DOI: 10.3390/nu18193159 ISSN: 2072-6643

Effect of a Phenotype-Tailored Lifestyle Intervention on Caloric Intake and Macronutrient Composition in Adults with Obesity: A Secondary Analysis of the PHENO-Diet Trial

Maria Antonia Espinosa, Lindsey Sefried, Jose Villamarin, Lizeth Cifuentes, Thomas Fredrick, Diego Anazco, Sima Fansa, Jamie Friend, Thomas M. Rieck, Michael Camilleri, Maria D. Hurtado Andrade, Matthew M. Clark, Donald D. Hensrud, Andres Acosta

Background/Objectives: A 12-week trial demonstrated that phenotype-tailored lifestyle intervention (PLI) was associated with greater weight loss than standard lifestyle intervention (SLI); however, differences in dietary patterns between groups were not examined. Methods: This secondary analysis evaluated caloric and macronutrient intake at 12 weeks in adults with obesity (BMI ≥ 30 kg/m2), classified into four energy-balance phenotypes (abnormal satiation, reduced postprandial satiety, low energy expenditure, emotional eating) (n = 77) who underwent PLI or SLI. Dietary intake was assessed by 24-h recalls and verified by a blinded registered dietitian. Results: Groups were compared at 12 weeks adjusting for baseline values. PLI was associated with lower total energy intake than SLI (1273 ± 182 vs. 1894 ± 176 kcal/day; p = 0.018), due to lower carbohydrate (464 ± 76 vs. 818 ± 73 kcal/day; p = 0.002) and fat intake (479 ± 76 vs. 697 ± 73 kcal/day; p = 0.036); absolute protein intake did not differ. However, protein comprised a greater proportion of energy intake in PLI (26.8 ± 1.6%) than SLI (19.8 ± 1.5%; p = 0.003), and dairy protein contributed more to total energy in PLI (6.8 ± 0.8% vs. 3.7 ± 0.7%; p = 0.004). Conclusions: PLI was associated with lower energy intake, mainly due to reduced carbohydrate and fat intake, and a higher relative contribution of protein, consistent with phenotype-based recommendations rather than macronutrient prescriptions. These findings suggest a potential role for personalized dietary approaches in obesity management. However, the non-randomized design does not allow causal inferences.