Article Text
PDF
Abstract
Introduction A body shape index (ABSI) is independently associated with mortality in general population, but studies on the predictability of ABSI in the risk of mortality in patients with type 2 diabetes (T2D) are limited. We aimed to examine the independent and joint association of ABSI, body mass index (BMI), waist circumference (WC), waist-to-hip ratio (WHR), waist-to-height ratio (WHtR), and body roundness index (BRI) with mortality in patients with T2D.
Research design and methods The study included 11 872 patients (46.5% women) aged 30 years and older and who took part in diabetes care management program of a medical center in Taiwan. Body indices were evaluated by anthropometric measurements at baseline between 2001 and 2016, and their death status was followed up through 2021. Multivariate Cox regression models were used to assess the effect of body indices on mortality.
Results During a mean follow-up of 10.2 years, 560 cardiovascular disease (CVD) deaths and 3043 deaths were recorded. For ABSI, WC, WHR, WHtR and BRI, all-cause mortality rates were statistically significantly greater in Q4 versus Q2. For BMI and WHtR, all-cause mortality rates were also statistically significantly greater in Q1 versus Q2. The combination of BMI and ABSI exhibited a superiority in identifying risks of all-cause mortality and CVD mortality (HRs: 1.45 and 1.37, both p<0.01).
Conclusions Combined use of ABSI and BMI can contribute to the significant explanation of the variation in death risk in comparison with the independent use of BMI or other indices.
- Diabetes Mellitus, Type 2
- Body Mass Index
- Obesity
- Waist Circumference
Data availability statement
All data relevant to the study are included in the article or uploaded as supplementary information.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
WHAT IS ALREADY KNOWN ON THIS TOPIC
Numerous studies explored the effects of various body indices simultaneously on mortality, but most of them examined the independent effect of body indices.
A Body Shape Index (ABSI) is independently associated with mortality in general population, but studies focused on ABSI in the risk of mortality in patients with type 2 diabetes (T2D) are limited.
WHAT THIS STUDY ADDS
Using a large cohort of persons with T2D, our study observed that the magnitude of joint effects of body mass index (BMI) and ABSI on mortality exceeded that of their individual effects and was the largest combination effect compared with other combinations.
The combination of BMI and ABSI had a 1.45-fold increase in overall mortality and a 1.37-fold increase in cardiovascular disease death risk.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
ABSI combined with BMI as a baseline assessment may serve as a superior obesity index of mortality risk stratification in clinical practice, especially among patients with T2D.
We suggested that the additional use of ABSI as an indicator of health management among patients with T2D is recommended in clinical practice.
Introduction
Diabetes mellitus (DM) has become a global health problem. The prevalence of DM has increased substantially and reached 9.3% in 2019, which corresponds to 463 million patients globally; this number is expected to rise to 700 million (10.9%) by 2045.1 Between 1990 and 2019, a total of 134 million deaths due to high fasting plasma glucose (FPG) were recorded globally. The number of deaths increased from 2.91 million in 1990 to 6.50 million in 2019.2 The global disease burden attributable to diabetes was 67.9 million in 2017 with a projection of 79.3 million disability-adjusted life years in 2025.3 In Taiwan, the total population with DM and the incidence of DM increased by 66% and 19% from 2005 to 2014, respectively.4 The prevalence of diabetes was 10.1% in 2014 and is projected to increase to 13.1% in 2035.4 5 The disease burden attributable to DM was 145.7 million disability-adjusted life years in 2017.6 Given that DM may cause premature death or disability, it is important to prevent and delay the poor progression of DM.
Previous studies have shown a strong and direct association between DM and obesity.7 8 The increased prevalence of DM and obesity is rising rapidly.2 9 10 Obesity and DM are termed as a ‘21st Century epidemic’ by WHO. The vertiginous rise in obesity triggers a parallel upward swing in DM statistics.11 12 The transition from obesity to diabetes may be due to a progressive defect in insulin secretion combined with a gradual increase in insulin resistance.13 DM and obesity are well-known risk factors of incident cardiovascular disease (CVD).14 15 Concurrent diabetes and obesity have great adverse effects on health.16 According to the result of a large type 2 DM cohort study with 10-year follow-up, obesity was a negative prognostic factor for survival of patients <65 years.17 DM prevention and management have been the focus of attention. Therefore, identifying obese patients with type 2 DM and who are at high risk for adverse outcome is an important first step toward adopting preventive measures.
WHO has defined obesity based on anthropometric index as a body mass index (BMI) ≥30 kg/m2.18 Based on nationally representative samples of adults aged 19 years and above from three consecutive waves of Nutrition and Health Survey in Taiwan (1993–1996, 2005–2008, and 2013–2014), the global age-standardized prevalence of obesity increased from 3.2% in 1993–1996 to 10.8% in 2013–2014 in men, and from 6.4% to 14.9% in women during the corresponding period.19 Given its simple and practical use, BMI has been recommended for prediction of adverse outcomes in clinical practice.20 However, BMI has been criticized for not reflecting individual fat distribution.21 A large-scale study compared the association of cardiovascular risk with BMI, waist circumference (WC), and waist-to-hip ratio (WHR) in more than 10 000 patients with diabetes and concluded that BMI is not the best predictor of cardiovascular events and mortality in diabetics.22 According to the Expert Panel on Population and Prevention Science, the measurement of WC or WHR can indicate visceral adipose tissue or intra-abdominal fat, which may be more deleterious than overall overweight or obesity.20 As suggested by the limitations of BMI, previous studies have confirmed that other body indices, such as WC, WHR, and waist-to-height ratio (WHtR), are superior to BMI as indicators of risks of mortality and CVDs.23–25
A body shape index (ABSI) and body roundness index (BRI) are two novel body indices.26 27 The former is calculated based on WC adjusted for height and weight.27 The latter is quantified by the individual body shape in a height-independent manner.26 ABSI is positively associated with visceral adiposity, which is independent of BMI of patients with type 2 diabetes (T2D).28 BRI predicts body fat and the percentage of visceral adipose tissue better than BMI.26 Although previous studies showed that the combination of BMI and ABSI achieves better mortality risk stratification than the combination of BMI and alternative indices of abdominal obesity in general populations,27 29 the prediction of ABSI and BRI in relation to mortality of patients with T2D has not been fully elucidated. Only the Fremantle Diabetes Study in Australia explored the association between ABSI and mortality in T2D.30 The study reported that ABSI was a better index of central obesity than WC, BMI, or WHR for the prediction of mortality risk based on a small cohort of 1296 patients with T2D. Therefore, we examined the independent and joint association between six obesity indices (ABSI, BMI, WC, WHR, WHtR, and BRI) and mortality in patients with T2D using a large cohort of persons with T2D.
Methods
Data source
A retrospective cohort study was conducted including all enrollees in the diabetes care management program (DCMP) of a medical center in Taiwan. The DCMP is a nurse case management program for patients with diabetes. We acquired baseline characteristics, including sociodemographic factors, lifestyle behaviors, diabetes-related variables, body indices, biomarker, comorbidities, hypertension medications, cardiovascular medications, and hyperlipidemia medications from the DCMP database. All-cause and CVD mortality were defined from the data set of Health and Welfare Statistics Application Center, Ministry of Health and Welfare. The vital status was determined through linkage with Death Dataset of National Health Department.
Study subjects
All patients with T2D in the DCMP from 2001 were included. Patients diagnosed with DM in accordance with the criteria established by the American Diabetes Association (International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) code 250) during the baseline and subsequent years were included as study subjects. Patients were excluded from the sample if they were diagnosed with type 1 diabetes (ICD-9-CM code 250.x1/x3) (n=448), aged under 30 years (n=504), followed up for <1 year (n=246) and lacked body indices values (n=5050). We also excluded participants with missing data of sociodemographic factors, lifestyle behaviors, diabetes-related variables, biomarkers, comorbidities, and medication use (n=253). Of the 18 373 patients considered for inclusion, 12 125 were eligible, and 11 872 were included in the final analysis (online supplemental figure S1). The index date was defined as the first date that the patient’s body index can be defined in the database. Each individual was followed up from the date of entry until December 2021 or until an event of death.
Supplemental material
Measurements
Body indices
The anthropometric measurements included height (cm), body weight (kg), WC (cm) and hip circumference (HC; cm). Height and body weight were measured on an autoanthropometer (Super-View, HW-666) with the participants being shoeless and wearing light clothing. WC was measured mid-way point between the inferior margin of the last rib and the crest of ilium in the horizontal plane when the participant was standing. HC was measured as the distance around the pelvis at the point of maximal protrusion of the buttocks. The WC and HC were measured by trained assistants using the same equipment. ABSI was calculated using the following equation: 
.27 BRI is calculated using the following equation: 
.26 BMI was calculated as weight (kg) divided by height squared (m2). WHR and WHtR were, respectively, calculated as WC divided by HC and WC divided by height.
Sociodemographic factors, lifestyle factors, diabetes-related variables, and biomarkers
These variables included age, sex, smoking habits, alcohol drinking, physical activity, duration of diabetes, and types of diabetes treatment. Smoking habits, alcohol drinking, and physical activity are categorized into two levels: yes and no. The number of days between dates for DCMP enrollment and onset diabetes was calculated as the duration of diabetes. Then the number of days was divided by 365 days. Measured HbA1c, FPG, low-density lipoprotein cholesterol, high-density lipoprotein cholesterol (HDL-C), total cholesterol and triglycerides at the index date of DCMP were also retrieved.
Medications
The types of diabetes treatment were divided into four groups: diet or exercise, oral hypoglycemic agents, insulin injection, and oral hypoglycemic agents plus insulin injection. Oral hypoglycemic agents, such as biguanide, sulfonylurea, thiazolidinedione, and meglitinide, and insulin therapy were extracted from electronic medical records. Drug-related variables included hypertension, hyperlipidemia, and cardiovascular medications. Drug-related variables were divided into two responses (yes or no) based on electronic medical records from the DCMP database.
Comorbidities
Comorbidities consisted of hypertension, hyperlipidemia, stroke, coronary artery disease, severe hypoglycemia, peripheral neuropathy, neuropathy, nephropathy, diabetic ketoacidosis, and hyperglycemic hyperosmolar non-ketotic coma (HHNK) and were retrieved from the DCMP from medical files and interview at baseline. All comorbidities were each divided into two groups: yes versus no.
Outcome ascertainment
The major outcomes were total and CVD deaths, which were identified through record linkages with the National Death Datasets, provided by the Taiwan Ministry of Health and Welfare. The time of follow-up began from recruitment (index date) to death or the end of follow-up (December 2021). CVD deaths included deaths due to CVD (ICD-9-CM codes 390–398, 401–405, 410–414, 420–438, and 440, and ICD-10-CM codes I01–I02, I05–I15, I20–I25, I27, I30–I52, and I60-I71).
Statistical analysis
Proportions were presented for categorical variables, and means and SDs were determined for continuous variables. The differences between subgroups of vital status were determined by χ2 test and Student’s t-test. The relationship among six body indices was assessed by Pearson correlation coefficients stratified by sex. Participants were divided into four groups based on gender-specific body index quartiles. The HRs of predictor variables were estimated using Cox proportional hazards models to estimate the hazard of mortality. Cox proportional hazards models for the time of death were fitted to obtain age and sex-adjusted HRs initially. Then, the models were adjusted for potential confounding variables of lifestyle behaviors and biomarkers and finally for comorbidities and diabetic complication. Multivariable-restricted cubic spline models were used to evaluate potential non-linear associations between the continuous change in body indices and mortality. For estimation of the capability in predicting mortality among these body indices, the receiver operating characteristic (ROC) curves were used. The comparisons between different areas under the curves of ROC were made by a non-parametric method developed by DeLong et al.31 In addition, the interaction of each body index with demographic and diabetes-related factors in risk of death was examined to determine whether significant modifiers were present, and in such case, the risk of death for quartiles of each body index stratified by modifiers was further calculated. Lastly, the joint effect of overall obesity (BMI) with other central obesity indices on mortality was explored. Two categories for each body index were used by grouping two quartiles with similar risks of death. Therefore, for BMI, WC, WHR, WHtR, and BRI, the lowest and highest quartiles were combined, and the two intermediate quartiles were combined. For ABSI, the lowest two quartiles were combined, and the same was applied to the highest two quartiles. Then, the joint effect of BMI categories with the five central obesity index categories on mortality was estimated with Cox proportional hazards model adjusted for potential confounding variables. All analyses were performed using SAS V.9.4 (SAS Institute).
Results
Over the mean 10.2 years’ follow-up (SD=3.8), 3043 deaths were observed among 11 872 patients with T2D. Table 1 presents the bivariate associations between baseline characteristics and all-cause mortality and CVD mortality. Significant differences were observed in sex, age, duration of diabetes, type of diabetes treatment, ABSI, BMI, WC, WHR, WHtR, BRI, HbA1c, FPG, HDL-C, hypertension, stroke, coronary artery disease, severe hypoglycemia, peripheral neuropathy, nephropathy, HHNK, and medication use between persons who were alive and dead while significant differences were found in age, smoking status, duration of diabetes, type of diabetes treatment, ABSI, BMI, WHR, WHtR, BRI, HbA1c, FPG, HDL-C, hypertension, stroke, coronary artery disease, severe hypoglycemia, peripheral neuropathy, nephropathy, HHNK, and hypertension and cardiovascular medication use between persons who died of CVD and who did not. Online supplemental table S1 shows the Pearson correlation coefficient between obesity indices stratified by sex. The results showed that WHR had the highest correlation coefficient with ABSI (r=0.53 for men, r=0.64 for women, both p<0.001).
The comparisons of baseline characteristics of the diabetic cohort according to mortality status (n=11 872)
Table 2 displays the all-cause and CVD mortality risks across quartile subgroups for ABSI, BMI, WC, WHR, WHtR, and BRI obtained by Cox regression analysis. Compared with the second quartile (Q2), all-cause mortality significantly increased in ABSI Q4 (HR: 1.40, 95% CI 1.26 to 1.55), BMI Q1 (HR: 1.25, 95% CI 1.14 to 1.38), WC Q4 (HR: 1.11, 95% CI 1.01 to 1.23), WHR Q3 (HR: 1.13, 95% CI 1.02 to 1.25), WHR Q4 (HR: 1.26, 95% CI 1.14 to 1.40), WHtR Q1 (HR: 1.12, 95% CI 1.00 to 1.25), WHtR Q4 (HR: 1.13, 95% CI 1.02 to 1.25), and BRI Q4 (HR: 1.13, 95% CI 1.02 to 1.25), whereas CVD mortality increased in ABSI Q4 (HR: 1.36, 95% CI 1.06 to 1.73).
HRs of all-cause and CVD mortality for body indices in patients with type 2 diabetes (n=11 872)
Figure 1 shows the area under the ROC curves for six body indices with inclusion of baseline characteristics for the prediction of all-cause and CVD mortality. In addition, we compared the predictive capability of six body indices in addition to baseline characteristics when assessing all-cause and CVD mortality. Compared with the model with baseline characteristics, models with ABSI, BMI, or WHR showed better capability to predict all-cause mortality (p<0.05). Online supplemental figures S2 and S3 show the multivariable-adjusted restricted cubic spline plots of the HR for all-cause and CVD mortality according to six body indices by sex. The positive near-linear relationships of all-cause and CVD mortality with ABSI were found, whereas BMI, WC, WHR, WHtR, and BRI showed non-linear U-shaped associations.
The areas under the receiver operating characteristic curves for (A) all-cause mortality and (B) CVD mortality. Baseline characteristics included age, sex, smoking, alcohol drinking, exercising, duration of diabetes, and types of diabetes treatment. ABSI, a body shape index; BMI, body mass index; BRI, body roundness index; CVD, cardiovascular disease; WHR, waist-to-hip ratio; WHtR, waist-to-height ratio.
The significant interactions of diabetes duration with BMI and ABSI on CVD mortality were detected (p=0.02 and 0.01, respectively). Figure 2 reveals the adjusted mortality risk based on body index quartiles stratified by median year of diabetes duration (<3 and ≥3 years). The multivariate-adjusted HR for the highest quartiles of ABSI and WHR was significantly associated with all-cause mortality in patients with diabetes duration <3 years; the highest quartile of ABSI and the lowest quartile of BMI were significantly associated with all-cause and CVD mortality in patients with diabetes duration ≥3 years. This finding implied that the body indices exerted stronger effects on mortality in individuals with diabetic duration of 3 years or more than in individuals with a diabetic duration of less than 3 years.
Adjusted HRs of all-cause and CVD mortality by body indices according to duration of diabetes (<3 or ≥3 years). *P<0.05; **p<0.01; ***p<0.001. ABSI, a body shape index; BMI, body mass index; BRI, body roundness index; CVD, cardiovascular disease; WHR, waist-to-hip ratio; WHtR, waist-to-height ratio.
Combined assessment of overall obesity (BMI) and central obesity (ABSI, WC, WHR, WHtR, and BRI) identifies obesity patterns associated with mortality risk. Table 3 presents the joint effects of BMI and other body indices (ABSI, WC, WHR, WHtR, and BRI) on mortality. For the combination of BMI and ABSI, patients in the two intermediate quartiles of BMI and the lowest two quartiles of ABSI were considered as a reference group. Compared with the patients in the reference group, the adjusted HR was 1.45 (95% CI 1.30 to 1.61) for all-cause mortality and 1.37 (95% CI 1.06 to 1.77) for CVD mortality in patients with the combination of BMI in Q1 and Q4 with ABSI in Q3 and Q4. Compared with the combination of BMI and other body indices, the highest adjusted HRs for all-cause and CVD mortality were observed in the combination of BMI and ABSI.
Adjusted HRs of all-cause mortality and CVD mortality based on BMI and other body indices in patients with type 2 diabetes
Discussion
This study conducted a large cohort study that assessed the independent and joint associations of body indices in the prediction of all-cause and CVD mortality among patients with T2D. The present results indicated that BMI, WC, WHR, WHtR, and BRI showed U-shaped associations of all-cause mortality, whereas the risk of all-cause and CVD mortality increased monotonically with the increase in ABSI. When BMI combined with other central obese indices, the combination of BMI and ABSI can predict the risk for all-cause and CVD mortality. Moreover, only ABSI can independently predict all-cause mortality in patients with T2D with diabetes duration of <3 or ≥3 years. To test whether ABSI is independently and significantly associated with mortality in various BMI subgroup populations, we further analyzed the associations between ABSI and mortality stratified by BMI quartile subgroups; we found that ABSI is independently and significantly associated with all-cause mortality in four BMI quartile subgroups, but ABSI is not independently and significantly associated with CVD mortality in four BMI quartile subgroups (online supplemental table S2).
According to a meta-analysis of 38 studies focusing on general population, ABSI was significantly associated with all-cause mortality and outperformed BMI and WC as an accurate predictor,32 but the findings were inconsistent across studies.33 34 No association between ABSI and mortality was found in Chinese male33 and hemodialysis patients.34 Several studies have examined the effect of ABSI on mortality in patients with diabetes. Tate et al used an Australian cohort of 1282 patients with T2D and observed that ABSI was significantly associated with mortality in men after adjustment for confounders (diabetes duration, ethnicity, and smoking) but not in women. The lack of an association for women may be explained by the limited study power due to the small sample size.30 To improve the limitation of prior work of Tate et al, we assessed the association between ABSI and mortality using a large cohort including 11 872 patients with T2D. We observed that ABSI was significantly associated with all-cause mortality in men and women, and the results were consistent after adjustment of biomarkers and comorbidity. Regarding the addition of ABSI in the prediction of mortality, Dhana et al discovered that the addition of ABSI to traditional risk factors in general populations did not improve the area under the ROC curves.35 Conversely, the present study revealed that ABSI provided a significant and the largest added predictive capability among body indices (ABSI, BMI, WC, and WHR). Therefore, ABSI is a substantial marker of obesity independent of traditional anthropometric indices (such as BMI, WC, WHR, and WHtR) in patients with T2D and is a risk factor for death.
Numerous studies explored the effects of various body indices simultaneously on mortality, but most of them examined the independent effect of body indices.35–37 Only two studies examined the combined effect of BMI with other body indices on mortality in general populations,27 29 but no research investigated people with T2D. Thus, scholars should examine the effect on T2D because the effect of body shape on the risk of death differs between people with and without T2D.38 The present study elucidated the independent effect of six obesity indices and joint contribution of BMI and other abnormal body indices to the risk of mortality in T2D.
We observed that the magnitude of joint effects of BMI and ABSI on all-cause mortality exceeded that of their individual effects and was the largest combination effect compared with other combinations. The results are consistent with those of a large population-based study in Europe, which showed that subjects in the combination group of high ABSI with severe obesity (BMI≥35 kg/m2) or underweight (BMI<18.5 kg/m2) were at the highest risk of death compared with their reference group (low ABSI and normal weight).29 Thus, ABSI can be used to shed light on the predictive capability of mortality, which cannot be attributed solely to traditional anthropometric indices. Given that ABSI reflects abdominal adiposity, it has been reported that ABSI predicted MRI-defined abdominal adipose tissue depots such as total intra-abdominal adipose tissue and subcutaneous abdominal adipose tissue in non-obese Asian Indian adults, whereas other central obesity measures of hip index and anthropometric risk index did not.39 In addition, it has been reported that ABSI is independent of BMI, and more accurately than WC to be a substantial marker of arterial stiffness in Japanese adults.28 Among persons who received bariatric surgery in the USA, baseline ABSI could predict change in mortality risk at 3-year after surgery, but baseline BMI did not.40
For the traditional anthropometric index, the present study reported an increase in all-cause mortality in the lowest quartile of BMI (equivalent to underweight) but not the upper two quartiles (equivalent to overweight and obese) compared with the second quartile of BMI (equivalent to normal weight), and this finding is consistent with that of a previous study reporting the ‘obesity paradox’.41 It means the overweight or obesity is paradoxically associated with a lower or not higher risk of mortality.41 However, the obesity paradox for survival among individuals with T2D has been observed in several research42–44 but not in other studies.45 46 Although several studies have shown that this paradox may be explained by smoking as an effect modifier, the U-shape of the BMI–mortality relationship for the obesity paradox was observed in current and former smokers but not in never smokers46 47; we did not observe the moderating effect of smoking behavior. Notably, the risk of death increased with the increase in ABSI in the present study. No obesity paradox with death was observed when the ABSI was applied to measure obesity. The risk of death due to obesity measured by ABSI is intuitive because of the approximately exponential linear relationship between them.
Interestingly, we observed the interaction between diabetes duration and ABSI in the risk of death while no significant interactions between diabetes duration and other body indices were found. In patients with a short diabetes duration, a significant risk of all-cause death was observed only in the highest quartile of ABSI, and in patients with long diabetes duration, the risk of death was shown in the upper two quartiles of ABSI. Thus, the threshold of ABSI for the risk of death differs between patients with short and long durations of diabetes. In patients with long duration of diabetes, the risk of death begins with relatively small ABSI values. Mechanisms in the pathogenesis of T2D include insulin resistance and impaired insulin secretion by β-cell dysfunction48; long diabetes duration significantly increases insulin resistance, implying that insulin action leads to progressive deterioration after the onset of diabetes.49 50 The non-esterified fatty acids released from adipose tissue in obese subjects are linked with insulin resistance and β-cell dysfunction,51 and the visceral adipose tissue may magnify these dysfunctions.52 These pieces of evidence may support our study findings, which suggest that patients with long diabetes duration need to pay attention to abdominal obesity and properly control their body shape because individuals in the third quartile of ABSI are associated with the increased risk of death compared with those whose diabetes duration is greater than 3 years but not with those with a diabetes duration less than 3 years. Furthermore, patients with a short diabetes duration showed a high risk of death in the highest quartile of ABSI, but no increase in risk of death was found in other body indices.
This study has several strengths, which included a relatively large sample size, the use of hospital electronic health records53 and National Death Datasets, standardized measurements of body indices, and a 10-year retrospective follow-up period. In addition, we considered lifestyle behaviors, such as smoking, alcohol consumption, and physical activity, biomarkers and complications that may confound the associations observed in the study. The present study also encountered several limitations. First, body indices were measured at baseline. Thus, the effect of change in body indices over time on the risk of mortality was not examined in the present investigation. Therefore, to reduce the effects of potential reverse causation on our results, we excluded subjects with a follow-up period of less than 1 year. Second, no image assessment, which has been considered the gold standard for visceral obesity, was conducted for body indices.54 Visceral obesity is also a clinically important index for the identification of individuals at high risk of obesity-related health conditions.55 56 Third, our study subjects were from a single hospital center, and whether our results are generalizable to the entire population with T2D in Taiwan remains unknown. Our study recruited subjects with similar age and sex distributions from the diabetic population of Taiwan. Thus, selection bias was unlikely. Our results may be generalizable to several diabetic populations with characteristics similar to those of our study sample and thus can help in informed decision-making and treatment intensification in clinical settings.
Conclusion
Our study showed that ABSI alone is useful in discriminating all-cause and CVD mortality risks in patients with T2D. ABSI combined with BMI as a baseline assessment may serve as a superior obesity index of mortality risk stratification in clinical practice, especially among patients with T2D. The additional use of ABSI as an indicator of health management among patients with T2D is recommended in clinical practice.
Data availability statement
All data relevant to the study are included in the article or uploaded as supplementary information.
Ethics statements
Patient consent for publication
Ethics approval
This study was approved by the Ethical Review Board of China Medical University Hospital (CMUH110-REC2-214).
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
Contributors T-CL, C-IL and C-CL were responsible for drafting the article and the conception and design of the study. T-CL, C-IL and S-YY acquired and analyzed the data. C-SL and C-HL interpreted the data. All authors revised the manuscript and approved the final version. T-CL submitted the study and is responsible for the overall content as guarantor.
Funding This study was supported primarily by the Ministry of Science and Technology of Taiwan (MOST 107-2314-B-039-049 & MOST 108-2314-B-039-039 & MOST 109-2314-B-039-031-MY2 & MOST 110-2314-B-039-021 & MOST 111-2314-B-039-018), National Science and Technology Council (NSTC 112-2314-B-039-042) and China Medical University Hospital (DMR-111-145).
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.