BACKGROUND: Despite substantial progress in reducing pregnancy-related preventable morbidity and mortality, these remain unacceptably high in developing countries. In 2016, the World Health Organization (WHO) revised recommendations for antenatal care (ANC) from a 4-visit model to a minimum of 8 ANC contacts to reduce perinatal mortality further and improve women's experience of care. The guidelines also recommend that the first ANC visit (ANC-1) should occur during the first trimester. OBJECTIVES: To describe the uptake of routine ANC and its associated factors in South Africa (SA) prior to the 2016 WHO recommendations, when the country recommended 4 ANC visits, to bring to light potential challenges in achieving the current recommendations. METHODS: Secondary data analyses were performed from 3 facility-based, cross-sectional national surveys, conducted to measure 6-week mother-to-child transmission of HIV and coverage of related interventions in SA. These surveys recruited mother-infant pairs attending selected public primary healthcare facilities for their infants' 6-week immunisation in 2010, 2011 -2012 and 2012 -2013. Quantitative questionnaires were used to gather sociodemographic and antenatal-to-peripartum information from Road to Health cards and maternal recall. The inclusion criteria for this secondary assessment were at least 1 ANC visit, the primary outcome being uptake of ≥4 ANC visits. A multivariable logistic regression model was used to: (i) identify maternal factors associated with ANC visits; and (ii) establish whether receiving selected ANC activities was associated with frequency or timing of ANC-1. RESULTS: Of the 9 470, 9 646 and 8 763 women who attended at least 1 ANC visit, only 47.5% (95% confidence interval (CI) 45.4 -49.6), 55.6% (95% CI 53.2 -58.0) and 56.7% (95% CI 54.3 -59.1) adhered to ≥4 ANC visits, while 36.0% (95% CI 34.5 -37.5), 43.5% (95% CI 42.0 -45.1) and 50.8% (95% CI 49.3 -52.2) attended ANC-1 early (before 20 weeks' gestation) in 2010, 2011 -2012 and 2012 -2013, respectively. Multiparity and lower socioeconomic status were significantly associated with non-adherence to the 4-visit ANC recommendation, while a later survey year, higher education, being married, >19 years old, HIV-positive, planned pregnancy and knowing how HIV is transmitted vertically were strongly related to ≥4 ANC visits. The number of women who received selected ANC activities increased significantly with survey year and ≥4 ANC visits, but was not associated with timing of ANC-1. CONCLUSIONS: Despite increases in the uptake of ≥4 ANC visits and early ANC-1 rates between 2010 and 2013, these practices remain suboptimal. Adhering to ≥4 ANC visits improved coverage of selected ANC activities, implying that strengthening efforts to increase the uptake of ANC from at least 4 to 8, could improve overall outcomes.