The steel rod of the artificial hip had been slathered in antibiotic cement, an effort to ward off a nasty infection that had gone to the bone.

But when it had to be removed, the metal-and-plastic contraption was not disposed of as medical waste. Instead, it now sits on a shelf at Drexel University, alongside more than 7,000 artificial knees, hips, and other tokens of infirmity from around the country.

For Steven M. Kurtz, director of Drexel's Implant Research Center, the unusual collection is an invaluable resource in his quest to make the technology better.

"We study the tough things that happen to good implants," the biomedical engineer said.

Well over 90 percent of joint replacements last at least a decade, and with continued improvements in the technology, the newest models are expected to last even longer.

But nagging problems persist.

Among Kurtz's findings: a type of plastic used in hip and knee implants in the late 1990s degrades over time. Since then, manufacturers have moved to prevent the problem by using plastics impregnated with antioxidants.

The most common reasons for joint replacement failure are infection, a loosening of the device, or in the case of hip replacements, a dislocation. Much less common are cases in which the material in the device fails.

Kurtz wants to see all of it.

Blood, bone, and steel

Day after day, the discarded devices arrive in cardboard boxes from 15 hospitals in eight states, sometimes with pieces of bone attached, or with traces of the patient's blood still visible.

"We get everything they take out," said Kurtz, who also works at Exponent, an engineering consulting firm.

Research coordinator Eliza Fredette and Genymphas Higgs, a doctoral student in biomedical engineering, don protective gear to handle the donations. Each one is cleaned, dried, and carefully cataloged.

Other hospitals maintain collections of old implants. But Drexel's is unusual in its size and the fact that it is fed by many institutions, said Noreen J. Hickok, associate professor of orthopedic surgery at Thomas Jefferson University.

"Somebody like Steve is able to, early on, start saying 'Oh, my god, there's a problem here,' " said Hickok, who collaborates with Kurtz on research.

Conversely, he can see which approaches are working better - say, if a certain type of implant is less likely to harbor an infection, she said.

In the Philadelphia area, Penn Presbyterian Medical Center and the Rothman Institute have contributed implants to the Drexel collection.

Another source is the University of Tennessee Health Science Center, which has sent Kurtz more than 50 implants from various clinics across that state.

William M. Mihalko, a professor of orthopedics at the University of Tennessee, praised the Drexel effort, though he said a comprehensive national database would be even better.

"We could possibly avoid implant design variables that are problematic or raise the red flag about an issue sooner rather than later," Mihalko said.

Among the topics Kurtz has explored with the donated devices is the use of different materials. One study found that hip implants with ceramic heads develop less corrosion than those made with a cobalt chrome alloy.

In another study, he and Higgs looked at the impact of "modularity," meaning the number of component parts. Manufacturers fashion their devices with a variety of "tapers" and other interchangeable fittings so a surgeon can readily customize an implant to the patient's anatomy.

But there can be trade-offs. In hip implants with metal heads, increased modularity in the stem portion of the device was associated with more corrosion and "fretting," tiny scratches caused by back-and-forth motion, the study found.

'A ton of antibiotics'

Infection is an especially tricky problem, said Jefferson's Hickok. Despite all the precautions taken by the surgical team to maintain a sterile site, the joint can later become colonized with bacteria that are already in the patient, such as from tooth decay, she said.

She and Kurtz are part of a team working on a possible post-surgical solution for a type of spinal implant: "smart" devices with reservoirs of antibiotics that could be released after surgery, using ultrasound.

If it works, that would be a big upgrade over current methods, Kurtz said.

"The state of the art is to use a ton of antibiotics and use really careful operating-room techniques," he said.

But the most important factor in success is not the implant itself, but who is putting it in, said Kurtz, who came to Drexel in 2000. Find someone with lots of experience, he advised.

"You always choose your surgeon first," the engineer said. "Whatever implant they happen to be comfortable using, then use that implant."

One thing is clear: Kurtz will have an increasing amount of material to analyze. From 2000 to 2010, the number of hip replacements in patients over 45 soared from 138,700 to 310,800, according to a 2015 analysis by the National Center for Health Statistics.

Kurtz attributes the increase to longer life spans, the aging of the baby boom generation, and the obesity epidemic, as heavier bodies cause more wear and tear on joints.

Ideally, the newest generation of artificial joints will last for decades. But for those that must be removed, the overseer of Drexel's odd assortment of steel and plastic stands ready to have a look.