Sir

In your article “Better adherence vital in AIDS therapy” (Nature 390, 326; 1997) you stressed the importance of persuading HIV patients to follow the complex new drug treatments.

We agree, but our data show that other factors may also be of importance, as shown by experience in Brazil where, since 1996, the government has made combined therapy available to all Brazilians with AIDS.

The Ambulatorio da Providencia Outpatient Clinic and Support House in Rio de Janeiro, run by the Roman Catholic Church, has worked on HIV infection since 1985. During that time, there has been a constant increase of HIV infection among our patients. The clinical population profile encompasses the poor and the ‘social outcasts’ of the city: slum (favela) dwellers, street children, sex workers, transsexuals and beggars. Some 700 HIV-positive/AIDS patients are followed up by the clinic and most of them are homeless or living in favelas. Some 25 homeless AIDS patients can be housed at the Santo Antonio Support House.

In 1990 we started a prospective study comparing survival time between homeless and formerly homeless housed individuals with AIDS in Rio de Janeiro. HIV-1-positive homeless subjects, HIV-positive volunteers who previously lived on the streets and agreed to live at our support house, and a control group of people living with relatives and who came from the poorest section of Rio society were followed up to 1997.

Survival time was calculated from the date of AIDS diagnosis until the date of death. We divided the study into two phases. The first phase was from 1990 to 1995, when combined therapy was not available (we used monotherapy and/or double therapy) and the second from 1996 when triple therapy was made available by the government.

In the first phase, the mean survival of the homeless group (27 patients) was 8.2 months (range 1 to 33), in the formerly homeless group (26 patients) 17.8 months (1 to 48) and in the control group (59 patients) 18.3 months (2 to 60). The results from the log-rank test for pairs of survival curves revealed significant statistical differences between the groups of homeless and formerly homeless subjects (P = 0.0018) and the groups of homeless and housed subjects (P = 0.0001). In contrast, there was no significant statistical difference between the survival curves for the groups of formerly homeless and housed subjects (P = 0.9704). Better nutrition and hygiene, more frequent medical and psychological care together with controlled medicine intake, occupational therapy, decrease in promiscuity, alcohol and drug abuse and the provision of religious support may explain the difference in survival time observed in this phase of the study.

The second phase of the study has shown no deaths since the beginning of triple therapy (mean 13 months) for the formerly homeless (10) and housed (12) patients; all the patients are clinically well, CD4 counts have increased and viral loads are not detectable or below 700 copies (indicating efficacy of therapy). In contrast, the survival time of homeless subjects has remained the same as that observed previously, because we do not give combined therapy if we cannot monitor adherence to treatment.

The use of combined therapy has increased the cost of each patient on therapy at the support house. A new problem has emerged because the number of clinically well patients discharged from the house has been very low, consequently increasing the list of HIV-positive homeless waiting to be taken into the support house.

These findings highlight the severe impact of the HIV-1 epidemic and its treatment in people from the poorest part of our society. The results of this study should be the reason for some cautious optimism by governments and encourage them to pursue measures to deal with longer life of patients and resocialization of social outcasts.