Comparisons of conventional and novel anthropometric obesity indices to predict metabolic syndrome among vegetarians in Malaysia

The study investigates which anthropometric obesity indices best predict metabolic syndrome (MetS) among Malaysian vegetarians and establishes sex-specific optimal cut-offs. MetS, a cluster of abdominal obesity, elevated BP, fasting glucose, triglycerides, and low HDL-c, has high global prevalence and is rising in many regions including Malaysia. Conventional indices (BMI, BF%, WHtR) and newer indices (LAP, VAI, ABSI, BRI) have shown variable performance across populations, and some require laboratory lipids. Vegetarians tend to have different body composition and lipid profiles compared to omnivores, and women and men exhibit differential MetS risk, suggesting that cut-offs derived from general populations may not apply to vegetarians. The purpose is to compare conventional vs novel indices for MetS prediction and to derive sex-specific cut-offs tailored to Malaysian vegetarians.

Prior work shows BMI cannot distinguish fat from muscle or visceral from subcutaneous adiposity, while BIA-derived BF% differentiates lean and fat mass but may not capture fat distribution. WHtR correlates with central adiposity and cardiometabolic risk and is proposed as a simple early-risk indicator, yet it does not separate subcutaneous from visceral fat. Novel indices like LAP and VAI target visceral adiposity but rely on blood lipids, potentially limiting large-scale use; ABSI and BRI aim to capture body shape/roundness without lipids, with mixed predictive results across populations. Dietary patterns influence MetS risk; vegetarians generally show lower MetS prevalence and healthier lipid profiles than non-vegetarians, but specific cut-offs for vegetarians are lacking. Sex differences in body composition and lipid profiles warrant sex-specific thresholds. These gaps motivate a focused evaluation in vegetarians with sex-specific cut-off determination.

Design and participants: Cross-sectional study among adult vegetarians (≥18 years, vegetarian ≥2 years), recruited via cluster sampling from nine randomly selected Buddhist and Hindu community centers in Kuala Lumpur and Selangor, Malaysia. Exclusions: pregnancy/lactation, medication for dyslipidemia, diabetes, or hypertension, and non-fasting on collection day. Of 355 consented, 273 met criteria (96 men, 177 women). Ethics approval: Universiti Putra Malaysia JKEUPM FPSK (FR16) P023; written informed consent obtained. Data collection: Self-reported demographics (age, sex, ethnicity), smoking (GATS), alcohol (NHANES FFQ items), physical activity (WHO GPAQ; categorized as insufficient, moderate, high). Anthropometrics followed ISAK protocols: height (SECA213, 0.1 cm), weight (TANITA HD306, 0.1 kg), BMI calculated (kg/m²), waist circumference (Lufkin W606PM, midpoint between lower costal border and iliac crest at end-expiration). Body fat percentage measured via Omron HBF-306-E handheld BIA. BP measured twice (Omron HEM-7121) after 5-minute rest; average used. Biochemistry: 10 mL overnight fasting venous blood for fasting blood glucose (FBG), triglycerides (TG), and HDL-c analyzed on Olympus AU640 analyzer. Indices computed: BMI = weight/height²; WHtR = WC/height; LAP (men) = [WC − 65] × TG; LAP (women) = [WC − 58] × TG; VAI (sex-specific formulas using WC, BMI, TG, HDL-c); ABSI = WC/(BMI^(2/3) × height^(1/2)); BRI per published geometric model. MetS definition: Joint Interim Statement (2009); MetS present if ≥3 of: abdominal obesity (WC ≥90 cm men, ≥80 cm women, Asian cut-offs), SBP ≥130 or DBP ≥85 mmHg, FBG ≥5.6 mmol/L, TG ≥1.7 mmol/L, HDL-c <1.0 mmol/L men or <1.3 mmol/L women. Statistical analysis: Normality by skewness (±2). Continuous data as mean ± SD or median (IQR); categorical as n (%). ROC analyses (SPSS v24) to derive AUC, sensitivity (Sn), specificity (Sp), PPV, NPV for each index by sex. AUCs compared using paired ROC design (NCSS v20.0.3). Sex-specific optimal cut-offs derived via Youden’s Index (YI). Significance at p < 0.05.

- Sample: 273 vegetarians (96 men, 177 women). Overall MetS prevalence 24.2% (men 29.2%; women 21.5%). - General characteristics: In both sexes, those with MetS had higher weight, WC, BMI, BF%, WHtR, LAP, VAI, SBP, DBP, FBG, TG and lower HDL-c (all p < 0.05). - Male ROC results: • Conventional indices: BMI AUC 0.777; BF% AUC 0.782; WHtR AUC 0.825 (all p = 0.0001). WHtR had the largest AUC within conventional indices; proposed WHtR cut-off 0.541. • Novel indices: LAP AUC 0.923 (cut-off 41.435), VAI AUC 0.864; ABSI AUC 0.652 (p = 0.060, not significant), BRI AUC 0.583 (p = 0.064, not significant). LAP outperformed VAI and others with higher Sn, Sp, PPV, NPV. - Female ROC results: • Conventional indices: BMI AUC 0.847 (cut-off 24.05), BF% AUC 0.785, WHtR AUC 0.863 (cut-off 0.532); all p = 0.0001. WHtR had largest AUC among conventional. • Novel indices: LAP AUC 0.920 (cut-off 21.743), VAI AUC 0.882, ABSI AUC 0.636 (p = 0.010), BRI AUC 0.537 (p = 0.480, not significant). LAP showed the best performance. - Direct comparison of best conventional vs novel indices: • Men: LAP AUC 0.923 vs WHtR AUC 0.825; difference 0.098 (11.9%), p = 0.016. • Women: LAP AUC 0.920 vs WHtR AUC 0.863; difference 0.057 (6.6%), p = 0.012. - Optimal cut-offs for Malaysian vegetarians: • WHtR: 0.541 (men), 0.532 (women). • LAP: 41.435 (men), 21.743 (women). - Additional observations: A larger proportion were classified obese by WC than by BMI, suggesting BMI may misclassify adiposity in this group.

The study demonstrates that among Malaysian vegetarians, WHtR is the most effective conventional anthropometric index and LAP is the most effective novel index for identifying MetS in both sexes. LAP significantly outperformed WHtR, likely because LAP incorporates both WC and TG, thus covering two MetS components and more directly reflecting visceral adiposity, which is strongly implicated in cardiometabolic risk. WHtR’s superiority over BMI and BF% within conventional indices aligns with its ability to capture central adiposity and adjust for height, enhancing early risk detection. ABSI showed limited utility (significant only in women with modest AUC), and BRI performed poorly in both sexes, suggesting limited applicability without population-specific recalibration. The derived sex-specific cut-offs for WHtR and LAP are higher or different from those reported in general or non-vegetarian populations, supporting the premise that dietary pattern and associated body composition differences necessitate tailored thresholds. Practically, LAP may be preferable in clinical settings where laboratory lipids are available, while WHtR offers a cost-effective alternative for large-scale screening without laboratory data.

LAP is the best predictor of MetS among Malaysian vegetarians, with WHtR as a practical alternative for large epidemiological surveys. Sex-specific optimal cut-offs identified were: WHtR 0.541 (men) and 0.532 (women); LAP 41.435 (men) and 21.743 (women). These findings provide vegetarian- and sex-specific thresholds to improve early identification of MetS. Future research should validate these cut-offs prospectively, assess applicability across different ethnicities and dietary patterns, and explore refinement of indices like ABSI and BRI for local populations.

- Cross-sectional design precludes causal inference between obesity indices and MetS. - Youden’s Index optimizes for two diagnostic groups and does not address intermediate states (e.g., prehypertension), potentially delaying identification of at-risk individuals. - Findings are limited to Malaysian vegetarians; generalizability to non-vegetarians or other countries may be constrained by dietary, anthropometric, and racial/ethnic differences. - No stratification by ethnicity within sex due to limited sample size, despite Malaysia’s multi-ethnic context where body composition may vary by ethnicity. - Novel indices requiring lipids (LAP, VAI) may be less feasible for large-scale, immediate screenings. - Mixed performance of ABSI and BRI suggests potential need for population-specific recalibration.