IMPORTANCE: Elevated lipoprotein(a) (Lp[a]) levels are associated with atherosclerotic cardiovascular diseases. The association between high Lp(a) levels and human longevity phenotypes is, however, controversial. OBJECTIVE: To examine whether genetically determined Lp(a) levels are associated with parental life span and chronic disease-free survival (health span) and the association between Lp(a) levels and long-term, all-cause mortality risk. DESIGN, SETTING, AND PARTICIPANTS: In this genetic association study, cross-sectional mendelian randomization (UK Biobank [2006-2010] and LifeGen Consortium) and prospective analyses (European Prospective Investigation Into Cancer and Nutrition (EPIC)-Norfolk [1993-1997, with patients followed up to 2016]) were conducted using individual-level data on 139 362 participants. The association between a weighted genetic risk score of 26 independent single-nucleotide polymorphisms at the LPA locus on parental life span using individual participant data from the UK Biobank, as well as with summary statistics of a genome-wide association study of more than 1 million life spans (UK Biobank and LifeGen), were examined. The association between these single-nucleotide polymorphisms and the age at the end of the health span was tested using summary statistics of a previous genome-wide association study in the UK Biobank. The association between Lp(a) levels and all-cause mortality in the EPIC-Norfolk study was also investigated. Data were analyzed from December 2018 to December 2019. EXPOSURES: Genetically determined and measured Lp(a) levels. MAIN OUTCOMES AND MEASURES: Parental life span, health span, and all-cause mortality. RESULTS: In 139 362 white British participants (mean [SD] age, 62.8 [3.9] years; 52% women) from the UK Biobank, increases in the genetic risk score (weighted for a 50-mg/dL increase in Lp[a] levels) were inversely associated with a high parental life span (odds ratio, 0.92; 95% CI, 0.89-0.94; P = 2.7 × 10-8). Using the Egger-mendelian randomization method, a negative association between LPA single-nucleotide polymorphisms and parental life span (mean [SD] Egger-mendelian randomization slope, -0.0019 [0.0002]; P = 2.22 × 10-18) and health span (-0.0019 [0.0003]; P = 3.00 × 10-13) was noted. In 18 720 participants from EPIC-Norfolk (5686 cases), the mortality risk for those with Lp(a) levels equal to or above the 95th percentile was equivalent to being 1.5 years older in chronologic age (β coefficient [SE], 0.194 [0.064]). CONCLUSIONS AND RELEVANCE: The results of this study suggest a potential causal effect of absolute Lp(a) levels on human longevity as defined by parental life span, health span, and all-cause mortality. The results also provide a rationale for trials of Lp(a)-lowering therapy in individuals with high Lp(a) levels.