Reports

This report is based on medical evidence presented at sanctioned medical congress, from peer reviewed literature or opinion provided by a qualified healthcare practitioner. The consumption of the information contained within this report is intended for qualified Canadian healthcare practitioners only.

JOURNAL CLUB 2008 - Cardiology

December 2008

Editorial Overview:

David Fitchett, MD, FRCPC

Division of Cardiology, St. Michael’s Hospital, Associate Professor of Medicine, University of Toronto, Toronto, Ontario

The SEAS (Simvastatin and Ezetimibe in Aortic Stenosis) trial tested a completely new hypothesis: that cholesterol lowering would reduce event rates in patients with aortic valve stenosis and no known coronary artery disease. That aggressive lipid lowering in this study failed to demonstrate an improved outcome was not a surprise to many of us. The hypothesis that led to the SEAS study was drawn primarily from the common pathological features of atherosclerosis and aortic valve disease suggesting that lipid lowering might prevent the progression of aortic valve disease in the same way lipid reduction is beneficial for individuals with atherosclerosis. The results of SEAS were therefore unpredictable. Although SEAS provides prospective evidence that lipid lowering does not reduce progression and complications of asymptomatic aortic stenosis, it should not alter the well-established role of lipid control in the management and prevention of coronary artery disease.

The SEAS (Simvastatin and Ezetimibe in Aortic Stenosis) study was presented at the 2008 European Society of Cardiology (ESC) Congress and published simultaneously (Rossebø et al. N Engl J Med 2008;359(13):1343-56). The study tested the hypothesis that intensive cholesterol lowering would reduce the risk of major cardiovascular (CV) events in patients with mild to moderate asymptomatic aortic valve stenosis. The study was unable to validate the hypothesis. Over a median follow-up of 52.2 months, an intensive lipid-lowering strategy that reduced LDL-C by 61.3% compared to placebo did not reduce the composite end point of aortic valve and ischemic events.

The 61% LDL reduction achieved with simvastatin 40 mg plus ezetimibe 10 mg is consistent with previous observations of the efficacy of the combined agents. In patients with coronary artery disease, an LDL-C reduction of this magnitude is associated with a substantial reduction in CV events. Yet the SEAS trial tested the ability of lipid lowering to reduce the risk of events associated with valve disease, including the need for valve replacement and the risk of heart failure associated with aortic stenosis progression. Consistent with the trial design, events associated with the aortic valve were far more common than those associated with vascular disease, and aortic valve replacement represented almost 80% of all events included in the primary outcome. The primary message from SEAS, therefore, is that intensive lipid lowering is not effective in reducing the clinical risks posed by valve disease.

When ischemic CV events were assessed separately, intensive lipid lowering in the SEAS trial with the combination of simvastatin/ ezetimibe compared to placebo was associated with a significant 22% (P=0.02) reduction in CV events. The most frequent ischemic CV event was the need for revascularization by coronary artery bypass grafting (CABG) at the time of aortic valve replacement, which was reduced by intensive lipid lowering (placebo 10.8% vs. simvastatin/ezetimibe 7.2%) over the more than four years of follow-up. The rate of non-fatal MI was also lower in the active treatment group (placebo 2.8% vs. simvastatin/ ezetimibe 1.8%).

Figure 1.

Although the SEAS trial demonstrates that aggressive lipid lowering did not reduce the most common clinical events associated with aortic stenosis, it is also possible to draw a second and related conclusion that ischemic CV events are not the most common source of complications for patients with valvular aortic stenosis.

LDL lowering with statins has successfully reduced CV events in almost every population evaluated so far. Consequently, there has been a desire to test the impact of lipid lowering for other conditions, such as aortic stenosis. The extension of treatment for new indications is facilitated by the safety of statins, with a favourable benefit:risk ratio even when the clinical advantage is modest. However, as statins are tested outside of indications where their mechanism of action is well established, the opportunity for benefit will be less predictable and less likely to demonstrate the benefit observed in vascular diseases. In the specific case of aortic valve disease, calcification and degeneration in advanced disease is unlikely to be affected by lipid lowering.

Whilst the absence of a significant protective effect from intensive lipid lowering on aortic valve disease in SEAS was not surprising, the association of an increased risk of cancer and active therapy was unexpected. Yet analysis of multiple clinical trials of lipid lowering with over 90,000 subjects shows no hint of an increased cancer risk associated with statin therapy. The SEAS trial is a small study with only 1800 subjects, and the observation had many inconsistencies suggesting that the increased incidence of cancer is a chance finding. Firstly, fatal cancer developed very rapidly within three years. As Dr. Terje Pedersen (the study steering committee chair) stated, “I am not aware of any data that would support the hypothesis that ezetimibe causes fatal cancer within three years after the onset of treatment. I do not know of anything that can do that, except maybe a Chernobyl catastrophe.” Secondly, there was no clustering of cancer type; moreover, the absence of a cluster weakens the likelihood
onship.

Figure 2.

<img2811|center>

The surprising finding of the increased cancer risk in the SEAS trial led to an early safety analysis of two large ongoing clinical trials that are evaluating the outcome benefits of the combination of simvastatin and ezetimibe that was published in the same edition of the New England Journal of Medicine as the SEAS study (Peto et al. N Engl J Med 2008;359(13):1357-66). This analysis evaluated the cancer rates in the SHARP (Study of Heart and Renal Protection) and IMPROVE-IT (Improved Reduction of Outcomes: Vytorin Efficacy International Trial) databases. The analysis included the SHARP study (9264 patients with a mean follow-up of 2.7 years), the IMPROVE-IT study (11,353 patients with a mean follow-up of one year) and the SEAS study (1873 patients with a mean follow-up of 4.3 years). The analysis showed no credible evidence to support an increased cancer risk with ezetimibe in combination with simvastatin on rates of cancer.

The results of SEAS reduces the likelihood of lipid lowering being a successful strategy in the management of patients with aortic stenosis. However, it is appropriate to wait for the results of the ASTRONOMER (Aortic Stenosis Progression Observation Measuring Effects of Rosuvastatin) study, which is addressing the same question before drawing definitive conclusions. It is important to emphasize that the results of SEAS do not in any way challenge the proven relationship between lipid lowering and the reduction in coronary artery disease events, especially in patients at elevated CV risk. The current guideline-recommended goals for cholesterol control should be followed. Treatment should be initiated with statins and the drug titrated to the optimally tolerated dose. If the LDL-C target is not achieved, then combination therapy with agents such as ezetimibe should be considered.

Summary

The failure of lipid lowering to prevent the progression of calcific aortic valve stenosis in the SEAS study was not unexpected. These findings do not challenge the LDL hypothesis for the prevention of coronary heart disease events. Furthermore, the SEAS trial demonstrated the powerful efficacy of the combination of simvastatin and ezetimibe to lower LDL-C by more than 60%. Whilst statins are first-line therapy for achieving treatment goals, adjunctive therapies, including ezetimibe, are appropriate when treatment targets are not reached with statin therapy alone.

questions and answers

Panel

Kwan-Leung Chan, MD, FRCPC, Cardiologist, University of Ottawa Heart Institute

David Fitchett, MD, FRCPC, Division of Cardiology, St. Michael’s Hospital

Gordon Hoag, MD, PhD, Royal Jubilee Hospital

Peter H. Jones, MD, Center for Cardiovascular Disease Prevention, Methodist DeBakey Heart Center

Koon Teo, MBBCH, PhD, LRCP, MRCP, FRCP, FRCPC, Department of Clinical Epidemiology and Biostatistics, McMaster University

Were you surprised that intensive LDL reductions were not associated with significant benefit on the composite primary end point of the SEAS trial?

Dr. Jones: This result was not surprising. The trial was weighted to detect a reduction in the need for aortic valve surgery rather than on CV outcomes and benefit against aortic valve disease from lipid lowering was not certain. There was some evidence from smaller studies that cholesterol lowering might reduce valve complications, possibly through antiinflammatory effects, but the support for this hypothesis was not particularly strong. The investigators rolled the dice. As we look for new indications for statins outside of their proven areas of benefit, we are likely to see more of these types of studies. Aortic valve disease is life-threatening and valve replacement is costly and risky, so protection against progression with a statin would have been an important benefit, but the absence of benefit was not particularly surprising.

Dr. Teo: Yes, I was surprised with the results because I strongly believed the study hypothesis that LDL lowering would result in substantial reduction in progression of the aortic stenosis. However, I also believed that this hypothesis has to be vigorously tested and proven before being accepted in clinical practice, which is the point of studies like SEAS. In the meantime, one may speculate on why the hypothesis was not proven. Was it due to the patients selected? Was it related to the extent of disease in the valves? Perhaps the co-interventions, particularly the many cases of CABG that resulted in aortic valve replacement, diluted the advantage of statins. However, there may simply be no relationship between LDL reductions and aortic stenosis progression.

Dr. Fitchett: Ultimately, the evidence for a benefit from LDL lowering in patients with aortic stenosis was weak. I think protection from LDL lowering against progression of aortic valve disease would have been more surprising than the evidence that it did not provide protection. The study confirms that these patients die primarily of progressive valve disease, which is not affected by lipid lowering, rather than from vascular events, which is. As there is no obvious mechanism by which LDL lowering would reduce pathology in the valve, the results cannot be considered that surprising.

Dr. Hoag: The LDL reductions on the SEAS trial were most interesting. Although it was a negative trial, it may not reflect the benefits of LDL reduction but probably represents a design issue and so the results were not really surprising. The results may not be conclusive until we better understand the pathophysiology of aortic valve disease.

Dr. Chan: The primary end point in SEAS was a composite of ischemic and aortic stenosis-related events. Although we would expect a 61% reduction in LDL-C to reduce the risk of vascular events, the absence of overall benefit in SEAS is likely due to the fact that aortic stenosis-related events were more common. In fact, vascular events were reduced on lipid lowering therapy when these events were evaluated separately. It may have been more useful to understand the effects of intensive lipid lowering by looking at ischemic and aortic stenosis-related events as separate endpoints.

Do the results of SEAS weaken the importance of lowering LDL for its current indications?

Dr. Teo: No. The importance of LDL lowering in reducing vascular events has been proven beyond doubt by many trials.

Dr. Chan: The results of the SEAS study should only be applicable to patients with aortic stenosis and should not mitigate the effectiveness of lowering LDL-C in the prevention of ischemic events in patients at risk for CV events identified in the previous trials.

Dr. Fitchett: Not in the least. The correlation between the reduction in LDL and the reduction in the risk of vascular events has been well proven in dozens of controlled studies with remarkably consistent results. The SEAS study of lipid lowering in patients with aortic valve disease really has no relevance to the wellproven CV risk reductions associated with lipid lowering.

Dr. Jones: The lipid-lowering hypothesis is no longer a hypothesis but a proven concept in CV risk reduction. The lipid hypothesis relates specifically to vascular disease. Lack of effect in patients with aortic valve disease is not relevant to the ability of lipid reductions to reduce the risk of thrombotic events. In SEAS, lipid lowering did reduce CV events, particularly the need for coronary artery bypass graft procedures so if anything, it reinforces the relationship between lipid lowering and CV risk reduction.

Dr. Hoag: The benefit of reducing LDL as a strategy for reducing CV risk has been consistent across almost all studies. The SEAS data do not change this perspective. I do not think the SEAS study has provided a definitive answer regarding the effect of lipid lowering in aortic stenosis, so this is a question that is unresolved.

Do you believe that the degree of LDL lowering correlates with the degree of risk reduction and that a target of <2 mmol/L is appropriate in patients at high risk for CV events?

Dr. Jones: If you are at high risk for CV events, an LDL target of <2.0 mmol/L is achievable, it is safe, and it is associated with increased protection relative to higher LDL levels. All of the data we have obtained so far, including data generated by the recent JUPITER study, show that lower is better regardless of the value. In my opinion, guidelines are likely to be simplified to recommend an LDL of <2.5 mmol/L for anyone at increased risk of CV disease and <2.0 mmol/L in those at high risk. For those who can go even lower, there appears to be additional protection.

Dr. Teo: Yes, again, the results of many secondary prevention trials tell us the target is <2.0 mmol/L.

Dr. Fitchett: A target of <2.0 mmol/L was established by the TNT (Treating to New Targets) trial, which confirmed additional risk reductions relative to higher LDL levels. The current guidelines in Canada, which call for a target of <2.0 mmol/L in high-risk patients, are evidence-based. In post-hoc analyses from the lipid-lowering studies, it appears that LDL levels far below 2.0 mmol/L provide additional incremental benefit, but we certainly can be confident that patients should be treated to the established targets.

Dr. Chan: Clinical trials show a strong correlation between the reduction in LDL and the reduction in risk of vascular events. The message from these data is that the target for high-risk patients is an LDL <2 mmol/L. Even lower levels may be more effective. We now have data from JUPITER, which achieved a median LDL level of 1.4 mmol/L, that levels substantially <2 mmol/L are safe.

Dr. Hoag: The SEAS study does not change the evidence that patients at high risk for CV events benefit from LDL levels <2 mmol/L. There are also natural experiments in nature, such as genetic hypobetalipoproteinemia and hypocholesterolemia associated with some PCSK9 deficiency states, that inform us very low levels of LDL are safe. The SEAS trial does not alter that. The LDL treatment goal of <2 mmol/L is appropriate in reducing CV risk.

Subsequent analyses of the association between active therapy and increased cancer risk in the SEAS study have suggested it is a chance finding. Are there any other examples of studies that found an increased risk of cancer which also turned out to be a chance finding (whether in cardiology or other areas of medicine)?

Dr. Jones: There have been a number of previous studies in which data have generated concern about cancer that was later dispelled. In the CARE (Cholesterol and Recurrent Events) trial, which included a relatively small proportion of women, there were 10 breast cancers in the statin arm vs. zero in the placebo. Now, we have more than 10 years of data to discredit this association. In the recent JUPITER trial, there was actually a reduction in cancer deaths among those on the statin. In addition to the fact that we would not expect to see any change in cancer risk over the relatively short periods of a lipid-lowering trial, the analysis by Peto et al. that appeared in the same issue of the New England Journal of Medicine with the SEAS results was compelling. There is really no signal of a cancer risk with lipid lowering in the enormous pool of long-term data we now have available.

Dr. Fitchett: This is not the first time that there has been an association between a therapy and cancer that was subsequently found to be due to chance. When results of SEAS was made available, a number of studies using much larger pools of data attempted to confirm the link between the active treatment and cancer, but these studies were uniformly negative. It is notable that there was also no strong pattern between the treatments and a specific type of cancer. The relationship between lipid lowering and cancer makes little sense from a mechanistic perspective and the patterns of the cancers in SEAS were not consistent with what is known about malignancy.

Dr. Teo: The risk of cancer cannot be dismissed but in reviewing the SEAS study and the accompanying paper by Prof. Richard Peto, I do not believe that the association between active treatment and increased cancer risk in this study is proven. One can remember in the early 1990s when it was believed that there was increased risk of cancer with lipidlowering therapy, which was subsequently found not to be correct.

Dr. Hoag: This is an issue that has been evaluated multiple times, including early on in the 1988 Canadian Cholesterol Consensus Conference, held in Ottawa. Although there have been suggestions that statins increase the risk of various adverse outcomes, including cancer, it has been clearly demonstrated in the large body of evidence we now have available that these were chance findings and incorrect. When the data were revisited after results of SEAS were reported, it was widely concluded again that statins are not associated with an increased cancer risk.

Is there any epidemiologic or clinical evidence to support a protective effect of phytosterols on cancer? If so, how strong is the evidence?

Dr. Fitchett: The theory that inhibition of dietary sterols by ezetimibe may have a role in causing cancer is not any more strongly supported by experimental data than it is by the clinical data. There is simply no pattern that ties any of this together.

Dr. Jones: Ezetimibe prevents absorption of all sterols, not just cholesterol, in the intestine. This produced the hypothesis that ezetimibe may inhibit absorption of plant sterols important for preventing neoplastic growth. However, there are a number of problems with this hypothesis. One is that we would not expect any change in dietary absorption to alter rates of cancer in two years. Also, the most common cancer in SEAS was of the skin. Such a theory would make more sense if the cancer was of the gastrointestinal system or the prostate, particularly if there was a clear pattern for an increased risk of one of these cancers. None of this really adds up.

Dr. Hoag: Any specific protective effect of phytosterols on cancer is not a portion of the literature I am familiar with. From a population perspective, there are certain European countries that have supplementation of certain food sources with phytosterols in an attempt to reduce the burden of CV disease. I am not aware of any literature that suggested this has led to any increased cancer.

In patients unable to achieve current recommended LDL goals on statin therapy alone, do you feel that ezetimibe remains an appropriate add-on medication to achieve targets?

Dr. Teo: I believe in the lipid hypothesis and therefore that reduction of LDL results in decreased vascular risk in secondary prevention. Thus, if current recommended LDL goals on maximum therapy alone are not achieved, I believe that ezetimibe remains an important add-on to achieve the targets.

Dr. Fitchett: The evidence supporting current LDL targets is overwhelming. Statins are well tolerated and very effective, but not all patients can achieve goals on these agents alone. In those who cannot, ezetimibe is also well tolerated and very effective for additional reductions in LDL. This drug is certainly indicated when patients have not reached goals despite lifestyle changes and maximum doses of statins.

Dr. Chan: It is important to achieve a marked reduction in LDL-C levels in patients at high risk of CV events, and fortunately we now have potent statins to achieve this in many of these patients. In the patients in whom adequate reduction of LDL-C levels (<2mmol/L and at least a 50% reduction regardless of the starting level) are not achieved on statins, ezetimibe can be used in addition to a statin. We eagerly await the results of the IMPROVE-IT study to confirm that the addition of ezetimibe reduces ischemic events in an evidence-based trial.

Dr. Jones: There are certainly individuals who cannot reach LDL goals on statins alone. In my practice, I consider ezetimibe among several adjunctive therapies to get patients to goals. In general, I add a bile acid resin first, ezetimibe second and then niacin, although this order might change depending on other considerations, but patients who do not achieve the target reductions on a statin alone should be given additional therapy.

Dr. Hoag: Ezetimibe is one of several appropriate strategies to achieve LDL targets. With any therapy, including statins and ezetimibe, it is important to be prepared for variability in response. This can be achieved by re-evaluating lipid levels after a short course of therapy to verify that the individual response is similar to that observed in the general population. I think it is important to recognize interpatient differences in the major components of drug and cholesterol metabolism, including absorption. This is useful in providing a more individual care strategy and intervention program for patients.