We conducted a cross-sectional study that included all patients admitted with ACS to the Cardiology Department of a tertiary referral hospital (Unidade Local de Saúde Gaia/Espinho) from January 2022 to December 2023. Lp(a) levels were assessed upon admission, and patients were divided into two groups: those with elevated Lp(a) (>100 nmol/L) and those with lower Lp(a) (≤100 nmol/L) levels. Beyond this dichotomous classification, we examined the continuous relationship between Lp(a) concentration and outcomes. For clinical clarity, we set a reference threshold of 100 nmol/L – a value increasingly acknowledged in cardiology literature as clinically significant – allowing us to compare patients with elevated levels to those with concentrations at or below this marker.11

Since January 2022, Lp(a) measurements have been opportunistically included in the standard analytical workup for all ACS patients admitted to our hospital. The decision to request Lp(a) testing was made by the attending cardiologists during the initial assessment upon hospital admission. This assessment was performed alongside routine blood tests, including complete blood count, renal function, liver enzymes, lipid profile, glycated hemoglobin (HbA1c), and thyroid function.

Lp(a) levels were measured at our center's laboratory using a particle-enhanced turbidimetric immunoassay, conducted on the Roche Cobas® Integra Z701 analyzer, in accordance with the manufacturer's instructions. The results are expressed in nanomoles per litre (nmol/L).

For the primary analysis, an Lp(a) cut-off of 100 nmol/L was used to define elevated Lp(a) levels, in accordance with the recommendations of the National Lipid Association (NLA) guidelines.11 Secondary analyses explored additional thresholds of 75 nmol/L and 175 nmol/L to assess the sensitivity of findings across clinically relevant cut-off points: 75 nmol/L representing a more conservative threshold suggested for primary prevention populations, and 175 nmol/L corresponding to the eligibility criterion for emerging Lp(a)-lowering therapies in clinical trials.15

Patient demographic and clinical data were extracted from hospital records. Diabetes mellitus, arterial hypertension, obesity, smoking, and dyslipidemia (defined as LDL-cholesterol levels ≥130 mg/dL) were considered traditional cardiovascular risk factors. Premature CAD was defined as occurring in men aged ≤55 years and women aged ≤60 years.16

Categorical variables were described as counts and percentages. Continuous variables were described as median and interquartile range (IQR). Valid percentages were reported, and missing data were not imputed.

Lp(a) group comparisons were made using the Chi-squared test for categorical variables and Independent T or Mann–Whitney U tests for continuous variables.

Spearman's correlation coefficients were used to evaluate the relationships between Lp(a) levels and other lipid parameters.

We used R software (version 4.1.2) for all statistical analyses; this meant the data were analyzed using both numerical and graphical methods. Additionally, a p-trend was computed to evaluate the associations between continuous Lp(a) values and independent variables using linear regression analysis. This approach allowed us to assess the direction and significance of potential dose-response relationships rather than simply comparing categorical groups, providing insight into whether higher Lp(a) levels were associated with linear changes in the clinical variables of interest. Statistical significance was set at p<0.05.

This study was approved by the Ethics Committee of Unidade Local de Saúde Gaia e Espinho (document number 134/2024).

Among the patients included, 238 (61.3%) had lower Lp(a) levels (≤100 nmol/L), while 150 (38.7%) had elevated Lp(a) levels (>100 nmol/L) (Table 1). In the secondary analysis, 209 patients (53.9%) had Lp(a) levels at or below 75 nmol/L, whereas 179 patients (46.1%) had levels above 75 nmol/L. Furthermore, 306 patients (78.9%) had Lp(a) levels below 175 nmol/L, whereas 82 patients (21.1%) had levels above 175 nmol/L (Table 2).

In the global sample, the median Lp(a) was 62.0 (18.4, 153.8) nmol/L, median total cholesterol was 164.0 (141.0, 196.0) mg/dL, median LDL-C was 100.0 (74.3, 125.0) mg/dL, median HDL-C was 41.0 (35.0, 49.0) mg/dL and median triglyceride level was 112.0 (80.0, 147.0) mg/dL. There were no significant differences in the lipid profiles between the elevated Lp(a) and lower Lp(a) groups (Table 1).

No correlation was found between Lp(a) and LDL-C (p=0.217), HDL-C (p=0.291), total cholesterol (p=0.272), or triglyceride (p=0.905) levels (Table 3 and Figure 1).

Figure 1.

Correlation between lipoprotein(a) and other lipid parameters. Panel A, Correlation between Lp(a) and LDL-C levels; Panel B, Correlation between Lp(a) and HDL-levels; Panel C, Correlation between Lp(a) and total cholesterol levels; Panel D, Correlation between Lp(a) and triglyceride levels.

The median age of the patients was 63 years (IQR, 55–73). The age distribution did not differ significantly between the lower and elevated Lp(a) groups (p=0.4) (Table 1).

A total of 25.5% of patients had premature CAD. There were no significant differences in Lp(a) levels between patients with non-premature CAD and those with premature CAD (Table 1).

Regarding the presence of traditional cardiovascular risk factors, 53.4% of patients had arterial hypertension, 49.7% had dyslipidemia, 25.5% had diabetes, 25% were obese, 32.5% were active smokers and 20.1% were previous smokers. There were no differences in the prevalence of traditional cardiovascular risk factors between the elevated Lp(a) and lower Lp(a) groups (Table 1).

Overall, when considering the sum of traditional cardiovascular risk factors, 10.6% did not display traditional cardiovascular risk factors, 23.2% had one traditional cardiovascular risk factor, 28.9% had two traditional cardiovascular risk factors, 23.7% had three traditional cardiovascular risk factors, 10.6% had four traditional cardiovascular risk factors and 2.1% had five traditional cardiovascular risk factors (Supplementary Table 2).

Lp(a) values remained consistent across the different traditional cardiovascular risk factor categories, showing no differences (p=0.7) (Figure 2A and Supplementary Table 2).

Figure 2.

(A) Lipoprotein(a) values according to the number of traditional cardiovascular risk factors. (B) Lipoprotein(a) values according to the number of traditional cardiovascular risk factors and sex.

This cohort comprised 19.6% women and 80.4% men. Women tended to have higher Lp(a) levels [47% of female patients with elevated Lp(a) vs. 37% of male patients with elevated Lp(a)] (p-trend=0.014) (Table 1).

When applying the 75 nmol/L cut-off, men represented 84% of patients with Lp(a) levels at or below 75 nmol/L and 77% of those with levels above 75 nmol/L, whereas women accounted for 16% and 23%, respectively. The difference in sex distribution between these two groups was not statistically significant. Considering the 175 nmol/L cut-off, the proportion of men decreased from 83% among patients with Lp(a) levels below 175 nmol/L to 72% among those with levels above this threshold, while the proportion of women increased from 17% to 28%, showing a statistically significant difference (p=0.030) (Table 2).

Women showed higher Lp(a) levels than men in the group of patients with no traditional cardiovascular risk factors (p=0.003), but no significant sex-related differences were noted in patients with one or more traditional cardiovascular risk factors (Figure 2B).

Previously diagnosed CAD (20.4%) was the most common comorbidity, followed by a family history of CVD (8.8%) and peripheral artery disease (7%) (Table 1).

Patients with a history of CAD had higher Lp(a) values (p-trend=0.003) (Table 1), particularly when using the 175 nmol/L threshold, where a difference was observed between groups (p=0.010) (Table 2).

Patients with a history of atrial fibrillation (AF) or flutter showed a trend toward lower Lp(a) levels (p-trend=0.028) (Table 1). This difference reached significance when patients were stratified using the 75 nmol/L (p=0.016) and 175 nmol/L (p=0.029) thresholds (Table 2).

Patients who had been previously treated with high-intensity statins, either alone or in combination with ezetimibe, showed a tendency toward higher Lp(a) levels (p-trend=0.032 and p-trend=0.014, respectively) (Table 1). When patients were stratified according to the 175 nmol/L Lp(a) cut-off, there were also a number of patients in the higher Lp(a) group treated with high-intensity statins (p=0.045) and ezetimibe (p=0.003) (Table 2).

First, our findings showed that Lp(a) levels were heterogeneously distributed within the population, with no clear associations with most individual characteristics or traditional cardiovascular risk factors studied. These results emphasize the importance of universal screening, as recommended by the EAS and the European Society of Cardiology (ESC), to prevent underestimation of CV risk.10,16 Second, we found that more than one-third of the ACS patients had Lp(a) levels greater than 100 nmol/L, representing a relatively high proportion of individuals at increased CV risk, which is consistent with previous literature.11 Notably, the median Lp(a) level in our study (62.0 nmol/L, IQR 18.4–153.8) appears higher than in the general population (19.6 nmol/L, IQR 7.6–74.8) and in other studies, such as Nissen et al., who reported a median of 42.0 nmol/L (IQR 15.0–155.4) in a large multicentre epidemiological cohort of ASCVD patients from 48 countries.17,18 Similarly, Altmann et al. found median Lp(a) levels of 37.05 nmol/L (Min–Max 0.70–886.00) and 28.70 nmol/L (Min–Max 1.60–626.00) in patients with myocardial infarction (MI) aged <60 years and ≥60 years, respectively.19 Additionally, a recent Portuguese study by Saraiva et al. reported exceptionally high Lp(a) values (median 100.50 nmol/L, IQR 19.30–224.70) in 255 ASCVD patients, with a median age of 57.07 years, who were predominantly men (78.4%).20 These comparisons suggest a potentially higher prevalence of elevated Lp(a) in our cohort, warranting further investigation to better characterize Lp(a) levels in the Portuguese population.

This study also provides important information for ongoing phase III trials of Lp(a)-lowering therapies. We observed that 21.1% of our patients had Lp(a) levels >175 nmol/L, a threshold that may qualify them for Lp(a)-lowering therapies, such as pelacarsen, which is currently being evaluated in the Lp(a) HORIZON trial.21 This highlights a potential opportunity for further CV risk reduction through targeted interventions.

The Lp(a) measurement method remains controversial, primarily due to the structural polymorphisms of Apo(a) which result in isoforms of varying sizes.22 The preferred method for Lp(a) quantification is immunoassay using polyclonal antibodies targeting the highly polymorphic kringle IV type 2 domain of Apo(a) component of Lp(a) particles. Given the significant heterogeneity in Apo(a) isoforms, nmol/L is the recommended unit of measurement over mg/dL to avoid over- or underestimating Lp(a) levels.10,11 A key strength of our study is the adoption of this standardized method, enhancing the reliability and comparability of our findings.

Women exhibited higher Lp(a) levels than men, particularly in those who did not exhibit traditional cardiovascular risk factors. This difference may be influenced by age, as previous studies have suggested postmenopausal increases in Lp(a) levels. For instance, Simony et al., using data from the Copenhagen General Population Study, reported that women aged 50–89 years had mean Lp(a) levels 17% higher than men, with a 27% increase post-menopause and a 12% reduction during hormone replacement therapy.23 Given the established link between high Lp(a) levels and premature ASCVD in women, our findings reinforce the need to measure Lp(a) in female patients, particularly those without traditional cardiovascular risk factors.

Shiyovich et al. previously demonstrated that among patients <50 years presenting with a first MI and no prior ASCVD, high Lp(a) was associated with an increased risk of ACS, independent of traditional cardiovascular risk factors burden.24 Notably, in our study we did not observe significant differences in lp(a) levels across groups with increasing numbers of traditional cardiovascular risk factors, suggesting that elevated Lp(a) levels may be found in ACS patients regardless of traditional risk factors.

Another key finding was that only 11% of patients with Lp(a) greater than 100 nmol/L had a documented family history of CVD, suggesting that Lp(a) screening should not solely rely on family history. Moreover, only 24% of patients with high Lp(a) levels had previously diagnosed CAD, reinforcing the need for Lp(a) screening in asymptomatic individuals. Previous research has shown that most young patients with elevated Lp(a) had ASCVD risk scores inferior to 7.5% before experiencing their first MI, further questioning the reliance on standard risk calculators when deciding on Lp(a) measurement.25

An unexpected finding was that patients with a history of AF or atrial flutter had lower Lp(a) levels. The relationship between Lp(a) and AF remains unclear. Some authors hypothesize that Lp(a) infiltration into atrial tissue may induce gene expression changes, leading to microcalcifications, cellular damage, and electrical remodeling.26

Given the established causal link between elevated Lp(a) and calcific aortic valve disease, particularly aortic stenosis, future studies should also explore the potential correlation between Lp(a) levels and echocardiographic findings of aortic sclerosis or stenosis in ACS populations.

Patients who were previously treated with high-intensity statins and ezetimibe tended to have higher Lp(a) levels. A systematic review and meta-analysis of randomized statin trials found that statin therapy did not significantly increase Lp(a) levels nor modify the associated CVD risk.27 Similarly, ezetimibe, whether used alone or in combination with a statin, has been shown to have no impact on plasma Lp(a) levels.28 These findings emphasize the presence of residual CVD risk in patients with high Lp(a) levels despite lipid-lowering therapy. They may also reflect that individuals with higher baseline Lp(a) levels – due to their increased inherent CV risk – are more likely to have experienced a greater number of atherosclerotic events, which in turn leads to more frequent initiation or intensification of lipid-lowering therapy.

This study faced certain limitations. Firstly, the relatively small sample size, which did not involve a power calculation, might have been inadequate to detect significant relationships between variables. Nonetheless, our main objective was to describe Lp(a) levels in patients with ACS. Second, during the initial implementation of our Lp(a) measurement protocol, adherence was inconsistent, with an increase in compliance observed after the first year of protocol implementation. As a result, only 42% of all ACS admissions during the study period were tested for Lp(a), according to the attending cardiologist's decision. This opportunistic sampling approach may have introduced selection bias, limiting the external validity of our findings and particularly the estimation of the true prevalence of elevated Lp(a) levels. Third, our cohort was predominantly male, which reflects the real-world ACS incidence but limits generalizability to female population. Fourth, data were retrospectively collected from hospital records, making the accuracy of the data dependent on the quality of clinical documentation.