Cone-beam computed tomography (CBCT) and radiographic outcomes of crestal sinus elevation, performed using mineralized human bone allograft, were analyzed to correlate results with maxillary sinus size. A total of 60 sinus augmentations in 60 patients, with initial bone ≤5 mm, were performed. Digital radiographs were taken at surgical implant placement time up to post-prosthetic loading follow-up (12-72 months), when CBCT evaluation was carried out. Marginal bone loss (MBL) was radiographically analyzed at 6 months and follow-up time post-loading. Sinus size (BPD), implant distance from palatal (PID) and buccal wall (BID), and absence of bone coverage of implant (intra-sinus bone loss--IBL) were evaluated and statistically evaluated by ANOVA and linear regression analyses. MBL increased as a function of time. MBL at final follow-up was statistically associated with MBL at 6 months. A statistically significant correlation of IBL with wall distance and of IBL/mm with time was identified with greater values in wide sinuses (WS ≥ 13.27 mm) than in narrow sinuses (NS < 13.27 mm). This study is the first quantitative and statistically significant confirmation that crestal technique with residual ridge height <5 mm is more appropriate and predictable, in terms of intra-sinus bone coverage, in narrow than in WS.