Study type and design

An analytical cross-sectional study was conducted using the data from the 2017–2018 Food and Nutrition Surveillance by Life Stages (VIANEV) survey. The VIANEV survey has national representativeness and includes data collected on anthropometric measurements, lifestyle habits, and biochemical analyses of Peruvian adults aged 18–59 years, with a fasting period of at least 9–12 hours [20]. The survey excluded adults who consumed food prior to biochemical assessments, those with gastrointestinal diseases affecting diet, and those with anatomical conditions preventing accurate anthropometric measurement. Further methodological details of the survey are available in the official technical report [20].

The survey collected information across three domains: Metropolitan Lima (the capital of Peru), the rest of the urban area and rural areas, using a stratified, multistage, probabilistic, and independent sampling design. Sampling was conducted in two stages: first, by randomly selecting clusters as the primary sampling unit, and subsequently by randomly selecting households with adult members as the primary unit. The VIANEV 2017–2018 survey was conducted among a subsample of adults who also participated in the National Household Survey 2017 (ENAHO). ENAHO constitutes one of the largest population-based surveys in Peru and widely used to inform health decision-making and guide public policy in the country. In contrast to ENAHO, which focuses primarily on demographic and economic aspects, VIANEV participants were asked a broader set of health-related questions [21].

Study population

Adults aged 30–59 years were included in the present analysis. Participants younger than 30 years were excluded because the development and validation of the Framingham risk score and the estimation of 10-year cardiovascular risk were performed in adults aged 30–74 years [14]. Therefore, estimating 10-year cardiovascular risk in younger individuals would not reflect a real risk [22]. Additionally, participants with missing values in any of the variables of interest were excluded.

Variables

Exposure variable: obesity.

Obesity was defined using the RFM index, calculated from height (cm), WC (cm), and sex of participants. WC was measured using a 200-cm measuring tape with 1-mm precision, positioning it at the midpoint between the lower margin of the last rib and the upper border of the iliac crest. Height was assessed with a fixed wooden stadiometer [20]. The following calculation was performed [11]:

where sex was coded as 0 for men and 1 for women. Cut-off points of ≥40% for women and ≥30% for men were applied to define obesity as a dichotomous variable (No/Yes). The selection of these cut-off points was based on prior validation studies for obesity diagnosis and mortality prediction [23], as well as their use in a Peruvian population-based survey [24].

Statistical analysis

All statistical analyses were performed using Stata version 18 (StataCorp LLC). To obtain additional socioeconomic information of VIANEV participants, the VIANEV 2017–2018 database was merged with the ENAHO 2017 survey using strata, clusters, dwellings, and households as linkage variables.

Given the characteristics of the VIANEV survey, the analyses accounted for sample weights and the complex survey design. All analyses were stratified by sex to account for sex-specific differences in body composition. Participant characteristics were described using absolute frequencies, weighted frequencies (with 95% confidence intervals [95% CI]), and measures of central tendency (means and medians) with their respective measures of dispersion (standard deviation and interquartile range).

To assess the association between RFM and the Framingham risk score, crude and adjusted generalized linear models (GLMs) from the gamma family with a logarithmic link function were employed. All covariates were included in the adjusted models based on their epidemiological relevance. This model was chosen because it is appropriate for positively skewed continuous outcomes such as Framingham score values. The strength of association was expressed as β coefficients with their 95% CI and standard errors. Additionally, as a sensitivity analysis, GLMs were also fitted considering RFM as a continuous variable and categorized into tertiles derived from our own sample, with the highest tertile representing the greatest RFM values.

Ethical considerations

The VIANEV 2017–2018 survey data are publicly available and do not contain information that allows identification of participants. The dataset was accessed for research purposes on April 29, 2025, through the Peruvian National Open Data Platform (https://datosabiertos.gob.pe/dataset/estado-nutricional-en-adultos-de-18-59-a%C3%B1os-per%C3%BA-2017-%E2%80%93-2018). Similarly, ENAHO data are publicly available at https://proyectos.inei.gob.pe/microdatos/. At no point did the authors have access to identifiable information about the participants. The interviewees provided their verbal consent to participate. Therefore, approval from an institutional ethics committee was not required.

Comparison with previous studies

The findings of this study revealed that the presence of obesity defined by RFM was associated with a 48% and 61% higher estimated 10-year cardiovascular risk in men and women, respectively, compared with non-obese individuals. Additionally, when stratifying by RFM tertiles, a dose-response relationship was observed, with the estimated 10-year cardiovascular risk in the highest tertile being 84% in men and 60% in women, compared with the lowest tertile. These results are consistent with previous evidence supporting the clinical utility of RFM in assessing cardiovascular risk, both in cross-sectional and prospective studies. For instance, in a cross-sectional analysis of more than 11,000 Chinese adults, each standard deviation increase in RFM was associated with higher odds of CVDs in men (OR: 1.66; 95% CI: 1.36–2.02) and women (OR: 1.26; 95% CI: 1.08–1.47) [28]. Similarly, in U.S. adults, RFM was significantly associated with CVDs in both sexes, with 4% higher odds in men and 3% higher odds in women when comparing the highest versus the lowest quintile of RFM [29]. Moreover, prospective studies have documented that higher RFM is associated with greater incidence of heart failure [30], as well as cardiovascular mortality [12]. Despite variability in the magnitude of the association between RFM and CVDs across studies, which may be attributed to differences in how outcomes are measured and defined, taken together, these findings support the potential clinical utility of RFM in cardiovascular risk assessment.

Potential explanatory mechanisms

The observed association between RFM and estimated CVDs risk may reflect both pathophysiological and methodological factors. RFM incorporates height and WC in estimating body fat percentage [11]. Waist circumference is a marker of central adiposity distribution, particularly visceral adiposity [31], which has been widely associated with cardiovascular risk factors [32]. Visceral adipose tissue is metabolically active and characterized by a more lipolytic and inflammatory profile than subcutaneous adiposity, and is associated with lipid and glucose metabolism disturbances [33]. Greater visceral adiposity has also been associated with higher blood pressure [34]. Consistent with this mechanistic framework, a longitudinal study in Peruvian adults reported that RFM-defined obesity was associated with a higher incidence of hypertension over five years, which represents a major component of cardiovascular risk, thereby supporting the relevance of RFM in cardiometabolic risk assessment [35]. Taken together, increased central adiposity has been closely related to several components of cardiovascular risk [14]. Consistent with this, it has been documented that for every 10 cm increase in WC, CVDs risk increases by 3.4% in women and 4% in men [36]. From a methodological perspective, RFM has been validated against DEXA, the gold standard for assessing body composition [9]. Furthermore, RFM is based on the height-to-waist ratio, essentially the inverse of the WHtR, which has shown a strong association with CVDs risk [36]. However, our results should be interpreted with caution as the Framingham risk score has neither been validated nor recalibrated for the Peruvian population. Chile is one of the few countries in the region that has undertaken efforts to recalibrate the model for its own population [37]. Since the score was originally developed using baseline cardiovascular event rates from a U.S. cohort, the absolute levels of predicted 10-year risk may differ substantially from those in the Peruvian context and other Andean countries. Nevertheless, the model likely preserves the relative ranking of individuals according to higher or lower estimated cardiovascular risk.

Sex-based differences

Additionally, our findings show a stronger association between obesity and estimated cardiovascular risk in women, which is consistent with evidence indicating a higher burden of cardiovascular risk factors in women compared to men [38]. In Peru, studies have consistently reported a higher prevalence of obesity and metabolic syndrome among women, who also have markedly higher odds of abdominal obesity, reinforcing sex-specific vulnerability in adiposity-related cardiometabolic risk [39,40]. Together, these findings may help contextualize why increases in RFM are associated with larger relative changes in estimated 10-year cardiovascular risk among women in our study.

On the other hand, epidemiological studies in Latin America [4] have reported a higher incidence of CVDs in men than in women (655 vs. 531 per 100,000 individuals), and additional evidence indicates that Peruvian men present worse cardiovascular health than women [5,41]. Of note, compared with previous reports, our results address related but not identical dimensions of cardiovascular risk. Evidence from other Andean populations shows that despite women having a higher prevalence of risk factors, men experienced higher cardiovascular event incidence, suggesting that estimated risk may not fully reflect observed outcomes [42]. This pattern suggests that a greater burden of risk factors does not necessarily translate into higher event incidence, and that sex-specific patterns of estimated cardiovascular risk may not align perfectly with observed outcomes. These findings do not contradict our results, as the Framingham Risk Score estimates predicted cardiovascular risk based on multiple factors, which may vary independently of actual event incidence. While our study assessed estimated 10-year risk, previous studies examined observed cardiovascular events. Both approaches are complementary and highlight the need for longitudinal studies in Peru to determine whether adiposity has differential cardiometabolic effects by sex and how these translate into actual outcomes.

Implications for public health and clinical practice

In the Region of the Americas, particularly in low- and middle-income countries such as Peru, CVDs remain one of the leading causes of mortality. In response, and within the framework of the Global Action Plan for the Prevention and Control of Noncommunicable Diseases [43], the World Health Organization has promoted the Global HEARTS Initiative [44], and its regional adaptation, HEARTS in the Americas, which aims to strengthen health service performance for CVDs prevention and control. A main component of this strategy is the estimation of 10-year cardiovascular risk and the use of simple, accessible tools for its calculation in resource-limited settings [45]. Given that the use of simple anthropometric indicators is key in these strategies, it is relevant to evaluate their relationship with cardiovascular risk estimates in local populations. In this context, our findings suggest that obesity defined by RFM is associated with higher 10-year cardiovascular risk, as estimated by the Framingham risk score, which may be useful for future research exploring the potential role of RFM as an anthropometric marker in cardiovascular risk stratification. Although BMI is widely used to define obesity and is also associated with CVDs [46], it has important limitations that may lead to misclassification when detecting adults with excess body fat [47]. A high BMI does not necessarily reflect greater adiposity, as it is derived from total body weight and does not differentiate by sex [48]. These findings seek to contribute to the growing body of evidence supporting the need to implement alternative diagnostic criteria for obesity that can more accurately estimate cardiometabolic risk in Peru. To advance this effort, several key steps are required. First, drawing on the experience of countries such as the United States [49] and the United Kingdom [50]—where WC and WHtR measurements are endorsed by clinical guidelines for estimated cardiovascular risk assessment—there is a need to recalibrate RFM for the Peruvian context, given the country’s distinct ethnic characteristics. This requires population-based studies that include a valid reference standard for adiposity, such as DEXA, as well as cardiometabolic outcomes that can inform the establishment of appropriate cut-off points. Subsequently, considering the simplicity and very low cost of measuring WC and height in primary care settings, implementing RFM through pilot studies would help quantify the cardiovascular risk that BMI commonly underestimates. The main barriers to implementation would likely include the need to update clinical guidelines and potential resistance among clinical personnel when adopting a new adiposity estimator.

Study limitations

This study has several limitations that should be considered when interpreting our findings. The analysis was based on secondary data from the VIANEV 2017–2018 survey; therefore, recording errors or missing information cannot be ruled out, as the data were collected for purposes other than this research. In addition, due to the cross-sectional design of the study, it is not possible to establish causality between exposure and outcome. Nevertheless, these results may represent an initial step toward the development of more complex methodologies aimed at identifying potential causal pathways between variables. Moreover, the characteristics of excluded participants may have influenced the strength of the reported associations. We also conducted a sensitivity analysis to evaluate whether the association retained its direction and significance when modifying the nature of the exposure variable (RFM). Furthermore, the measurement of anthropometric variables, although standardized, may vary across interviewers or equipment. Some self-reported variables, such as lifestyle habits or living conditions, may have been influenced by social desirability or recall bias. The lack of potentially confounding variables, such as diet quality, energy intake or more specific physical activity measurement, could also result in residual confounding, affecting the magnitude and direction of the estimated associations due to its strong influence on body fat percentage and components of the Framingham Risk Score; nevertheless, the adjustment models included indirect measures of diet quality (daily consumption of five or more portions of fruits and vegetables) and physical activity (daily hours of sedentary behavior) in order to reduce confounding, although residual confounding cannot be completely excluded. Although the Framingham risk score is widely used to estimate 10-year cardiovascular risk, its accuracy when applied outside the original U.S. population may lead to over- or underestimation of risk. In addition, the lack of validation in the Peruvian population may limit its applicability in this context. Nevertheless, the score is currently recommended by the Peruvian Ministry of Health for cardiovascular risk stratification [16], which supports its use in local epidemiological analyses. Finally, since the data were collected before the COVID-19 pandemic, patterns of obesity and cardiovascular risk may have shifted thereafter, restricting the direct applicability of our findings to the current context.