Drs Collignon and Sorrell have drawn our attention to another group of patients who have disseminated candidiasis with macronodular skin lesions. It is fascinating that these authors, as well as Dupont and Drouhet,1 have described mini-epidemics of this form of candidiasis occurring within days of intravenous use of "batches" of heroin. The latter authors attempted to culture Candida from the very heroin that resulted in the infection, but were unable to do so.
The patients reported in our study denied illicit drug use. Therefore, it is not readily apparent why our alcoholic patients developed a unique form of candidiasis that was quite similar to the lesion described in drug addicts. Needle sharers have a notoriously high prevalence of serologic markers for parenterally transmitted viruses. For unclear reasons, alcoholics also have an unusually high prevalence of markers of hepatitis B virus.2 Perhaps the common link between alcoholics and drug