Background
Obesity rates have increased in the past few decades along with an increase in comorbid disease states related to obesity and health care expenditures [
1]. In the general population, being overweight (BMI 25–29.9 kg/m
2) and class I obesity (BMI 30–34.9 kg/m
2) are associated with lower mortality [
2]. Some studies have also suggested that measures of abdominal adiposity such as waist circumference (WC) and waist-to-hip ratio might be better predictors of adverse outcomes in the general population [
3,
4]. Higher BMI is associated with an increased risk for development and progression of kidney disease [
5,
6]. However, reports indicate that higher BMI is associated with lower mortality in those with established CKD (reverse epidemiology) [
7,
8]. Recent population based studies have shown that WC rather than BMI might be a better measure for predicting mortality in those with CKD [
9]. This could be attributed to the inability of BMI to distinguish muscle mass and fat mass as differential associations between muscle mass, fat mass and mortality might exist. However, studies examining the relationship between body fat, lean body mass (measured using dual-energy x-ray absorptiometry [DEXA], CT scan or other modalities) and mortality in those with CKD are lacking.
Several studies in the general population reported a lower risk for death among individuals with higher levels of physical activity [
10‐
13]. Analyses using NHANES III data (1988–1994) suggested that inactive or insufficient physical activity levels are associated with lower death among those with and without kidney disease [
14,
15]. However, these analysis lacked fat mass and fat free mass data and whether higher physical activity levels have similar benefits despite accounting for fat mass and lean body mass is unknown. We hypothesized that obesity as measured by - higher percent fat mass, lower lean body mass and lower leisure time physical activity (LTPA), are associated with increased mortality in those with and without CKD. Therefore, we studied the associations between adiposity measures (BMI, WC and percent body fat), lean body mass and LTPA with all-cause mortality among a nationally representative sample of US adults.
Discussion
In this nationally representative data from the general US population, various measures of adiposity including total and percent body fat were not associated with death in those with and without kidney disease. Higher lean body mass was associated with 18% lower risk for death in those without kidney disease but lean body mass was not associated with death in CKD. Irrespective of kidney function, those who did not meet LTPA goals were at higher risk for death in this cohort. The associations between adiposity measures seem to be similar irrespective of age, gender and race, but gender did modify the associations between LTPA and death in those without CKD (higher risk in females but not in males).
In the general population, pooled analysis of several studies reported that in comparison to normal weight category, being overweight (BMI 25–29.9 kg/m
2) or with class I obesity (BMI 30.0-34.9 kg/m
2) were not associated with higher mortality [
2]. Overall, our results relating to BMI were similar to previous reports, but we did not find an association between WC and death as reported in earlier studies [
23‐
26]. These differences could be attributed to the shorter follow-up data available in NHANES compared to other reports. Absence of association between percent and total body fat and mortality in the general population may be explained by the differential associations between visceral and subcutaneous adiposity [
27,
28]. Body fat distribution is one of the major determinants of metabolic health, and visceral adiposity has a stronger correlation with metabolic abnormalities and cardiovascular disease than subcutaneous adipose tissue [
27‐
29]. Visceral fat is metabolically active and is an important site for adipokines such as adiponectin and leptin, that can modulate inflammation and insulin resistance, and impart cardiovascular risk [
30].
Even though studies have shown a relationship between obesity and initiation and progression of kidney disease, studies linking obesity (defined as a BMI ≥30 kg/m
2) and mortality have shown paradoxical results in the CKD population [
7,
8,
31]. Thus the lack of association between higher BMI and mortality is not surprising. BMI, a composite measure of muscle, and subcutaneous and visceral adipose tissue, is widely used to diagnose obesity in clinical practice. Therefore, it has low sensitivity and is likely to give unreliable results in CKD patients who are often old, frail, and lack the muscle mass of healthy individuals. In addition, lower GFR is associated with fluid retention, which further confounds the utility of anthropometric measurement such as BMI in CKD. Data from the REGARDS cohort suggested higher mortality rates for those with higher WC but not with high BMI [
9]. However, we did not observe a higher risk for death with higher WC and this may be attributed to the small sample size of the NHANES sample (2153 vs. 5805 in the REGARDS cohort). Preliminary data also indicate that visceral fat may be linked to coronary artery vascular calcification in CKD, highlighting the need for future studies assessing visceral fat in the CKD population [
32,
33].
Higher lean body mass, a reflection of higher muscle mass, was associated with lower mortality. Few studies have reported this finding in other cohorts without kidney disease and our findings emphasize the importance of preserving muscle mass in the aging population [
34]. Higher physical activity is associated with higher lean body mass, and this may be negatively associated with cardiovascular risk factors including diabetes and hypertension, which could explain the lower risk for death. Even though studies using urinary creatinine measurements (a proxy for muscle mass) reported higher risk for death in diabetic nephropathy patients with lower muscle mass, we did not find such associations with lean body mass in the CKD population [
35]. Future studies should examine the associations between muscle mass assessed using MRI or CT, with death, in those with and without kidney disease.
Physical activity improves insulin sensitivity and endothelial function and lowers inflammatory markers thereby rendering cardiovascular benefits [
36,
37]. Higher levels of physical activity are linked to an overall beneficial impact on cardiovascular disease and all-cause mortality in the general population [
10‐
12,
38,
39]. Greater physical activity is associated with lower albuminuria in nondiabetic women and higher levels of physical activity are associated with a lower risk of decline in kidney function among adults >65 years after accounting for other comorbid conditions [
40]. In the NHANES III cohort, physical inactivity was associated with increased mortality in CKD and non-CKD populations [
14,
15]. Cumulatively, the data suggests that physical activity level, a surrogate measure of their physical fitness, may adversely affect outcomes in those with and without kidney disease. In addition, there are recent reports that longer sitting time is associated with cardiovascular disease and death in the general population, however the lack of consistent data collection in NHANES over the years precluded further analyses [
41].
Studies enrolling diabetics and heart failure patients have reported that both fitness and fatness influence cardiovascular risk factors [
42,
43]. A higher physical fitness levels among those with a higher BMI is associated with a lower prevalence of cardiovascular risk factors and mortality, compared to those with normal BMI and lower fitness level [
43‐
46]. This may also partly explain the obesity paradox noted in heart failure and diabetic patients. It is unclear if fitness and fatness are differentially associated with the individual cardiovascular risk factors and outcomes in CKD. Physical fitness (based on VO
2max measures) was assessed in only a limited number of NHANES participants precluding further sensitivity analysis to address this issue.
Strengths of this analysis include the availability of a nationally representative data sample with adequate representation of various ethnic groups in the United States, and availability of longitudinal data to study mortality. In addition, the use of DEXA to directly determine body composition provides a measure that has been validated and correlates highly with other reference methods [
47]. However, this analysis is subject to limitations that include the use of a single UACR measurement in the NHANES surveys, which may lead to misclassification of CKD, especially among participants with early-stage CKD. Understanding the limitations, we used eGFR from single serum creatinine measurement. Further, the number of participants with advanced CKD is relatively small (Additional file
1: Table S1) and whether these associations exist among those with advanced kidney disease needs to be confirmed in future studies. In addition, availability of long-term follow-up data might uncover associations that may have been missed in the current study, as some of the ill-effects of adipose tissue take many years to manifest. In addition, we did not obtain visceral adiposity data, which would have helped distinguish the harmful effects of visceral adiposity over subcutaneous adiposity. When we adjusted for diabetes and hypertension, the results remained similar for the non-CKD population, but lost statistical significance for those with CKD (HR 1.32, 95% 0.99, 1.76). These factors were not adjusted for in the multivariate models as lie in the causal pathway between obesity, low physical activity and death in CKD population [
48].
Competing interests
SDN is supported by a career development award from the National Center for Research Resources and the National Center for Advancing Translational Sciences, National Institutes of Health (Grant #TR000440). JPK is supported by the National Institutes of Health - RO1 DK089547-01. JDS is supported by NIH/NIDDK (R01 DK085185-01A1) and NIH/NIMH (P60MD00265). Its contents are solely the responsibility of the authors and do not necessarily represent the official view of NCRR or NIH. Information on Re-engineering the Clinical Research Enterprise can be obtained from
http://nihroadmap.nih.gov/clinicalresearch/overview. The authors have no relevant financial interest in the study.
Authors’ contributions
Concept and design of the study: SDN, SA, JDS. Data analysis: SDN, JDS, SA. Interpretation of data and critical revision for intellectual content: SDN, JPK, JDS, SA, Writing the final manuscript, final approval of version to be published: SDN, JPK, JDS, SA. SDN and SA had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. All authors read and approved the final manuscript.