Niaspan, HDL and Assumptions
A week and a half ago the National Institutes of Health pulled the plug on a research study assessing the effect of long-acting niacin (Niaspan) on the risk of heart attack and stroke among 3500 patients who had a combination of coronary disease and low levels of the protective form of cholesterol known as high-density lipoprotein (HDL). It turns out that preliminary evaluation of the collected data in the Atherothrombosis Intervention in Metabolic Syndrome with Low HDL Cholesterol/High Triglyceride and Impact on Global Health Outcomes (AIM-HIGH) study revealed that Niaspan not only had no beneficial effect but may also have increased the risk of stroke among the volunteers who were randomly selected to be on this medication.
To fully understand what kind of impact this news has you’ll need some background information. We learned many years ago that having a high level of LDL puts you at risk for heart disease and that treatment of LDL results in a precipitous fall in your likelihood of suffering a heart attack or stroke. The statin medications are now the standard of therapy for lowering LDL and have been proven beyond all reasonable doubt to decrease cardiovascular disease and prolong life.
HDL is known as the protective form of cholesterol. People born with high levels have a fairly low risk of cardiovascular disease; people cursed with low HDL tend to do worse and more frequently suffer heart attacks and strokes. This relationship has led many clinicians to assume that artificially raising HDL in the blood will result in decreased risk of heart disease, but, as we’re learning, this might not necessarily be the case.
Let’s say you have two people with identical health characteristics who differ only in their HDL level: patient X has an HDL of 25 mg/dl and patient Y is at 90 mg/dl. It’s established that Mr. X’s risk of a heart attack is considerably higher than Mr. Y’s, but what would happen if you were to artifically raise the HDL level in Mr. X from 25 mg/dl to 90 mg/dl with some sort of medication that does nothing more than induce a change in HDL? Would Mr. X now be at a risk level that is identical to Mr. Y?
Most people assume the answer is yes, but in doing so they confuse data from retrospective population studies with the type of answers we derive from prospective treatment trials. The real question here is this: Is a low HDL level the actual culprit in this equation or is low HDL simply a marker of higher risk?
Analogous to this would be a scenario involving hair color and risk of skin cancer. We know that people with red hair are at higher risk than brunettes of developing skin cancer in response to sun exposure, but do we assume that the hair color is the causative agent in this relationship? Could we cut the risk of skin cancer by rounding up all the redheads and dying their hair brown?
Questioning the precise relationship between cholesterol levels and heart disease is not without precedent. Statins do an excellent job in reducing both LDL levels and risk of heart problems, but how are these two effects related? Does the drop in LDL cause the drop in heart attacks? Again, this is what most people assume. However, history throws a small wrench into this logic.
Many years ago—before the development of lovastatin, pravastatin, and simvastatin, but after the damaging effect of high cholesterol was discovered—surgeons took a crack at lipid therapy by excising part of the small bowel in patients with very high cholesterol levels. The distal ileum is responsible for absorption of cholesterol-rich bile secreted by the liver into the gastrointestinal tract. Removal of this part of the piping results in a large volume of cholesterol ending up in the toilet instead of being recycled in the body. The procedure was highly successful in lowering LDL levels to a degree similar to our potent statin medications. The ultimate effect on cardiac outcomes, however, was surprisingly disappointing. In the largest study to evaluate this mode of therapy, the researchers needed five years to see even the slightest divergence in heart attack rate between the patients who underwent the invasive surgery and those who didn’t.
A similar tale exists with a more modern therapy. Ezetimibe (Zetia) chemically blocks the reabsorption of cholesterol in the bowel and is also a relatively potent LDL treatment. Many of you who follow scientific news know that the large studies evaluating the effects of ezetimibe treatment in people with heart disease have largely been disappointing. Surprising results like these lead us to believe that, while LDL lowering is important, the real issue is how you do it: LDL lowering with a statin is beneficial; via any other mechanism it’s less helpful.
Let’s now return to the HDL issue. Low HDL is bad and high HDL is good, but does raising HDL produce any benefit other than making the doctor feel better when she sees your cholesterol results? It should be noted that it has become almost commonplace for doctors to assume a benefit to Niaspan in low HDL patients, a “standard of practice” based solely on the assumptions I’ve identified above. Just a few weeks ago I was chided by a sales representative who markets Niaspan because I don’t prescribe this medication as much as other doctors do. When I pressed her for hard evidence supporting this therapy she was at a loss and returned to her original argument that I should embrace this approach because that’s what everyone else does.
Well, perhaps that will change. The National Heart Lung and Blood Institute sponsored the AIM-HIGH study in order to determine if artificially raising HDL level could cut the risk of heart attack and came up with a negative result. There is already chatter on medical websites speculating that we should take our patients off this medication or ban its use altogether. Such drastic action is, of course, premature. AIM-HIGH was only one of several trials underway to more definitively answer the HDL issue and we need to await more information before we can draw any conclusions on the utility—or lack thereof—of targeting HDL with Niaspan.