Thirty-nine patients (24 men and 15 women) were enrolled in the study. Their demographic details are shown in Table 1. Twenty (51%) had previous exposure to antiretrovirals, and 19 (49%) were antiretroviral-naive. Nine (23%) had previous exposure to zidovudine (ZDV) monotherapy during 1995 and 1996. Thirteen patients had previous exposure to protease inhibitor (PI) treatment in 1996 and 1997; all had received saquinavir. Three patients had previous exposure to nonnucleoside reverse transcriptase inhibitors (nonnucleoside analogues): 1 patient self-discontinued nevirapine (NVP) after 1 week of therapy because of symptoms of methadone withdrawal; the second patient discontinued NVP because of a grade 2 cutaneous reaction; and the third patient failed to continue efavirenz (EFV) after 1 week of therapy and failed to return to the clinic for 3 months.

For the program of directly observed HAART, the combination of antiretrovirals was individually chosen, and an attempt was made to simplify the regimen using once-daily dosing where possible. For example, if stavudine (d4T), didanosine (ddI), and EFV were prescribed for a patient, he was given a daily dose of ddI and EFV, along with the morning dose of d4T, before receiving his methadone. The patient was then given his evening capsule of d4T to take at home.

All 39 patients received triple-drug antiretroviral therapy: 35 received a regimen including 2 nucleoside analogues (16, d4T plus ddI; 13, ZDV plus lamivudine [3TC]; 6, d4T plus 3TC) and either a PI (7, nelfinavir [NFV]) or a nonnucleoside analogue (18, EFV; 10, NVP). A triple-nucleoside analogue combination of ZDV, 3TC, and abacavir was prescribed for 4 patients.

Four of the original 39 patients (10%) discontinued treatment before their 3-month review. In 2 patients receiving a combination that included dual nucleoside analogues and EFV, significant hepatic dysfunction developed, which necessitated discontinuation of treatment. Another patient -- also receiving EFV -- discontinued all medications against medical advice, because of the sudden onset of anxiety and agitation. Another patient initially attended regularly but, after 3 months of successful therapy and a recurrence of a depressive illness, failed to return to either the medical or the drug treatment clinic.

Significant diarrhea, initially unresponsive to antidiarrheals, developed in 1 patient receiving NFV; however, his symptoms resolved spontaneously after 3 weeks of antidiarrheal therapy. Eight patients receiving EFV complained of methadone withdrawal symptoms during the first month of therapy. Symptoms developed in 4 patients in the first 3 days of treatment; since their symptoms were considered a direct CNS effect of EFV, their methadone dosage was not altered. Symptoms developed in the other 4 patients during the first 2 weeks. Clinically, all 4 patients had signs of methadone withdrawal. Their methadone dosage was adjusted in increments of 10 mg/d until their symptoms resolved. Of the 35 remaining patients, 31 were still attending and receiving DOT at 6 months, with 27 of these still attending at 1 year.

Figure 1 demonstrates the increases in CD4⁺ cell counts over time. Immunologic and virologic outcomes for treatment-naive and treatment-experienced patients have been amalgamated because there was no significant difference found between the groups. The incremental mean CD4⁺ cell count changes from baseline were significant at 3, 6, and 12 months: 27 cells/µL (P < .01), 39 cells/µL (P < .001), and 43 cells/µL (P < .001), respectively.

Using an intent-to-treat analysis, the percentage of patients with HIV RNA levels below the limit of detection (less than 50 copies/mL, Roche Ultrasensitive assay) was 33%, 46%, and 36% at 3, 6, and 12 months, respectively. Using an on-treatment analysis, this increased to 38%, 60%, and 58%, respectively (Figure 2).