Improved method for detecting knee replacement infections based on extended combined Tc-99m-white blood cell/bone imaging

The purpose of this study was to evaluate whether an extension of the imagi ng time to 24 h post-injection improves the diagnostic accuracy of teclinet ium-99m-hexamethylpropyleneamine oxime (Tc-99m-HMPAO) leucocyte imaging in detecting knee replacement infections. Thirty patients were studied, with i nfection confirmed in eight (27%) and excluded in 22 on the basis of clinic al and microbiological findings. Leucocyte imaging was carried out at 2-4 h (routine images) and at 24 h (late images) post-injection. For comparison, bone imaging with technetium-99m-hydroxydiphosphonate (Tc-99m-HDP) was car ried out at arterial, soft tissue and metabolic phases. Late leucocyte imag ing was found to be more sensitive (100% vs. 87.5%) and more specific (82% vs. 77%) than routine leucocyte imaging in detecting infections. All the bo ne imaging methods showed a sensitivity of 100%, whereas the specificity va ried from only 5% to 23%. All procedures had high negative predictive value s (NPVs) (94 to 100%) for excluding infection. However, the positive predic tive value (PPV) was only 28 to 32% for bone imaging and 58% for routine le ucocyte imaging, whereas late leucocyte imaging showed a PPV of 67% and a d iagnostic accuracy of 87%. The data indicate that late leucocyte imaging ma y be superior to routine leucocyte imaging for examining patients with symp tomatic knee replacements. ((C) 2001 Lippincott Williams & Wilkins).