Lipoprotein(a) and aortic valve stenosis: work in progress  Free

Graphical Abstract

Schematic illustration of the observational and genetic evidence that LDL-cholesterol (LDL-C) and lipoprotein(a) [Lp(a)] might be involved in the pathogenesis of aortic valve stenosis and/or aortic valve calcification. Recent trials with statins had no effect on the aortic valve outcomes. Questions concerning trials to be planned with Lp(a)-lowering drugs are listed. OxPL, oxidized phospholipids.

Open in new tabDownload slide

This editorial refers to ‘Lipoprotein(a) is associated with the onset but not the progression of aortic valve calcification’, by Y. Kaiser et al., https://doi.org/10.1093/eurheartj/ehac377.

The first studies indicating that lipoprotein(a) [Lp(a)] might be associated with aortic valve stenosis were published in the 1990s (reviewed in Arsenault and Kamstrup¹ and Koschinsky and Kronenberg²). However, major attention has been drawn to this association by a genome-wide association study in 2013 published by Thanassoulis and colleagues which ushered in a new era: the authors observed that genetic variation in the LPA locus, that goes hand in hand with high Lp(a) concentrations, is associated with aortic valve calcification across multiple ethnic groups and with incident clinical aortic stenosis.³ This association between Lp(a) and aortic valve stenosis is even stronger than with any other cardiovascular outcome.¹ Numerous studies followed in the subsequent years. For example, data from Copenhagen of >77 000 individuals including 454 cases of incident calcific aortic valve stenosis revealed that individuals with Lp(a) values between the 90th and 95th percentile (i.e. 65–90 mg/dL) had a two-fold risk of developing an incident stenosis, and those with Lp(a) above the 95th percentile (>90 mg/dL) had even a 2.9-fold increased risk compared with those with Lp(a) below the 22nd percentile.⁴ Genetic data in this study as well as in the EPIC Norfolk study supported the causality between high Lp(a) concentrations and aortic valve stenosis.^4,5 That means that genetic variants which are associated with a lifelong exposure to high Lp(a) concentrations can be found significantly more often in patients with aortic valve stenosis; this type of study is called a Mendelian randomization study.⁶ A further investigation with six case–control studies including 9459 cases of calcific aortic valve stenosis and 428 722 controls observed that a weighted genetic risk score associated with high Lp(a) concentrations was linked to calcific aortic valve stenosis independently of the coronary artery disease status.⁷ All these data together support a causal effect of high Lp(a) concentrations on the prevalence and incidence of aortic valve stenosis.

So far, so good. However, does this mean that Lp(a) should be a therapeutic target to fight against this disease, the prevalence of which is predicted to increase by ∼300% by 2050? This is even more important since it is the last major cardiovascular condition without any medical therapy capable of slowing disease progression, beyond surgical or transcatheter valve replacement. Some might fear that they are experiencing déjà-vu as observed with LDL-cholesterol. Observational studies described an association between LDL-cholesterol and incident aortic stenosis, and a weighted genetic risk score as a measure of the genetic predisposition to high LDL-cholesterol provided support for a causal association with aortic stenosis.⁸ However, three randomized controlled trials using statins did not detect a significant benefit of lowering LDL-cholesterol on the progression of aortic stenosis. Many reasons for these negative findings were discussed such as a too short or a not sufficiently pronounced lowering of LDL-cholesterol or off-target effects of statins such as pro-osteogenic properties of the statins used in these trials.⁹ It has furthermore been speculated that high LDL-cholesterol might trigger the initiation of the disease but then has no major influence on the progression of the disease. This has been shown in the MESA study observing that traditional risk factors are associated with onset of aortic valvular calcification but not progression during a relatively short observation period of 2.4 years.¹⁰

What is the evidence we have for Lp(a) as a risk factor for aortic valve stenosis? The epidemiological and genetic evidence is quite strong that high Lp(a) is associated with the development of aortic valve stenosis. Lp(a) is an important carrier of oxidized phospholipids which are considered as an key culprit for the development of aortic valve stenosis. Earlier imaging studies reported that increased Lp(a) and oxidized phospholipids in elderly patients with advanced aortic valve stenosis are associated with increased valvular calcification activity and confirmed faster rates of disease progression using both computed tomography (CT) calcium scoring and echocardiography. This translated into an increased incidence of aortic valve replacement and death.¹¹ Furthermore, Capoulade and colleagues showed that Lp(a) concentrations and oxidized phospholipids are independently associated with faster progression of aortic stenosis.¹²

In the current issue of the European Heart Journal, Kaiser and colleagues¹³ changed this picture and assessed whether high Lp(a) levels are associated with incidence and progression of aortic valve calcification in a different setting. They had access to a subgroup of 922 individuals from the long-term population-based Rotterdam Study who had undergone non-enhanced cardiac CT imaging at baseline and after a median follow-up of 14 years. They made three important observations. (i) Each 50 mg/dL higher Lp(a) concentration increased the risk for a baseline aortic valve calcification by 43%. (ii) They defined a new-onset aortic valve calcification when a subject had an Agatston score >0 on the follow-up scan in the absence of aortic valve calcification on the first scan; almost 60% (n = 415) of the 702 individuals without aortic valve calcification at baseline showed calcifications at the follow-up scan. Each 50 mg/dL higher Lp(a) concentration increased the risk for a new-onset calcification by 30%. (iii) Progression of aortic valve calcification was observed in 220 individuals and was defined as the absolute difference in Agatston units between the baseline and follow-up scan when in both scans signs of calcifications were observed. The Agatston score in these individuals increased from baseline 52 to 242 units at the end of the 14 years of follow-up. Most importantly, there was no association of Lp(a) with progression of the disease. The authors concluded that Lp(a) lowering might be most effective in pre-calcific stages of aortic valve disease. As in an earlier study,¹⁴ the authors assumed that high Lp(a) and other traditional risk factors might be involved in the initiation phase of aortic valve stenosis. Following endothelial damage, Lp(a) as well as oxidized phospholipids and autotaxin carried by Lp(a) might enter the valvular tissue, resulting not only in inflammation but also in an osteogenic transformation of valvular interstitial cells with first calcium deposits.¹⁵ During the propagation phase, increasing calcium deposition in the valve with further damage might be independent from the initiating risk factors [including Lp(a)] and accelerates disease progression as observed by Kaiser et al.¹³ These findings are not necessarily in line with those of Capoulade et al.⁷ or those of Zheng et al.¹¹ for reasons that are not clear. These latter two studies investigated patients who already had aortic stenosis at baseline and both reported in the top tertile of Lp(a) a higher rate of progression of aortic stenosis, calcification, and aortic valve replacement therapy during a shorter follow-up period of ∼1–5 years (depending on the endpoint). In the present study by Kaiser et al.,¹³ the subgroup of the 220 patients who already had aortic valve calcifications at baseline are derived from a population-based study with a very long follow-up of 14 years. This study excluded participants who underwent valvular replacement therapy. However, a sensitivity analysis including these participants did not change the results. Furthermore, 1573 of the 2524 originally investigated study participants died or were lost to follow-up, and only 922 were finally available for the analysis. It is not clear whether a higher proportion of patients with calcifications at baseline and high Lp(a) concentration might have developed a progression of the disease and some of them might even have died due to the consequences of aortic valve stenosis during this long observation period of 14 years and the relatively high average age of 68 years at baseline.

The observed findings in the present population-based study by Kaiser and colleagues¹³ might have an impact on the design of a future intervention study, raising several questions (Graphical Abstract). Which patients should be selected for inclusion in the study? If the observations hold true, the inclusion of participants who already have signs of calcification might already be in a too advanced stage to reverse the disease. However, those are the patients with the most need for an intervention. If these patients are excluded, one might have to include patients with very high Lp(a) concentrations but no signs of initiation of aortic valve calcification. This will require a large sample size and follow-up duration to observe the required number of events. On the other hand, an intervention study will include only individuals with Lp(a) concentrations of, for example, above the 90th or 95th percentile. The subgroup in which progression has been investigated by Kaiser et al.¹³ probably included only ∼20 individuals with these high concentrations. An extrapolation from these data for future trials in which only individuals with these high concentrations will be included therefore must be considered with caution. A further decision will be how to define progression of aortic valve disease: will it be aortic valve replacement therapy or the development of aortic valve calcification? From the study by Kaiser et al., we see that a major fraction of ∼60% developed this latter endpoint in 14 years of observation and started at an age of ∼66 years.¹³ On the one hand, a lower frequency can be expected in a typical trial duration. On the other hand, patients will be selected on very high Lp(a) concentrations and therefore a larger number of events can be expected.

In any case, these findings, together with earlier studies, are an important basis for the planning of interventional studies which might help to avoid a similar experience to that seen in the LDL-cholesterol-lowering trials that targeted the endpoint aortic valve stenosis (Graphical Abstract). Selecting patients without first aortic valve calcifications, deciding on high Lp(a) concentrations (e.g. above the 90th percentile) as an inclusion criterion, a sufficient long observation period, and probably ‘intermediate’ endpoints on the way to valvular replacement therapy will be important key points for the planning of the study. However, having more data on this topic would make planning of intervention studies easier. What we know with 100% certainty is that we definitely need drugs to keep this common disease under control.

Data Availability

No new data were generated or analysed in support of this research.

Funding

None.

References

Author notes