotopic dilution nor percentTBF derived from BIA were related to TBM (r2 = 0.04,SEE = 3.54, P = 0.47; r2 = 0.02, SEE = 7.95, P = 0.65,respectively).

There was a strong relationship between energy den-sity and BCI (r2 = 0.99, SEE = 0.05, P < 0.001; Fig. 5).Energy density was related to percent TBF based on BIA(r2 = 0.56, SEE = 5.81, P = 0.005; Fig. 6a), but wasnot related to percent TBF based on isotopic dilution (r2

= 0.19, SEE = 3.37, P = 0.16; Fig. 6b). Similarly, BCIwas related to BIA-derived percent TBF (r2 = 0.40, SEE= 6.20, P = 0.02; Fig. 6c), but not percent TBF derivedfrom isotopic dilution (r2 = 0.05, SEE = 3.53, P = 0.44;Fig. 6d). Body condition index was related to kg TBM (r2

= 0.51, SEE = 19.98, P = 0.004), but energy density wasnot related to kg TBM (r2 = 0.25, SEE = 20.2, P = 0.10).

DiscussionOur results indicate that BIA may be useful to detect

large-scale variations in the body composition of wild

Table 1. Mean ( ± SE) body mass, total body water (TBW), and total body fat (TBF) from female polar bears(Ursus maritimus) captured on the sea ice of the southern Beaufort Sea in April 2014–2016. Bears were bothdosed with deuterium oxide (isotopic dilution) and sampled by bioelectrical impedance analysis (BIA).

Isotopic dilution BIA

Age class N Body mass (kg) TBW (kg) TBF (kg) TBW (kg) TBF (kg)

Ad 12 183.9 ± 6.4 107.4 ± 3.1 35.3 ± 2.9 116.7 ± 5.0 30.8 ± 4.8Subad 2 133.6 ± 7.0 75.1 ± 3.2 29.5 ± 2.5 82.6 ± 4.7 25.2 ± 0.9

Ursus 28(2):171–181 (2017)


Fig. 3. Relationships between (a) kg total body water (TBW) determined by bioelectrical impedance analysis(BIA) and isotopic dilution (Dilution; y = 28.49 + 0.66x), (b) percent TBW determined by BIA and isotopic dilution,(c) kg total body fat (TBF) determined by BIA and isotopic dilution (y = 25.30 + 0.31x), and (d) percent TBFdetermined by BIA and isotopic dilution from female polar bears (Ursus maritimus) captured on the sea ice ofthe southern Beaufort Sea in April 2014–2016. ‘SEE’ is standard error of the estimate.

polar bears, but may not reflect the fine-scale individualvariations in body composition that were apparent usingisotopic dilution. Total body water and TBF measuredusing BIA and isotopic dilution were related when val-ues were unscaled by body mass, which suggests thatboth methods distinguished differences between overalllarger and smaller TBW and TBF pools resulting fromdifferences in body mass. Additionally, there was no dif-

ference in percent TBF using either method in a pairwisecomparison. However, there was no relationship betweenpercent TBW or TBF estimated by BIA and isotopic di-lution. This suggests that BIA may provide a reasonablemeasure of mean body composition among groups of po-lar bears (e.g., between sexes and bears of different repro-ductive statuses), but errors resulting from measurementsin the field may affect the resolution of body composition

Ursus 28(2):171–181 (2017)


Fig. 4. Relationships between (a) total body mass (TBM) and kg total body water (TBW) determined by isotopicdilution (Dilution; y = 8.95 + 0.53x), (b) TBM and kg TBW determined by bioelectrical impedance analysis (BIA;y = 10.68 + 0.58x), (c) TBM and kg total body fat (TBF) determined by isotopic dilution (y = −11.91 + 0.26x), and(d) TBM and kg TBF determined by BIA from female polar bears (Ursus maritimus) captured on the sea ice ofthe southern Beaufort Sea in April 2014–2016. ‘SEE’ is standard error of the estimate.

estimates, thus limiting the utility of BIA in discerningthe condition of individual bears within groups (e.g., indi-viduals of the same sex, reproductive status, and season)or in monitoring changes within the same individual over

Fig. 5. Relationship (y = −3.14 + 0.17x) between anenergy density model and a residual body conditionindex (BCI) from 8 adult female polar bears (Ursusmaritimus) captured 12 times on the sea ice of thesouthern Beaufort Sea in April 2014–2016. ‘SEE’ isstandard error of the estimate.

time. Bowen et al. (1999) similarly found BIA suitablefor measuring mean TBW among groups of grey seals(Halichoerus grypus), but a poor measure for discerningTBW of individual seals. Thus, caution may be neededin attempting to use BIA as a long-term monitoring toolfor polar bears.

The use of BIA for estimating TBW is based on equa-tions in which resistance readings and body length arerelated to TBW estimated from isotopic dilution (Far-ley and Robbins 1994). In our study, the relationshipbetween resistance and TBW derived from isotopic di-lution was considerably weaker than previously reported(Hoffer et al. 1969, Farley and Robbins 1994, Gales et al.1994, Bowen et al. 1999). Other studies have shown thatrelationships between BIA and TBW derived using iso-topic dilution can provide suitable measures of TBF inbrown and black bears (Farley and Robbins 1994, Hilder-brand et al. 1998). However, calculations of TBF arebased on relationships with percent TBW (Farley andRobbins 1994), and given the weak relationship we foundbetween percent TBW using BIA and isotopic dilution,

Ursus 28(2):171–181 (2017)


Fig. 6. Relationships between (a) an energy density model and percent total body fat (TBF) determinedby bioelectrical impedance analysis (BIA; y = −12.05 + 2.15x), (b) an energy density model and percent TBFdetermined by isotopic dilution (Dilution), (c) a residual body condition index (BCI) and percent TBF determinedby BIA (y = 25.24 + 8.99x), and (d) a BCI and percent TBF determined by isotopic dilution from female polarbears (Ursus maritimus) captured on the sea ice of the southern Beaufort Sea in April 2014–2016. ‘SEE’ isstandard error of the estimate.

we would expect a similarly weak relationship in TBFpredictions.

Field conditions in our study, which are more extremefor polar bears compared with other bear species, mayhave affected estimates of TBW using BIA. Moisture,ambient temperature, body temperature, and gut fill areknown to affect BIA measurements (Farley and Robbins1994, Hilderbrand et al. 1998, Marken Lichtenbelt 2001)and may be particularly problematic when using BIAon polar bears on the sea ice in the spring. For exam-ple, impedance is known to decline with both skin andbody temperature (Marken Lichtenbelt 2001). Ambienttemperatures in this study varied by 31◦C, rectal temper-atures of bears varied by 3◦C, and bears were processedwhile lying on snow and ice. Though other studies haveused plastic sheets underneath animals to potentially re-duce the loss of electrical current to the ground (Farleyand Robbins 1994, Atkinson and Ramsay 1995, Bowenet al. 1999, Sciullo et al. 2016), no data exist to assesswhether ground insulation improves the accuracy of re-sistance measures. Other factors that could affect BIA

and our other metrics are common issues in field stud-ies of ursids, whether conducted on the sea ice or land.Wet fur could affect BIA resistance measures as wellas inflate body mass measures, which would affect allof our metrics. However, it may be a less likely sourceof error in our study because bears were relatively dryand swam infrequently (0.3% of the time; A.M. Paganoet al., U.S. Geological Survey, unpublished data). Addi-tionally, gut fill could inflate body mass measures, whichwould similarly affect all of our metrics. We palpatedanimals’ stomachs to assess gut fill and excluded the oneanimal that was assessed to have fed recently. Urine inthe bladder has the potential to minimally affect TBWmeasurements using isotopic dilution (Nagy and Costa1980), but we were unable to account for this potentialsource of error. Lastly, field morphometric measures aredifficult to measure accurately and have been found to beassociated with considerable error among individual re-searchers (Cattet et al. 1997), which could further affectthe accuracy of measures using BIA, energy density mod-els, or BCI metrics; but such measures are not required

Ursus 28(2):171–181 (2017)


in estimating body composition using isotopic dilution.Although Hilderbrand et al. (1998) found strong correla-tions between BIA and whole-body chemical extractionin black and brown bears, they found stronger correla-tions using isotopic dilution and concluded that isotopicdilution was more suitable for field applications given thepotential sources of error in using BIA in the field. Our re-sults further support this conclusion, with the caveat thatuse of isotopic dilution more than doubled our typicalprocessing time of solitary female polar bears, requiringan average of 152 minutes compared with our typicalaverage of 69 minutes (A.M. Pagano et al., unpublisheddata).

Energy density and BCI also were not related to per-cent TBF derived from isotopic dilution. Percent TBFdiffered by <10% among individuals based on isotopicdilution, which may have been insufficient variation todetect significant relationships using indirect measuresof condition (i.e., energy density and BCI). Total bodyfat estimated from BIA was more closely correlated withenergy density and BCI than was TBF based on isotopicdilution, which may be a result of BIA, energy density,and BCI all incorporating morphometric measurementsin their derivations. Nevertheless, the relationship be-tween percent TBF derived from BIA and energy densitywas weak (r2 = 0.56) and similar to the relationship re-ported by Sciullo et al. (2016; r2 = 0.49) from a largerdata set of polar bears captured onshore in the West-ern Hudson Bay subpopulation. Energy density includesboth the energetic content of body fat as well as that ofnon-structural lean tissue (Molnar et al. 2009). Similarly,Cattet et al. (2002) validated the BCI against individualmasses of fat and skeletal muscle from dissected polarbears, which would include fat, protein, and some amountof water and would not include lipids contained at the cel-lular level. Hence, our weak correlations of percent TBFwith BCI and energy density may in part be related todifferences in the tissues being used to assess body con-dition. Additionally, both energy density and BCI met-rics were developed using data from bears of varyingreproductive status, from different seasons, and across awider range of body masses (and presumably composi-tion) than we used in the current study. Energy densityand BCI strongly correlated with one another, which wasalso reported by Rode et al. (2014a). Our results fromisotopic dilution suggest that, at fine scales (i.e., withinthe same individual or same group, e.g., individuals of thesame sex, reproductive status, and season), energy den-sity and BCI may not be predictive of an animal’s bodyfat. Nevertheless, energy density has been shown to bea strong predictor of reproductive output in polar bears

from the Western Hudson Bay subpopulation (Molnaret al. 2011). Further, both energy density and BCI havebeen found to vary in expected ways, such as femaleswith cubs-of-the-year having lower measures than lonefemales and females with older cubs, and among seasons(Molnar et al. 2009, Rode et al. 2014a, Obbard et al.2016). Hence, despite the lack of ability of these metricsto predict TBF at fine scales, energy density and BCIappear to be suitable for evaluating broader, population-level differences and changes in body condition.

Our results indicate that BIA, when used on female po-lar bears lying on the snow or ice, only coarsely measuresbody composition and is likely no more advantageous inevaluating polar bear body condition than is the use of anenergy density model or BCI, which only require stan-dard morphometric measurements. Most capture stud-ies routinely collect morphometric measurements, so en-ergy density models or BCI can also be used to monitorchanges in body condition from historical cross-sectionaldata sets. However, our results found both the energy den-sity model and BCI failed to predict the percent body fatof lone female polar bears, which suggests that thesemetrics should not be used to evaluate the condition ofpolar bears at fine physiological scales or in longitudinalstudies.

AcknowledgmentsThis work was supported by the U.S. Geological Sur-

vey’s Changing Arctic Ecosystems Initiative. We thankT. Atwood, T. Donnelly, G. Durner, K. Simac, and M.Spriggs for assistance in the field. We thank helicopterpilot F. Ross (Soloy Helicopters) for field support. G.Durner, G. Hilderbrand, and T. Williams provided valu-able input on earlier versions of the manuscript. We thankthe Associate Editor and 3 reviewers for their commentsthat improved the manuscript. Data from this study areavailable from the U.S. Geological Survey data portal:https://doi.org/10.5066/F7ZP447H. Use of trade namesis for descriptive purposes only and does not imply en-dorsement by the U.S. Government.

Literature citedARNOULD, J.P.Y. 1990. The energetics of lactation in po-

lar bears. Thesis, University of Saskatchewan, Saskatoon,Saskatchewan, Canada.

———, AND M.A. RAMSAY. 1994. Milk production andmilk consumption in polar bears during the ice-free pe-riod in western Hudson Bay. Canadian Journal of Zoology72:1365–1370.

ATKINSON, S.N., R.A. NELSON, AND M.A. RAMSAY. 1996.Changes in the body composition of fasting polar bears

Ursus 28(2):171–181 (2017)


(Ursus maritimus): The effect of relative fatness on proteinconservation. Physiological Zoology 69:304–316.

———, AND M.A. RAMSAY. 1995. The effects of prolongedfasting on the body composition and reproductive success offemale polar bears (Ursus maritimus). Functional Ecology9:559–567.

BARTHELMESS, E.L., M.L. PHILLIPS, AND M.E. SCHUCKERS.2006. The value of bioelectrical impedance analysis vs. con-dition indices in predicting body fat stores in North Amer-ican porcupines (Erethizon dorsatum). Canadian Journal ofZoology 84:1712–1720.

BOWEN, W.D., C.A. BECK, AND S.J. IVERSON. 1999. Bioelectri-cal impedance analysis as a means of estimating total bodywater in grey seals. Canadian Journal of Zoology 77:418–422.

CALVERT, W., AND M.A. RAMSAY. 1998. Evaluation of agedetermination of polar bears by counts of cementum growthlayer groups. Ursus 10:449–453.

CATTET, M. 1990. Predicting nutritional condition in blackbears and polar bears on the basis of morphological andphysiological measurements. Canadian Journal of Zoology68:32–39.

CATTET, M.R.L., S.N. ATKINSON, S.C. POLISCHUK, AND M.A.RAMSAY. 1997. Predicting body mass in polar bears: Ismorphometry useful? Journal of Wildlife Management61:1083–1090.

———, N.A. CAULKETT, M.E. OBBARD, AND G.B. STEN-HOUSE. 2002. A body condition index for ursids. CanadianJournal of Zoology 80:1156–1161.

CHERRY, S.G., A.E. DEROCHER, I. STIRLING, AND E.S.RICHARDSON. 2009. Fasting physiology of polar bears inrelation to environmental change and breeding behavior inthe Beaufort Sea. Polar Biology 32:383–391.

DEROCHER, A.E., N.J. LUNN, AND I. STIRLING. 2004. Polarbears in a warming climate. Integrative and ComparativeBiology 44:163–176.

FARLEY, S.D., AND C.T. ROBBINS. 1994. Development of twomethods to estimate body composition of bears. CanadianJournal of Zoology 72:220–226.

GALES, R., D. RENOUF, AND G.A.J. WORTHY. 1994. Use ofbioelectrical impedance analysis to assess body compositionof seals. Marine Mammal Science 10:1–12.

GAU, R.J., AND R. CASE. 2002. Evaluating nutritional conditionof grizzly bears via 15N signatures and insulin-like growthfactor-1. Ursus 13:285–291.

GREEN, A.J. 2001. Mass/length residuals: Measures of bodycondition or generators of spurious results? Ecology82:1473–1483.

HARLOW, H.J., T. LOHUIS, R.G. GROGAN, AND T.D.I. BECK.2002. Body mass and lipid changes by hibernating reproduc-tive and nonreproductive black bears (Ursus americanus).Journal of Mammalogy 83:1020–1025.

HILDERBRAND, G.V., S.D. FARLEY, AND C.T. ROBBINS. 1998.Predicting body condition of bears via two field methods.Journal of Wildlife Management 62:406–409.

———, S.G. JENKINS, C.C. SCHWARTZ, T.A. HANLEY, AND

C.T. ROBBINS. 1999. Effect of seasonal differences in dietarymeat intake on changes in body mass and composition inwild and captive brown bears. Canadian Journal of Zoology77:1623–1630.

———, C.C. SCHWARTZ, C.T. ROBBINS, AND T.A. HANLEY.2000. Effect of hibernation and reproductive strategy onbody mass and condition of coastal brown bears. Journal ofWildlife Management 64:178–183.

HUNDERTMARK, K.J., AND C.C. SCHWARTZ. 2002. Evaluationof bioelectrical impedance analysis as an estimator of moosebody composition. Wildlife Society Bulletin 30:915–921.

HWANG, Y.T., S. LARIVIERE, AND F. MESSIER. 2015. Evalu-ating body condition of striped skunks using non-invasivemorphometric indices and bioelectrical impedance analysis.Wildlife Society Bulletin 33:195–203.

KUSHNER, R.F. 1992. Bioelectrical impedance analysis: A re-view of principles and applications. Journal of the AmericanCollege of Nutrition 11:199–209.

MARKEN LICHTENBELT, W.D.V. 2001. The use of bioelectri-cal impedance analysis for estimation of body composition.Pages 157–182 in J.R. Speakman, editor. Body compositionanalysis of animals: A handbook of non-destructive meth-ods. Cambridge University Press, New York, New York,USA.

MCKINNEY, M.A., T. ATWOOD, R. DIETZ, C. SONNE, S.J. IVER-SON, AND E. PEACOCK. 2014. Validation of adipose lipidcontent as a body condition index for polar bears. Ecologyand Evolution 4:516–527.

MOLNAR, P.K., A.E. DEROCHER, T. KLANJSCEK, AND M.A.LEWIS. 2011. Predicting climate change impacts on polarbear litter size. Nature Communications 2:186.

———, T. KLANJSCEK, A.E. DEROCHER, M.E. OBBARD, AND

M.A. LEWIS. 2009. A body composition model to estimatemammalian energy stores and metabolic rates from bodymass and body length, with application to polar bears. TheJournal of Experimental Biology 212:2313–2323.

NAGY, K.A., AND D.P. COSTA. 1980. Water flux in animals:Analysis of potential errors in the tritiated water method.American Journal of Physiology 238:R454–R465.

OBBARD, M.E., M.R.L. CATTET, E.J. HOWE, K.R. MIDDEL,E.J. NEWTON, G.B. KOLENOSKY, K.F. ABRAHAM, AND C.J.GREENWOOD. 2016. Trends in body condition in polar bears(Ursus maritimus) from the Southern Hudson Bay subpop-ulation in relation to changes in sea ice. Arctic Science2:15–32.

———, ———, T. MOODY, L.R. WALTON, D. POTTER, J. IN-GLIS, AND C. CHENIER. 2006. Temporal trends in the bodycondition of Southern Hudson Bay polar bears. ClimateChange Research Information Note, No. 3. Ontario Ministryof Natural Resources, Sault Ste. Marie, Ontario, Canada.

PATYK, K.A., C. DUNCAN, P. NOL, C. SONNE, K. LAIDRE, M.OBBARD, Ø. WIIG, J. AARS, E. REGEHR, L.L. GUSTAFSON,AND T. ATWOOD. 2015. Establishing a definition of polarbear (Ursus maritimus) health: A guide to research and

Ursus 28(2):171–181 (2017)


management activities. Science of the Total Environment514:371–378.

PILFOLD, N.W., A.E. DEROCHER, I. STIRLING, AND E.RICHARDSON. 2015. Multi-temporal factors influence preda-tion for polar bears in a changing climate. Oikos 124:1098–1107.

PITT, J.A., S. LARIVIERE, AND F. MESSIER. 2006. Condi-tion indices and bioelectrical impedance analysis to predictbody condition of small carnivores. Journal of Mammalogy87:717–722.

POLISCHUK, S.C., R.J. NORSTROM, AND M.A. RAMSAY. 2002.Body burdens and tissue concentrations of organochlorinesin polar bears (Ursus maritimus) vary during seasonal fasts.Environmental Pollution 118:29–39.

R CORE TEAM. 2014. R: A language and environment for sta-tistical computing. R Foundation for Statistical Comput-ing, Vienna, Austria. http://www.R-project.org/. Accessed29 May 2015.

RAMSAY, M.A., AND I. STIRLING. 1988. Reproductive biologyand ecology of female polar bears (Ursus maritimus). Jour-nal of Zoology 214:601–634.

REGEHR, E.V., C.M. HUNTER, H. CASWELL, S.C. AMSTRUP,AND I. STIRLING. 2010. Survival and breeding of polar bearsin the southern Beaufort Sea in relation to sea ice. Journalof Animal Ecology 79:117–127.

———, N.J. LUNN, S.C. AMSTRUP, AND I. STIRLING. 2007.Effects of earlier sea ice breakup on survival and popula-tion size of polar bears in western Hudson Bay. Journal ofWildlife Management 71:2673–2683.

RICHMOND, C.R., W.H. LANGHAM, AND T.T. TRUJILLO. 1962.Comparative metabolism of tritiated water by mammals.Journal of Cellular Physiology 59:45–53.

RODE, K.D., S.C. AMSTRUP, AND E.V. REGEHR. 2010. Reducedbody size and cub recruitment in polar bears associated withsea ice decline. Ecological Applications 20:768–782.

———, A.M. PAGANO, J.F. BROMAGHIN, T.C. ATWOOD, G.M.DURNER, K.S. SIMAC, AND S.C. AMSTRUP. 2014a. Effectsof capturing and collaring on polar bears: Findings fromlong-term research on the southern Beaufort Sea population.Wildlife Research 41:311–322.

———, E. PEACOCK, M. TAYLOR, I. STIRLING, E.W. BORN,K.L. LAIDRE, AND Ø. WIIG. 2012. A tale of two polar bearpopulations: Ice habitat, harvest, and body condition. Pop-ulation Ecology 54:3–18.

———, E.V. REGEHR, D.C. DOUGLAS, G. DURNER, A.E. DE-ROCHER, G.W. THIEMANN, AND S.M. BUDGE. 2014b. Vari-ation in the response of an Arctic top predator experiencinghabitat loss: Feeding and reproductive ecology of two polarbear populations. Global Change Biology 20:76–88.

SCHAMBER, J.L., D. ESLER, AND P.L. FLINT. 2009. Evaluatingthe validity of using unverified indices of body condition.Journal of Avian Biology 40:49–56.

SCHULTE-HOSTEDDE, A.I., B. ZINNER, J.S. MILLAR, AND G.J.HICKLING. 2005. Restitution of mass–size residuals: Vali-dating body condition indices. Ecology 86:155–163.

SCIULLO, L., G.W. THIEMANN, AND N.J. LUNN. 2016. Com-parative assessment of metrics for monitoring the body con-dition of polar bears in western Hudson Bay. Journal ofZoology 300:45–58.

SPEAKMAN, J.R. 1997. Doubly labelled water: Theory and prac-tice. Chapman and Hall, London, England, UK.

———. 2001. Body composition analysis of animals: A hand-book of non-destructive methods. Cambridge UniversityPress, New York, New York, USA.

STEVENSON, R.D., AND W.A. WOODS. 2006. Condition indicesfor conservation: New uses for evolving tools. Integrativeand Comparative Biology 46:1169–1190.

STIRLING, I., AND A. DEROCHER. 1993. Possible impacts ofclimatic warming on polar bears. Arctic 46:240–245.

———, N. J. LUNN, AND J. IACOZZA. 1999. Long-term trendsin the population ecology of polar bears in western HudsonBay in relation to climatic change. Arctic 52:294–306.

———, AND N.A. ØRITSLAND. 1995. Relationships betweenestimates of ringed seal (Phoca hispida) and polar bear (Ur-sus maritimus) populations in the Canadian Arctic. Cana-dian Journal of Fisheries and Aquatic Sciences 52:2594–2612.

———, C. SPENCER, AND D. ANDRIASHEK. 1989. Immobiliza-tion of polar bears (Ursus maritimus) with Telazol R© in theCanadian Arctic. Journal of Wildlife Diseases 25:159–168.

THIEMANN, G.W., S.J. IVERSON, AND I. STIRLING. 2006. Sea-sonal, sexual and anatomical variability in the adipose tis-sue of polar bears (Ursus maritimus). Journal of Zoology269:65–76.

VIRGL, J.A., AND F. MESSIER. 1992. Seasonal variation inbody composition and morphology of adult muskrats incentral Saskatchewan, Canada. Journal of Zoology 228:461–477.

VONGRAVEN, D., J. AARS, S. AMSTRUP, S.N. ATKINSON, S. BE-LIKOV, E.W. BORN, T.D. DEBRUYN, A.E. DEROCHER, G.DURNER, M. GILL, N. LUNN, M.E. OBBARD, J. OMELAK,N. OVSYANIKOV, E. PEACOCK, E. RICHARDSON, V. SAHA-NATIEN, I. STIRLING, AND Ø. WIIG. 2012. A circumpolarmonitoring framework for polar bears. Ursus MonographSeries 5:1–66.

WILLIAMS, R., G.A. VIKINGSSON, A. GISLASON, C. LOCKYER,L. NEW, L. THOMAS, AND P. HAMMOND. 2013. Evidence fordensity-dependent changes in body condition and pregnancyrate of North Atlantic fin whales over four decades of varyingenvironmental conditions. ICES Journal of Marine Sciences.doi:10.1093/icesjms/fst059.

WINSTANLEY, R.K., W.A. BUTTEMER, AND G. SAUNDERS.1999. Fat deposition and seasonal variation in body com-position of red foxes (Vulpes vulpes) in Australia. CanadianJournal of Zoology 77:406–412.

Received: October 11, 2016Accepted: September 28, 2017Associate Editor: Obbard

Ursus 28(2):171–181 (2017)

Evaluating methods to assess the body condition of female ...

Documents

Transcript of Evaluating methods to assess the body condition of female ...