In the Simvastatin and Ezetimibe in Aortic Stenosis (SEAS) trial, combined ezetimibe (10 mg) and simvastatin (40 mg) decreased low-density lipoprotein cholesterol levels by 50% and ischemic cardiovascular event (ICE) risk by 22% compared to placebo. A larger decrease in ICE risk might have been expected for the degree of lipid-lowering observed. This analysis investigated relations between changes in lipoprotein components (LCs), and ICE risk decrease in the SEAS trial in all patients, by severity of aortic stenosis (AS), and compared to results of other clinical trials. A total of 1,570 patients with baseline aortic jet velocity (JV) data, baseline and 1-year low-density lipoprotein cholesterol, high-density lipoprotein cholesterol, and apolipoprotein B, and no ICEs during the first year were included in the analysis. Relations between on-treatment measurements of 1-year LCs and time-to-ICE occurrence were assessed in all patients and in JV tertiles (<2.8, 2.8 to 3.3, and >3.3 m/s). Observed and predicted ICE risk decreases were compared by Cox model. Decreases in LCs after 1 year of ezetimibe plus simvastatin were associated with decreased ICE risk in all patients and in the 2 lower JV tertiles (p <0.05 to <0.001) but not in tertile 3. In JV tertiles 1 and 2, ICE risk decreased by 47% and 36%, respectively, was reasonably well predicted by all LCs, and was consistent with findings from meta-regression analyses in other populations. In conclusion, the degree of lipid lowering by ezetimibe plus simvastatin may predict the extent of ICE risk decrease in patients with mild AS, but ICE risk prediction in patients with more severe AS is confounded by AS-associated cardiovascular events and a shorter interval of exposure to lipid lowering.
In the Simvastatin and Ezetimibe in Aortic Stenosis (SEAS) trial, decreases of 50% in low-density lipoprotein cholesterol and 22% (95% confidence interval [CI] 3 to 37) in the composite ischemic cardiovascular event (ICE) end point (cardiovascular death, nonfatal acute myocardial infarction, coronary artery bypass grafting [CABG], percutaneous coronary intervention, hospitalization for unstable angina pectoris, nonhemorrhagic stroke) resulted after treatment with the combination of ezetimibe (10 mg) and simvastatin (40 mg) compared to placebo. Given the substantial extent of low-density lipoprotein cholesterol lowering and improvements in other lipoprotein components (LCs) in SEAS, a greater decrease in ICE risk might have been expected, based on findings in the Cholesterol Treatment Trialists’ (CTT) meta-analysis of 14 statin trials and epidemiologic studies in populations other than those with aortic stenosis (AS). The SEAS population is unique compared to the CTT meta-analysis population because of the strict inclusion criteria and a relatively low atherogenic lipoprotein profile and ICE risk. Thus, factors that differentiate the AS population in SEAS from other populations may have influenced these findings. In SEAS, baseline severity of AS, assessed by baseline peak aortic jet velocity (JV), influenced the effect of lipid-lowering therapy on cardiovascular events. The present report examined relations between changes in LC and ICE risk decrease in all patients and based on AS severity as stratified by JV tertiles. Analyses were made separately by including and excluding CABG from the ICE end point. In addition, the observed ICE risk decrease in the total population and in JV tertiles was compared to that predicted by the degree of low-density lipoprotein cholesterol lowering, observed at the 1-year time point within SEAS, and compared to data from the CTT meta-analysis.
Methods
The SEAS study design and main outcome have been published. In short, men and women (45 to 85 years of age) with asymptomatic mild-to-moderate AS (Doppler measured JV ≥2.5 and ≤4.0 m/s) and the absence of heart conditions requiring therapy at baseline were included. Patients with known coronary, cerebral, or peripheral arterial disease, diabetes mellitus, endocrine disorders, or active liver disease or patients taking/deemed to require lipid-lowering therapy to decrease atherosclerotic cardiovascular risk were excluded.
Patients were randomized to receive ezetimibe plus simvastatin or placebo. Median study duration was 52 months; patients were followed for ≥4 years. Events were classified by an independent, blinded, end-point classification committee. Safety and lipid end points were assessed every 6 months. Echocardiography was assessed yearly and before valve surgery; only baseline and 1-year measurements were evaluated in this analysis. The SEAS Echocardiography Core Laboratory (Haukeland University Hospital, Bergen, Norway) performed all echocardiographic measurements and only core laboratory values were used for analyses. Baseline JV measurements were obtained in 1,763 of the 1,873 SEAS patients (missing values in 110 patients). Associations between on-treatment LCs and subsequent ICEs were evaluated by including all patients in a time-varying Cox regression analysis where LCs were updated from the previous visit and then by a method where only end points occurring after the 1-year visit were included. In this analysis, all patients were excluded who died or had an ICE before the 1-year time point (chosen because all LCs were included in the 1-year assessments) and who did not have complete 1-year measurements of low-density lipoprotein cholesterol, high-density lipoprotein cholesterol, and apolipoprotein B. Data are presented for the fixed-covariate approach only (thought to be more clinically relevant), whereas deviations found using the time-varying approach are briefly described. The study population was classified by JV tertiles with split points at 2.79 and 3.33 m/s.
A Cox proportional hazards model was used to estimate the slope of the association between 1-year LC levels and time to subsequent first occurrence of ICEs (interaction tests revealed the relations between LCs at 1 year and ICEs to be nonsignificantly different between treatments) and to test interaction effects by a multiplicative model term between treatment and JV tertile. Because this approach loses statistical power when excluding ICE cases occurring during the first year, an additional analysis model was run using a time-varying covariate Cox regression model, where follow-up started at baseline. LCs were measured at baseline and at 8- and 24-week visits and every 1/2 year thereafter. Because the CABG proportion of ICEs was large and determined by a different mechanism compared to an ordinary lipid-outcome trial, a separate analysis of ICE risk excluding CABG cases was performed.
Mean differences between treatment groups in LCs at 1 year were calculated and used to approximate the observed average difference in LC exposure between groups, relevant only in the time-fixed regression model. One-way analysis of variance was used to test for differences in 1-year LCs between JV tertiles. Predicted ICE hazard ratios (HRs) and 95% CIs were calculated by the formula: exponential ([β ± 1.96 × SE] × D), where D is the SD of the respective LC variable at 1 year, β is the nonstandardized regression coefficient of LCs, and SE is the SE of β. This calculation assumes that the risk decrease is proportional to LC change on the log scale when using the slope from the regression equation. The predicted value was then compared to the observed value for each LC and JV tertile. External comparisons of ICE risk decrease were made using data from the CTT meta-analysis.
Results
A total of 1,570 patients in the SEAS trial who had complete data for baseline JV, baseline and 1-year low-density lipoprotein cholesterol, high-density lipoprotein cholesterol, and apolipoprotein B measurements, and no ICEs during the first year were included in the main analyses. A special analysis with time-varying covariates was also performed in those 1,763 patients who had a valid baseline assessment of peak aortic JV measured at the core laboratory. Patient baseline characteristics were similar to those in the original SEAS study cohort (n = 1,873). Patients were on average 67 years of age ( Table 1 ). Mean JVs at baseline were 2.5, 3.1, and 3.7 m/s for tertiles 1, 2, and 3, respectively. Mean baseline LC levels were generally similar among the JV tertiles as were other baseline variables except for C-reactive protein, which increased with higher JV (p = 0.02).
| Baseline and Year-1 Variables | Tertile 1 (n = 534) | Tertile 2 (n = 522) | Tertile 3 (n = 514) | ||||||
|---|---|---|---|---|---|---|---|---|---|
| Baseline | E + S | Year-1 Placebo | Baseline | E + S | Year-1 Placebo | Baseline | E + S | Year-1 Placebo | |
| Age (years) | 67 ± 9 | — | — | 67 ± 10 | — | — | 67 ± 10 | — | — |
| Men | 318 (60%) | — | — | 320 (61%) | — | — | 323 (63%) | — | — |
| Body mass index (kg/m 2 ) | 27 ± 4 | — | — | 27 ± 4 | — | — | 27 ± 4 | — | — |
| Systolic blood pressure (mm Hg) | 145 ± 20 | — | — | 144 ± 20 | — | — | 144 ± 21 | — | — |
| Peak transaortic jet velocity (m/s) ⁎ | 2.5 ± 0.2 | 2.7 ± 0.4 | 2.7 ± 0.4 | 3.1 ± 0.2 | 3.2 ± 0.4 | 3.2 ± 0.4 | 3.7 ± 0.3 | 3.9 ± 0.5 | 3.8 ± 0.5 |
| Low-density lipoprotein cholesterol | |||||||||
| mmol/L | 3.6 ± 0.9 | 1.4 ± 0.6 | 3.4 ± 0.8 | 3.6 ± 0.9 | 1.4 ± 0.8 | 3.4 ± 0.8 | 3.5 ± 0.9 | 1.4 ± 0.5 | 3.4 ± 0.9 |
| mg/dl | 139 ± 34 | 53 ± 22 | 133 ± 33 | 138 ± 34 | 54 ± 31 | 132 ± 32 | 137 ± 36 | 53 ± 20 | 132 ± 35 |
| Non–high-density lipoprotein cholesterol | |||||||||
| mmol/L | 4.3 ± 1.0 | 1.9 ± 0.6 | 4.1 ± 0.9 | 4.2 ± 1.0 | 1.9 ± 0.9 | 4.1 ± 0.9 | 4.2 ± 1.0 | 1.8 ± 0.6 | 4.0 ± 1.0 |
| mg/dl | 165 ± 38 | 72 ± 24 | 159 ± 36 | 163 ± 37 | 72 ± 35 | 157 ± 37 | 161 ± 39 | 71 ± 23 | 156 ± 38 |
| Apolipoprotein B † | |||||||||
| g/L | 1.3 ± 0.3 | 0.7 ± 0.2 | 1.3 ± 0.3 | 1.3 ± 0.3 | 0.7 ± 0.3 | 1.3 ± 0.3 | 1.3 ± 0.3 | 0.7 ± 0.2 | 1.2 ± 0.3 |
| mg/dl | 131 ± 29 | 69 ± 19 | 127 ± 28 | 131 ± 27 | 69 ± 26 | 126 ± 26 | 130 ± 28 | 69 ± 19 | 125 ± 29 |
| Total cholesterol/high-density lipoprotein cholesterol | 4.1 ± 1.5 | 2.3 ± 0.5 | 3.9 ± 1.2 | 4.0 ± 1.2 | 2.3 ± 0.8 | 3.8 ± 1.0 | 4.0 ± 1.1 | 2.2 ± 0.5 | 3.9 ± 1.1 |
| C-reactive protein | |||||||||
| nmol/L | 33 ± 41 | — | — | 39 ± 91 | — | — | 47 ± 92 | — | — |
| mg/L | 3.5 ± 4.3 | — | — | 4.1 ± 9.5 | — | — | 4.9 ± 9.6 | — | — |
| Creatinine | |||||||||
| μmol/L | 94 ± 16 | — | — | 93 ± 15 | — | — | 94 ± 16 | — | — |
| mg/dl | 1.1 ± 0.2 | — | — | 1.1 ± 0.2 | — | — | 1.1 ± 0.2 | — | — |
| Glucose | |||||||||
| mmol/L | 5.3 ± 0.8 | — | — | 5.3 ± 0.8 | — | — | 5.3 ± 0.9 | — | — |
| mg/dl | 95 ± 15 | — | — | 96 ± 14 | — | — | 95 ± 16 | — | — |
| Hypertension | 289 (54%) | — | — | 262 (50%) | — | — | 250 (49%) | — | — |
| Smoking | |||||||||
| Current | 91 (17%) | — | — | 104 (20%) | — | — | 102 (20%) | — | — |
| Previous | 178 (33%) | — | — | 194 (37%) | — | — | 196 (38%) | — | — |
⁎ n = 1,476 year-1 measurements.
Changes in LCs from baseline to 1 year were slightly smaller in JV tertile 3 compared to 1 and 2, possibly reflecting natural history or more patients with aortic valve replacement during the first year in that tertile ( Table 1 ) because of more severe AS. All LCs were markedly decreased by treatment, and percent decreases were similar (p >0.20) across JV tertiles. Mean baseline low-density lipoprotein cholesterol levels were higher than the current recommended target of 3.0 mmol/L (115 mg/dl) for asymptomatic subjects with ≥3 risk factors. Treatment decreased low-density lipoprotein cholesterol to well below the recommended levels across JV tertiles. Lengths of follow-up ± SDs with interquartile ranges were 1,235 ± 221.1 (interquartile range 1,153 to 1,335), 1,232.3 ± 236.9 (interquartile range 1,153 to 1,351), and 1,212.6 ± 275.9 (interquartile range 1,147 to 1,361) in the 3 tertiles, respectively.
After the 1-year time point, there were 237 first end-point cases of ICEs in the 1,570 patients at risk (15.1%; Table 2 ). The CABG component of this end point constituted the greatest proportion of ICEs (n = 119, 50.2%), and these procedures coincided with an aortic valve replacement in all cases but 1, which occurred after a preceding isolated valve replacement. Smaller numbers of ICE cases were observed in tertiles 1 and 2 compared to tertile 3. Incidences of CABG cases compared to other ICEs were also lower in tertiles 1 and 2 than in tertile 3 (p <0.05), whereas those for the composite events of fatal or nonfatal myocardial infarction or unstable angina were lower in tertile 3 than in tertiles 1 and 2 (p <0.05). Incidences of other events (percutaneous coronary intervention, stroke, and cardiovascular death excluding fatal myocardial infarction) were lower in tertile 3 than in 1 and similar in tertiles 2 and 3. Aortic valve replacements were graded and much more frequent in the 2 higher tertiles than in tertile 1.
| Tertile 1 (n = 534) | Tertile 2 (n = 522) | Tertile 3 (n = 514) | |
|---|---|---|---|
| Aortic valve replacement ‡ | 54 (10%) | 133 (25%) | 255 (50%) |
| Aortic valve replacement ‡ (from year 1 onward) | 54 (10%) | 132 (25%) | 242 (47%) |
| Patients with ischemic cardiovascular events | 57 (11%) | 61 (12%) | 119 (23%) |
| Distribution of type of first ischemic cardiovascular event ⁎ | |||
| Coronary artery bypass grafting § | 20 (35%) | 29 (48%) | 70 (59%) |
| Fatal or nonfatal myocardial infarction or unstable angina § | 14 (25%) | 13 (21%) | 10 (8%) |
| Others † | 23 (40%) | 19 (31%) | 39 (33%) |
⁎ Percentages are relative to the number of patients with ischemic cardiovascular events.
† Includes percutaneous coronary intervention, stroke, and cardiovascular death (excluding fatal myocardial infarction).
Table 3 lists HRs per SD increase in LC by the time-fixed mode of analysis using the 1-year levels as covariates by JV tertiles. Results are given for all patients in the analysis population and in baseline JV tertiles. No heterogeneity was found between the 2 treatment groups in these relations (all p values >0.20); consequently, the 2 groups were pooled for analysis purposes. All LCs were significantly associated with ICE risk in the 2 lowest JV tertiles, whereas none were significantly related to ICE risk in the highest tertile. HRs were almost 50% lower (weaker associations) in tertile 2 compared to tertile 1 for the 4 LCs. All LCs showed significant interactions with tertile category related to ICE risk. When all ICEs were used in the time-varying approach, these relations were still significant for total patients and for tertile 1. However, estimates were lower than in the time-fixed model, and the significant relations in tertile 2 were lost (data not shown). Only apolipoprotein B and non–high-density lipoprotein cholesterol showed significant interactions between JV tertiles and LCs on ICE risk (p <0.05).
| Lipoprotein Component | Tertile 1 | Tertile 2 | Tertile 3 | Interaction p Value ⁎ | Total HR (95% CI) | ||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| SD | HR (95% CI) | p Value | SD | HR (95% CI) | p Value | SD | HR (95% CI) | p Value | |||
| Low-density lipoprotein cholesterol (mmol, mg/dl) | 1.3 | 1.5 (1.2–2.0) | 0.001 | 1.3 | 1.3 (1.0–1.7) | 0.030 | 1.3 | 1.0 (0.9–1.2) | 0.610 | 0.010 | 1.3 (1.0–1.5) |
| Non–high-density lipoprotein cholesterol (mmol, mg/dl) | 1.4 | 1.6 (1.3–2.1) | <0.001 | 1.4 | 1.3 (1.0–1.7) | 0.021 | 1.4 | 1.0 (0.9–1.2) | 0.750 | 0.002 | 1.3 (1.1–1.5) |
| Apolipoprotein B (g/L, mg/dl) | 0.4 | 1.7 (1.3–2.1) | <0.001 | 0.4 | 1.4 (1.1–1.8) | 0.009 | 0.4 | 1.1 (0.9–1.3) | 0.379 | 0.004 | 1.4 (1.1–1.6) |
| Total cholesterol / high-density lipoprotein cholesterol | 1.2 | 1.6 (1.3–2.0) | <0.001 | 1.2 | 1.3 (1.0–1.6) | 0.028 | 1.2 | 1.0 (0.9–1.2) | 0.689 | 0.003 | 1.2 (1.1–1.4) |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
