In this nationally representative study of 7411 young- to middle-aged adults, we found that leg adiposity, as measured via leg/total% and leg fat%, as well as corresponding arm adiposity measures were inversely associated with DBP and SBP, respectively. Arm and leg lean/fat ratio, on the other hand, were positively associated with SBP. Among the three measures of leg adiposity, only leg/total% demonstrated a significant, 30% decreased odds of hypertension when comparing T3 (≥36% for male, ≥41% for female) to T1 (< 31.5% for male, < 36% for female). Furthermore, leg/total% decreased the odds of IDH and SDH, but not ISH relative to normotension. BMI modified the association between leg/total% and hypertension, with low/normal BMI individuals showing significantly more protective odds than overweight and obese individuals. Interestingly, arm fat%, not arm/total%, was inversely associated with SBP and overall hypertension. Consequently, leg/total% had a higher predictive value than arm/total% but was insignificantly lower than waist circumference and age in predicting hypertension.
Our findings are in line with existing literature on the association between leg adiposity and various cardiometabolic conditions [13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32]. While prior studies have generally not focused on BP or hypertension as a primary outcome, they have consistently reported inverse associations between leg adiposity measures and metabolic risk factors. Zhang et al. , in a study of NHANES 1999 to 2006 participants with DXA scans, found that, regardless of ethnicity, leg adiposity indices (leg/total%, leg fat%, leg fat/truncal fat ratio) were all inversely associated with metabolic syndrome. Leg/total%, however, had the strongest association. Other studies have similarly found inverse associations between leg/total% and cardiovascular risk factors [18,19,20], cardiovascular disease risk and all-cause mortality [21,22,23,24,25,26]. In fact, Han et al. , using the Korean NHANES, showed that leg/total% displayed the strongest association with atherosclerotic cardiovascular disease risk. Interestingly, studies that additionally assessed arm adiposity have found no significant association, and in some cases positive associations, with metabolic risk factors [17, 20, 32, 33]. Our findings also corroborate a positive association between arm/total% and SBP and hypertension that was later attenuated by adjustment for cardiometabolic risk factors.
There continues to be variability in data concerning sex differences. Hu et al. , among White and Black Americans, showed that females generally had more protective associations of leg adiposity with metabolic syndrome criteria than males, Sakai et al.  demonstrated no difference between males and females for SBP or DBP after adjusting for menopausal status. While we also found no significant difference between sexes, we observed more protective odds of hypertension in females than males.
Of the four studies that have investigated the association between appendicular adiposity and BP/hypertension [18, 19, 25, 31], only one was conducted in a US representative population, all used 140/90 as the hypertension threshold, and none have commented on hypertension subtype. Nevertheless, our findings corroborate those of the four studies. Sakai et al.  found a significant, inverse association between leg fat mass and hypertension (≥140/90) as well as a mean 0.22 mmHg decrease in DBP and 0.16 mmHg decrease in SBP per 1 kg increase in leg fat mass in 4256 Japanese men aged 20 to 79 years. The authors also found an increase in BP for every 1 kg increase in lean leg mass, a counterintuitive finding that we also observed . Increased lean mass, we speculate, may be associated with increased serum testosterone, an androgen that has been inconclusively linked to increased BP and cardiovascular disease . Zhang et al.  showed a significant inverse correlation between leg/total%, but not leg fat/leg mass ratio, and SBP and DBP. Han et al. , in the Korean NHANES population, found that those in T1 of leg/total% had nearly 3.5 times the odds of hypertension (≥140/90) compared to T3, although models were not adjusted for truncal adiposity or arm adiposity measures. Finally, Yan et al. , in a Chinese population, similarly demonstrated that leg/total% was primary negative risk factor for hypertension (≥140/90).
They clarified that this association was stronger in the non-obese population, a finding that we also demonstrated. This would seem to suggest that, even in those with low/normal BMI, body distribution of fat is important. However, after adjusting for truncal fat and waist circumference, leg/total% continues to be a negative predictor of hypertension in low/normal BMI individuals. This suggests that, in addition to body distribution of fat, there exist a physiological difference between lower extremity adiposity and truncal or upper extremity adiposity. Conventionally, we have treated truncal fat as the primary source of fat that drives of metabolic disease; however, these studies and ours lead to the speculation that measurement of leg fat is just as important, and its effects independent of, truncal fat.
Not only is leg/total% inversely associated with hypertension, it is also preferentially inversely associated with DBP and associated hypertension subtypes. This suggests that leg adiposity may a play a role in regulation of DBP, which is determined in large part by peripheral vascular resistance in arteriolar vessels. Although the biological mechanism behind leg fat’s protective role is unknown, several studies have speculated and found that (1) leg fat is located primarily in subcutaneous tissue, as opposed to visceral tissue; (2) adipocytes in the lower extremities had lower free fatty acid turnover and lower rates of lipolysis; and (3) this decreased free fatty acid turnover may lead to decreased fatty acid concentrations in the blood and downregulation of triglyceride production [18, 19, 38].
We hypothesize that, in addition to decreased sympathetic and renin-angiotensin-aldosterone axis activation associated with subcutaneous fat, this downregulation of triglyceride production, which is corroborated by inverse associations between leg/total% and triglyceride levels in studies assessing metabolic syndrome criteria, leads to decreased endothelial cell damage in arterioles and maintenance of vessel elasticity and compliance. This is further supported by Lee et al. , which found that every SD increase in leg/total% was associated with a 1.03 m/sec decrease in brachial-ankle pulse wave velocity, a measure of arterial stiffness.
There are several strengths of this study. We used a set of recent, large, nationally representative surveys with gold standard measurements of fat and lean mass through DXA scans. We comprehensively accounted for covariates to further confirm the independent association of leg/total% and BP. By having a younger study population, we also minimized unmeasured chronic disease effects and effects of menopause on women. We utilized the 2017 ACC/AHA guidelines for hypertension, which not only includes more individuals in its hypertension diagnosis but emphasizes the notion that leg/total% may be important in those who are seemingly healthy.
Despite these strengths, there are limitations as well. This is a cross-sectional study so causality cannot be determined. It may be that those who are hypertensive tend to accumulate fat elsewhere. Further, while it is generalizable to those in the US aged 20 to 59 years, these results may not be generalizable to older adults ≥60 years who also have disproportionately higher hypertension prevalence. While DXA scans are the gold standard measurement for body fat distribution, it cannot distinguish between subcutaneous and intramuscular fat in the legs, limiting our interpretation of results. Furthermore, visceral fat is still more clinically relevant and more strongly associated with cardiovascular disease risk, posing the question as to how clinically useful our results may be. Also, although our results are statistically significant, they may or may not be clinically significant. Although the absolute changes in BP are miniscule, it is important to note these BP differences are on a population level and across 1% increments of appendicular adiposity. There may be residual confounding that is unaccounted for in this analysis, especially due to limited information on medication use and comorbidities (e.g., polycystic ovarian syndrome in women, hormone therapy).