Transplant Puzzle Probed

Scientists may be close to understanding why organ transplant recipients are especially likely to get a form of cancer called Kaposi's sarcoma. The disease affects one in 200 transplant patients in the United States -- 400 to 500 times the usual incidence. It usually starts as a slow-growing skin tumor, but in 40 percent of cases it spreads to other organs and becomes deadly.

Scientists have known for a decade that Kaposi's is caused by a herpes virus -- a virus that can remain dormant and then, for poorly understood reasons, can stir from its sleep and cause disease.

One explanation for the Kaposi's in organ recipients was that the virus was hitching rides on organs from infected donors. An alternative possibility was that the recipients had the virus all along but had kept it suppressed until they started taking immune system-suppressing drugs to protect the new organs.

To settle the issue, researchers in Italy, Israel and Germany conducted genetic tests on cancer cells taken from six women who got Kaposi's after receiving organ transplants from men. In four of them, the cancer cells had a Y chromosome, found only in male cells. Tests on two additional transplant patients with Kaposi's found that one of the cancers had the genetic fingerprint of the donor, the team reports in this week's advance online edition of the journal Nature Medicine.

The findings suggest that many cases of Kaposi's in organ recipients are seeded by healthy, infected donors, then blossom into full-blown disease because of immune suppression. It may be worthwhile to test potential donors and screen out those who carry the virus, the researchers say. They also suggest that recipients might benefit from transfusions of specialized white blood cells from donors to keep any transplanted viruses under control.

-- Rick Weiss

Big Earthquakes Dated in L.A.

There's good news and bad news for Southern California. A hidden fault running beneath downtown Los Angeles has caused immense earthquakes. But they only happen about every 3,000 years -- and they don't have to happen at all.

"It's Mother Nature's choice," said University of Southern California earthquake geologist James F. Dolan. "The fault stores up all this energy, and can release it in lots of moderate-sized earthquakes or a few large ones."

Earth scientists have long known about the Puente Hills blind thrust fault that runs east-west from northern Orange County to Beverly Hills and that is overlain by two miles of folded river sediment. Reporting in the journal Science this week, Dolan and two associates for the first time showed how the dates of previous earthquakes could be determined by the depth of the sediment layers.

"The river wants to have a smooth gradient, and if you perturb that with an earthquake, the river fills up the downhill side, and the sediments are thicker," Dolan said. The researchers used radiocarbon dating to determine that "big" earthquakes had occurred about 3,000 years ago, between 3,000 and 6,000 years ago, 7,500 years ago and about 11,000 years ago.

By big, Dolan meant that the quakes expended about 15 times the energy of Los Angeles's 1994 Northridge quake, which caused $60 billion in damage -- one of the most costly natural disasters in U.S. history. Northridge was not a Puente Hills quake, but Whittier Narrows in 1987 -- relatively inexpensive at $600 million -- was.

Bigger, he pointed out, means the quake will last longer and be felt over a wider area, but the amount of damage is unpredictable. Little earthquakes expend high-frequency energy: more hazardous to small structures, like houses. Big earthquakes have low-frequency energy: more dangerous for skyscrapers, bridges and dams.

-- Guy Gugliotta

Eyewitness Testimony Doubts

Eyewitness testimony plays an important part in many criminal trials, but it's a notoriously tricky and unreliable form of evidence. Among the things that can affect a person's confidence in what and who they saw during a crime are what they're told by police about those recollections.

A team of psychologists led by Gary L. Wells at Iowa State University demonstrated that convincingly in an experiment described in a recent Journal of Experimental Psychology.

The researchers prepared a 60-second videotape purportedly showing a man on a roof dropping what appears to be a bomb down an air shaft. They showed the tape to 253 volunteers, who were then asked to pick out the bomber from six photographs. Unknown to the volunteer witnesses, a picture of the actor playing the bomber was not in the array. Nevertheless, every volunteer picked a suspect.

After making their choice, some were told they made the right choice, some told they made the wrong one, and some were told nothing. They were then asked how well they remembered what they saw in the video.

The people who were told they picked the right suspect had much greater confidence in virtually all aspects of their recollection, and 23 percent said they were at least 90 percent sure of details. Those given no feedback were much less confident, with only 2 percent saying they were at least 90 percent sure.

The researchers said the findings affirm the recommendation that police line-ups be "double-blind," with neither the witness nor the investigator accompanying the witness told whether the right suspect was chosen.

-- David Brown