Lipoproteins and calcific aortic valve disease: Hardening evidence?

Calcific aortic valve disease has a prevalence of >25% in those ≥65 years of age,¹ and may progress to aortic stenosis, in which calcification and fibrosis increase leaflet rigidity to an extent that outflow of blood from the left ventricle is obstructed. Severe aortic stenosis has a surprisingly high prevalence of ~4% in those ≥65 years of age.^2,3 In the pre-surgical valve replacement era, symptomatic aortic stenosis was a highly mortal disease with a median survival ~2 years.⁴ Fortunately, both surgical aortic valve replacement and more recently transcatheter aortic valve implantation have conferred dramatic survival benefit for those with severe aortic stenosis.^5,6 Nonetheless, despite their efficacy, valve replacement by either means is neither prevention nor cure.

For many years, calcific aortic valve disease was considered to be a degenerative process.⁷ However, in the 1990s and early 2000s, histologic and immunohistochemical studies of human aortic valve lesions suggested strongly that the disease is an active process, by demonstrating co-localization of chronic inflammatory cells and lipids with calcium;⁸ deposition of apolipoproteins (apos) of “atherogenic” lipoprotein particles, including apo (a), B and E;^9,10 and active expression of calcification mediators.^11,12 Later, apo A-I - the predominant protein on “anti-atherogenic” HDL particles - was found to be present at lower concentrations in stenotic, as opposed to control, human aortic valves.¹³

Subsequently, the opportunity to identify potential risk factors for aortic valve calcification in large, epidemiologic cohorts was created by development and validation of electron beam and multidetector computed tomography scanning as a tool for quantifying aortic valve calcification.^14,15 Findings from these studies further supported the concept that calcific aortic valve disease is an active and potentially modifiable process by demonstrating, for example, associations of traditional “atherogenic” risk factors (e.g., diabetes and metabolic syndrome,^16,17 hypertension¹⁸ and lower HDL levels¹⁷) with increased prevalence^16,18 and incidence¹⁷ of aortic valve calcification. Mendelian randomization studies also provided strong evidence of roles for both Lp(a)¹⁹ and LDL²⁰ in risk not just for aortic valve calcification but also for its late, obstructive stage, i.e., incident aortic stenosis.

In this issue of ATVB, Bortnick et al.²¹ report the results of their comprehensive analyses of the potential relationships of various measures of “anti-atherogenic” HDL with risks for incident and progressive aortic valve calcification in the Multi-Ethnic Study of Atherosclerosis (MESA). Importantly, they find that several, potentially “anti-atherogenic” markers, including increases in HDL cholesterol (HDL-C), HDL particle numbers (HDL-P), large HDL-P, as well as HDL-C without apo C3, are associated with decreased risk for aortic valve calcification incidence and progression. The study has many strengths, including the large number of participants with serial computed tomography scans (N=6,814 at baseline, N=5,886 at one of two interim timepoints, and N=3,304 at the final timepoint), with a long median follow-up of 8.9 years, and state-of-the-art statistical methodology.

The implications of these findings are a bit less clear.

First, as the authors readily acknowledge, their many analyses were not adjusted for multiple comparisons. Thus, there is risk for an inflated false-positive rate. Moreover, in a model including both HDL-C and HDL-P, only HDL-C retained a statistically significant association with lower aortic valve calcification risk. This latter finding is consistent with an earlier study, albeit in the same cohort and with shorter follow-up, that lower HDL-C levels were associated with higher aortic valve calcification prevalence.¹⁷ At a minimum, the findings of Bortnick et al.²¹ give us reasonable confidence that higher HDL-C is associated with lower aortic valve calcification incidence and progression.

Second, while their finding that higher HDL-C concentration is associated with lower aortic valve calcification incidence and progression²¹ may reflect a protective effect of HDL-C, e.g., through enhanced cholesterol efflux, there are other potential explanations. For example, it may instead be that those with lower HDL-C - and, by implication, higher risk for incident and progressive aortic valve calcification - have a greater preponderance of “pro-atherogenic” factors, such as insulin resistance. In addition, smaller HDL particles of patients with coronary artery disease²² are enriched in HDL proteins including apo A-IV and complement factors C3, C4B and C9, each of which recently has been localized to calcifying microdomains in human aortic stenosis.²³

Other HDL-associated proteins might also promote calcific aortic valve disease, specifically the amyloidogenic proteins transthyretin and serum amyloid A. A recent study identified that 16% of aortic stenosis patients undergoing transcatheter aortic valve implantation have transthyretin-associated cardiac amyloidosis.²⁴ Transthyretin has also been shown to bind to apo A-I and cleave its C-terminal portion, thereby both inhibiting the ability of apo A-I to promote cholesterol efflux and increasing its propensity to form amyloid fibrils.²⁵ In the present study, Bortnick, et al.²¹ were unable to identify an association of an HDL particle-associated marker, GlycA, with aortic valve calcification incidence or progression. However, serum amyloid A is not only an acute phase reactant,²⁶ but it also is carried on HDL,²² is fibrillogenic, and has been shown to mediate HDL particle binding to aortic sinus extracellular matrix in mice.²⁷ Thus, while highly speculative, it is conceivable that HDL-associated serum amyloid A “traps” HDL in extracellular matrix and promotes amyloid fibril formation in human aortic valves.

Finally, what are the therapeutic options if, based on the association of higher HDL-C with lower AVC incidence and progression, HDL-C raising is contemplated as a potential therapy? Based on the results of HDL-C-raising trials in atherosclerosis, this strategy is not promising. While cholesteryl ester transfer protein inhibitors raise HDL-C levels, they have largely been abandoned as a therapy for atherosclerosis.^28,29 Similarly, while niacin also effectively raises HDL-C, its addition to statin therapy showed no benefit over statins alone in the AIM-HIGH Trial.³⁰ Moreover, niacin recently has been shown to increase pro-atherogenic proteins carried on HDL particles.³¹

In summary, the authors of the current report are to be applauded for their careful and comprehensive analyses as well as their convincing demonstration that study of HDL in calcific aortic valve disease merits real attention. Their findings do us all a service by raising many intriguing questions worthy of further study, especially since we still lack any proven medical therapy for calcific aortic valve disease. Hopefully, answers to those questions may someday lead us to that medical “holy grail” of a definitive cure or preventive treatment.