Utility of the Z-score of log-transformed A Body Shape Index (LBSIZ) in the assessment for sarcopenic obesity and cardiovascular disease risk in the United States

Body mass index (BMI) has limited accuracy for predicting cardiovascular diseases (CVD) and is not capable of identifying sarcopenic obesity, the combination of sarcopenia (an age-associated decline in muscle mass and physical function) and obesity. To overcome this, the z-score of the log-transformed A Body Shape Index (LBSIZ) was recently introduced as a measure of obesity using waist circumference, height, and weight. We aimed to investigate the association of LBSIZ with sarcopenic obesity and CVD, and propose appropriate cut-off values using the National Health and Nutrition Examination Survey 1999–2016 data. Of 92,062 participants, 40,468 adults (≥20 years) were included. Overall area under curve (AUC) of LBSIZ was 0.735 (95% confidence interval [CI]: 0.716–0.754) for sarcopenic obesity, and 0.695 (95% CI: 0.687–0.703) for CVD. The subgroup analysis of ethnicity/race showed similar results. Waist circumference (WC), BMI, conicity index, body roundness index (BRI), Clinica Universidad de Navarra-Body Adiposity Estimator (CUN-BAE), new BMI, and waist to height ratio (WHtR) showed a negative association with sarcopenic obesity, while LBSIZ and conicity index showed a positive association. The AUC of LBSIZ was significantly higher for sarcopenic obesity than that of conicity index (p < 0.001). The AUC of LBSIZ was significantly higher for CVD than those of parameters including WC, BMI, BRI, CUN-BAE, new BMI, and WHtR (p < 0.001). The AUC for conicity index alone was comparable to that of LBSIZ for CVD. Overall LBSIZ cut-off was 0.35 for both sarcopenic obesity (sensitivity, 65.3%; specificity, 71.5%) and CVD (sensitivity, 63.3%; specificity, 66.6%). These results may be useful not only to identify sarcopenic obesity, but also to conduct CVD risk assessment in the clinical setting.

www.nature.com/scientificreports www.nature.com/scientificreports/ in muscle mass and physical function, and may synergistically worsen the adverse effects of obesity, leading to higher disability, morbidity and mortality 7 .
Accurate assessment of obesity is required for the prevention and treatment thereof. Obesity can be assessed by directly measuring body fat via computed tomography (CT), magnetic resonance imaging, DEXA, and positron emission tomography (PET)-CT 8 . However, these methods are costly and have limitations that make the use of these modalities for diagnosing obesity in real clinical settings challenging; instead, indirect indices of obesity are used. Body mass index (BMI) has long been used, as it is easy to measure and calculate. However, BMI has limited accuracy for predicting the amount and distribution of body fat and is not capable of identifying sarcopenic obesity 9 . It is also limited in its ability to clinically predict the risk of chronic diseases such as CVD [10][11][12] . To overcome these limitations, A Body Shape Index (ABSI), which is a formula that uses waist circumference (WC), height, and weight, has recently been introduced 13 . However, ABSI has limited clinical usefulness due to not having cut-off points for identifying individuals at high risk for obesity-related diseases 14,15 . Therefore, we have previously proposed the z-score of the log-transformed ABSI (LBSIZ), which overcomes the limitations of the ABSI, using representative samples from Korea [15][16][17] . However, data is lacking in terms of its assessed usefulness in other populations.
Therefore, this study aimed to propose a LBSIZ formula for each race, using representative samples from the United States (US). We then examined its relationship with both sarcopenic obesity and CVD risk, compared to other obesity parameters, and provided appropriate cut-off values to identify individuals at high risk for sarcopenic obesity and CVD.

Discussion
This study investigated the association of LBSIZ with both sarcopenic obesity and CVD, using a representative US sample, and found that LBISZ showed superior association thereto, compared with other weight-, and WC-related obesity measures. The study also provided appropriate cut-off values of LBSIZ, to be able to identify individuals at high risk for sarcopenic obesity or CVD, irrespective of sex and race, and to improve its clinical usefulness in practice.
A number of epidemiologic studies showed heterogeneous results regarding the association between traditional BMI and CVD [10][11][12] . This heterogeneity might be due to the limitations in defining obesity based on BMI, which does not differentiate fat from lean mass, nor does it consider the distribution of adipose tissue 1,18-25 . As central deposition of adipose tissues due to obesity became known to be a major cause of CVD-related mortality and morbidity, WC has emerged as an important complement to BMI, as an indicator of visceral adiposity, metabolic risk, and increased morbidity and mortality [26][27][28][29][30] . The National Cholesterol Education Program-Adult Treatment Panel III (NCEP-ATP III) criteria use WC instead of BMI to define metabolic syndrome 31 . However, there are insufficient data regarding the appropriate WC values to define obesity among the different age groups and sexes 19,28,32 . Due to these limitations of BMI and WC, many researchers have explored other obesity indices 33,34 .
In 2012, Krakauer et al. proposed a new obesity index, ABSI 12 , using the WC, weight, and height data from the NHANES 1999-2004. Several studies reported that ABSI predicted premature mortality and CVD more  www.nature.com/scientificreports www.nature.com/scientificreports/ effectively than did BMI or WC 13,35,36 . However, other studies reported that the role of ABSI has been challenged as a risk predictor of mortality, cardiovascular diseases, and metabolic syndrome 32,37,38 . A recent meta-analysis of 38 studies reported ambiguous results, where ABSI outperformed BMI and WC in predicting all-cause mortality, but underperformed in predicting hypertension, and diabetes mellitus 39 . In addition, several studies reported that ABSI had limitations in predicting fat mass 40 .
Although LBSIZ is a revised measure of abdominal obesity based on ABSI, it was better in predicting the development of CVD, than BMI or WC, and even improved the predictability of the Framingham risk score for CVD events in a prospective Korean cohort 15 . In another population-based study using the Korea NHANES data, LBSIZ showed a linear relationship with CVD 17,41 . Interestingly, this study showed that LBISZ has a superior association with CVD in a representative US sample, compared with other obesity parameters. Although the mechanism is unclear, we suspect that the association of obesity parameters with body composition might have played an important intermediary role thereto. In this regard, therefore, the result of LBSIZ having a positive association with FMI, and a negative association with ASMI, is consistent with the results of the previous Rotterdam study 14 . Conversely, other obesity parameters, associated positively with both fat mass and ASMI, could not identify sarcopenic obesity, or the presence of low muscle mass accompanied by a high fat mass 42 .
A complex (albeit not fully elucidated) interplay of several underlying mechanisms are responsible for the development of sarcopenic obesity; fat accumulation, which is related to the increase of proinflammatory cytokines, oxidative stress, and insulin resistance, might cause muscle fiber atrophy and mitochondrial dysfunction, leading to the development and progression of sarcopenia. Likewise, sarcopenia could worsen obesity through the decline of physical activity and energy expenditure, thereby resulting in further sarcopenia, leading to a vicious cycle of atrophy, aggravating their effects on metabolic, and functional abnormalities 43,44 . While a clear definition of sarcopenic obesity was not available, several studies showed that it was related to metabolic diseases and physical disability [45][46][47] ; a few studies about the relationship between sarcopenic obesity and CVD, and mortality have been performed [48][49][50] . According to a recent meta-analysis, sarcopenic obesity was assessed to be associated with a 24% increased risk of all-cause mortality 51 . This study used Baumgartner's definition of sarcopenic obesity 42 , and showed that LBSIZ is the only measure of obesity related to sarcopenic obesity among all the obesity parameters that were considered. These results indicate the usefulness of LBSIZ in screening sarcopenic obesity, unlike other obesity measures.
LBSIZ has limitations in its use in clinical settings, or in epidemiological studies, considering the calculation of LBSIZ is highly complicated. Therefore, we present a simpler formula that can be used to estimate LBSIZ for each race, using a large-scale dataset from the NHANES 1999-2016. In addition, the estimated cut-off value can be used as a clinical standard, thereby facilitating easier clinical use, irrespective of sex or race. Considering the increase in ORs for sarcopenic obesity and CVD after the median LBSIZ in the restricted cubic spline regression plots, the cut-off value corresponding to the 65 th percentile of LBSIZ can effectively help to identify individuals at a high risk of sarcopenic obesity and CVD.
Despite our interesting findings, there are a few limitations in this study. First, this was a cross-sectional study, and therefore, the causal relationship between obesity and CVD remains to be examined further, using cohort data. Secondly, this study did not analyze mortality, and may have missed fatal CVD events due to a lack of data. Third, we only examined weight-, and WC-related obesity measures, omitting hip circumference-related measures due to data availability. Finally, potential confounders are yet to be further examined to elucidate the LBSIZ-related pathophysiological mechanism.
In conclusion, LBSIZ showed a stronger association with both sarcopenic obesity and CVD, compared with other obesity parameters. These results may be useful when conducting a CVD risk assessment in clinical settings; the proposed LBSIZ cut-off values have the potential to be a useful clinical standard.

Materials and Methods
study population. Data were collected from the NHANES dataset between 1999 and 2016. Exclusion criteria were as follows: those aged ≤20 years, those with missing data (CVD questionnaire, anthropometric, or laboratory data), or those who were not Hispanic, non-Hispanic whites, or non-Hispanic blacks. Finally, 40,468 of 92,062 participants were included in this study (Fig. 1). Because the DEXA data were available between 1999 and 2005, 11,780 participants were included in the subgroup analysis for body composition.
Measurements of obesity parameters and body composition. WC was measured using a measuring tape at the upper-lateral border of the iliac crest 52 . BMI was defined by the weight in kilograms, divided by the height in meters squared (kg/m 2 ). Conicity Index, BRI, CUN-BAE, new BMI, and WHtR were calculated based on the earlier-suggested formula 13,32,53,54. . We calculated the LBSIZ based on the regression [ln(waist) = a 0 + a 1 ln(weight) + a 2 ln(height) + δ], standardizing the waist values according to weight and height. After respective estimation for each race, the log-transformed ABSI (LBSI) was calculated using the following equation: log [waist/(exp(a 0 ) × weight a1 × height a2 ]. In the next step, the z-score of the log-transformed ABSI (LBSIZ) was calculated using the mean of the LBSI and standard deviation (SD) of LBSI: LBSIZ = (LBSI-mean(LBSI))/SD(LBSI). More details are provided in the supplementary Excel file, and are referenced in a previous study 16 .
Assessment of body composition was performed by whole body DEXA using a Hologic QDR 4500 A fan beam X-ray bone densitometer (Hologic Inc., Marlborough, MA, USA). All DEXA scans were analyzed using Hologic Discovery software (version 12.1, Hologic Inc.) to measure total and regional body composition, including bone mineral content lean body mass, fat mass, and % body fat. ASM was defined as the sum of the total lean mass of both the arms and legs. ASM index was defined as ASM divided by the square of the height. Fat mass index was defined as the total fat mass divided by the square of the height.
Definition of sarcopenic obesity and CVD. Sarcopenic obesity was defined as ASMI < 7.26 kg/m 2 , and % body fat >27% in men, or ASMI < 5.45 kg/m 2 , and % body fat >38% in women, based on the definitions by Baumgartner 42 A structured questionnaire was used to investigate CVD. A patient was deemed to have CVD if they had at least one of the following conditions: angina pectoris, coronary heart disease, myocardial infarction, congestive heart failure, or cerebrovascular disease 55 . statistical analysis. For summary statistics, we presented the mean with a 95% CI, or prevalence (%) according to ethnicity/race. Continuous variables were assessed using one-way analysis of variance, and categorical variables were assessed using the Pearson's chi-square test. We estimated the distribution of LBSIZ, and verified the correlations between the obesity indices. A receiver operating characteristic (ROC) curve was used to analyze the correlation between sarcopenic obesity or CVD, and each of the obesity parameters, and the de Long's test was used to identify obesity indices that were significantly superior 56 . We validated the model using 1,000 samples generated through the bootstrap method. The cut-off value of LBSIZ was determined as the value with the highest Youden's index, conditional on having a sensitivity and specificity greater than 60% 57 . A multivariate logistic regression analysis was performed to determine the OR for sarcopenic obesity and CVD. Furthermore, the OR of LBSIZ for sarcopenic obesity and CVD was analyzed using restricted cubic spline splits with five knots. The analyses were performed using SPSS software version 24.0 (IBM Inc., Armonk, NY, USA), R (version 3.1.0, R Foundation for Statistical Computing, Vienna, Austria) and Stata version 15 (Stata Corporation, College Station, Texas, USA). A P-value of 0.05 was considered statistically significant. ethics statement. The study protocol was approved by the institutional review board of Kangnam Sacred Heart Hospital (IRB No. HKS 2017-07-007). All participants volunteered, and provided written informed consent prior to their enrolment. All participants' records were anonymized before being accessed by the authors. All methods were carried out in accordance with the approved guidelines and regulations.