AI/ML-Enabled Precision Nutrition Modestly Improves Anthropometric Outcomes and Systolic Blood Pressure: A Systematic Review and Meta-Analysis.

چکیده اصلی

Precision nutrition tailors dietary guidance to individual biological, genetic, metabolic, and behavioral characteristics. Artificial intelligence (AI), machine learning (ML), wearables, sensors, mobile apps, and multi-omics tools increasingly support this approach, but evidence on their effectiveness, feasibility, implementation, and equity remains fragmented. This review evaluated the effectiveness of AI/ML-enabled analytical approaches and technologies in precision nutrition across dietary, metabolic, behavioral, implementation, and equity outcomes. We conducted a PRISMA-guided systematic review and meta-analysis of studies published from January 2015 to December 2025 across six databases. Eligible studies examined AI/ML, digital tools, wearables, mobile platforms, or multi-omics approaches for personalized nutrition across the lifespan. Two reviewers independently screened studies, extracted data, and assessed bias. Random-effects meta-analyses were conducted where outcomes were comparable; remaining studies underwent qualitative and thematic synthesis. Overall, 262 studies met the inclusion criteria. Meta-analyses indicated modest pooled reductions in body weight (-1.97 kg), BMI (-0.57 kg/m2), waist circumference (-2.41 cm), and systolic blood pressure (-4.05 mmHg), whereas no significant pooled effects were observed for lipid outcomes or glycemic markers. Corresponding 95% confidence intervals and 95% prediction intervals were -3.02 to -0.92 and -6.69 to 2.75 kg for body weight, -0.99 to -0.15 and -2.43 to 1.29 kg/m2 for BMI, -4.75 to -0.07 and -11.30 to 6.49 cm for waist circumference, and -6.90 to -1.19 and -13.24 to 5.15 mmHg for systolic blood pressure. Dietary intake outcomes were generally small and inconsistent, although vegetable intake showed a modest improvement. Certainty of evidence was generally low to very low because of risk-of-bias concerns, imprecision, and substantial heterogeneity; prediction intervals for several statistically significant pooled effects crossed the null. Subgroup findings were exploratory and underpowered. Technology-enabled precision nutrition may modestly improve anthropometric outcomes and systolic blood pressure, but longer, well-reported studies are needed to clarify the clinical impact of precision nutrition technologies. SYSTEMATIC REVIEW REGISTRATION: This review was registered with PROSPERO under registration number CRD420251241063.

متن کامل اصلی