The study found that optimal cut-off values for various anthropometric measures differ between men and women. For men, the optimal cut-offs were 85.0 cm for waist circumference (WC), 93.5 cm for hip circumference (HC), 0.89 for waist-hip ratio (WHR), and 0.53 for waist-to-height ratio (WHtR) at the 25% body fat threshold. For women, the corresponding values were 81.1 cm for WC, 97.2 cm for HC, 0.84 for WHR, and 0.51 for WHtR at the 35% body fat threshold. Women exhibited higher sensitivity for WC, HC, WHR, and WHtR compared to men, but men had higher specificity for WHR and WHtR measures. These findings highlight the importance of considering sex-specific differences when using anthropometric indices to assess obesity and related health risks.

Unlike the higher waist circumference cut-off values, the optimal cut-off values for WHR and WHtR to define obesity in this study closely align with previous findings: a WHR of 0.88 for males and 0.82 for females, and a WHtR of 0.49 for males and 0.50 for females, as reported by Sinaga et al. (2018) [22]. Similarly, the reported waist circumference (WC) values were higher compared to those of Indonesian adults [24], as well as the waist circumference values for Sri Lankan women (WC: 76 cm) [68]. Unlike the current study, if the body fat thresholds were lowered from 35 to 33%, the corresponding cut-off values were aligned and found to be 80 cm for WC, 95 cm for HC, and 0.83 for WHR among Sri Lankan adult women [23]. Similarly, lower waist circumference cutoffs of 75.6 cm and 80.5 cm for men, and 71.5 cm and 81.5 cm for women of Nigerian and Cameroonian origin, respectively, were used for identifying hypertension [45]. Waist circumference cutoffs of ≥ 91.0 cm for men and ≥ 82.3 cm for women among Botswanans were determined to be the optimal cutoff points for abdominal obesity, predicting at least two other indicators of metabolic syndrome in both genders [69].

The waist to height ratio cut off values proposed in the current study for both in men and women were in line with the index of central obesity (ICO) cutoffs obtained by considering the waist circumference cutoffs suggested by IDF and average heights from various country sources ranged from 0.51 to 0.58 among males and 0.47 to 0.54 among females [18]. The small difference in ICO cutoffs between males and females suggests a common ICO cutoff could be applicable for both genders, regardless of country of origin [18]. The current study’s proposed WHtR cut-off values of 0.53 for men and 0.51 for women align with the widely acknowledged boundary of 0.5 proposed for both sexes. This supports the prior recommendation to ‘Keep your waist circumference less than half your height’ - a simple, inexpensive screening method for identifying increased health risks [21].

The WC cut off values reported in the current study were lower than a WC cut off 93 cm and 84 cm for Thais men and women respectively [28]. Likewise, for Latin Americans, the optimal waist circumference (WC) thresholds were determined to be 94 cm for men and between 91 and 92 cm for women [30]. In Brazil, the waist circumference (WC) cut-off was 90.2 cm for men and 92.3 cm for women [26]. Moreover, a study in Thailand [28]revealed that higher waist circumference (WC) cut-off values of 93 cm for men and 84 cm for women.

The differences in circumference cut-off values may be attributed to various factors, including demographic variations such as age and ethnicity, the specific study participants [22, 23, 26, 28, 30], variations in body composition assessments methods [22, 23, 26, 28, 30], and the thresholds used to define the body fat level, and variation in the waist and hip circumference measurement protocols [23, 26, 30]. Nonetheless, similar patterns of association were observed between health outcomes and all waist circumference protocols across sample size, sex, age, race and ethnicity, and the waist circumference measurement protocol had no substantial influence on the association between waist circumference, all-cause and CVD‐specific mortality, and risk of CVD and diabetes [70].

Most importantly, the optimal cut-off values for waist circumference (WC) and waist-hip ratio (WHR) to define obesity in the current study were lower compared to the cut-off values recommended by the World Health Organization (WHO), International Diabetes Federation (IDF), the National Cholesterol Education Program (NCEP-ATP III), and national institute of health(NIH) for both men and women [6, 12, 13, 47]. This reaffirms the relevance of ethnic and context-specific anthropometric cut-off values, which further indicates the impacts of variations in age, sex, and ethnicity on body composition [6].

Notwithstanding, studies in Africa recommend lower waist circumference (WC) and waist-to-hip ratio (WHR) cutoff values for central obesity [71]. However, due to the insufficient strength of evidence [71], these new threshold values have not yet been established as they have been for Asian populations [47, 72]. Consequently, some countries adhered to one or more recommended waist circumference (WC) and waist-to-hip ratio (WHR) cutoff values, while others established their own specific guidelines. However, the rationale behind the choice and implementation of these specific recommendations often remains unclear [6]. In particular, a consensus statement from the International Diabetes Federation recommending the use of European data until more specific data for Sub-Saharan Africa become available [72].

Furthermore, the WHO [6] recommended a WC cut-off of > 94 cm for men and > 80 cm for women, while the WHR cut-off was ≥ 0.90 for men and ≥ 0.85 for women. The IDF proposed sex- and ethnic-specific WC cut-off values. For Europid men, the cut-off is > 94 cm, while for Asian, Chinese, and Japanese men, it is > 90 cm. Similarly, the WC cut-off for Europid, Asian, Chinese, and Japanese women is > 80 cm [47]. In accord with the NCEP-ATP III guidelines, the WC cut-off values for men and women are > 102 cm and > 88 cm, respectively [13]. This acknowledged that applying a single waist circumference cut-off point may underestimate the prevalence of central obesity in shorter individuals [20]. The need for distinct gender and race-specific cut-off points to define central obesity is likely due to differences in average height across these populations [19]. Thus, the current findings suggested that population-specific optimal cut-off values may be necessary to accurately define obesity using anthropometric measures, as the thresholds recommended by international organizations may not be appropriate for all populations.

The proposed cut-off values for waist circumference (WC) and waist-to-height ratio (WHtR) demonstrated the highest performance in detecting obesity, with excellent diagnostic accuracy for both men and women. Specifically, WC, HC, and WHtR had AUCs of 0.93 (95% CI: 0.90–0.95) among women. These findings align with a previous study in Southwest Ethiopia [22], which found waist-to-height ratio and waist circumference had larger ROC curve areas, indicating higher sensitivity and specificity for detecting obesity in both men and women. In contrast, waist-to-hip ratio had a relatively lower area covered under the ROC curve. Similarly, a study among Sri Lankan women demonstrated that WC exhibited higher discriminatory power to detect obesity as compared to WHR [23]. An Indonesian study [24] reported that WC and WHtR showed superior discriminatory performance, with WC being the most effective indicator of body fat percentage (%BF) in both genders. Contrarily, WHR was identified as the least effective indicator, with the smallest area under the ROC curve.

The new proposed cutoff values for waist circumference, hip circumference, and waist-to-height ratio had sensitivity ranging 70–93% and specificity of 83–94% in accurately identifying obesity based on body fat percentage. A study in Southwest Ethiopia [22] found optimal cutoffs of 83.7 cm for waist circumference and 0.49 for waist-to-height ratio had high sensitivity (85–86%) and specificity (90–92%) for detecting obesity in men. For women, a WC cut off of 78.0 cm and 0.50 for WHtR showed the sensitivity and specificity ranged from 84 to 86% and 85–87% respectively. For Sri Lankan adults [68], a WC cutoff of 76.2 cm showed 82.5% sensitivity and 75.9% specificity in women, while a cutoff of 85.0 cm showed 85.4% sensitivity and 85.4% specificity in men. In the Korean population [25], the WC cutoffs were 89.8 cm for men (84.7% sensitivity, 91.7% specificity) and 86.1 cm for women (83.6% sensitivity, 62.9% specificity), based on visceral fat area. Among Latin Americans [30], a WC threshold of 94 cm for men (89.8% sensitivity, 80.2% specificity) and a range of 91–92 cm for women (75.9–74.5% sensitivity, 71.7–74.5% specificity) were found when considering a threshold level of > 100 cm² for visceral adiposity area. These exemplifies that, compared to international thresholds [6, 13, 47], the country-specific cutoffs for waist circumference, waist-hip ratio, and waist-height ratio showed improved performance in diagnosing obesity. The evidence showed substantial population-dependent variations in the association between disease risk and measures of abdominal obesity, when considering ethnicity-specific cut-off points [6]. Due to this relevancy, the International Atherosclerosis Society (IAS) and International Chair on Cardiometabolic Risk (ICCR) consensus statement recommend that the WC measurements should become a vital sign in clinical practices. Since waist circumference is associated with health outcomes within all BMI categories independent of sex and age [15].

The current study determined the cut-off values for underweight, normal weight, and overweight body weight categories. Accordingly, for men, the waist circumference cut-offs were: ≤ 74.55 cm (underweight), 74.56–78.95 cm (normal), and 79.0–85.0 cm (overweight). For women, the cut-offs were: ≤ 68.25 cm (underweight), 68.26–79.59 cm (normal), and 79.60–81.10 cm (overweight). Consistently, the healthy weight for Asian Indian adults was defined by using a WC cut-off of 75.8 for men and 70 cm for women, and a WHR cut-off of 0.79 for men and 0.74 for women [73]. However, the methodology and study population differed from the current one. The optimal BMI-specific WC thresholds for white women, Africa American women, white men, and Africa American men were as follows: 72, 76, 82, and 78 cm for normal-weight (18.5–24.9 kg/m 2 ); 87, 85, 95, and 92 cm for overweight (25–29.9 kg/m 2 ) participants respectively [17]. However, the reported WC thresholds for elevated cardiometabolic risks are based on existing BMI classifications, rather than directly on body adiposity levels as determined in this study.

The waist-to-hip ratio (WHR) cut-offs are as follows: ≤ 0.86 (underweight), 0.86–0.87 (normal), and 0.88–0.89 (overweight) for men; and ≤ 0.79 (underweight), 0.80–0.83 (normal), and 0.83–0.84 (overweight) for women. These proposed values align with a study reported in Nigeria [74], where good health in men is defined by a WHR of less than 0.85, and better health is indicated by a value between 0.85 and 0.89. For women, good health corresponds to a WHR of less than 0.75, while better health falls between 0.75 and 0.79. Understanding WHR aids in managing body fat and reducing the risk of obesity and cardiovascular diseases, as an increasing WHR signals a higher risk if left unaddressed [74].

The waist-to-height ratio thresholds were: ≤0.43 (underweight), 0.44–0.47 (normal), and 0.48–0.53 (overweight) in men; ≤0.43 (underweight), 0.44–0.50 (normal), and 0.50–0.51 (overweight) in women. The proposed cutoff values are in ranges, with a widely accepted WHtR boundary of 0.5, meaning waist circumference should be less than half of height [21]. Furthermore, these proposed WHtR values were also signifies the different action which would be considered as per the Ashwell’s recommendation. According to Margaret Ashwell’s Shape Chart for practical screening based on WHtR, four regions guide patient actions: WHtR less than 0.4 suggests ‘Take Care’; between 0.4 and 0.5 indicates ‘Healthy Shape’; between 0.5 and 0.6 advises ‘Consider Action’; and above 0.6 signals ‘Probability Your Health Is at Risk [21].

Widely, the reported findings indicate that the local cut-off values have enhanced the effectiveness of anthropometric measurements in assessing body adiposity based on body fat percentage. Nonetheless, slight differences observed in the performance of circumference measurements could be attributed to variations in cutoff values, methods used to assess body composition and threshold used to define obesity, age, and ethnic diversity among the study participants [23, 26, 28, 30]. Indeed, direct studies linking body fat measurements to morbidity and mortality are limited, making it difficult to establish precise cut-off points for obesity based on body fat measurements. Thus, there is currently no clear consensus on the cut-off point for obesity solely based on body fat measurements [65].

The proposed cut-off points for waist circumference, waist-hip ratio, and waist-to-height ratio have several potential uses: surveillance and screening for obesity and associated health risks; diagnosis and treatment decisions in clinical settings, and assessing the value and effectiveness of interventions for individual patients and at the community level [6]. Furthermore, the relevance to public health lies in the ability to predict disease burden and guide prevention efforts, rather than just predicting mortality. These measures and cut-offs can inform policy decisions on the need for and effectiveness of interventions to reduce obesity-related health risks and associated costs [6].

Altogether, the current findings have significant implications for the understanding and assessment of weight-related health conditions. The disparities observed between our study and the existing WHO [6] classification highlight the need for potential revisions in the classification criteria for overweight and obesity. Furthermore, it also emphasizes the need to consider ethnic-specific cut-off values when assessing overweight and obesity among Ethiopians. Therewithal, further research and consideration should be given to these discrepancies to ensure accurate and appropriate categorization of weight-related health conditions for both men and women.

Despite the fact that, this study has notable strengths. Firstly, it benefits from a large sample size, increasing the applicability of the results. Moreover, being a population-based study, it provides unique insights into establishing anthropometric cut-off values in Ethiopia. The inclusion of participants from both urban and rural areas ensures diversity in the population representation. Additionally, the use of bioelectrical impedance as a reliable reference method for body composition assessment adds practicality. The study’s adoption of standard receiver operating characteristic (ROC) techniques strengthens the determination of cut-off values for each anthropometric measurement. However, the body composition assessment method used in the study may be less reliable compared to more advanced methods such as multicompartment models [75]. While multicompartment models offer accurate body composition measurements for validating other methods, their limited availability, high cost, and radiation exposure restrict their widespread regular use [75]. Meanwhile, unlike central or abdominal adiposity, which can be accurately measured using techniques like CT, MRI, or DEXA [76], the current study used total body fat percentage to establish the cutoff values. Errors in anthropometric measurements are still possible, despite this study implementing rigorous precautions, including extensive training for data collectors, calibration of equipment, and standardization of procedures to minimize such errors. Although we do not anticipate significant differences among the various ethnic groups in Ethiopia, as most distinctions are related to language and culture, not all ethnic categories were included.