BackgroundThe relationship between body mass index (BMI) and mortality in older adults diminished. It is necessary to examine other factors that may accurately predict mortality in older adults. The visceral adiposity index (VAI) is an uncomplicated marker specific to the gender that incorporates anthropometric data and lipid profiles. VAI has been proposed as a marker of visceral adipose tissue dysfunction and of the related cardiometabolic risk. The aim of this study was to evaluate the link of VAI with all-cause mortality among the elderly.MethodsThe present prospective cohort study included data from 1999 to 2014 provided by the National Health and Nutrition Examination Survey (NHANES) in the United States. NHANES participants at or above the age of 65 were included. Data collection was carried out by taking face-to-face interviews, mobile-physical examinations, and lab tests. From the start of the survey to the end of December 2015, mortality-related follow-up statistics are available. The shape of the link between VAI and all-cause mortality was investigated using a restricted cubic spline model. Univariate- and multivariate-adjusted Cox proportional hazard models were estimated for VAI, and the results were presented as regression coefficients and 95% confidence intervals (CI).ResultsThe 82,091 NHANES participants represented 442.2 million non-institutionalized residents of the United States. A total of 11,173 older individuals (representing 23.3 million; aged 73.4 ± 5.8 years; 56.3% women, 82.7% non-Hispanic Whites, 6.8% non-Hispanic Blacks, and 3.3% Mexican Americans) were included in the study. During the 80-month follow-up period, 4466 fatalities were reported, including 825 deaths from cancer, 867 from heart disease, and 211 from cerebrovascular disease. The restricted cubic spline model demonstrated a robust J-shaped link between VAI and all-cause mortality, revealing a significant decrease in risk within the lower range of VAI, which attained the lowest risk close to 1.7. With VAI greater than 1.7, the risk of mortality increased with the increase of VAI (P for non-linearity = 0.025). In the multivariate-adjusted model, the risk of all-cause mortality was 0.73 (0.56-0.97) and 1.05 (1.01-1.09) in participants with VAI less than 1.7 and VAI greater than or equal to 1.7, respectively.ConclusionThis investigation is a population-based cohort study with high sample sizes and a long-term in older individuals follow-up that showed a J-shaped link between VAI levels and all-cause mortality. Understanding the independent roles of VAI in the relationship between BMI and mortality is crucial to understanding the obesity paradox phenomenon.