The purpose of this study is to compare the safety and effectiveness of 2 forms of stavudine (d4T). One form is taken once a day (extended release) and the other form is taken twice a day (immediate release).

Patients are randomized to receive blinded stavudine extended release (d4T ER) or immediate release (d4T IR) formulation. Randomization is balanced by screening HIV viral load of less than 30,000 copies/ml or at least 30,000 copies/ml and by site. Patients also receive open-label efavirenz (EFV) and lamivudine (3TC). Efficacy is evaluated over 48 weeks of the dosing period and safety is evaluated over the entire dosing period (56 weeks).

Inclusion Criteria

Patients may be eligible for this study if they:

• Are at least 18 years old. Patients living in USA, Puerto Rico, Australia, Brazil, Singapore, and Thailand can be at least 12 years old (need consent of parent or guardian if under 18).
• Have a viral load of at least 2,000 copies/ml within 21 days of study entry.
• Have a CD4 count of at least 100 cells/mm3 within 21 days of study entry.
• Agree to use a barrier method of birth control (such as condoms) during the study.
• Are available for follow-up for at least 56 weeks.

Exclusion Criteria

Patients will not be eligible for this study if they:

• Are pregnant or breast-feeding.
• Have taken nonnucleoside reverse transcriptase inhibitors (NNRTIs) or protease inhibitors (PIs) for more than 30 days and within 14 days of study entry.
• Have a new opportunistic (HIV-related) infection or condition requiring treatment.
• Have acute (early) HIV infection.
• Have diarrhea (at least 6 loose stools/day for at least 7 days in a row) within 30 days prior to study entry.
• Abuse alcohol or drugs.
• Have active hepatitis within 30 days prior to study entry.
• Have a history of peripheral neuropathy (a condition affecting the nervous system).
• Cannot take medications by mouth.
• Are allergic to certain antiviral drugs.
• Need to take certain medications that should not be taken with EFV.
• Have certain other conditions or prior treatments that might affect the study.