This article reviews literature on the epidemiology, pathogenicity, and control of HIV and Mycobacterium tuberculosis coinfection. Regarding pathogenicity, immune system deterioration makes HIV-infected people more likely to develop active tuberculosis on primary or secondary exposure to the bacillus or to suffer reactivation of latent infections, and to experience considerably higher rates of extrapulmonary manifestations, relapses, and death. Regarding epidemiology, as of 1990 there were an estimated 3 million people coinfected with HIV and M. tuberculosis, with some 300,000 active tuberculosis cases and 120,000-150,000 tuberculosis deaths occurring annually among those coinfected. Over 500,000 coinfected people are thought to reside in the Americas, over 400,000 of them in Latin America. In general, the impact of coinfection is evident. Relatively high and increasing prevalences of HIV infection have been detected among tuberculosis patients around the world, and tuberculosis has become a frequent complication of AIDS cases. Moreover, there is no longer any doubt that coinfection obstructs tuberculosis prevention and control. Among other things, it affects BCG vaccination policies, suggests the need to administer preventive chemoprophylaxis to HIV-infected individuals at high risk of harboring or contracting tuberculosis infections, and complicates both detection and treatment of active tuberculosis cases. The recent proliferation of M. tuberculosis strains resistant to multiple drugs, most notably in the United States, compounds the problem. Tuberculosis prevention and control are still technically and economically feasible. However, more must be done to establish surveillance programs with laboratory support. More research is needed to determine what case prevention measures are best-suited to current circumstances and the HIV/AIDS presence. More effective preventive treatment regimens that are well tolerated, well complied with, and do not pose the risk of multiresistance need to be devised. More health workers need to be trained to suspect tuberculosis and to conduct timely and appropriate tests confirming this diagnosis. And finally, more must be done to standardize the types and durations of the various curative treatment regimens employed.