Purpose: To compare the clinical outcome of immediately loaded cross-arch maxillary prostheses supported by zygomatic implants vs conventional implants placed in augmented bone. Materials and methods: In total, 71 edentulous patients with severely atrophic maxillas without sufficient bone volumes for placing dental implants, or when it was possible to place only two implants in the anterior area (minimal diameter 3.5 mm and length of 8 mm) and less than 4 mm of bone height subantrally, were randomised according to a parallel group design to receive zygomatic implants (35 patients) to be loaded immediately vs grafting with a xenograft, followed after 6 months of graft consolidation by placement of six to eight conventional dental implants submerged for 4 months (36 patients). For immediate loading, zygomatic implants had to be inserted with an insertion torque superior to 40 Ncm. Screw-retained metal reinforced acrylic provisional prostheses were provided, to be replaced by definitive Procera Implant Bridge Titanium prostheses (Nobel Biocare, Göteborg, Sweden), with ceramic or acrylic veneer materials 4 months after initial loading. Outcome measures were: prosthesis, implant and augmentation failures, any complications, quality of life (OHIP-14), patients' number of days with total or partial impaired activity, time to function and number of dental visits, assessed by independent assessors. Patients were followed up to 1 year after loading. Results: No augmentation procedure failed. Five patients dropped out from the augmentation group. Six prostheses could not be delivered or failed in the augmentation group vs one prosthesis in the zygomatic group, the difference being statistically significant (difference in proportions = -16.5%; P = 0.045; 95% CI: -0.34 to -0.01). Eight patients lost 35 implants in the augmentation group vs two patients who lost four zygomatic implants, the difference being statistically significant (difference in proportions = -20.1%; P = 0.037; 95% CI: -0.38 to -0.02). A total of 14 augmented patients were affected by 22 complications, vs 28 zygomatic patients (40 complications), the difference being statistically significant (difference in proportions = 34.8%; P = 0.005; 95% CI: 0.12 to 0.54). The 1-year OHIP-14 score was 3.93 ± 5.86 for augmented patients and 3.97 ± 4.32 for zygomatic patients with no statistically significant differences between groups (mean difference = 0.04; 95% CI: -2.56 to 2.65; P = 0.747). Both groups had significantly improved OHIP-14 scores from before rehabilitation (P < 0.001 for both augmented and zygomatic patients). On average, the number of days of total infirmity was 7.42 ± 3.17 for the augmented group and 7.17 ± 1.96 for the zygomatic group, the difference not being statistically significant (mean difference = -0.25; 95% CI: -1.52 to 1.02; P = 0.692). The number of days of partial infirmity were on average 14.24 ± 4.64 for the augmented group and 12.17 ± 3.82 for the zygomatic group, the difference being statistically significant (mean difference = -2.07; 95% CI: -4.12 to -0.02; P = 0.048). The mean number of days that needed to have a functional prosthesis was 444.32 ± 207.86 for augmented patients and 1.34 ± 2.27 for zygomatic patients, the difference being statistically significant (mean difference = -442.98; 95% CI: -513.10 to -372.86; P < 0.001). The average number of dental visits was 19.72 ± 12.22 for augmented patients and 15.12 ± 5.76 for zygomatic patients, the difference not being statistically significant (mean difference = -4.61; 95% CI: -9.31 to 0.92; P = 0.055). Conclusions: Preliminary 1-year post-loading data suggest that immediately loaded zygomatic implants were associated with statistically significantly fewer prosthetic failures (one vs six patients), implant failures (two vs eight patients) and time needed to functional loading (1.3 days vs 444.3 days) when compared to augmentation procedures and conventionally loaded dental implants. Even if more complications were reported for zygomatic implants, they proved to be a better rehabilitation modality for severely atrophic maxillae. Long-term data are absolutely needed to confirm or dispute these preliminary results.