Some studies of high altitude populations argue that stature reduction results from caloric, rather than hypoxic, stress. However, tradeoff models of oxygen and glucose metabolism predict that in hypoxemia, glucose metabolism will be downregulated. We used tradeoff assumptions in two hypotheses: First, that hypoxia targets leg segment growth differentially, and second, that proportions of leg segments partition the impact of high altitude into hypoxemic and energetic components. A group of 113 Han and Tibetan middle school children at 3100 m aged 8 to 11 were measured for segment anthropometries, skinfolds, vital capacity, blood oxygen saturation, and percent body fat. MANOVA showed that Tibetan children were significantly larger and fatter than Han children. Independent of ethnicity or caloric status, absolute and relative tibia length was significantly reduced in children with lower blood oxygen saturation. Height, chest circumference, sitting height, tibia length, and ankle diameter were greatest in fatter children, independent of ethnicity or blood oxygen. For children of either ethnicity with the lowest blood oxygen, size as well as proportion was impacted. These results support the tradeoff model. Caloric reserves and ethnicity independently affect total skeletal size. Oxygen saturation and ethnicity affect leg proportions. In hypoxemia, body fat has less impact on growth than when ample oxygen is present. Therefore, we should qualify the claim that size in high altitude populations stems from nutritional stress. The findings also suggest that decanalization may have different meanings and outcomes depending on which body segments contribute to the effect.

To assess cross-cultural relations between dietary intake and plasma lipoproteins, we randomly selected 222 men and 243 women from the urban and rural areas of Puriscal, Costa Rica; related their dietary composition (assessed by a food-frequency questionnaire), fitness level, and body fat to plasma lipids, apolipoproteins, and low density lipoprotein (LDL) particle size; and compared these data with those from a subsample of 280 adults from the Framingham Offspring Study. Total cholesterol and LDL cholesterol levels were significantly (p<0.0001) higher in Framingham (207 and 137 mg/dl, respectively) than in Puriscal (184 and 114 mg/dl, respectively) residents. Elevated triglyceride and apolipoprotein (apo) B levels (25% and 16% higher), low HDL cholesterol and apo A-I levels (12% and 29% lower), and smaller LDL particles (17%) were more frequent in Puriscal than in Framingham residents. Urban Puriscal residents had a significantly lower fitness level; increased body fat, total cholesterol, and triglyceride levels; decreased HDL cholesterol in men; and higher apo B levels in women compared with rural Puriscal residents. Body fat, animal fat, and saturated fat intakes were significantly correlated with total cholesterol, LDL cholesterol, and apo B levels in both men and women in Puriscal. Intakes of protein and animal fat were higher among urban (10.7% and 14.1%, respectively) compared with rural (8.9% and 9.9%, respectively) Puriscal residents and in Framingham (16.0% and 20.8%, respectively) compared with Puriscal residents. No significant differences were found in dietary cholesterol. Saturated fat (largely from palm oil in Puriscal) intakes were significantly different among the three groups: rural Puriscal, 10.7% of calories; urban Puriscal, 11.6%; and Framingham residents, 12.9%. These data indicate that the more atherogenic plasma lipid profile among urban compared with Puriscal residents was largely explained by increased adiposity, decreased fitness level, and higher saturated fatty acid intake. Puriscal residents consumed less animal fat and more carbohydrate than did Framingham residents, and these differences were associated with a 21% lower LDL cholesterol level, a 12% lower HDL cholesterol level, a 29% lower apo A-I level, a 25% higher triglyceride level, a 16% higher apo B level, and a 17% smaller LDL particle size. Some of these cross-cultural differences may be due to differences in ethnic background and physical activity as well. (Arteriosclerosis and Thrombosis 1991;ll:1089-1099)

Increased general and abdominal obesity has been independently associated with diabetes, increased risk of stroke, and coronary artery disease (CAD). It is more prevalent in developed countries and in urban areas of nonindustrialized nations than in less developed and rural areas. To evaluate the associations between general and abdominal obesity (as determined by total body fat, waist to hip ratio, umbilical to triceps ratio, and umbilical to subscapular ratio) with glucose, plasma lipoproteins, apolipoprotein (apo) A-I and B concentrations, and low density lipoprotein (LDL) particle size (LDL 1-7), we randomly selected 222 men and 243 women from rural and urban areas of Puriscal, Costa Rica. Abdominal obesity, as assessed by the waist to hip ratio, was independently and significantly associated with higher triglyceride levels (p<0.01) and with lower high density lipoprotein cholesterol levels (p<0.05) in men and women and with higher glucose levels (/?<0.05) and smaller LDL particle size (p<0.01) in women. Abdominal obesity, as assessed by the umbilical to subscapular ratio, was independently and significantly associated with higher total cholesterol (p<0.005) and apo B (p<0.01) levels. Umbilical to triceps ratio was positively associated with blood pressure in men. Urban men had increased general and abdominal obesity (/><0.0001), number of cigarettes smoked per day (p<0.0001), and diastolic blood pressure (p<0.05) and had a decreased fitness level (p<0.0001) as well as higher (/?<0.05) plasma glucose, triglyceride, and total cholesterol concentrations and lower (p<0.05) apo A-I and HDL cholesterol levels compared with rural men. The differences between rural and urban women were not as striking. Urban women had increased general and abdominal obesity, glucose, and apo B levels (/?<0.05) and a decreased fitness level (/><0.0001). Our data indicate that general and abdominal obesity, increased cigarette smoking, diastolic blood pressure, and decreased fitness level are more prevalent in an urban than in a rural area in Costa Rica, particularly in men. The higher prevalence of such risk factors in the urban area is associated with a more atherogenic plasma lipoprotein profile. (Arteriosclerosis and Thrombosis 1991;ll:1077-1088) G eneral and abdominal obesity has been independently associated with increased levels of cardiovascular disease (CVD) risk factors and mortality and with increased death rates for all causes.1 -5 Although the prevalence of CVD risk factors is lower in developing countries than in industrialized nations, 6 the incidence of mortalityFrom the Lipid Metabolism Laboratory (H.C., L

Socioeconomic status is systematically related to the level of fatness, and therefore the incidence of obesity, in a total community survey of nearly 5,000 adults. Among males with more than 12 years of schooling, the average thickness of 4 fatfolds is 10% greater, amounting to about 2 kg of total fat, than those with 8 years or less of education. In females, however, the opposite trend is observed, those in the higher educational group averaging 20% thinner fatfolds, or about 5.5 kg total fat, than females in the lower educational group. These findings confirm the need for standards of obesity that take socioeconomic status into account.

Body fat distribution, its continuity from childhood (4-6 years) to 30 years of age, and its link to that of parents is described in a longitudinal study population. A computer-assisted image analysis technique was used to measure body fat distribution (as measured by waist and hip diameters) from somatotype photographs. Pearson correlation coefficients were used to assess the association between the waist/hip diameter ratio (WHDR) and the body mass index (BMI) at all ages and between the WHDR of parents and offspring. Both Pearson correlation coefficients and the Foulkes-Davis tracking index were used to assess tracking of the WHDR from childhood to age 30 years. Stepwise regression analyses were performed to determine the predictability of the WHDR at age 30 years from WHDRs in childhood and adolescence. Among both sexes the WHDR decreased and BMIs increased with age from childhood to 18 years, then both increased from age 18 to 30 years. The correlation between the WHDR and BMI was significant only at 30 years for males (r = 0.37; p < 0.05) and during all stages of adolescence in females (p < 0.05). Age-to-age correlations were high for both sexes (p < 0.0001), and remained significant over a span of up to 25 years. Using the Foulkes-Davis tracking index, tracking from the year of peak height velocity to 30 years (a span of approximately 20 years) was strong for both sexes. Parent and child WHDRs were correlated and differed by sex. Father-son correlations were not significant in childhood, but reached significance in early adolescence (2 years before the year of peak velocity) and remained significant to 30 years (p < 0.05). Mother-daughter correlations were significant at all ages (p < 0.05). The level of the WHDR at the peak of the pubertal growth spurt (year of peak height velocity) predicted up to 58% of the variance in males and 51% of the variance in females of the WHDR at 30 years. We conclude that the adult WHDR (a proxy for the waist/hip ratio) becomes evident by the year of the pubertal growth spurt in height.