The research chief of Roche Holding AG said the company's experimental drug to raise good HDL cholesterol has the potential to generate annual sales of $10 billion and may have advantages over a rival medicine being developed by Merck & Co Inc .

This is the order of magnitude we are looking at, Jean-Jacques Garaud said in an interview on Friday, when asked if his dalcetrapib drug could produce annual sales in the $10 billion range.

That would make it a mega-blockbuster heart drug in the same league as today's biggest cholesterol fighter, Pfizer Inc's

Lipitor. Lipitor and other current top-selling cholesterol drugs work instead by lowering bad LDL cholesterol.

Garaud, head of Pharma Research and Early Development for the Swiss drugmaker, said Roche and partner Japan Tobacco Inc <2914.T> hope to seek regulatory approvals by 2013 for dalcetrapib.

He said dalcetrapib, if approved, would probably reach the market two years before Merck's own HDL-boosting drug, called anacetrapib. They both belong to a new family of drugs called CETP inhibitors.

Researchers last month said dalcetrapib raised HDL levels by an average of 31 percent in a mid-stage study, without affecting levels of LDL choleterol. It is now being tested among tens of thousands of patients in large Phase III studies to see if it can prevent heart attacks and stroke.

Merck's anacetrapib, by contrast, boosted HDL levels by a whopping 138 percent in a mid-stage trial described in November, while also cutting LDL levels by almost 40 percent.

Neither drug caused high blood pressure, a problem that helped doom Pfizer's own CETP inhibitor, called torcetrapib. Pfizer scrapped its product in late 2006 after it was linked to deaths in a costly late-stage trial.

Torcetrapib, which Pfizer had expected to capture peak annual sales of $10 billion or more, boosted HDL levels by about 60 percent.

Although the HDL-raising ability of Merck's drug appears far more impressive at first glance than Roche's product, some researchers have speculated it might be preferable to avoid overly dramatic increases in HDL cholesterol. They say maybe less is more, given the failure of torcetrapib.

Garaud said the main consideration may be cholesterol efflux, meaning how well LDL is removed from the bloodstream as a result of boosting HDL.

Even though their (anacetrapib's) HDL elevation is higher, that might not be the best marker for potential activity, he said. What is important is how much cholesterol we pull from the blood. We think ours is better than any other drug.

(Reporting by Ransdell Pierson; editing by Lisa Von Ahn and Andre Grenon)