1
Department of Laboratory Medicine, Changi General Hospital, Singapore
2
Department of Medicine, Yong Loo Lin School of Medicine, National University of Singapore, Singapore
Corresponding author details:
Aw TC
Department of Laboratory Medicine
Changi General Hospital
Singapore
Copyright:
© 2020 Aw TC, et al. This is
an open-access article distributed under
the terms of the Creative Commons
Attribution 4.0 international License, which
permits unrestricted use, distribution, and
reproduction in any medium, provided the
original author and source are credited.
Lipoprotein (a) [Lp(a)] is increasingly being used to assess cardiovascular disease
(CVD) risk, with a lower Lp(a) associated with a lower CVD risk. However, many issues
in the analysis of Lp(a) targets remain. High Lp(a) levels in CVD merely describe an
association as the causative pathophysiologic mechanism of Lp(a) in CVD is still unknown.
The extreme polymorphism of Lp(a) creates many issues in the immunochemical analysis of
Lp(a), resulting in great variation between Lp(a) assays with no universal standardization.
Furthermore, the units of measurement of Lp(a) (mg/dL or nmol/L) have yet to be
standardized, and there is no appropriate conversion factor between units. Despite new
therapies that can lower Lp(a), international treatment targets and guidelines have yet to
be solidified. In addition, lower Lp(a) levels occur in diabetics, complicating the predictive
value of Lp(a) for CVD in this population. As such, further research is required before Lp(a)
can be ready for primetime.
Lipoprotein (a); Lipids; Risk prediction;Cardiovascular Disease; Diabetes
Lp(a): Lipoprotein (a)
CVD: Cardiovascular disease
SD: Standard deviation
apoB: Apolipoprotein B
apo(a): apolipoprotein (a)
DM2: Type 2 diabetes mellitus
OR: Odds ratio
Lipoprotein (a) or Lp(a) is seeing a resurgence as a cardiovascular disease (CVD)
risk biomarker [1,2]. A search of PubMed (accessed February 2020) using “Lp(a)” as the
keyword resulted in 10056 publications (Figure 1). From a peak of 2191 publications
in the late 90s interest in Lp(a) began to wane only to pick up a decade later with new
developments. Its association with cardiovascular disease (CVD) risk continues to b reaffirmed in recent studies [1]. Genetic studies underscore the link between Lp(a) and CVD
risk [2], where a one SD lower Lp(a) was associated with a 29% decreased CVD risk. The
recent 2019 National Lipid Association statement [3]recommends that Lp(a) be used to
further refine risk assessment as well as during statin treatment in patients with high CVD
risk. In this article, we will review the current literature regarding the viability of Lp(a)
assessment.
Figure 1: Number of Lp(a) publications in PubMed from 1969-
2019
Lp(a) is a complex lipoprotein particle, and its physiological function is not completely understood [4]. It is composed of a lipoprotein particle similar in protein and lipid composition to LDL with a core of cholesteryl esters and triglycerides but contains a modified apolipoprotein B (apoB) moiety such that a unique protein, apolipoprotein (a) [apo(a)] is covalently linked to the apoB by disulfide bonds. The apo(a) molecule is highly polymorphic in size and glycosylation. Apo(a) contains 10 subtypes of Kringle IV (KIV) in its structure, 1 copy of Kringle V (KV) and an inactive protease domain, resulting in >40 different isoforms of different Lp(a) sizes [5].The rate of apo(a) synthesis is inversely related to the molecular mass, and therefore individuals with lower molecular mass isoforms have higher Lp(a) levels [6]. This variation in Lp(a) structures also leads to tremendous variation among commercial Lp(a) assays, with some studies [7] showing variations between -8% to +22% among different assays. Size polymorphism of Lp(a) creates many problems in its immunochemical analysis since suitable antibodies are needed to bind to all the myriad variants of apo(a). The size variability in Lp(a) isoforms also presents problems in the search for Lp(a) lowering treatments. In a study [8] comparing the niacin-laropiprant treatment to placebo, the double monoclonal antibody-based ELISA reference method used to assess Lp(a) showed that proportional reductions in Lp(a) depended on Lp(a) isoform size. In patients with low isoform size, the mean proportional reduction in Lp(a) with treatment was only 18% compared to 31% in groups with larger isoforms.
The units for which Lp(a) is measured are not universally
standardized. Reference materials created for Lp(a) are measured
in nmol/L. However, even in recent meta-analyses [9], Lp(a)
measurements are reported in density units (mg/dL). The latest
2019 European recommendations for dyslipidemia management
[10] explicitly states that conversion between molar and mass
concentrations for Lp(a) is dependent on the Lp(a) isoform size
and concentration. Many studies assume a conversion factor for
Lp(a) of roughly 2.4 from mg/dL to nmol/L (i.e. 0.4 from nmol/L to
mg/dL) [11]. Marcovina [11] also found that the conversion factor
for converting Lp(a) values from nmol/L to mg/dL differed for 2
different assays (2.02 vs. 1.67) and was substantially lower than the
factor of 2.4 earlier suggested. Additionally, the Lp(a) molar/mass
ratios are threshold, method, and isoform dependent [12]. As such,
a simple conversion factor of 2.4 to convert Lp(a) from mg/dL to
nmol/L is not appropriate, and the two measurements should not be
used interchangeably.
Recent studies continue to have shown that elevated baseline Lp(a) above the 80th percentile of the general population is a strong risk factor for CVD independent of LDL [9,13]. In 2010, the European Society of Cardiology recommends that an Lp(a) >50mg/dL be considered a risk factor [14], based on the <80th percentile of Lp(a) (<50mg/dL) from the Copenhagen General Population Study [15]. In a 2018 meta-analysis, an Lp(a) cut point of ≥50mg/dL has also been proposed for CVD risk [9]. Canadian guidelines are less stringent, with a high Lp(a) level defined as >30mg/dL [16]. However, Lp(a) levels are also different between populations and ethnic groups [17]. Thus a universal Lp(a) target may not be applicable across different populations. An international consensus target for Lp(a) management is awaited. Previously, there were few therapies to reduce Lp(a) levels. Today PCSK9 inhibitors [18] and hepatocytedirected antisense oligonucleotides [19,20] have been proven to effectively lower Lp(a) levels in a dose-dependent manner by 6.2- 46.7% and 35-80% respectively. However, clinical practice guidelines on how to use these newer agents have yet to be developed. To date, there have been no randomized clinical trials to show that lowering Lp(a) alone results in a significant lowering of CVD risk. Furthermore, the exact pathophysiology of how Lp(a) contributes to CVD has remained elusive [21].
The role of Lp(a) in diabetes mellitus (DM) remains enigmatic.
In the recent study of 143,087 Icelanders [1], an association was
found between very low Lp(a) (<3.5nmol/L) and DM2 [odds ratio
(OR) of 1.44]. This finding was also found in a study of Lp(a) in a
Korean population [21] (OR for DM2 = 0.323 in subjects with Lp(a)
67.2±31.4mg/dL vs. 7.6±1.7mg/dL). Even in older studies [22], much
lower Lp(a) values were found in diabetic men (mean 12.3mg/dL)
and women (mean 15.1mg/dL) compared to non-diabetics (mean
Lp(a) of 16.8 and 22.0mg/dL in males and females respectively). The
meta-analysis in the Bruneck study [23] involving four prospective
cohorts also showed that the risk of DM2 was higher in the lowest two
quintiles of Lp(a) concentrations compared to the highest quintile
(mean Lp(a) of 3.3mg/dL vs. 62.9mg/dL). However, the literature on
whether a higher Lp(a) level is associated with a lower risk of DM2
is unclear. In the EPIC-Norfolk cohort [24], a 1 SD increase in log
Lp(a) was not associated with any reduction of DM2 risk (OR = 1.03),
suggesting that elevated Lp(a) levels did not help lower DM2 risk.
On the other hand, a prospective study [25] of two separate study
populations (Women’s Health Study and Copenhagen City Heart
Study) found that the inverse association between Lp(a) concentration
and DM2 persisted in both populations and was independent of
other known risk factors such as hypertension, HbA1c, body mass
index, and triglycerides. Furthermore, the incidence of prediabetes,
insulin resistance, and hyperinsulinemia increase with decreasing
Lp(a) levels [26], further hinting that Lp(a) may be involved in the
pathogenesis of prediabetes/impaired glucose tolerance. Further
research is required to elucidate the exact mechanism of how Lp(a)
may contribute to the causation of DM2. Therefore, the lower levels
of Lp(a) found in DM2 impacts the use of Lp(a) to predict CVD risk in
these subjects; lowering Lp(a) in DM2 may, in fact, be detrimental. A
study [27] that analyzed trends in glycemic control, blood pressure,
hepatic steatosis and beta-cell function in diabetics found that those
with lower Lp(a) (5-8nmol/L) had significantly worse outcomes
than those with higher Lp(a) values. On the other hand, higher Lp(a)
values (26-78mg/dL) were found to be associated with a greater risk
of retinopathy [28] and CVD (43-75mg/dL) [29] in diabetic patients.
Lp(a) shows promise as a CVD biomarker, but further research is
required to understand its pathogenetic mechanisms. Treatments are
available today that can lower Lp(a), but population-specific targets
and clinical guidelines still need to be established. We understand
more of the pathobiology of Lp(a), but its measurement has yet to
be standardized and harmonized. Thus, Lp(a) remains a work in
progress.
Copyright © 2020 Boffin Access Limited.