Tree-related microhabitats (hereafter “TreMs”) play a key role in forest biodiversity. However, harvesting may cause their erosion. In North America, knowledge about TreMs is still lacking but defect surveys are largely available in managed forests. The objectives of our study were: (1) to demonstrate that defect surveys can be a reliable resource to identify TreMs; and, (2)to evaluate the capacity of silvicultural treatments to maintain TreM abundance and diversity according to a harvest gradient.To achieve these objectives, we identified TreMs from a defect survey performed the year a harvest gradient was applied to20 plots, including uncut control, shelterwood treatments removing 50%, 43% and 36% of basal area, and clearcut (4 plots/treatment). The density and composition of TreMs were then compared based on treatments. Overall, 38% of defectsactually corresponded to TreMs, confirming that tree-defects can be used as TreM proxies. Bark loss was the most abundantTreM. While there was practically no TreM in clearcuts, all shelterwood treatments initially maintained TreM diversity anddensity at the same values found in uncut control plots. Shelterwood systems, especially those maintaining a continuouscover, could therefore prove helpful to sustain TreMs and their biodiversity in managed forests.