Prosthetic Joints—Loose or Infected?

abstract & commentary

Synopsis: Prosthetic joint infections were identified using cultures taken by surgeons during surgery. The study found that surgeons should take at least three careful cultures and have them separately processed to determine if infection is present. Evaluating blood studies and culture fluid before surgery may also be beneficial.

Source: Atkins BL, et al. Prospective evaluation of criteria for microbiological diagnosis of prosthetic-joint infection at revision arthroplasty. J Clin Microbiol 1998;36:2932-2939.

Atkins and colleagues in oxford, england, studied 297 cases of prosthetic knee and hip revisions over 17 months. Because of their concern that Gram stain results were not reliable and that specimen collection and culture methods were not well standardized, they asked their surgeons to take cultures at surgery, which included a swab of joint fluid, capsular tissue, femoral and tibial membrane, and other abnormal tissue. Cultures were taken using separate sterile instruments for each specimen and processed for aerobic and anaerobic organisms. An average of 4.06 specimens was taken per patient.

Their definition of infection was the presence of five or more polymorphonuclear leukocytes per high power field on histology exam provided the patient did not have an inflammatory process such as rheumatoid arthritis. One pathologist read all histology specimens. Based on this criterion, 41 patients (13.8%) were diagnosed with infection—even though 35% of those had no growth on culture. Gram staining had a sensitivity of 6% and specificity of 99.7% for a positive culture and 12% and 98.8%, respectively for positive histopathology.

The culture results showed a variety of organisms, the most frequent being coagulase-negative staphylococci followed by corynebacterium, then propionibacterium, Staphylococcus aureus, and streptococci. The number of positive cultures from each patient was also examined. The histology results correlated well with microbiology results when at least three specimens were found to contain the same bacteria. A mathematical model was constructed from the data, which showed a relative likelihood of infection of 0.7 if one culture is positive. For two positive cultures showing the same organism, the likelihood was 4.3, and for three or more positive cultures, 25.9.

There is little information on the clinical evaluation, treatment, or outcome data for patients in the study, although Atkins et al do point out that their policy is to remove the prosthesis for six weeks before replacing it if a definite infection is established with positive histology.

Comment by Alan Tice, MD, FACP

The problem of appropriate treatment of prosthetic joint infections is a frequent and important one. This study provides some useful results and underscores the value of cultures in confirming an infection. It seems that all too often, infectious disease consultations are requested after the joint is replaced and a single culture of the wound grows a skin organism after three days.

A diagnosis of infection with a prosthetic joint may be easy in some cases but clinical indicators may not be able to distinguish between loosening of the joint and infection. It is not unusual for a surgeon to be unpleasantly surprised at surgery. The situation is further complicated by coagulase-negative staphylococci, which are not only the most frequent contaminants but the most likely pathogens as well. In addition, coagulase-negative staphylococci may hide within the slime they create on prosthetic joint surfaces and not produce significant apparent disease.

The Gram stain has been used by some surgeons as an indicator of infection but prior studies, and this one, point out it is a poor and unreliable one. The use of histopathology to make a diagnosis of a prosthetic joint infection may provide a measure of accuracy and standardization, but it is not a timely one and also has limitations in that only 65% of cultures are positive, despite meeting histopathology criteria. Even with two or three positive cultures for the same organism, the histology was not always positive.

Atkins et al allude to the possible value of the specific bacteria in making a diagnosis of infection but do not provide clinical data. Certainly, the recovery of a S. aureus would make a serious infection more likely than a Propionibacterium. It would be nice to know if the culture or histopathology results correlate with outcomes of the infections.

This study suggests surgeons should use a protocol to take at least three careful cultures (one joint fluid and one tissue) at surgery and that they be separately processed. This would be helpful in many cases although the cost of the cultures should be considered.

The problem still remains as to whether to put in a new joint at the time of surgery, especially if there is some suspicion of infection. Some cases are obviously clinically infected. A Gram stain may be useful if pus is present but may not be helpful otherwise. Histopathology studies for inflammatory cells and cultures are not timely. The best approach may be a careful evaluation before surgery with blood studies (ESR, C-reactive protein) and cultures of joint fluid. There may also be benefit to aspirates or biopsies of abnormal areas on radiographic studies or where the ends of the prosthesis may be loose or particularly painful.