Resting energy expenditure (REE) accounts for 70–80% of total daily energy expenditure and is tightly regulated, exhibiting relatively little variation within an individual over time. Epidemiologic survey research on the inter-relationship of REE with traits involved in the cardiovascular metabolic risk pattern and other chronic diseases has been focused primarily on the potential relationship with obesity [1]. The majority of energy required in the resting state is devoted to protein synthesis and maintenance of the electro-chemical gradient across cell membranes [2]. As a consequence, REE is highly correlated with lean body mass although it can be up-regulated by increases in circulating thyroxin and catecholamines [3‐5]. REE also responds to variation in energy intake, for example, during periods of sustained fasting REE declines to compensate for lower energy intake along with profound changes in the hormone milieu [6].

Because of its important role in the energy budget, inter-individual variation in REE has been examined extensively as a potential contributor to susceptibility to obesity. A series of widely cited studies conducted among the Pima Indians in the US Southwest has reported a significant negative correlation between REE and risk of obesity [7, 8]. However, our group found equivalent REE levels after adjustment for body size and composition in a study comparing African-Americans and rural Nigerians, and no relationship to risk of future weight gain [9]. This null REE-obesity association is consistent with previously reported research finding no effect for either REE, or sleeping EE, for future weight gain among obesity-resistant or obesity-prone women [10, 11].

As a measure of metabolic activity in the resting state, it is reasonable to speculate that a range of systemic metabolic functions might be influenced by, or associated with, REE [12]. In a previous report, we identified a strong relationship between REE and blood pressure (BP) after controlling for measures of body size and composition in population-based samples in the US and Nigeria [13]. This relationship has also been described in a case-control study of obese and non-obese individuals [14], as well as in indigenous Siberians [15]. A recent report found after 4.5 years of follow-up in healthy, overweight African American and white women, that REE was a significant predictor of systolic blood pressure, independent of adiposity [12].

Adiposity has long been recognized as a risk factor for hypertension [13, 16, 17]. However, the underlying physiological mechanisms that link excess body fat stores to BP are poorly understood [13]. Obesity is a highly complex syndrome that is a consequence of metabolic alterations and lifestyle patterns that include, but are not limited to, insulin resistance, hyper-leptinemia, elevated BP, a high-calorie/high-sodium intake and reduced physical activity [13, 18]. Co-linearity of these relationships makes it difficult to isolate specific causal effects. For example, sodium intake is associated with BP in cross-sectional studies, although the effect is reduced or eliminated by control for body mass index (BMI) [19]. The increased energy intake among the obese makes higher salt consumption obligatory; whether control for BMI is “over-control” is therefore an open question [20]. Abnormalities in the renin-angiotensin system and increased levels of various regulatory hormones, like insulin and leptin, have been consistently described to take place among the obese, although the causal link to hypertension is tenuous. A number of other metabolic abnormalities, have been reported among obese vs. non-obese hypertensive individuals, although these associations are relatively inconsistent [13, 21]. Generally, it is assumed that body mass itself, driven largely by the size of body fat stores, is the primary determinant of hypertension risk. However, it must be recognized that an obligatory increase in lean body mass occurs in parallel with total body weight, which in the general population is highly correlated with fat mass [13, 21].

As part of an international comparative study of the evolution cardio-metabolic risk factors, we have been investigating the effects of energy expenditure and obesity on BP in populations of the African diaspora. Using community-based samples in Nigeria, Jamaica, and the United States, we previously demonstrated that metabolic processes represented by REE appear to mediate the effect of body size on BP [13]. In the current study, we aimed to confirm that finding in four additional independent population samples in the African diaspora. In this second phase the community-based samples were also drawn from a wide spectrum of social and cultural contexts. The findings reported here strongly confirm the association between REE and BP.

For the present study, the analytical data set included all participants who underwent REE and body composition measurements at the same time, and had valid BP measurements. The final sample included 289 participants of which 69 were from Ghana, 72 from South Africa and 72 from Seychelles, and 76 from the US.

Baseline participant characteristics are presented in Table 1. Women were slightly older than men (33.6 ± 5.9 vs. 35.0 ± 6.0 yrs., p < 0.05), and at all sites had on average higher BMI than men and in all but 1 site weighed more than men. As expected, men had higher mean fat free mass levels. Men also had higher mean systolic and diastolic BP measurements. The mean systolic BP (SBP) among men was 121.9 ± 15.4 mmHg, and among women was 111.4 ± 16.2 mmHg. The mean diastolic BP (DBP) among men was 73.3 ± 12.2 mmHg, and among women was 71.8 mmHg. Men and women from the US, and South Africa presented with the highest SBP and DBP’s. Women presented with higher resting pulse values (p < 0.001), but which were similar across sites among both men and women (p = NS). With regards to PA levels, across all sites, participants spent approximately 33.9 ± 27.7 min/d in moderate-and-vigorous activity (1-min bouts), very little time doing any vigorous activity (3.4 ± 6.8 min/d) and approximately 4 h a day being sedentary (228.5 ± 50.8 min/d).

Biochemical results are presented in Table 2. The majority of participants presented with healthy biochemical measures. The mean glucose levels across all sites were 93.2 ± 19.1 mg/dL, with participants in Ghana presenting with the highest values (100.6 ± 9.7 mg/dL). Mean insulin levels were 17.6 ± 10.5 mlU/L and lowest among the Ghanaians (13.7 ± 6.0 mlU/L). CRP values were highest among the 2 African sites, compared to the US and Seychelles, whereas measures of lipid metabolism were lowest among South Africans and Ghanaians.

Data concerning participants’ medical background and behaviors are presented in Table 3. Overall, 12.8% of the participant sample had hypertension. Among men; 24% of Americans were hypertensive compared to 18% for Seychellois, 11% for South Africa and only 3% among Ghanaians (p = 0.094). Similarly, 24% of US women were hypertensive compared to 16% of South Africans, 3% of Seychellois and Ghanaian women (p < 0.01). 45.7% of the participant population met the U.S. General Guidelines for PA. As expected, a significantly higher percentage of men met the U.S. General Guidelines than women (65 vs. 35%, p < 0.001). 31.6% of the participant sample indicated that they smoke, with more men indicating that they smoked compared to women (45.4 vs. 14.5%). Approximately 46.7% of the participant population reported that they consumed alcohol, with twice as many men indicating alcohol consumption than women (64.6 vs. 32.1%). Men on average reported consuming 1.8 ± 3.0 units/day and women reported 0.68 ± 1.7 units/day.

Partial correlations were examined between REE, SBP, DBP, pulse, weight, fat-free mass, and age for men and women and are represented in Table 4. Among men, REE (kcal/d) was significantly correlated with both SBP and DBP (mmHg), weight (kg), fat free mass (kg) and the adipokines (leptin and adiponectin). Similarly, among women, REE was significantly correlated with the same variables, and additionally pulse (bpm), but not adiponectin.

Multiple regression analyses were conducted to determine relationships between BP and REE, age, gender, body mass index, fat free and fat mass, resting pulse, and location of site. We present two models, Model 1 included BMI, and Model 2 included fat and fat free mass levels. These results are shown in Table 5. For Model 1, after adjusting for BMI, age, gender, pulse, and site, having higher REE (kcal/d) was significantly associated with greater systolic BP (β = 0.019, p = 0.001) and strongly associated with diastolic BP (β = 0.006, p = 0.056) BP (mmHg). Similarly, adjusting for fat and fat free mass, instead of BMI, indicated that those with higher REE (kcal/d) also had significantly higher systolic BP (β = 0.019, p = 0.001) and a strong association with diastolic BP (β = 0.007, p = 0.056).

In this report, we found an independent association between REE and BP among five populations with different mean BP levels. As noted earlier, the obesity syndrome should be conceptualized as a complex, inter-correlated set of external exposures and internal physiologic responses [13, 18]. Separating cause of effect relationships is therefore inherently difficult. Our findings suggest that the measure of REE captures a physiologic process more tightly associated with BP than basic estimates of weight-for-height, as modeled by BMI. The most likely candidate linking REE with BP is circulating catecholamines [44]. Although inconsistent, a substantial body of literature has demonstrated elevated catecholamines in hypertensive subjects [45‐47]. Some investigators have likewise argued that increased sympathetic activity secondary to hyperinsulinemia is a feature of the obesity syndrome [48, 49]. Palatini et al. [50] found that among 3 international adult cohorts, fasting insulin and post 75-g glucose load correlated positively with higher heart rates. There is also some evidence that obese adults are more likely to present with increased urinary noradrenaline levels [51] and plasma levels [52]. Weight loss studies provide evidence supporting the catecholamine hypothesis in obesity. Among twenty-five obese individuals studied during a short-term weight loss intervention, a reduction in plasma norepinephrine was closely associated with reductions in REE and BP [14]. Leptin has also been associated with thermogenesis and hypertension and is markedly elevated in the obese [53]. While not a significant predictor in the current study, the role of other inflammation markers such as CRP has also been cited as a significant predictor of future CVD. For example, several studies find a relationship between adipocytokines, CRP and risk for CVD [54, 55]. Seven et al. [54] explored the relationship between leptin, adiponectin, CRP and risk for CVD in over 6500 adults (30-60 yrs). After 11 years of follow-up only CRP was found to be significantly associated with an increased CVD risk. Similarly, Barbero et al. [55] in a meta-analysis of almost 105,000 adults followed for 5 years, found that CRP, together with history of diabetes, past myocardial infarction, and being male were the most significant predictors for future CVD.

In attempting to disentangle correlated physiologic relationships, it can be useful to consider contextual patterns to determine whether they support or contradict the specific hypothesis being examined. REE would require increased cardiac output, which in the long-term could predispose to hypertension [56, 57]. For example, catecholamines increase with age, as does BP, and elevated heart rate has long been recognized as a predictor of hypertension [56]. Physical activity and training effects, however, are inconsistent with a straightforward causal link between REE and BP. REE tends to increase with aerobic fitness, while BP and pulse rate decline [58, 59]. Likewise, weight lifters and other athletes have marked increases in lean body mass, but avoid the obesity-related effect of high BP [60]. We have also demonstrated that the association between BMI and BP is weaker among the obese than the non-obese, suggesting that lean body mass is the more influential component of total body size [61]. It is nonetheless possible that a catecholamine-mediated effect that leads to increases in both REE and BP, could be over-ridden by other compensatory mechanisms after aerobic training. A simpler non-mechanistic explanation of why REE remains significantly associated with BP in our analyses, and BMI is not, could be a result of differential precision of the measurements. BMI captures about 40–50% of the variation in lean body mass, while repeat REE measurements for that person have a correlation of 0.95, reflecting very high precision. In multivariate models, a secondary variable measured with little error can achieve significance “independently” of the underlying causal effect, particularly if other covariates in the model, which may have a true causal link with the outcome, lack precision.

In our study, while we attempted to ensure consistency and reliability of measurements, particularly with regards to the measurement of BP and REE across our research sites, our study is not without limitations. For example, we have no measure of circulating catecholamines, an issue which we have now addressed in our current studies. In order to maximize comparability of REE measurements we used the same indirect calorimeter in three of the four research sites, and conducted regular methanol burns as described by Cooper et al. [62]. With regards to the measurement of BP, the same model Omron was used in all sites and all clinic staff underwent the same in-house training prior to the start of the data collection period, and as previously described by Luke et al. [26]. Due to the nature of multi-site research, there is always the potential for site-to-site variability in some measures. The investigators of the current study worked hard to minimize that variability.

In summary, our study provides further evidence of a strong association between REE and BP, and a null or modest role of BMI or other adiposity measures on BP when REE is considered, which has not been fully appreciated so far. These findings expand our understanding of the complex of metabolic and physiologic abnormalities observed in the obesity syndrome.