This study compared three radionuclide techniques in distinguishing musculoskeletal infection from noninfectious inflammation. METHODS Thirty-five orthopedic patients with suspected musculoskeletal infection were examined using three radionuclide techniques in sequence: triphasic bone scintigraphy, 99mTc radioleukocytes (99mTc-WBC) scintigraphy and 99mTc human immunoglobulin (99mTc-Hig) scintigraphy. Two "early" and "late" acquisitions were performed, at 4-6 hr and 20-24 hr postinjection, respectively. Patients who were diagnosed as suffering from noninflammatory lesions became the controls. We calcu"late"d for all studies one index of inflammation (Infl) as the ratio between counts in the uptake area and counts in an equal area of normal tissue. RESULTS The "early" radiolabeled leukocytes and "late" Hig scintigraphy allowed the greatest ability to distinguish between infections and noninfectious inflammations (p < 0.011 and p < 0.016) with a sensitivity of 96.6% and 96.5% and specificity of 71% and 100%, respectively. Hig and radioleukocytes allowed distinguishing infections from noninflammatory diseases at both examinations. CONCLUSIONS The "early" radioleukocyte scintigraphy allowed us to separate infections from noninfectious inflammations. In contrast, the same result can be obtained only with the "late" scan in the Hig study, but Hig mapped the spread of the inflammation into soft tissues better. Hig might be an alternative to radioleukocytes because of its simple preparation, similar accuracy and safety.