Archief Roodbont

Medscape contributor Linda Brookes, MSc, spoke with Dr. Harris about the COMBOS study.

Medscape: What is non-HDL-C and why is it important in the context of the COMBOS study?

Dr. Harris: Triglycerides are primarily carried by very-low-density lipoproteins (VLDL). People with elevated triglycerides typically have an associated increase in VLDL remnant particles. Because VLDL remnants appear to have atherogenic potential similar to that of LDL, The VLDL-C level can be added to the low-density lipoprotein cholesterol (LDL-C) value to produce a new index of risk and secondary target of therapy: non-HDL-C. The non-HDL fraction in people with high triglycerides (ie, a large number of VLDL remnant particles) is greatly enriched with VLDL-C relative to LDL-C compared with patients with a normal triglyceride level. The sum of LDL-C and VLDL-C, the non-HDL-C value, is really important in predicting risk because these people with elevations in non-HDL-C are at high risk even if their LDL-C is not terribly high. We are focusing more and more on non-HDL-C as a risk factor. It is a very good marker of risk, particularly in people with high triglycerides, and even more in women, because high triglyceride (or high non-HDL-C) levels in a woman attribute more risk for heart disease than the same level does for a man.

Medscape: Is it as easy to measure non-HDL-C as LDL-C?

Dr. Harris: It is very simple. Just subtract HDL-C from the total cholesterol value. Both measurements are routine in the laboratory.

Medscape: Is non-HDL-C as good as or even better than LDL-C at predicting cardiovascular disease?

Dr. Harris: There is growing consensus that non-HDL-C is better. People have just now started looking at it, but it makes sense, because it includes LDL-C and adds additional cholesterol from particles that we know can cause heart disease, such as intermediate-density lipoprotein (IDL, the other name for VLDL remnants) and VLDL, and puts them all into one number. So as people look more carefully, they are finding that the non-HDL-C predicts as well as, if not better than, LDL-C.

Medscape: Is non-HDL-C still as reliable a predictor in people with triglyceride levels as high as 500 mg/dL?

Dr. Harris: Non-HDL-C is not as useful when triglyceride levels are ≥ 1000 mg/dL, because people with levels this high have other problems that are not heart disease-related but are related to the risk for pancreatitis. However, there are not many people with triglycerides > 500 mg/dL, let alone > 1000 mg/dL. But yes, in patients with triglycerides up to 500 mg/dL, the non-HDL-C is quite useful.

Medscape: The patient population in the COMBOS study had LDL-C levels within 10% of goal but triglyceride levels between 200 and 499 mg/dL. Are these patients that physicians encounter often?

Dr. Harris: Yes; there are many more patients like this than patients with triglycerides > 500 mg/dL.

Medscape: Presumably they would have had dietary counseling at the start of their original statin treatment.

Dr. Harris: Yes, and that would not change.

Medscape: So these patients start on a statin and get the LDL-C level down, but they are left with a residual atherogenic component. Would there be any other option than omega-3 ethyl esters (Lovaza)?

Dr. Harris: Another drug that targets LDL-C, such as ezetimibe, would not make much of a difference in the level of non-HDL-C. It might drop it another 10 or 20 points, but in people with triglycerides in the 200-500 mg/dL range, much of the non-HDL-C is from the VLDL, particularly remnant VLDL, not the LDL which ezetimibe affects. Patients with triglycerides in this elevated range and taking statins typically have LDL well under control, so adding another LDL-targeting drug (like ezetimibe) will not be helpful.

Omega-3 ethyl esters (Lovaza, formerly Omacor), was first submitted to the FDA because there was good evidence that it could lower triglycerides significantly in people with triglycerides > 500 mg/dL. That was the study that we did in our laboratory over 10 years ago.^[3] We found that Lovaza 4 × 1 g capsules per day over 4 months lowered mean triglyceride concentrations about 45% in people whose triglycerides averaged around 800 mg/dL. These data were strong enough for the FDA to approve the indication for treating that patient population, and Lovaza (Omacor) was approved in the United States in 2005. At the time of the approval for the treatment of very high triglycerides, the FDA also issued an Approvable Letter citing the requirements for Lovaza to be approved for the treatment of elevated non-HDL-C and triglycerides (200-499 mg/dL) in adult patients concurrently taking a statin at or near LDL-C treatment goal, as an adjunct to diet. The COMBOS study was carried out to address this patient population. It sought to show that Lovaza would not reverse the effect of the statin and that a good overall lowering in non-HDL-C could be achieved. LDL-C and VLDL-C fell by 9.0% in the Lovaza/simvastatin group, and fell by 2.2% in the simvastatin/placebo group, a net difference of about 7%; and, as expected, Lovaza produced a significant fall of 29.5% in triglycerides in these patients whose LDL-C was already well controlled with simvastatin (40 mg).

Medscape: The NCEP/ATP III goal for non-HDL-C is 30 mg/dL higher than the LDL-C goal. What would the LDL-C goal have been in these patients?

Dr. Harris: In some of the patients it would have been 100 mg/dL and in others 130 mg/dL, depending on the other CHD risk factors present (eg, hypertension, smoking, personal CHD history).

Medscape: The results of the study appear to show a slight increase in LDL-C. How can this be explained?

Dr. Harris: Modest increases in plasma levels of LDL-C are seen with Lovaza monotherapy in people with high triglycerides (≥ 500 mg/dL), and the FDA wanted to be sure that this did not occur in patients on statins. Ideally LDL-C levels would decrease or at least not change when Lovaza was added. Because the study was so large, however, the clinically insignificant change in LDL-C (an increase of 0.7%) was statistically significant when compared with the 2.8% drop in LDL-C observed in the placebo group. When the results of the study were submitted to the FDA in June, the FDA issued an Approval Letter permitting the addition of select study data within the clinical studies section of the label for Lovaza, but it did not approve the additional indication. So it is a confusing situation, to be able to talk about the COMBOS data (because they are in the package insert), but not to have Lovaza indicated in this population, What is not confusing, however, is the effect of Lovaza on non-HDL-C, which fell by 9% in the Lovaza/statin group compared with 2% in the simvastatin/placebo group, for a net decrease of about 7%.

Looking at the bottom line in the study, the Lovaza group started with median LDL-C 91 mg/dL at baseline and finished the study at 88 mg/dL. This looks like a decrease, but because the median percent changes were reported it looks like an increase. This is a puzzling message. The median percent change is the percent change that was in the middle of the group. In other words, for each individual in the study, a percent change in LDL-C was calculated from baseline to end of treatment. Those percent changes were ranked from the greatest to the least. The value in the middle happened to be a person who had a 0.7% increase. So that is recorded as the median percent change, even though the median LDL value actually decreased. If you look at the placebo group, the median LDL-C went from 88 mg/dL to 85 mg/dL, the same decrease as was seen in the Lovaza group; but it is a -2.8% median percent change in the placebo group, when compared with the median 0.7% increase in the Lovaza group, that together produced the final reported net increase 3.5% in LDL-C. It is only an increase because LDL-C in the placebo group decreased, not because LDL-C in the Lovaza group increased. Because the sample size was so large (254 patients), this small difference was statistically significant, and thus the authors had to conclude that the addition of Lovaza caused a significant increase in LDL-C. Another problem was that the P value for the difference between groups was .052, which does not actually meet the < .05 criterion for declaring statistical significance. Hence, strictly speaking, there was no statistically significant increase in LDL-C in this study. This would have been even more evident had the study not been so large. If the study had enrolled perhaps 20 less people, that P value would have probably been clearly in the nonsignificant zone. It would not have reached statistical significance and the investigators could have concluded that they did not get a significant change in LDL-C. Here is an example where statistical significance and clinical significance diverge.

Medscape: So it was a statistical accident?

Dr. Harris: Yes, in a way. The study enrolled a lot of people so that it could show a significant non-HDL-C effect; but you do not want to have so many people that you can detect a tiny difference in LDL, and they actually did. So it was a rare example, I think, of a study being too large. I also think it is very important to point out that a percent increase in LDL-C may be meaningless and driven more by the denominator than the numerator. For example, a 5-mg/dL increase in a person whose LDL is 100 mg/dL is a 5% increase, but the same increase in another person with LDL-C 50 mg/dL is a 10% increase. It is still just a 5-mg/dL change, but the percent increase can be huge, depending on the starting level. As the denominator gets smaller, the percent increase becomes huge. The LDLs were relatively low in this study (because of the simvastatin) and that contributed to this outcome. However, that is the way the FDA wanted to see the data, I believe.

Medscape: The patients in the COMBOS study were not particularly high-risk and they had already lowered their LDL-C.

Dr. Harris: Their LDL-C levels were at target at baseline, and after Lovaza they were all still at the same target. A typical target for these patients is 100 mg/dL, so on average, they were below target.

Medscape: Some of the patients switched to simvastatin from a different statin at the start of the study. Would that have had any effect on LDL-C levels?

Dr. Harris: Not likely. Because most patients under "real life" clinical care only imperfectly comply with their statin prescription, it's likely that the subjects' LDL-C was under better control in the study than before it. This is because not only was the simvastatin (40 mg) provided at no cost, but patients also are more compliant with medications when they are in a clinical trial than when they are not. So although their LDL-C levels were probably improved compared with pre-study, I seriously doubt that this would have altered the study's outcome.

Medscape: Has Lovaza been studied with other statins besides simvastatin?

Dr. Harris: Yes; in a small study from Australia^[4] some years ago it was shown to be effective in combination with atorvastatin. Actually, a study similar to COMBOS is being carried out right now with atorvastatin and Lovaza. Two hundred patients with high to very high triglycerides and non-HDL-C above NCEP/ATP III goals were randomized -- after a 4-week diet-only lead-in period -- to atorvastatin 10 mg/day plus either Lovaza 4 g per day or placebo for 8 weeks. After the initial 8-week treatment period, the dose of open-label atorvastatin is titrated to 20 mg per day for an additional 4 weeks, and at week 12, a second titration of atorvastatin to 40 mg is maintained for an additional 4 weeks. The primary outcome measure, again, is change in non-HDL-C, with changes in other lipids and biomarkers as secondary outcome measures. That study will close in about a couple of weeks and then the data have to be analyzed. The results are expected within a year.

Medscape: Would you expect the outcome to be the same with an equivalent dose of another statin plus Lovaza?

Dr. Harris: Yes. As alluded to earlier, there have been other studies with a variety of statins plus Lovaza, and the outcome was much the same in terms of the effects on non-HDL-C. In one study^[5] the addition of Lovaza lowered LDL-C even further, which was surprising. The patients were on various dosages of different statins. In another study,^[4] when they were on atorvastatin adding a different patient population, there was no change in LDL-C. I do not think there will be a change in LDL-C in the Lovaza/atorvastatin study either.

Medscape: Are there any other lipid results in COMBOS that were of particular interest?

Dr. Harris: The apolipoprotein B (apo B) is a different way of looking at non-HDL-C because the particles that constitute non-HDL-C (LDL, IDL, and VLDL) all contain 1 apo B molecule per particle, so changes in apo B follow those in non-HDL-C. A statistically significant decrease was seen in apoB, and there was also a significant rise in HDL-C. This is a really nice pattern of add-on therapy for an agent like Lovaza that has no side effects and no drug interactions.

Medscape: Do statins and omega-3s have any additive effects?

Dr. Harris: I think they do. If you have a new patient who has high LDL-C and high triglycerides whom you put on a statin and Lovaza from the start, you are going to get a combined effect, because the statin is going to have some triglyceride-lowering effect, with Lovaza having a larger effect on top of that. What is suggested from the COMBOS study is that even though the triglycerides come down with statins (not measured directly in this study), the addition of 4 g of Lovaza will lower them further.

Medscape: The incidence and nature of adverse effects were reported to be similar in both groups in the trial. Fasting blood glucose was significantly increased with Lovaza/simvastatin compared with placebo/simvastatin (P < .002); was this clinically significant?

Dr. Harris: I do not think so. Compared with the net benefit of the fall in non-HDL-C and the rise in HDL-C, a 5-mg/dL change in glucose, which was about what it was, is irrelevant, especially when no change in fructosamine was observed.

Medscape: So would there be any reason not to prescribe Lovaza as an add-on therapy in a similar type of patient to those of the COMBOS study population?

Dr. Harris: There are reasons why you might not. There are convenience issues; some people find that taking 4 capsules a day is inconvenient. Some people get a "fishy burp" after taking the capsules, but this can be avoided by taking the capsules at bedtime. The alternatives are fibrates or niacin, both of which have more drug-drug interactions and more potential side effects than Lovaza. Lovaza has no medically significant side effects, and there are other data suggesting that omega-3 fatty acids may reduce risk for arrhythmias and potentially even lower risk for Alzheimer's disease and depression. There are therefore significant "pleiotropic effects" of Lovaza that you would expect to benefit from in addition to triglyceride lowering.

Supported by an independent educational grant from Reliant Pharmaceuticals

References

# posted by roodbont @ 2:22 AM