Lifecycle nutrition and cardiovascular health: the aged.

As the world's population ages, cardiovascular health becomes increasing important. The ageing process gradually leads to a decline in the structure and function of the cardiovascular system. Other factors associated with ageing can hasten this decline, for instance, lifestyles that have become more sedentary. Additionally, the prevalence of hypertension, dyslipidaemia and diabetes, major risk factors for cardiovascular disease increase with age. Nutrition throughout the lifecycle can help prevent the development of these conditions and appropriate food habits instigated later in life can improve the management of these conditions and their impact on cardiovascular health.

Candidate foods in the asia-pacific region for cardiovascular protection: nuts, soy, lentils and tempe.

Cross-cultural and intervention studies increasingly point the way to seeds like nuts, soy and lentils, and products of them like tempe being cardioprotective. Soy and its products (like tofu, tempe, soy drinks and soy desserts) are historically and currently some of the most important foods in the Asian region where diets remain predominantly plant-based. The mechanisms by which these seeds may protect populations against cardiovascular disease are several. They include the minimisation of classical risk factors like positive energy balance leading to obesity, hypertension. dyslipidemia and insulin resistance with hyperglycaemia. However, in addition, they provide compounds like n-3 fatty acids, isoflavones and arginine which are only now recognised for their ability to optimise other pathways which connect lifestyle to cardiovascular disease--like oxidant status, vascular reactivity and myocardial electrical stability and proneness to dysrhythmia. Thus, once an Asian food culture changes on its emphasis on these plant foods, it may place its consumers at cardiovascular risk.

Evidence-based nutrition and cardiovascular disease in the Asia-Pacific region.

The Asia-Pacific region is undergoing a major change in both food and health patterns, with a connection between the two more than likely. Evidence for certain traditional Asia-Pacific foods as protective agents against chronic non-communicable disease and cardiovascular disease (CVD), in particular, is growing at a time when their usage diminishes. The nature of the evidence to establish relevant Asia-Pacific food-health linkages will include randomised placebo-controlled clinical trials, but is much more extensive and meaningful. Okinawans have probably achieved one of the most successful food cultures from a health point of view and serve as a reference point for the Asia-Pacific region. The expert working party has produced, in November 2000, the 'Okinawan Recommendations on Nutrition and CVD in the Asia-Pacific region'.

Can skin wrinkling in a site that has received limited sun exposure be used as a marker of health status and biological age?

OBJECTIVES: to determine if skin wrinkling in a site that had received limited sun exposure may be a marker of health status and biological age. DESIGN: population-based, cross-sectional study. PARTICIPANTS: we evaluated the health status of representative samples of elderly Greek-born people living in Melbourne, Greeks living in rural Greece, Anglo-Celtic Australians living in Melbourne and Swedes living in Sweden. We carried out microtopographic assessment of their skin and measured plasma dehydroepiandrosterone concentrations. METHODS: we derived activities of daily living, well-being, memory and general health status scores from a cross-cultural questionnaire. We measured skin wrinkling using cutaneous microtopographic methods and plasma dehydroepiandrosterone by enzyme immuno-assay. RESULTS: skin wrinkling was positively correlated with age (r(s)=0.27, P<0.0001) and negatively with body mass index (r(s)=-0.19, P<0.0001). Therefore, all analyses were controlled for these variables. Plasma dehydroepiandrosterone was higher in smokers than non-smokers (2.86 vs 2.08; P<0.001) and men had significantly higher plasma dehydroepiandrosterone than women (2.74 vs 1.69; P<0.0001). In the pooled data, skin wrinkling was negatively associated with general health score (r(s)=-0.13, P<0.01) and activities of daily living score (r(s)=-0.14, P<0.05) after controlling for age, body mass index and smoking. These associations were more pronounced in women. Finally, those with the least skin wrinkling had the highest dehydroepiandrosterone level (r(s)=-0.12, P=0.06) after adjusting for age, smoking and sex. CONCLUSION: skin wrinkling in a site with limited sun exposure might be used as a marker of health status and, to some extent, biological age--particularly for women.

Skin wrinkling: can food make a difference?

OBJECTIVES: This study addressed whether food and nutrient intakes were correlated with skin wrinkling in a sun-exposed site. METHODS: 177 Greek-born subjects living in Melbourne (GRM), 69 Greek subjects living in rural Greece (GRG), 48 Anglo-Celtic Australian (ACA) elderly living in Melbourne and 159 Swedish subjects living in Sweden (SWE) participating in the International Union of Nutritional Sciences IUNS "Food Habits in Later Life" study had their dietary intakes measured and their skin assessed. Food and nutrient intakes were assessed using a validated semi-quantitative food frequency questionnaire (FFQ). Skin wrinkling was measured using a cutaneous microtopographic method. RESULTS: SWE elderly had the least skin wrinkling in a sun-exposed site, followed by GRM, GRG and ACA. Correlation analyses on the pooled data and using the major food groups suggested that there may be less actinic skin damage with a higher intake of vegetables (r(s)=-0.31, p<0.0001), olive oil (r(s)=-0.29, p<0.0001), fish (r(s)=-0.24, p<0.0001) and legumes (r(s)=-0.16, p<0.0001), and lower intakes of butter (r(s)=0.46, p<0.0001) and margarine (r(s)=0.24, p<0.001), milk products (r(s)=0.16, p<0.01) and sugar products (r(s)=0.12, p<0.01). Similar findings were obtained using regression analyses, except fish was no longer significant; 32% of the variance for actinic skin damage was predicted by six out of the ten major food groups. In particular, a high intake of vegetables, legumes and olive oil appeared to be protective against cutaneous actinic damage (collectively explaining 20% of the variance); a high intake of meat, dairy and butter appeared to be adverse (explaining <5% of the variance). Prunes, apples and tea explained 34% of variance amongst ACA. CONCLUSION: This study illustrates that skin wrinkling in a sun-exposed site in older people of various ethnic backgrounds may be influenced by the types of foods consumed.

Laminin inhibition of beta-amyloid protein (Abeta) fibrillogenesis and identification of an Abeta binding site localized to the globular domain repeats on the laminin a chain.

beta-Amyloid protein (Abeta) is a major component of neuritic plaques and cerebrovascular amyloid deposits in the brains of patients with Alzheimer's disease (AD). Inhibitors of Abeta fibrillogenesis are currently sought as potential future therapeutics for AD and related disorders. In the present study, the basement membrane protein laminin was found to bind Abeta 1-40 with a single dissociation constant, K(d) = 2.7 x 10(-9) M, and serve as a potent inhibitor of Abeta fibril formation. 25 microM of Abeta 1-40 was incubated at 37 degrees C for 1 week in the presence of 100 nM of laminin or other basement membrane components, including perlecan, type IV collagen, and fibronectin to determine their effects on Abeta fibril formation as evaluated by thioflavin T fluorometry. Of all the basement membrane components tested, laminin demonstrated the greatest inhibitory effect on Abeta-amyloid fibril formation, causing a ninefold inhibition at 1 and 3 days and a 21-fold inhibition at 1 week. The inhibitory effects of laminin on Abeta fibrillogenesis occurred in a dose-dependent manner and were still effective at lower concentrations. The inhibitory effects of laminin on Abeta 1-40 fibril formation was confirmed by negative stain electron microscopy, whereby laminin caused an almost complete inhibition of Abeta fibril formation and assembly by 3 days, resulting in the appearance of primarily amorphous nonfibrillar material. Laminin also caused partial disassembly of preformed Abeta-amyloid fibrils following 4 days of coincubation. Laminin was not effective as an inhibitor of islet amyloid polypeptide fibril formation, suggesting that laminin's amyloid inhibitory effects were Abeta-specific. To identify a potential Abeta-binding site(s) on laminin, laminin was first digested with V8, trypsin, or elastase. An Abeta-binding elastase digestion product of approximately 120-130 kDa was found. In addition, a approximately 55 kDa fragment derived from V8 and elastase-digested laminin interacted with biotinylated Abeta 1-40. Amino acid sequencing of the approximately 55 kDa fragment identified a conformationally dependent Abeta-binding site within laminin localized to the globular repeats on the laminin A chain. These studies demonstrate that laminin not only binds Abeta with relatively high affinity but is a potent inhibitor of Abeta-amyloid fibril formation. In addition, further identification of an Abeta-binding domain within the globular repeats on the laminin A chain may lead to the design of new therapeutics for the inhibition of Abeta fibrillogenesis.

Body fat distribution and lipids profile of elderly in southern Jakarta.

A cross-sectional study on 222 elderly subjects was carried out at Health Centers in 10 subdistricts in south Jakarta, Indonesia. The anthropometric data (body mass index (BMI), body fat distribution), fasting blood glucose, serum total cholesterol, low density lipoprotein (LDL) cholesterol and triglycerides were assessed. There was a positive correlation between body fat distribution and serum lipid concentration (total cholesterol, LDL cholesterol and triglycerides). Body fat distribution appears to be a stronger determinant of serum lipids than BMI.

The sulfate moieties of glycosaminoglycans are critical for the enhancement of beta-amyloid protein fibril formation.

Our previous studies have demonstrated that perlecan and perlecan-derived glycosaminoglycans (GAGs) not only bind beta-amyloid protein (Abeta) 1-40 and 1-42, but are also potent enhancers of Abeta fibril formation and stabilize amyloid fibrils once formed. However, it was not determined which moieties in perlecan heparan sulfate GAG chains may be responsible for the observed effects and whether other GAGs were also capable of a similar enhancement of Abeta fibril formation as observed with perlecan GAGs. In the present study, thioflavin T fluorometry (over a 1-week period) was used to extend our previous studies and to test the hypothesis that the sulfate moiety is critical for the enhancing effects of heparin/heparan sulfate GAGs on Abeta 1-40 fibrillogenesis. This hypothesis was confirmed when removal of all sulfates from heparin (i.e., completely desulfated N-acetylated heparin) led to a complete loss in the enhancement of Abeta fibrillogenesis as demonstrated in both thioflavin T fluorometry and Congo red staining studies. On the other hand, removal of O-sulfate from heparin (i.e., completely desulfated N-sulfated heparin), and to a lesser extent N-sulfate (i.e., N-desulfated N-acetylated heparin), resulted in only a partial loss of the enhancement of Abeta 1-40 fibril formation. These studies indicate that the sulfate moieties of GAGs are critical for enhancement of Abeta amyloid fibril formation. In addition, other sulfated molecules such as chondroitin-4-sulfate, dermatan sulfate, dextran sulfate, and pentosan polysulfate all significantly enhanced (greater than twofold by 3 days) Abeta amyloid fibril formation. These latter findings indicate that deposition and accumulation of other GAGs at sites of Abeta amyloid deposition in Alzheimer's disease brain may also participate in the enhancement of Abeta amyloidosis.

Thiamine deficiency is prevalent in a selected group of urban Indonesian elderly people.

This cross-sectional study involved 204 elderly individuals (93 males and 111 females). Subjects were randomly recruited using a list on which all 60-75 y-old-people living in seven sub-villages in Jakarta were included. The usual food intake was estimated using semiquantitative food frequency questionnaires. Hemoglobin, plasma retinol, vitamin B-12, red blood cell folate and the percentage stimulation of erythrocyte transketolase (ETK), as an indicator of thiamine status, were analyzed. Median energy intake was below the assessed requirement. More than 75% of the subjects had iron and thiamine intakes of approximately 2/3 of the recommended daily intake, and 20.2% of the study population had folate intake of approximately 2/3 of the recommended daily intake. Intakes of vitamins A and B-12 were adequate. Biochemical assessments demonstrated that 36.6% of the subjects had low thiamine levels (ETK stimulation > 25%). The elderly men tended to have lower thiamine levels than the elderly women. The overall prevalence of anemia was 28.9%, and the elderly women were affected more than the elderly men. Low biochemical status of vitamins A, B-12 and RBC folate was found in 5.4%, 8.8 % and 2.9% of the subjects, respectively. Dietary intakes of thiamine and folate were associated with ETK stimulation and plasma vitamin B-12 concentration (r = 0.176, P = 0.012 and r = 0.77, P = 0.001), respectively. Results of this study suggest that anemia, thiamine and possibly vitamin B-12 deficiency are prevalent in the elderly living in Indonesia. Clearly, micronutrient supplementation may be beneficial for the Indonesian elderly population living in underprivileged areas.

Are the advantages of the Mediterranean diet transferable to other populations? A cohort study in Melbourne, Australia.

A prospective cohort study, involving 141 Anglo-Celts and 189 Greek-Australians of both sexes aged 70 years or more, was undertaken in Melbourne, Australia. The objective was to evaluate whether adherence to the principles of the Mediterranean diet affects survival of elderly people in developed non-Mediterranean countries. Diet was assessed using an extensive validated questionnaire on food intake. A one unit increase in a diet score, devised a priori on the basis of eight key features of the traditional common diet in the Mediterranean region, was associated with a 17% reduction in overall mortality (two-tailed P value 0.07). Mortality reduction with increasing diet score was at least as evident among Anglo-Celts as among Greek-Australians. We conclude that a diet that adheres to the principles of the traditional Mediterranean diet is associated with longer survival among Australians of either Greek or Anglo-Celtic origin.

Chronic energy deficiency and relative abdominal overfatness coexist in free-living elderly individuals in Ho Chi Minh City, Vietnam.

As part of the Cross-Cultural Research on Nutrition of Older Subjects (CRONOS) project, the aim of this cross-sectional study was to observe the nutritional status and body composition of free-living elderly and middle-aged people in a low-income area of Ho Chi Minh City, Vietnam. Anthropometric data (weight; height; arm span; four skinfold thicknesses; left upper mid-arm, abdomen, hip and calf circumferences) were collected from 50 Vietnamese men and 50 Vietnamese women aged between 35 and 44 years and 50 men and 50 women aged between 60 and 74 years who lived in Village 2, District 4, Ho Chi Minh City. The prevalence of chronic energy deficiency (CED) in the elderly men (BMI = 19.9 ± 2.8 kg/m2) was higher (52 vs 34%) than that in the middle-aged men (BMI = 19.2 ± 3.4 kg/m2). Fat-free mass and calf circumference of the elderly were lower than those of the middle-aged (p<0.001). However, the Vietnamese elderly had higher body fat content, higher abdomen-to-hip ratios and lower fat-free mass than their younger counterparts (p<0.05). In particular, 36% of middle-aged women and 20% of elderly women were classified with 'relative abdominal overfatness'. Vietnamese middle-aged and older adults are shorter and thinner than their counterparts in American and European countries but overfatness increases with age, particularly in women. Urban elderly are vulnerable to both undernutrition and overnutrition, both of which merit consideration in the geriatric care system in Vietnam.

Food intake and eating patterns of Indonesian elderly before the 1998 economic crisis.

The Nutrient and Metabolic Study of Indonesian Elderly (NUMSIE) was conducted in part to identify differences in eating patterns and in food and energy intakes between elderly people residing in urban metropolitan Jakarta (JAK) and in urban non-metropolitan Semarang (SEM) in order to investigate the prevalence of food and energy deficiencies. Data on food intake were collected from 212 JAK elderly and 238 SEM elderly aged 60 years and over using a quantitative food frequency questionnaire (FFQ). Although most of the elderly lived with their families or extended families, a large proportion of the subjects were eating alone, especially in the SEM sample. Jakartan elderly had significantly higher intakes of most food groups, except for added sugar and cow's milk. Total food intake of JAK subjects was also significantly higher (p<0.0001) than that of SEM subjects. The ratio of plant to animal food was lower among SEM elderly due principally to their higher intake of milk. Thirty percent of both JAK and SEM elderly consumed less than the recommended amounts of cereals, followed by vegetables and fruits (10%, 47% JAK; 22%, 75% SEM, respectively). Finally, it was found that the range of daily energy intakes was higher in JAK (1251-2079 kcal) than in SEM (939-1579 kcal). This suggests that SEM elderly were more likely to be energy deficient than were JAK elderly. While the results of this study indicate that food and energy intakes may be inadequate in Indonesian elderly, especially in non-metropolitan areas, more analyses are required to ascertain the true prevalence of malnutrition in this age group using anthropometric and blood measurements.

Clinical trials in nutrition.

Trials of nutritional intervention in a wide range of health and disease states, preventive and therapeutic, are required. Not only has the emergence of chronic non-communicable disease (CNCD) with acknowledged nutritional pathogenesis created this imperative need, but so also have other conditions which, previously, had not been regarded as nutritionally based. Among the latter are health problems associated with ageing: the menopause, a decline in immune function, and a decline in cognitive function. At the same time, there is a new set of materno-foetal and infant nutrition issues for investigation which relate to new food exposures and the long-term effects of nutritionally mediated gene expression. The emergence of the new food science of phytochemicals with human biological importance also sets the scene for their evaluation in traditional diets and novel foods. Such trials are more complex than comparable pharmacotherapeutic studies because of the complexity of food chemistry, as well as the food behavioural changes which may accompany a nutritional intervention, and the general problem of there not being a 'gold standard' for food intake methodology. Choice of study population is also a key issue in relation to the extrapolation of findings from a particular trial, with population representativeness being an advantage. In order to obtain useful information on manageable sample sizes, either intermediate end-points (short of morbidity and mortality) need to be studied or high-risk groups (such as the aged) need to be recruited. There are some unique ethical issues which must inform clinical nutrition trials. These include certain preventive imperatives like the right to be fed, the risks in disruption of food cultures and the need for food security and sustainability. Rapid changes in the food supply do, however, make such trials more important, while the value of food-health knowledge that cannot be obtained by trial must still be appreciated.

Agreement of skinfold measurement and bioelectrical impedance analysis (BIA) methods with dual energy X-ray absorptiometry (DEXA) in estimating total body fat in Anglo-Celtic Australians.

OBJECTIVE: To compare percentage total body fat (%BF) estimated by the four skinfold thickness measurement (SKF) and single-frequency bioelectrical impedance analysis (BIA) methods using three different sets of equations, to that assessed by the dual energy X-ray absorptiometric (DEXA) method using a Lunar DPX densitometer. DESIGN: Cross-sectional study. SUBJECTS: An Anglo-Celtic Australian population of 66 males and 130 females (age: 26-86 y). MEASUREMENTS: %BF by anthropometry, BIA using three different sets of equations and DEXA. RESULTS: Mean %BF assessed by DEXA (%BF(DEXA)) was similar to that estimated by SKF (%BF(SKF)) in males, while %BF(DEXA) was slightly higher in females. %BF estimated by BIA (%BF(BIA)) was significantly lower than %BF(DEXA) in females, regardless of equations used for calculation, while the level of agreement between BIA and DEXA in estimating %BF in males was dependent on prediction equations used for calculation of %BF(BIA). A better agreement was obtained from the use on the prediction equations of Segal et al (1988), compared to other two sets of equations. The agreement between SKF or BIA and DEXA declined with increasing %BF. CONCLUSIONS: There was a good agreement between DEXA and SKF, and slightly less so between DEXA and BIA, in estimating %BF in an Anglo-Celtic adult population. The agreement in most cases, however, was dependent on the degree of body fatness. In comparison to DEXA, both SKF and BIA, with the use of the equations of Segal et al (1988), are applicable to estimate %BF in an Anglo-Celtic Australian population.

Validation of predicted total body water and extracellular water by multi-frequency impedance in young Indonesian adults.

To investigate the validity of the prediction of body water compartments by multi-frequency impedance in Indonesians, 42 adults (19 males and 23 females) were recruited in Jakarta, the capital of Indonesia. Extracellular water (ECW) and total body water (TBW) were measured with dilution techniques and predicted from body impedance at 1 and 100 kHz, respectively, using prediction formulas developed in a Dutch population. Furthermore, body weight, body height and body density were measured and as a measure of body build TBW/height and ECW/height were calculated. Percent ECW of body weight in both males (26 +/- 3%) and females (23 +/- 2%) was higher compared to the Dutch reference population, and also the body water distribution (ECW/TBW) was higher (0.46 +/- 0.04) compared to the reference population. TBW/height and ECW/height appeared to be considerably lower in the Indonesian group compared to the reference group, indicating a more slender body shape. TBW was overestimated by 1.8 +/- 2.4 and 2.5 +/- 1.3 kg in males and females, respectively. ECW was underestimated in males by 0.1 +/- 1.2 kg and slightly overestimated in females (0.1 +/- 1.0 kg). After corrections for differences in body build and body water distribution the bias of the predicted TBW was markedly reduced from 2.2 +/- 1.9 to 1.3 +/- 1.8 kg (males and females combined). For ECW the bias after correction for body water distribution and body build was 0.3 +/- 0.6 kg (males and females combined). It is concluded that multi-frequency impedance is suitable for the prediction of body water compartments at the population level but, due to differences in body build and body water distribution, validation in a larger group of Indonesian subjects is needed to avoid systematic prediction errors.

Prevalence of Helicobacter pylori in a representative Anglo-Celtic population of urban Melbourne.

The aims of this study were to assess the prevalence of Helicobacter pylori and its relationship with different epidemiological factors in an Anglo-Celtic Australian population in the Melbourne urban area. Two hundred and seventy-three (120 men and 153 women with a mean age of 55.6 and range of 20 to 80 years) of 396 eligible subjects randomly sampled from the telephone directory were studied. An ELISA technique was used to detect H. pylori immunoglobulin (Ig)G antibody and self-administered questionnaires were completed. The overall seroprevalence of H. pylori was 38% and increased with age from 18% (20-30 years old) to 53% (over 70 years; P < 0.0001). The acquisition of H. pylori infection was 1% per year. The prevalence of H. pylori was 48% in men and 30% in women (P < 0.01). The frequency of H. pylori was also associated with low-income levels and current smoking, but was not associated with peptic ulcer disease history. The prevalence of H. pylori infection in a representative Australian population was found to be similar to other developed countries. The risk factors for H. pylori infection include age, male sex, low household income and a smoking habit. No correlation between H. pylori status and dyspepsia symptoms were observed.

Platelet fatty acids and peripheral blood lymphocyte subsets in an institutionalized elderly population.

The associations between platelet fatty acids and peripheral blood lymphocyte subsets were studied in 78 institutionalized elderly individuals (27 men and 51 women), aged 67 to 100. Platelet fatty acids were assessed by gas chromatography, and peripheral blood lymphocyte subsets were quantitated by immunophenotyping using flow cytometry. It was found that women had a higher number of total T-cells (CD3), T-helper (CD3+4+) cells, and B-cells (CD19). However, no gender differences were observed in the percentages of lymphocyte subsets. In elderly men, after adjusting for age and fatty acid intake, the platelet concentration of omega-3 polyunsaturated fatty acids was positively related to the percentage of CD3 and CD3+4+ bearing lymphocytes (rs = 0.59, P < 0.05; and rs = 0.55, P < 0.05, respectively), and the concentration of total saturated fatty acids was also positively associated with the percentage of B (CD19) cells (rs = 0.50, P < 0.05). However, similar relationships were not observed in elderly women. No significant associations were found between trans fatty acids and any of the lymphocyte subsets in the study population. These findings suggest that fatty acids may be related to immune function. Any effects may be important in the host immune defence, especially in elderly individuals.

Cross-cultural comparison of Lp(a) profiles.

Lipoprotein(a) (Lp(a)) and apolipoprotein(a) (apo(a)) phenotypes as genetic markers for coronary heart disease (CHD) have been the focus of great interest in recent times. Included in this study were four Australian populations comprising 348 Anglo-Celtic Melburnians (157 men and 191 women), 339 Chinese Melburnians (169 men and 170 women), 402 South Asian Melburnians (216 men and 186 women) and 394 Aboriginal Australians from Western Australia (175 men and 219 women). Plasma Lp(a) concentrations were more highly skewed towards the lower range in the Chinese and Aboriginal groups than in the Anglo-Celtics and South Asians. Approximately 33% of Anglo-Celtics, 20% Aboriginals, 13% Chinese and 44% South Asians had plasma Lp(a) levels above the generally accepted risk threshold values of 300 mg/L. In Aboriginals and Chinese, the S4 apo(a) phenotype predominated while in Anglo-Celtics and South Asians, the highest frequency occurred in the S3 phenotype. In the S4 phenotype, Lp(a) values varied between the four populations but there was no significant difference in concentration between gender.

Measurements of body fat in Indonesian adults: Comparison between a three-compartment model and widely used methods.

Body composition was assessed in Indonesian male (n = 18) and female (n = 23) students using densitometry (underwater weighing), deuterium oxide dilution, skinfold thickness measurements, bioelectrical impedance analysis (BIA) and a prediction equation based on the body mass index. From body density and total body water percentage body fat (BF%) was calculated using a three-compartment body composition model. Percentage body fat obtained by this three-compartment model was regarded as the reference value and BF% obtained by the single methods were compared with this value. Mean differences (± SD) in BF% from the threecompartment model minus the single methods were -1.1 ± 2.1 for densitometry, +1.1 ± 1.6 for deuterium oxide dilution, +1.3 ± 2.8 for skinfold thickness measurement, +2.8 ± 4.3 for BIA and +3.4 ± 4.8 for body mass index in males. In females these values were +0.1 ± 1.7, +0.2 ± 1.4, +3.6 ± 3.3, +3.6 ± 2.4 and +8.7 ± 2.0 BF%, respectively. Correlation coefficients between different methods were high and significant (P < 0.05 in males, P < 0.001 in females). This study shows that the single predictive methods have considerable mean and individual biases compared with the three-compartment model and all predictive methods underestimated body fat in the studied subjects. It is concluded that the development of population-specific prediction formulas may be necessary.