A Novel Personalized Systems Nutrition Program Improves Dietary Patterns, Lifestyle Behaviors and Health-Related Outcomes: Results from the Habit Study

Abstract

Personalized nutrition may be more effective in changing lifestyle behaviors compared to population-based guidelines. This single-arm exploratory study evaluated the impact of a 10-week personalized systems nutrition (PSN) program on lifestyle behavior and health outcomes. Healthy men and women (n = 82) completed the trial. Individuals were grouped into seven diet types, for which phenotypic, genotypic and behavioral data were used to generate personalized recommendations. Behavior change guidance was also provided. The intervention reduced the intake of calories (-256.2 kcal; p < 0.0001), carbohydrates (-22.1 g; p < 0.0039), sugar (-13.0 g; p < 0.0001), total fat (-17.3 g; p < 0.0001), saturated fat (-5.9 g; p = 0.0003) and PUFA (-2.5 g; p = 0.0065). Additionally, BMI (-0.6 kg/m²; p < 0.0001), body fat (-1.2%; p = 0.0192) and hip circumference (-5.8 cm; p < 0.0001) were decreased after the intervention. In the subgroup with the lowest phenotypic flexibility, a measure of the body's ability to adapt to environmental stressors, LDL (-0.44 mmol/L; p = 0.002) and total cholesterol (-0.49 mmol/L; p < 0.0001) were reduced after the intervention. This study shows that a PSN program in a workforce improves lifestyle habits and reduces body weight, BMI and other health-related outcomes. Health improvement was most pronounced in the compromised phenotypic flexibility subgroup, which indicates that a PSN program may be effective in targeting behavior change in health-compromised target groups.

Keywords: dietary intervention; healthy lifestyle; mixed meal tolerance test; personalized nutrition; systems biology.

Figure 1

Study design overview. The screening visit, run-in period and 10-week intervention (personalized advice and meals; phase 1) of a single-arm, multi-phase study. Screening consisted of anthropometric measurements and a screening questionnaire. After screening, participants had mid- and end-point visits/contacts during the run-in and intervention period. Participants completed an at-home challenge test and sample collection (weeks 0, 10 and 20; including DNA at week 0 only), anthropometric and body composition assessments (all weeks), electronic questionnaires (all weeks except 5), coaching (weeks 10, 15 and 20), and were distributed an activity tracker (Fitbit; week 0).

Figure 2

Study flow diagram. A total of 168 participants were screened/consented and healthy men and women were enrolled in the study (n = 107). A total of 82 participants completed phase 1 (personalized advice + meals intervention period; through week 20). Of the 25 participants that did not complete phase 1, four were lost to follow-up and 21 withdrew from the study. Data from 73 participants were included in the PP analysis. Abbreviations: PP, per protocol; QOL, quality of life.

Figure 3

Data used in the health space model. Abbreviations: HDL, high-density lipoprotein; LDL, low-density lipoprotein; TG, triglycerides. Postprandial markers were measured at 30 and 120 min after challenge beverage consumption.

Figure 4

Boxplots of body weight (kg), protein intake (energy %), calorie intake (kcal), LDL cholesterol (mmol/L) and total cholesterol (mmol/L), grouped according to personalized diet type; dark grey box plots represent group A (n = 48) and light grey box plots represent group G (n = 22). Subgroup specific statistically significant differences are noted (** p < 0.01; *** p < 0.001; **** p < 0.0001) over time, except for calories where a statistical difference for the PP population is indicated.

Figure 5

Boxplots of health space scores for baseline (week 0, n = 73), end of run-in (week 10, n = 63) and end of intervention (week 20, n = 49), grouped according to age (left) and personalized diet type (right). A lower score on the health space is considered healthier.