OBJECTIVE: Elevated lipoprotein(a) [Lp(a)] is a causal, independent risk factor for cardiovascular disease and aortic stenosis. We aimed to define the prevalence and patient thresholds of elevated Lp(a) levels in the United States. APPROACH AND RESULTS: We analyzed Lp(a) levels in 532 359 subjects from 2 data sets: (1) in 531 144 subjects from a referral laboratory and (2) in 915 patients from a tertiary referral center. Lp(a) mass levels were measured by immunoturbidometric assays in both centers and expressed as mg/dL. At the referral laboratory, the median age (interquartile range) of the subjects was 57.0 (46-67) years, and 51.9% were female. Lp(a) levels were skewed rightward as expected. The mean±SD levels were 34.0±40.0 mg/dL, and median (interquartile range) levels were 17 (7-47) mg/dL, with range 0 to 907 mg/dL. Lp(a) levels at 75%, 80%, 90%, 95%, 99%, and 99.9% percentiles were >47, >60, >90, >116, >180, and >245 mg/dL, respectively. At the referral laboratory, Lp(a) levels >30 and >50 mg/dL were present in 35.0% and 24.0% of subjects, respectively, and at the tertiary referral center, 39.5% and 29.2%, respectively. Females had higher mean (SD) (37.0 [42.7] versus 30.7 [36.7]; P<0.0001) and median (interquartile range) (19 [8-53] versus 15 [7-42]; P<0.0001) Lp(a) than males. CONCLUSIONS: This is the largest database to assess the distribution of Lp(a) and is derived from patients as opposed to general populations. Lp(a) levels >30 and >50 mg/dL were fairly common, particularly in a tertiary care setting. These data may inform consensus documents, guidelines, and therapeutic cutoffs for Lp(a)-mediated cardiovascular risk.