Transcription

182Asia Pac J Clin Nutr 2012;21 (2):182-190Original ArticleA digital Calliper for training and study purposesMaria T Restivo PhD1, Teresa F Amaral PhD1,2, Maria F Chouzal PhD1, Celina P Leão PhD3,Rita S Guerra MSc1,2, Elisa Marques PhD4, Joaquim Mendes PhD1, Manuel Quintas PhD1,Jorge Mota PhD41UISPA-IDMEC Campus FEUP, Faculty of Engineering, University of Porto, Rua Dr Roberto Frias, 4200465 Porto, Portugal2Faculty of Nutrition and Food Sciences, University of Porto, Rua Dr Roberto Frias, 4200-465 Porto,Portugal3Scholl of Engineering, University of Minho, Campus de Azurém, Portugal4Research Centre in Physical Activity, Health and Leisure, Faculty of Sport Sciences, University of Porto,Rua Dr Plácido Costa 91, 4200-450 Porto, PortugalA quick and valid method for evaluating percentage body fat is based on the use of skinfold callipers. However,limitations associated to their use and characteristics led the authors to improve a traditional calliper (Harpenden)and to integrate it with a software application. Such a measuring system, LipoTool, is meant to have better accuracy and reliability, including data processing and digital recording at a very low cost. At first, a sample of 49older adults was used to evaluate the performance of LipoTool by comparing its results to those obtained with thetraditional Harpenden calliper. A strong positive association in %BF was achieved. This digital sensing systemwas later improved by incorporating wireless communication between the calliper and the software application,adding other functionalities. The software application works in any computer and is flexible to incorporate newcoming models, linear regressions or new algorithms. This new system was validated against the standard DualEnergy X-Ray Absorptiometry system, using a sample of 40 adults with positive results. This solution is a validand reliable alternative to traditional reference callipers, simplifying the percentage of body fat evaluation andproviding a more effective use in daily practice with less expenditure of time and resources. Its implementedguided procedure turns it into a precious training tool based on a non-invasive, portable device, and not requiringspecial individual preparation. Ongoing activities are focused on the design of a new mechanical structure, withnovel functionalities and for exploring other studies.Key Words: body fat percentage, digital skinfold calliper, training system, wireless data communication,automated processINTRODUCTIONIndividual changes in nutritional status are the mostcommon health problem with high impact in society atindividual, social and economic levels. Malnutrition, obesity or even the co-existence of both is a major worldhealth problem.The nutritional problems affecting industrialized countries are related with high energy food intake, leading tooverweight, obesity and associated co-morbidities.The assessment of percentage body fat (%BF) is highlyrelevant for the evaluation of nutritional status, growthand childhood development, disease impact and sportspractice.1Different methods are available for evaluating malnutrition and obesity. The Dual-Energy X-Ray Absorptiometry (DXA) method is one of the most widely appliedtechniques for %BF studies.2 This method, as othersbased on imaging, relies in very expensive, bulky andinvasive equipment unsuitable for daily diagnosis.1A widespread method, simple and non invasive, forevaluating subcutaneous fat tissue is based on the use oflow cost and portable equipment, such as the skinfoldcalliper (SC),3,4 which allows for quantitative indirectevaluation of fat tissue by measuring skinfolds. The SCprovides a validated diagnostic method, using skinfoldmeasurement for evaluating %BF,1 based on the twocompartment model which divides the body into fat massand fat-free mass.Nevertheless, SC measurements have to be performedby highly trained health technicians following standardprocedures5,6 in order to minimize subjective decisionsand individual variability.All different types of SC accepted as valid devices inthe nutritional area present limitations, making them unatCorresponding Author: Dr Maria T Restivo, UISPA-IDMECCampus FEUP, Faculty of Engineering, University of Porto,Rua Dr Roberto Frias, 4200-465 Porto, Portugal .Tel: 351225081701, 351913503009; Fax: 351222074241Email: [email protected] received 10 June 2011. Initial review completed 16December 2011. Revision accepted 11 January 2012.

LipoTool183Figure 1. LipoTool with a Harpenden SC adapted with a digital dial indicator.tractive and so less used in daily practice.In fact, the needle movement around the analogue display during tests, due to tissue yield under pressure,makes the measurement difficult and leads to subjectivereadings. Also, it is highly subjective to mental counting(2, 3 or 4 seconds) once the calliper jaws touches the foldsurfaces and then the jaws pressure is slowly released.7Manual data registration means time consumption oreven requires a second technician responsible for registering data. The oral transmission process of the measureddata is also subject to errors. Finally, the natural interaction with the individual under assessment influences thesubjective mental evaluation of the time interval and thevisual readout of the rotating needle.To assess body density and %BF, regression equationsare required after the skinfold thickness measurement.There are many equations, developed for different ages,gender, races or specific groups, requiring the selection ofthe most appropriate, based on its validity within thepopulation under study. These calculations are an additional task making the assessment of %BF even moredifficult and time consuming.Those problems and the need to improve measurementaccuracy, reliability, data processing and digital recording,using today’s technology, have led to the development ofa new solution for %BF assessment based on the skinfoldthickness measurement, the LipoTool, comprising amodified SC, Adipsmeter, and a LabVIEW software application, LipoSoft.The Adipsmeter is based on a Harpenden calliper, recognized in the specialized literature,3,5,6,8,9 mechanicallyadapted and electronically instrumented with a digital dialindicator. A data communication interface allows the calliper communication with a software application offeringa user-friendly interface. The application developed formonitoring, digital recording and assessing body composition may also integrate a database. A second iteration ofthe system incorporates digital sensing and a wirelesscommunication system with the software application.The digital sensing accurate system significantly reduces evaluation subjectivity and considerably increaseschecking task efficiency, making it an excellent trainingtool. The features of the system prototype make it a precious tool for training nutritional technicians, trainees forresearch work and for field use in large population studiesdue to its recording capabilities in database.MATERIAL AND METHODSLipoTool systemThe material to be described in this work is an integratedsystem named LipoTool. It consists of an adapted Harpenden skinfold calliper, Adipsmeter, and a software application, LipoSoft, for monitoring, measuring and recording subcutaneous fat tissue in a systematic, accurateand expeditious way. The Harpenden is used because it isa benchmark SC recommended by the International Society for the Advancement of Kinanthropometry (ISAK).8In the first version the digital developed system incorporates a Harpenden calliper mechanically adapted forallowing the automatic skinfold thickness measurement,by replacing the analogue dial indicator by a digital one.It communicates with the PC software applicationthrough a cable interface (Figure 1).10Digital information eliminates reading errors that mighthappen when measuring with a rotating needle movingaround the analogue scale, as in the original Harpenden.The needle movement has not constant speed, varyingaccording to the body tissues composition. This makesmeasurement difficult during the measuring time intervalof 2 to 4 seconds, selected in the software application,after pressuring the tissue, according to established procedure. The counting time, the needle position readoutand its manual registration, all traditionally subjectivelyperformed by the technician, are now done digitally andautomatically.Two miniature switches have been added to the calliperto allow user interactivity with the computer application.The software application has been developed in LabVIEW and presents a user-friendly graphical interface. Itoffers an intuitive guidance through the established procedure for %BF assessment, avoiding subjective measurement and calculation errors – mentally counting theprotocol time interval and reading the position of themoving needle in an analogue display. A visual graphicalevaluation of the measurement is monitored. The %BFvalue is available after performing the skinfold thicknessmeasurement and according to the selected equationsfrom the specialized literature offered by the software.6,8It also integrates a database allowing individual data recording and updating. This application has six sequentiallevels: entry, individual search, new individual identification, individual data update, measurement and processingdata panels, through the selection of equations for bodycomposition determination.

184MT Restivo, TF Amaral, MF Chouzal, CP Leão, RS Guerra, E Marques, J Mendes, M Quintas and J MotaThe LipoSoft application was then adapted to the wireless data communication protocol requirements, namedLipoSoft ZigBee. This new application has also otherstructural improvements. The access to the individualgraphical history record for tracking the evolution ofhis/her body composition has been added.At the data processing level, other features were introduced. So, according to the selected method, the technician will be notified if any type of skinfold measurementhas not been carried out, indicating which ones shouldadditionally be measured. According to individual history,his/her %BF maximum values, a first analysis of the present result is possible, even for a less experienced technician. At the end of the tests a printed report is produced.For safeguarding data an "encrypt" function was created.Figure 2. LipoTool integrating sensing and wireless communication system.At the individual search level the LipoSoft applicationlooks for the personal details in the database, according tothe individual process number introduced. If no information is found, the application switches to the new individual identification panel. If the individual is already registered, the update panel opens and new tests are performedaccording to the selected method. If measurements are notadjusted to the selected method, the technician is informed and the application automatically redirectsher/him to the adequate menu level. The LipoSoft notablyreduces evaluation subjectivity, increases the efficiencyand is easily updatable.However, negative aspects were outlined: the solutionis heavier than the traditional calliper and the interfacecable causes handling unbalance. A new iteration tries toovercome those features, looking for a new digital, miniaturized, light and low cost sensing system. In parallel, thenew prototype should integrate a wireless data transmission solution for the computer (Figure 2), making thesystem easy to use in daily clinical practice and in largepopulation samples.The digital dial indicator incorporated in the calliperwas replaced by a miniaturized encoder, ensuring a contactless skinfold thickness measurement and loweringmaintenance costs. This provides the device with a resolution (0.35mm) three times better than that typically required by these systems (better than 1mm). The new finalprototype weights less than the original Harpenden.To overcome undesired effects from the data transmission cable (causing important interference with the individual position relatively to the computer and introducingunbalance in the technician manipulation) the ZigBeetechnology was used for the wireless communication system. The wireless receiver is connected to a USB computer input. A rechargeable battery is incorporated forpowering the system with 9h autonomy, throughUSB/USB, car battery, USB/220V A.C. connections,flexible for field use.Performance evaluationThe performance of the first version of LipoTool wasevaluated through a cross-sectional study using a sampleof 49 elderly Caucasian subjects, aged from 61 to 92years, on a voluntary bases. This study was conducted inaccordance with the internationally agreed ethical principles for the conduct of medical research (Declaration ofHelsinki). All the participants were informed of the studypurposes as well as the different procedures. Verbal informed consent from all the subjects was witnessed andformally recorded. This elderly group reports participating in some kind of physical exercise, at least twice aweek: 37%, 41%, 12% and 10% respectively two, three,four and five or more times per week.The %BF was estimated based on skinfold measurements performed by the original Harpenden, used as thetest reference, and with the Adipsmeter. Both calliperswere pre-calibrated.The height and weight were measured according totechniques and procedures described in the literature.11Height was measured, with subjects standing barefoot,with a stadiometer (Seca 708 model and resolution of0.01 m). The weight was measured with a scale of 0.5 kgof resolution with subjects wearing underwear and anexam gown. The BMI was then estimated using these twocharacteristics, BMI weight (kg)/(height (m))2.The measurement of skinfold in the triceps, biceps,subscapular and iliocristale parts of the human body weremade on the right side and in triplicate,9 with both callipers, by an Anthropometrist accredited by ISAK. Bodydensity and %BF were estimated using the mean of thethree measured values for each calliper. Body density wasestimated using the equations of Visser et al.12 and %BFwas estimated using the equation of Brozek et al.13 Themean %BF was estimated based on Harpenden measurements and from LipoTool results.The total error (TE) was calculated as:TE (y yL ) n2Hwhere yH are the estimated values based on Harpendenmeasurements and yL are the values coming from LipoTool.

LipoTool185Table 1. Characterization of the two samples used (1: elderly people and 2: adults samples) (mean sd†)Age (years)Height (m)Weight (kg)BMI (kg/m2)†Females(n1 34)(n2 25)72.9 5.5740.9 11.51.53 0.061.58 0.0561.4 7.1758.3 9.426.3 3.4423.4 3.7Males(n1 15)70.4 4.091.68 0.0575.1 7.7726.6 2.18Total(n2 15)24.7 2.91.77 0.0578.4 12.525.1 4.1(n1 49)72.1 5.251.58 0.0965.6 9.6826.4 3.09(n2 40)34.8 12.11.65 0.1165.9 14.424.0 3.9Sd-Standard deviationTable 2. Association between %BF estimated by the LipoTool and by the Harpenden - elderly sample.Total (n 49)HarpendenLipoToolHarpenden–LipoToolFemales (n 34)HarpendenLipoToolHarpenden–LipoToolMales (n 15)HarpendenLipoToolHarpenden–LipoTool†Mean sdrr2InterceptSlopesx-†TE‡38.0 5.8537.9 5.830.082 0.460.997*0.9940.190**0.993*0.4580.4641.7 1.8341.5 2.000.127 0.510.969*0.939-2.60**1.06*0.5020.3629.6 1.1329.6 1.19-0.021 0.300.967*0.935-0.386**1.01*0.3140.29sx-, Standard error estimation; ‡TE, Total error; *p 0.01; **p 0.05.The correlation coefficient, r, was used to evaluate theassociation between the values of %BF estimated by thetwo systems. To describe the agreement between themeasurements obtained by both callipers, a Bland-Altmanplot was done.14In a second stage, the performance of the new integrated system LipoTool incorporating a wireless communication protocol was also tested with a sample of 40Caucasian adults aged from 22 to 58 years. This studywas conducted in accordance with the internationallyagreed ethical principles for the conduct of medical research (Declaration of Helsinki).Most adults declared practicing some form of physicalexercise (72.5%, being 45% men and 27.5% women) distributed as follows: 27.5%, 10%, 10% and 25% respectively two, three, four and five or more times per week.The %BF was estimated with the original Harpendenand the LipoTool. Both callipers were previously calibrated. These measurements were performed and compared with parallel DXA measurements, as a gold standard system. The height and weight were measured andBMI calculated as previously described. The triceps, subscapular, suprailiac and thigh measurements were obtained with the two callipers. The %BF was estimatedusing the equations of Peterson et al.15The correlation coefficient, r, was used to evaluate theassociation between the values of %BF estimated by thetwo systems, and the values estimated by DXA (%BFDXA). The analysis was made considering the total sample and by gender.RESULTSThe descriptive statistic of the elderly sample is shown inTable 1. The elderly men were 13.7 kg heavier and 0.15m higher than the elderly women. However, these meandifferences (p 0.001) were not reflected in the mean val-ues of BMI for these two groups (women 26.3 kg/m2,men 26.6 kg/m2), where the variances were consideredhomogenous. The difference between those values wasnot significant.A strong positive association between the values of the%BF, with skinfold measurements made with both callipers, was achieved (r 0.997; p 0.001). The slope of theregression line is near the unit (0.993), strengthening theagreement between the %BF for both systems.Considering error estimations, it is possible to concludethat both the standard error (sx-) and the Total Error TEpresenting values less than 0.5%, meaning that the LipoTool, and so the Adipsmeter, can be accepted to be as accurate as the original Harpenden (Table 2).As observed in Figure 3, the differences (%BF Harpenden-%BF LipoTool) tend to increase with the increase ofmean %BF values. This type of variability appears with theintrinsic difficulty of measuring larger skinfolds and limitation of the equipment. In fact, with increase of fat thickness the skinfold is under lower pressure due to the nonparallelism between the jaw surfaces of the skinfold calliper.16This analysis proved that the new system performanceis similar to the original Harpenden. This means thatchanges on the original mechanical structure of Harpenden did not interfere with the main Harpenden characteristics, adding new capabilities associated with the intrinsic digital condition.The descriptive statistic of sample two is displayed inTable 1. The adult men were 20.1 kg heavier and 0.19 mhigher than the adult women. However, these mean differences (p 0.001) were not reflected in the mean valuesof BMI for these two groups (women: 23.4 kg/m2, men:25.1 kg/m2), where the variances were considered homogenous. Notice that the age range of the male samplewas from 22 to 33 years old whereas the range for the

186MT Restivo, TF Amaral, MF Chouzal, CP Leão, RS Guerra, E Marques, J Mendes, M Quintas and J MotaFigure 3. Measurement error Bland-Altman plot for the Harpenden and Adipsmeter v0 %BF values. (---) mean and 95% limits of agreement (elderly sample).Table 3. Association between the %BF estimated by LipoTool and Harpenden against the DXA (adults sample).Mean sdrr2InterceptSlopesx-†TE‡Total (n 40)DXAHarpendenLipoTool27.4 8.5929.5 8.2929.2 .712.63-3.463.16Females (n 25)DXAHarpendenLipoTool32.0 5.634.0 6.033.8 912.84-3.493.30Males (n 15)DXAHarpendenLipoTool19.8 7.221.9 5.721.4 091.92-3.412.91†sx-, Standard error estimation; ‡TE, Total error; * p 0.01.female sample was higher, from 22 up to 58 years (thiscan be confirmed by the obtained higher value for thestandard deviation, 11.5 years).The average %BF estimated by DXA is within thenormal range for both men and women.17 The differentvalues obtained for the %BF DXA dispersion, was notsignificant.There was a strong correlation between the average%BF estimated by DXA and the average %BF estimatedeither with Harpenden (%BF Harpenden), r 0.946 (p 0.001) or with LipoTool (%BF LipoTool), r 0.952(p 0.001) (Table 3). However, the female and male correlations were significantly different when analysed separately, r 0.887 and 0.948, respectively, with LipoTool (p 0.001), showing a lower correlation in women than inmen. Similar behaviour was obtained with Harpenden(Table 3).The regression line slope is near one for Harpendenand LipoTool, 0.914 and 0.937 respectively, strengthening the agreement between them with comparison to thereference DXA.The values of TE obtained slightly higher than 3 couldbe accepted (between 3 and 4% fat).18Bland-Altman shows %BF values by gender forDXA/LipoTool [Figure 4 (a)] and, for DXA/Harpenden[Figure 4 (b)].The limits of agreements for all samples and by genderare summarized in Table 4. The differences (%BF DXA%BF Harpenden) according to Bland & Altman14 plotsshow that %BF Harpenden overestimate %BF DXA (dif-

LipoTool1874a.4b.Figure 4. Bland-Altman plot for the %BF values (a) Adipsmeter v1/DXA; (b) Harpenden/DXA. (---) mean and 95% limits of agreement(adult sample).ference mean value -2.12) and presents higher agreementlimits (only 1 in 40 adults was outside the agreement limits). Similar behaviours occur for the %BF DXA-%BFLipoTool, however, with a mean value closer to the zerovalue (-1.77).Analysing the differences between DXA estimatedvalues and those estimated by both systems, it can be observed that the LipoTool introduces an improvement, inaverage, of 0.35%. LipoTool measures slightly lower%BF when compared with the original Harpenden, and soslightly closer to DXA, as presented in the results of Table 3 (means’ difference are not statistically significant, p 0.05). A deeper discussion of results can be found inAmaral et al.19DISCUSSIONAlthough in the present study all measurements weretaken by a trained technician, the innovations introducedare supplying considerable aids mainly for technicianswith a lower degree of practice, not requiring readingagility, objectivity in time counting and additional skillsin combining these tasks with a simultaneous interactionwith the individual.The simplicity, quickness, reliability and accuracy ofthe skinfold measurement process, as well as the totaldata recording in a database and, finally, the instantaneous %BF evaluation are the strong features of the system.Its capability for guiding technicians of low proficiencythrough the procedure also makes it a valuable trainingsystem.

188MT Restivo, TF Amaral, MF Chouzal, CP Leão, RS Guerra, E Marques, J Mendes, M Quintas and J MotaTable 4. Mean differences 95% limits of agreement between SCs and DXA %BF estimations, and between the twosystems used (mean sd†) – adults sample.%BF DXA-%BF Harpenden%BF DXA-%BF LipoTool%BF Harpenden-%BF LipoTool†Females (n 25)-2.07 2.87-1.84 2.790.23 0.37Males (n 15)-2.19 2.71-1.64 2.480.55 0.49Total (n 40)-2.12 2.77-1.77 2.650.35 0.45sx-, Standard error estimation.Moreover, it is possible to monitor step by step themeasurements quality, accepting or rejecting them if convenient. The requirements concerning the techniciantraining degree are certainly less demanding.At research level, the system makes available its database for later studies. It also permits easy introduction ofnew algorithms for %BF evaluation or other functionalities.The main negative aspects of the first solution, unbalanced handling caused by the interface cable for communication and the calliper weight, was overcome by theintroduction of a wireless communication system and thedecrease of original calliper weight.The second solution offers better performance, increasing the simplicity of use and handling, presenting a betterweight balance and even weight reduction.Its battery autonomy is of 9h with a very flexible andeasily rechargeable system. Considering the digital sensorization and the data wireless communication systemadded to the traditional calliper, the cost/benefit ratio isattractive. In fact, three working days of a traditional datarecording technician will be enough to support the cost ofadded electronics to adapt the commercialized Harpendento the new system.The use of SC for %BF assessment is based on simpleand well established procedures, although with wellknown errors and limitations, particularly some of thesubjective type. The increase in accuracy offered by theLipoTool strengthened the use of SC as a valid and morereliable method. In fact, it removes from the evaluationperformed with the original SC, the inherent subjectivitynow overcome by digital measurement, automated procedure and wireless data transmission combined capabilities.The integration of a database is also valuable. The finalprintable report is another important issue for summarizing test results. It is the authors' belief that such system,of low cost, will contribute to simplifying the procedureof %BF evaluation, not only by promoting their larger usein clinical practice and in other applications, but also bysupplying database resources and further research in thearea. This system for measuring skinfold thickness hasbeen patented - PT 103721.20From the above discussion the LipoTool performanceremains equivalent to that of the Harpenden calliper, exhibiting an improvement of 0.35% when referenced to theDXA. This result has no clinical significance. However, itpoints out the interest of further development to improvethis difference and consequently to gain clinical relevance.ConclusionsThe LipoTool, as the traditional SC, is a simple, portable,non-invasive and low cost diagnostic system, based on askinfold method, for evaluating the %BF. Although itsperformance is very similar to that of the traditional Harpenden calliper, according to the present study its resultsbecame closer to those obtained by DXA system. Additionally, the wireless communication between the digitally sensorized calliper and a precious step by stepguided software application, significantly reduces the%BF evaluation subjectivity and increases the task efficiency and quality control avoiding the involvement ofhighly trained technicians; and is an important tool fortraining purposes. In fact, many of its features were devised from its use for training purposes. In general thesystem represents an upgrade of the skinfold measurement procedure for better reliability, sensitivity and accuracy. Integration of a database guarantees the access tothe individual historic data. Therefore, the software implementation of the measurement protocol is easily adjusted to any updates from future studies, which represents to the authors one of the most relevant aspect forother future research. Finally, this system could be easilyadjusted to be an assessment tool for teaching purposes.However it faces serious constraints coming from theuse of the original Harpenden mechanical structure: themeasuring range is not adjustable for obese individuals;the pressure between jaws is not constant as it is supposedto be (in theory); the jaws surfaces are not parallel to theskinfold in all the opening range; the structure has inherent weight and strongly determines the final cost.Although not statistically significant, analysis showedLipoTool exhibits better agreement when compared withboth original Harpenden and DXA. In any case, the SEEis less than 3% which resulted in the new system beingaccepted as accurate,18 and always slightly lower for theLipoTool than for the Harpenden. So, the developed system facilitates the measurement procedure, data collectionin a database system, availability for later use makes thesystem important for training purposes, without increasing random error.In order to overcome most of those problems a newprototype has been under development with a completelynew structure of novel characteristics and better digitalsensing and data wireless communication in order to get alower price and also reliable hardware. At the present itneeds to be validated.Other additional benefits could possibly come fromfuture studies for producing algorithms which could introduce simplicity to the numerous equations needed forthe calculations according to the individual age, genderand race.ACKNOWLEDGEMENTSThe present studies and developments are due to normal active-

LipoToolties in UISPA, IDMEC-Pólo FEUP, CIAFEL and it was alsosupported by University of Porto. E. M. was supported by adoctoral grant from FCT-MCTES and R. S. G. was supportedby a research grant from FCT-MCTES, within IDMEC – PóloFEUP.AUTHOR DISCLOSURESAll authors declared there is no conflict of interest.REFERENCES1. Heymsfield SB, Lohman TG, Wang Z, Going SB. HumanBody Composition, 2nd ed. Champaign, IL: Human Kinetics;2005.2. Heymsfield SB. Development of imaging methods to assessadiposity and metabolism. Int J Obes (Lond). 2008;32:S7682.3. Jelliffe DB, Jelliffe EFP, Zerfas A, Neumann CG. Community Nutritional Assessment with Special Reference to LessTechnically Developed Countries. New York: Oxford University Press; 1989.4. Krämer HJ, Ulmer HV. Two-second standardization of theHarpenden Caliper. Eur J Appl Physiol Occup Physiol. 1981;46:103-4.5. Fidanza F. Nutritional Status Assessment. London, UK:Chapman & Hall; 1991.6. Gibson RS. Principles of Nutritional Assessment, 2nd ed.New York: Oxford University Press; 2005.7. Lee RD, Nieman DC. Nutritional Assessment, 4th ed. Boston:McGraw Hill Higher Education; 2007.8. Norton K, Olds T. Anthropometrica, Sidney, Australia:University of New South Wales Press; 1996.9. Marfell-Jones M, Olds T, Stewart A, Carter L. InternationalStandards for Anthropometric Assessment. Potchefstroom,South Africa: International Standards for AnthropometricAssessment; 2006.10. Restivo MT, Quintas MR, Chouzal MF, Mendes J, AmaralT and Maia M. Virtual instrument for monitoring, digital recording and assessing body composition, CD-Rom Proceedings of the Internati

penden skinfold calliper, Adipsmeter, and a software ap-plication, LipoSoft, for monitoring, measuring and re-cording subcutaneous fat tissue in a systematic, accurate and expeditious way. The Harpenden is used because it is a benchmark SC recommended by the International