OBJECTIVES: (1) To measure and compare endo-sinus bone levels around implants randomly placed with an osteotome sinus floor elevation (OSFE) procedure in grafted (control) and non-grafted (test) sinuses, (2) to evaluate the OSFE efficacy with short, tapered, and chemically modified hydrophilic surfaced implants in extremely atrophic maxillae, (3) to show that fused corticals may constitute a complication risk. MATERIAL AND METHODS: The TE(®) SLActive 8 mm-long implants (Straumann AG) were placed using an OSFE procedure in 4 mm or less of bone height. Healing time before prosthetic rehabilitation was 10 weeks. One year after implant placement, bone levels were measured on standardized periapical radiographs. RESULTS: Thirty-seven (17 tests, 20 controls) implants were placed in 12 patients with a mean maxillary residual bone height (RBH) of 2.4 ± 0.9 mm. Before loading, two control implants failed (RBH 1.4 and 1.2 mm); two others rotated at loading (one test, RBH 0.9 mm; one control, RBH 1.5 mm) but were uneventfully loaded after three additional months of healing. These adverse events and complications occurred when implants were placed in merged corticals. Endo-sinus bone gain was 3.9 ± 1.0 and 5.0 ± 1.3 mm for the test and control groups (P = 0.003). The 1-year success rate was 100% and 90%, respectively (P = 0.49). CONCLUSION: Although more bone is gained when grafting material is used, this may not be required to promote endo-sinus bone gain. The OSFE procedure with or without grafting material could be efficient when the RBH is ≤4 mm. However, when both corticals merged, the risk of complication could increase.