FDA looking into Vytorin and cancer risk, but interim analysis reassuring for patients to continue with medication
Rockville, MD - The Food and Drug Administration (FDA) has issued a statement informing clinicians about its investigation into the possible link between Vytorin, a combination of ezetimibe and simvastatin (Merck/Schering-Plough Pharmaceuticals), and an increased risk of cancer [1]. At this point, however, the FDA states the review should not prompt patients to stop taking Vytorin or any other cholesterol-lowering drug.
The ongoing safety review is based on findings from the Simvastatin and Ezetimibe in Aortic Stenosis (SEAS) trial, a study that missed its primary end point when the combination failed to reduce the composite end point of aortic-valve and cardiovascular events. Although there was a reduction in ischemic events, significantly more subjects treated with Vytorin developed and died from all types of cancer when compared with those taking placebo during the five-year study.
Presented in July by steering committee chair Dr Terje Pedersen (Ulleval University Hospital, Oslo, Norway), the SEAS data prompted a review of interim data of two other ongoing trials of Vytorin. The analysis of the Study of Heart and Renal Protection (SHARP) and the Improved Reduction in High-Risk Subjects Presenting with Acute Coronary Syndrome (IMPROVE-IT), performed by Sir Richard Peto (Clinical Trials Service Unit [CTSU], Oxford, UK), showed no increased risk of cancer with the combination of simvastatin plus ezetimibe. Based on these interim data, the FDA said patients should not stop taking Vytorin.
The FDA expects to see a final SEAS study report from the drug sponsors in three months, and completion of its review of the clinical trial data and other information will take another six months. "As soon as this review is complete, FDA will communicate our conclusions and recommendations to the public," according to the FDA's report.
In its letter, the FDA adds it is aware of previous studies suggesting a link between low on-treatment cholesterol levels and an increased risk of cancer. Data from large prospective statin studies, however, have shown no difference in cancer incidence between the active and placebo arms. A study published this week and reported by heartwire confirmed the inverse association between on-treatment LDL-cholesterol levels and cancer in statin-treated patients, but a similar relationship between LDL-cholesterol levels and incident cancer among control patients was also observed. Investigators concluded that statins, despite significantly reducing LDL-cholesterol levels, are not associated with an increased risk of cancer.
US House Energy and Commerce Committee continues to ask questions
In July, the US House Energy and Commerce Committee sent a letter to the FDA asking for the results of the Peto analysis. Now the committee has sent letters to Merck and Schering-Plough asking for the same analysis, and according to a report in the Wall Street Journal, committee chair Rep John Dingell (D-MI), and Rep Bart Stupak (D-MI), chair of the Subcommittee on Oversight and Investigations, say they are "troubled" by the confusing data the companies released about the study [2]. The confusion stems from a discrepancy in the cancer numbers presented by Peto and numbers issued by the sponsors. The lawmakers also questioned Peto's independence as a consultant to the study, because Dr Rory Collins, also from the CTSU where Peto is located, is carrying out the SHARP trial, a study funded by Merck and Schering-Plough.
Merck and Schering-Plough refuse to release the Peto analysis to the committee without assurances that it won't be leaked. The discrepancy with the cancer data, they said, is the result of different time periods measured, calling the suggestion that the independent interim analysis was influenced by Merck and Schering-Plough insulting to Peto and entirely untrue.
Expect more news on Vytorin soon. Data from the SEAS trial will be presented at the European Society of Cardiology (ESC) 2008 Congress in Munich, Germany. The presentation is scheduled for Tuesday, September 2, 2008 during an ESC hotline session.
 |
 |
 |
 |
|
 |
 |
 |
 |
 |
|
|
August 23, 2008 07:55 (EDT)
|
|
 |
 |
 |
 |
 |
|
biological plausibility
Ezetimibe reduces absorption of plant sterols. Plant sterols have been linked to reduced cancer risk in some epidemiological studies, in vitro studies, and animal studies. Here is a link to the most recent review article, by Bradford and Awad:
http://www3.interscience.wiley.com/journal/114107323/abstract?CRETRY=1&SRETRY=0
I am not at all reassured by Richard Peto's analysis of cancer incidence in SHARP and IMPROVE-IT. The patients in those trials were only randomized 1-2 years ago, not long enough for most cancers to develop. In addition, if you pool the cancer deaths from the three trials, there is a statistically significant increase in cancer deaths in the ezetimibe group.
Given the substantial questions regarding the safety and efficacy of ezetimibe, the FDA should expedite its review. In addition, patients on Vytorin or Zetia should strongly consider switching to a statin, including if necessary a higher dose or more potent statin, to control their cardiovascular risk. |
|
 |
|
|
|
August 24, 2008 07:50 (EDT)
|
|
|
|
|
|
|
|
give the latency period of most carcinogens....
Given the latency period of most carcinogens, which can be measured in decades, how could a drug like ezetimibe produce an increase in cancer death over a follow-up measured in thirty some months?
Answer: it can't. Patients who got cancer in SEAS had subclinical disease to begin with. Smaller trials tend to increase the risk of chance bias (remember the NINDS trial of tPA for stroke??) -- meaning chance imbalance in confounders between groups (in this case, inherent cancer risk or subclinical neoplasia).
The above mechanism of action regarding plant sterols is interesting and warrants attention. Plant sterols also reduce cardiovascular risk (patients with the highest serum levels have low event rates), so if there is a causal link, ezetimibe should also be increasing c.v. event rates too. In SEAS, the opposite happened. |
|
|
|
|
|
August 24, 2008 03:32 (EDT)
|
|
|
|
|
|
|
|
response to D Hackam
In SEAS there was a 22% risk reduction for atherosclerotic events. However, SEAS was a trial of ezetimibe 10/simva 40 vs placebo. Therefore, all of the benefit could have been from simva. We really don't know. There was a 61% decrease in LDL in the Vytorin group, which makes the 22% risk reduction kind of underwhelming (less than in HPS and 4S).
Whether plant sterols increase or decrease cardiovascular risk is controversial. Oliver Weingarten has done some interesting work on this, including an article that was published last year in JACC, "Vascular effects of dietary supplementation with plant sterols."
I do not understand how the reduction in the secondary endpoint in SEAS, even if it could be attributed to ezetimibe, would tell us anything about plant sterols and cancer risk. It is quite possible that the reduction in serum plant sterols decreases cardiovascular risk but increases cancer risk.
The fact that ezetimibe decreases absorption of plant sterols is well known. In fact, it is used as a treatment for sitosterolemia. |
|
|
|
|
|
August 24, 2008 07:20 (EDT)
|
|
|
|
|
|
|
|
No evidence that EZ reduced CV rates in SEAS
D Hackam, you state that if the theory about plant sterols was correct then EZ should be increasing the rate of CV events. In the SEAS trial we don't know that that did not happen. Perhsps EZ is increaing events? There was no comparison of Vytorin with Simva in the trial. Perhaps all the risk reduction (which was very low considering the LDL reduction) came from simva? Maybe the risk reduction in CV events would of been even lower with simva alone and EZ was increasing the risk of an event. The only way we can know that EZ is not increasing risk is if there was an EZ/placebo or Vytorin/simva study. The ENHANCE (Vytorin/simva) study showed no advantage in CIMT reduction for the Vytorin arm.
A second point is that the cholesterol guidelines state to use plant stanols and sterols early. If one can get the benefit of a 10-14% LDL reduction with stanols/sterols (with low economic costs and no health risks) there does not seem to be a need to use ezetimibe.
All lipid subfraction data show niacin to benefit the lipid profile to a greater degree than other compounds. There are some differences between statins, but the star of the subfraction show is niacin.
There was also an article in 2005 comparing the endothelial benefits of 10mg/simva versus 10 mg EZ (similar LDL reductions of 15.5% and 15.6%). There were no endothelial benefits for EZ.
My interest is experimental design and statistical analysis so I can make no comment on the mechanism of action in regards to the increased cancer rates, but it is an interesting hypothesis. |
|
|
|
|
|
August 24, 2008 08:23 (EDT)
|
|
|
|
|
|
|
|
most likely sterols are a surrogate marker
The studies showing that sterols are associated with reduced cardiovascular and cancer risk are typically observational, and are very similar to dozens of studies published over decades showing that beta-carotene, tocopherol, vitamin C, etc, etc in the blood are associated with reduced risk. When these were subjected to rigorous RCTs, these substances did not decrease c.v. or cancer risk (and in some cases actually increased it -- eg beta carotene/vitamin A in smokers, and vitamin E in heart failure).
Most likely high plant sterols in serum is a surrogate marker for healthier behavior in non-smoking, non-alcohol-imbibing, less-red-meat eating people who exercise and watch their weight (and therefore in many ways are at lower risk of cancer). They could well be an epiphenomenon. The mechanism you postulate is interesting but needs to be proven in a large randomized trial.
The finding in SEAS is not dissimilar to similar findings in PROSPER with pravastatin, which were later disproved in both large individual patient meta-analyses of pravastatin and in 15-year follow-up of WOSCOPS. I believe what we are seeing is a chance finding, but I am waiting for data from IMPROVE-IT and SHARP for further safety verification of ezetimibe. |
|
|
|
|
|
August 25, 2008 09:06 (EDT)
|
|
|
|
|
|
|
|
Primum non nocere
While the FDA is investigating ezetimbe, why would physicians and patients take the risk unless all other safer alternatives were tried and failed? Unless the benefit/risk ratio still favors taking ezetimbe, why accept any doubt? Primum non nocere. |
|
|
|
|
|
August 25, 2008 10:54 (EDT)
|
|
|
|
|
|
|
|
it's the LDL achieved (not the drug) that counts
A devil's advocate position:
In meta-regression after meta-regression of randomized trials, it does not matter HOW you lowered LDL but HOW MUCH you lowered it.
Therefore, if you achieved striking reductions - with diet, statins, bile acid sequestrants, niacin, a combination of the above, or even ileal bypass - cardiovascular risk was proportionately reduced.
This is why I think that adding ezetimibe on top of a statin to achieve LDL < 70 (or using ezetimibe as first choice in statin-intolerant) is still a very appropriate option.
In this case, the key is not the drug but the risk factor (LDL). |
|
|
|
|
|
August 25, 2008 12:57 (EDT)
|
|
|
|
|
|
|
|
HDL/meta-analysis and such
It had always been assumed that raising HDL by whatever method available was appropriate. We now know better.
Rosuvastatin and Simva/EZ both were entered into longterm and very large clinical trials at approximately the same time. The rosuva trial was stopped 3 years early due to benefit. The Simva/EZ trial? Recently, announced an extension of two additional years.... A drug that gets up to 61% LDL reduction needs this much work to show benefit over simva? The question is not whether the combination can lower events, but whether the combination is better than than the generic simva going solo at reducing events.Five plus years into the product there is no evidence of benefit of the combination over simva. |
|
|
|
|
|
August 25, 2008 02:30 (EDT)
|
|
|
|
|
|
|
|
JUPITER:IMPROVE-IT::Apples:Oranges
JUPITER and IMPROVE-IT are two very different trials (placebo-controlled vs. active-comparator) in two very different patient populations (high hsCRP but otherwise healthy vs. post-ACS). No way you can compare the two studies/timelines like that. |
|
|
|
|
|
August 25, 2008 08:19 (EDT)
|
|
|
|
|
|
|
|
the devils in the details....
being an advocate (Dan, sic)
I agree it's interesting to see nicotinic acid, then bile sequestrant, then bypass (not coronary but ileal) and then finally statin showing CV event reduction and in some cases CV mort and all cause mort reductions. Fibric acids have shown benefit in some studies predominantly Met Syn or DM pts (primary and/or secondary).
Now we have primary 'low risk' populations with high hscrp showing benefit with monotherapy. (jupiter and agree apples and oranges - many lipid studies can't be cross compared as the pts had different base risk, different profiles/demographics, different lipid values wide ranging)
I guess the real debates are:
1. If using combination therapy - what would be best (statin/eze, statin/bile resin, statin/gastric bypass, statin/N-3, statin/nicotinic acid, etc. etc..) Our approach again is to treat by the guidelines - if LDL and NHDL can't be reached with optimal statin therapy then eze, nico, feno, bile resins are options. (i am not avoiding stanols)
The next question which has been raised for years and now resurfaces is: How low to go with LDL 100, 70, 40, etc...? How to you measure risk? (imaging, expanded lipids, other markers)
In the perfect world we would have the NIH, ACC, AHA, NHLBI, CDC study that had baseline statin LDL to 70 and 50 by baseline statin in very high risk pts and then add either placebo, feno, nico, bile, n-3, eze and see which combo worked best. WE would have simple imaging studies to see if knowing and treating the image lowered events. We would have simple cardiac markers like expanded lipid (nmr, vap, berk) and hscrp and lppla2 and see if knowing and treating these values appropriately would reduce events.
But that's our santa clause wish list and if we are really good perhaps the insurance industry who has sworn to help us be healthy or our govt who looks out for our health and welfare will do such a thing (oops, now the devil has entered the discussion again - well it is in the details) |
|
|
|
|
|
August 28, 2008 12:33 (EDT)
|
|
|
|
|
|
|
|
Just give me one trial
I would like to see any clinical evidence that ezetimibe is efficacious in reducing clinical events prior to prescribing it. |
|
|
|
|
|
August 28, 2008 01:04 (EDT)
|
|
|
|
|
|
|
|
That sounds reasonable
With so many agents in the lipid arena showing event reduction and vascular benefits - I think that is a reasonable request and concern.
We will know by 2012. |
|
|
|
|
|
August 28, 2008 10:53 (EDT)
|
|
|
|
|
|
|
|
Approach I take
If someone is on ezetimibe with a statin and at goal, I will not stop the drug unless they ask it to be stopped. I am not starting anyone on zetia with a statin until IMPROVE-IT comes out. I have started using more niacin with greater success and tolerance than I expected.
Interesting that I gave a talk to GPs for a board review and what I told them was that for now, I would reserve ezetimibe for patients unable to tolerate/take higher dose statin AND failed fibrate or niacin to attempt to get numbers to goal.
Thoughts? What is everyone else doing with patients already on ezetimibe? Not able to tolerate high dose statin and combo with fibrate or niacin?
Daniel |
|
|
|
|
|
August 28, 2008 03:11 (EDT)
|
|
|
|
|
|
|
|
thoughts
I think fibrate would be more for lowering trigs or raising HDL, whereas ezetimibe works differently - lowering LDL. I do not see these two therapies as similar, but in some sense I would use them interchangeably (fibrates preferred over ezetimibe) to modify cardiovascular risk by going after different lipoproteins.
Interestingly I admitted an 80 y.o. man over the weekend on lipitor 40 mg and ezetimibe 10 mg who presented with pneumonia and rhabdomyolysis with acute renal failure. He is getting better without dialysis, but I wonder if combination lipid-lowering therapy was the culprit. (He may have an underlying malignancy causing pancytopenia as well)
Niaspan is a real pain here because it is not covered under provincial formulary. My experience with prescribing it in patients over 65 in the past couple months has entailed writing repeatedly to the government for coverage, and giving lots of free samples.
My approach is similar to yours - I rarely stop it, but I do try to maximize the statin based on the accrued experience of 7 hard endpoint RCTs (of low vs high statin dosing) and 3 meta-analyses thereof. |
|
|
|
|
|
August 28, 2008 10:54 (EDT)
|
|
|
|
|
|
|
|
low LDL cholesterol and cancer
I think it is the best policy I personaly following in my practice. Patients divided into 3 groups when lipids found high. Group I treated for 2 months if they respond and LDL lowered to arround 100 mg/dl then cut the dose to half for a month and stop drugs keeping in mind the patient is strict on diet free of fat and carring good exercises.
Group II intermitant treatment one month treatment every 3months with diet and LDL control.
Group III Familial dyslipidemia with very high risk of CVS disease should carry on with treatment with regular blood lipids check-up every 3 months. |
|
|
|
|
|
August 29, 2008 12:23 (EDT)
|
|
|
|
|
|
|
|
the chosen drug does count
It doesn't matter how you lower the LDL, it just matters that you do?
I challenge anyone who believes this theory to use picnogenol, garlic, and/or guggul lipid as first line therapy to lower cholesterol. Each of these three compounds lower LDL. . .and have no effect on mortality.
Who here would use alpha-blockers as first line treatment for HTN? Alpha-blockers lower BP but aren't so hot at preventing HTN related complications.
It does matter how you achieve a goal.
It does matter what medications you pick.
Zetia may well fall into the "it lowers LDL but doesn't do much else" category. Time will tell, but until I have hard data, and until this noisy little side effect of cancer gets cleared up, I'm not using it.
|
|
|
|
|
|
 |
Comment 
Share 
Cite
Print
Text size
