High-Z nanoparticles incorporated in tumors have great potential to enhance radiotherapy treatments by increasing the absorbed radiation dose in the vicinity of the nanoparticle. However, current Monte Carlo simulations estimate dose enhancement metrics in 1D, neglecting the anisotropy of the energy deposition around the nanoparticle. This could have radiobiological consequences and affect treatment accuracy. Accurate knowledge of the spatial distribution of the dose or other microdosimetric quantities is necessary for predicting radiobiological outcomes using biophysical models. In this study, we utilized a new combination of two Monte Carlo Variance Reduction Techniques (VRTs) to estimate microdosimetric quantities at the nanoscale, taking into account the anisotropy of the microdosimetric distributions. Our results show significant anisotropy in multi-events frequency-mean specific energy, as well as in microdosimetric distributions. These findings highlight the importance of considering anisotropy in microdosimetric distributions when predicting radiobiological outcomes and suggest that current 1D simulations may not accurately represent the real microdosimetric effects. Our work contributes to the development of multiscale dosimetry approaches for radiotherapy treatments with high-Z nanoparticles.