I thank Carucci and Halvorsen for their comments. I agree that the lack of pathologic proof makes it impossible to be 100% certain that we are dealing with a case of bacillary angiomatosis (BA). However, the patient had multiple blood cultures, cultures of body fluids, and bone marrow biopsy, none of which demonstrated evidence of other fungal or mycobacterial infection. Bartonella species were grown on skin and blood cultures.

To the best of my knowledge, hyperattenuating lymphadenopathy in patients with acquired immunodeficiency syndrome with BA has been described in 12 cases [14]. The details of scanning technique and contrast infusion were not given in these case studies. A nonhelical or single-slice computed tomographic (CT) scanner would have been used. The largest and most recent of these studies [4] indicates that contrast enhancement may be variable. Because vascular lesions such as hemangioma may show variable enhancement depending on the time of scan, it is not certain that hyperenhancing nodes would necessarily be seen with rapid scanning using multichannel CT scanners. I am therefore uncomfortable making categorical statements about enhancement patterns of nodal disease in BA. Necrotic or low attenuation centers have been reported in adenopathy/mesenteric mass associated with hepatic peliosis, particularly after treatment [1, 5].

Although most cases of BA with skin lesions respond to antibiotic therapy [6], peliosis indicates a systemic form of the infection which may be fatal [7]. Causes of death include disseminated intravascular coagulation and liver failure [2, 6]. Our patient had a low fibrinogen level, a low platelet count, prolonged prothrombin time, and elevated D-dimer. A premortem diagnosis of disseminated intravascular coagulation was made. In many cases of suspected peliosis, as in the case I presented, liver biopsy may be dangerous. Attempted drainage of such lesions was reported to be fatal in one case [8].

The main learning point of this report is for the radiologist to raise this diagnosis given the appropriate CT findings so that empiric treatment may be started before culture results are available. In the presented case, BA was not diagnosed until several days after admission and the patient died despite starting appropriate therapy.