Effects of polygenic risk score communication on health outcomes: systematic review and meta-analysis

Abstract Objective The purpose of this systematic review and meta-analysis is to summarize evidence from all RCTs to-date on the efficacy of polygenic risk score (PRS) communication in changing health outcomes. Design Systematic review and meta-analysis. Data sources Cochrane Central Register of Controlled Trials (CENTRAL) and Pubmed, from inception to March 2025. Study selection Randomized controlled trials comparing disclosure versus non-disclosure of PRS results. Intervention included comunication of risk information from ≥ 2 single nucleotide polymorphism (SNPs). Methods Pairs of reviewers conducted screening, extracted data and assessed risk of bias. Meta-analyses were conducted using inverse variance-weighting with fixed and random effects models. Otto-SR was used to verify screening and data extraction. Risk of bias was assessed with the Cochrane Risk of Bias 2 (RoB-2) tool. All measurable health-related outcomes were considered eligible. Results Of 7,830 articles retrieved, 27 RCTs were eligible. PRSs mainly predicted risks for cancer (n=9 RCTs), cardiovascular diseases (n=8), and diabetes (n=6). 21 RCTs targeted primarily healthy populations, 3 at-risk populations and 3 individuals who had already developed a disease and PRS predicted complications. 15/26 RCTs concluded in their abstracts with favorable claims about the PRS, with only 5/15 justifying it with any statistically significant results. 9/26 RCTs had high risk of bias. Meta-analysis revealed no statistically significant effects on any measured outcome, among 22 outcomes tested in 2 or more trials. Standardized mean differences (SMDs) (95% CI) for dietary outcomes were −0.11 (−0.23; 0.01) for daily energy intake, 0.08 (−0.15; 0.31) for daily fat intake and −0.11 (−0.28; 0.06) for alcohol consumption. For physical activity, SMD was −0.01 (−0.13; 0.11). Relative risks were 1.12 (0.77; 1.61) for screening attendance, 1.50 (0.98; 2.29) for statin use, and 0.95 (0.32; 2.79) for disease incidence. For psychological outcomes, SMDs were −0.02 (−0.13; 0.08) for anxiety, −0.06 (−0.23; 0.10) for worry, −0.10 (−0.40-0.19) for perceived risk, and −0.05 (−0.23; 0.13) for depression. For clinical outcomes, mean differences were −2.01 (−8.27; 4.26) for total cholesterol, −3.64 (−7.88; 0.60) for LDL cholesterol, −0.21 (−2.65; 2.23) for HDL cholesterol, −1.88 (−4.17;0.42) for diastolic blood pressure, −1.26 (−4.44; 1.92) for systolic blood pressure, −0.12 (−0.64; 0.39) for BMI and −0.33 (−0.87; 0.20) for weight. 69 outcomes had been reported in only a single trial (18/69 primary ones), and of those 2 had statistically significant results at p<0.05 as primary outcomes and 3 as secondary outcomes. Conclusions Overall, despite frequent promising claims, the disclosure of PRS typically did not lead to meaningful changes in behavioral, psychological or clinical outcomes. Systematic review registration OpenScience Framework: doi.org/10.17605/OSF.IO/28V6J Summary box Section 1: What is already known on this topic - Polygenic risk scores (PRSs) represents one of the most promising approaches of personalized medicine where communication of genetic information is expected to improve behavior and health outcomes. - Previous reviews have evaluated the communication of genetic risk but none have evaluated yet the effect of specifically PRS. - Since PRSs implementation in clinical practice is contemplated, rigorous evaluation of their impact is valuable for further implementation. Section 2: What this study adds - PRS communication effects were close to null for a large set of outcomes, demonstrating no improvement in preventive behaviours, including screening adherence or clinical measures. - High heterogeneity in RCTs in the field is present with small sample sizes, short follow-up periods, and many self-reported outcomes. - A gap is evident between the theoretical promise of PRS-guided prevention and its lack of documented real-world effectiveness.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.