our months after a young Minnesota man died from routine knee surgery, federal officials said yesterday that they had turned up 25 more cases of serious bacterial infections in people who received similar operations over the past several years.

The officials, from the Centers for Disease Control and Prevention in Atlanta, said a single tissue-processing company had been involved in 14 of the 26 cases, including the fatal one. The agency also said the tissue that had fatally infected a Minnesotan who was having kneecap repair surgery, Brian Lykins, 23, had not been refrigerated until 19 hours after the donor's death, by which time lethal bacteria from his intestines might have spilled into his body.

The disease control centers' report, which offered recommendations for improving tissue processing procedures, did not name the company, but it has been widely identified as CryoLife Inc. of Kennesaw, Ga. Roy Vogeltanz, CryoLife's vice president for corporate communications, acknowledged that half the patients with serious infections had received implants from the company. But he said, "The C.D.C. doesn't know and we don't know that the infections were related to the implants."

Surgery is always risky, Mr. Vogeltanz said, and infections can come from many sources, including operating rooms. But he said CryoLife would consider the disease control agency's recommendations "and see if they improve tissue processing quality, given our own knowledge and experience."

CryoLife is the largest of half a dozen companies that collect and prepare tissue for transplant. The tissue includes heart valves, tendons and bones but not whole organs.

The federal report sheds light on a fast-growing industry that is coming under increasing scrutiny from regulators. In 1999, 650,000 grafts were distributed by tissue processors, said Dr. Daniel Jernigan, an epidemiologist in charge of the investigation. Few complications have been reported, Dr. Jernigan said, but no one knows if contamination is rare or simply underreported. The problem stands out in knee surgery because most patients are young and healthy. But older people who receive heart valves, veins or other soft tissue could be getting sick from implants, he said, with their doctors failing to make the connection because the patients tend to be frail and ill.

Sterilization procedures vary by company. Companies may follow practices laid out by the American Association of Tissue Banks, a voluntary organization, or create their own. They are inspected by the Food and Drug Administration, but are required to show only that they are following written procedures, said Lenore Gelb, an agency spokeswoman.

The agency is writing a new rule that will govern good practices, but it will not be very specific, Ms. Gelb said, adding, "We won't outline every step of the way."

Dr. Marion Kainer, an epidemiology officer at the disease control agency, said tissue to be donated could be sterilized in several ways. The safest is to bombard it with gamma radiation, which kills fungi, viruses and bacteria, Dr. Kainer said; the problem is that radiation weakens tissues like cartilage and tendons used to repair damaged joints.

To make sure that unirradiated tissue is free of germs, companies place both the replacement knee and a small sliver of cartilage called companion tissue in a solution designed to kill all pathogens. The recipes for these solutions vary by company and are jealously guarded, Dr. Kainer said.

After some time in the solution, the implantable tissue is taken out and packaged; the companion tissue is placed into various culture baths to see if any bacteria can be grown. The problem is that the sliver of cartilage is still covered with antibiotics while it is being tested, Dr. Kainer said. While most common bacteria will already be dead, some less common but deadly kinds of bacteria can escape detection.

One such group, called clostridia, live in suspended animation as spores. Antibiotics have no effect on them. Clostridial spores were found in half the patients with serious infections, Dr. Jernigan said.

Once such spores get into a warm, wet environment with no air — like a swollen knee joint after surgery — the bacteria spring to life, producing a deadly toxin that at the very least chews up healthy cartilage in the affected knee.

"We suggested several ways the company can minimize such contamination," Dr. Kainer said.

One is to take tissue cultures before the material is placed in solution. Also, she said, because bacteria like clostridia are difficult to find, technicians should look for common bacteria found in the human bowel that can serve as surrogates for the more dangerous species.

When a person dies, the intestinal wall eventually breaks down and gut bacteria spill into the bloodstream and may lodge in distant tissues. This decomposition can be delayed by refrigeration, Dr. Kainer said.

In Mr. Lykins's case, the donor had been dead 19 hours before refrigeration, Dr. Jernigan said, and should not have been used. But as the market grows for donor tissue, he added, "more tissues are being collected later and later."