If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Corresponding author. Division of Angiology, Department of Internal Medicine 2, Medical University of Vienna, Waehringer Guertel 18-20, 1090, Vienna, Austria.
The predictive capability of Lp(a) for cardiovascular mortality was evaluated over a long-term period in two cohorts of symptomatic peripheral artery disease (PAD).
•
Lp(a) was not associated with cardiovascular mortality in both cohorts.
•
No specific pattern of lesion site (iliacal, femoral, below the knee, multivessel) for endovascular repair was detected with elevated Lp(a) levels.
•
This data suggests that elevated Lp(a) does not impose an additional risk to patients with symptomatic PAD and high rates of statin treatment.
Abstract
Background and aims
Despite low LDL-C goals, the residual risk for further cardiovascular (CV) events in patients with peripheral artery disease (PAD) remains high. Lipoprotein (a) (Lp(a)) is a known risk factor for PAD incidence, but little is known regarding the outcome in patients with symptomatic PAD. Thus, this study investigates Lp(a) and CV mortality in PAD after endovascular repair.
Methods
A total of 1222 patients with PAD in two cohorts according to Lp(a) assay in nmol/L (n = 964, Lip-LEAD-A) or mg/dl (n = 258, Lip-LEAD-B) were followed up for 4.3 (IQR 3.0–5.6) or 7.6 (IQR 3.2–8.1) years. Lp(a) was measured before endovascular repair for either intermittent claudication (IC) or critical limb ischemia (CLI). Outcome information was obtained from the federal death registry.
Results
In Lip-LEAD-A, 141 CV-deaths occurred (annual calculated CV-death rate 3.4%), whereas 64 CV-deaths were registered in Lip-LEAD-B (annual calculated CV-death rate 3.3%). After adjustment for traditional CV risk factors Lp(a) was neither associated with outcome in Lip-LEAD-A (highest tertile HR 1.47, 95%CI [0.96–2.24]) nor in Lip-LEAD-B (highest tertile HR 1.34 [0.70–2.58]). Subanalyses for IC (HR 1.37 [0.74–2.55]; HR 1.10 [0.44–2.80], CLI (HR 1.55 [0.86–2.80], HR 3.01 [0.99–9.10]), or concomitant coronary artery disease (CAD; HR 1.34 [0.71–2.54]; HR 1.21 [0.46–3.17]) failed to show a significant association between Lp(a) and CV-mortality.
Conclusions
In this large-scale cohort of symptomatic PAD no association of elevated Lp(a) with CV mortality was found over a median observation period of 5 years. Thus, an even longer study including asymptomatic patients is warranted.
]. This particularly extends to patients with peripheral artery disease (PAD), a CV disease that can be seen as a manifestation of systemic and advanced atherosclerotic disease and burden. Lowering of LDL-C, the most relevant objective of secondary prevention in the lipid spectrum, was shown to reduce CV events in statin trials [
Low-density lipoprotein cholesterol lowering with evolocumab and outcomes in patients with peripheral artery disease: insights from the FOURIER trial (further cardiovascular outcomes research with PCSK9 inhibition in subjects with elevated risk).
] on lipid-lowering treatment, effects of lipoprotein particles heterogeneity on atherosclerotic disease progression, and thus morbidity and mortality, might be more relevant.
Lipoprotein(a) (Lp(a)), an LDL particle with apolipoprotein(a) attached to apolipoprotein B100, might be one candidate. Lp(a) bears several proatherogenic properties in advance to the LDL particles mediated effects such as selective binding of apo(a) to the extracellular matrix, and the amount of oxidized phospholipids, covalently bound to apo(a) [
]. Epidemiologic, cross-sectional, and prospective data, as well as Mendelian randomization studies, link Lp(a) to both morbidity and mortality in CV disease in primary and secondary prevention [
]. In two large European cohorts in the primary setting, the EPIC-Norfolk study and the Edinburgh Artery Study, Lp(a) levels were similarly associated with incidental PAD [
Novel risk factors for systemic atherosclerosis: a comparison of C-reactive protein, fibrinogen, homocysteine, lipoprotein(a), and standard cholesterol screening as predictors of peripheral arterial disease.
]. In the secondary prevention setting for patients with PAD, three studies reported a significant association between higher Lp(a) levels and major adverse limb events (MALE) or major adverse cardiac events (MACE) [
Association of serum lipoprotein (a) with the requirement for a peripheral artery operation and the incidence of major adverse cardiovascular events in people with peripheral artery disease.
Association of serum lipoprotein (a) with the requirement for a peripheral artery operation and the incidence of major adverse cardiovascular events in people with peripheral artery disease.
Association of serum lipoprotein (a) with the requirement for a peripheral artery operation and the incidence of major adverse cardiovascular events in people with peripheral artery disease.
] showed data for a combined cohort of patients with abdominal aortic aneurysm and lower extremity artery disease. The other studies that reported MACE or MALE had low event rates and limited total numbers of included patients [
]. To our knowledge, no further study evaluated CV mortality and Lp(a) in a specific PAD cohort.
Thus, this study evaluates a possible association of Lp(a) in patients with PAD and long-term CV mortality after endovascular repair either for critical limb ischemia (CLI) or intermittent claudication (IC) in a single-center retrospective cohort.
2. Patients and methods
2.1 Patients and study design
The Vienna Lip-LEAD (LIPids in Lower Extremity Artery Disease) study was designed to evaluate lipid profiles and lipid treatment goals in patients with peripheral artery disease undergoing endovascular repair. The study was approved by the institutional ethics committee and follows the Declaration of Helsinki and contemporary Good Clinical Practice guidelines. All patients treated with endovascular repair of the lower limb arteries at the inpatient department of the Division of Angiology, Medical University of Vienna, were included retrospectively from 1/Jan/2013 to 31/Dec/2018. Only patients with a first admission during this time at the inpatient department were analyzed in this study; further admissions of the same patient for endovascular repair or any other reason were not included. PAD is used as the term for lower extremity artery disease (LEAD) in this study. Patients with missing values of Lp(a) or covariates, an eGFR<15 mg/dl, and prior organ transplantation were excluded from this analysis (flow-chart, Fig. 1). Endovascular repair was indicated either by symptomatic PAD (Fontaine stage II-IV) or in Fontaine stage I in the case of critical stenosis of a previously implanted bypass graft or stent. For diagnostic work-up, all patients received both a hemodynamic assessment (oscillometry and ankle-brachial index measurement) and further imaging (either ultrasonography or contrast-enhanced CT/MR scans) for endovascular repair planning. Ankle-brachial index (ABI) measurement, as defined by the ratio of the brachial and ankle blood pressure according to the TASC II criteria [
], was performed by experienced and specially trained staff with Doppler sonography (ELCAT, Wolfratshausen, Germany). Ultrasonography was performed by specially trained technicians at the Division of Angiology. CT- or MR-scans were performed at the Department of Radiology, Medical University of Vienna, according to standard examination protocols.
Routine laboratory parameters were measured on the day of admission, including kidney function, liver function, and a lipid profile (total cholesterol, HDL-C, and Lp(a)) by the laboratory department. LDL-C was calculated according to the Friedewald formula. Lp(a) was included in this routine laboratory testing and measured on a Roche Diagnostics platform. Lp(a) was measured in mg/dL until 17/Feb/2014 and after assay change of the central laboratory in nmol/L. Since mg/dl and nmol/L levels of Lp(a) cannot simply be transformed due to different molar masses according to KIV-2 kringle repeats [
Quantifying atherogenic lipoproteins: current and future challenges in the era of personalized medicine and very low concentrations of LDL cholesterol. A consensus statement from eas and eflm.
], and no assay comparison was performed, patients were separately analyzed in a mg/dL (Lip-LEAD-B) cohort and a nmol/L (Lip-LEAD-A) cohort.
Medical history including previous myocardial infarction, stroke, PAD, cerebrovascular disease/carotid stenosis, diabetes mellitus, arterial hypertension, and smoking history was recorded from available medical records. Data on medication was registered from the discharge letter for endovascular repair.
Information on patients’ survival was obtained via queries to the central death registry of the federal department of statistics (Statistik Austria) on 31/Aug/2021. ICD-codes of the I category were categorized as CV-death.
Medical artwork was used from Servier medical arts under a creative commons license (CCA 3.0 unported license).
2.2 Definition of parameters
Peripheral artery disease (PAD), was diagnosed by ABI measurement with a resting ABI <0.9 and clinically graded according to the Fontaine classification. The grade of stenosis was assessed by imaging either with ultrasonography, CT, or MR angiography. Fontaine stage II (both IIa and IIb) was defined as intermittent claudication (IC) and both Fontaine stage III (resting pain) and IV (ulcers) were defined as critical limb ischemia (CLI).
Diabetes mellitus was defined according to the current ADA guidelines of 2021 [
] by either an HbA1c > 6.5 rel.%, previously known diabetes mellitus, and/or previous antidiabetic medication. In the case of SGLT2i, no other indication besides diabetes mellitus type 2 existed in the country of this study until the end of inclusion (December 2018).
Coronary artery disease was defined as previously known myocardial infarction, percutaneous coronary intervention, or ischemic cardiomyopathy.
Hypertension was defined according to medical records (discharge summaries) and verified by the intake of antihypertensive treatment.
Statin doses were calculated as atorvastatin equivalency doses for better comparison according to Naci et al. [
Dose-comparative effects of different statins on serum lipid levels: a network meta-analysis of 256,827 individuals in 181 randomized controlled trials.
Lesion sites were subjected to four categories: 1 iliac region including kissing stenting of the common iliac artery; 2 femoral-popliteal regions; 3 below the knee repair; 4 multiple regions (multisite).
2.3 Statistical analysis
The entire statistical analysis was performed with SPSS 27.0 (SPSS Inc. Chicago, IL, USA). Figures were drawn with GraphPad Prism 9.0 (GraphPad Software Inc., San Diego, CA, USA). Data are presented as mean ± standard deviation (SD) or median and percentiles (25th, 75th), as appropriate. A two-sided p-value <0.05 was defined as statistically significant. Student's t-tests or nonparametric equivalents and chi-square tests were applied, as appropriate. Lp(a) levels were divided into tertiles for better comparison. Kaplan-Meier analyses were assessed with a log-rank test and performed with tertiles, respectively. Cox regression analyses were performed categorically with the same grouping. A multivariable-adjusted model was performed with traditional CV risk factors (age, sex, T2DM, LDL-C, HDL-C, arterial hypertension, ezetimibe usage, BMI, Fontaine stage, and CKD-EPI). To evaluate a possible non-linear association of Lp(a) with CV-death, this model was included as a restricted cubic spline with 3 knots in the model. Therefore, R Version 4.0.3 (R Foundation for Statistical Computing, Vienna, Austria) and packages “rms”, “survminer” and “ggplot2” were utilized. Subanalyses were performed with Cox-regression analyses for intermittent claudication, critical limb ischemia, the combination of coronary artery disease and PAD. Subanalyses for LDL-C control <70 mg/dl and <55 mg/dl were only performed in Lip-LEAD-A due to the small sample number as well as overfitting and underpowering issues in Lip-LEAD-B.
3. Results
3.1 Baseline characteristics
3.1.1 LIP-LEAD-A (nmol/L) group
A total of 964 patients were included in the final analysis and divided into tertiles. Traditional CV risk factors (age, T2DM, hyperlipidemia, smoking) were equally distributed between the tertiles. However, a significantly lower systolic blood pressure was found in the highest vs. the lowest tertile (p = 0.033). Furthermore, a steady significant increase in the percentage of women in higher tertiles was found, ranging from 32.2% in the lowest group to 48.3% in the highest group (overall p<0.001, 1 vs. 3 p<0.001). BMI was higher in the lower tertiles, but only a trend was observed between the highest vs. the lowest tertile overall p = 0.044, 1 vs. 3 p = 0.013). Similar to most traditional CV risk factors, preexisting CV disease, including previously known PAD, stroke, myocardial infarction, CAD, and atrial fibrillation, were equally distributed between the groups. Regarding laboratory values, significantly higher levels of LDL-C (overall p<0.001, 1 vs. 3 p<0.001), HDL-C (overall p = 0.012, 1 vs. 3 p = 0.003), and serum total cholesterol (overall p<0.001, 1 vs. 3 p<0.001) were found in the highest tertile. CV secondary prevention treatment was similar in all groups apart from ezetimibe usage. The highest group had significantly higher prescription rates of ezetimibe while overall the percentage was low (6.6% lowest vs. 11.8% highest tertile, overall p = 0.025, 1 vs. 3 p = 0.021). A detailed overview can be seen in Table 1.
Table 1Baseline characteristics of Lip-LEAD-A.
1st Tertile
2nd Tertile
3rd Tertile
p-value
1 vs. 3
n = 320
n = 323
n = 321
Lipoprotein (a) nmol/l
7 (7, 8)
22 (16, 34)
180 (108, 230)
<0.001
<0.001
Age (yrs)
69 ± 10
68 ± 12
70 ± 11
0.097
0.529
Female Sex n(%)
103 (32.2%)
123 (38.1%)
155 (48.3%)
<0.001
<0.001
Systolic Blood Pressure mmHg
148 ± 24
145 ± 24
139 ± 19
0.088
0.033
Diastolic Blood Pressure mmHg
77 ± 10
77 ± 12
75 ± 12
0.335
0.516
BMI kg/m2
27.4 ± 5.4
27.0 ± 4.7
26.4 ± 4.9
0.044
0.013
Fontaine stage n(%)
Fontaine 1/2
234 (73.4%)
242 (74.9%)
229 (71.3%)
0.825
0.838
Fontaine 3/4
85 (26.6%)
81 (25.1%)
92 (28.7%)
Previous PAD n(%)
97 (30.3%)
106 (32.8%)
99 (30.8%)
0.770
0.885
Ankle-Brachial-Index
0.57 ± 0.20
0.59 ± 0.19
0.55 ± 0.19
0.113
0.217
Toe-Brachial-Index
0.58 ± 0.23
0.59 ± 0.22
0.55 ± 0.20
0.108
0.072
T2DM n(%)
138 (43.1%)
135 (41.8%)
134 (41.7%)
0.923
0.724
Art. Hypertension n(%)
281 (87.8%)
292 (90.4%)
290 (90.3%)
0.474
0.304
Hyperlipidemia n(%)
269 (84.1%)
283 (87.6%)
278 (86.6%)
0.407
0.363
Active Smoking n(%)
103 (32.2%)
110 (34.1%)
116 (36.1%)
0.135
0.067
CAD n(%)
94 (29.4%)
107 (33.1%)
105 (32.7%)
0.535
0.361
Stroke n(%)
26 (8.1%)
24 (7.4%)
28 (8.7%)
0.834
0.785
Myocardial Infarction n(%)
36 (11.3%)
56 (17.3%)
43 (13.4%)
0.078
0.409
Atr. Fibrillation n(%)
37 (11.6%)
38 (11.8%)
34 (10.6%)
0.882
0.695
Serum Creatinine mg/dl
1.06 ± 0.40
1.06 ± 0.41
1.04 ± 0.42
0.724
0.430
eGFR (CKD-EPI) ml/min/1.73m2
85.1 ± 25.1
87.0 ± 27.1
87.1 ± 26.4
0.540
0.319
LDL-C mg/dl
82.5 ± 34.2
90.0 ± 37.3
93.1 ± 39.7
0.001
<0.001
HDL-C mg/dl
48 ± 16
50 ± 17
53 ± 19
0.012
0.003
HbA1c rel.%
6.3 ± 1.1
6.4 ± 1.3
6.2 ± 1.1
0.142
0.655
Triglycerides mg/dl
127 (98, 177)
121 (90, 162)
125 (89, 172)
0.221
0.337
Serum total Cholesterol mg/dl
161 ± 41
168 ± 45
174 ± 46
0.001
<0.001
C-reactive Protein mg/dl
0.28 (0.13, 0.67)
0.35 (0.15, 1.05)
0.34 (0.12, 0.83)
0.105
0.338
ACEi n(%)
135 (42.2%)
141 (43.7%)
107 (33.3%)
0.015
0.021
AT2 n(%)
118 (36.9%)
116 (35.9%)
138 (43.0%)
0.136
0.114
Beta-Blocker n(%)
154 (48.1%)
160 (49.5%)
159 (49.5%)
0.919
0.721
Antihypertensive n(%)
266 (83.1%)
257 (79.6%)
261 (81.3%)
0.512
0.548
ASS n(%)
271 (84.7%)
285 (88.2%)
284 (88.5%)
0.276
0.160
Statin n(%)
275 (85.9%)
295 (91.3%)
281 (87.5%)
0.092
0.550
Statin class usage n(%)
Simvastatin
149 (46.6%)
166 (51.4%)
154 (48.0%)
0.686
0.980
Rosuvastatin
39 (12.2%)
39 (12.1%)
36 (11.2%)
Atorvastatin
81 (25.3%)
84 (26.0%)
86 (26.8%)
Fluvastatin
5 (1.6%)
3 (0.9%)
4 (1.2%)
Lovastatin
1 (0.3%)
3 (0.9%)
1 (0.3%)
Ezetimibe n(%)
21 (6.6%)
22 (6.8%)
38 (11.8%)
0.025
0.021
Atorvastatin Equivalency Dosages mg
33.4 ± 25.7
34.6 ± 25.7
33.6 ± 25.8
0.807
0.931
Patients are presented in tertiles according to their Lp(a) nmol/L level. BMI body mass index. Data are shown as mean ± standard deviation or median and interquartile range, as applicable. p < 0.05 (two-sided) was considered statistically significant.
A total of 258 patients were included in the final analysis of this group. Patients were divided into tertiles of Lp(a) for better comparison. Similar to the nmol/L group traditional risk factors were equally distributed between the tertiles and no statistically significant difference in sex and BMI was found. Furthermore, previous CV diseases were equally distributed between the tertiles, and percentages of medication for secondary prevention did not differ between the groups. A detailed overview can be seen in Table 2. No significant increase in HDL-C, LDL-C, or TC was seen (in difference from the first group).
Table 2Baseline characteristics of Lip-LEAD-B.
1st Tertile
2nd Tertile
3rd Tertile
p-value
1 vs. 3
n = 85
n = 89
n = 84
Lipoprotein (a) mg/dl
3 (3,5)
16 (11, 22)
68 (52, 108)
<0.001
<0.001
Age (yrs)
68 ± 12
70 ± 10
68 ± 11
0.550
0.927
Female Sex n(%)
29 (34.1%)
27 (30.3%)
32 (38.1%)
0.561
0.590
BMI kg/m2
26.2 ± 4.6
26.4 ± 4.6
26.8 ± 4.4
0.386
0.395
Fontaine stage n(%)
Fontaine 1/2
64 (75.3%)
68 (76.4%)
57 (67.9%)
0.423
0.360
Fontaine 3/4 CLI
21 (24.7%)
21 (23.6%)
27 (32.1%)
Previous PAD n(%)
37 (43.5%)
42 (47.2%)
46 (54.8%)
0.329
0.144
Ankle-Brachial-Index
0.61 ± 0.21
0.55 ± 0.21
0.58 ± 0.15
0.602
0.560
Toe-Brachial-Index
0.52 ± 0.17
0.51 ± 0.13
0.34 ± 0.2
0.107
0.093
T2DM n(%)
32 (37.6%)
41 (46.1%)
37 (44%)
0.506
0.397
Art. Hypertension n(%)
82 (96.5%)
78 (87.6%)
80 (95.2%)
0.046
0.688
Hyperlipidemia n(%)
79 (92.9%)
79 (88.8%)
77 (91.7%)
0.611
0.756
Active Smoking n(%)
33 (38.8%)
35 (39.3%)
29 (34.5%)
0.776
0.562
CAD n(%)
24 (28.2%)
31 (34.8%)
34 (40.5%)
0.246
0.094
Stroke n(%)
11 (12.9%)
10 (11.2%)
9 (10.7%)
0.894
0.654
Myocardial Infarction n(%)
9 (10.6%)
17 (19.1%)
19 (22.6%)
0.105
0.035
Atr. Fibrillation n(%)
6 (7.1%)
13 (14.6%)
6 (7.1%)
0.153
0.983
Serum Creatinine mg/dl
1.0 ± 0.4
1.1 ± 0.52
1.1 ± 0.5
0.164
0.226
eGFR (CKD-EPI) ml/min/1.73m2
89.9 ± 25.1
81.9 ± 27.1
85.9 ± 27.4
0.146
0.329
LDL-C mg/dl
88.1 ± 39.1
88.9 ± 34.7
93.4 ± 34.2
0.600
0.360
HDL-C mg/dl
51 ± 16
50 ± 17
50 ± 18
0.841
0.832
HbA1c rel.%
6.2 ± 1.0
6.3 ± 1.1
6.3 ± 1.1
0.509
0.243
Triglycerides mg/dl
133 (90, 179)
129 (92, 185)
124 (94, 171)
0.865
0.586
Serum total Cholesterol mg/dl
169 ± 44
169 ± 46
175 ± 42
0.642
0.383
C-reactive Protein mg/dl
0.24 (0.14, 0.61)
0.31 (0.12, 0.65)
0.32 (0.12, 1.07)
0.675
0.386
ACEi n(%)
40 (47.1%)
33 (37.1%)
35 (41.7%)
0.410
0.143
AT2 n(%)
30 (35.3%)
31 (34.8%)
32 (38.1%)
0.891
0.706
Beta-Blocker n(%)
38 (44.7%)
44 (49.4%)
46 (54.8%)
0.425
0.191
Antihypertensive n(%)
66 (77.6%)
75 (84.3%)
71 (84.5%)
0.412
0.254
ASS n(%)
78 (91.8%)
81 (91%)
80 (95.2%)
0.529
0.360
Statin n(%)
70 (82.4%)
80 (89.9%)
75 (89.3%)
0.260
0.197
Statin class usage n(%)
Simvastatin
29 (34.1%)
29 (32.6%)
23 (27.4%)
0.126
0.168
Rosuvastatin
6 (7.1%)
8 (9.0%)
16 (19.0%)
Atorvastatin
33 (38.8%)
36 (40.4%)
35 (41.7%)
Fluvastatin
1 (1.2%)
4 (4.5%)
1 (1.2%)
Lovastatin
1 (1.2%)
3 (3.4%)
0 (0%)
Ezetimibe n(%)
3 (3.5%)
2 (2.2%)
8 (9.5%)
0.068
0.114
Atorvastatin Equivalency Dosages mg
31.82 ± 24.51
35.59 ± 27.02
28.49 ± 19.88
0.153
0.334
Patients are presented in tertiles according to their Lp(a) mg/dl level. BMI body mass index. Data are shown as mean ± standard deviation or median and interquartile range, as applicable. p<0.05 (two-sided) was considered statistically significant.
During the observation time (4.3 years, IQR 3.0–5.6) a total of 326 fatal events occurred (overall death rate 33.8%, calculated annual death rate 7.8%). Furthermore, a total of 141 CV deaths were observed (overall CV-death rate 14.6%, calculated annual death rate 3.4%). In Kaplan-Meier curves no significant difference could be seen between the tertiles for CV-death (log-rank p = 0.244, Fig. 2A) or all-cause death (log-rank p = 0.654).
Fig. 2(A) KM curves for CV-death in Lip-LEAD-A for tertiles according to Lp(a) levels.
No significant event rate was seen between tertiles (log-rank p = 0.244), (B) KM curves for CV-death in Lip-LEAD-B. for tertiles according to Lp(a) levels. No significant event rate was seen between tertiles (log-rank p = 0.321).
The functional form between nmol/L Lp(a) and log hazards was further explored using restricted cubic splines with 3 knots. No significant deviation from linearity was observed (Supplemental Fig. 1).
Log transformed Lp(a) levels were subjected to univariate and multivariable-adjusted Cox-regression analyses (age, sex, T2DM, LDL-C, HDL-C, arterial hypertension, ezetimibe usage, BMI, Fontaine stage, and CKD-EPI). Both the univariate (hazard ratio 1.07, 95% confidence interval 0.95–1.21) and the multivariable-adjusted model (HR 1.10, 95% CI 0.97–1.24) did not show an association for CV-death for an increase of one-unit log transformed Lp(a). Furthermore, patients above the threshold of 175 nmol/L Lp(a) were not associated with CV death in the same continuous analyses (crude HR 1.00, 95% HR 0.99–1.00; multivariable model HR 1.00, 95% HR 0.99–1.01).
Lp(a) tertiles were further subjected to the same Cox-regression model. Neither in crude (highest tertile HR 1.29, 95% CI 0.86–1.95) nor in multivariable-adjusted models (highest tertile HR 1.47, 95% CI 0.96–2.24) was a significant association seen for CV death (Table 3).
Table 3Cox-Regression analyses for CV-death in Lip-LEAD-A shown in tertiles. Data are presented as Hazard ratio and 95% confidence interval. Models were adjusted first for age and sex and further multivariable adjusted for traditional CV risk factors (T2DM, LDL-C, HDL-C, arterial hypertension, ezetimibe usage, BMI, Fontaine stage and CKD-EPI).
When stratified according to controlled LDL-C (<70 mg/dl; highest tertile HR 0.99, 95% CI 0.50–1.96) or optimal LDL-C (<55 mg/dl; highest tertile HR 0.98, 95% CI 0.33–2.85) no significant associations for CV-death were found in multivariable adjusted Lp(a) tertile analyses (Table 3).
In a subanalysis for patients with endovascular repair due to intermittent claudication (Fontaine stage II; highest tertile HR 1.37, 95% CI 0.74–2.55) vs. critical limb ischemia (Fontaine stage III, IV; highest tertile HR 1.55, 95% CI 0.86–2.80) Lp(a) groups were not associated with CV-death as well in multivariable models (Table 3).
Likewise, in the subcohort of patients with concomitant coronary artery disease (highest tertile 1.34 HR, 95% CI 0.71–2.54) no association was found in multivariable models (Table 3).
3.2.2 LIP-LEAD-B (mg/dL) group
During the observation time of the mg/dl group (median 7.6 years, IQR 3.2–8.1) a total of 123 fatal events were registered (overall death rate 47.7%, calculated annual death rate 6.3%). Furthermore, a total of 64 CV-death were registered (overall CV death rate 24.8%, calculated annual death rate 3.3%. In the Kaplan-Meier analysis no association with overall mortality (log-rank p = 0.883) or CV-death was found (log-rank p = 0.321, Fig. 2B).
The functional form between mg/dl Lp(a) and log hazards was further explored using restricted cubic splines with 3 knots. No major deviation from linearity was observed (Supplemental Fig. 2).
Both univariate and multivariable Cox-regression analyses did not show an association for CV-death (HR 1.05 (95%CI 0.86–1.30); HR 1.08 (95% CI 0.86–1.34) for an increase of one-unit log-transformed Lp(a) levels. Similar, in categorical analyses for tertiles of Lp(a) no associations both in univariate (highest tertile HR 1.31, 95% CI 0.70–2.43) and multivariable-adjusted analyses (highest tertile HR 1.34, 95% CI 0.70–2.58) were seen (Table 4). Furthermore, in subanalyses for intermittent claudication (Fontaine stage II, highest tertile HR 1.10, 95% CI 0.44–2.80), critical limb ischemia (Fontaine stage III, IV, highest tertile HR 3.01, 95% CI 0.99–9.10), and a subcohort of concomitant CAD (highest tertile HR 1.21, 95% CI 0.46–3.17) no association was found (Table 4).
Table 4Cox-Regression analyses for CV-death in Lip-LEAD-B shown in tertiles. Data are presented as Hazard ratio and 95% confidence interval. Models were adjusted first for age and sex and further multivariable adjusted for traditional CV risk factors (T2DM, LDL-C, HDL-C, arterial hypertension, ezetimibe usage, BMI, Fontaine stage and CKD-EPI).
To evaluate a typical pattern of lesions with high Lp(a) levels, locations of endovascular repair were categorized into 4 categories: iliac, femoral including popliteal artery, below the knee, and multiple region repair. Lp(a) levels did not significantly differ between those four sites (p = 0.881, Fig. 3A).
Fig. 3(A) Lp(a) levels in Lip-Lead-A and lesion site. No significant difference of means between iliacal, femoral, below the knee, or multivessel revascularization was seen (overall p = 0.881). (B) Lp(a) in Lip-LEAD-B and lesion site. No significant difference of means between iliacal, femoral, below the knee, or multivessel revascularization was seen (overall p = 0.129).
The endovascular repair site of the mg/dL group was subjected to the same analysis as the nmol/l group. No significant difference was seen (p = 0.129, Fig. 3B).
4. Discussion
This study evaluates Lp(a) and its association with all-cause and CV-death in patients with symptomatic PAD after endovascular repair in two separated cohorts according to the assay measurement method of Lp(a). Neither the Lip-LEAD-A (nmol/L) nor the Lip-LEAD-B (mg/dL) cohort showed a significant association with Lp(a). Furthermore, sub-analyses for LDL-C control, only IC, only CLI, or concomitant CAD did not show any significant association with long-term outcome. Likewise, no significant differences in Lp(a) levels and target lesion site (iliac, femoral, BTK, or multisite repair) were found in Lip-LEAD-A or Lip-LEAD-B.
The evidence of a causal relationship between Lp(a) and a markedly increased risk of CV disease is large. Since the first Mendelian randomization study, linking Lp(a) to CV disease [
Associations between lipoprotein(a) levels and cardiovascular outcomes in black and white subjects: the Atherosclerosis Risk in Communities (ARIC) Study.
]. However, results in the secondary prevention setting do not support a clear-cut association. A meta-analysis of O'Donoghue et al. evaluated Lp(a) and outcome as well as MACE in three large RCTs (PEACE, CARE, and PROVE IT-TIMI 22) [
]. This initial analysis failed to show a significant outcome result. A positive outcome was reported only after a second step pooled meta-analysis for four additional RCTs and three observational studies. Yet, the heterogeneity of this meta-analysis was high. Furthermore, all RCTs of this meta-analysis included only CAD disease patients.
In PAD, to our knowledge, only one study by Golledge et al. evaluated mortality in the secondary prevention [
Association of serum lipoprotein (a) with the requirement for a peripheral artery operation and the incidence of major adverse cardiovascular events in people with peripheral artery disease.
] but did not report any association. However, the patients of this study had either atherosclerotic PAD or abdominal aortic aneurysms and no specific data for the respective diseases was given. While a recent study linked MALE to high Lp(a) after femoral artery endarterectomy [
Association of serum lipoprotein (a) with the requirement for a peripheral artery operation and the incidence of major adverse cardiovascular events in people with peripheral artery disease.
] linked high Lp(a) need for revascularization in patients with PAD, our study did not find any specific pattern of an affected lesion site.
A major difference between the profound signal of a relationship between high Lp(a) in the primary versus secondary setting can be seen in LDL-C levels and statin treatment. In the aforementioned meta-analyses of the CARE, PEACE and PROVE-IT TIMI 22 study [
] LDL-C and statin treatment percentages were higher than in the usual primary prevention setting. This might have led to a diminished effect of Lp(a) in these cohorts, which can also be seen in our Lip-LEAD cohort with even higher rates of statin treatment and better LDL-C control. Likewise, in the primary prevention setting, both the Women's Health study [
Lipoprotein(a), measured with an assay independent of apolipoprotein(a) isoform size, and risk of future cardiovascular events among initially healthy women.
Apolipoprotein (a) size and lipoprotein (a) concentration and future risk of angina pectoris with evidence of severe coronary atherosclerosis in men: the Physicians' Health Study.
] only showed an association with Lp(a) and outcome (MACE) in participants with high LDL-C (>121–160 mg/dl). Furthermore, since Lp(a) is genetically determined and thus elevated at birth, the observation period in our cohorts is short in comparison to the lifelong elevated Lp(a).
However, an “index event bias” as defined by Dahabreh and Kent cannot be ruled out by the null finding of this study [
]. Index event bias describes a paradoxical null finding with a risk factor after selection for this risk factor for recurrent events in a specific patient cohort. A similar index event bias can be seen with the “smoker paradox” that describes a beneficial effect of smoking compared to non-smoking after myocardial infarction [
Evaluation of paradoxic beneficial effects of smoking in patients receiving thrombolytic therapy for acute myocardial infarction: mechanism of the "smoker's paradox" from the GUSTO-I trial, with angiographic insights. Global Utilization of Streptokinase and Tissue-Plasminogen Activator for Occluded Coronary Arteries.
Another major difference of our cohort was that only symptomatic patients either with IC or CLI were analyzed. Previous data show that more than 50% of patients with PAD are, in fact, asymptomatic (Fontaine stage I) [
The risk of disease progression in peripheral arterial disease is higher than expected: a meta-analysis of mortality and disease progression in peripheral arterial disease.
]. It seems likely that symptomatic patients feature additive life-limiting risk factors, such as infections and limited physical activity both by IC or amputations aside from higher rates of other CV events [
]. Latter mentioned problems are associated with reduced life expectancy but are naturally not as prevalent in asymptomatic as in those without symptoms. Thus, a dedicated study on Lp(a) in asymptomatic PAD would be of high interest.
Nevertheless, our data suggest that Lp(a) levels are not associated with fatal CV outcome in a large cohort of patients with symptomatic PAD. This finding aligns with the aforementioned studies with adequate LDL-C control in other CV diseases. However, the upper limit of confidence intervals for the categorical analyses on CV-death in both cohorts is higher than 1.5 and a trend for higher HR with higher tertiles is present in some analyses. Thus, a moderate association cannot be ruled out by this study. Since outcome RCTs on Lp(a) lowering in the secondary prevention setting are currently conducted, results of these highly anticipated studies might bring more clarity into this field of uncertainty.
This study has several limitations. Firstly, despite a two-cohort approach, the study was only performed at a single center. Secondly, treatment in a tertiary care hospital might lead to a selection bias. Thirdly, since symptomatic PAD accounts for less than half of the PAD prevalence [
] worldwide, an evaluation of only symptomatic patients might present another selection bias. Fourthly, Lp(a) was measured in an “acute-phase” reaction previous to endovascular repair, at least for CLI, which could have obscured baseline Lp(a) levels. Fifthly, no apo(a) isoform size and SNP analyses were performed in this cohort considering Lp(a). Sixthly, no outcome evaluation on MACE or MALE was performed.
However, this study has several strengths to consider. Firstly, the sample size in comparison to similar PAD studies is large. Secondly, control of traditional risk factors is high in comparison to other cohorts. Thirdly, follow-up time is extensively long and only the hard outcome of fatal CV events was included in this analysis.
In conclusion, this study does not demonstrate an association with CV-death or a specific pattern of lesion site with Lp(a) in patients with symptomatic PAD. However, while this is in contrast to most studies on the secondary prevention of CAD and a large proportion of patients with PAD remains asymptomatic, further analyses, especially in the latter cohort and extended long-term follow-up studies are warranted.
CRediT authorship contribution statement
Bernhard Zierfuss: Conceptualization, Methodology, Validation, Formal analysis, Investigation, Resources, Data curation, Writing – original draft, Visualization. Clemens Höbaus: Data curation, Methodology, Validation, Resources, Data curation, Writing – review & editing, Visualization, Supervision. Anna Feldscher: Investigation, Writing – review & editing. Antonia Hannes: Investigation, Writing – review & editing. Daniel Mrak: Formal analysis, Validation, Writing – review & editing. Renate Koppensteiner: Resources, Writing – review & editing. Herbert Stangl: Validation, Writing – review & editing. Gerit-Holger Schernthaner: Supervision, Resources, Project administration, Writing – review & editing, Supervision.
Declaration of competing interests
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Appendix A. Supplementary data
The following are the Supplementary data to this article:
Low-density lipoprotein cholesterol lowering with evolocumab and outcomes in patients with peripheral artery disease: insights from the FOURIER trial (further cardiovascular outcomes research with PCSK9 inhibition in subjects with elevated risk).
Novel risk factors for systemic atherosclerosis: a comparison of C-reactive protein, fibrinogen, homocysteine, lipoprotein(a), and standard cholesterol screening as predictors of peripheral arterial disease.
Association of serum lipoprotein (a) with the requirement for a peripheral artery operation and the incidence of major adverse cardiovascular events in people with peripheral artery disease.
Quantifying atherogenic lipoproteins: current and future challenges in the era of personalized medicine and very low concentrations of LDL cholesterol. A consensus statement from eas and eflm.
Dose-comparative effects of different statins on serum lipid levels: a network meta-analysis of 256,827 individuals in 181 randomized controlled trials.
Associations between lipoprotein(a) levels and cardiovascular outcomes in black and white subjects: the Atherosclerosis Risk in Communities (ARIC) Study.
Lipoprotein(a), measured with an assay independent of apolipoprotein(a) isoform size, and risk of future cardiovascular events among initially healthy women.
Apolipoprotein (a) size and lipoprotein (a) concentration and future risk of angina pectoris with evidence of severe coronary atherosclerosis in men: the Physicians' Health Study.
Evaluation of paradoxic beneficial effects of smoking in patients receiving thrombolytic therapy for acute myocardial infarction: mechanism of the "smoker's paradox" from the GUSTO-I trial, with angiographic insights. Global Utilization of Streptokinase and Tissue-Plasminogen Activator for Occluded Coronary Arteries.
The risk of disease progression in peripheral arterial disease is higher than expected: a meta-analysis of mortality and disease progression in peripheral arterial disease.