Current body mass index (BMI) strata likely misrepresent the accuracy of true adiposity in older adults. Subjects with normal BMI with elevated body fat may metabolically have higher cardiovascular and overall mortality than previously suspected. We identified 4,489 subjects aged ≥60 years (BMI = 18.5 to 25 kg/m 2 ) with anthropometric and bioelectrical impedance measurements from the National Health and Nutrition Examination Surveys III (1988 to 1994) and mortality data linked to the National Death Index. Normal weight obesity (NWO) was classified in 2 ways: creation of tertiles with highest percentage of body fat and body fat percent cutoffs (men >25% and women >35%). We compared overall and cardiovascular mortality rates, models adjusted for age, gender, smoking, race, diabetes, and BMI. The final sample included 1,528 subjects, mean age was 70 years, median (interquartile range) follow-up was 12.9 years (range 7.5 to 15.3) with 902 deaths (46.5% cardiovascular). Prevalence of NWO was 27.9% and 21.4% in men and 20.4% and 31.3% in women using tertiles and cutoffs, respectively. Subjects with NWO had higher rates of abnormal cardiovascular risk factors. Lean mass decreased, whereas leptin increased with increasing tertile. There were no gender-specific differences in overall mortality. Short-term mortality (<140 person-months) was higher in women, whereas long-term mortality (>140 person-months) was higher in men. We highlight the importance of considering body fat in gender-specific risk stratification in older adults with normal weight. In conclusion, NWO in older adults is associated with cardiometabolic dysregulation and is a risk for cardiovascular mortality independent of BMI and central fat distribution.
Using data from the Third National Health and Nutrition Examination Surveys, a subset of participants with normal body mass index (BMI; 18.5 to 25 kg/m 2 ) but elevated percentage of body fat among United States adults was shown to be at higher risk of cardiometabolic dysregulation, endothelial dysfunction, insulin resistance, metabolic syndrome, and possibly mortality. With age-related alterations in body composition, it is unclear whether these conclusions apply in a noninstitutionalized older population. Our aim was to ascertain whether high body fat in subjects aged ≥60 years contributes to cardiometabolic dysregulation and mortality, similar to that reported in the general population.
Methods
We identified participants aged ≥60 years from the National Health and Nutrition Examination Surveys III (1988 to 1994) and its publically available mortality dataset. The survey is complex, stratified survey design that oversamples minorities and older adults, representative of noninstitutionalized adults of the United States. Mortality data were linked using a unique identifier based on death certificate data from the National Death Index and updated through December 31, 2006. Study design, sampling characteristics, and survey components are available on http://www.cdc.gov/nchs/nhanes.htm and were downloaded in January 2012. The survey was administered by the Centers for Disease Control and was exempt from the Institutional Review Board review because of the de-identified data and strict nature of the data procedures.
Of the 39,695 sampled subjects, we excluded 31,320 aged <60 years. All sampled participants were asked to be interviewed and examined by a doctor at a mobile examination unit. Trained staff performed interviews and asked questions directly to subjects or where appropriate to their proxies and data collection was automated. All interviews were conducted by field staff in both English and Spanish and they received intensive initial training with retraining throughout the survey. Of the remaining examined subjects (n = 6,178), we excluded participants without complete anthropometric measurements and contraindications to bioelectrical impedance (n = 1,694) leaving 4,484 subjects in our sample. All races (non-Hispanic white, non-Hispanic black, and Hispanic American) were included. Date of birth was self-reported both at the initial screening questionnaire and also at examination and verified against an age verification chart, with differences reconciled per protocol. Age at screening was used for baseline measurements, and all included participants were eligible for mortality follow-up.
Weight was measured in kilograms on a calibrated digital scale, and standing height was measured in centimeters with the participant’s back to a stadiometer’s backboard, with the subject’s weight distributed evenly on both feet after deep inhalation. Waist circumference was measured by palpating the right iliac crest, crossing the midaxillary line, and placing the measure around the trunk’s body at minimal respiration in the standing position. Hip circumference was ascertained at the side seams of the pants above the hips, at maximum extension of the buttocks, with the tape measure held snuggly. All procedures were performed on the right side of the body except where physically incapable of doing so. For quality assurance, replicates and data review were performed. Waist-hip ratio was obtained using the ratio of waist and hip circumference. BMI was calculated as weight (kg) divided by height (m) squared and grouped into standard categories (≤18.5, 18.5 to 25, 25 to 29.9, and ≥30 kg/m 2 ). Only subjects in the normal BMI category (18.5 to 25 kg/m 2 ) were included (n = 1,528).
Using a Valhalla 1990B Bio-Resistance Body Composition Analyzer (Valhalla Scientific, San Diego, California), bioelectrical impedance values were obtained. Subjects were asked to avoid eating or drinking anything except water during the fasting period. Participants had a single tetrapolar measurement of resistance and reactance at 50 kHz taken between the right wrist and ankle while in supine position. Reactance values were not used in subsequent analyses. Data were converted using gender-specific predictive equations from a separate independent cohort to RJL resistance values to determine fat-free mass and total body water. Converted RJL resistance values were used to calculate impedance index of stature squared divided by the resistance previously mentioned. Calculated total body fat was the difference between lean mass and weight (kg). The quotient of total body fat and weight multiplied by 100 was termed percent body fat.
Lipids were measured using commercially available reagents (Cholesterol/HP, cat. no. 816302 and triglycerides/GPA, cat. no. 816370; Boehringer Mannheim, Indianapolis, Indiana). High-density lipoprotein cholesterol was measured in a clear supernatant after precipitating other lipoproteins with heparin and MnCl 2 and removing excess Mn 2+ by precipitation with NaHCO 3 . Standard assay measured glucose (Sigma Chemical, St. Louis, Missouri), and a radioimmunoassay kit measured insulin (Pharmacia Diagnostics, Uppsala, Sweden). Both apolipoprotein B and apolipoprotein A1 were measured by radial immunodiffusion in the first 8.2% of specimens during the first 5 months of the study and by rate immunonephelometry for the remaining period, and leptin was measured using radioimmunoassays (Linco Research, St. Charles, Missouri). Measurement of C-reactive protein used a modified latex-enhanced assay (Behring Diagnostics, Westward, Massachusetts). The homeostatic model assessment index model calculated insulin sensitivity and β-cell function according to an online computer program at http://www.dtu.ox.ac.uk/homacalculator/ .
We classified subjects as having hypertension according to previously published criteria (blood pressure ≥140/90 mm Hg) or patients with a self-reported diagnosis of hypertension on antihypertensive medications. Dyslipidemia was defined if subjects satisfied any of the following: total cholesterol ≥240 mg/dl, high-density lipoprotein ≤40 mg/dl in men and ≤50 mg/dl in women, or low-density lipoprotein ≥160 mg/dl or those with a self-reported diagnosis of dyslipidemia on lipid-lowering medications. Subjects were diabetic if they reported current usage of antidiabetic medications (insulin and oral medications), a self-reported diabetes diagnosis, and/or if their glycosylated hemoglobin was ≥6.5%, and/or if their fasting morning plasma glucose was ≥126 mg/dl. Metabolic syndrome was defined if subjects had ≥3 components: waist circumference (men ≥102 cm, women ≥88 cm) for central obesity, triglycerides (≥150 mg/dl or treatment for elevated triglycerides), high-density lipoprotein cholesterol (men <40 mg/dl, women <50 mg/dl, or on treatment), blood pressure (≥130/85 mm Hg or on treatment in patients with a self-reported history of hypertension), and fasting morning glucose ≥100 mg/dl or those with self-reported diabetes. Smokers answered, “Do you smoke cigarettes”, whereas “ever smokers” included subjects answering “Have you ever smoked at least 100 cigarettes in your lifetime”.
Mortality data consisted of causes of deaths after both the ninth revision of the International Statistical Classification of Disease, Injuries and Causes of Death guidelines (deaths before 1999) and the tenth revision (deaths after 1999). Codes I00 to I78 classified cardiovascular deaths. Person-months were calculated in months from date of interview to date of death or most recent vital status record. Mortality data were complete in >99% of our sample.
We classified subjects as having normal weight obesity (NWO) using 2 separate methods—tertiles and cutoffs. The tertile method divided the subject cohort into 3 groups, based on gender-specific percentage of body fat. In men, the boundaries were <19.5, 19.5 to 24.0, and >24%; in women, they were <29.6, 29.6 to 33.6, >33.6%. Subjects with NWO were defined as those with elevated percentage of body fat based on the upper tertile of body fat percent and a BMI of 18.5 to 25.0 kg/m 2 . Alternatively, NWO was defined as having a BMI of 18.5 to 25.0 kg/m 2 and body fat percent of >25% for men and >35% for women for the cut-off method.
All data were merged into 1 dataset according to survey procedures. Continuous variables are represented as means (standard errors) and categorical variables as counts (percentage). Baseline characteristics of the overall cohort and by age category (60 to 69.9, 70 to 79.9, and ≥80 years) were examined and compared using nonparametric testing for categorical variables and analysis of variance for continuous variables. Three models ascertaining all-cause and cardiovascular-specific mortality were assessed. Model 1 adjusted for age, gender, race, diabetes, and BMI. Model 2 substituted BMI for waist circumference. Model 3 included both variables. Cox proportional hazard modeling modeled the effect of our primary predictor percent body fat (continuous variable) on overall and cardiovascular mortality and replicated using body fat percent cutoffs (low percentage of body fat = referent). The proportional hazard model assumption was not fulfilled as it was not constant for body fat percentage over different time periods. Hence, we performed 2 separate analyses using a censoring time cutoff of 140 person-months and reported hazard ratios (95% confidence interval) for each period. All estimates were weighted and calculated using an SAS (Version 9.2, SAS Institute, Cary, North Carolina) procedure SURVEYPHREG that accounts for a complex survey design, using pseudo-strata, pseudo-primary sampling units, and sampling weights provided by the National Institute for Health Statistics. To account for the complex study design, procedure SURVEYMEANS or SURVEYFREQ were used. Taylor series linearization approach was used for variance estimation. Statistical significance was defined as p <0.05.
Results
Tables 1 and 2 list the gender-specific baseline characteristics of our cohort. Mean age was 70 years. Prevalence of NWO was 27.9% and 21.4% in men and 20.4% and 31.3% in women using tertiles and body fat cutoffs, respectively. Demographic, anthropometric, metabolic, and cellular parameters were all associated with NWO in both genders. Lean mass decreased and leptin levels increased with increasing tertile of body fat. Insulin resistance and sensitivity, C-reactive protein, and β-cell function differed in women with NWO only using tertiles, but using percent body fat cutoffs, differences were observed in both men and women. No differences were observed by age-group (data not shown). Table 3 explores the relation of NWO with metabolic syndrome and its components. There were 902 deaths primarily because of cardiovascular disease (419 deaths [46.5%]). Mortality estimates are listed in Table 4 . No differences were observed in either gender in both short- (<140 person-months) and long-term (≥140 person-months) overall mortality, regardless of the method used to identify NWO (p >0.05). Using percent body fat did not demonstrate any worse mortality in either gender. Using cutoffs, in all models, short-term (<140 person-months) cardiovascular mortality was significantly higher in women with NWO than in those without it, whereas long-term mortality was no different. Interaction terms were nonsignificant in all models. Men had no differences in short-term mortality but had markedly higher long-term cardiovascular mortality ( Table 4 ).
| Variables | Men | Women | ||||||
|---|---|---|---|---|---|---|---|---|
| Low Body Fat Percent Tertile, n = 251 | Normal Body Fat Percent Tertile, n = 260 | High Body Fat Percent Tertile, n = 252 | Age-Adjusted p Value | Low Body Fat Percent Tertile, n = 252 | Normal Body Fat Percent Tertile, n = 260 | High Body Fat Percent Tertile, n = 253 | Age-Adjusted p Value | |
| Age (yrs) | 70.8 ± 1.0 | 69.6 ± 0.5 | 69.6 ± 0.4 | — | 71.5 ± 0.90 | 71.8 ± 0.47 | 70.1 ± 0.6 | — |
| Non-Hispanic white | 168 (80.3) | 203 (86.1) | 185 (83.8) | 214 (94.2) | 210 (89.3) | 211 (92.0) | ||
| Non-Hispanic black | 77 (12.6) | 50 (8.3) | 62 (11.7) | 0.25 | 35 (4.5) | 44 (5.9) | 37 (5.6) | 0.052 |
| Mexican American | 6 (7.1) | 7 (5.6) | 5 (4.4) | 3 (1.3) | 6 (4.7) | 5 (2.4) | ||
| Weight (kg) | 65.1 ± 0.9 | 68.1 ± 0.5 | 68.9 ± 0.7 | 0.007 | 54.0 ± 0.5 | 57.2 ± 0.4 | 59.5 ± 0.5 | <0.001 |
| Body mass index (kg/m 2 ) | 21.5 ± 0.2 | 22.8 ± 0.1 | 23.4 ± 0.1 | <0.001 | 21.0 ± 0.1 | 22.5 ± 0.1 | 23.5 ± 0.1 | <0.001 |
| Waist circumference (cm) | 84.4 ± 0.7 | 89.0 ± 0.5 | 92.5 ± 0.7 | <0.001 | 78.1 ± 0.4 | 83.0 ± 0.5 | 87.0 ± 0.5 | <0.001 |
| Waist-hip ratio | 0.93 ± 0.01 | 0.96 ± 0.01 | 0.99 ± 0.01 | <0.001 | 0.86 ± 0.01 | 0.89 ± 0.01 | 0.91 ± 0.01 | <0.001 |
| Hip circumference (cm) | 90.7 ± 0.5 | 93.0 ± 0.3 | 93.8 ± 0.3 | <0.001 | 91.2 ± 0.3 | 93.8 ± 0.4 | 96.0 ± 0.6 | <0.001 |
| Body fat (%) | 16.0 ± 0.3 | 21.8 ± 0.1 | 26.7 ± 0.1 | <0.001 | 25.8 ± 0.3 | 31.7 ± 0.1 | 36.2 ± 0.2 | <0.001 |
| Body fat (kg) | 10.5 ± 0.2 | 14.9 ± 0.1 | 18.0 ± 0.2 | <0.001 | 13.9 ± 0.2 | 18.2 ± 0.1 | 21.5 ± 0.3 | <0.001 |
| Lean mass (kg) | 54.6 ± 0.8 | 53.2 ± 0.4 | 51.0 ± 0.54 | 0.002 | 39.8 ± 0.4 | 39.1 ± 0.3 | 37.9 ± 0.3 | <0.001 |
| Systolic blood pressure (mm Hg) | 130.9 ± 2.1 | 133.3 ± 1.5 | 136.8 ± 1.3 | 0.03 | 130.3 ± 2.1 | 136.3 ± 1.7 | 137.8 ± 1.6 | 0.004 |
| Diastolic blood pressure (mm Hg) | 72.2 ± 1.1 | 74.0 ± 0.8 | 75.5 ± 0.7 | 0.06 | 70.0 ± 0.7 | 71.9 ± 0.6 | 73.2 ± 0.6 | 0.006 |
| Total cholesterol (mg/dl) | 192.7 ± 3.4 | 208.8 ± 2.7 | 218.3 ± 2.6 | <0.001 | 215.2 ± 3.4 | 225.7 ± 2.7 | 233.3 ± 2.7 | 0.003 |
| High-density lipoprotein cholesterol (mg/dl) | 51.5 ± 1.4 | 49.7 ± 1.3 | 49.0 ± 0.9 | 0.31 | 61.6 ± 1.6 | 59.6 ± 1.0 | 60.5 ± 1.5 | 0.64 |
| Low-density lipoprotein cholesterol (mg/dl) | 119.1 ± 4.1 | 129.5 ± 4.2 | 131.3 ± 2.3 | 0.26 | 142.4 ± 5.1 | 131.1 ± 3.1 | 139.2 ± 3.2 | 0.22 |
| Triglycerides (mg/dl) | 114.5 ± 6.2 | 134.3 ± 10.3 | 152.3 ± 5.8 | 0.003 | 120.3 ± 5.7 | 142.7 ± 6.7 | 147.1 ± 8.3 | 0.02 |
| Glucose (mg/dl) | 100.5 ± 1.4 | 100.1 ± 1.4 | 105.3 ± 3.3 | 0.40 | 94.0 ± 1.0 | 104.6 ± 4.2 | 97.8 ± 1.3 | 0.004 |
| ApoB/ApoA1 ratio | 0.69 ± 0.01 | 0.72 ± 0.01 | 0.86 ± 0.002 | <0.001 | 0.67 ± 0.01 | 0.74 ± 0.001 | 0.74 ± 0.01 | <0.001 |
| Insulin resistance | 0.91 ± 0.13 | 0.89 ± 0.03 | 0.98 ± 0.05 | 0.61 | 0.76 ± 0.003 | 0.99 ± 0.06 | 0.96 ± 0.03 | <0.001 |
| Insulin sensitivity (%) | 165.2 ± 8.4 | 128.8 ± 3.9 | 124.2 ± 5.0 | 0.052 | 156.9 ± 5.3 | 128.9 ± 5.3 | 120.1 ± 2.8 | <0.001 |
| β-cell function (%) | 64.5 ± 2.5 | 70.6 ± 1.0 | 69.9 ± 1.8 | 0.28 | 71.1 ± 1.5 | 70.0 ± 1.7 | 78.8 ± 2.2 | 0.02 |
| Leptin (fg/L) | 3.0 ± 0.1 | 4.2 ± 0.3 | 5.4 ± 0.3 | <0.001 | 7.3 ± 0.3 | 9.5 ± 0.3 | 12.5 ± 0.5 | <0.001 |
| C-reactive protein (mg/dl) | 0.4 ± 0.1 | 0.4 ± 0.04 | 0.53 ± 0.06 | 0.14 | 0.43 ± 0.08 | 0.34 ± 0.02 | 0.43 ± 0.04 | 0.01 |
| Variables | Men | Women | ||||
|---|---|---|---|---|---|---|
| Body Fat <25%, n = 563 | Body Fat >25%, n = 200 | Age-Adjusted p Value | Body Fat <25%, n = 597 | Body Fat >25%, n = 168 | Age-Adjusted p Value | |
| Age (yrs) | 70.2 ± 0.6 | 69.5 ± 0.5 | — | 71.5 ± 0.6 | 69.9 ± 0.7 | |
| Non-Hispanic white | 409 (82.9) | 147 (85.3) | 497 (92.0) | 138 (91.2) | ||
| Non-Hispanic black | 138 (10.4) | 51 (12.1) | 0.48 | 88 (5.0) | 28 (6.6) | |
| Mexican American | 16 (6.7) | 2 (2.6) | 12 (3.0) | 2 (2.1) | 0.79 | |
| Weight (kg) | 66.6 ± 0.6 | 69.7 ± 0.9 | 0.01 | 55.8 ± 0.4 | 60.1 ± 0.4 | <0.001 |
| Body mass index (kg/m 2 ) | 22.2 ± 0.1 | 23.5 ± 0.1 | <0.001 | 21.9 ± 0.1 | 23.7 ± 0.1 | <0.001 |
| Waist circumference (cm) | 87.0 ± 0.6 | 93.1 ± 0.7 | <0.001 | 81.2 ± 0.3 | 87.9 ± 0.6 | <0.001 |
| Waist-hip ratio | 0.95 ± 0.004 | 0.99 ± 0.01 | <0.001 | 0.88 ± 0.004 | 0.91 ± 0.01 | <0.001 |
| Hip circumference (cm) | 91.8 ± 0.4 | 94.5 ± 0.4 | <0.001 | 92.7 ± 0.3 | 96.8 ± 0.6 | <0.001 |
| Body fat (%) | 19.4 ± 0.3 | 27.3 ± 0.2 | <0.001 | 29.5 ± 0.3 | 37.2 ± 0.2 | <0.001 |
| Body fat (kg) | 13.0 ± 0.2 | 19.0 ± 0.3 | <0.001 | 16.5 ± 0.2 | 22.4 ± 0.2 | <0.001 |
| Lean mass (kg) | 53.6 ± 0.5 | 50.6 ± 0.7 | 0.001 | 39.3 ± 0.3 | 37.8 ± 0.3 | <0.001 |
| Systolic blood pressure (mm Hg) | 132.2 ± 1.6 | 137.7 ± 1.6 | 0.012 | 134.2 ± 1.3 | 136.6 ± 1.6 | 0.062 |
| Diastolic blood pressure (mm Hg) | 73.4 ± 0.6 | 75.4 ± 0.9 | 0.12 | 71.1 ± 0.5 | 73.8 ± 0.9 | 0.02 |
| Total cholesterol (mg/dl) | 201.4 ± 2.5 | 220.9 ± 3.3 | <0.001 | 221.3 ± 2.1 | 237.1 ± 3.7 | <0.001 |
| High-density lipoprotein cholesterol (mg/dl) | 50.6 ± 1.1 | 48.5 ± 1.1 | 0.24 | 60.6 ± 0.9 | 60.3 ± 1.8 | 0.79 |
| Low-density lipoprotein cholesterol (mg/dl) | 124.4 ± 2.6 | 133.0 ± 2.2 | 0.07 | 136.3 ± 3.3 | 140.9 ± 3.5 | 0.35 |
| Triglycerides (mg/dl) | 124.0 ± 6.0 | 161.9 ± 8.0 | 0.002 | 132 ± 4.3 | 154.0 ± 12.2 | 0.09 |
| Glucose (mg/dl) | 100.5 ± 1.0 | 106.2 ± 4.2 | 0.23 | 99.2 ± 2.1 | 97.8 ± 1.4 | 0.72 |
| ApoB/ApoA1 ratio | 0.72 ± 0.01 | 0.87 ± 0.001 | <0.001 | 0.70 ± 0.004 | 0.77 ± 0.01 | <0.001 |
| Insulin resistance | 0.88 ± 0.06 | 1.06 ± 0.6 | 0.05 | 0.89 ± 0.04 | 0.96 ± 0.03 | 0.054 |
| Insulin sensitivity (%) | 147.8 ± 7.2 | 110.3 ± 2.4 | <0.001 | 139.6 ± 4.5 | 119.0 ± 2.2 | <0.001 |
| β-cell function (%) | 66.7 ± 1.3 | 73.9 ± 2.0 | 0.02 | 72.0 ± 1.2 | 79.4 ± 2.1 | <0.001 |
| Leptin (fg/L) | 3.7 ± 0.1 | 5.6 ± 0.4 | <0.001 | 8.9 ± 0.3 | 13.4 ± 0.6 | <0.001 |
| C-reactive protein (mg/dl) | 0.4 ± 0.03 | 0.6 ± 0.1 | 0.012 | 0.39 ± 0.04 | 0.43 ± 0.04 | 0.54 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
