ARTICLE

Physiological responses and energy expenditure duringcompetitive fencingRaffaele Milia, Silvana Roberto, Marco Pinna, Girolamo Palazzolo, Irene Sanna, Massimo Omeri,Simone Piredda, Gianmario Migliaccio, Alberto Concu, and Antonio Crisafulli

Abstract: Fencing is an Olympic sport in which athletes fight one against one using bladed weapons. Contests consist of three3-min bouts, with rest intervals of 1 min between them. No studies investigating oxygen uptake and energetic demand duringfencing competitions exist, thus energetic expenditure and demand in this sport remain speculative. The aim of this study wasto understand the physiological capacities underlying fencing performance. Aerobic energy expenditure and the recruitment oflactic anaerobic metabolism were determined in 15 athletes (2 females and 13 males) during a simulation of fencing by using aportable gas analyzer (MedGraphics VO2000), which was able to provide data on oxygen uptake, carbon dioxide production andheart rate. Blood lactate was assessed by means of a portable lactate analyzer. Average group energetic expenditure during thesimulation was (mean ± SD) 10.24 ± 0.65 kcal·min−1, corresponding to 8.6 ± 0.54 METs. Oxygen uptakeand heart rate were alwaysbelow the level of anaerobic threshold previously assessed during the preliminary incremental test, while blood lactate reachedits maximum value of 6.9 ± 2.1 mmol·L−1 during the final recovery minute between rounds. Present data suggest that physicaldemand in fencing is moderate for skilled fencers and that both aerobic energy metabolism and anaerobic lactic energy sourcesare moderately recruited. This should be considered by coaches when preparing training programs for athletes.

Key words: oxygen uptake, carbon dioxide production, blood lactate, heart rate, anaerobic threshold.

Résumé : L’escrime est un sport olympique dans lequel les athlètes s’affrontent avec des armes blanches. Une épreuve comprend3 assauts de 3 minutes chacun avec 1 minute de repos entre les assauts. Il n’y a pas d’études traitant de la consommationd’oxygène et des exigences énergétiques durant les compétitions d’escrime; la dépense énergétique et les exigences énergétiquesde ce sport ne sont encore que spéculatives. Cette étude se propose d’évaluer les capacités physiologiques durant une perfor-mance a l’escrime. On évalue la dépense d’énergie en aérobiose et la sollicitation du métabolisme anaérobie lactique chez15 athlètes (2 femmes, 13 hommes) au cours d’une simulation a l’escrime en utilisant un analyseur de gaz portatif (MedGraphicsVO2000) pour déterminer la consommation d’oxygène, la production de gaz carbonique et le rythme cardiaque. On évalue laconcentration sanguine de lactate au moyen d’un analyseur de lactate portatif. La dépense énergétique moyenne au cours de lasimulation est (moyenne ± écart-type) de 10,24 ± 0,65 kcal·min−1, soit 8,6 ± 0,54 METs. La consommation d’oxygène et le rythmecardiaque se situent toujours sous le seuil anaérobie évalué précédemment au cours d’un test d’effort progressif; la sanguine delactate atteint une valeur maximale de 6,9 ± 2,1 mmol·L−1 durant la dernière minute de récupération entre les épreuves. D’aprèsles résultats de cette étude, les exigences énergétiques de l’escrime sont modérées pour des escrimeurs compétents et les deuxsources d’énergie, aérobie et anaérobie lactique, sont sollicitées modérément. Les entraîneurs devraient prendre en compte cesdonnées au moment de préparer des programmes d’entraînement pour les athlètes. [Traduit par la Rédaction]

Mots-clés : consommation d’oxygène, production de gaz carbonique, lactate sanguin, rythme cardiaque, seuil anaérobie.

IntroductionFencing is a combat sport between 2 athletes who fence each

other using 1 of 3 types of weapon (the foil, the sabre, and theépée, each contested with different rules) and who are protectedby specific fencing clothing, mask, gloves, and plastrons. Fencingis usually practiced indoors and requires skill, technique, andtactical excellence for success. The competition platform measures14 m in length with a width of between 1.5 and 2 m. A judgepresides over the bout with the aid of an electrical scoring appa-ratus connected to the target of the fencer (Roi and Bianchedi2008). The International Federation of Fencing (Federation Inter-nationale D’Escrime) claims 149 member countries worldwide. Forinstance, there are about 80 000, 60 000, 30 000, and 16 000 active

athletes in France, Germany, the United States, and Italy, respec-tively.

A fencing international tournament may last between 9 and11 h. Bouts, during which each athlete is engaged in fighting,represent only about 18% of the total tournament time and, in thisperiod, the effective fight time is between 17 and 48 min. There-fore, the length of dynamic phases is unpredictable, but is usuallyof short duration. Indeed, it has been reported that the durationof every action may be very short and intensive (less than 1 s) or itmay last more than 60 s. Bouts are characterized by submaximalintensity preparatory movements, followed by intensive move-ments of short duration with the aim to touch the opponent (Roiand Bianchedi 2008; Roi and Pittaluga 1997). Taking into consid-eration these facts, energy requirements in fencing competitions

Received 26 May 2013. Accepted 16 September 2013.

R. Milia, S. Roberto, M. Pinna, G. Palazzolo, I. Sanna, A. Concu, and A. Crisafulli. Department of Medical Sciences, Sports Physiology Laboratory,University of Cagliari (Italy), Via Porcell 4, 09124 Cagliari, Italy.M. Omeri and S. Piredda. National Italian Fencing Federation, Rome, Italy.G. Migliaccio. Regional School of Sport of Sardinia, Italian Olympic Committee, Cagliari, Italy.Corresponding author: Antonio Crisafulli (e-mail: crisafulli@tiscali.it).

324

Appl. Physiol. Nutr. Metab. 39: 324–328 (2014) dx.doi.org/10.1139/apnm-2013-0221 Published at www.nrcresearchpress.com/apnm on 26 September 2013.

App

l. Ph

ysio

l. N

utr.

Met

ab. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y C

alif

Dig

Lib

- I

rvin

e on

12/

17/1

4Fo

r pe

rson

al u

se o

nly.

can vary unpredictably, depending on the duration of both thepreparatory and intensive phases. Therefore, from a physiologicalpoint of view, fencing appears to be an intermittent physicallydemanding sport, with short phases of maximal or supramaximalintensity spaced by recoveries. Hence, it is likely that both aerobicand anaerobic energy systems are recruited during a match.

To obtain an understanding of the physiological capacities un-derlying fencing performance it would be useful to know theenergetic demand during a real fight. However, to the best of ourknowledge, the energy requirement during a real fencing compe-tition has never been studied. This information would providebenchmarks for the monitoring athletes’ training. This investiga-tion was devised to study energetic demand during a fencing com-petition and to test the hypothesis that fencing is an activity thatrecruits both aerobic and anaerobic energy systems. In particular,we were interested in measuring aerobic energy expenditure dur-ing a competition and in discovering whether, and to what extent,anaerobic glycolysis is recruited. This information would allowcoaches to design specific training programs able to induce thespecific adaptations required by fencing.

Materials and methods

SubjectsFifteen fencers (2 female, 13 male) that regularly participated in

competitions over the last 4 years were enrolled. Mean ± SD of age,height, and body mass were 21.4 ± 6.9 years, 176.7 ± 10.6 cm, and68.5 ± 12.9 kg, respectively. All subjects were skilled athletes whotrained for 10–12 h a week and had been involved in regular train-ing programs for at least 5 years. In the last year, 6 of them hadparticipated in international competitions, while the other 4 innational tournaments. Thus, our group is representative of mid-upper level athlete fencers.

The study was performed in accordance with the Declarationof Helsinki and was approved by our local ethics committee(University of Cagliari). Written informed consent was obtained fromall participants.

Experimental protocol

Preliminary testAll athletes underwent a preliminary incremental exercise test

on a motorized treadmill (Runrace, Technogym, Forlì, Italy) toassess their anaerobic threshold (AT) and maximal oxygen uptake(VO2max). The test consisted of linear increases of running velocityof 1 km·h−1 every minute, starting at 8 km·h−1, until exhaustion,which was considered as the exercise level at which the subjectwas unable to maintain the running speed, i.e., muscular fatigue.The treadmill slope was set at 1% to compensate for the lack of airfriction (Jones and Doust 1996). Achievement of VO2max was con-sidered as the attainment of at least 2 of the following criteria:(i) a plateau in oxygen uptake (VO2) despite increasing speed(<80 mL·min−1); (ii) respiratory exchange ratio (RER) above 1.10;and (iii) heart rate (HR) ± 10 beats·min−1 of predicted maximum HRwas calculated as 220 – age (Howley et al. 1995). During the pre-liminary test, AT was determined using the V-slope method,which detects AT by using computerized regression analysis of theVO2 slopes vs. the carbon dioxide production (VCO2) plot duringexercise (Beaver et al. 1986), while VO2max was calculated as theaverage VO2 during the final 30 s of the exercise test.

Fencing simulation testOn a different day to the preliminary test, with an interval of at

least 3 days, all subjects underwent a simulated fencing match ona standard fencing platform. Following a 15-min warm-up eachindividual then rested on a bench until cardiorespiratory vari-ables returned to their pre-exercise levels. Recovery was consid-ered complete when HR was below 10 beats·min−1 compared withthe pre-exercise level and when RER was less than 0.9. The last

3 min of seated rest were used to collect values of variables at restand following this period the simulated match was started.

The simulation consisted of 3 bouts, each lasting 3 min andfollowed by 1 min of recovery, during which the subject sat on abench. A further 3 min of recovery was allowed at the end of thesimulation. Hence, the whole simulation test lasted a total of18 min: 3 min of rest before beginning the match, 3 bouts eachlasting 3 min spaced by 3 min of recovery (for a total of 12 min),and 3 min of final recovery. All experiments were conducted be-tween 0900 and 1400 h. Subjects consumed a light meal at least 2 h(interval 2–5 h) before exercising. Meal consisted of a typical Ital-ian breakfast with toasted bread, biscuits, fruit jam, honey, andmilk. Subjects could eat ab libitum. Subjects were also asked toavoid caffeine and alcohol the day before tests were scheduled.

Variables

Assessment of respiratory variables and HRVO2, VCO2, pulmonary ventilation (VE), and HR values were

obtained throughout the preliminary and simulation tests bymeans of a portable metabolic system (MedGraphics VO2000, St.Paul, Minn., USA), which provides a 3-breath average of variablesthrough telemetric transmission. This system has been shown tobe reliable and to have positive agreement compared with a stan-dard metabolic cart for laboratory use (Byard and Dengel 2002;Olson et al. 2003). The device weighs about 1.2 kg and includes themetabolic unit, battery pack, harness, chest belt for HR monitor-ing, face mask, and breathing valve. It is worn on the subject’schest with a harness, without limiting the athlete’s movements.Prior to testing, the VO2000 was calibrated according to manufac-turer’s instructions. During the simulation test, the face mask forgas analysis was placed over the nose and mouth and the fencingmask was then fitted over the face mask.

Measurement of aerobic energy expenditure and anaerobicglycolysis

During the simulation test aerobic energy expenditure (EE, ex-pressed as kcal·min−1) was calculated by utilizing the Weir equa-tion (Weir 1949; Mansell and Macdonald 1990):

EE � 3.941 × VO2 � 1.106 × VCO2

This equation was used when the RER was <1, while an oxygencaloric equivalent of 5.04 was used when EE became >1. In thiscase it was assumed that all aerobic energy was derived fromcarbohydrate oxidation.

Blood samples were obtained with a finger prick and bloodlactate (BLa) concentration was measured at rest, at the end ofeach round and at the third and fifth minutes of the final recoveryby using a portable lactate analyzer (Lactate Pro, Arkray Inc., Kyoto,Japan), to evaluate the time course of lactate concentration.

Statistical analysisData were averaged over 3 min during the rest period before the

simulation bout, during rounds, and during postsimulation recov-ery, while a 1-min average was employed for the recovery periodsbetween rounds. In this manner, information on the time courseof studied variables was gathered and differences among the var-ious protocol periods were detected. Responses are reported asmeans ± SD. Comparisons between periods were performed us-ing repeated measures ANOVA, followed by Neuman–Keuls posthoc test when appropriate. Significance was set at a P valueof <0.05. Descriptive statistics were performed on each variablebefore ANOVA to confirm the assumptions of normality by meansof the Kolmogorov–Smirnov test. The � level was set at P < 0.05.Statistics were calculated employing commercially available soft-ware (Graph-Pad Prism).

Milia et al. 325

Published by NRC Research Press

App

l. Ph

ysio

l. N

utr.

Met

ab. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y C

alif

Dig

Lib

- I

rvin

e on

12/

17/1

4Fo

r pe

rson

al u

se o

nly.

ResultsAll subjects completed the study protocol. Table 1 shows the

results of the incremental preliminary test. Figures 1 to 3 depictresults of the ANOVA test applied to the mean value of variablesduring the various protocol periods. HR (Fig. 1A) increased duringthe fencing bouts in comparison with rest. This HR elevation waspresent during the whole test, including the recovery phases be-tween bouts and the 3 min of final recovery after the test. HR wasbelow the AT level previously assessed during the preliminaryincremental test for the entire period of the simulation, with theexception of the last bout. HR progressively increased throughoutthe whole combat, as testified by the fact that this parameter washigher during fight 3 in comparison with fight 1. VO2 (Fig. 1B) roseduring the combat as compared with rest, without significantdifferences between rounds and recoveries, with the exception ofthe final recovery, when VO2 decreased towards resting level. VO2always remained below the AT level.

Very similar behaviour was observed between VCO2 and VE(Fig. 2A and 2B), since they both increased throughout the test toreturn towards baseline during the final period of recovery. Asreported for HR and VO2, both VCO2 and VE remained below ATvalues throughout the combat simulation.

Finally, Fig. 3A shows EE time course, which obviously resem-bled that of VO2. It is noteworthy that during the whole test EEaveraged 10.24 ± 0.65 kcal·min−1, i.e., 8.6 ± 0.54 METs. Figure 3Bdepicts BLa behaviour. This variable was seen to increase abruptlyeven during the first round to maintain a stable level throughoutthe test. BLa reached its maximum during the final recovery min-ute between rounds, when it reached a value of 6.9 ± 2.1 mmol·L−1.

DiscussionThis investigation aimed to study energetic demand during a

fencing competition. From the present data it appears that aero-bic energy source is only moderately recruited during fencing.Indeed, during the fight, VO2 and HR remained below the AT levelpreviously assessed during the preliminary incremental test,with the exception of HR during the last fight bout, when thisparameter was slightly increased with respect to AT. Very sim-ilar behaviour was also observed for VE and VCO2, thus confirm-ing that fencing imposed moderate respiratory and metabolicstress only.

To the best of our knowledge this is the first study to assessenergy requirements during fencing. Thus, the present informa-tion could be useful for coaches to devise specific training pro-grams able to develop the specific adaptations required byfencing. Notwithstanding the fact that aerobic energy source ap-peared to be only moderately recruited and that athletes per-formed below the level of AT, it is noteworthy that the energyexpenditure seemed to be high, since EE was on average about10 kcal·min−1 (i.e., 600 kcal·h−1) during combat. This phenomenoncan be explained by taking into consideration that the mean ATlevel of the fencers enrolled in the present investigation was highin relation to VO2max, as it was on average about 78% of VO2max.Hence, even though athletes performed below the AT level, itshould be considered that their EE was high in absolute terms.This elevated AT in relation to VO2max bore witness to the conceptthat our group was constituted by well-trained athletes. It is to benoted that this energy expenditure very likely underestimated the

real energy requirement since it did not take into account theenergy derived from the anaerobic metabolism.

Concerning BLa measures, it seems that the lactic anaerobiccapacity was moderately activated to support the energy re-quirements of combat. Indeed, this parameter remained above6 mmol·L−1 throughout the combat simulation, with a peak of6.9 mmol·L−1 during the final recovery. This occurrence appears tobe at odds with the finding that athletes performed below thelevel of AT. This outcome can probably be ascribed to the fact that,during combat, athletes used the arm muscles to a higher level ofrecruitment than during the incremental test employed to assessAT. It is well known that arms have a greater composition infast-twitch fibers than legs and that these kinds of fibers producea greater quantity of lactate than the slow-twitch ones (Saltin andGollnick 1983). Therefore, it is reasonable to predict a greater BLaaccumulation when athletes use arms as compared with legs.

Table 1. Mean group values ± SD of maximum oxygen uptake (VO2max; expressed as absolute and indexed by body mass values), maximum heartrate (HRmax), oxygen uptake at anaerobic threshold (VO2AT), and heart rate at anaerobic threshold (HRAT) reached by subjects during thepreliminary incremental test.

VO2max (mL·min·kg−1) VO2max (mL·min−1) HRmax (beats·min−1) VO2AT (mL·min−1·kg−1) VO2AT (mL·min−1) HRAT (beats·min−1)

Mean 46.3 3168.6 194.5 37.9 2498.5 173.3SD 5.2 558.4 9.6 5.5 424.9 9.2

Fig. 1. Group heart rate (HR) values (A) and oxygen uptake (VO2; B)during the various periods of the fencing match. The level ofanaerobic threshold is identified by a horizontal line. Values aremeans ± SD (n = 15). *, p < 0.05 vs. rest; †, p < 0.05 vs. final recovery;‡, p < 0.05 vs. recovery 1; §, p < 0.05 vs. recovery 2; ¶, p < 0.05 vs.fight 1.

326 Appl. Physiol. Nutr. Metab. Vol. 39, 2014

Published by NRC Research Press

App

l. Ph

ysio

l. N

utr.

Met

ab. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y C

alif

Dig

Lib

- I

rvin

e on

12/

17/1

4Fo

r pe

rson

al u

se o

nly.

Thus, despite the fact that fencers performed below the AT levelpreviously determined during a running test, the relatively highBLa value during the exercise bouts can be explained by takinginto consideration that during the simulated fighting, it is likelytheir arm muscles were more engaged. It is of interest that HRshowed a progressive increase throughout the whole combat, astestified by the fact that this parameter was higher during fights 2and 3 in comparison with fight 1. This phenomenon was not evi-dent for the other gathered metabolic variables. This occurrencesuggests that there was a sort of dissociation between HR and themetabolic requirements of the exercise being performed, whichcaused an increase in HR over the real metabolic engagement.Very similar HR behaviour was described in research reportingthat when the exercise is characterized by alternate phases ofmaximal exercise and recovery, then HR provides overestimatedvalues of energy expenditure (Crisafulli et al. 2006a, 2009). In fact,both maximal and supramaximal bouts of exercise impact pro-foundly on cardiovascular homeostasis since they modify cardiacpre-load, after-load, and contractility, which stress the cardiovas-cular regulatory systems and induce compensatory tachycardia(Crisafulli et al. 2004, 2006b). Hence, the present report furtherstrengthens the concept that the use of HR monitoring to assessthe intensity of exercise may be unreliable in activities that in-volve repeated bouts of maximal and supramaximal exercise andlead to a massive recruitment of anaerobic glycolysis.

Another finding deserving attention is that none of the studiedvariables returned to rest level during the 3 min of final recovery.

This suggests that a 3-min recovery was insufficient to completelyrecover from the strain that was due to exercise. In conclusion,the data from our study suggest that fencing is a moderately phys-ically demanding activity for skilled fencers, such as those studiedin the present investigation. Both the aerobic energy metabolismand anaerobic lactic energy sources are moderately recruited.Thus, present data support the concept that successful fencingperformance depends more on skill and technique than the supe-rior aerobic and anaerobic capacities. Further study is needed tobetter clarify this point. Moreover, HR can be misleading whenestimating energy expenditure in this kind of competition sincethis parameter showed a progressive increase that was dissociatedto oxygen uptake during combat. Finally, interval periods be-tween rounds do not allow for complete recovery. Given the in-complete recovery between bouts, it is important that athletesdevelop specific training programs able to improve this ability. Allthese facts should be considered by coaches when preparing train-ing programs for athletes. Since the scientific literature on fenc-ing is not particularly abundant, future research is needed tobetter characterize this sport from a physiological point of view.

AcknowledgementsThis study was supported by the University of Cagliari, the Ital-

ian Ministry of Scientific Research, and the Regional School ofSport of Sardinia, Italian Olympic Committee (Italy). The authorswish to thank Mr. Barry Mark Wheaton for his editorial assistance.

Fig. 2. Group carbon dioxide production (VCO2) values (A) andpulmonary ventilation (VE; B) during the various periods of thefencing match. The level of anaerobic threshold (AT) is identified bya horizontal line. Values are means ± SD (n = 15). *, p < 0.05 vs. rest;†, p < 0.05 vs. final recovery.

Fig. 3. Group energy expenditure (EE) values (A) and blood lactate(BLa; B) during the various periods of the fencing match. The level ofanaerobic threshold (AT) is identified by a horizontal line. Values aremeans ± SD (n = 15). *, p < 0.05 vs. rest; †, p < 0.05 vs. final recovery.

Milia et al. 327

Published by NRC Research Press

App

l. Ph

ysio

l. N

utr.

Met

ab. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y C

alif

Dig

Lib

- I

rvin

e on

12/

17/1

4Fo

r pe

rson

al u

se o

nly.

ReferencesBeaver, W.L., Wasserman, K., and Whipp, B.J. 1986. A new method for detecting

anaerobic threshold by gas exchange. J. Appl. Physiol. 60(6): 2020–2027.PMID:3087938.

Byard, A.D., and Dengel, D.R. 2002. Validity of a portable metabolic measure-ment system. Med. Sci. Sports Exerc. 34(5): S149. doi:10.1097/00005768-200205001-00829.

Crisafulli, A., Carta, C., Melis, F., Tocco, F., Frongia, F., Santoboni, U.M., et al.2004. Hemodynamic responses following intermittent supramaximal exer-cise in athletes. Exp. Physiol. 89(6): 665–674. doi:10.1113/expphysiol.2004.027946. PMID:15328308.

Crisafulli, A., Pittau, G.L., Lorrai, L., Cominu, M., Tocco, F., Melis, F., et al. 2006a.Poor reliability of heart rate monitoring to assess oxygen consumption dur-ing field training. Int. J. Sports Med. 27(1): 55–59. doi:10.1055/s-2005-837504.PMID:16388443.

Crisafulli, A., Tocco, F., Pittau, G., Lorrai, L., Porru, C., Salis, E., et al. 2006b. Effectof differences in post-exercise lactate accumulation in athletes’ hemodynam-ics. Appl. Physiol. Nutr. Metab. 31(4): 423–431. doi:10.1139/h06-017. PMID:16900232.

Crisafulli, A., Vitelli, S., Cappai, I., Milia, R., Tocco, F., Melis, F., et al. 2009.Physiological responses and energy cost during a simulation of a Muay Thaiboxing match. Appl. Physiol. Nutr. Metab. 34(2): 143–150. doi:10.1139/H09-002.PMID:19370044.

Howley, E.T., Bassett, D.R., and Welch, H.G. 1995. Criteria for maximal oxygenuptake: review and commentary. Med. Sci. Sports Exerc. 27(9): 1292–1301.doi:10.1249/00005768-199509000-00009. PMID:8531628.

Jones, A.M., and Doust, J.H. 1996. A 1% treadmill grade most accurately reflectsthe energetic cost of outdoor running. J. Sports Sci. 14: 321–327. doi:10.1080/026404196367796, 10.1080/02640419608727717. PMID:8887211.

Mansell, P.I., and Macdonald, I.A. 1990. Reappraisal of the Weir equation forcalculation of metabolic rate. Am. J. Physiol. Regul. Integr. Comp. Physiol.258(27): R1347–R1354. PMID:2360685.

Olson, T.P., Tracy, J.E., and Dengel, D.R. 2003. Validity of a low-flow pneumotachand portable metabolic system for measurement of basal metabolic rate.Med. Sci. Sports Exerc. 35(5): S143. doi:10.1097/00005768-200305001-00795.

Roi, G.S., and Bianchedi, D. 2008. The science of fencing: implications for per-formance and injury prevention. Sports Med 38(6): 465–481. doi:10.2165/00007256-200838060-00003. PMID:18489194.

Roi, G.S., and Pittaluga, I. 1997. Time-motion analysis in women’s swordfencing. In Proceedings of the 4th IOC Congress on Sport Sciences. Oct.22–25; Monaco: 66.

Saltin, B., and Gollnick, P.D. 1983. Skeletal muscle adaptability: significance formetabolism and performance. In Handbook of Physiology, Section 10, SkeletalMuscle, Chapter 19. Edited by L.D. Peachy. American Physiological SocietyBethesda, Md., USA. pp. 555–631.

Weir, J.B. 1949. New methods for calculating metabolic rate with special refer-ence to protein metabolism. J. Physiol. 109: 1–9. PMID:15394301.

328 Appl. Physiol. Nutr. Metab. Vol. 39, 2014

Published by NRC Research Press

App

l. Ph

ysio

l. N

utr.

Met

ab. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y C

alif

Dig

Lib

- I

rvin

e on

12/

17/1

4Fo

r pe

rson

al u

se o

nly.