Bearing steels are heat treated to obtain martensitic microstructures that provide high hardness necessary for good rolling contact fatigue performance. For the most common bearing steel, SAE 52100, without specific thermal treatments such as stabilizing tempering, the microstructure consists of unstable phases that can evolve in service leading to detrimental dimensional variations. A previous publication of the 10th ASTM bearing steel conference held in 2014 focused on explaining and modeling the dimensional variations induced by thermal aging of SAE 52100 and the role of retained austenite in the expansion occurring in service. The influence of a few heat treatment parameters on retained austenite content was also briefly discussed based on the literature. Very few publications are available in the literature regarding the influence of heat treatment parameters on dimensional stability. An experimental study was conducted at NTN-SNR to better understand this topic. It was based on the same methodology presented in 2014 but with different initial heat treatments. The influence of austenitizing parameters such as time and temperatures, but also of the cooling conditions during quench, were examined. Different experimental techniques were used to precisely qualify and quantify microstructural evolutions: thermoelectric power measurements, synchrotron X-ray diffraction, and quench dilatometry. Through this study, it was concluded that the widely held belief that the amount of retained austenite is the only parameter to assess expansion in service was too simple. In some cases, heat treatments with similar retained austenite contents led to significant differences in dimensional stability. This article describes such results, depending on the initial heat treatment, and proposes an explanation on the microstructural phenomena leading to these differences.