Introduction

In Latin American and Caribbean countries such as Chile, Mexico and Cuba, the proportion and the number of the population over 60 y has increased steadily. Even though there is an imminent urbanization phenomenon in the region, an important sector of the population still lives in the rural areas with very different lifestyles especially regarding physical activity.^1,2 In rural areas of developing countries, free-living independent individuals over 60 y can be particularly active which may be reflected in their body composition and health status.^3,4

Body composition in this population group can be a sensitive indicator of health or disease. In a previous study in elderly subjects, we found an inverse correlation between physical activity, systolic and diastolic blood pressure and body fat, as well as a direct correlation between body mass index and 2-h glucose level after an oral glucose tolerance test.⁵ Furthermore, body composition is a sensitive indicator of health and nutritional status. Regarding energy expenditure (EE), it is also important to adjust EE for fat-free mass (FFM) or metabolically active tissue. Body composition measurements can reflect the changes in fat or lean tissue after dietary interventions or a physical activity program. The results of physical activity pattern of rural elderly could be taken into account to design physical activity programs, or interventions to improve health conditions in this sector of the population of either rural or urban areas.

There are many techniques to measure body composition, and the accuracy and reliability of these techniques are critical.⁶ Reference techniques such as the measurement of total body water (TBW) by deuterium oxide dilution can be a good option in developing regions that do not have access to the gold standard methodology like the four-compartment models. There is discussion about what techniques are appropriate to assess body composition in the elderly. Today it is recognized that the measurement of TBW or hydrometry is valid in the elderly, as long as the age and gender effects on the hydration factor are taken into account.^7,8,9 The main changes in body composition are: the reduction of FFM^10,11,12 and loss of bone minerals specially in women,^13,14 a decrease in TBW⁷ and an increase in the proportion of body fat accompanied by a more central distribution.^10,15,16,17

With respect to hydrometry as the method to measure FFM in the elderly, the debate had been if age affects the hydration of the FFM component. Ritz.¹⁸ has estimated a hydration factor for elderly subjects looking at independent measurements of FFM and TBW reported in the literature and from his own measurements. He concluded that the hydration of FFM is not affected by healthy aging and not different from those of younger adults, 73.42.4% and 73.22.4%, respectively. There is limited information based on multicompartment methods, such as four-compartment model, to better estimate hydration of the FFM in the elderly and this could be a limitation. Considering these results and the little information about body composition on rural elderly of many countries in Latin America and the Caribbean, we measured TBW by deuterium oxide dilution.

The purpose of this study was to determine body composition in free living and healthy elderly subjects >60 y from rural areas of Chile, Cuba and Mexico using deuterium oxide dilution and bioelectrical impedance (BIA) and to develop and cross-validate a predictive equation for this group of subjects for future use of BIA as a field technique.

Methods

Experimental protocol and subjects

In order to include only healthy subjects, a medical examination, oral glucose tolerance test, routine blood and urine analysis were performed. As a result, 133 out of 182 male and female subjects were selected to participate. Subjects were physically independent as evaluated by the daily activity scale of Katz et al¹⁹ and SENECA²⁰ and free-living conditions in their rural community. In Cuba, the study was done with elderly residents in 'Las Terrazas', Candelaria Pinar del Río; in Chile, 'Municipalidad de Paine'; and in Mexico, 'Ejido La Victoria', 'El Tazajal' and 'San Pedro del Saucito' (rural sectors of the municipality of Hermosillo, Sonora). The studies were conducted during the spring and summer of 1998–2000. In Addition to health screening, anthropometry, BIA and deuterium dilution technique measurements were performed. All measurements were done in the morning between 07.00 and 09.00 h after a minimum 11 h fast.

The subjects were informed of the study and signed the appropriate consent forms complying with the regulations of the Ethical Committee from each participating institution—in Cuba, Instituto de Nutrición e Higiene de los Alimentos, La Habana; in Chile, Instituto de Nutrición y Tecnología de los Alimentos, Universidad de Chile, Santiago; and in México, Centro de Investigación en Alimentación y Desarrollo, A.C., Hermosillo, Sonora.

Health screening

Health status was evaluated by a medical examination. Additionally, an oral glucose tolerance test was included. For this purpose, blood glucose was measured by HemoCue System (HemoCue AB, Angelholm, Sweden).²¹ Diagnostic criteria for type II diabetes were based on WHO recommendations.²² Hemoglobin was measured by HemoCue System (HemoCue AB, Angelholm, Sweden). Also, a multistix 10 SG from Bayer (Bayer Diagnostics, Bayer from Mexico) was used to screen fasting urine for glucose, ketones, bilirubin, pH, proteins, leukocytes and blood. Blood pressure was measured and evaluated according to WHO criteria.²³

Type II diabetics, hypertensive subjects and individuals with cardiovascular history, and presence of edema, as well as the individuals with exercise limiting noncardiac disease (arthritis, peripheral vascular disease, cerebral vascular disease) were excluded from the study. None of the participating subjects were taking diuretics or other medication that could alter body composition, or reported important weight changes during the 6 months prior to the study (2 kg).

Anthropometry

In Mexico, body weight was measured with a digital electronic scale of 150 0.05 kg capacity (ADN FV-150 K, Japan) and standing height by means of a Holtain stadiometer of 205 0.1 cm capacity (Holtain Ltd, Dyfed, UK). In Chile, weight was measured with a SECA scale of 1500.1 kg capacity and standing height by a SECA stadiometer (Modern Measurement Devices, San Antonio, TX, USA). In Cuba, body weight was measured with a digital electronic scale of 1500.1 kg capacity (Soehnle, Soehnle-Frères S.A., Duppigheim) and standing height with a Holtain Anthropometer (Holtain Ltd, Dyfed, UK). Recumbent length was measured in all subjects in supine position using a Holtain Anthropometer (Holtain Ltd, Dyfed, UK) adapted to a measuring bed. Body mass index (BMI) (kg/m²) was calculated based on weight and standing height. Skinfolds were measured using a skinfold caliper (Holtain Ltd, Crymych, Dyfed, Wales, UK) according to the recommendations given by Durnin.²⁴ Waist circumference was measured in supine position at the umbilicus level, using a fiber glass measuring tape (Lafayette Instruments Company Inc., USA). Hip circumference was measured on a standing position at the level of the most prominent part of the gluteus.^25,26 Waist/hip ratio (WHR) was determined from these measurements. Weight measurements were performed with subjects dressed in a bathing suit.

Body composition by total body water

In this study, hydrometry was considered as the reference method. TBW was determined by the isotope dilution technique using deuterium oxide. An accurately weighed dose of 3.0 g of deuterium oxide (99.8 at%; Icon Stable Isotopes, Morion, NY, USA) was given orally to subjects. The dose flask was rinsed with 30 ml of water and this rinse water also ingested to ensure completeness.

Previous to the dosage, the subjects provided a saliva sample for the determination of deuterium background in body water. After a 3.5 h period, another saliva sample was collected. Saliva samples were collected in special plastic vials, doubled sealed and stored at -20°C until analysis by mass spectrometry. To determine the deuterium content, 0.4 ml of saliva in triplicates was centrifuged at 4000 rpm to separate solids from the sample.

Isotopic equilibration

The samples were evacuated previously to the injection of the reference gas, hydrogen (99.9%). Platinum served as the catalyst, either in coated rods (Finnigan Co., Germany) or in grenalla (platinum 5 wt% on alumina; Aldrich Chem Co., USA). The catalyst helps in the exchange of hydrogen and deuterium, until reaching a steady state depending on temperature, in which deuterium content in the gas is similar to that in the original biological sample. The samples remained at a constant temperature during the time to achieve equilibrium.²⁷ The isotopic concentration was measured by isotope-ratio mass-spectrometry (IRMS) in a HYDRA, Europe Scientific (Crewe, UK), in the Laboratory of Energy Metabolism and Stable Isotopes, at INTA, University of Chile. The values obtained were expressed with relation to standards measured against V-SMOW and SLAP (International Atomic Energy Agency, Vienna, Austria). The precision was 0.3–0.4 ppm.

TBW was determined by deuterium oxide dilution. According to this procedure, the volume of the compartment is equal to the amount of tracer added to the compartment divided by the concentration of the tracer in that compartment.²⁸ The TBW was corrected, considering a 4% for deuterium exchange with nonaqueous compartment in the body.²⁹ FFM was calculated from TBW, assuming that FFM has a hydration constant of 0.732.³⁰ Fat mass was estimated from the difference between body mass and FFM.

Bioelectrical impedance

Resistance (R) and reactance (Xc) were measured with a BIA analyzer (Model BIA-101, RJL Systems, Detroit, MI., USA). Resistance and reactance were measured according to the manufacturer's instructions. Volunteers were instructed to lie supine with their hands at their side and with their legs separated. The skin surface was cleaned with ethanol, and the electrodes were placed on the dorsum of the right foot and hand. The system was calibrated periodically with a resistor (500 ) and measurements were continued as long as the instrument readings remained within 4  of the calibration resistor.

Statistical analysis

Data were analyzed using the statistical program NCSS 1997 (Number Cruncher Statistical System for Windows, Kaysville UT, USA). All the results are expressed as meanstandard deviation (s.d.). ANOVA was used to examine the differences between groups (country and sex). In this case, the results are expressed as meanstandard error (s.e.).

The total sample size (n=133) was used in a split-sample internal cross-validation. In this approach, the sample is split randomly into subsamples of equal or similar size. The regression equation is developed using one sample and is cross-validated in the other sample. Model selection was carried out using the 'All Possible Regressions' procedure. This method guarantees to find the model having the largest R² and the smallest standard error of the estimate (s.e.e.).

Mallow's Cp statistic was used to optimize model selection. Multiple regression was used to analyze the relation between FFM as dependent variable with country or ethnicity (Chile=1, Cuba=2 and Mexico=3), sex (women=0 and men=1), body weight (kg), standing height/resistance (cm²/) and reactance () as independent variables. Multicollinearity was analyzed by regression diagnostics using the condition number (CN <30) and the variance inflation factor (VIF <10).

The accuracy of the new BIA equation was tested by the Student's paired t-test (to test the significance of the difference between the hydrometry chosen as the reference method and the new BIA equation) and then by regression procedures. The comparison was considered accurate if the regression between FFM by hydrometry and new BIA equations had a slope not significantly different from 1.0, or an intercept not significantly different from zero.

Precision was assessed by the model R² and the s.e.e. from the estimate from the regression procedures described above. A Bland and Altman plot³¹ of the difference between FFM by hydrometry and FFM by the new BIA equation, against the mean of the two measurements, was used to evaluate bias. This procedure tests the hypothesis that methodological error is randomly distributed across the spectrum of body fat content, as indicated by a nonsignificant regression between technique error and body fat content.

Results

In this study, 182 subjects (60–93 y) from Cuba, Chile and Mexico were evaluated. A total of 26 (14.2%) subjects were diagnosed with type II diabetes and excluded. A further 23 subjects were excluded because of referred pathology upon medical examination.

The results of anthropometry and body composition of 133 healthy elderly are presented. Table 1 depicts the physical characteristics and body composition of the subjects. Both Mexican men and women were 3 cm taller than the Chilean and Cuban subjects. However, the Cubans were the group that had significantly lower BMI (P<0.02) averaging about 22.5 in men and women, as well as less body fat and FFM. The Mexican group also presented a higher WHR but was only significant in women (P<0.01). The Chilean and the Mexican groups presented higher values of skinfolds compared to the Cuban group. The sum of four skinfolds showed an overall tendency to be lower in the Cubans. Also the body fat distribution based upon skinfolds showed that the Cuban men and women had the lowest sum of subscapular and suprailiac skinfolds (26.8 mm), but significantly different only from the Chileans (43.5 mm) (P<0.001). There were no differences in the arm sum distribution (sum of triceps and biceps skinfolds).

Table 1 - Age, physical characteristics and body composition in a group of free-living healthy subjects aged >60 y from rural regions of Cuba, Chile and Mexico.

Full table

A randomized split sample was obtained from the 133 subjects from the three countries. One of them was used to develop the equation for prediction of FFM by BIA and the other sample was used to validate the equation. Physical and body composition characteristics are presented in Table 2. There were no significant difference in physical characteristics and body composition variables in the two randomized split samples. The predictive equation was obtained by stepwise multiple regression (Table 3). The independent variables selected that most accurately estimated FFM were height²/R, weight, sex, reactance and country as shown in Table 3. The R² was 0.944 and the root mean square error (RMSE) was 2.08 kg.

Table 2 - Some physical characteristics of elderly in a split sample.

Full table

Table 3 - Prediction equation for fat-free mass (n=67).

Full table

The means.d. of the prediction of FFM was 44.229.18 vs 44.6210.09 resulting in a mean difference of 0.4073.84 kg. The t-test paired-comparisons of the difference between the measured FFM by dilution and that predicted by new equation based on BIA and anthropometry were found to be not significant (P=0.39) (Table 4). The new BIA equation was cross-validated against the second split sample comparing the individual predicted values with those measured by hydrometry in this independent sample. There was no significant difference with the line of identity. The intercept (=–0.32) was not different from zero (P=0.89) and the slope (=1.06) not different from 1.0 (P>0.90). The R² was 0.86, the RMSE was 3.86 kg of FFM and the pure error was 3.83. The results are shown in Table 3 and depicted in Figure 1.

Figure 1.

Comparison of FFM by hydrometry (deuterium oxide dilution) against BIA using the new equations developed in 67 healthy elderly men and women from Cuba, Chile and Mexico. The dotted line is the line of identity (regression slope=1.0, regression intercept=0).

Full figure and legend (55K)

Table 4 - Measured FFM by TBW and predicted with new equation based on BIA and anthropometry in group B.

Full table

Bland and Altman analysis showed that the bias (expressed as the mean of the difference in FFM measured by hydrometry and estimated from new BIA equation) was small (0.4073.84 kg of FFM); the limits of agreement, defined as mean2 s.d.s, were from -7.3 to 8.1 kg of FFM. Further, there was no significant correlation between the difference and the mean of two measurements by hydrometry and new BIA equation (r=0.24; P=0.051) (Figure 2).

Figure 2.

Differences in FFM measured by hydrometry (deuterium oxide dilution) and estimated by BIA using the new equations developed in 67 healthy elderly men and women from Cuba, Chile and Mexico.

Full figure and legend (81K)

The cross-validated equation was tested in the Cuban, Chilean and Mexican elderly groups and in the three countries together. There were no significant differences in the mean FFM derived from the predicted equation and the values measured by hydrometry in the individual countries or in the three countries as a group (Table 5).

Table 5 - Measured FFM by TBW and predicted with new equation based on BIA and anthropometry in women and men from three countries.

Full table

Discussion

Latin America and other developing regions of the world are becoming steadily urban. However, many people still live in rural or semi-rural areas, and have important differences in health-care, food availability and physical activity. In this study, we evaluated the body composition of free-living elderly men and women, with physical independence from rural areas of Cuba, Chile and Mexico.

It is important to clarify that the rural areas were classified as such, in spite of having improvements in the environmental conditions. Nevertheless, the lifestyle and type of activities carried out by the men and the women of this study were mainly of an agricultural nature. Women were involved in house chores, without the modern conveniences of urban dwellers. A few of the women were also involved in agricultural work. Houses are large and have big yards and gardens to keep. All these factors can have an effect on physical activity, energy expenditure and body composition.¹

The Cuban group showed important differences in body weight, height, BMI and central fat distribution. These data agree with the percent body fat differences obtained by hydrometry, where both Cuban men and women had significant lower levels of body fat (8–10%). There were no anthropological evaluations of ethnicity in the three groups and there must be substantial differences in the admixture with regard to Indian-Spanish in the Mexican and Chilean groups, compared to African-Spanish mixture in Cuba. Parallel studies using the doubly labeled water technique in these groups have shown higher physical activity level in the Cuban group compared to that of the Chilean and Mexican groups.³² Also, we have found that the occupational activities in the Cuban group are of great intensity and the number of hours spent in these activities is higher.³² We have no evidence of energy intake differences between countries for elderly groups. However, limitations in food availability due to the economic situation in recent years in Cuba³³ compared to that of Chile and Mexico cannot be discarded.

Central fat distribution by waist circumference was very different in the Cubans. In this group, the waist in men and women was 81.3 and 80.5 cm, respectively, which was about 8 cm less than the Chilean and Mexican groups. The WHR did not show this as clearly. Further, looking at skinfold measurements as indicators of body fat and distribution, the Cubans showed less accumulation of central fat implying a lesser risk for type II diabetes and cardiovascular disease.³⁴ Additionally, these differences could also be attributed to possible heavier work loads, and to the fact that La Sierra del Rosario in Cuba is a hilly location, where public transport is inexistent and people must continuously go up and down these hills, which is not the case of the rural communities studied in Chile and Mexico.

With respect to design and validation of equation based on BIA and anthropometry, the random split procedure allowed the development of a BIA prediction equation, considering hydrometry as reference method. This new equation was tested and cross-validated in the other half of the subjects. Height²/R, ethnicity or country, body weight, sex and reactance were the variables that best predicted FFM.

There were no differences in physical characteristics between the sample used for developing the equation and the validation sample. This fact indicates that the randomization procedure was adequate. The prediction equation obtained was satisfactory in terms of regression parameters with a high R², and a low s.e.e. as has been shown in other studies with similar equations.^35,36 The results of regression diagnostics showed absence of multicollinearity in the prediction equation.

The estimates were not significantly different with respect to reference method which was the FFM measured by hydrometry in the validation sample. Comparison with the line of identity showed that the new BIA equation was accurate and precise as determined by the model R² and RMSE of the estimate. Agreement between FFM measured by hydrometry and estimated from new equation was good as indicated by a small bias (0.4073.84 kg of FFM); however, the limits of agreement, defined as mean2 s.d.s, were from -7.27 to 8.09 kg of FFM. This interindividual bias is to be expected with indirect body composition techniques like predictive equations which work best at a population level than at the individual level.

The use of bioelectrical impedance and anthropometry in different populations has increased in the last few years.^{35,36,37,38,39} It is especially important in the case of the elderly, where very few equations are available.^35,36,38 In this case, we have developed and validated a prediction equation, based upon free-living elderly subjects from rural areas of Cuba, Chile and Mexico. The body weight range in elderly men was from 42.3 to 104.6 kg and in elderly women was from 37.5 to 87.0 kg. The range in standing height was from 1.50 to 1.77 m and from 1.42 to 1.66 in men and women, respectively and the BMI (with standing height) range, in both women and men of this elderly group from the three countries, was 16.5–35.4. The mean of BMI in equation developed sample and validation sample was 25.1.

When we applied the equation to each country, comparing with the values obtained by hydrometry, the differences ranged from -652 to 514 and +200 g of FFM when applied to the three countries together. None of the differences between the methods reached significance. Proper validation in this case is not possible, because the equation contains an intrinsic part of the sample of each individual country, or for all the three groups. Further research in totally independent samples from each region, with subjects of similar characteristics, would be warranted.

The FFM and fat mass obtained in this study are similar to those values reported by other studies in elderly subjects using a two-compartment model. The Chilean men show similar percent body fat to that of subjects from The Netherlands of similar age, in spite of considerable differences in weight and height.^35,40 The Chilean and Mexican women show similar percent body fat to that of Dutch women, with the same described differences.⁴¹

In the case of elderly men from USA, they had around 5% more of body fat, than Chilean and Mexican elderly men.³⁸ Anthropometry and body composition characteristics of Cuban elderly men and women were different from Chilean and Mexican elderly, and from the elderly from developed countries.

Conclusion

The new BIA equation was accurate, precise and showed a reasonable agreement with the validation sample with no systematic error. The use of this equation could improve the estimates of body composition for the elderly population for these regions, as well as those from other Latin American countries, but caution should be taken when applying such equation at an individual level.

References

1. Esparza J, Fox C, Harper IT, Bennett PH, Schulz LO, Valencia ME, Ravussin E. Daily energy expenditure in Mexican and USA Pimas Indians: low physical activity as a possible cause of obesity. Int J Obes Relat Metab Disord 2000; 24: 55–59. | Article | PubMed | ChemPort |
2. Valencia ME, Bennett PH, Ravussin E, Esparza J, Fox C, Schulz LO. The Pimas Indian in Sonora, México. Nutr Rev 1999; 11: S55–S58.
3. González S, Alemán-Mateo H, Hernández-Triana M, Salazar G, Sánchez V, Basabe B, Valencia ME. Total energy expenditure in elderly subjects from a rural community in Cuba, by questionnaire and resting metabolic rate (factorial method), compared with the double-labelled water method as gold standard [Abstracts]. Ann Nutr Metab Suppl 2001; 45: 354.
4. Torún B. Patrones de actividad física en América Central. In: Peña M, Bacallao J (eds). La Obesidad en la Pobreza. Un nuevo reto para la salud pública, Organización Panamericana de la Salud; 2000. pp 33–43.
5. Alemán-Mateo H, Esparza J, Mauro E. Valencia. Antropometría y composición corporal en personas mayores de 60 años: importancia de la actividad física. Salud Publica Mex 1999; 41: 309–316.
6. Levenhagen DK, Borel MJ, Welch DC, Piasecki JH, Piasecki DP, Chen KY, Flakoll PJ. A comparison of air displacement plethysmography with three other techniques to determine body fat in healthy adults. JPEN 1999; 23: 293–299.
7. Schoeller DA. Changes in total body water. Am J Clin Nutr 1989; 50: 1176–1181. | PubMed |
8. Fuller NJ, Sawyer MB, Laskey MA, Paxton P, Elia M. Prediction of body composition in elderly men over 75 years of age. Ann Hum Biol 1996; 23: 127–147.
9. Goran MI, Toth MJ, Poehlman ET. Assessment of research-based body composition techniques in healthy elderly men and women using the 4-compartment model as a criterion method. Int J Obes Relat Metab Disord 1998; 22: 135–142. | Article | PubMed | ChemPort |
10. Novak LP. Aging, total body potassium, fat-free mass, and cell mass in males and females between ages 18 and 85 years. J Gerontol 1972; 27: 438–443. | PubMed | ChemPort |
11. Forbes GB, Halloran E. The adult decline in lean body mass. Hum Biol 1976; 48: 162–173.
12. Steen B. Body composition and aging. Nutr Rev 1988;46:45–51. | PubMed |
13. Mazess RB. On aging bone loss. Clin Orthop Res 1982; 165: 239–252.
14. Heaney RP, Gallagher JC, Johnston CC, Neer R, Parfitt AM, Whedon GD. Calcium nutrition and bone health in the elderly. Am J Clin Nutr Suppl 1982; 36: 986–1013.
15. Noppa H, Andersson M, Bengtsson C, Bruce A, Isaksson B. Body composition in middle-aged women with special reference to the correlation between body fat mass and anthropometric data. Am J Clin Nutr 1979; 32: 1388–1395.
16. Durnin JV, Womersley J. Body fat assessed from total body density and its estimation from skinfold thickness: measurements on 481 men and women aged from 16 to 72 years. Br J Nutr 1974; 32: 77–97. | Article | PubMed | ISI | ChemPort |
17. Borkan GA, Hults DE, Gerzof SG, Robbins AH. Comparison of body composition in middle-aged and elderly males using computed tomography. Am J Phys Anthropol 1985; 66: 289–295. | PubMed |
18. Ritz P. Body water spaces and cellular hydration during healthy aging. Ann NY Acad Sci 2000; 904: 474–483. | PubMed |
19. Katz S, Downs TD, Cash HR, Grotz RC. Progress in development of the index of ADL. Gerontologist 1970; 10: 20–30. | PubMed | ISI | ChemPort |
20. Euronut SENECA investigators. Life-style: physical activities and activities of daily living. Eur J Clin Nutr Suppl 1991; 45: 139–151.
21. Ashworth L, Gibb I, Alberti KGMM. HemoCue: evaluation of a portable photometric system for determining glucose in whole blood. Clin Chem 1992; 38: 1479–1482. | PubMed |
22. WHO. Diabetes mellitus. Report of a WHO Study Group. Technical Report Series No. 727. World Health Organization, Geneva: 1985.
23. World Health Organization (WHO). Arterial hypertension. Technical Report No. 628. Expert Committee Report, Geneva, 1978.
24. Durnin JVGA. Anthropometric methods for assessing nutritional status. In: Horowitz A, Macfadyen DM, Munro H, Scrimshaw NS, Steen B, Williams TF (eds). Nutrition in the elderly. Oxford University Press: Oxford; 1989. pp 15–32.
25. Fox C, Esparza J, Nicolson M, Bennett PH, Schulz LO, Valencia ME, Ravussin E. Plasma leptin concentrations in Pima Indians living in drastically different environments. Diabetes Care 1999; 22: 413–417. | Article | PubMed | ChemPort |
26. Fox CS, Esparza J, Nicolson M, Bennett PH, Schulz LO, Valencia ME, Ravussin E. Is a low leptin concentration, a low resting metabolic rate, or both the expression of the 'thrifty genotype' ? Results from Mexican Pima Indians. Am J Clin Nutr 1998; 68: 1053–1057.  | PubMed |
27. Scrimgeour CM, Rollo MM, Mudambo MKT, Handley LL, Prosser SJ. A simplified method for deuterium/hydrogen isotope ratio measurements on water samples of biological origin. Biol Mass Spectrom 1993; 22: 383. | Article | PubMed | ISI | ChemPort |
28. Edelman IS, Leibman J. Exchange of water between blood and tissues: characteristics of deuterium oxide equilibration in body water. Am J Physiol 1952; 171: 279–296.
29. Forbes GB. Human body composition. In: Growth, aging, nutrition, and activity. Springer-Verlag; New York; 1987.
30. Pace N, Rathburn EN. Studies on body composition. III: the body water and chemically combined nitrogen content in relation to fat content. J Biol Chem 1945; 158: 685–691. | ISI | ChemPort |
31. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986; 1: 307–310. | PubMed | ISI | ChemPort |
32. Alemán Mateo Heliodoro. Nivel de actividad física, requerimiento de energía y composición corporal en personas activas de la tercera edad, residentes de una región rural de Cuba, Chile y México. Tesis de Doctorado, CIAD, A.C., Hermosillo, México, 2000.
33. Porrata C, Rodríguez-Ojeda A, Jiménez S. La transición epidemiológica en Cuba. In: Peña M, Bacallao J (eds). La Obesidad en la Pobreza. Un nuevo reto para la salud pública, Organización Panamericana de la Salud; 2000. pp 57–72.
34. Bouchard C, Després JP, Mauriege P. Genetic and nongenetic determinants of regional fat distribution. Endocr Rev 1993; 14: 72–93. | Article | PubMed | ISI | ChemPort |
35. Deurenberg P, van der Kooy K, Leenen R, Evers P, Hulshof T. Assessment of body composition by bioelectrical impedance in a population aged >60 y. Am J Clin Nutr 1990; 51: 3–6. | PubMed | ISI | ChemPort |
36. Roubenoff R, Baumgartner RN, Harris TB, Dalla GE, Hannan MT, Economos ChD, Stauber PM, Wilson PWF, Kiel DP. Application of bioelectrical impedance analysis to elderly population. J Gerontol A. 1997; 52: M129–M136.
37. Bussolotto M, Ceccon A, Sergi G, Giantin V, Beninca P, Enzi G. Assessment of body composition in elderly: accuracy of bioelectrical impedance analysis. Gerontology 1999; 45: 39–43. | PubMed |
38. Dupler TL, Tolson H. Body composition prediction equations for elderly men. J Gerontol A 2000; 55: M180–M184.
39. Kyle UG, Genton L, Karsegard L, Slosman DO, Pichard C. Single prediction equation for bioelectrical impedance analysis in adults aged 20–94 years. Nutrition 2001; 17: 248–253. | Article | PubMed | ISI | ChemPort |
40. Deurenberg P, Saris WHM, Voorrips LE, van Staveren WA. The assessment of body composition in the elderly. Age Nutr 1993; 4: 34–39.
41. Broekhoff C, Voorrips MP, Witvoet GA, van Staveren WA, Deurenberg P. Relative validity of different methods to assess body composition in apparently healthy elderly women. Ann Nutr Metab 1992; 36: 148–156. | PubMed |

Acknowledgements

This study was supported by CONACYT (25728-M) from Mexico. We thank the volunteers of this study and Inocencio Higuera C, Juan P Camou, Erik Díaz B, Ramón Figueroa, Soledad Figueroa, Osmany Cienfuegos and José Ramón Porrata Mauri. We are also indebted to Julián Esparza Romero, María Esther Hernández, Ingrid Rolón, Rocío Berlangas, Milagros Marcia Velásquez, Jimmy Hernández, Ana Cristina Gallegos Aguilar, Nayeli Macías Morales, María de los Angeles Sánchez and Alejandrina Cabrera for their technical assistance. We appreciate Silvia Y Moya, Ana María Calderón and Elaine Rush for their valuable suggestions and comments on the manuscript.