WASHINGTON - Scientists and drug companies have failed to notify the National Institutes of Health about six deaths that occurred in gene therapy experiments in the past 19 months, keeping details of the deaths from becoming public, according to interviews with researchers and federal officials.

The deaths are the first in gene therapy to come to light that were purposely withheld from the NIH, one of two federal agencies charged with overseeing the safety of the controversial field of medical research, which seeks to cure diseases by giving patients new genes.

The lack of disclosure provides new evidence of a shift toward secrecy in gene therapy, traditionally one of the more open fields of medical research, and reflects escalating efforts by gene therapy companies to weaken federal reporting regulations.

Confirmation of the six deaths follows revelations last week of a death and two serious illnesses in gene therapy patients that were reported to the NIH with the unprecedented insistence that they be kept confidential, in defiance of a longstanding agency policy of public disclosure.

The six deaths occurred in heart studies headed by two leading gene researchers - Ronald Crystal of the New York Hospital-Cornell Medical Center in Manhattan and Jeffrey Isner of Tufts University in Boston. The two are racing to be the first to grow new blood vessels around blocked ones as an alternative to bypass surgery.

Crystal and Isner said that, unlike the widely reported death of a teenage patient at the University of Pennsylvania in September, they believe that the fatalities in their studies were not directly caused by the gene therapy but by complications stemming from the patients' underlying illnesses.

Because they decided the deaths weren't caused by gene therapy, they argued, federal regulations don't require them to notify the NIH - a new interpretation of those regulations that stands in sharp contrast to the one held by NIH officials and a decade of practice.

The researchers said they reported the deaths to the Food and Drug Administration, which keeps such information secret.

But NIH officials in the federal office that oversees gene therapy are adamant that even deaths not initially believed to have been caused by the therapy must be reported to the NIH and made public, because often it is not clear until later whether the therapy actually caused the deaths.

''It may take five, six, seven patients ill, or 20 patients, before you find out, `Hey, this is also happening in other people's trials,''' said Amy Patterson, who heads the NIH Office of Recombinant DNA Activities, which oversees gene therapy studies. ''And if you don't know what's going on in other people's trials, then you can't put two and two together.''

The FDA can suspend a study if it determines that the therapy being tested is dangerous, but the agency makes such information public only if the therapy eventually is approved or with the permission of the study's sponsor.

Most of the new deaths are coming to light because federal officials put out a plea for gene researchers nationwide to report any undisclosed deaths or illnesses, after the death of the teenager at the University of Pennsylvania. His death is thought to be the first directly caused by gene therapy, and NIH officials are looking for indications of similar problems in other studies that may not have been attributed to the treatments themselves.