BACKGROUND Earlier antiretroviral therapy initiation can reduce the incidence of HIV-1. This benefit can be offset by increased transmitted drug resistance (TDR). We compared the preventive benefits of reducing incident infections with the potential TDR increase in East Africa. METHODS A mathematical model was constructed to represent Kampala, Uganda, and Mombasa, Kenya. We predicted the effect of initiating treatment at different immunological thresholds (<350, <500 CD4 cells/μl) on infections averted and mutation-specific TDR prevalence over 10 years compared to initiating treatment at CD4 cell count below 200 cells/μl. RESULTS When initiating treatment at CD4 cell count below 350 cells/μl, we predict 18 [interquartile range (IQR) 11-31] and 46 (IQR 30-83) infections averted for each additional case of TDR in Kampala and Mombasa, respectively, and 22 (IQR 17-35) and 32 (IQR 21-57) infections averted when initiating at below 500. TDR is predicted to increase most strongly when initiating treatment at CD4 cell count below 500 cells/μl, from 8.3% (IQR 7.7-9.0%) and 12.3% (IQR 11.7-13.1%) in 2012 to 19.0% (IQR 16.5-21.8%) and 19.2% (IQR 17.1-21.5%) in 10 years in Kampala and Mombasa, respectively. The TDR epidemic at all immunological thresholds was comprised mainly of resistance to non-nucleoside reverse transcriptase inhibitors. When 80-100% of individuals with virological failure are timely switched to second-line therapy, TDR is predicted to decline irrespective of treatment initiation threshold. CONCLUSION Averted HIV infections due to the expansion of antiretroviral treatment eligibility offset the risk of transmitted drug resistance, as defined by more infections averted than TDR gained. The effectiveness of first-line non-nucleoside reverse transcriptase inhibitor-based therapy can be preserved by improving switching practices to second-line therapy.