Background: In 2010, South Africa launched an ambitious national HIV counseling and testing (HCT) campaign. Routine opt-out HCT may impact retention in care, as it is expected to result in a higher proportion of asymptomatic patients with lower levels of immunodeficiency who may not be motivated to seek HIV care. We examined retention in care before and after implementation of routine opt-out HCT at a primary care clinic. Methodology: In January 2011, Witkoppen Health and Welfare Centre in Johannesburg, South Africa, implemented routine opt-out HCT for all adult (≥18 years) clinic clients with unknown HIV status and those testing negative >3 months ago, regardless of reason for clinic visit. Prior to 2011, clinic clients were tested upon request (voluntary CT), if they accessed antenatal care services, TB services, or if the health care provider requested HCT due to symptoms compatible with HIV infection. We compared patient characteristics and retention in care among clinic clients newly diagnosed with HIV from January-June 2010 (pre-intervention) and February-July 2012 (post-intervention). Results: In total, 1092 individuals were newly diagnosed during the pre-intervention period and 967 post-intervention. During both periods, median age was 32 years and almost half had immigrated to South Africa from neighboring countries. Most new HIV diagnoses were women (67.2% in 2010 and 65.9% in 2012), but the proportion of pregnant women decreased from 26.6% in the pre-intervention to 16.2% in the post-intervention period (p<0.001). Surprisingly, the median CD4 count at time of HIV diagnosis was similar during both periods: 253 (IQR 128-398 cells/μl) in 2010 and 268 (IQR 139.5- 408 cells/μl) in 2012 (p=0.15). The proportion of individuals eligible for ART was 66.7% in the pre-intervention period and 64.1% in the post-intervention period (p=0.21). Within the first 12 months following HIV diagnosis more than half of all individuals newly diagnosed with HIV were lost to follow up: 58.4% during the pre-intervention and 54.6% during the post-intervention. Adjusting for age, CD4 count, gender, pregnancy and nationality, the hazard of LTFU was not impacted by the change in policy (HR 0.97, 95% CI: 0.86-1.09). Conclusions: Implementation of routine opt-out HCT at the primary care facility level did not increase baseline CD4 count or proportion of people eligible for ART, and did not impact the risk of loss to follow up among those newly diagnosed with HIV. Retention in care after HIV diagnosis, however, was extremely low, with more than half of those newly diagnosed lost to care by one year. To be successful, expansion of HIV testing should be coupled with effective strategies to increase retention in care among those newly diagnosed.