We thank the correspondents for their comments. Dr. Harindra correctly notes that cerebrovascular disease should be included in the differential diagnosis of focal neurologic lesions in HIV-infected patients, particularly when the deficits are transient or apoplectic in onset. We and others [1, 2] have reported several mechanisms of cerebrovascular disease in these patients, including nonbacterial thrombotic endocarditis and hemorrhage associated with thrombocytopenia. Vasculitis may be associated with syphilis, herpes zoster, and other viral infections, but the frequency of these disorders in HIV infection is unclear. Although the role of HIV infection in cerebrovascular disease is unknown, Brew and Miller [3] have reported that transient ischemic attacks in HIV-infected patients may respond to antiretroviral therapy.

Dr. Stricker's report of cryoglobulinemic neuropathy is of interest, given that several other autoimmune complications may develop in patients with HIV infection, including idiopathic thrombocytopenic purpura, inflammatory demyelinating polyneuropathy, and polymyositis. One should be cautious, however, in inferring a true association between cryoglobulinemic neuropathy or other neurologic disorders and HIV infection from limited case reports because their concurrence may be coincidental.

The diagnosis and management of neurosyphilis in HIV-infected patients is problematic. Several authors have noted that HIV may alter the natural history of syphilis and its response to conventional therapy [4]. Further, the neurologic symptoms and cerebrospinal fluid profiles of neurosyphilis and HIV infection have considerable overlap. The cerebrospinal fluid VDRL test is not adequately sensitive to be relied on alone for excluding the diagnosis of neurosyphilis [5]. We have therefore suggested that one may presume the diagnosis of neurosyphilis and initiate empiric treatment in an HIV-infected patient who has neurologic signs consistent with neurosyphilis, reactive cerebrospinal fluid, and a positive serum VDRL test result.

Dr. Pulido and colleagues indicate that the serum VDRL test result may be falsely positive in HIV-infected patients and suggest that other treponemal antibody tests be used. Although the sensitivity and specificity of these assays in HIV-infected patients have not been established, we agree that the use of specific treponemal antibody tests may assist in confirming the diagnosis of neurosyphilis. We believe, however, that a risk–benefit analysis favors treating some falsely seropositive patients for neurosyphilis rather than withholding therapy from persons with false-negative results.