In children and young adults with osteosarcoma, limb-sparing surgery is a better choice than amputation—if the postoperative course runs smoothly. But although survival rates have increased dramatically over the past two decades—from about 20% to 75%—and although physiologic function is significantly better with limb-sparing surgery than it is with amputation, the postoperative course is fraught with danger, at least as reported in a study done at St. Jude Children's Research Hospital.
The study in the August 1 issue of Cancer by Aditya Gauer, MD, and colleagues surveyed the charts of 103 patients who underwent limb-sparing surgery (LSS) at St. Jude from April 1982 to January 2001. Patients had an average of four infections each: focal bacterial in 67% and bacteremia in 21%.
In addition, 21% of patients developed orthopedic-device infections, with 16% of the latter requiring limb amputation even after long courses of antibiotics.
The study was designed to understand the epidemiology of and risk factors for the high incidence of infection following limb-sparing surgery, noted the senior author, Elisabeth Adderson, MD, Associate Member of the Department of Infectious Diseases.
“No previous studies have been done, so it's hard to compare, but it's my impression that the national infection rate is anywhere from zero to 60 percent. A lot of these estimates are more than a decade old, and with new aggressive protocols for treating infection and new surgical techniques, the rate is decreasing.”
Martin Malawer, MD, Professor of Orthopedic Surgery at Georgetown University and a consultant to the National Cancer Institute, was asked in an interview about his experience at Georgetown and NCI over the past 25 years: “After more than 200 LSS procedures for bony sarcoma using orthopedic devices for the knee, hip, and shoulder, we have had a 6% infection rate overall, and we were able to save 95% of all extremities for 10 to 15 years. There were about 6% late amputations due to infection.”
Also asked for a comment, Jeffrey Eckardt, MD, Professor of Orthopedic Surgery and the Helga and Walter Oppenheimer Chair of Musculoskeletal Oncology at UCLA, said his experience has shown an infection rate even lower: “I've done more than 620 of these procedures since 1980 and have had only one acute infection. The overall rate at our institution is less than 2%. It's the same for ODI [orthopedic-device infections]. We just don't see infection, and I can't understand why they're having so much trouble at St. Jude. A big rate like that tells me they're doing something wrong.”
Patients with infection at St. Jude were twice as likely to be black than white and to have tumors of the tibia rather than the femur. Dr. Adderson said she doesn't know why blacks are at greater risk.
“Maybe they have more advanced disease when they're diagnosed and are more immunocompromised by the time of surgery, or maybe they're genetically more susceptible to certain types of bacteria. No research has been done about this.”
Dr. Malawer said he knows of no racial difference. However, he works in Washington, DC, where the population is close to 90% black, so there is not much basis for comparison. Dr. Eckardt echoed him: “I've never heard of such a thing. We have no racial distinctions.”
Two Types of Infections
The report from St. Jude defined an episode of wound infection as a group of signs and symptoms that prompted antibiotic therapy even though the specific causative organism was not always identified.
There were two types: local or soft tissue (no evidence of bone involvement, which was successfully treated with topical or short-term systemic antibiotics), and orthopedic-device infections, diagnosed if bacteria or fungi were isolated from a normally sterile site such as bone or joint fluid.
Early-onset infections occur within 90 days of surgery; late onset, after 90 days. The most common infectious agents were Streptococcus pneumoniae, Staphylococcus aureus, and Enterococcus.
At St. Jude, the patients (78% of whom had osteosarcoma; 14% of whom had Ewing sarcoma) included in the study were followed for a median of 5.4 years after limb-sparing surgery, and they were an average of 13.6 years old.
Almost half (44%) were female, and almost all had a central venous catheter inserted when cancer was diagnosed. Ninety-five percent received adjuvant or neoadjuvant chemotherapy, and 12% had local radiotherapy.
Most patients received prophylactic antibiotics such as trimethoprim-sulphamethoxazole, and since 1992, almost all were treated with recombinant granulocyte colony-stimulating factor to prevent neutropenia.
Early-onset orthopedic-device infections developed in nine patients an average of 19 days postop. One patient was treated with antibiotic therapy alone, but the rest required surgical drainage and a prolonged course of antibiotics. Six had to have an amputation.
According to Dr. Malawer, “Almost all these early infections are secondary to poor wound healing and can be avoided by good soft tissue coverage of the orthopedic device. This is especially important for the gastrocnemius muscle flap for tumors around the knee. If your surgical technique is good, you can reduce the early infection rate to less than 1%.”
Dr. Eckardt agreed: “If you do the right kind of soft tissue closure, you won't get an infection. It's probably the single most important factor.”
Dr. Malawer also said that intermediate infections, which begin at about three months and continue until chemotherapy is complete, usually derive from indwelling central lines where bacteria tend to colonize.
“At the first sign of line sepsis, remove it, give antibiotics for three to six weeks, and don't put the line back until the infection is gone,” he said. “That's the way to prevent ODI, but I have noticed a marked hesitancy on the part of oncologists to remove the lines—often with bad results.”
Late-onset orthopedic-device infections in 13 of the St. Jude patients began a median of 13 months postop with wound drainage, dehiscence, swelling of the extremity, and erythema.
Six patients who initially responded to antibiotics had a recurrence 140 days later. One patient was cured with removal of the orthopedic device and systemic antibiotic therapy, and three needed amputation.
Risk Factors for Orthopedic-Device Infections
The authors identified risk factors for orthopedic-device infections: being black rather than white; having a tumor of the tibia rather than the femur; and developing a wound infection from the limb-sparing surgery.
Dr. Malawer also identified risk factors: poor skin and muscle coverage of the orthopedic device, which leads to skin flap necrosis; use of induction chemotherapy, which compromises the immune system; and lack of surgical experience.
He added that using allografts results in a very high infection rate and a graft failure rate of 10% to 20%.
“Most surgeons have given up allografts because they are inherently more risky and prone to infection and other complications. Overall survival of allografts and the limb in which they are used is less than 35% at five years. This compares with 90% for orthopedic devices.”
What Does This Mean?
Patients who develop orthopedic-device infections, the most serious postop complication, are likely to require subsequent amputation—the very thing that limb-sparing surgery was performed to circumvent.
The authors of the St. Jude study attribute the high incidence of such infections to a number of factors, the two most important of which are immunosuppression subsequent to chemotherapy and use of an indwelling central venous catheter.
In fact, said Dr. Adderson, anything that impairs the immune system will automatically predispose patients to infection. “And it's awfully hard to beef it up when it's compromised.”
Dr. Eckardt said that chemotherapy should not be restarted until the patient recovers from surgery. The protocol at UCLA after a diagnosis of osteosarcoma is three months of chemotherapy, followed by surgery with three weeks of recovery, and then another six months of chemotherapy.
Other risk factors include anemia, neutropenia, poor nutrition, prolonged breaches of epithelial integrity, use of contaminated cadaveric allografts, and substantial tissue loss that necessitates skin and other grafts, all of which can seed infection in the orthopedic device.
What can be done to prevent infection and its prolonged, expensive treatment, as well as the heartbreak that results from amputating a limb that could have been saved?
▪ Bone allografts should be used sparingly, if at all, and kept as sterile as possible. The patient should receive prophylactic antibiotics.
▪ Dr. Adderson said that surgeons are now using orthopedic devices that are less prone to infection. “They are made of new polymers that not only do not allow bacteria to adhere but are impregnated with short-term antibiotics, as is the cement that fixes the device,” she said.
▪ Antibiotic therapy should be instituted immediately upon the onset of focal infection, especially in the tibia because of the relative paucity of soft tissue covering it.
▪ A clinical diagnosis of orthopedic-device infection should be followed up with invasive diagnostic techniques such as biopsy and culture of deep tissue or the orthopedic device.
▪ Before dental or other invasive procedures, patients who have orthopedic devices should receive prophylactic antibiotics.
▪ Providing hyperoxia and maintaining normal blood glucose levels during surgery can reduce the frequency or severity of infections.
▪ Cover the device with a good layer of muscle and skin—that is the most important factor in preventing orthopedic-device infections, Dr. Malawer said. “If the skin does become infected, it's a ‘nothing problem.’ You just debride it and put on a graft. It heals well.”
Dr. Adderson described infection prophylaxis on the near horizon: orthopedic devices that can be lengthened with magnets—and without surgery—as the child grows; a vaccine for Staphylococcus aureus; and a biosensor imbedded in the orthopedic device that would monitor the presence of bacteria, thus warning of early infection.
Treatment, especially of orthopedic-device infections, is problematic. Aggressive early and/or prolonged antibiotics may reduce the likelihood of chronic infection and other complications, but the risks of exposing patients to broad-spectrum antibiotics for a long time are well known.
The only other alternative is more surgery. But another operation subjects an ill, immunocompromised patient to the potential for poor wound healing and the risk of advanced cancer if adjuvant chemotherapy is delayed further.
Still, Dr. Adderson and her colleagues concluded, it is extremely unlikely that an orthopedic-device infection will be cured without removal of the affected orthopedic device, and prolonged efforts to treat such infections with antibiotics alone may not be appropriate except in exceptional circumstances.