In the last decades we’ve made major strides in heart attack prevention through the use of blood pressure medications, smoking cessation, and statins – a family of cholesterol-lowering medications that have been proven to prevent heart attacks and strokes. Despite these advances, heart attacks remain the leading cause of death in the US. New medications to further decrease heart attack risk are being eagerly sought.

Allow me a brief digression to explain three important fat molecules in your blood. LDL is called “bad cholesterol” because it best predicts heart attack and stroke risk, meaning people with an elevated LDL have a higher risk of stroke and heart attack than people with normal LDL. Statins lower LDL. HDL is called “good cholesterol” because people with elevated levels have lower risk of stroke and heart attack. So low HDL is a risk factor for heart attacks. Finally, triglycerides are another fat molecule elevated levels of which predict increased risk of heart attack.

Given the enormous success of statins in lowering LDL and preventing strokes and heart attacks, hopes were high that a medication that raised HDL would have similar benefits. Niacin is known to raise HDL and lower triglycerides and has been used for many years in many patients to do just that. So the NIH decided to perform a study to measure its ability to prevent heart attacks.

The AIM-HIGH trial (All large clinical trials now have contrived, long, overly clever acronym names. You don’t want to know what AIM-HIGH stands for.) enrolled over 3,400 patients with a history of cardiovascular disease, low HDL and high triglycerides. That means they were at high risk of a future stroke or heart attack. Importantly, they all received simvastatin (Zocor) and had excellent LDL control. Half of the patients were randomized to also take extended release niacin (Niaspan) and the other half took placebo.

The patients were followed for an average of 32 months to see if the niacin group would have fewer strokes and heart attacks. As expected, the niacin group had higher HDL and lower triglycerides than the placebo group. But there was no difference in the numbers of strokes and heart attacks between the two groups. In fact, the niacin group had a tiny (but statistically significant) increase in strokes. This lack of benefit and small increase in harm led to the trial being stopped 18 months earlier than expected.

There was much wailing and lamentation, particularly at Abbott Laboratories which makes Niaspan and sold over $900 million of it last year. (Let that number roll around in your head for a minute while contemplating that the drug has no proven benefit. We skeptical evidence-based physicians who never prescribed it last year will be rewarded with an internal sense of accomplishment, a reward somewhat less tangible than $900 million.) Abbott is now in full damage-control spin mode suggesting that Niaspan may still have benefits in patients other than those enrolled in the trial. Sure. The trial was specifically designed to find a benefit by testing niacin on patients most likely to benefit – those at highest risk for heart attack and with high HDL and low triglycerides. Niacin didn’t help them, but it might help, who exactly? People with nasal allergies? People who need the Niaspan tablets to wedge under the short leg of a wobbly table?

Sadly, this isn’t the first bit of bad news in our attempts to help people by raising their HDL. A group of medicines called fenofibrates, which includes TriCor and Trilipix, also raise HDL and lower triglycerides and in recent trials have also failed to prevent heart attacks.

So what are we to make from these negative results? If low HDL is a risk factor for heart attacks, why doesn’t raising HDL decrease heart attacks? The answer is that risk factors are not causes; correlation is not causation. The fact that people who have low HDL have more frequent heart attacks than those with normal HDL doesn’t mean that the low HDL is causing the heart attacks. Some other unknown factor may cause both the HDL decline and the heart attacks, which would make HDL simply a marker of risk but with no effect on the heart attacks themselves. For example, ice cream sales in rural towns might be correlated with livestock deaths. So ice cream sales might be useful for predicting how many livestock die on a given day, but will banning ice cream lead to improved livestock longevity? No. Both the ice cream sales and the livestock deaths are caused by another factor – very hot days.

The many news articles covering this story (links below) seem not to understand this point. They state things like “lowering LDL prevents heart attacks” or “statins prevent heart attacks by lowering LDL” but we don’t know that. We have no idea how statins prevent heart attacks. We only know that they do prevent heart attacks and that they do lower LDL. There is no way to know if statins prevent heart attacks through lowering LDL or through some other mechanism. In fact, since estrogen lowers LDL and doesn’t prevent heart attacks, that suggests that LDL (like HDL) may be a marker of risk, not the cause.

So we can focus on risk factors to identify patients at risk, but for an effective treatment we must insist on a therapy that affects clinical outcomes (like strokes or heart attacks) not just risk factors. Niacin just failed that important test.

Learn more:

New York Times article: Study Questions Treatment Used in Heart Disease

NPR health blog: Study: Boosting Good Cholesterol With Niacin Did Not Cut Heart Risks

Wall Street Journal article: UPDATE: NIH Stops Clinical Study Involving Abbott’s Niaspan

NIH press release: NIH stops clinical trial on combination cholesterol treatment

For an explanation of the important difference between modifying risk factors and changing clinical outcomes, see my post Merck Knows More about Zetia than They’re Telling Us

Correction on December 2, 2011: In the 1960s the Coronary Drug Project (CDP) showed that niacin was superior to placebo in the prevention of stroke and heart attacks. The statement in this post that niacin has never been proven to prevent strokes and heart attacks is incorrect and has been crossed out. I regret the error. More details about the CDP and niacin are in this post. The title of this post might have been somewhat more accurate if appended with “in patients taking statins”.