Tubercular abscesses are relatively common complications of tubercular spondylodiscitis. Fifty-one patients with suspected abscesses were selected from a group of 97 patients with tubercular spondylodiscitis and submitted to US. In 10 cases CT was performed before US and detected 7 abscesses, all of them confirmed by US. In the extant 41 cases, CT followed US; in 13 cases only US poorly visualized ilio-psoas muscles. As for the group of 23 patients who underwent both CT and US, if the former method is assumed as the reference gold standard, overall US sensitivity is 97% (1 false negative) and its specificity is 100%. In all cases where US findings were accurate and specific enough, CT was not performed; the patients were followed every seventh month and no abscesses found. US showed abscesses in the iliac fossa in 20 cases, along the psoas fascia in 6 and in the thighs in 3 cases. Two cases of gluteal localization were observed, together with 1 Grynfelt's triangle abscess, 1 Petit's triangle and 1 Scarpa's triangle abscesses; finally, 1 abscess was found in the knee. The most common appearance of tubercular abscesses is a hypoechoic and inhomogeneous pattern; sometimes caseum makes the abscess solid and hyperechoic. Calcifications were unusual in our series. All patients were submitted to percutaneous drainage under US guidance. The results proved US to allow the early and unquestionable diagnosis of tubercular abscesses and to confirm clinical suspicion. Moreover, US is also useful to guide percutaneous drainage and to follow the patients after drainage. As for CT, it remains the method of choice to depict vertebral involvement, but, in our series, it exhibited no significant advantages over US in the study of abscessual lumbar collections.