The juvenile-onset, adolescent-onset and adult-onset obese.

As shown in more than 8000 proband-parent pairs derived from a total-community sample and followed in longitudinal fashion, the 5-year incidence of obesity (new cases per 5-year period) approximates 8 percent for the juvenile-onset, adolescent-onset and adult-onset obese alike. Parents of juvenile-onset (ages 5-9), adolescent-onset (10-19) and adult-onset obese (20-39) tend to be of above-average fatness level, +0.25Z scores, overall, regardless of the age at onset of obesity in their progeny. Except for the parents of the juvenile-onset obese, educational level of the parents tends to be below average for the sample as a whole. These new data acquired in longitudinal context and explored in retrospective-prospective fashion do not substantiate the notion that different onset ages of obesity indicate separate etiologies and different family constellations.

Fatness and obesity of the parents of obese individuals.

As shown in 1419 pairings of obese probands with their parents drawn from a larger series of greater than 9000 proband-parent pairings, the fathers and mothers of obese probands are of increased fatness level (+0.27 Z scores) and more often obese than expected (odds ratio 1.50 overall). However, the tendency towards increased fatness and a greater prevalence of obesity among the parents of obese probands bears a curvilinear relationship to the age of the proband, being least when the probands are young, peaking when the sons and daughters are teen-agers, and declining thereafter. Parents of lean probands in turn tend to be lean themselves (averaging -0.25 Z scores) and least often obese when their progeny are teen-aged. As shown in a two-generational context, familial obesity is best demonstrated in adolescents and their parents, either reflecting years spent in common or a specific etiology for adolescent-onset obesity.

Educational level, fatness, and fatness differences between husbands and wives.

In 1017 husband-wife pairs aged 20-49 y, both fatness and weight of the wives is inversely and linearly related to educational level whereas the husbands' summed skinfold measurements bear a curvilinear, parabolic relationship to years of schooling. Overall, fatness differences between husbands and wives diminish with increased education, and beyond 13 y of education husbands often exceed the subcutaneous fat thickness of their wives. When the education of one spouse is held constant at 9-12 y, women who marry up to men of greater education are systematically leaner and women who mary down are both fatter and heavier. These findings in spouse pairs sharing a common family income suggest that assortative mating with respect to educational level accounts for much of the socioeconomic effect in fatness including the fatness differences between women who marry down and those who marry up. These data also demonstrate that women are not fatter than men at all socioeconomic levels.

Effect of skinfold levels on lipids and blood pressure in younger and older adults.

As shown in 5507 white participants in a total population sample, the level of fatness is systematically related to lipid levels and to blood pressure levels in older adolescents, younger adults, and older adults of both sexes. At all three age levels, the fatter subjects were highest in serum cholesterol, serum triglycerides, and systolic and diastolic blood pressure and more often hypertensive. The effect of fatness level on the four risk factors was similar for all four skinfolds regardless of location. Fatness level is thus related to lipid levels and blood pressure levels in both younger and older subjects, and there is no evidence that "central" or "peripheral" or upper body and truncal skinfolds are more directly related to these risk factors even after the sixth decade.

Evidence against functional differences between "central" and "peripheral" fat.

As shown in 1639 men and 1851 women in the 20-49 y age range, four commonly used skinfold measurements (triceps, subscapular, iliac, and abdominal) are all highly covariant and do not show clear evidence of upper-body and lower-body or central and peripheral groupings. Moreover, the four skinfold measurements are similarly related to systolic and diastolic blood pressures and to total serum cholesterol at each decile of skinfold thickness and to a degree that strikingly parallels summed skinfold thickness (sigma sf). The changing relative contributions of the different skinfold measurements to the summed skinfold thicknesses with increasing levels of fatness also reveal no evidence of consistent anatomical or topographical groupings by body segment or distance from the body core.

Differential rates of fat change relative to weight change at different body sites.

As shown in 620 adults followed over a 5-year period, rates of change in outer fat differ from skinfold to skinfold and, to a lesser extent, between the sexes. At each of four sites and for men and women respectively fatness change rates per kilogram of weight change are much the same for heavier and lighter individuals, for fatter and leaner subjects, and during weight gain as compared with weight loss. In both sexes the abdominal and iliac sites are more sensitive to overall changes in weight and fatness.