This study results do not disprove the HDL hypothesis. Rather they prove the lack of efficacy of fenofibrate to raise HDL.

There was not a significant difference (p<0.05) betwen the grops in on-treatment HDL in most of the years with the only exception being year1 and exit visit. And even then the differences were  40.4 vs 39.4 and w 41.2 vs 40.5 - hardly worth discussing. So all this money was spent to prove that fenofibrate did not raise HDL more than placebo (at least once you are on a statin. WOW!

It seems that with what we presently know we should treat with max statin to target non-HDL-C and then add niacin (11 year follow-up of CDP, Arbiter,Fats) before adding fenofibrate for TG>200. Hopefully AIM-HIGH and HPS-Thrive will give better results.

     I reserve use of fenofibrate for patients with low HDL and High TG levels since that is the only scenario in which fibrates raise HDL.  Fibrates ( and here I mean fenofibrate, since statins and gemfibrozil together leade to rhabdomyolysis ) are of no help in raising HDL in the low HDL and normal TG scenario.  In the subgoup described by Dr Bilazarian in which HDL was low and TG were high, benefit was seen.  Since most women would not be characterized by low HDL and high TG, fenofibrate would not be indicated.

     It has long been known that maximal regression of atheromatous plaque occurs when LDL is lowered simultaneous with HDL being raised, once LDL has been lowered, then addition of fenofibrate to raise HDL and lower TG is in my opinion useful and I have not seen any harm done in my long-running lipid clinic (since 1974).  My experience has been summarized in  Amer J Cardiol, 2008; 7: 372-373 (August).

Thank you for the comments

I agree that the HDL hypothesis is not disproved and for our patients I hope the HDL Hypothesis will become dogma after the publications of AIM-High and HPS THRIVE and the CETP inhibitor trials but for now the fenofibrate is difficult to justify.  I am surprised that the presentation of ACCORD  came with a recommendation by the PI to use fibrates in the low HDL and High Triglyceride subset even though this non prespecified subset was not statistically significant p = 0.06.  The lay press on cardiologists ignoring evidence based medicine (A Simple Health-Care Fix Fizzles Out in WSJ February 11, 2010  - http://online.wsj.com/article/SB10001424052748703652104574652401818092212.html

Might be a problem here again with the 1.5 billion in fibrate sales.

I am further concerned about the hazard in women and the renal effects which seems to be ignored thus far.  We'll see what the FDA has to say to provide guidance for physicians and patients.  Althogh Dr. Freeman's lipid experinece dwarfs mine these RCT findings require reflection and reconsideration.

The Lipid arm of the ACCORD trial had  only 17% of the Atherogenic Dyslipidemia I see in my Dyslipidemic patients and in my Diabetics with dyslipidemia.

Most of Dylipidemic patients are way above  the 17% and have Triglyceride over 300 and HDL below 25. So ACCORD did not represent most of my dyslipide my pateints.

Hypertriglycedemia definetely shifts the LDL particles to the more small atherogenic size.

I have had great success with FENOFIBRATES and most of my patients cannot tolerate NIACIN

Moreover if you analyze the subset of the 17% Dylipidemic patient there was a 31 % reduction in the primary or secondary endpoint if I am not mistaken.

So to me this trial was not representative of the vast amount of patient with Dyslipidemia I see in my practice

So in my practice I use a lot of fenofibrates with Statins and will add Lovaza , welchol and Zetia to get patients top goal

Adding Fibrates to TG levels and HDL levels in the Lipid arm of the ACCORD trial did not make sense to me and I did not expect to see any benefits

As far as the FIELD trial was concerned there was a 61% addition of STATINS  in the Fenofibrate arm which masked the benefits of Fenofibrate

I would note that the Helsinki Heart Study showed great cardiovascular benefit if the CT:HDL ratio exceeded 5.0 and if TG>300.  VA-HIT showed benefit in patients with low HDL and high TG.  True, most of these patients were men, but as I have already observed the low HDL-high TG scenario is mostly seen in men.  I do not rercall any harm done to the few women with this lipid profile.

I would invite Dr Bilazarian to read my point-counterpoint article in the upcoming issue of Journal of Clinical Lipidology or the Readers; Comments section ot the American Journal of Cardiology, 2008; 7: 372-373.  This article gives my approach.  Although it does not discuss fibrate therapy, it shows how I use the decades old finding that the best regression of athdromatous occurs when LDL is lowered simultaneously with raising of HDL. 

Topic of lipid management is both rewarding & frustrating.Literature supports your viewpoint, whatever it may be!

I think that a practicing cardiologist & physician requires a specific flow chart for various groups of patients for management e.g.isolated low HDL-c group, or isolated high triglycerides group.

In each group use of combination therapyshould also be defined.

At present the plethora of trials indicate the confusion rather than clarity of thinking in the minds of invetigators.

Are we being remote-managed by the pharma industry?

Two quick epidemiological points.

 First, subgroup analyses are historically underpowered (trial is powered to the primary endpoint in the entire cohort), and therefore the p value of 0.06 for the TG/HDL subgroup that Seth mentions is well within the p<.10 for a positive subgroup interaction test recommended by most trialists/epidemiologists.

Second, what does a p value of .06 versus a p value of .05 or .04 tell us anyway?  If you bear in mind all the positive data on this subgroup from HHS, BIP, LEADER, VA-HIT, FIELD, etc, the post-test probability of a positive finding actually INCREASES given a p value of .06 for the interaction chi square test.  Here one is taking a Bayesian perspective and not viewing the data in strict, frequentist-like isolation (which most statisticians believe is a major mistake made by clinicians).

To All,

     I wonder about the concept of residual risk.  If I recall correctly, this concept derives from the various statin trials that used a fixed dose of statin and noted a 30-40% reduction in cardiac events. Had the trials allowed dose titration, would there still have been residual risk?  The approach that I use and advocate, as referenced in my preceeding comments, has little, if any, residual risk, and cardiovascular events in my treated patients are few and far between--and makor events are rare.

I strongly disagree with the position that pts cannot tolerate niacin - I have at least 100 pts taking IR nicain 2-3x/day - only ~1 out of 20-30 patients just cannot tolerate it. Some of these pts will do fine on Niaspan. 

The key is to strongly stress the tremendous benifit of statins plus niacin. I have not had a single c-v event in over 5 years in pts on this regimen. Also when I can get their chol/HDL ratio under 3.0, we usually see regression on serial CIMT.

I tell patients to take half of an aspirin, have a meal and then take the niacin after the meal. Also I stress that the flushing is a benign nuisance which becomes much less with time - some pts are up to 1000mg tid.

 IMO, niacin is the most underutilized medicine we have in the prevention of c-v events.