High-density lipoprotein(a) (Lp(a)) is a well-established independent risk factor for atherosclerotic cardiovascular diseases (ASCVD). However, the interaction between Lp(a), low-density lipoprotein cholesterol (LDL-C), and polygenic risk score (PRS) in cardiovascular diseases has been the subject of relatively limited research. The present study included a total of 346,751 participants from the UK Biobank. According to the guideline of Lp(a), the study subjects were divided into 3 groups: the first group was <75 mmol/L (n = 272,643), the second group was 75 to 125 mmol/L (n = 35,792), and the third group was >125 mmol/L (n = 38,316). Elevated Lp(a) levels were associated with a progressively increased risk of overall cardiovascular events (CVEs), including ischemic stroke (IS), coronary heart disease (CHD), angina pectoris, and myocardial infarction (MI). In contrast, the risks of atrial fibrillation (AF) and heart failure (HF) decreased with higher Lp(a) levels. Additive interaction analyses revealed significant synergistic effects between Lp(a) and LDL-C for CHD (relative excess risk interaction [RERI] = 0.081, attributable proportion of interaction [AP] = 0.046, synergy index [SI] = 1.117), angina pectoris (RERI = 0.112, AP = 0.055, SI = 1.121), and MI (RERI = 0.183, AP = 0.079, SI = 1.161), with MI showing the strongest synergy. Incorporating PRS further amplified these effects, and the RERI (CHD: RERI = 0.721; angina pectoris: RERI = 0.781; MI: RERI = 1.318) and SI (CHD: SI = 2.218; angina pectoris: SI = 1.97; MI: SI = 2.326) were significantly higher than those of the interaction model containing only Lp(a) and LDL-C. In conclusion, Lp(a) and LDL-C show a significant synergistic effect in ASCVD, and this effect is more prominent in individuals with a higher PRS, suggesting that dual lipid management should be strengthened for such populations. While AF and HF may require alternative risk factor management.
Lipoprotein(a) (Lp[a]) is a specialized form of low-density lipoprotein cholesterol (LDL-C) assembled in the liver from LDL-C, comprised of a LDL-C particle and a unique glycoprotein known as apolipoprotein(a), which is covalently linked to the apolipoprotein B of LDL particle via disulfide bridge. And the levels of Lp(a) vary significantly across different populations. Recent studies indicate that approximately 20% of the global population exhibits elevated Lp(a) levels, which are strongly associated with an increased incidence of atherosclerotic cardiovascular disease (ASCVD) and aortic stenosis. , High Lp(a) levels are recognized as a causal risk factor for cardiovascular disease, akin to the established risks associated with elevated LDL-C cholesterol levels. ,,, Moreover, emerging research has demonstrated that elevated Lp(a) levels independently and linearly predict future cardiovascular events (CVEs). ,, Although numerous studies have demonstrated the association between Lp(a) and an increased risk of cardiovascular disease, research on the interaction between Lp(a) and LDL-C, as well as its relationship with polygenic risk scores (PRS), remains limited. Lp(a) is unique as it is entirely controlled by genetics and negligibly influenced by diet. Therefore, this study aims to investigate the relationship between Lp(a) levels and various CVEs in the general population, we hope to clarify these associations and provide further evidence for the growing role of Lp(a) in cardiovascular risk evaluation and management.
Methods
Study population
The UK Biobank is a population-scale longitudinal cohort study that recruited 500,000 people aged between 37 and 73 years in the UK from 2006 to 2010. The UK Biobank received ethical approval from the North West Multi-Centre Research Ethics Committee (REC reference: 11/NW/0382). This research was conducted by using the UK Biobank under Application Number 15255. Participant selection for the current study is depicted in Figure 1 .
Research cohort construction flowchart.
Definitions
Health information is collected from self-reported medical history and physical measurements at recruitment, and ongoing health information is collected through associated electronic health records.
CAD was defined based on International Classification of Disease edition 10 (ICD 10): I20 to I25. HF was defined ICD 10 revision codes of I11.0, I13.0, I13.2, I50.0, I50.1, and I50.9. AF was defined based on 10th ICD: I48.0 to I48.9).
Statistical analysis
Patient data are presented as either categorical or continuous variables. The Shapiro–Wilk test was employed to evaluate the normality of continuous variables. Variables that followed a normal distribution are described using the mean and standard deviation (SD), and group comparisons were conducted using the t test. Variables with skewed distributions are summarized using the median and interquartile range, and differences between groups were analyzed using the Mann–Whitney test. Categorical variables are expressed as proportions, with group differences assessed by the Pearson chi-square test.
According to the relevant guidelines for Lp(a), we analyzed Lp(a) as a categorical variable. The specific group divisions are as follows: Group 1 has Lp(a) levels <75 nmol/L, Group 2 has Lp(a) levels ranging from 75 to 125 nmol/L, and Group 3 has Lp(a) levels >125 nmol/L. And univariate and multivariate Cox regression models were used to assess the association between Lp(a) and various CVEs (IS, n = 5,953; angina pectoris, n = 11,694; MI, n = 8,560; CHD, n = 23,220; AF, n = 21,325; HF, n = 9,516; CVD, n = 89,808). Model 1 was not adjusted at all. Model 2 was adjusted for variables age, sex, race, smoking, drinking, education, physical, BMI, diabetes, hypertension, systolic blood pressure (SBP), diastolic blood pressure (DBP), lipid, triglyceride, insulin, hemoglobin A1c (HbA1c), estimated glomerular filtration rate (EGFR), antihypertensive drugs and aspirin. Based on model 2, stratified analysis was further conducted by combining LDL-C levels and PRS. To explore whether there is an additive and multiplicative interaction between Lp(a) and LDL-C, additive and multiplicative models were constructed. The additive interaction was determined by measuring whether the estimated value of the combined effect of Lp(a) and LDL-C exceeded the sum of their individual independent effects. In addition, the interaction effect between Lp(a), LDL-C and PRS was further analyzed by stratifying based on PRS. The cumulative effect of the interaction was evaluated using indicators such as RERI, AP, SI and multiplicative. When the 95% confidence intervals (95% CI) of RERI and AP included 0, and the 95% CI of SI included 1, it indicated that there was no significant additive interaction effect. A hazard ratio (HR) >1 with 95% CI excluding 1 indicates synergistic interaction; HR <1 with 95% CI excluding 1 suggests antagonistic interaction in multiplicative interaction. When the 95% CI includes 1, no statistically significant interaction.
Analyses were performed using R Version 4.2.3 (R Foundation for Statistical Computing, Vienna, Austria), with p values ≤0.05 considered statistically significant. And the standardized mean difference (SMD) <0.1, the difference between groups is generally considered to be statistically negligible.
Results
Baseline clinical characteristics
Table 1 demonstrates the clinical baseline characteristics of the different subgroups of Lp(a), with 346,751 individuals in the entire cohort, of whom 152,375 (43.94%) were male and 194,376 (56.06%) were female, with a mean age of 56.49 years. There were 272,643 (78.63%) individuals with Lp(a) levels <75 nmol/L, 35,792 (10.32%) individuals with Lp(a) levels between 75 nmol/L and 125 nmol/L, and 38,316 (11.05%) individuals with Lp(a) levels >125 nmol/L. And the results showed that except for a few variables (such as race and LDL-C) whose SMD were slightly higher than 0.1, the SMDs of most baseline variables were less than 0.1, indicating good baseline balance among the 3 groups ( Table 1 ).
Table 1
Baseline characteristics of cohort
| Variable Names | Overall | Missing | Lp(a) <75 mmol/L | 75 ≤Lp(a) ≤125 mmol/L | Lp(a) >125 mmol/L | p-value | SMD |
|---|---|---|---|---|---|---|---|
| n | N = 346,751 | n = 272,643 | n = 35,792 | n = 38,316 | |||
| Sex, n (%) | 0 | <0.001 | 0.045 | ||||
| Male | 152,375 (43.94) | 119,859 (43.96) | 16,324 (45.61) | 16,192 (42.26) | |||
| Female | 194,376 (56.06) | 152,784 (56.04) | 19,468 (54.39) | 22,124 (57.74) | |||
| Race, n (%) | 0.4 | <0.001 | 0.149 | ||||
| White | 325,464 (94.27) | 257,038 (94.67) | 32,390 (90.96) | 36,036 (94.46) | |||
| Mixed race | 2,187 (0.63) | 1,718 (0.63) | 272 (0.76) | 197 (0.52) | |||
| Black | 11,866 (3.44) | 9,488 (3.49) | 1,457 (4.09) | 921 (2.41) | |||
| Asian | 5,744 (1.66) | 3,259 (1.20) | 1,491 (4.19) | 994 (2.61) | |||
| Physical activity, n (%) | 5.1 | 0.507 | 0.007 | ||||
| Low | 40,894 (12.42) | 32,281 (12.47) | 4,115 (12.13) | 4,498 (12.37) | |||
| Moderate | 125,390 (38.09) | 98,558 (38.07) | 12,989 (38.30) | 13,843 (38.08) | |||
| High | 162,897 (49.49) | 128,073 (49.47) | 16,813 (49.57) | 18,011 (49.55) | |||
| SBP | 141.25 ± 19.422 | 5.6 | 141.254 ± 19.407 | 140.71 ± 19.537 | 141.728 ± 19.406 | <0.001 | 0.035 |
| DBP | 84.22 ± 10.412 | 5.6 | 84.191 ± 10.393 | 84.177 ± 10.596 | 84.467 ± 10.371 | <0.001 | 0.019 |
| BMI | 26.556 (24.001 to 29.637) | 0.4 | 26.542 (23.992 to 29.619) | 26.597 (23.937 to 29.78) | 26.574 (24.113 to 29.633) | 0.004 | 0.011 |
| Education, n (%) | 1 | <0.001 | 0.034 | ||||
| Others | 227,158 (66.20) | 178,487 (66.16) | 23,048 (65.11) | 25,623 (67.51) | |||
| College and above | 115,991 (33.80) | 91,310 (33.84) | 12,350 (34.89) | 12,331 (32.49) | |||
| Smoking status, n (%) | 0.5 | <0.001 | 0.03 | ||||
| Never | 193,405 (56.05) | 151,768 (55.93) | 20,495 (57.55) | 21,142 (55.48) | |||
| Former | 115,649 (33.52) | 91,316 (33.65) | 11,389 (31.98) | 12,944 (33.97) | |||
| Now | 35,997 (10.43) | 28,247 (10.41) | 3,730 (10.47) | 4,020 (10.55) | |||
| Alcohol consumption, n (%) | 0.2 | <0.001 | 0.035 | ||||
| Never | 15,181 (4.39) | 11,784 (4.33) | 1,841 (5.16) | 1,556 (4.07) | |||
| Former | 11,588 (3.35) | 9,091 (3.34) | 1,224 (3.43) | 1,273 (3.33) | |||
| Now | 319,133 (92.26) | 251,125 (92.33) | 32,610 (91.41) | 35,398 (92.60) | |||
| Hypertension, n (%) | 0 | <0.001 | 0.018 | ||||
| No | 212,148 (61.18) | 167,058 (61.27) | 22,021 (61.52) | 23,069 (60.21) | |||
| Yes | 134,603 (38.82) | 105,585 (38.73) | 13,771 (38.48) | 15,247 (39.79) | |||
| Diabetes, n (%) | 0 | <0.001 | 0.022 | ||||
| No | 325,897 (93.99) | 256,185 (93.96) | 33,524 (93.66) | 36,188 (94.45) | |||
| Yes | 20,854 (6.01) | 16,458 (6.04) | 2,268 (6.34) | 2,128 (5.55) | |||
| Laboratory tests | |||||||
| TG | 1.468 (1.036 to 2.127) | 0 | 1.484 (1.047 to 2.147) | 1.4 (0.984 to 2.053) | 1.421 (1.017 to 2.046) | <0.001 | 0.051 |
| HDL-C | 1.411 (1.185 to 1.685) | 8.5 | 1.409 (1.182 to 1.683) | 1.395 (1.169 to 1.673) | 1.437 (1.214 to 1.709) | <0.001 | 0.07 |
| LDL-C | 3.564 (3.017 to 4.146) | 0 | 3.545 (2.999 to 4.128) | 3.553 (3.015 to 4.124) | 3.708 (3.166 to 4.28) | <0.001 | 0.122 |
| Lp(a) | 44.387 ± 48.991 | 0 | 21.496 ± 17.674 | 99.72 ± 14.49 | 155.585 ± 18.657 | <0.001 | 5.188 |
| SCR | 69.9 (61.1 to 80.2) | 0 | 69.8 (61 to 80.1) | 70.4 (61.5 to 80.8) | 70 (61.1 to 80.3) | <0.001 | 0.024 |
| Cys-C | 0.881 (0.8 to 0.973) | 0.1 | 0.881 (0.801 to 0.974) | 0.877 (0.795 to 0.97) | 0.879 (0.801 to 0.97) | <0.001 | 0.017 |
| UA | 299.3 (247.9 to 356.4) | 0.1 | 299.3 (247.9 to 356.6) | 299.6 (247.3 to 356.1) | 299.2 (248.5 to 355) | 0.491 | 0.004 |
| BUN | 5.23 (4.46 to 6.09) | 0 | 5.24 (4.47 to 6.1) | 5.17 (4.4 to 6.04) | 5.25 (4.5 to 6.11) | <0.001 | 0.037 |
| GLU | 4.918 (4.592 to 5.291) | 8.5 | 4.919 (4.594 to 5.294) | 4.904 (4.577 to 5.278) | 4.92 (4.596 to 5.283) | 0.248 | 0.006 |
| ALB | 45.231 ± 2.603 | 8.5 | 45.259 ± 2.602 | 45.113 ± 2.602 | 45.14 ± 2.602 | <0.001 | 0.037 |
| TC | 5.71 (5.005 to 6.459) | 0 | 5.689 (4.985 to 6.439) | 5.669 (4.977 to 6.408) | 5.894 (5.184 to 6.63) | <0.001 | 0.124 |
| eGFR | 90.598 ± 13.805 | 0.1 | 90.568 ± 13.798 | 91.036 ± 13.926 | 90.402 ± 13.731 | <0.001 | 0.031 |
| Medication | |||||||
| Aspirin, n (%) | 0.3 | 0.011 | 0.015 | ||||
| No | 312,707 (90.45) | 245,895 (90.45) | 32,379 (90.78) | 34,433 (90.13) | |||
| Yes | 33,016 (9.55) | 25,959 (9.55) | 3,287 (9.22) | 3,770 (9.87) | |||
| Antihypertensive drugs, n (%) | 0.2 | <0.001 | 0.021 | ||||
| No | 286,704 (82.87) | 225,416 (82.86) | 29,818 (83.53) | 31,470 (82.34) | |||
| Yes | 59,250 (17.13) | 46,619 (17.14) | 5,881 (16.47) | 6,750 (17.66) | |||
| Lipid-lowering drugs, n (%) | 0.2 | <0.001 | 0.058 | ||||
| No | 303,251 (87.66) | 238,731 (87.76) | 31,690 (88.77) | 32,830 (85.90) | |||
| Yes | 42,703 (12.34) | 33,304 (12.24) | 4,009 (11.23) | 5,390 (14.10) | |||
| Insulin, n (%) | 0.2 | 0.548 | 0.004 | ||||
| No | 342,996 (99.14) | 269,688 (99.14) | 35,397 (99.15) | 37,911 (99.19) | |||
| Yes | 2,958 (0.86) | 2,347 (0.86) | 302 (0.85) | 309 (0.81) |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
