chatGPT:
Summary
This paper argues that familial longevity is partly explained by reduced inherited risk for coronary artery disease (CAD). The authors use the Leiden Longevity Study, which includes multigenerational families, and the Leiden 85-plus study. They quantify familial longevity using the Longevity Relatives Count (LRC) score, which measures the proportion of ancestors who survived into the top 10% of their birth cohort. They then test whether people with more long-lived ancestors have lower polygenic risk scores for diseases that commonly cause death in the Netherlands.
The central finding is that CAD polygenic risk is lower in descendants of long-lived families. Among 18 disease polygenic scores, CAD was the clear signal: increasing LRC score was associated with decreasing CAD-PGS. Those with at least 30% long-lived ancestors had about 0.26 standard deviations lower CAD-PGS than those with no long-lived ancestors. Other disease PGSs did not show significant associations after correction.
The authors then ask whether this lower CAD genetic burden helps explain the delayed cardiovascular disease onset seen in long-lived families. They report that each standard deviation increase in CAD-PGS was associated with increased CVD incidence, while the high-LRC group had substantially delayed CVD onset. Mediation modelling suggested that lower CAD-PGS explained about 14% of the delay in CVD onset using Cox frailty models, and about 20% using accelerated failure time models.
They also perform pathway analysis on genes contributing to the CAD-PGS. This identifies several enriched clusters, including cholesterol metabolism, platelet degranulation, focal adhesion, AGE-RAGE, Rho GTPase, RAP-RAS-MAPK, and epidermal growth factor signalling. They then construct pathway-specific PGSs, including a cholesterol metabolism PGS based on 52 SNPs in 30 genes.
The cholesterol metabolism PGS was associated with mortality after age 90 in both cohorts. In the Leiden Longevity Study nonagenarian siblings, each standard deviation increase in the cholesterol metabolism PGS increased yearly mortality risk by about 12%, corresponding to a median survival difference of about 2.5 years between lower and higher genetic-risk groups. The Leiden 85-plus replication showed a similar but smaller effect. The signal persisted, though weakened, after removing APOE-related SNPs, suggesting the cholesterol-metabolism link is not simply an APOE artefact.
Main novelty
The paper’s novelty is not simply that CAD risk is lower in long-lived people; that has been suggested before. Its more specific novelty is the multigenerational quantitative design.
First, the authors do not merely compare centenarians with controls. They use the LRC score to measure how strongly longevity clusters in a family, reducing the problem of “phenocopies”: people who live long for non-heritable reasons.
Second, they systematically choose disease PGSs based on major causes of death in the Netherlands, rather than selecting a few candidate diseases ad hoc. This gives the CAD result more weight because it emerges from a broader disease screen.
Third, they test a mediation model: lower CAD-PGS is not just associated with longevity ancestry, but appears to explain part of the delayed cardiovascular disease onset in descendants of long-lived families.
Fourth, they move from a whole CAD-PGS to a pathway-level cholesterol metabolism PGS, then test whether that pathway predicts survival in independent very-old cohorts. That is a useful attempt to convert a statistical genetic signal into a more biologically interpretable one.
Critique
The strongest part of the study is the design. The Leiden Longevity Study is unusually valuable because it contains family structure, ancestor survival information, genotypes, and prospective morbidity follow-up. The use of birth-cohort-specific survival percentiles is also sensible, because “long-lived” means something different for people born in different periods.
However, the central claim should be read carefully. The paper shows that low CAD genetic burden is one contributor to familial healthy aging, not that it is the main cause of longevity. Even the authors’ mediation estimate leaves roughly 80–86% of the delayed CVD onset unexplained by CAD-PGS. That remaining fraction could involve environment, diet, smoking, socioeconomic status, rare variants, structural variants, protective alleles, gene-environment interaction, or non-CAD biology.
A second limitation is that the study is heavily European/Dutch in structure. Polygenic scores are ancestry-dependent, and effects derived from European GWASs often transfer poorly across ancestries. So the result is probably most directly applicable to populations genetically and environmentally similar to the Dutch cohorts studied.
Third, the mediation analysis is useful but not definitive causal proof. CAD-PGS is fixed at conception, which strengthens causal inference relative to many biomarkers, but mediation through observed CVD onset still depends on modelling assumptions, ascertainment of disease, competing mortality, and environmental confounding of the LRC–CVD relationship.
Fourth, the disease screen may under-detect non-CAD biology. The authors use available GWASs for common fatal diseases, but PGS quality varies substantially by disease. CAD GWASs are large and powerful, while some other disease PGSs may be weaker. Therefore, the finding that only CAD emerges may partly reflect better statistical power for CAD genetics rather than exclusive biological importance.
Fifth, the cholesterol metabolism pathway result is interesting but needs caution. Pathway-PGS construction from CAD-associated loci is not the same as measuring cholesterol metabolism mechanistically. The ChoMet-PGS is still a genetic risk instrument derived from CAD loci, and “cholesterol metabolism” may capture lipoprotein biology, vascular biology, APOE-adjacent effects, or other linked mechanisms. The APOE-removal sensitivity analysis helps, but it does not fully prove the pathway mechanism.
Overall assessment
This is a strong and well-motivated paper. Its best-supported conclusion is:
Families enriched for exceptional longevity carry lower common-variant genetic risk for CAD, and this lower CAD risk explains a meaningful but minority fraction of their delayed cardiovascular disease onset.
The more speculative but interesting conclusion is:
Cholesterol/lipoprotein biology may be one of the key genetically mediated pathways linking cardiovascular resilience to survival beyond age 90.
It is a good contribution to the “longevity as disease-risk avoidance” model, but it does not rule out active protective mechanisms, non-CAD mechanisms, environmental inheritance, mitochondrial/metabolic resilience, inflammation control, or rare protective variants.