2025 Volume 32 Issue 4 Pages 421-438
Aims: To investigate the distribution of lipoprotein(a) (Lp(a)) and its association with atherosclerotic cardiovascular disease (ASCVD) in Japanese patients at high risk for ASCVD using a health insurance database.
Methods: Between July 2013 and June 2021, patients eligible for ASCVD prevention according to the 2017 Japan Atherosclerosis Society (JAS) guidelines with documented Lp(a) test results were extracted from the Medical Data Vision claims database and divided into three groups: primary prevention high-risk (Group I), secondary prevention (Group II) and secondary prevention high-risk (Group III). Data on lipid levels, cardiovascular morbidity risk factors and lipid-lowering treatments were extracted.
Results: Of 700,580 patients with documented low-density lipoprotein cholesterol (LDL-C), 2,967 (0.42%) were tested for Lp(a). In 2,170 eligible patients, the median [interquartile range] serum concentration of Lp(a) was 13.9 [7.5-24.6] mg/dL, with 151 patients (7.0%) above the recommended risk threshold of ≥ 50 mg/dL. Lp(a) levels increased with risk across all prevention groups. Being in the highest Lp(a) quintile (Q5) was associated with an increased frequency of ASCVD (28.9% versus 18.9% in the lowest quintile (Q1) for unstable angina; 18.7% versus 10.1% for myocardial infarction; 27.9% versus 17.0% for ischemic stroke). In the secondary prevention groups, the proportion of patients meeting an LDL-C target of ＜70 mg/dL decreased from 30.2% in Q1 to 19.0% in Q5 for Group II and from 32.9% to 16.3% for Group III.
Conclusions: Despite a high prevalence of Lp(a) ≥ 50mg/dL in Japanese patients at high risk for ASCVD, it found that the Lp(a) testing rate was very low.
The prevention of cardiovascular disease is a significant public health issue in Japan1, 2), against a backdrop of a rising incidence of atherosclerotic cardiovascular diseases (ASCVD) and hypercholesterolaemia3, 4). Risk factors for ASCVD include modifiable lifestyle factors such as high-fat diets and sedentarism, and intrinsic factors related to cholesterol and triglyceride metabolism. One of the most important of these intrinsic risk factors is lipoprotein(a) (Lp(a)), which is a unique lipoprotein with pro-atherogenic, pro-thrombotic and pro-inflammatory properties5-7). High levels of circulating Lp(a) are associated with significant increases in the risk of ASCVD and ASCVD-related mortality8-14). Lp(a) levels are genetically determined and, unlike LDL-cholesterol (LDL-C), are not significantly influenced by diet or other lifestyle factors6, 15). Different genetic variants affecting Lp(a) levels exist in different human populations leading to geographical variation in circulating Lp(a) levels, which are relatively low in East Asian populations such as in Japan15). The relevance of Lp(a) to strategies for the prevention of ASCVD was for a long time unclear, since no consensus existed on appropriate risk thresholds for plasma Lp(a) levels5, 6, 16) and standardised methods for dosing have proved challenging to establish16). Moreover, lipid-modifying drugs do not significantly modify Lp(a) levels5), except proprotein convertase subtilisin kexin type 9 (PCSK9) inhibitors17). However, it is increasingly recognised that high plasma Lp(a) should be considered as part of the individual risk assessment when setting treatment goals for lipid-lowering therapy in primary or secondary ASCVD prevention16).
In 2017, the Japan Atherosclerosis Society (JAS) published revised guidelines for the prevention of ASCVD (JAS GL2017), which set explicit treatment goals for LDL-C, now stratified on the patient’s risk profile18). We have recently performed a large study on rates of achievement of the 2017 JAS Guidelines (GL) these treatment goals19). using a Japanese hospital-based insurance claims database (Medical Data Vision; MDV; Tokyo, Japan)20). In this population of 395,314 patients with documented LDL-C test results between July 2018 and June 2021, we observed that JAS GL2017 LDL-C goals were unmet in a large proportion of patients (34.4% of patients in the primary prevention high-risk setting, 39.4% of those in the secondary prevention setting and 74.6% of those in the secondary prevention high-risk setting) and that the majority of eligible patients were not prescribed statins. We have now extended these observations to an evaluation of Lp(a) and its influence on LDL-C levels.
The objectives of this study were to investigate the distribution of serum Lp(a) and its association with ASCVD in Japanese patients at high risk for ASCVD using real-world data from a health insurance database.
This cross-sectional, retrospective study used data from a large hospital-based insurance claims database in Japan (Medical Data Vision; MDV; Tokyo, Japan)20). Data was extracted for all patients with at least one documented laboratory measure of LDL-C between 1st July 2013 and 30th June 2021. The index date was defined as the date of the most recent LDL-C measure. A baseline period of twelve months before the index date was defined for data collection on medical history. The methods follow those of our previous study of lipids in the MDV database, where further information is available19).
Study PopulationThe study population consisted of adult (≥ 18 years) patients with at least one measure of LDL-C and one measure of Lp(a) during the study period and who were considered to be at high risk of ASCVD, as defined in the JAS GL2017 18). These were defined as at least one claim at any time before the index date for angina pectoris (AP), myocardial infarction (MI), ischaemic stroke (IS), peripheral artery disease (PAD), diabetes mellitus (DM), chronic kidney disease (CKD) or familial hypercholesterolaemia (FH). These conditions were defined from the relevant International Classification of Diseases 10th Edition (ICD-10) codes documented on the claim. These ICD-10 codes are provided in Supplementary Table 1. In addition, patients had to have at least one healthcare claim (irrespective of its nature) every six months during the baseline period.
| Disease | ICD-10 codes |
| Angina pectoris | I20.0, I20.1, I20.8, I20.9 |
| Myocardial infarction | I21.0-9, I22.0-9, I23.0-8, I24.0-9, I25.2 |
| Coronary artery disease |
I20.0, I20.1, I20.8, I20.9, I21.0-4, I21.9, I22.0, I22.1, I22.8, I22.9, I23.0-6, I23.8, I24.0, I24.1, I24.8, I24.9, I25.1-6, I25.8, I25.9 Z951, Z955 (procedure codes) |
| Ischaemic stroke | I63.0-9, I65.0, I65.1, I65.2, I65.3, I66.0-9 |
| Peripheral artery disease | I70.0-9, I71.0-9, I72.0-9, I73.0-9, Z95.8 |
| Aortic aneurism | I71.0-6, I71.9 |
| Aortic valve stenosis | I06.0, I06.2, I08.0, I08.2, I08.3, I08.8, I35.0, I35.2, Q230, Q24.4, Q25.3 |
| Mitral valve stenosis | I05.0, I05.2, I08.0, I08.1, I08.3, I08.8, I34.2, Q232, Q233 |
| Dyslipidaemia | E10.6, E11.6, E14.6, E78.0-9 |
| Hypercholesterolaemia | E14.6, E78.0 |
| Familial hypercholesterolaemia | 272001, 8845523, 8845524, 8831271 (diagnosis code) |
| Diabetes mellitus | E10.0-9, E11.0-9, E12.0-9, E13.0-9, E14.0-9, O24.4 |
| Renal disease | E10.2, E11.2, E13.2, E14.2, E85.0, I12.0, I12.9, M100.9, M32.1, M34.8, N00.2-4, N00.7, N00.9, N01.2, N01.4, N01.7, N01.9, N02.8, N02.9, N03.0, N03.2-4, N03.7, 03.9, N04.0-4, N04.6, N04.9, N05.1-9, N07.8, N07.9, N11.9, N14.0-4, N17.0-2, N17.8, N17.9, N18.1-5, N18.9, N19, N26, N27.0, N27.1, N27.9, N28.0,N28.1, N28.9, N39.1, N99.0, Q60.0-2, Q60.5, Q61.0-5, Q61.8, Q62.0, Q62.1, Q62.5 |
| Chronic kidney disease | N18.1-4, N189 |
| Hyperuricaemia | E790 |
| Liver disease | B00.8, B15.0, B15.9, B16.2, B16.9, B17.1, B17.2, B17.8, B17.9, B18.1, B18.2, B18.9, B19.0, B19.9, B25.1, B26.8, B27.0, B33.8, B58.1, B65.9, B66.0, B66.1, B66.3, B67.0, B67.5, B67.8, B83.8, B89, E10.6, E11.6, E13, E13.0-9, E14.6, E55.0, E72.2, E74.0, E80.2, E83.1, E88.8, K45.8, K65.0, K65.8, K70.0-4, K70.9, K71.0-3, K71.6-9, K72.0, K72.9, K73.0, K73.2, K73.8, K73.9, K74.0, K74.1, K74.3-6, K75.0, K75.2-4, K75.8, K75.9, K76.0-5, K76.7-9, K918 |
| Chronic obstructive pulmonary disease | J43.0-2, J43.9, J44.0, J44.1, J48.8, J48.9 |
The study population was divided into a number of ASCVD risk subgroups, as defined in the JAS GL2017 18). Patients with at least one claim for IS, PAD, DM, CKD or FH without a claim for coronary heart disease (AP or MI) at any time before the index date were assigned to the primary prevention high-risk subgroup (Group I) and those with at least one such claim for coronary heart disease (AP or MI) to the secondary prevention group (Group II). Patients within the secondary prevention group who had, in addition, a claim for MI or unstable angina (UA) within the twelve-month baseline period, a claim for FH at any time before the index date, or a claim for DM together with a claim for IS, PAD or CKD at any time before the index date were identified and these together formed the high-risk secondary prevention group (Group III).
Data Extraction and Study VariablesAge, gender, body mass index (BMI) and smoking status were documented at the index date. The most recent laboratory test results, up to and including the index date, were documented for lipid-related analytes (total, LDL-C, high-density lipoprotein cholesterol (HDL-C) and non-HDL-C and triglycerides). Patients were assigned to classes based on serum LDL-C concentrations, using thresholds of ＜120 mg/dL, ＜100 mg/dL and ＜70 mg/dl. These thresholds represent the recommended LDL-C treatment goals for Groups I, II and III respectively in the JAS GL2017 18). Other analytes documented when available were glycosylated haemoglobin (Hb1Ac), serum creatinine, estimated glomerular filtration rate (eGFR), blood urea nitrogen (BUN) and blood cell counts.
Certain morbidities documented in the database up to and including the index date were identified from the appropriate ICD-10 diagnosis codes (Supplementary Table 1). These included ASCVD presentations (coronary artery disease, AP, UA, MI and recent acute coronary syndrome), lipid disorders (dyslipidaemia, hypercholesterolaemia and FH), other cardiovascular disorders (hypertension, chronic heart failure, IS, PAD), structural heart disorders (aortic valve stenosis, mitral valve stenosis and aortic aneurism), DM, renal disease (a composite of diabetic nephropathy, renal amyloidosis, hypertensive renal disease, malignant nephrosclerosis, atherosclerotic kidney disease, shock-induced renal failure, CKD, kidney disease associated with gout, scleroderma or lupus, glomerulonephritis, nephrotic syndrome, drug-induced or toxic kidney disease, or organic or structural renal disease), CKD (defined as CKD explicitly in the ICD-10 code or uraemic neuropathy, uraemic polyneuropathy, uraemic cardiomyopathy, uraemic pericarditis, uraemic encephalopathy, uraemic lung, renal retinopathy, erythropoiesis stimulating drug hyporesponsive anaemia or end-stage renal failure), liver failure (a composite of viral hepatitis, parasitic liver disease, cirrhosis, diabetes-associated liver disease, alcohol-related liver disease, toxic liver disorders, autoimmune liver diseases, steatosis, fibrosis, cirrhosis, jaundice, hepatic coma and cholangitis) hyperuricaemia and chronic pulmonary disease (COPD).
Medications were identified from the appropriate European Pharmaceutical Marketing Research Association Anatomical Therapeutic Chemical (ATC) codes in the database. The only treatments retained were those prescribed for at least four consecutive weeks up to and including the index date, without any gap of ＞3 days between the estimated end of one prescription (defined as the date of the prescription plus the number of days that the prescription covered) and the beginning of the following one. In particular, statins, ezetimibe and PCSK9 inhibitors were documented. Statin treatment regimens were classified as standard statins (fluvastatin, pravastatin and simvastatin) or intensive statins (atorvastatin, rosuvastatin and pitavastatin) and statins in combination with ezetimibe21).
Statistical AnalysisPatients with at least one Lp(a) measurement were compared to patients with no such measurement. Categorical variables were compared using the χ² test and continuous variables using Student’s t-test or the Mann-Whitney U-test as appropriate.
The study population was divided into five equivalent-sized quintiles by serum Lp(a) levels. The association of Lp(a) levels with morbidities, other serum lipids, achievement of LDL-C goals and treatments was evaluated using the Cochran-Armitage trend test for categorical variables. For continuous variables, the analysis of variance (ANOVA) linear trend test was employed for normal distributions or the Jonckheere-Terpstra trend test for non-normal distributions.
Multivariate regression analysis was performed to identify variables potentially related to serum Lp(a) concentrations. The dependent variable was the proportion of patients in the fifth (uppermost) quintile of Lp(a) levels. Gender and age by class were included in the model as forced variables. Independent variables included in the model were eGFR (per 5 mL/min/1.73m2 incremental decrease), BUN (per 1 mg/dL incremental increase ), LDL-C (per 10 mg/dL incremental increase), HDL-C (per 10 mg/dL incremental increase), triglycerides (per 10 mg/dL incremental increase) and HbA1c (per 1% incremental increase), the presence of any of the morbidities of interest (except familial hypercholesterolaemia, for which the number of patients affected was anticipated to be very low), and treatment with standard statins, intensive statins and ezetimibe. A forward stepwise regression was performed, with variables retained in each step if they were significantly associated with serum Lp(a) levels at a probability threshold of 0.1. Separate multivariate analyses were performed for the three risk subgroups (Groups I to III).
EthicsThe study complied with all relevant international and national legislation on medical research and data privacy. In particular, it complied with the Declaration of Helsinki (Fortaleza Revision, 2013) and the Japanese Act on the Protection of Personal Information (Act No. 57, 2003 and next revisions). Furthermore, since the data was anonymised prior to extraction, the Japanese Pharmaceuticals and Medical Devices Agency guidelines for conducting pharmacoepidemiological research using medical databases, which specify when ethics approval and informed consent are required, do not apply.
During the study period, 700,580 patients had at least one LDL-C measure, constituting the source cohort. Of these, 2,967 (0.42%) also had a measure of Lp(a) during the study period. After the exclusion of 797 patients who were ineligible, the remaining 2,170 patients constituted the study population. Of these, 1,293 (59.6%) were assigned to Group I (primary prevention) and 877 (40.4%) to Group II (secondary prevention). Within Group II, 712 patients (81.2%) constituted Group III (secondary prevention – high risk). A study flowchart is provided in Fig.1. The characteristics of the overall population and subgroups are presented in Table 1.
Percentages are calculated with the patient numbers on the line above as the denominator. The three groups were constituted on the basis of ASCVD risk factors as specified in the Methods.
ASCVD: atherosclerotic cardiovascular disease; LDL-C: low-density lipoprotein cholesterol; Lp(a): lipoprotein(a).
| Variable | No Lp(a) measure (n=697,613) | ≥ 1 Lp(a) measure (n=2,967) | p-value | Study population | By statin treatment＊ | |||||
|---|---|---|---|---|---|---|---|---|---|---|
| Overall (n=2,170) | Group I (n=1,293) | Group II (n=877) | Group III (n=712) | Treated (n=783) | Non-treated (n=1385) | p-value | ||||
| Gender (%) | ＜0.0001 | 0.0007 | ||||||||
| Women | 336,909 (48.3%) | 1,144 (38.6%) | 811 (37.4%) | 530 (41.0%) | 281 (32.0%) | 220 (30.9%) | 329 (42.0%) | 480 (34.7%) | ||
| Men | 360,704 (51.7%) | 1,823 (61.4%) | 1,359 (62.6%) | 763 (59.0%) | 596 (68.0%) | 492 (69.1%) | 454 (58.0%) | 905 (65.3%) | ||
| Age (years) | ＜0.0001 | ＜0.0001 | ||||||||
| Mean±SD | 65.1±18.3 | 66.5±15.6 | 67.1±14.5 | 63.9±15.4 | 71.9±11.4 | 72.3±11.1 | 68.9±12.0 | 66.2±15.6 | ||
| ＜50 years | 137,795 (19.8%) | 416 (14.0%) | 251 (11.6%) | 219 (16.9%) | 32 (3.6%) | 21 (2.9%) | 49 (6.3%) | 201 (14.5%) | ||
| 50-59 years | 91,252 (13.1%) | 391 (13.2%) | 291 (13.4%) | 204 (15.8%) | 87 (9.9%) | 69 (9.7%) | 107 (13.7%) | 184 (13.3%) | ||
| 60-69 years | 135,683 (19.4%) | 754 (25.4%) | 599 (27.6%) | 383 (29.6%) | 216 (24.6%) | 173 (24.3%) | 236 (30.1%) | 362 (26.1%) | ||
| 70-79 years | 171,036 (24.5%) | 767 (25.9%) | 582 (26.8%) | 299 (23.1%) | 283 (32.3%) | 230 (32.3%) | 231 (29.5%) | 351 (25.3%) | ||
| ≥ 80 years | 161,847 (23.2%) | 639 (21.5%) | 447 (20.6%) | 188 (14.5%) | 259 (29.5%) | 219 (30.8%) | 160 (20.4%) | 287 (20.7%) | ||
| Body mass index (kg/m²) | N=212,377 | N=1,556 | ＜0.0001 | N= 942 | N=508 | N=434 | N=383 | N= 293 | N= 649 | ＜0.0001 |
| Mean±SD | 22.6±4.5 | 24.2±5.3 | 24.4±5.3 | 24.4±5.7 | 24.3±4.7 | 24.4±4.7 | 25.6±4.7 | 23.8±5.5 | ||
| 25-29.9 kg/m² | 41,942 (19.7%) | 377 (24.2%) | 244 (25.9%) | 125 (24.6%) | 119 (27.4%) | 109 (28.5%) | 101 (34.5%) | 143 (22.0%) | ||
| ≥ 30 kg/m² | 11,911 (5.6%) | 190 (12.2%) | 120 (12.7%) | 68 (13.4%) | 52 (12.0%) | 47 (12.3%) | 48 (16.4%) | 72 (11.1%) | ||
| Smoking status | N=201,197 | N=1440 | N= 842 | N=466 | N=376 | N=326 | N= 264 | N= 578 | ||
| Past or present smoker (n, %) | 72,392 (36.0%) | 591 (41.0%) | ＜0.0001 | 330 (39.2%) | 195 (41.8%) | 135 (35.9%) | 119 (36.5%) | 106 (40.2%) | 224 (38.8%) | 0.70 |
| Brinkman smoking index (median, Q1-Q3) | 0 [0.0 - 400.0] | 0 [0.0 - 516.0] | ＜0.0001 | 0 [0.0 - 500.0] | 0 [0.0 - 480.0] | 0 [0.0 - 562.0] | 0 [0.0 - 564.0] | 0.0 [0.0 ; 470.0] | 0.0 [0.0 ; 500.0] | 0.83 |
| Laboratory tests | ||||||||||
| Total cholesterol (mg/dL) | N=498,877 | N=2,606 | ＜0.0001 | N= 2,024 | N=1,240 | N=784 | N=659 | N= 732 | N= 1292 | ＜0.0001 |
| Mean±SD | 188.1±43.1 | 174.7±40.2 | 173.3±39.6 | 180.4±38.7 | 161.9±38.3 | 159.1±37.6 | 165.1±35.3 | 177.9±41.1 | ||
| HDL-C (mg/dL) | N=639,430 | N=2,937 | ＜0.0001 | N= 2,153 | N=1,283 | N =870 | N=705 | N= 778 | N= 1373 | 0.34 |
| Mean±SD | 57.9±18.0 | 53.2±16.8 | 53.6±17.1 | 55.2±17.9 | 51.4±15.6 | 50.1±15.3 | 53.2±15.8 | 53.9±17.8 | ||
| LDL-C (mg/dL) | ＜0.0001 | ＜0.0001 | ||||||||
| Mean±SD | 108.8±34.4 | 99.6±32.8 | 97.2±31.8 | 102.7±31.1 | 89.1±31.0 | 87.1±30.5 | 88.3±26.8 | 102.3±33.2 | ||
| Non-HDL-C (mg/dL) | N=498,721 | N=2,606 | N= 2,024 | N=1,240 | N =784 | N=659 | N= 732 | N= 1292 | ＜0.0001 | |
| Mean±SD | 130.6±40.0 | 121.7±37.9 | 119.6±36.6 | 125.2±36.2 | 110.7±35.6 | 108.8±35.0 | 111.9±32.2 | 124.0±38.3 | ||
| Total triglycerides (mg/dL) | N=665,020 | N=2,944 | ＜0.0001 | N= 2,154 | N= 1,287 | N= 867 | N= 702 | N= 781 | N= 1371 | 0.0320 |
| Median [Q1-Q3] | 103 [73.0 - 150.0] | 112 [78.0 - 162.0] | 112.5 [79.0 - 163.0] | 111 [78.0 - 167.0] | 115 [80.0 - 157.0] | 115 [80.0 - 159.0] | 121.0 [84.0 ; 171.0] | 109.0 [77.0 ; 160.0] | ||
| HbA1c (%) | N=506,544 | N=2,701 | ＜0.0001 | N= 2,082 | N= 1,249 | N= 833 | N= 692 | N= 760 | N= 1320 | 0.93 |
| Mean±SD | 6.1±1.1 | 7.4±2.0 | 7.4±1.8 | 7.5±1.8 | 7.1±1.7 | 7.2±1.7 | 7.4±1.6 | 7.4±1.9 | ||
| Blood urea nitrogen (mg/dL) | N=654,679 | N=2,929 | ＜0.0001 | N= 2,143 | N= 1,275 | N= 868 | N= 707 | N= 772 | N= 1369 | 0.46 |
| Mean±SD | 17.9±11.8 | 18.0±10.2 | 18.8±10.5 | 17.8±10.1 | 20.2±10.9 | 21.1±11.6 | 18.5±9.5 | 18.9±11.1 | ||
| eGFR (mL/min/1.73m²) | N=674,481 | N=2,934 | ＜0.0001 | N= 2150 | N= 1,277 | N= 873 | N= 709 | N= 774 | N= 1374 | 0.0050 |
| Median [Q1-Q3] | 68.7 [54.5 - 82.7] | 66.5 [50.9 - 81.8] | 64.5 [48.4 - 78.8] | 68.6 [53.8 - 83.7] | 57 [42.5 - 72.1] | 55.1 [39.8 - 69.5] | 62.1 [47.4 ; 76.5] | 65.6 [48.6 ; 80.3] | ||
| Morbidities | ||||||||||
| Dyslipidaemia | 205,840 (29.5%) | 1,765 (59.5%) | ＜0.0001 | 1,533 (70.6%) | 779 (60.2%) | 754 (86.0%) | 631 (88.6%) | 781 (99.7%) | 750 (54.2%) | ＜0.0001 |
| Hypercholesterolaemia | 97,492 (14.0%) | 1,071 (36.1%) | ＜0.0001 | 962 (44.3%) | 427 (33.0%) | 535 (61.0%) | 453 (63.6%) | 606 (77.4%) | 354 (25.6%) | ＜0.0001 |
| Familial hypercholesterolaemia | 1,051 (0.2%) | 19 (0.6%) | ＜0.0001 | 14 (0.6%) | 7 (0.5%) | 7 (0.8%) | 7 (1.0%) | 5 (0.6%) | 7 (0.5%) | 0.77 |
| Hypertension | 260,685 (37.4%) | 1,978 (66.7%) | ＜0.0001 | 1,704 (78.5%) | 427 (33.0%) | 535 (61.0%) | 453 (63.6%) | 665 (84.9%) | 1039 (75.0%) | ＜0.0001 |
| Coronary artery disease | 109,014 (15.6%) | 985 (33.2%) | ＜0.0001 | 896 (41.3%) | 19 (1.5%) | 877 (100.0%) | 712 (100.0%) | 424 (54.2%) | 472 (34.1%) | ＜0.0001 |
| Angina pectoris | 94,052 (13.5%) | 925 (31.2%) | ＜0.0001 | 852 (39.3%) | 0 (0.0%) | 852 (97.1%) | 695 (97.6%) | 417 (53.3%) | 435 (31.4%) | ＜0.0001 |
| Unstable angina pectoris | 18,611 (2.7%) | 525 (17.7%) | ＜0.0001 | 489 (22.5%) | 0 (0.0%) | 489 (55.8%) | 457 (64.2%) | 281 (35.9%) | 208 (15.0%) | ＜0.0001 |
| Myocardial infarction | 31,142 (4.5%) | 339 (11.4%) | ＜0.0001 | 314 (14.5%) | 0 (0.0%) | 314 (35.8%) | 272 (38.2%) | 179 (22.9%) | 135 (9.7%) | ＜0.0001 |
| Recent acute coronary syndrome | 19,920 (2.9%) | 438 (14.8%) | ＜0.0001 | 411 (18.9%) | 0 (0.0%) | 411 (46.9%) | 411 (57.7%) | 233 (29.8%) | 178 (12.9%) | ＜0.0001 |
| Chronic heart failure | 106,579 (15.3%) | 695 (23.4%) | ＜0.0001 | 628 (28.9%) | 166 (12.8%) | 462 (52.7%) | 391 (54.9%) | 278 (35.5%) | 350 (25.3%) | ＜0.0001 |
| Aortic valve stenosis | 6,676 (1.0%) | 49 (1.7%) | ＜0.0001 | 44 (2.0%) | 10 (0.8%) | 34 (3.9%) | 29 (4.1%) | 23 (2.9%) | 21 (1.5%) | 0.0240 |
| Mitral valve stenosis | 9,648 (1.4%) | 38 (1.3%) | 0.58 | 7 (0.3%) | 2 (0.2%) | 5 (0.6%) | 2 (0.3%) | 3 (0.4%) | 4 (0.3%) | 0.71 |
| Aortic aneurism | 9,648 (1.4%) | 38 (1.3%) | 0.63 | 36 (1.7%) | 7 (0.5%) | 29 (3.3%) | 23 (3.2%) | 11 (1.4%) | 25 (1.8%) | 0.48 |
| Ischaemic stroke | 63,704 (9.1%) | 625 (21.1%) | ＜0.0001 | 514 (23.7%) | 292 (22.6%) | 222 (25.3%) | 190 (26.7%) | 184 (23.5%) | 330 (23.8%) | 0.86 |
| Peripheral artery disease | 54,145 (7.8%) | 412 (13.9%) | ＜0.0001 | 386 (17.8%) | 152 (11.8%) | 234 (26.7%) | 209 (29.4%) | 162 (20.7%) | 224 (16.2%) | 0.0080 |
| Diabetes mellitus | 184,641 (26.5%) | 2,090 (70.4%) | ＜0.0001 | 1,824 (84.1%) | 1,171 (90.6%) | 653 (74.5%) | 601 (84.4%) | 671 (85.7%) | 1152 (83.2%) | 0.12 |
| Hyperuricaemia | 608,23 (8.7%) | 478 (16.1%) | ＜0.0001 | 440 (20.3%) | 207 (16.0%) | 233 (26.6%) | 208 (29.2%) | 165 (21.1%) | 275 (19.9%) | 0.50 |
| Renal disease | 93,496 (13.4%) | 641 (21.6%) | ＜0.0001 | 581 (26.8%) | 310 (24.0%) | 271 (30.9%) | 249 (35.0%) | 214 (27.3%) | 366 (26.4%) | 0.65 |
| Chronic kidney disease | 232,556 (33.3%) | 1,176 (39.6%) | ＜0.0001 | 1,394 (64.2%) | 719 (55.6%) | 675 (77.0%) | 596 (83.7%) | 530 (67.7%) | 863 (62.3%) | 0.0120 |
| Liver disease | 115,640 (16.6%) | 732 (24.7%) | ＜0.0001 | 684 (31.5%) | 418 (32.3%) | 266 (30.3%) | 236 (33.1%) | 274 (35.0%) | 408 (29.5%) | 0.0080 |
| Chronic obstructive pulmonary disease | 18,762 (2.7%) | 93 (3.1%) | 0.13 | 90 (4.1%) | 34 (2.6%) | 56 (6.4%) | 45 (6.3%) | 36 (4.6%) | 54 (3.9%) | 0.43 |
＊Two patients treated with a PCSK9 inhibitors and statins combination were excluded from this analysis. There were no patients treated with PCSK9 inhibitors monotherapy
eGFR: estimated glomerular filtration rate; HbA1c: haemoglobin A1c; HDL-C: high density lipoprotein cholesterol; LDL-C: low density lipoprotein cholesterol; Lp(a): lipoprotein(a); SD: standard deviation; Q1: first quartile; Q3: third quartile.
The characteristics of the patients were additionally compared between the 2,967 patients with an Lp(a) measurement and the remaining 697,613 patients in whom only LDL-C and not Lp(a) was measured (Table 1). Significant differences (p＜0.001) were observed for most variables evaluated. Patients with at least one Lp(a) measure were older and more frequently men, overweight or obese, and current or former smokers than in those in whom Lp(a) was not measured. Large differences were observed for the proportion of patients with ASCVD, lipid disorders and DM: for AP, 31.2% in patients with an Lp(a) measure vs. 13.5% in those without, 17.7% vs. 2.7% respectively for UA, 11.4% vs. 4.5% respectively for MI, 14.8% vs. 2.9% respectively for a recent acute coronary syndrome (ACS), 21.1% vs. 9.1% respectively for IS, 13.9% vs. 7.8% respectively for PAD, 36.1% vs. 14.0% respectively for hypercholesterolaemia and 70.4% vs. 26.5% respectively for DM.
Table 1 also presents comparison of patients treated and non-treated with statins. Significant differences (p＜0.001) were observed for a majority of evaluated variables. Patients treated with statins were older and more frequently women, and overweight or obese than those non-treated with statins. Those treated with statins had lower total cholesterol, LDL-C and non-HDL-C values as well as higher triglycerides compared to those non-treated with statins. Large differences were observed for the proportion of patients with ASCVD, and lipid disorders: for AP, 53.3% in patients treated vs. 31.4% non-treated with statins, 35.9% vs. 15.0% respectively for UA, 22.9% vs. 9.7% respectively for MI, 29.8% vs. 12.9% respectively for a recent ACS and 77.4% vs. 25.6% respectively for hypercholesterolaemia.
Lipoprotein(a) LevelsIn the 2,170 patients of the study population, the mean serum concentration of Lp(a) was 19.6±19.5 mg/dL and the median concentration was 13.9 mg/dL (interquartile range: 7.5 – 24.6 mg/dL]. The distribution of serum concentrations of Lp(a) in all patients is presented in Fig.2A. The distribution is strongly skewed towards lower values. Serum concentrations ≥ 30 mg/dL were reported in 377 patients (17.4%), ≥ 50 mg/dL in 151 patients (7.0%), and ≥ 70 mg/dL in 74 patients (3.4%) in the overall population. Figs.2B, 2C and 2D illustrate the distribution of Lp(a) levels in the three risk prevention groups: the distribution was shifted to higher values moving from Group I to Group II (p=0.03; Mann-Whitney U-test) and Group III (p=0.005). Median [interquartile range] Lp(a) serum concentrations increased from 13.8 [7.0-23.2] mg/dL in Group I to 14.8 [8.3-27.0] mg/dL in Group III (Fig.2). Median [interquartile range] Lp(a) serum concentrations were higher among patients treated with statins compared to untreated patients (14.7 [8.1-26.3] mg/dL vs. 13.4 [7.0- 23.1] mg/dL, p=0.002; Fig.2E and 2F). Similarly, the proportions of patients with Lp(a) levels higher than the different thresholds (≥ 30 mg/dL, ≥ 50 mg/dL and ≥ 70 mg/dL) were higher in treated than in untreated patients (Fig.2E and 2F).
Lipoprotein(a) levels were ＞200 mg/dL in 1 patient in Group I (280 mg/dL).
IQR: interquartile range; SD: standard deviation.
p-value (Group I vs Group II)=0.033; p-value (Group I vs Group III)=0.003; p-value (Treated vs Non-treated with statins)=0.002.
The distribution of Lp(a) levels by quintile is presented in Fig.3. The overall study population had between 430 and 439 patients per quintile. The upper limit of Lp(a) levels ranged between 6.0 mg/dL (Group I) and 7.1 mg/dL (Group III) in the first quartile and the lower limit between 27.2 mg/dL (Group I) and 30.7 mg/dL (Group III).
Data are presented as mean values with standard deviations. The boundaries for each quintile were as follows for the overall population: first quintile: ≤ 6.3 mg/dL, second quintile: 6.4 – 11.0 mg/dL, third quintile: 11.1 – 17.2 mg/dL, fourth quintile: 17.3 – 27.7 mg/dL, fifth quintile: ≥ 27.8 mg/dL, for the primary prevention group: first quintile: ≤ 6.0 mg/dL, second quintile: 6.1 – 10.6 mg/dL, third quintile: 10.7 – 16.8 mg/dL, fourth quintile: 16.9 – 27.0 mg/dL, fifth quintile: ≥ 27.1 mg/dL, for the secondary prevention group: first quintile: ≤ 7.0 mg/dL, second quintile: 7.1 – 11.7 mg/dL, third quintile: 11.8 – 18.2 mg/dL, fourth quintile: 18.3 – 29.2 mg/dL, fifth quintile: ≥ 29.3 mg/dL and for the secondary high-risk prevention group: first quintile: ≤ 7.1 mg/dL, second quintile: 7.2 – 12.0 mg/dL, third quintile: 12.1 – 18.8 mg/dL, fourth quintile: 18.9 – 30.6 mg/dL, fifth quintile: ≥ 30.7 mg/dL.
The relationship between Lp(a) levels and morbidities is shown in Table 2. In the study population, the trend test demonstrated significantly higher Lp(a) levels in patients with atherosclerotic disease, who made up a higher proportion of patients in the higher Lp(a) quintiles compared to the lower quintiles. This was the case for patients with CAD (from 37.0% in the first quintile to 45.3% in the fifth quintile), AP (from 35.6% in the first quintile to 43.2% in the fifth quintile), UA (from 18.9% in the first quintile to 28.9% in the fifth quintile), MI (from 10.1% in the first quintile to 18.7% in the fifth quintile), recent ACS (from 15.4% in the first quintile to 24.5% in the fifth quintile), chronic heart failure (from 24.8% in the first quintile to 35.6% in the fifth quintile) and IS (from 17.0% in the first quintile to 27.9% in the fifth quintile), as well as for patients with CKD (from 57.5% in the first quintile to 69.7% in the fifth quintile). The opposite relationship was observed for patients with DM (from 87.8% in the first quintile to 79.9% in the fifth quintile) or liver disease (from 38.2% in the first quintile to 27.7% in the fifth quintile) who were over-represented in the lower Lp(a) quintiles.
| MORBIDITY | 1st Lp(a) quintile | 2nd Lp(a) quintile | 3rd Lp(a) quintile | 4th Lp(a) quintile | 5th Lp(a) quintile | p-valuea |
|---|---|---|---|---|---|---|
| Dyslipidaemia | 288 (66.2%) | 292 (67.4%) | 305 (69.5%) | 314 (73.0%) | 334 (77.1%) | ＜0.001 |
| Hypercholesterolaemia | 173 (39.8%) | 171 (39.5%) | 183 (41.7%) | 207 (48.1%) | 228 (52.7%) | ＜0.001 |
| FH | 2 (0.5%) | 2 (0.5%) | 3 (0.7%) | 3 (0.7%) | 4 (0.9%) | 0.34 |
| Hypertension | 344 (79.1%) | 332 (76.7%) | 329 (74.9%) | 342 (79.5%) | 357 (82.4%) | 0.12 |
| Coronary artery disease | 161 (37.0%) | 170 (39.3%) | 188 (42.8%) | 181 (42.1%) | 196 (45.3%) | 0.01 |
| Angina pectoris | 155 (35.6%) | 161 (37.2%) | 178 (40.5%) | 171 (39.8%) | 187 (43.2%) | 0.02 |
| Unstable angina | 82 (18.9%) | 76 (17.6%) | 108 (24.6%) | 98 (22.8%) | 125 (28.9%) | ＜0.0001 |
| Myocardial infarction | 44 (10.1%) | 54 (12.5%) | 73 (16.6%) | 62 (14.4%) | 81 (18.7%) | 0.0003 |
| Recent ACS | 67 (15.4%) | 60 (13.9%) | 94 (21.4%) | 84 (19.5%) | 106 (24.5%) | ＜0.0001 |
| Chronic heart failure | 108 (24.8%) | 114 (26.3%) | 130 (29.6%) | 122 (28.4%) | 154 (35.6%) | 0.001 |
| Aortic valve stenosis | 11 (2.5%) | 7 (1.6%) | 6 (1.4%) | 10 (2.3%) | 10 (2.3%) | 0.90 |
| Mitral valve stenosis | 2 (0.5%) | 1 (0.2%) | 3 (0.7%) | 1 (0.2%) | 0 (0.0%) | 0.29 |
| Aortic aneurism | 7 (1.6%) | 6 (1.4%) | 6 (1.4%) | 9 (2.1%) | 8 (1.8%) | 0.54 |
| Ischaemic stroke | 74 (17.0%) | 96 (22.2%) | 105 (23.9%) | 118 (27.4%) | 121 (27.9%) | ＜0.0001 |
| Peripheral artery disease | 66 (15.2%) | 67 (15.5%) | 82 (18.7%) | 82 (19.1%) | 89 (20.6%) | 0.13 |
| Diabetes | 382 (87.8%) | 357 (82.4%) | 377 (85.9%) | 362 (84.2%) | 346 (79.9%) | 0.011 |
| Hyperuricaemia | 100 (23.0%) | 80 (18.5%) | 83 (18.9%) | 85 (19.8%) | 92 (21.2%) | 0.72 |
| Renal disease | 94 (21.6%) | 105 (24.2%) | 115 (26.2%) | 115 (26.7%) | 152 (35.1%) | ＜0.0001 |
| Chronic kidney disease | 250 (57.5%) | 272 (62.8%) | 272 (62.0%) | 298 (69.3%) | 302 (69.7%) | ＜0.0001 |
| Liver disease | 166 (38.2%) | 127 (29.3%) | 142 (32.3%) | 129 (30.0%) | 120 (27.7%) | 0.004 |
| COPD | 17 (3.9%) | 17 (3.9%) | 18 (4.1%) | 14 (3.3%) | 24 (5.5%) | 0.39 |
ACS: acute coronary syndrome; COPD: chronic obstructive pulmonary disease; FH: familial hypercholesterolaemia; Lp(a): Lipoprotein(a) levels.
a Cochran-Armitage trend test.
Variables potentially associated with high serum Lp(a) concentrations were evaluated with multivariate logistic regression analysis independently in each of the three groups. The findings are presented as Forest plots in Fig.4. The odds ratios (ORs) for being in the highest Lp(a) quintile did not differ from unity for age or gender in any of the three subgroups. Markers of impaired kidney function (reduced eGFR in Groups I (OR: 1.056 [95% confidence interval (CI) 1.025-1.089])and Group II (OR: 1.077 [95% CI 1.034-1.122]) and increased BUN in Group III (OR: 1.022 [95% CI 1.007-1.038]) were associated with an increased probability of being in the highest Lp(a) quintile, as was elevated LDL-C (OR: 1.085 [95% CI 1.024-1.150] in Group II and OR: 1.099 [95% CI 1.032-1.170] in Group III) and use of intensive statins (OR: 1.795 [95% CI 1.242-2.596] in Group II and OR: 1.972 [95% CI 1.316-2.956] in Group III) in Groups II and III and a diagnosis of unstable angina in Group II (OR: 1.858 [95% CI 1.275-2.706]). On the other hand, DM (in Group III) (OR: 0.558 [95% CI 0.343-0.908]) or elevated HbA1c (in Group I) (OR: 0.871 [95% CI 0.794-0.956]) was associated with a lower probability of being in the highest Lp(a) quintile, as was liver disease in Group I (OR: 0.708 [95% CI 0.515-0.975]). The covariates selected for Groups II and III were similar because of the high proportion of overlap between those two groups.
Data are expressed as odds ratios (with their 95% confidence intervals) for being in the fifth (highest) quintile of serum lipoprotein(a) levels.
The covariates presented in the Forest plot correspond to variables identified in the stepwise regression model process. Gender and age were included a priori as forced variables in all models. The following covariates were considered in the statistical model building: eGFR (per 5 mL/min/1.73m² incremental decrease), BUN (per 1 mg/dL incremental increase), LDL-C (per 10 mg/dL incremental increase), HDL-C (per 10 mg/dL incremental increase), triglycerides (per 10 mg/dL incremental increase), HbA1c (per 1 percentage point incremental increase), CAD (yes/no), dyslipidaemia (yes/no), angina pectoris (yes/no), unstable angina (yes/no), myocardial infraction (yes/no), recent ACS (yes/no), ischaemic stroke (yes/no), PAD (yes/no), aortic valve stenosis (yes/no), mitral valve stenosis (yes/no), aortic aneurism (yes/no), diabetes (yes/no), hyperuricaemia (yes/no), hypertension (yes/no), chronic heart failure (yes/no), COPD (yes/no), liver disease (yes/no), any standard statins (yes/no), any intensive statins (yes/no).
ACS: acute coronary syndrome; BUN: blood urea nitrogen; CAD: coronary artery disease; COPD: chronic obstructive pulmonary disease; eGFR: estimated glomerular filtration rate; HbA1c: glycated haemoglobin A1c; HDL-C: high density lipoprotein cholesterol; LDL-C: low density lipoprotein cholesterol; PAD: peripheral artery disease.
In the 2,170 patients of the study population, median serum lipid concentrations were 171 mg/dL [IQR: 146 – 197] for total cholesterol (TC), 94 mg/dL [IQR: 75 – 117] for LDL-C, 51 mg/dl [IQR: 42 – 62] for HDL-C and 112.5 mg/dL [IQR: 79 – 163] for total triglycerides. These concentrations were very similar across all Lp(a) quintiles except those for LDL-C (Fig.5).
HDL-C: high-density lipoprotein cholesterol; LDL-C: low-density lipoprotein cholesterol; Lp(a): lipoprotein(a); p-values for trend were presented. The lines represent the best-fit linear regression through the median values for the quintiles.
The rate of achievement of LDL-C goals by Lp(a) quintile in each of the three Groups is presented in Fig.6. In Group I (primary prevention – high risk), a similar proportion of patients achieved the goal of ＜120 mg/dL in all quintiles (ranging from 67.7% in Quintile 4 to 76.8% in Quintile 1). A similar proportion of patients across quintiles achieved the more stringent thresholds of ＜100 and ＜70mg/dL. In Group II, the proportions of patients achieving the goal of ＜100 mg/dL were similar, ranging from 64.8% in Quintile 3 to 72.1% in Quintile 1. However, the proportion meeting the more stringent target of ＜70 mg/dL decreased by more than a third across quintiles from 30.2% in Quintile 1 (but 33.5% in Quintile 2) to 19.0% in Quintile 5 (p-value for trend 0.007). This was particularly evident in Group III, where ＜70 mg/dL is the management goal, where the proportion decreased by a factor of two from 32.9% in Quintile 1 (but 36.0% in Quintile 2) to 16.3% in Quintile 5 (p-value for trend 0.003).
Data are presented by quintile of serum lipoprotein(a) level. For each group, the column in blue represents the low-density lipoprotein cholesterol (LDL-C) goal in each risk prevention group.
;%0A%09%09%09newWindow.document.open();%0A%09%09%09newWindow.document.write('<img src=%22./Graphics/32_64953_1.png%22>');%0A%09%09%09newWindow.document.close();%0A%09%09)
: LDL-C ＜70 mg/dL (goal for Group III);
;%0A%09%09%09newWindow.document.open();%0A%09%09%09newWindow.document.write('<img src=%22./Graphics/32_64953_2.png%22>');%0A%09%09%09newWindow.document.close();%0A%09%09)
: LDL-C ＜100 mg/dL (goal for Group II);
;%0A%09%09%09newWindow.document.open();%0A%09%09%09newWindow.document.write('<img src=%22./Graphics/32_64953_3.png%22>');%0A%09%09%09newWindow.document.close();%0A%09%09)
: LDL-C ＜120 mg/dL (goal for Group I).
p-values (Cochrane-Armitage trend test):
Group I: target ＜70 mg/dl: 0.09; target ＜100 mg/dl: 0.14; target ＜120 mg/dl: 0.05
Group II: target ＜70 mg/dl: 0.007; target ＜100 mg/dl: 0.37; target ＜120 mg/dl: 0.65
Group III: target ＜70 mg/dl: 0.003; target ＜100 mg/dl: 0.44; target ＜120 mg/dl: 0.85
Overall: target ＜70 mg/dl: 0.020; target ＜100 mg/dl: 0.26; target ＜120 mg/dl: 0.43.
Lipid-lowering treatments are described by quintile of serum Lp(a) concentrations in Table 3. Only standard statins, intensive statins and statins in combination with ezetimibe were evaluated. These were the only treatments prescribed to at least 50 patients in all study groups and quintiles combined. Overall, 162 patients (7.5%) were prescribed a standard statin, 623 patients (28.7%) an intensive statin, and 76 patients (3.5%) a statin in combination with ezetimibe. Nonetheless, 1,152 patients (53.1%) were prescribed no treatment. No clear trends in prescription patterns were observed across Lp(a) quintiles, except for intensive statins, for which prescription rates increased somewhat in Group I (from 15.8% in the first quintile to 22.8% in the fifth quintile) and in Group II (from 41.4% in the first quintile to 52.6% in the fifth quintile) but not in Group III.
| 1st Quintile | 2nd Quintile | 3rd Quintile | 4th Quintile | 5th Quintile | p-valueb | |
|---|---|---|---|---|---|---|
| Group I | N= 259 | N= 264 | N= 257 | N= 257 | N= 256 | |
| Standard statins | 21 (7.6%) | 28 (10.5%) | 20 (8.0%) | 25 (9.7%) | 17 (7.1%) | 0.79 |
| Intensive statins | 44 (15.8%) | 51 (19.1%) | 46 (18.5%) | 57 (22.1%) | 55 (22.8%) | 0.029 |
| Statins + ezetimibe | 3 (1.1%) | 3 (1.1%) | 2 (0.8%) | 2 (0.8%) | 0 (0.0%) | 0.15 |
| No treatmenta | 182 (65.5%) | 161 (60.3%) | 159 (63.9%) | 147 (57.0%) | 145 (60.2%) | 0.14 |
| Group II | N= 179 | N= 173 | N= 176 | N= 175 | N= 174 | |
| Standard statins | 6 (3.8%) | 15 (9.0%) | 10 (5.4%) | 14 (7.9%) | 6 (3.1%) | 0.55 |
| Intensive statins | 65 (41.4%) | 58 (34.9%) | 81 (44.0%) | 65 (36.5%) | 101 (52.6%) | 0.031 |
| Statins + ezetimibe | 12 (7.6%) | 9 (5.4%) | 11 (6.0%) | 11 (6.2%) | 23 (12.0%) | 0.10 |
| No treatment | 66 (42.0%) | 76 (45.8%) | 66 (35.9%) | 78 (43.8%) | 72 (37.5%) | 0.34 |
| Group III | N= 143 | N= 150 | N= 135 | N= 143 | N= 141 | |
| Standard statins | 6 (4.7%) | 13 (10.4%) | 3 (2.0%) | 12 (8.4%) | 4 (2.4%) | 0.22 |
| Intensive statins | 42 (33.1%) | 39 (31.2%) | 61 (40.9%) | 47 (32.9%) | 70 (41.7%) | 0.13 |
| Statins + ezetimibe | 12 (9.4%) | 7 (5.6%) | 9 (6.0%) | 11 (7.7%) | 22 (13.1%) | 0.15 |
| No treatment | 49 (38.6%) | 54 (43.2%) | 52 (34.9%) | 57 (39.9%) | 60 (35.7%) | 0.47 |
Standard statins correspond to fluvastatin, pravastatin or simvastatin and intensive statins to atorvastatin, rosuvastatin or pitavastatin.
a No lipid-lowering treatments prescribed within four weeks of the index date.
b Cochran-Armitage trend test
Two patients were prescribed a PCSK9 inhibitor, in combination with a statin, in Group I and the fifth quintile of serum Lp(a) measures. Nine patients received nicotinic acid derivatives distributed across all Lp(a) quintiles (four in Group I, five in Group II and four in Group III).
This study reported Lp(a) levels and cardiovascular morbidity in a large population of 2,170 patients fulfilling criteria for ASCVD prevention according to the JAS GL2017 19). In our population, the median Lp(a) serum concentration was 13.9 mg/dL. Target thresholds for serum Lp(a) of ＜30 mg/dL and ＜50 mg/dL have been proposed as being appropriate for Japanese populations6). A threshold of ≥ 70 mg/dL has been used in the Phase III Lp(a)HORIZON clinical trial22, 23), which corresponds to around the 75th percentile of Lp(a) levels observed in the large multinational Lp(a)HERITAGE study of Lp(a) levels in patients with a history of ASCVD24). The proportion of patients exceeding these thresholds in our study was 17.4% (30 mg/dL), 7.0% (50 mg/dL) and 3.4% (70 mg/dL). Our data on the distribution of Lp(a) are broadly consistent with a previous study of 342 patients with ACS undergoing percutaneous coronary intervention in Japan, which reported a median Lp(a) level of 16.0 mg/dl, with 26.3% of patients with Lp(a) ＞30 mg/dL and 10.5% had Lp(a) ＞50 mg/dL6). These values are somewhat lower than those reported in European populations25, 26); much higher values have been reported in African or African-American populations15, 24, 26).
To our knowledge, the testing rate of Lp(a) in the Japanese population has never been studied. Our sample of 2,170 patients with a documented measure of serum Lp(a) represents only a very small proportion (＜0.5%) of the ＞700,000 patients with risk factors for ASCVD identified in the MVD database. Measurement rates were lower (＜0.1%) in the considered ASCVD risk groups during the selection period from 2018 to 2021 presented in the previous paper19). This low testing rate may be partly due to a perception that the frequency of elevated Lp(a) in the Japanese population is relatively low15). However, similarly low levels of Lp(a) testing have also been reported in patients receiving lipid-lowering drugs for prevention of ASCVD in the United States (0.06%)27). A recent real-world study in the United States reported testing rates of 0.1% in the primary prevention setting and 0.03% in the secondary prevention setting, even though ≥ 28% of patients tested had serum Lp(a) levels ≥ 125 nmol/L28). This low rate may be due to the fact that, for many years, there was no consensus on the clinical significance of Lp(a), on standardization of assay method or on clinically-relevant cut-off thresholds16, 29); in addition, the availability of the measure was not anticipated to influence treatment decisions24). However, the utility of measuring serum Lp(a) is now recognised and recommended in practice guidelines16, 30, 31). For example, the 2022 JAS ASCVD prevention guidelines (JAS GL2022) state that Lp(a) level should be measured in patients with diagnosed dyslipidaemia (LDL-C ≥ 140 mg/dL, HDL-C ＜40 mg/dL, fasting triglycerides ≥ 150 mg/dL, or non-HDL-C ≥ 170 mg/dL)32). In our study, compared with patients who were not tested, tested patients typically had a history of ASCVD or presented major risk factors for ASCVD. For example, 14.8% of tested patients had a recent ACS compared with 2.9% in the untested group, and 59.5% of tested patients had dyslipidaemia compared with 29.5% in the untested patients. This might suggest that Lp(a) testing was prescribed once patients had developed overt cardiovascular disease rather than as part of a prevention strategy.
We observed that Lp(a) levels were higher in patients in Groups II and III (secondary prevention) than in Group I (primary prevention), which may reflect the well-characterised association between high circulating Lp(a) and incidence of ASCVD5). Further support for this is provided that the frequency of ASCVD, as well as chronic heart failure and CKD, rose with increasing serum Lp(a) levels. The strongest associations were observed for UA, whose frequency rose from 18.9% in the lowest Lp(a) quintile to 28.9% in the highest quintile, MI (from 10.1% to 18.7%) and recent ACS (from 15.4% to 24.5%). However, no obvious common threshold between quartiles could be identified where risk increases for all disorders. The proportion of patients with each disorder of interest was always highest in the highest quartile, whose lower limit was around 30 mg/dL in all three groups. These findings are consistent with numerous findings over the last thirty years indicating an association between elevated Lp(a) and cardiovascular mortality8, 11).
With respect to variables associated with elevated Lp(a), we did not observe any such association for age and gender. This is not consistent with a large worldwide study of the demographics of the distribution of Lp(a) levels24), which found that Lp(a) levels were higher in women than in men, as well as in subjects younger than 65 years, although the differences were relatively modest. Our failure to find such an association may represent a particularity of the Japanese population or reflect the fact that the majority (72.8%) of our patients were aged ≥ 65 years. Nonetheless, higher Lp(a) levels in women have also been described in an earlier Japanese study 33). On the other hand, we observed that Lp(a) levels were increased in patients with renal impairment (defined by reduced eGFR) and decreased in those with liver disease, and this is consistent with many previous findings34, 35). It has been suggested that the increase in serum Lp(a) levels in patients with renal impairment is due to reduced clearance of Lp(a)35, 36). The liver is the site of apolipoprotein(a) synthesis, so it is to be expected that Lp(a) levels are decreased in patients with liver injury35). The inverse association between DM (or elevated HbA1c) and Lp(a) levels has also been reported in several previous studies30, 37-39). The reason for this association is unclear, but may involve mutations or polymorphisms in genes important for the development of type 2 DM that also affect the expression of the gene encoding apolipoprotein(a), such as the mature-onset diabetes of the young (MODY) gene40). Data from large clinical trials indicate that certain statins, notably atorvastatin, may increase Lp(a) levels, perhaps by up-regulating apolipoprotein(a) gene expression41). Our observation of an association between intensive statin use and high Lp(a) would be consistent with these findings.
Given their risk profile, all these patients should receive lipid-lowering treatment according to JAS GL2017 and JAS GL2022 18, 32). These guidelines specify treatment goals for LDL-C stratified by risk. It should be noted that the JAS GL2022 32) guidelines set more stringent LDL-cholesterol targets for certain patient groups (for example in primary prevention in patients with diabetes and one other risk factor) than was the case in the 2017 guidelines, on which our definitions of risk categories were based. We observed that high serum levels of Lp(a) were associated with a reduced proportion of patients achieving the most stringent goal of LDL-C ＜70 mg/dL in the secondary prevention setting, with ＞65% of patients failing to achieve their goals. In Group III, the proportion of patients meeting this goal was two times lower for the fifth quintile of serum Lp(a) levels compared to the first quintile. In the primary prevention setting, the proportion of patients who achieved the treatment goal of LDL-C ＜120 mg/dL was not markedly different across serum Lp(a) quintiles.
No associations were observed between serum levels of Lp(a) and levels of cholesterol or triglycerides except LDL-C, as expected from its status as an independent risk factor for atherosclerosis14). High serum levels of Lp(a) were associated with higher LDL-C, which could be explained by the fact that Lp(a) contains cholesterol, which was measured as LDL-C42).
With respect to treatment, no lipid-lowering treatment prescription was documented for most patients (53.1%), contrary to JAS GL2022 32) and consistent with our previous study in the MDV database19). There was little evidence that Lp(a) levels influenced prescribing, with overall treatment rates being similar across Lp(a) quintiles for all ASCVD risk groups. There was some indication that intensive statins were used more frequently in patients in higher Lp(a) quintiles, although the difference was modest. Our study provided no evidence that lipid-lowering therapy was systematically intensified in patients with elevated serum Lp(a) to ensure that LDL-C treatment goals were achieved. However, this is the recommended management strategy in current European and North American practice guidelines30, 43). Although statins are not recognised to reduce Lp(a) levels, there is evidence that this can be achieved with PCSK9 inhibitors17, 44) or nicotinic acid derivatives45-47). In addition, PCSK9 inhibitors have been shown to provide a further reduction in ASCVD risk in patients with well-controlled LDL-C and elevated Lp(a)48). However, the present study indicates that these treatments are not widely offered to patients with elevated Lp(a) and a history of ASCVD or at risk of ASCVD in Japan.
Current European and North American guidelines recommend once-in-a-lifetime testing in certain patients with a family or a personal history of ASCVD30, 41). Given the low Lp(a) testing rate in Japan (＜1% of patients undergoing serum cholesterol testing), more widespread testing in patients at risk for ASCVD in Japan is no doubt merited. Given that Lp(a) levels are not modified by lifestyle measures and no Lp(a) lowering medications are available, the only therapeutic measure that could be recommended in patients with elevated Lp(a) would be to adopt a more intensive LDL-cholesterol-lowering treatment strategy30, 41).
The study presents a number of strengths and limitations. The strengths include the relatively large number of participants identified, particularly given the low Lp(a) testing rate in Japan and the possibility of comparing Lp(a) levels with other lipid parameters, clinical events and treatments prescribed. The limitations include the inclusion of patients from a limited set of MDV hospitals, which may limit the representativeness of the sample. In addition, the MDV database principally covers large acute-care hospitals, and patients who are prescribed lipid-lowering treatments in community clinics or by their home doctors may not be identified in this study. Thirdly, the prescription data provide no insight into patient adherence or to whether the medication was taken. The failure to meet LDL-C goals may be due to poor adherence rather than to treatment ineffectiveness. Also, we do not know the reason why patients had Lp(a) measured or not, which may influence the characteristics of patients with Lp(a) measured in this study. Finally, information was not available on the assay methods used, which may influence the interpretation of the data and comparison with other studies.
In this population of patients at risk for ASCVD, the rate of Lp(a) testing was very low, despite the high prevalence of a serum Lp(a) level over the cut-off of ≥ 50 mg/dl, especially among secondary prevention patients. The prevalence of coronary artery disease, PAD, and IS was higher in higher Lp(a) quintiles. Especially in the most at-risk patients (those in Group III with elevated Lp(a)), the achievement rate of the target LDL-C goal was low. These findings highlight the need for more widespread Lp(a) testing and more intense lipid-lowering treatment in patients with high Lp(a) in order to ensure that LDL-C goals are met and that ASCVD risk is thus effectively controlled.
The authors thank Salsabil Touzeni, Renata Majewska, Céline Quélen and Yoshie Onishi (Putnam) for data acquisition and analysis, and their contribution to the study design and interpretation of the data for preparing this manuscript. The authors also thank Dr. Adam Doble (Foxymed, Paris) for medical writing support.
This study was funded by Novartis Pharma K.K. Japan.
Emi Fujii, Yuri Takahashi, Mitsutoshi Toda, Kazuma Iekushi are employees of Novartis Pharma K.K. Junya Ako reports speaking honoraria from Amgen and Novartis. Shizuya Yamashita reports consulting fees from Hayashibara Co. and Immuno-Biological Laboratories Co., Ltd., Payment or honoraria for lectures, presentations, speakers’ bureaus, manuscript writing or educational events from Kowa, Novartis and Otsuka, and participation on a Data Safety Monitoring Board or Advisory Board from Novartis and Otsuka.
