Background and purpose — In the last decade, the direct anterior approach (DAA) for total hip arthroplasty (THA) has become more popular in the Netherlands. Therefore, we investigated the learning curve and survival rate of the DAA in primary THA, using data from the Dutch Arthroplasty Register (LROI). Patients and methods — We identified all patients who received a primary THA using the DAA in several highvolume centers in the Netherlands between 2007 and 2019 (n = 15,903). Procedures were ordered per surgeon, using date of operation. Using the procedure number, operations were divided into 6 groups based on the number of previous procedures per surgeon (first 25, 26–50, 51–100, 101–150, 151–200, > 200). Data from different surgeons in different hospitals was pooled together. Revision rates were calculated using a multilevel time-to-event analysis. Results — Patients operated on in group 1–25 (hazard ratio [HR] 1.6; 95% CI 1.1–2.4) and 26–50 (HR 1.6; CI 1.1–2.5) had a higher risk for revision compared with patients operated on in group > 200 THAs. Between 50 and 100 procedures the revision risk was increased (HR 1.3; CI 0.9–1.9), albeit not statistically significant. From 100 procedures onwards the HR for revision was respectively 1.0 (CI 0.6–1.6) and 0.8 (CI 0.5–1.4) for patients in operation groups 101–150 and 151–200. Main reasons for revision were loosening of the stem (29%), periprosthetic infection (19%), and dislocation (16%). Interpretation — We found a 64% increased risk of revision for patients undergoing THA using the DAA for the first 50 cases per surgeon. Between 50 and 100 cases, this risk was 30% increased, but not statistically significant. From 100 cases onwards, a steady state had been reached in revision rate. The learning curve for DAA therefore is around 100 cases.