Data source and study population
Beginning in 1999, continuous NHANES is conducted by the National Center for Health Statistics in the Centers for Disease Control and Prevention. It includes a series of two-year cross-sectional nationally representative survey of US civilian noninstitutionalized population.17,18 Each survey consists of interview and examination. For each individual, an interview is completed in his/her home, and a health examination is conduted in a mobile examination center. The interview collects demographic, socioeconomic, dietary, and other questionnaire-related data. The examination collects medical, dental, and physiological, and laboratory data. Using a multistage stratified complex probability sampling design, NHANES oversamples older adults, low-income individuals, and certain racial/ethnic groups; participants were assigned weights to account for their unequal sampling probability and nonresponse. All the participants gave informed consent, and the survey was approved by the National Center for Health Statistics Institutional/Ethics Review Board.
In the present study, the sample of participants were chosen from the three most recent cycles of NHANES conducted in 2011-2016. The choice of cycles was determined by data availability of SAD, the only outcome measure in the present study: 2011-2012 represents the first cycle of collecting SAD measurements, and 2015-2016 is the most recent cycle for which data are available. Participants aged less than 18 years who were interviewed only but not examined, who were pregnant during the examination, and/or who had SAD measurements missed, were excluded from the study based on the following considerations. SAD, similar to BMI and WC, may have different definitions or meanings between children and adults, and the purpose of this study was to examine sex differences in SAD among adults. Participants with ‘interview only’ missed the examination of SAD which was our focus measure in the study. Pregnant women experienced dramatic changes in physiology and body shape, and their normal levels of SAD could not be reflected. After applying the above exclusion criteria, we had 14,054 individuals in the final study sample, including 6,975 men and 7,079 women.
Sagittal abdominal diameter
SAD was measured by a trained examiner when the participant was in the supine position on an examination table (Figure 1).19 An abdominal caliper of proper size with lower and upper arms (Holtain Model 609XL, Seritex Inc, NJ, USA) were used to establish the external distance between the front of the abdomen and the small of the back at the iliac level line. The participant was first asked to lie down on the table, bend his or her knees at a 90 degree angle with feet resting flat on the table and arms crossed over the chest. The right and left iliac crests were located, and a line perpendicular to the table on the uppermost lateral border of the right ilium was drawn. A measuring tape was extended over the abdomen without compressing the skin from the left iliac crest to the mark on the right iliac crest. A horizontal line was drawn around 5 cm long, on the abdomen along the iliac level line on the top left edge of the tape.
The caliper’s lower arm was inserted under the small of the back making sure the upper arm exceeds the participant’s abdominal diameter. The shaft of the caliper was adjusted in a vertical position. The caliper’s upper arm was slid down to lightly touch the abdomen with the edge aligned with the iliac level line mark after the participant took in a gentle breath, slowly let the air out and then paused. The measurement was taken when the participant was in the resting phase and the caliper’s shaft was in the vertical position. Up to four SAD readings were taken following the above procedure. The SAD for each person was calculated as an average over the readings.
Socio-economic characteristics
Socio-economic characteristics considered in this study were all self-reported, including age, sex, race, education, birth place, and household income. They were chosen from a large number of socio-economic variables using model selection techniques and were only significant variables in relation to SAD. Age was categorized as under 30, 30-39, 40-49, 50-59, and 60 years or more. Race/ethnicity included non-Hispanic white, non-Hispanic black, non-Hispanic Asian, Hispanic, and other. The level of education was categorized as high school below, high school graduate/GED or equivalent, and college or above in terms of years in school. Birth place had two categories: born in the US and born in other countries. The family income was grouped into 4 categories: less than $20,000, $20,000-$44,999, $45,000-$74,999, and $75,000 or more. Weight and height were measured using standardized techniques and equipment. BMI was calculated as weight in kilograms divided by the square of height in meters, and then categorized as normal weight (BMI less than 25.0 kg/m2), overweight (BMI greater than 25.0 and less than 29.9 kg/m2), and obesity (BMI greater than or equal to 30.0 kg/m2).
Statistical analysis
The NHANES supporting and analytical guidelines for surveys 2011-2016 were followed.20 Stratum, cluster and weight design techniques for survey data were incorporated into data analysis to ensure the representativeness and generalization of the estimates. All the data analyses were performed on PC with windows 10 using survey procedures in SAS version 9.4 (SAS Institute Inc, Cary, NC).
Percentages and standard errors were calculated for categorical variables for men and women to examine sex difference in socio-economic characteristics of subjects. The significance of difference in percentages was tested using weighted χ2 tests. To assess unadjusted sex difference in SAD, means and standard errors of SAD were calculated for men and women in each group define by the categories of age, race, education, birth place, household income, and BMI. Weighted independent t tests were used to compare the means of SAD between men and women.
Weighted multiple regression was performed with SAD as a dependent variable and socio-economic characteristics as independent variables including age, sex, race/ethnicity, education, birth place, household income, and BMI. Regression coefficients and standard errors with 95% confidence intervals (CIs) were estimated and adjusted associations of sex with SAD were examined. Wald χ2 tests were useed to examine the significance of parameter estimates. The model was then conducted separately in women and men to identify socio-economic correlates of SAD and the sex differences in associations of the correlates with SAD were test by using weighted independent t tests.