New ways of controlling cholesterol, including possibly directly injecting "good" HDL cholesterol into patients, need to be studied following the failure of promising treatments from Eli Lilly, Pfizer Inc and Roche Holding AG, according to top heart researchers.

Lilly in October halted a 12,000-patient study of its experimental drug evacetrapib, an oral medication that in smaller earlier studies slashed "bad" LDL cholesterol and doubled levels of HDL.

But improved cholesterol levels did not prevent heart attacks and strokes, diminishing hopes for the approach to treating heart disease - by raising HDL through blockage of a protein called CETP.

Roche in 2012 scrapped its own CETP inhibitor after it also failed to help patients. Pfizer's similar drug was discontinued in 2006 after being linked to deaths in trials.

Although Merck & Co continues to develop its own CETP inhibitor in a 30,000-patient study expected to be completed next year, researchers on Sunday said the failures of the Lilly, Roche and Pfizer drugs bode poorly for it.

"Merck's drug is the fourth shot on goal for CETP inhibitors, but with disappointment or lack of success for the other agents you have to be increasingly pessimistic" about the class of drugs, said Dr. Stephen Nicholls, deputy director of the South Australian Health and Medical Research Institute in Adelaide, Australia. He was a lead investigator for the failed trial of Lilly's drug.

Nicholls and Dr. Steve Nissen, the head of cardiology for the Cleveland Clinic, who co-lead the evacetrapib study, on Sunday reviewed the baffling evacetrapib data in a presentation at the annual scientific sessions of the American College of Cardiology in Chicago.

"This drug lowered LDL by 37 percent and raised HDL by 130 percent and had absolutely no effect" on preventing deaths and heart attacks, Nissen said in an interview.

Although other ways of raising HDL cholesterol might eventually prove protective, Nissen said all attempts so far have been fruitless.

Nicholls said he remains hopeful of future HDL therapies and is testing whether artificial HDL can be made in the laboratory and injected directly into high-risk heart patients. "There is enthusiasm it may be able to shrink plaque" in heart arteries, he said.

He said he is studying variations of that approach with French drugmaker Cerenis Therapeutics and the Medicines Company.

Nicholls said another possible approach would be to instruct the liver to make more HDL.