Discussion

Are advanced lipoprotein methodologies that evaluate the density and size of LDL and HDL beneficial in further assessing risk in these patients?

Theodore Mazzone, MD, FACP: That is a controversial area. I believe that in routine clinical use, outside of research trials, non-HDL cholesterol and perhaps the measurement of apoB gives the needed information.

Henry Ginsberg, MD: Approximately 90% of patients with insulin resistance and diabetes have small dense LDL. Therefore, their LDL cholesterol level is not as indicative of how many LDL particles they have compared with a person who has a similar LDL cholesterol level but does not have diabetes or insulin resistance and has normal triglycerides. Knowing this, the approach to treating the patient with insulin resistance or diabetes is to lower the LDL cholesterol more than would be done in a person without diabetes or to consider non-HDL cholesterol or apoB. Estimating particle number is not necessary, as non-HDL cholesterol or apoB serve as an index of particle number.

Does apoB measurement add any value in risk assessment in a patient with diabetes on a statin who has an LDL level of 90 mg/dL, an HDL level of 36 mg/dL, and high triglycerides?

Alan R. Tall, MD: There may be some issues with standardization of apoB measurements across the country. Therefore, I believe it is more feasible to measure non-HDL cholesterol.

Mazzone: Patients with hypertriglyceridemia and low HDL are the population in which non-HDL cholesterol can be beneficial for predicting residual risk, as shown by population studies. Non-HDL gives the same index of risk as apoB, as it includes non-LDL atherogenic particles such as VLDL and IDL. Furthermore, because there are standardization issues with apoB, I prefer non-HDL for measuring residual risk.

Are apoB target levels of 80 mg/dL and 90 mg/dL in patients with diabetes low enough in terms of the atherosclerotic burden associated with atherogenic lipoproteins?

Ginsberg: No. Target values are chosen for several reasons, and these targets should be attainable. However, the goals that are set are often not achieved by the practicing physician. I believe an apoB target of 70 mg/dL is more appropriate than a target of 80 mg/dL. There are some epidemiological data that suggest a lower target is preferable. However, apoB has not been widely measured in large studies. The apoB/apoA-I ratio has been measured. However, this was mainly in cross-sectional and case control studies. I believe that if a person with diabetes achieves an LDL of 70 mg/dL, the apoB will probably not be much higher than 80 mg/dL or 85 mg/dL. This is an ongoing debate. The IMPROVE-IT study is planned to explore LDL cholesterol in the 50 mg/dL range compared with levels in the 60 mg/dL range. If the lower LDL cholesterol targets yield positive results, apoB targets may be more appropriately set at 70 mg/dL or 75 mg/dL.

Tall: The IMPROVE-IT trial will provide important answers if it reaches its endpoint. In epidemiological data, the curve relating LDL cholesterol to risk is not linear down to zero. Although it continues to decrease, the curve begins to flatten at LDL cholesterol levels of approximately 90 mg/dL.

Mazzone: However, there does not appear to be a disadvantage of achieving low LDL cholesterol levels. A clinician contemplating whether the statin dose should be reduced for a patient whose apoB level became 65 mg/dL after initiating the statin is not unusual. The answer is no. Developing therapeutic targets also includes consideration of drug costs and potential side effects. These are in constant flux as drugs become less expensive and new treatments are approved.

Robert H. Eckel, MD: Goals for HDL were established by the American Diabetes Associa­tion, not by the National Cholesterol Education Program (NCEP) Adult Treatment Panel Guidelines (ATP) III. The ATP guidelines are currently being updated, and ATP IV will take HDL targets into consideration.

What would be your choice of a second drug in the patient who has achieved LDL levels of 85 mg/dL but has triglycerides of 240 mg/dL and HDL of 36 mg/dL?

Mazzone: I may be more conservative than many of my colleagues, as I may not add a second drug to that patient’s treatment regimen.

Tall: I would consider using niacin. However, this is hampered by the recent early stopping of the AIM-HIGH study of high-dose extended-release niacin administered with statins, when the rate of clinical events was the same in both niacin and placebo-treated patients.¹² The Data and Safety Monitoring Board concluded that evidence indicated that continuing the trial would not produce a change in outcome, and the trial was stopped in May 2011.

Ginsberg: Although the Action to Control Cardiovascular Risk in Diabetes (ACCORD) lipid trial of fibrates added to statins in patients with type 2 diabetes had a negative overall result, the prespecified subgroup of patients who had triglycerides > 204 mg/dL and HDL < 34 mg/dL experienced a benefit in the primary outcome of first occurrence of nonfatal MI, nonfatal stroke, or death from cardiovascular causes. There was a nearly significant (P = .06) interaction between dyslipidemia status and efficacy of fenofibrate to reduce the primary endpoint.¹³ That subgroup analysis was consistent with subgroup analyses of several monotherapy fibrate studies. These patients should receive a fibrate, not patients with triglycerides of 150 mg/dL. Many patients with diabetes have triglycerides of 130 mg/dL to 150 mg/dL but have very low HDL levels, especially patients from different racial groups. Until recently I thought niacin would have been a good choice for a second agent in that group, but with the results of the AIM-HIGH study, I am uncertain. However, it is important to note that the results of AIM-HIGH have yet to be published.

What are the prospects of a cholesterol efflux capacity assay reaching the stage of clinical applicability?

Henry Ginsberg, MD: I do not know if there are plans to commercialize that assay. The technique adds to our understanding, but I do not think it will be generally helpful because, although it was independently predictive on a statistical basis, the resulting receiver operating characteristic curve revealed a minimally increased ability to predict coronary disease. In a patient with coronary disease and very high HDL, it may be interesting to see if a defect is shown.

Patients with acute coronary syndrome (ACS) are not the only group for whom HDL may become acutely dysfunctional; anyone who has atherosclerosis or who has had a cardiovascular event may have HDL that is bearing the burden of constantly removing atherogenic compounds from the vessel wall. This is an important research area that requires considerable additional investigation.

Can HDL particles be pro-inflammatory?

Ginsberg: This concept originated from observation of patients with an HDL of
60 mg/dL or 70 mg/dL who still have CAD. One group that has been studying this showed that some HDLs did not display the protective activities typically observed in laboratory models.¹⁶ However, other groups have not been able to validate this assay and there may be a variety of other explanations for these patients having coronary disease. Nevertheless, as the complexity of HDL is better understood, it is possible that a total HDL cholesterol level might not always be as revealing as population epidemiology suggests it is.

Alan R. Tall, MD: I believe it is likely that HDL becomes dysfunctional in certain circumstances. For example, it was recently shown that HDL loses its ability to stimulate NO release from endothelial cells in the setting of ACS.¹⁷ An important question is whether the HDL dysfunction is the primary defect or whether it is a secondary defect. I believe that under certain conditions, for example when people have ACS, HDL may be overwhelmed, becoming filled with oxidized lipids that exert pro-inflammatory functions. Therefore, I believe these are likely secondary events. The primary abnormality is not that dysfunctional HDL is causing disease, but it is simply an issue of the HDL becoming overwhelmed.

Based on results of the AIM-HIGH study, what should we recommend for patients with low LDL levels being treated with niacin?

Henry Ginsberg, MD: A patient on
1.5 g/day niacin will have a reduction in LDL cholesterol. If niacin is added to the regimen of a patient whose LDL is not at goal, or is between 100 mg/dL and 70 mg/dL with high triglycerides and/or low HDL, their LDL will be further decreased. I still believe that adding niacin in these settings will add benefit. On the other hand, if the LDL cholesterol is 50 mg/dL, with high triglycerides and low HDL, I am not sure I would recommend using niacin at this point.

Theodore Mazzone, MD, FACP: The major question with niacin is whether it will be helpful in patients who are adequately treated with statins to the appropriate LDL goal.

Alan R. Tall, MD: Patients in AIM-HIGH were treated with statins and ezetimibe, not just statins. In the real world, it is unlikely that patients will achieve such low LDL levels. AIM-HIGH was really an experiment, because the investigators wanted to see if niacin would change HDL and triglycerides and the result would be good.

Ginsberg: Other large niacin studies such as the Heart Protection Study 2 (HPS2)-THRIVE trial are ongoing, exploring LDL, triglyceride, and HDL outcomes. HPS2-THRIVE is more of a typical clinical trial testing the efficacy of a drug. At least one-half of the patients in HPS2-THRIVE are in China and India, which may confound the outcome if these populations have differences in insulin resistance. In the meantime, we should not abandon niacin as a drug, but focus more on its ability to lower LDL and triglycerides while increasing HDL.

How much effect can lifestyle modifications have on HDL?

Mazzone: HDL can be modified to a modest extent by lifestyle changes such as weight loss, increased aerobic exercise, limited alcohol intake, and cessation of smoking. Depending on the patient population, these changes can achieve a 5% to 7% change in HDL. However, that change in HDL could have a substantial effect on future risk. Therefore, although it is important not to discard lifestyle changes, they may not be sufficient to achieve levels needed for an obese patient with type 2 diabetes, for example.

What have we learned from the apoA-I mimetics and apoA-I Milano? What is the status of this research?

Tall: ApoA-I Milano is a naturally occurring variant of apoA-I, with carriers having unusually low HDL (10 to 30 mg/dL), while supposedly having less atherosclerosis than would be expected. Data from preclinical studies with recombinant apoA-I suggested that it produced rapidly occurring anti-atherosclerotic effects. In a pilot clinical trial with recombinant apoA-I Milano in combination with phospholipid in patients with acute coronary syndrome, a regression in coronary atherosclerosis volume compared with baseline values in the active treatment group occurred.¹⁷ Results from this and other studies support the theory that HDL functionality can be more important than its concentration. Synthetic peptides based on the 243-amino acid apoA-I molecule are expensive to prepare and require large doses to be delivered by infusion. Part of the research effort is, therefore, developing small peptide mimetics that can produce similar results.

How important is apoC-3, and does it relate to the lipid disturbance we see in diabetes?

Ginsberg: ApoC-3 is secreted by the liver and circulates on both triglyceride-rich lipoproteins and HDL, moving back and forth between them. The reason why people with high triglycerides have high apoC-3 is not completely understood. Data suggest that apoC-3 synthesis and secretion by the liver are increased with insulin resistance. ApoC-3 inhibits lipoprotein lipase, which results in more triglyceride-containing lipoproteins.

In addition, recently published studies suggest that apoC-3 has direct effects on a variety of vessel wall activities including monocyte adherence and nitric oxide production.¹⁸ Although understanding of its role is expanding, at the moment it is not a clinically measurable variable. However, it is of interest to pharmaceutical companies. The fibrates lower apoC-3 production, and pioglitazone has been shown to lower apoC-3 production in the liver. This may be related to the mechanism by which some of these drugs reduce triglycerides.

Is statin therapy appropriate for patients with type 1 diabetes?

Mazzone: In the absence of a lot of supporting data, my approach is that when a patient reaches the age when myocardial infarction begins to be more common, I put them on statin therapy. I may put a patient on statin therapy after 20 years of type 1 diabetes, depending on their age.

Are the relative differences between dalcetrapib and anacetrapib important (eg, their different effects on HDL and LDL levels)?

Tall: I think it is likely that the difference will be important. However, at this time we do not know exactly how. There has been discussion about the function of HDL perhaps being different in response to dalcetrapib compared with anacetrapib. Also, it is clear that anacetrapib has a greater LDL-lowering effect. The lipoprotein (a) (LP(a)) effect of dalcetrapib has not yet been reported. In the DEFINE trial, anacetrapib lowered Lp(a) by 36% compared with placebo, which is the most potent LP(a)-lowering effect described for an agent to date. These effects will most likely be important considerations in the overall drug profile.

Are omega-3 fatty acids an option to treat lipid disorders in patients with diabetes?

Mazzone: Omega-3 fatty acids are primarily useful for managing hypertriglyceridemia and have little effect on LDL or HDL. Therefore, I use them in patients where that isolated abnormality is still of concern. I believe there is little safety risk with the purified omega-3 fatty acids, although they can be very expensive.