All scientists at the National Institutes of Health will be banned from any new outside collaborations with pharmaceutical or biotechnology companies for at least one year -- and all existing collaborations will have to be discontinued -- under a surprise shift in policy released late yesterday by agency officials.

The blanket moratorium represents a much more radical policy change than NIH officials recently said they would invoke, and one that could shake already flagging morale at the beleaguered agency. But its need became apparent after the agency's own conflict-of-interest investigation turned up more problems than had been anticipated, said Raynard S. Kington, NIH's deputy director and ethics chief.

"We've learned there are vulnerabilities in our system of oversight," Kington said in an interview -- vulnerabilities, he said, "that give us pause."

NIH has been under investigation by the House oversight subcommittee since December, when the Los Angeles Times reported that some agency scientists were engaged in lucrative collaborations with drug and biotech companies that posed at least the appearance of conflicts of interest. Investigations have also been launched by federal ethics offices.

The controversy at first appeared limited to a few cases in which scientists had engaged in activities that were legal and approved by appropriate overseers but which suffered from an appearance of conflict of interest -- such as the case of an official who, with approval from the highest legal official in the Department of Health and Human Services, received a cash award from a university that gets grants from the NIH.

But over the months, the number and severity of apparent ethics violations have steadily escalated.

In one example, congressional investigators reported that a researcher at the National Institute of Mental Health had been paid $517,000 in fees, honorariums and expense reimbursements by drug giant Pfizer Inc. over a five-year period without reporting that income to NIH officials, as is required.

In another, an NIH doctor was allegedly retained as an expert witness in several private product-liability lawsuits at a rate of $600 per hour, congressional investigators said -- again without required agency permission.

NIH Director Elias A. Zerhouni convened a blue-ribbon committee to look into the allegations early this year, and on the basis of that group's recommendations began to implement new tiers of ethics review for all scientists wishing to engage in outside consulting arrangements.

Zerhouni resisted calls by several members of Congress to ban those arrangements altogether, saying that to do so would make it too difficult to attract and retain high-quality researchers. But he banned all such collaborations by institute directors and others with control over grant decisions.

In June, as the clamor for change remained strong on Capitol Hill, Zerhouni proposed additional restrictions, including a ban on ownership of drug company or biotech stocks by some key employees; no membership on corporate boards; creation of a centralized registry of all outside arrangements; and prohibition of all paid consulting or speaking engagements at institutions that receive NIH funding.

"I have reached the conclusion that drastic changes are needed," he said at the time.

Yesterday's shift reflected in part ongoing disputes between the NIH and the Office of Government Ethics over how much the agency needs to do.

The new plan was submitted to OGE yesterday and is expected to go into effect as soon as that office approves the change, Kington said.

Kington would not reveal details about the number or depth of the ethics violations the agency's internal investigation has uncovered, saying the investigation is ongoing.

Perhaps 100 NIH scientists are involved in collaborations that will have to be discontinued, he said, noting that many scientists dropped their deals lately in light of all the attention. NIH legal staff has concluded that the agency has the authority to ban even those arrangements bound by contracts with companies, Kington said.

"We believe it's in the best interest of the agency," he said.

The NIH's Raynard S. Kington said the agency had oversight issues with scientists' work.