This study demonstrates that the relationship between BMI and body fat distribution differs by race/ethnicity. Furthermore, it adds to the literature by examining these variables in reproductive-aged Hispanic women. We observed that, for a given BMI, white women had the highest value for FMtotal
, and %FM while Hispanic women had the highest value for %FMtrunk
, and FMRtrunk-to-limb
, and the lowest %FMleg
. With the exception of FMleg
, black women had the lowest value for all other body fat distribution variables. In nearly all cases, the magnitude of the difference significantly increased with increasing BMI. Consistent with previous research (6
), these findings suggest that racial differences should be taken into account for obesity-related cut-off points and health risks in reproductive-aged women.
Several prior studies have similarly observed that white women are more susceptible to visceral obesity than their black counterparts, despite similar BMI, waist circumference, or waist-to-hip ratio (6
). Extending this research, we also showed significant differences in these variables between Hispanic women and their white and black counterparts in reproductive-aged women. Both white and Hispanic women were more likely to have body fat at different body sites while white women had greater amount of adiposity than their Hispanic counterparts except percent body fat mass. Our observation that white women were more likely to have greater percent body fat mass than black women is in contrast to the findings reported in Fernandez et al study (27
) among postmenopausal women. The authors observed that Hispanic women had more percent body fat mass than black women while they were similar in black and white women. We are unable to explain the reasons; however, some physiological or environmental changes could be ultimate factors, which should be examined in future studies. Based on the racial influence on body fat distribution and consistent with the notion of personalized medicine, women should not be assessed and treated with a one-size-fits-all approach. Indeed, as shown in this and other studies, it is evident that susceptibility to a particular body type (and thus CVD risk) may be partially determined by one’s race/ethnicity.
Central fat distribution has repeatedly been associated with CVD risk factors and metabolic complications (1
) while the reverse is true for leg fat (5
). Given the findings from this study and previous research showing that black women have less central fat distribution and a better lipid profile (35
) than white women, it would seem that black women may have a lower incidence of CVD and metabolic complications. However, the actual scenario is different; black women have a higher incidence of CVD, diabetes mellitus, and related morbidity than white women (36
). Furthermore, with regard to the relationship between central fat deposition and metabolic risk factors, Lovejoy and colleagues noted a generally weaker association in blacks relative to whites (22
). Lower peripheral insulin sensitivity in black women compared to white women (39
) could be one of the most important reasons behind this discrepancy. Other CVD risk factor indicators such as BMI, blood pressure, and glycosylated hemoglobin, which were reported to be significantly higher among black women than white women (36
) could also play an important role in this regard. In addition, genetics, environmental factors, dietary habits, and physical activity may play a crucial role in this regard and should be evaluated in longitudinal studies.
The differences by race/ethnicity observed in this study have important clinical and community health implications. For example, our finding that white women had the highest FMtrunk for a given BMI is contrary to the general perception that trunkal obesity is more prevalent among women of other racial/ethnic groups. Hispanic women, on the other hand, had higher trunk-to-limb ratio and trunk fat as percent of total fat. This finding suggests that prevention and intervention campaigns focusing on the dangers of trunkal obesity should target reproductive-aged women in general and white and Hispanic women in particular. Moreover, tailoring awareness materials to specific at-risk populations may also be warranted. For example, information targeting Hispanic women could address the importance of reducing their trunk-to-limb fat ratio as a measure to prevent CVD and metabolic complications.
It is undeniable that existing cut-off points of BMI can be used by a layperson easily to estimate the risk of obesity-related health problems. However, together with other studies (6
), our findings indicate that BMI-based guidelines to identify obesity are inadequate. Additional measures are required to counsel women of different race/ethnicity about their actual risk of morbidity based on their body fat distribution. For example, to reduce the risk of obesity-related morbidity, white and Hispanic obese women could be advised to reduce their BMI below 28–29 for BMI<30 target group (similarly 23–24 for BMI<25 target group) level as they have more body fat for a given BMI than black women. Similarly, waist circumference cut-off points for metabolic syndrome could also be revised for white and Hispanic women as they have more trunk fat for a given BMI than their black counterparts. Future studies are warranted to generate consensus on the exact and useable cut-off points for these two important and widely used parameters.
The findings that black and Hispanic women had significantly higher average BMI than white women actually offset their advantage of having lower body fat. Thus, there is no scope of complacency in the former two race/ethnic group of women with regard to their lower body fat for a given BMI. Rather, they should be targeted to reduce their body weight as they are in risk of gaining pregnancy-related extra body weight. Programs aimed at preventing weight gain during adolescence and reproductive-age along with minimizing racial disparities would have great overall impact. Parallelly, disparity in body fat distribution needs to be incorporated in all weight reduction programs.
The nonlinear relationship between BMI and body fat distribution variables observed in this study is similar to the results of Jackson and colleagues who found a quadratic relationship between %FM and BMI (45
), but is in contrast to Gallagher et al. who observed a linear relationship (46
). This discrepancy might be explained by differences in the upper range of BMI in the Gallagher et al. sample (≤35) compared to the Jackson et al. sample (≤40) and the current sample (≤49). Similar to the observations of Jackson et al., the influence of a BMI ≥35 on the curvilinear relationship is obvious in our study. In almost all cases, the relationship between body fat variables and BMI showed a linear relationship up to a BMI value of 35 kg/m2
Although this study adds to the growing literature on the importance of body fat distribution, several limitations should be noted. First, we did not collect information on related anthropometric measurements (waist circumference and waist-to-hip ratio), and visceral and subcutaneous adipose tissue of the abdomen separately, which could have given us additional insight about the racial influence on body fat distribution. Second, we were not able to include women over 300 pounds, due to the manufacturer’s instructions regarding the DXA table. In addition, women were not included if they were unable to receive hormonal contraceptives containing estrogen, or wished to become pregnant in ≤3 years due to the primary specific aims of the larger study. Together, these limitations could impact the overall generalizability of our findings and selection bias cannot be ruled out. The strengths of our study include use of DXA method to estimate the body fat variables which is well-validated and relatively large sample size with tri-ethnic women population.
In conclusion, our study demonstrated that racial differences are present in the relationship between BMI and body fat distribution. The findings generated in this study should be accommodated in obesity-related cut-off points and associated health risks. Furthermore, future research on body fat distribution and its relationship with CVD and metabolic risk factors should take into account racial differences, dietary habits, physical activity, and genetic and environmental factors.