Table 4 Summary of studies in secondary prevention (14 studies)
From: The relationship between Lp(a) and CVD outcomes: a systematic review
Study Details | Analysis Methods | Summary of findings |
|---|---|---|
4S Study [66] (n = 4444) Study design: RCT Follow-up (median): 5.4 yrs (range 4.9-6.3) Population description: Mixed gender aduts ≥ 18 yrs with history of CHD Risk of bias assessment overall a : Moderate Funding: Mixed (foundation/public) | Model: Logistic regresssion Variables: Age, gender and other variables; LDL-C not included CVD Outcomes: Death of any cause and MACE Lp(a) assay: Isoform independence – NR/unclear; frozen samples | Lp(a) comparison type: Categorical Comparisons included were: Lp(a) ≤ 38.25 units/l (ref) vs. 38.26-91 units/l, 91.1-289.75 units/l, ≥ 289.76 units/l No effect sizes reported, but all six logistic regression analyses reported a positive association, though this was only reported as statistically significant for three out of six comparisons (two simvastatin arm analyses and one placebo treatment arm). The authors concluded that Lp(a) independently predicts major coronary events as well as death in the secondary population. |
AIM–HIGH [67] (n = 3414) Study design: RCT Follow-up: 3 yrs (at trial termination) Population description: Mixed gender adults ≥ 45 yrs with established CVD and dyslipidemia Risk of bias assessment overall a : Low Funding: Mixed (induxstry/public) | Model: Cox proportional hazards Variables: Age, gender and other variables; LDL-C not included CVD Outcomes: Ischemic stroke or TIA Lp(a) assay: Isoform independence - NR; NR if fresh or frozen samples | Lp(a) comparison type: Categorical Four effect sizes reported, including two for ischemic stroke and two for ischemic stroke or TIA. Comparisons were between lowest tertile (reference) Lp(a) and moderate tertile or highest tertile (Lp(a) levels not defined). All showed a statistically significant positive association (same direction). The maximum effect size reported was HR 2.8, 95 % CI: 1.25 to 6.26 (lowest tertile vs. highest tertile) and the lowest HR 2.3, 95 % CI: 1.19 to 4.42 (lowest tertile vs. highest tertile) Results show an independent and significant association between ischemic stroke and elevated baseline Lp(a) [middle/highest tertile] |
Ezhov 2014 [68] (n = 356) Study design: Prospective cohort study Follow-up: 8.5 ± 3.5 yrs (range 0.9 -15.0 yrs) Population description: Mixed gender adults ≥ 18 yrs with stable CHD after sucessful CABG Risk of bias assessment overall a : Low Funding: NR | Model: Cox proportional hazards Variables: Age, gender and other variables; LDL-C not included CVD Outcomes: First cardiovascular event (non-fatal MI, cardiovascular death, coronary revascularization, or hospitalization for recurrent angina) Lp(a) assay: Isoform independent; fresh samples | Lp(a) comparison type: Categorical Two effect sizes reported, both were statistically significant showing a positive association (same direction) between Lp(a) and non-fatal MI or CD (< 30 mg/dl (reference) vs. ≥ 30 mg/dl: HR 2.98, 95 % CI: 1.76 to 5.03) and first ever major CVD event (< 30 mg/dl (ref) vs. ≥ 30 mg/dl: HR 3.47, 95 % CI: 2.48 to 4.85) These results show that Lp(a) concentration is independently associated with three-fold increase in risk of major adverse cardiovascular events within 15 years after CABG. |
GENERATION [69] (n = 483) Study design: Prospective cohort study Follow-up: 1.84 yrs Population description: Males adults ≥ 18 yrs admitted to hospital with the diagnosis of either stable angina, NSTACS or STEMI Risk of bias assessment overall a : Not enough information Funding: NR | Model: Cox proportional hazards Variables: Age, gender and other variables; LDL-C not included CVD Outcomes: Cardiac death; non-fatal MI; rehospitalisation for rest-unstable angina Lp(a) assay: Isoform independence – NR/unclear; frozen samples | Lp(a) comparison type: Categorical Four effect sizes reported for comparison of < 25 mg/dl (reference) vs. ≥ 25 mg/dl. Three were statistically significant showing a positive association (same direction): HR 3.31, 95 % CI: 1.33 to 8.22 (non-fatal MI); HR 2.09, 95 % CI: 1.16 to 4.12 (rehosptialisation for angina); HR 2.42, 95 % CI: 1.52 to 3.84 (CD, non-fatal MI, hospitalisation for unstable angina) One was NS: HR 1.27, 95 % CI: 0.48 to 3.34 (death) Authors concluded that high plasma levels of either CRP or Lp(a) may be associated with the incidence of late events after successful coronary stenting, but a more protracted latent period may be needed in order to manifest clinically the unfavorable influence of an elevated Lp(a) on atherosclerotic plaque instability. The authors also noted that there was not a statndardised analytic method for Lp(a) level determination. |
HERS [70] (n = 2759) Study design: RCT Follow-up: 4.1 yrs Population description: White elderly ≥ 50 yrs postmenopausal females with CHD – placebo group from RCT Risk of bias assessment overall a : Not enough information Funding: Industry | Model: Cox proportional hazards Variables: Age, gender and other variables; includes LDL-C CVD Outcomes: Unstable angina; primary CHD events including non-fatal MI and CHD death; MI Lp(a) assay: Isoform independent; NR if fresh or frozen samples | Lp(a) comparison type: Categorical 15 effect sizes reported for 5 different sets of CVD events. Only 2/15 analyses were statistically significant showing a positive association (same direction): 1st quartile (Lp(a) 0.0-7.0 mg/dl - reference) vs. 4th quartile (Lp(a) 55.0-236 mg/dl):HR 1.54, 95 % CI: 1 to 2.4 (primary CHD events, e.g non-fatal MI) 1st quartile (Lp(a) 0.0-7.0 mg/dl - reference) vs. 4th quartile (Lp(a) 55.0-236 mg/dl):HR 1.61, 95 % CI: 1.1 to 2.3 (CABG/PTCA) Overall, the authors concluded that Lp(a) is an independent risk factor for recurrent CHD in postmenopausal women. |
Ikenaga 2011 [71] (n = 410) Study design: Prospective cohort study Follow-up: 5 yrs Population description: Japanese mixed gender adults ≥ 18 yrs with PCI after MI: Lp(a) ≥ 40 mg/dl Risk of bias assessment overall a : Not enough information Funding: No financial support | Model: Cox proportional hazards Variables: Age, gender and other variables; LDL-C not included CVD Outcomes: MACE (cardiac death, MI and/or revascularisation for new lesions); revascularisation for new lesions Lp(a) assay: Isoform independence – NR/unclear; NR if fresh or frozen samples | Lp(a) comparison type: Categorical Two effect sizes reported and both statistically significant showing a positive association: ≤ 40 mg/dl (reference) vs. > 40 mg/dl: HR 1.64, 95 % CI: 1.31 to 2.06 (MACE) ≤ 40 mg/dl (reference) vs. > 40 mg/dl: HR 1.61, 95 % CI: 1.32 to 2.13 (revascularisation for new lesions) Results show that Lp(a) levels can significantly and independently predict the progression of non-culprit lesions after acute MI |
Konishi 2013 [45] (n = 330) Study design: Prospective cohort study Follow-up (median): 4.7 yrs Population description: Mixed gender Japenese adults ≥ 18 yrs undergoing PCI with achieved lipid targets: Lp(a) ≥30 mg/dl Risk of bias assessment overall a : High Funding: Public | Model: Cox proportional hazards and multivariable analysis Variables: Age, gender and other variables; LDL-C not included CVD Outcomes: All-cause death and ACS Lp(a) assay: Isoform independence – NR/unclear; NR if fresh or frozen samples | Lp(a) comparison type: Categorical Two effect sizes reported and both statistically significant showing a positive association (same direction): ≤ 30 mg/dl (reference) vs. ≥30 mg/dl: HR 1.68, 95 % CI: 1.03 to 2.7 (Cox proportional hazards) ≤ 30 mg/dl (reference) vs. ≥30 mg/dl: HR 2.47, 95 % CI: 1.19 to 5.06 (multivariable analysis) Results showed that high Lp(a) [ ≥30 mg/dL] could independently predict major adverse events |
Kwon 2015 [73] (n = 1494) Study design: Prospective cohort study Follow-up (mean): 4.4 (SD 2.6) yrs Population description: Mixed gender Korean adults ≥ 18 yrs with diabetes and a history of symptomatic CAD including IHD, stable/unstable angina, and MI Risk of bias assessment overall a : Moderate Funding: Public | Model: Cox proportional regression analysis Variables: Age, gender and other variables; LDL-C leven not included (hyperlipidemia defined as LDL-C of at least 130 mg/dL was included) CVD Outcomes: MACE Lp(a) assay: Isoform independent; NR if fresh or frozen samples | Lp(a) comparison type: Categorical Two effect sizes (adjusting for age, gender, hypertension, hyperlipidemia, smoking and extent of CAD) reported and both statistically significant showing a positive association (same direction) with risk of MACE: Tertile 1 (median 4.7 mg/dL; reference) vs. Tertile 2 (median 13.5 mg/dL): HR 1.54, 95 % CI: 0.68 to 3.50 Tertile 1 (median 4.7 mg/dL; reference) vs. Tertile 3 (median 38.8 mg/dL): HR 2.89, 95 % CI: 1.37 to 6.08 In addition, a survival probability plot according to Lp(a) tertile suggested that elevated Lp(a) level was associated with a worse prognosis (p = 0,008) after adjusting for age, gender, hypertension, hyper lipidemia, smoking and extent of CAD. Results suggested elevated Lp(a) is associated with worse outcomes (MACE) in type 2 diabetics patients with symptomatic CAD and has incremental prognostic value. |
LIPID [13] (n = 3949) Study design: RCT Follow-up: 6 yrs and 8 yrs Population description: Mixed gender White adults ≥ 18 yrs with history of CVD (Lp(a) 13.9-44.1 mg/dl) Risk of bias assessment overall a : Moderate Funding: Industry | Model: Cox proportional hazards Variables: Age, gender and other variables; LDL-C not included CVD Outcomes: Total CHD events (non-fatal MI, CHD death, unstable angina, coronary revascularization) Lp(a) assay: Isoform independent; frozen samples | Lp(a) comparison type: Categorical 36 effect sizes were reported across 3 different comparisons and 12 different CVD outcomes. 11/39 effect sizes were statistically significant analyses (all positive association (same direction). Effect size ranges were: From ≤ 13.9 mg/dl (reference) vs. >73.7 mg/dl: HR 1.21, 95 % CI: 1.07 to 1.36) To ≤ 13.9 mg/dl (reference) vs. >73.7 mg/dl: HR 1.45, 95 % CI: 1.2 to 1.75 28/39 analyses were NS including 3 at 8 yr follow-up and 28 at 6 yr follow-up Overall, the authors concluded that baseline Lp(a) and increased Lp(a) concentrations after one year were independently associated with future cardiovascular disease and CHD events. |
Park 2015 [72] (n = 161) Study design: Prospective cohort study (retrospective analysis of prospective registry data) Follow-up (median): 6 yrs (maximum: 8 yrs) Population description: Mixed gender adults ≥ 18 yrs undergoing PCI Risk of bias assessment overall a : Moderate Funding: NR | Model: Cox proportional hazards Variables: Age, gender and other variables; LDL-C included CVD Outcomes: MACE Lp(a) assay: Isoform independence – NR/unclear; NR if fresh or frozen samples | Lp(a) comparison type: Categorical Cox proportional hazards regression analysis adjusted for gender, age, diabetes mellitus, hypertension, hyperlipidemia, smoking, multivessel disease, minimal luminal diameter after PCI, reference vessel diameter after PCI, LDL-C, total lesion length, Lp(a) ≥ 50 mg/dL, showed that Lp(a) > 50 mg/dL (vs. Lp(a) ≤ 50 mg/dL) was significantly associated with the 3 yr adverse clinical outcomes including any MI, revascularization (target lesion revascularization (TLR) and target vessel revascularization (TVR)), TLR- MACEs, TVR-MACE, and All-MACEs. One significant effect size: OR 2.88, 95 % CI: 1.37 to 6.07. Authors concluded that high Lp(a) level ≥ 50 mg/dL in angina pectoris patients undergoing elective PCI with DES was significantly associated with binary restenosis and 3 yr adverse clinical outcomes in an Asian population. |
RESEARCH [74] (n = 161) Study design: Prospective cohort study Follow-up (median): 6 yrs (maximum: 8 yrs) Population description: Mixed gender adults ≥ 18 yrs undergoing PCI Risk of bias assessment overall a : Moderate Funding: NR | Model: Cox proportional hazards Variables: Age, gender and other variables; LDL-C not included CVD Outcomes: MACE Lp(a) assay: Isoform independent; frozen samples | Lp(a) comparison type: Categorical Eight effect sizes reported. Five out of eight effect sizes were statistically significant and all were positive association (same direction). These ranged from: Lp(a) tertile 1 (cut point 0.27 mg/dl - reference) vs. Lp(a) tertile 3 (> 1.83 mg/dl): HR 1.9, 95 % CI: 1 to 3.5 To Lp(a) tertile 1 (cut point 0.27 mg/dl - reference) vs. Lp(a) tertile 2 (cutpoint 1.83 mg/dl: HR 3.7, 95 % CI: 1.4 to 10.1 Three out of eight analyses were NS The results showed that high levels of Lp(a) were independently associated with a higher 1-year risk of MACE, |
Rosengren 1990 [75] (n = 155) Study design: Nested case-control study Follow-up: 6 yrs Population description: Males ≥ 50 yrs with MI or CHD death Risk of bias assessment overall a : Moderate Funding: Mixed (foundation/ public) | Model: Logistic regression Variables: Age, gender and other variables; LDL-C not included CVD Outcomes: CHD deaths and non-fatal MI Lp(a) assay: Isoform independence – NR/unclear; frozen samples | Lp(a) comparison type: Categorical One effect size reported for the comparison of controls (reference) vs. cases. This was reported as statistically significant difference for the unconditional likelihood estimate (0.0031) suggesting a positive association (same direction) The results show that serum Lp(a) concentration is an independent risk factor for subsequent MI or death from CHD |
TNT Study [76] (n = 1506) Study design: RCT Follow-up (median): 4.9 yrs Population description:Mixed gender adults ≥ 40 yrs who have experienced major cardiovascular events and are receiving statin treatment Risk of bias assessment overall a : High Funding: Pharma | Model: Cox proportional hazards regression Variables: Age, gender and other variables; LDL-C not included CVD Outcomes: CHD death; non-fatal, non-procedure-related myocardial infarction; resuscitated cardiac arrest; and fatal or nonfatal stroke. Lp(a) assay: Commercial assay; Immunoturbidimetric assay; Isoform dependence - NR/unclear; NR if fresh or frozen samples | Lp(a) comparison type: Categorical Three effect sizes reported. Two out of three effect sizes were statistically significant and all were positive association (same direction). One effect size was reported in whole population for Lp(a) with major CV events: HR 1.17, 95 % CI: 1.04 to 1.33. Other significant effect size was reported in subgroup atorvastatin 10 mg QD for Lp(a) with major CV events: HR 1.34, 95 % 1.12 to 1.6. No significant effect in Atorvastatin 80 mg subgroup 1.01 (95 % CI 0.85 to 1.20) Results suggest that higher plasma levels of Lp(a) are independently associated with an increased risk of recurrent events. |
Wehinger 1999 [77] (n = 2223) Study design: Prospective cohort study Follow-up: 1 yr Population description: Mixed gender adults ≥ 18 yrs successfully treated with intracoronary stent due to symptomatic CAD Risk of bias assessment overall a : Moderate Funding: NR | Model: Log-rank test Variables: Age, gender and other variables; LDL-C not included CVD Outcomes: Angiographic restenosis Lp(a) assay: Isoform independence – NR/unclear; fresh samples | Lp(a) comparison type: Categorical Four effect sizes reported. All were NS for the comparison between Lp(a) quintiles (2 to 3) vs. Lp(a) quintile 1. Results suggest that elevated Lp(a) levels did not influence the one-year clinical and angiographic outcome after stent placement. Thrombotic events and measures of restenosis were not adversely affected by the presence of high Lp(a) levels. |