The early results of total ankle arthroplasty were marred by inferior implant designs and uncertain surgical techniques1. These are the major reasons why so many foot and ankle surgeons were, and are today, reluctant to select arthroplasty over ankle arthrodesis. Today, proponents of total ankle arthroplasty have refined their surgical techniques, and improved implant designs have reduced formerly catastrophic failure rates. As the practice of total ankle arthroplasty matures, however, it faces the same hurdles seen during the development of other total joint arthroplasties: an increasing attention to epidemiological and risk factors associated with the surgery. Of particular concern among the complications of any joint replacement procedure such as total ankle arthroplasty is the risk of infection, which inevitably leads to implant failure and poor patient outcomes.
While we know that implant-associated infections are a major concern in total ankle arthroplasty, as the authors of the present study explain, we do not know as much about the risk factors in total ankle arthroplasty as we do about those in other implant procedures. We do not really have a good consensus on what the infection rate is. Thus, Kessler et al. conducted a case-control study of twenty-six patients with periprosthetic infection along with two control groups (twenty-six patients each) who did not develop infection. The clever design of their study allowed them to match each infection case with two controls and better account for the variables of age, sex, and date of surgery. Also wisely, the authors realized that risk factors for infection lie on a continuum (preoperative, operative, and postoperative) as well as in the patient demographics.
The major preoperative factors associated with infection were prior surgery at the site of infection and a low American Orthopaedic Foot & Ankle Society hindfoot score. Neither of these is a huge revelation, but they represent scientific confirmation of things that sound likely but have not been proven. The duration of surgery was the major operative factor associated with infection. Again, this is not a surprising finding in a study of periprosthetic infection, but we have not seen this confirmed with regard to total ankle arthroplasty. Likewise, prolonged wound dehiscence or a secondary disturbance in wound-healing was associated with infection risk.
I doubt that anyone would have seen these risk factors as unlikely candidates for a predisposition to eventual periprosthetic infection, but often the things “everybody knows” are wrong, and I do not think we should accept them dogmatically unless someone has demonstrated that they really are true. As the authors correctly noted, this is a case-control study, and as such it has some limitations of its own, especially the potential for sampling bias. However, all patients were followed as in a prospective study after the index surgery, and thus the cases and the controls were selected before the outcome of infection or no infection. This lends a great deal of validity to the study. The first major limitation of the study is not its design but its size—the absolute number of infections. This is one of the difficulties we all face in studying infection, and obviously in developing statistical studies. We are all working diligently to reduce the absolute number of infections, and have seen great success, but we generally need very large studies, probably from multiple centers, before we are able to compile a database that allows us to really dig into a multivariate analysis of risk factors for infection. The second major limitation is the lack of demographic data on the patients with regard to known risk factors for infections of other total joints, such as diabetes, obesity, use of tobacco, vascular compromise, cardiac and pulmonary status as well as a history of other infections and whether the patients are carriers of methicillin-resistant Staphylococcus aureus (MRSA).
I applaud the authors’ work. This report is a good example of the path we need to take in studying how to improve the outcomes of total ankle arthroplasty. It is a route that has already been well traveled by colleagues in hip and knee replacement, and I believe foot and ankle surgeons will use the knowledge gained through studies such as this one, and its successors, to drive down the rate of infection in a fashion similar to that seen in hips and knees. There is one catch: this road ends somewhere over a very distant horizon, or may never really end. Given the catastrophic consequences of infection in joint replacement, we will always be studying possible methods to reduce even small overall risks. This case-control study is an early step on that road.