Multi-Nutrient Analysis of Dietary Macronutrients with All-Cause, Cardiovascular, and Cancer Mortality: Data from NHANES 1999–2014

Macronutrients are a major component of the human diet. However, few studies have assessed their collective association with mortality. We sought to evaluate the associations of macronutrient intake with all-cause, cardiovascular, and cancer mortality in US adults using a multi-nutrient approach. This prospective cohort analysis used data from the National Health and Nutrition Examination Survey from the years 1999 to 2014. The participants included 33,681 US adults aged 20–85 years (52.5% female). The maximum follow-up time was 16.8 years, with a total of 4398 total deaths, including 772 cardiovascular deaths and 952 cancer deaths. The associations between mortality and dietary macronutrients were explored using three-dimensional generalized additive models, allowing for visual and statistical inference of complex nonlinear associations. Absolute macronutrient intake demonstrated a three-way interactive association with all-cause mortality (p < 0.001), cardiovascular mortality (p = 0.02), and cancer mortality (p = 0.05), adjusted for age, sex, ethnicity, socioeconomic status, dietary quality, and lifestyle. Compositionally, a high caloric diet composed of moderately high protein (20%), moderate fat (30%), and moderate carbohydrate (50%) levels was associated with the highest mortality risk. Across the total energy intake levels, lower mortality risk was observed in two separate regions consisting of higher protein (30%), higher carbohydrate (60%), and lower fat levels (10%) or lower protein (10%), moderate carbohydrate (45%), and higher fat levels (45%). These findings highlight a complex nonlinear and interactive association between macronutrients and all-cause mortality such that several distinct dietary compositions are associated with similarly high or low risk. Future research is needed to explore the drivers of these associations and whether they differ across varying dietary patterns and populations.