Deep infection in tumor endoprosthesis around the knee: a multi-institutional study by the Japanese musculoskeletal oncology group

Abstract

Background: The incidence of endoprosthesis failure has been well studied, but few studies have described the clinical characteristics of deep infection in tumor prostheses. This study aimed to analyze the characteristics of deep infection in tumor endoprostheses around the knee.

Methods: We analyzed clinical data of 57 patients with deep infections involving tumor endoprostheses around the knee enrolled from the Japanese Musculoskeletal Oncology Group. Profile of clinical presentation including time between surgery and infection, initial symptoms/blood tests and microbial cultures was evaluated. In addition pre-, intra-, and postoperative clinical factors influencing clinical presentation and treatment outcomes of infections were analyzed.

Results: Mean interval between the initial operation and diagnosis was 13 months, and mean time required for infection control was 12 months. The most common pathogen was Staphylococcus. Infection control rates were significantly higher when prostheses were removed rather than salvaged. Ten-year prosthesis survival and limb salvage rates were 41.6% and 75.6%, respectively. Analysis of underlying clinical factors suggested that soft-tissue condition significantly influenced the duration of the infection control period and likelihood of limb salvage.

Conclusions: Infection control is a prolonged process. Deep infection frequently results in amputation or prosthesis loss. Intensive analysis of clinical characteristics may aid infection control.

Figure 1

Kaplan-Meier curves showing duration from initial surgery until infection of each case.

Figure 2

Kaplan-Meier curves showing the duration of the interval between diagnosis of infection and the completion of infection control (A). Factors that might influence this period include tumor location (B), tumor extension (C) and infection presentation period (D).

Figure 3

Box-and-whisker plots showing the distribution of the number of surgical interventions required to control infection (A). Factors significantly influencing the number of surgeries included comorbidity as infectious risk, tumor origin (primary versus metastatic), application of gastrocnemius flap (in the tibia cases), and discharge/pus at the initial presentation of infection. Success rate for each surgical modality (B). The modalities in this figure were arranged in order from higher to lower success rate. Modalities with removal of prostheses were significantly more successful in infection control than those that left the prostheses in place (right two rows).

Figure 4

Kaplan-Meier survival curves showing prosthesis survival (A) and limb salvage (B) in patients with deep infections after insertion of endoprostheses. Factors that may significantly influence limb survival include extraarticular resection (C) and discharge/pus at the initial presentation of infection (D).