Obesity, Anthropometric Measures and Chronic Kidney Disease Complications

Chronic kidney disease (CKD) and obesity are major public health problems [1,2,3]. Epidemiological investigations have identified obesity as an independent risk factor for the development of CKD and the progression of CKD to ESRD in the general population [4]. Potential mechanisms that might mediate these associations include insulin resistance, increased inflammation, higher oxidative stress and endothelial dysfunction seen among those who are obese [5]. While body mass index (BMI), an anthropometric measure of whole body adiposity has been associated with the development of CKD, previous studies that examined the association of BMI and mortality showed paradoxical results in dialysis and renal transplant recipients [6,7]. Even though widely used, a major limitation of the BMI measure is its inability to differentiate fat mass and muscle mass which may have opposite associations with mortality [8,9].

Waist circumference (WC), a measure of visceral or abdominal adiposity is associated with subclinical inflammation independent of BMI [10]. In the general population, while some studies report that measures of abdominal adiposity are better predictors of cardiovascular disease and mortality, others claim that central adiposity measures such as WC and waist-to-hip ratio (WHR) do not provide more prognostic information than BMI alone [11,12,13]. In those with CKD, population-based studies suggest that measures of abdominal adiposity such as WC and WHR, rather than BMI, are better predictors of adverse outcomes such as mortality [14]. However, the association between the different measures of adiposity such as BMI and WC and complications arising from CKD has not been explored in previous studies. We hypothesized that obesity is associated with different CKD complications and that anthropometric measures such as BMI and WC have differential associations with CKD complications. As such, we studied the associations of WC and BMI with CKD complications such as anemia, secondary hyperparathyroidism, hyperphosphatemia, metabolic acidosis, hypoalbuminemia and hypertension among a nationally representative sample of US adults.

We examined data from the National Health and Nutrition Examination Survey (NHANES), a nationally representative, complex and multistage probability of the US civilian, noninstitutionalized population conducted by the National Center for Health Statistics. The National Center for Health Statistics Ethics Review Board approved the study protocol and each participant provided written informed consent. Participants in the NHANES were interviewed in their homes and underwent a standardized physical examination in a mobile examination center. Self-reported information on demographics, socioeconomic status, health conditions, health behaviors and routine site of healthcare were obtained during the interview. The examination component consists of medical, dental and physiological measurements, as well as laboratory tests administered by highly trained medical personnel. The four 2-year cycles of the continuous NHANES, 1999–2000, 2001–2002, 2003–2004 and 2005–2006, were combined for this analysis. The study included 2,853 participants who met the following criteria: 20 years of age and older who (a) underwent medical examination, (b) were not pregnant, (c) had BMI and WC measured, (d) had a BMI ≥18.5, (e) had serum creatinine level and albumin-to-creatinine ratio results, (f) were not on dialysis and (g) had an estimated glomerular filtration rate (eGFR) ≥15 ml/min/1.73 m².

Kidney Disease and Comorbidities

Participants with stage 1–4 CKD who were not on dialysis were included [stage 1–2 CKD: those with a urine albumin- to-creatinine ratio (UACR) ≥30–299 mg/g (microalbuminuria) and ≥300 mg/g (macroalbuminuria) with eGFR ≥60 ml/min/ 1.73 m² or stage 3–4 CKD: eGFR 15–59 ml/min/1.73 m²]. The eGFR was calculated according to the Chronic Kidney Disease Epidemiology Collaboration equations using calibrated creatinine [15,16]. Serum creatinine levels were corrected for the 1999–2000 and 2005–2006 surveys as suggested in the NHANES serum creatinine documentation [16]. The UACR was calculated from spot urine albumin and creatinine samples obtained. Diabetes was defined as self-reported if ever told by a doctor that the participant had ‘diabetes or borderline diabetes’. Hypercholesterolemia was defined as the presence of total cholesterol >200 mg/dl and/or use of cholesterol-lowering medications. Liver disease was defined by the answer ‘yes’ to the question, ‘Have you ever been told that you had any liver condition?’.

Adiposity Measures

Height and weight measurements were collected by trained health technicians using standardized techniques and equipment during the health examination conducted in each survey period. WC was measured midway between the lowest rib and the iliac crest with the participant standing. BMI was calculated as weight in kilograms divided by the measured height in meters squared and participants were classified as obese if they had a BMI ≥30. WC was categorized as high risk for men with a measured WC >102 cm and for women with a measured WC >88 cm. All others were considered low risk [17]. We also considered four different categories based on their BMI and WC data: (a) BMI ≥30 and high WC, (b) BMI ≥30 and low WC, (c) BMI <30 and high WC and (d) BMI <30 and low WC (fig. 1).

CKD Complications

Anemia was defined as hemoglobin <13.5 g/dl for males and <12 g/dl for females [18]. A serum bicarbonate level <21 mEq/l was defined as the presence of metabolic acidosis and parathyroid hormone (PTH) >70 pg/ml was defined as secondary hyperparathyroidism [19]. Hyperphosphatemia was defined as the presence of serum phosphorus >4.5 mg/dl [19]. Hypoalbuminemia was defined as a serum albumin level <4 g/dl. Hypertension was defined as the presence of systolic blood pressure >140 mm Hg and/or diastolic blood pressure >90 mm Hg or the use of antihypertensive medication.

Demographics, comorbidities and CKD complications were compared between participants with BMI ≥30 and <30 and with high (WC >102 cm in males and >88 cm in females) and low WC using t tests for continuous variables and Rao-Scott χ² tests for categorical variables. Results were reported as mean ± standard error for continuous variables and percent (95% CI) for categorical variables. We examined the associations between different anthropometric measures such as BMI (for every 2-point increase) and WC (for every 5-cm increase) and CKD complications while adjusting for age, gender, African American race, diabetes, hypertension, liver disease (for hypoalbuminemia outcome only), smoking, UACR and eGFR.

Logistic regression analysis was also used to assess the associations between individual CKD complications and BMI ≥30 while adjusting for the above-mentioned variables. A sensitivity analysis that compared BMI ≥30 to BMI 25–29.9 was also conducted. Similarly, logistic regression analysis was used to assess the associations between individual CKD complications and a high WC (>102 cm in males and >88 cm in females) while adjusting for the relevant confounding variables mentioned above. BMI and WC were not included in these models due to collinearity (unweighted correlation = 0.87).

In addition, logistic regression analysis categorized participants as: (a) BMI ≥30 and high WC (obese based on both criteria) versus participants with BMI <30 and low WC (nonobese based on both criteria) and (b) participants with BMI <30 and high WC versus participants with BMI <30 and low WC. This analysis was conducted while adjusting for the above-mentioned variables. We could not conduct a separate analysis for those with BMI ≥30 and low WC versus participants with BMI <30 and low WC due to the small sample size in the former (n = 22).

All analyses were performed using survey procedures with SAS version 9.2 for Unix (SAS Institute, Cary, N.C., USA), which accounts for the sampling design of NHANES and appropriately weighted participants in statistical models. Graphs were produced using R version 2.12.2 (The R Foundation for Statistical Computing, Vienna, Austria).

CKD participants with BMI ≥30 were younger [58.0 (0.7) vs. 62.7 (0.7) years] and had more non-Hispanic blacks (14.6 vs. 9.0%) than CKD participants with BMI <30. A higher proportion of CKD participants with BMI ≥30 had diabetes (p < 0.001; table 1). The mean WC of CKD participants with BMI ≥30 was 115.6 (0.6) cm compared to 92.4 (0.3) cm among CKD participants with BMI <30 (p < 0.001). The mean eGFR was 75.1 ml/min/ 1.73 m² for those with BMI ≥30 versus 71.2 (1.0) ml/min/ 1.73 m² for the BMI <30 group (p = 0.003). The UACR was higher among those with BMI ≥30 than among those with BMI <30 (p = 0.03). Table 1 lists additional participant characteristics based on their BMI.

CKD participants with high WC (>102 cm in males and >88 cm in females) were older [62.2 (0.6) vs. 58.6 (1.0) years] with a lesser proportion of males (37.9 vs. 52.8%) than CKD participants with a low WC. A higher proportion of those with high WC had diabetes and hyperlipidemia (p < 0.001; table 2). The mean BMI of CKD participants with high WC was 32.3 (0.2) compared to 23.9 (0.1) among those with low WC. The mean eGFR was 71.2 ml/min/1.73 m² for participants with high WC versus 75.7 (1.3) ml/min/1.73 m² for those with low WC (p = 0.004). The UACR was similar between those with high and low WC (p = 0.21). Table 2 lists additional participant characteristics based on their WC.

An increase in BMI by 2 points was associated with increased odds of having secondary hyperparathyroidism, hypoalbuminemia and hypertension (table 3). Similarly, an increase in WC by 5 cm was associated with increased odds of having secondary hyperparathyroidism, hypoalbuminemia and hypertension but also with lower odds of having anemia (table 3).

BMI ≥30 versus BMI <30

CKD participants with BMI ≥30 have higher odds of hypoalbuminemia and hypertension than CKD participants with BMI <30 when adjusted for relevant confounding variables (table 4). There was a nonsignificant increase in the odds for secondary hyperparathyroidism among those with BMI ≥30 (table 4). There were no associations between BMI ≥30 and hyperphosphatemia, metabolic acidosis and anemia. Similar associations were noted when BMI ≥30 was compared to BMI 25–29.9.

High versus Low WC

CKD participants with high WC have higher odds of hypoalbuminemia and hypertension and lower odds of having anemia than CKD participants with low WC when adjusted for relevant confounding variables (table 4). There were no associations between high WC and hyperphosphatemia, secondary hyperparathyroidism and metabolic acidosis.

BMI ≥30 and High WC versus BMI <30 and Low WC

CKD participants with BMI ≥30 and high WC (who would meet both the criteria of obesity) have higher odds of secondary hyperparathyroidism, hypoalbuminemia and hypertension than participants with BMI <30 and low WC when adjusted for relevant confounding variables (fig. 2).

BMI <30 and High WC versus BMI <30 and Low WC

CKD participants with BMI <30 and high WC were not associated with any increased likelihood of CKD complications when compared to participants with BMI <30 and low WC when adjusted for relevant confounding variables (fig. 2).

Our analysis of representative US adults with kidney disease shows that the anthropometric obesity measures such as BMI and WC are associated with different CKD complications. These associations are similar when examined as continuous variables. However, when examined as a categorical variable from standard definitions of obesity based on BMI and WC measures, there were differential associations. Participants with BMI ≥30 and high WC (obese based on both criteria) had higher odds of secondary hyperparathyroidism, hypoalbuminemia and hypertension than participants who had low WC and BMI <30 (nonobese based on both criteria). In addition, among those with BMI <30, the presence of high WC was not associated with any CKD complications.

Previous studies have explored the associations between various adiposity measures and incident CKD. Secondary analysis of the Cardiovascular Health Study and Atherosclerosis Risk in Communities Study showed that WHR but not BMI was associated with incident CKD and mortality [20]. In a nondiabetic cohort, WC and BMI but not WHR were associated with lower eGFR [21]. Studies exploring the associations of different anthropometric measures with kidney disease progression and mortality are limited. Secondary analysis of the Atherosclerosis Risk in Communities Study showed that the associations of obesity with mortality differed by the presence or absence of CKD, suggesting that kidney disease does not reverse the metabolic effects of obesity [22]. Recently, Kramer et al. [14] reported higher mortality rates for WC >94 cm in men and >80 cm in women, while no such associations were noted for higher BMI categories among 5,805 CKD participants. Our study adds to this growing body of evidence regarding the utility of these different adiposity measures in the CKD population. In addition, these findings highlight the different results that might be obtained when examining these anthropometric variables as continuous and categorical variables.

Associations between BMI ≥30 and secondary hyperparathyroidism have been reported previously, but the associations of BMI and WC with other CKD complications have not been explored. Kovesdy et al. [23] reported that higher BMI levels were associated with higher PTH levels among patients who were not dependent on dialysis. Similarly, secondary analysis of the Kidney Early Evaluation Program data showed that higher BMI was associated with high PTH levels independent of confounding variables [24]. WC data was not available in these studies. In addition to the findings from previous studies, we found higher odds of having PTH >70 pg/ml with both increasing BMI and WC. Some proposed explanations that may explain these associations include lower 25(OH)D levels and lower physical activity (thus lower sun exposure) noted among those with higher BMI and WC. In addition, the decreased bioavailability of vitamin D due to its deposition in adipocytes among those who are obese might also contribute to the higher rates of secondary hyperparathyroidism [25].

When examined as both continuous and categorical variables, participants with a high BMI and WC had higher odds of having hypoalbuminemia among CKD participants. Obesity is a state of inflammation and thus hypoalbuminemia might reflect the higher inflammatory burden among those who are obese rather than nutritional status alone [26,27]. We also examined whether the presence of BMI ≥30 and high WC were associated with higher Hs-CRP levels. Hs-CRP levels were higher among those with BMI ≥30 and high WC and this data supports the inflammation hypothesis (tables 1 and 2). We planned to conduct a sensitivity analysis by defining hypoalbuminemia as serum albumin <3.5 g/dl but the relatively small number of participants precluded this analysis.

As expected, the presence of higher BMI and WC were associated with increased odds of having hypertension among CKD participants, and this data is similar to that seen in the general population. However, there was a negative association between high WC and anemia noted in this analysis. To our knowledge, no studies have reported associations between anemia and obesity in both CKD and non-CKD populations except that a population-based study reported no significant associations between anemia and obesity in the absence of CKD [28]. The exact mechanisms that might explain these associations are unclear at present and these associations should be reexamined and confirmed in future studies.

BMI data is readily available in clinical settings and whether WC details provide additional prognostic information relating to CKD complications in this population is unclear. Hence, we examined the associations between BMI <30 and high WC and CKD complications. Our results suggest that BMI data alone might be sufficient to predict CKD complications in those with lower BMI. We could not conduct an analysis among participants with BMI ≥30 and low WC versus BMI <30 and low WC to examine the utility of BMI alone due to the small sample size.

One strength of this study was the inclusion of a nationally representative sample with adequate representation of various ethnic groups in the United States. This analysis is subject to limitations that include the cross-sectional study design of NHANES surveys that precludes the opportunity for studying temporal associations. Also, the use of a single UACR measurement in the NHANES surveys might lead to misclassification of CKD, especially among participants with early-stage CKD. Further, the number of participants with advanced CKD is minimal and whether these associations exist among those with advanced kidney disease needs to be examined in future studies. We also did not have body composition measures such as DEXA or computed tomography measures which would provide more detailed insight into the role of the body composition.

In summary, anthropometric measures such as BMI and WC are associated with secondary hyperparathyroidism, hypoalbuminemia and hypertension among those with CKD. Higher WC among those with BMI <30 is not associated with CKD complications. These associations suggest the need for closely monitoring the development of complications relating to CKD among those who are obese.

S.D.N. 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 No. RR024990). J.P.K. is supported by the National Institutes of Health (RO1 DK089547). J.D.S. is supported by NIH/NIDDK (R01 DK085185) and investigator-initiated grant support from PhRMA foundation, Genzyme and Roche Organ Transplant Research Foundation. M.J.S. is supported by K24 DK078204. The content of this paper is solely the responsibility of the authors and does not necessarily represent the official views of the NIH. The authors have no relevant financial interest in the study.