UPDATE 1-US experts urge limits for some drug safety trials

* IOM warns on randomized trials for drugs on market

* IOM says so-called real world trials can be sound

* FDA sought IOM’s advice ahead of Avandia meeting

* FDA staff divided on whether Glaxo trial should continue

* Advocates say Glaxo’s TIDE trial unethical (Adds details from FDA documents, Glaxo trial details, byline)

By Susan Heavey

WASHINGTON, July 9 (Reuters) - Clinical trials aimed at pinpointing emerging safety problems with drugs already on the market should only be conducted when other, less invasive research is not possible, a U.S. advisory group said on Friday.

The Institute of Medicine's report, requested by U.S. Food and Drug Administration officials, comes ahead of a highly-anticipated meeting next week on GlaxoSmithKline Plc's GSK.LGSK.N diabetes drug Avandia, that has come under fire for potential heart risks.

Glaxo is conducting a randomized, controlled trial called TIDE to compare Avandia to Takeda Pharmaceutical Co's 4502.T rival drug, Actos, after safety concerns emerged with Avandia that the FDA and its advisers concluded were unclear.

While the IOM did not target the Avandia trial in particular, FDA Commissioner Margaret Hamburg asked the independent, nonprofit medical group to weigh in with at least an initial finding on general ethical issues involving the safety of studying the risks of drugs already in use.

The FDA called in 2007 for the TIDE study, which aims to enroll 16,000 patients and end in 2015, although some critics say it could take longer.

Analyses of other data since then have convinced some researchers that Avandia’s risks are greater than those of Actos and have repeatedly called for the trial to be halted.

Glaxo has defended its drug, saying data overall shows it does not increase the risk of heart attack, stroke or death.

On Tuesday and Wednesday, FDA’s panel of outside experts will weigh numerous studies and analyses before recommending whether Avandia should remain on the market, be pulled from the market, or various options in between.

They also will be asked to weigh in on the TIDE trial.

In its report, the IOM said randomized controlled trials should only be done when “a responsible policy decision cannot be made based either on the existing evidence or on evidence from new observational studies.”

It also said studies of real world use of a drug -- known as observational studies and often done by reviewing insurance claims and other databases -- can yield strong data.

In separate FDA documents released on Friday, FDA staff scientists were split over whether the TIDE trial should continue just as they are divided over whether another major Glaxo trial, RECORD, showed excessive heart-attack risks with Avandia.

Some staff also pointed to studies done since 2007, including an FDA analysis of 52 trials on the drug, as well as a review of Medicare data that showed a greater risk of heart attack and other complications with Glaxo’s drug.

“Based on these findings, any proposed head-to-head trial of rosiglitazone vs. pioglitazone is unethical and exploitative,” agency reviewers David Graham and Kate Gelperin wrote.

Rosiglitazone is the generic name for Avandia, and pioglitazone is the generic name for Actos.

Another FDA scientist, clinical reviewer Karen Mahoney, said that the TIDE trial does not have many of the limitations of the earlier RECORD study and could be illuminating.

“This trial, if it continues to completion, has the potential to address the question of the cardiovascular safety of rosiglitazone more definitively,” she wrote in a separate memo. (Reporting by Susan Heavey; Editing by Tim Dobbyn)