Dietary patterns derived by reduced rank regression, macronutrients as response variables, and variation by economic status: NHANES 1999–2018

Abstract

Purpose: Macronutrient intakes vary across people and economic status, leading to a disparity in diet-related metabolic diseases. This study aimed to provide insight into this by: (1) identifying dietary patterns in adults using reduced rank regression (RRR), with macronutrients as response variables, and (2) investigating the associations between economic status and macronutrient based dietary patterns, and between dietary patterns with central obesity (waist circumference) and systemic inflammation (C-reactive protein [CRP]). Methods: 41,849 US participants from the National Health and Nutrition Examination Survey (NHANES), 1999–2018 were included. The percentages of energy from protein, carbohydrates, saturated fats, and unsaturated fats were used as response variables in RRR. Multivariable generalized linear models with Gaussian distribution were employed to investigate the associations. Results: Four dietary patterns were identified. Economic status was positively associated with both the high fat, low carbohydrate [βHighVsLow = 0.22; 95% CI: 0.16, 0.28] and high protein patterns [βHighVsLow = 0.07; 95% CI: 0.03, 0.11], and negatively associated with both the high saturated fat [βHighVsLow = -0.06; 95% CI: -0.08, -0.03] and the low alcohol patterns [βHighVsLow = -0.08; 95% CI; -0.10, -0.06]. The high saturated fat pattern was positively associated with waist circumference [βQ5VsQ1 = 1.71; 95% CI: 0.97, 2.44] and CRP [βQ5VsQ1 = 0.37; 95% CI: 0.26, 0.47]. Conclusion: Macronutrient dietary patterns, which varied by economic status and were associated with metabolic health markers, may explain associations between economic status and health.