OBJECTIVES: To describe the resistance mutations selected by a first-line regimen of zidovudine/lamivudine/tenofovir in the absence of real-time viral load monitoring. DESIGN: A substudy of 300 participants from the Development of Antiretroviral Therapy in Africa trial in Uganda and Zimbabwe, which compared managing antiretroviral therapy with and without laboratory monitoring. METHODS: Stored plasma samples from selected time points were assayed retrospectively for HIV-1 RNA. The pol gene in all baseline samples and those with HIV RNA >1000 copies per milliliter at weeks 24 and 48 were sequenced. RESULTS: The proportion with HIV RNA >1000 copies per milliliter increased from 15% at 24 weeks to 24% at 48 weeks. Eighteen of 31 (58%) genotyped samples at 24 weeks had ≥ 1 major nucleoside reverse transcriptase inhibitor-associated mutations compared with 41 of 47 (87%) at 48 weeks. Excluding 1 nonadherent patient, a mean of 2.0 (95% confidence interval: 1.3 to 2.8) thymidine analogue mutations (TAMs) developed between weeks 24 and 48 among 14 patients with HIV RNA >1000 copies per milliliter at both time points. K65R was detected in 8 of 63 (13%) patients and was negatively associated with number of TAMs (P = 0.01) but not viral subtype (P = 0.30). CONCLUSIONS: A high rate of acquisition of TAMs, but not of K65R, among patients with prolonged viraemia was observed. However, most patients were virologically suppressed at 48 weeks, and long-term clinical and immunological outcomes in the Development of Antiretroviral Therapy in Africa trial were favorable.