Non-communicable diseases deaths attributable to high body mass index in Chile

We estimated the proportion and number of deaths from non-communicable diseases (NCD) attributable to high body mass index (BMI) in Chile in 2018. We used data from 5927 adults from a 2016–2017 Chilean National Health Survey to describe the distribution of BMI. We obtained the number of deaths from NCD from the Ministry of Health. Relative risks (RR) and 95% confidence intervals per 5 units higher BMI for cardiovascular disease, cancer, and respiratory disease were retrieved from the Global BMI Mortality Collaboration meta-analyses. The prevalences of overweight and obesity were 38.9% and 39.1%, respectively. We estimated that reducing population-wide BMI to a theoretical minimum risk exposure level (mean BMI: 22.0 kg/m2; standard deviation: 1) could prevent approximately 21,977 deaths per year (95%CI 13,981–29,928). These deaths represented about 31.6% of major NCD deaths (20.1–43.1) and 20.4% of all deaths (12.9–27.7) that occurred in 2018. Most of these preventable deaths were from cardiovascular diseases (11,474 deaths; 95% CI 7302–15,621), followed by cancer (5597 deaths; 95% CI 3560–7622) and respiratory disease (4906 deaths; 95% CI 3119–6684). A substantial burden of NCD deaths was attributable to high BMI in Chile. Policies and population-wide interventions are needed to reduce the burden of NCD due to high BMI in Chile.


Scientific Reports
| (2021) 11:15500 | https://doi.org/10.1038/s41598-021-94974-z www.nature.com/scientificreports/ During the past three decades, Chile has experienced a fast economic growth and lifestyle transition 5,6 . These lifestyles changes have led the Chilean population to have one of the highest prevalences of obesity and other cardiovascular risk factors within the Latin American countries. In Chile, the prevalence of overweight or obesity reached 78% in 2017, among which, approximately, 4 million inhabitants (≥ 15 years) were living with obesity 7 . High prevalence of overweight and obesity may lead to a substantial increase in the burden of NCD [8][9][10] . For instance, it has been estimated that 8.7% (4394 out of 50,320 cases) of all cancer cases diagnosed in Chile in 2018 were attributable to high BMI 11 . Other studies have also estimated that obesity reduces life expectancy by 3.5 years 9 . However, national and subnational quantification of the burden of NCDs attributable to high BMI in Chile is unavailable. This quantitative information is timely and important to inform the development of interventions and public health policies aimed at counteracting the burden generated by high BMI.
In this study, we sought to estimate the proportion and number of deaths from NCD attributable to high BMI, using national and subnational representative data from Chile.

Material and methods
Study design. We obtained data from the National Health Survey of Chile 2016-2017 (National Health Survey-NHS). Briefly, the NHS was a cross-sectional household survey that included 6233 participants aged 15 years and over. NHS sampling method used a stratified and multistage selection of participants. Thirty strata were considered, which represented urban and rural areas of 15 geographical regions. In the multistage sampling, the selection was based on counties as the primary sampling unit, households as the secondary sampling unit, and, finally, one participant from selected households as the tertiary sampling unit. Sampling weights from the survey accounted for differences in selection probability and non-response rates, and the post-stratification adjustment allowed to expand the sample to the estimated population in Chile. Data collection was carried out between August 2016 and March 2017. One participant per household was randomly selected using a Kish computational algorithm and the response rate was 67%. The NHS 2016-2017 was funded by the Chilean Ministry of Health and approved by the Research Ethics Committee of the Faculty of Medicine of the Pontificia Universidad Católica de Chile (No. 16-019), and was performed in accordance with relevant guidelines and regulations. All participants gave their written consent before participating. Details on NHS 2016-2017 are available elsewhere 7 . In this study, we retrieved data from 5927 adults aged 20 to 96 years, who had comprehensive data available to measure weight and height.
The BMI was calculated as weight (in kg) by the square of height (in meters) 5,6 . Height was measured with a portable stadiometer with accuracy to the nearest 0.1 cm) and weight was measured with a digital scale (Tanita HD313) with an accuracy of 0.1 kg. Weight measurements were taken barefoot and the participants wore light clothing 7 . We estimated prevalences of underweight (≤ 18.5 kg/m 2 ), normal weight (18.5-24.9 kg/m 2 ), overweight (25.0-29.9 kg/m 2 ) and obesity (≥ 30.0 kg/m 2 ) 7 . We also calculated the mean and standard deviation of BMI. The NHS complex sample design was considered for calculating the prevalences of BMI categories and mean and standard deviation of BMI.
Relative risks (RR) and 95% confidence intervals (CI) per 5 units of BMI for all-cause, cardiovascular disease, respiratory disease, and cancer mortality were retrieved from the Global BMI Mortality Collaboration meta-analyses 14 .
Statistical analysis. Potential impact fraction (PIF) 15,16 for each NCD by sex was calculated using the following equation: where P is the mean and standard deviation (SD) of BMI in Chile in 2017, P* is the mean and standard deviation of BMI in the counterfactual scenarios (described below), RR is the relative risk per 1 kg/m 2 increase, and dx indicates the integration of RR according to BMI units (log-logit function was used to represent the dose-response relation between BMI and NCD 17,18 ). We considered two counterfactual scenarios of BMI distribution: To obtain the number of deaths from NCD attributable to high BMI, we applied PIF for cardiovascular disease, cancer and respiratory diseases to its respective number of deaths in Chile in 2018. Then, we summed the www.nature.com/scientificreports/ number of attributable deaths and divided it by total number of major NCDs (cardiovascular disease, cancer, and respiratory disease) and total number of deaths (all-cause mortality). We applied a simulation method in which repeated draws from PIF and the number of attributable/preventable deaths were generated considering the variability of BMI data. The simulation process was repeated 5000 times for each sex to calculate the mean and 95% confidence intervals (CI) of PIF and the number of attributable/preventable deaths.
We estimated that reducing population-wide BMI to a theoretical minimum risk exposure level (mean BMI 22 kg/m 2 and standard deviation 1) could prevent 21,977 deaths (95% CI 13,981 to 29,928) from NCD in Chile. These deaths represented about 31.6% of major NCD deaths (95% CI 20.1 to 43.1) and 20.4% of all deaths (95% CI 12.9 to 27.7) that occurred in 2018. Most of these preventable deaths were from cardiovascular disease   (Table 3). In men, the highest all-cause mortality PAF were found in the Magallanes/Antartica region (44.9%) followed by Los Lagos region (43.4%); in women, the highest all-cause mortality PAF was in Libertador B. O'Higgins region (41.6%) followed by Del Maule region (40.2%) (Fig. 2).

Discussion
In this study, we estimated the proportion and number of deaths from NCD attributable to high BMI using national representative data in Chile. We estimated that 21,977 deaths from NCD were attributable to high BMI, which represented about 31.6% of major NCD deaths and 20.4% of all deaths in 2018. Most of these preventable deaths were from cardiovascular disease, followed by cancer and respiratory disease. Regions with the highest burden of deaths attributable to high BMI were Los Lagos, Libertador B. O'Higgins and Maule.
Excess body weight is increasing and driving up its attributable burden throughout Chile, corroborating global findings for this specific risk factor 7 . Our study shows that high BMI played an important role in the national burden of NCD, especially those related to cardiovascular diseases. Prior studies have shown a positive association between excess body weight and mortality in four continents 14 . The present study has extended this existing evidence and reported the mortality burden of major NCDs using national and subnational representative data from Chile. Our results are similar to those found for other countries 14,19 . In Brazil, the decrease in BMI could prevent up to 65,721 deaths from NCDs. Approximately 25.3% of major NCD deaths and 14.9% of all deaths could be prevented each year in Brazil by reducing population-wide BMI to the theoretical minimum risk 14 .
In the United States, 280,000 deaths are attributed to excess body weight. Population attributable fractions for all-cause mortality from overweight or obesity were 19% in North America 14 . Furthermore, more than 60% of overweight individuals live in low-and middle-income countries, and, in these settings, the impact of high BMI on morbidity and mortality has been understudied 20 .
Major causes of death, including cardiovascular disease, and cancers are closely associated with high BMI 21 . Among these leading causes, a clear majority were due to cardiovascular diseases. We estimated that 41.8% of     17 . These findings provide insight into the underlying contribution of excess body weight to deaths from NCDs, indicating that reducing population-wide BMI may have a great contribution for the prevention of premature deaths from NCDs. Worldwide, it has been estimated that 15% of deaths from all-causes are due to a high BMI 24 . In United States the reported estimates ranged from 5 to 15% for all-cause mortality 25 . In South American countries, a previous study has calculated that a high BMI is responsible for approximately 12.3% of all deaths 26 . Worldwide, 4.7 million deaths from NCD were associated with a high BMI, accounting for 17.7% of all causes of death in 2017. This represents a 118.8% increase since 1990, with a projection of 5.5 million deaths in 2025 21 . Our findings of 20.3% of all-cause mortality due to high BMI is one of the largest reported in the literature so far. These results are likely due to the remarkably high prevalence of overweight and obesity in Chile (approximately 78%).
The association between high BMI and mortality has been widely reported in the literature [1][2][3] . There are several biological mechanisms through which high BMI increases the risk of NCD [27][28][29] . High BMI causes chronic systemic inflammation and higher sympathetic activity, which can contribute to insulin resistance and hypertension, respectively 27 , leading to endothelial dysfunction and atherosclerosis 30 . Thus, its effect is primarily mediated through other intermediary risk factors, such as hypertension, hypercholesterolemia, and hyperglycemia, the last two also known as metabolic risk factors 27 . www.nature.com/scientificreports/ Our study suggests a high mortality burden of NCDs related to BMI, which may be useful to improve public awareness and support population-wide interventions and public health policies to reduce excess body weight in Chile. Public health policy and other relevant stakeholders should pay consideration to this issue and prioritize their agenda, designing and implementing effective policies. Evidence-based interventions should be aimed at empowering and educating individuals, making them aware of the potentially detrimental consequences of their higher levels of adiposity 31 , and encouraging physical activity. New data and communication technologies could be exploited for the aim of both gathering BMI and other pertinent evidence and fostering changes, tracking variations over time 4 .
In this study, we used measured BMI from a representative sample of Chilean adults to estimate the burden of deaths from NCD attributable to high BMI. However, our study has several limitations. Although BMI is a useful indicator of overall body fat, it does not distinguish between lean and fat tissues 32 . However, it is unlikely that body fat measurements could be used extensively in clinical practice as BMI currently is due to the additional cost associated with this sort of measurement. Furthermore, we used RR estimates from a meta-analysis that included data from four continents but not from Chile 14 . It is unknown if these RR estimates are applicable to Chileans. However, these reliable RR estimates included never-smokers who had no prevalent diseases and excluded the first five years of follow-up to reduce reverse causality 14 . We applied the RR estimated for Chile's general population, regardless of disease and smoking status. NHS is not carried out annually to compare prevalence for 2016-2017 and 2018. We assumed that the distribution of BMI in 2018 is similar to the one obtained in the ENS 2016/2017.

Conclusions
The prevalence of overweight and obesity is high (approximately 78%) in the Chilean population, with some variability across regions. Such high BMI conferred, approximately, 21,977 deaths from NCD per year in Chile, which represents about 31.6% of major deaths from NCD and 20.4% of all deaths in 2018. These results suggest the need for public health programs aimed at maintaining a healthy weight in order to tackle the NCD burden in Chile. Strong public health policies aimed to reduce obesogenic environments are needed to decrease obesity and its health burden in Chile.

Data availability
The datasets generated and/or analysed during the current study are available in the database repository of the Epidemiology Department of the Chilean Ministry of Health: http:// epi. minsal. cl/ bases-de-datos/. Data are available upon reasonable request from the corresponding author.