High Lp(a) in midlife was linked to more aortic stenosis over the next 30 years
Table of Contents
1.0 Summary
Lipoprotein(a), or Lp(a), is an inherited cholesterol-carrying particle that raises the risk of heart attack and of aortic stenosis, a stiffening and narrowing of the valve that lets blood out of the heart. Most of the evidence for the valve link comes from genetic studies and from mostly White European populations. An analysis of the US Atherosclerosis Risk in Communities (ARIC) study, published in Arteriosclerosis, Thrombosis, and Vascular Biology in September 2026, followed 14,423 Black and White adults who were 45 to 64 at the start for a median of 27 years. People with an Lp(a) of 50 mg/dL or higher had about a 30% higher rate of aortic stenosis after accounting for deaths from other causes, and their 30-year incidence was 8.6% against 7.4%. The link was absent in Black participants, even though high Lp(a) was more than twice as common among them. Cases were identified from hospital billing codes rather than heart ultrasound, and no treatment has yet been shown to prevent aortic stenosis by lowering Lp(a).
2.0 Key Points
- ARIC measured Lp(a) in 14,423 adults (mean age 54, 54% women, 25% Black) between 1987 and 1989; 3,007 (21%) had a level of 50 mg/dL or higher.
- Over a median of 27 years, 1,113 participants (8%) developed aortic stenosis and 7,310 (51%) died.
- Lp(a) of 50 mg/dL or higher carried an adjusted subdistribution hazard ratio of 1.30 (95% CI 1.13 to 1.51), rising to 1.52 (1.21 to 1.91) when the outcome was limited to stenosis-specific codes and valve replacement; 30-year incidence was 8.6% against 7.4%.
- The authors estimate that Lp(a) at or above 50 mg/dL accounted for 6.5% of aortic stenosis cases in the cohort, roughly 1 in 15.
- In Black participants there was no association (hazard ratio 0.98, 0.71 to 1.36, from 77 cases), although 38% of them had Lp(a) of 50 mg/dL or higher compared with 16% of White participants.
3.0 Evidence base
Lp(a) is an LDL particle with an extra protein, apolipoprotein(a), attached. How much of it a person carries is set mostly by inheritance and changes little over life. Genetic studies and long-term cohorts, mostly in White European populations, had already tied levels above 50 mg/dL to aortic stenosis [1]. Two questions remained: whether the link holds in a racially diverse US population, and whether it survives a proper accounting for death from other causes, which ends follow-up before valve disease has time to appear [1].
ARIC enrolled 15,792 Black and White adults aged 45 to 64 in four US communities between 1987 and 1989. This analysis, by researchers at the National Heart, Lung, and Blood Institute, used the 14,423 who had an Lp(a) measured at that first visit and complete data on other risk factors. Their mean age was 54, 54% were women, 25% were Black and 8% had diabetes. In all, 3,007 (21%) had an Lp(a) of 50 mg/dL or higher [1].
A case of aortic stenosis was counted when a hospital discharge record through 2019 carried a diagnosis code for it or for aortic valve replacement. The analysis adjusted for age, sex, race, blood pressure and its treatment, diabetes, body mass index, HDL and LDL cholesterol, triglycerides, statin use and smoking. It used Fine-Gray regression, a method that treats death from other causes as a competing outcome instead of ignoring it [1].
Over a median of 27 years, 1,113 participants (8%) developed aortic stenosis and 7,310 (51%) died. Risk rose steadily as Lp(a) rose. Each standard-deviation increase carried a subdistribution hazard ratio of 1.13 (95% CI 1.05 to 1.20). An Lp(a) of 50 mg/dL or higher, compared with lower levels, carried 1.30 (1.13 to 1.51) [1].
Thirty-year cumulative incidence was 8.6% in the higher group and 7.4% in the lower group. People with higher Lp(a) were also more likely to die of other causes during follow-up, which pulls this estimate down [1]. When the outcome was narrowed to stenosis-specific diagnosis codes and valve replacement, the hazard ratio rose to 1.52 (1.21 to 1.91). The authors estimate that Lp(a) at or above 50 mg/dL accounted for 6.5% (3.2% to 9.8%) of aortic stenosis in the cohort, roughly 1 in 15 cases [1].
The association pointed the same way in men and women. In Black participants it did not appear: the hazard ratio was 0.98 (0.71 to 1.36), based on 77 cases, even though 38% of Black participants had Lp(a) at or above 50 mg/dL compared with 16% of White participants. The statistical test for a difference by race did not reach significance (P = .08) [1]. The authors point to the small number of cases, differences in access to care and referral, reliance on billing codes, and the type of Lp(a) test, which may misclassify people near the 50 mg/dL line differently by race. They add that the null result held for diagnosis codes alone, which do not depend on referral for valve surgery [1].
4.0 Where it fits in practice
For someone with a high Lp(a), this adds a second, slower reason the number matters. Alongside heart attack and stroke risk, it marks a modestly higher chance of valve disease decades later. The absolute difference is small, 8.6% against 7.4% over 30 years, and a high Lp(a) does not mean a person will develop aortic stenosis [1].
No medicine has been shown to prevent aortic stenosis by lowering Lp(a). The authors call for Lp(a)-lowering trials designed to test aortic stenosis outcomes in diverse populations [1].
Nothing changes in testing or treatment today. This site already recommends measuring Lp(a) once in a lifetime and lists calcific aortic valve disease among its consequences, in /clinical/04-risk-stratification.html. Measurement and interpretation are covered in /clinical/06-advanced-tools.html.
5.0 Open questions
Aortic stenosis was identified from hospital billing codes, without heart ultrasound to confirm it or to rule out valve disease already present at the start. That approach misses milder disease and undercounts the true number of cases [1].
The Lp(a) test measured protein mass in mg/dL and is affected by the size of the apolipoprotein(a) protein, which varies from person to person. It also does not capture the particle’s lipid content or oxidized phospholipids, which may contribute to risk [1]. The 50 mg/dL cutoff therefore does not carry over exactly to Lp(a) results reported in nmol/L.
Why the association was missing in Black participants is unresolved. The authors suggest repeating the analysis with an Lp(a) test that is not affected by apolipoprotein(a) size, in diverse cohorts [1].
The study is observational and included only Black and White adults, so residual confounding cannot be excluded and the results may not apply to other groups [1].
This summary is provided for clinical decision support only and does not replace individualized clinical judgment. Findings are summarized from the cited primary sources; consult those sources before changing practice.
Version History
| Version | Date | Description |
|---|---|---|
| 1.0.0 | 2026-10-04 | Initial release |
References
- Zubirán R et al. Lipoprotein(a) and Incident Aortic Stenosis Over 30 Years: The ARIC Study. Arteriosclerosis, thrombosis, and vascular biology. 2026-09-24. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13614517/?tool=EBI.