Collection of 4–5 independent peri-prosthetic tissue samples is recommended for microbiological diagnosis of prosthetic joint infections. Sonication of explanted prostheses has also been shown to increase microbiological yield in some centres. We compared sonication with standard tissue sampling for diagnosis of prosthetic joint and other orthopaedic device related infections. We used standard protocols for sample collection, tissue culture and sonication. Positive tissue culture was defined as isolation of a phenotypically indistinguishable organism from ≥2 samples; and positive sonication culture as isolation of an organism at ≥50 cfu/ml. We compared the diagnostic performance of each method against an established clinical definition of infection (Trampuz 2011), and against a composite clinical and microbiological definition of infection based on international consensus (Gehrke & Parvizi 2013). 350 specimens were received for sonication, including joint prostheses (160), exchangeable components (76), other orthopaedic hardware and cement (104), and bone (10). A median of 5 peri-prosthetic tissue samples were received from each procedure (IQR 4–5). Tissue culture was more sensitive than sonication for diagnosis of prosthetic joint and orthopaedic device related infection using both the clinical definition (66% versus 57%, McNemar's Χ2 test p=0.016) and the composite definition of infection (87% vs 66%, p<0.001). The combination of tissue culture and sonication provided optimum sensitivity: 73% (95% confidence interval 65–79%) against the clinical definition and 92% (86–96%) against the composite definition. Results were similar when analysis was confined to joint prostheses and exchangeable components; other orthopaedic hardware; and patients who had received antibiotics within 14 days prior to surgery. Tissue sampling appears to have higher sensitivity than sonication for diagnosis of prosthetic joint and orthopaedic device infection at our centre. This may reflect rigorous collection of multiple peri-prosthetic tissue samples. A combination of methods may offer optimal sensitivity, reflecting the anatomical and biological spectrum of prosthetic joint and other device related infections.
