Purpose: To evaluate if a different morphology of the implant-abutment connection (internal vs. external hexagon) is able to condition the behaviour of hard and soft peri-implant tissues. Materials and methods: Twenty patients with significantly unfavourable prognoses for their residual maxillary or mandibular dentitions were selected and rehabilitated with immediately loaded fixed full-arch rehabilitations in two different centres. Four to six implants with identical macro- and micro-topography were inserted in each arch: external hexagon implants (EHC) in one randomly selected side of the dental arch and internal hexagon implants (IHC) in the other side. Primary outcome measures were the success rates of the implants and prostheses. Any technical and biological complication was recorded. Secondary outcome measures were: peri-implant marginal bone level (MBL) changes, Plaque Index (PI), probing depth (PD) and bleeding on probing (BoP), evaluated at implant insertion and at 3, 6 and 12 months post-loading. Results: Forty-three EHC and 40 IHC implants were inserted in 20 patients. No patients dropped out. Two implants failed; one IHC after 3 months and one EHC after 6 months in two different patients (difference IHC vs. EHC at patient level: 0.06%; 95% CI: −1.9 to 2.1; P = 0.99). No prosthesis failed. No biological complications were identified and three loose prosthetic abutment screws were identified in three different patients (two EHC and one IHC); difference at patient level IHC vs. EHC: 2.1% (95% CI: −0.8 to 5; P = 0.43). Overall marginal bone loss was not significantly different between the two treatment groups (EHC vs. IHC) at any time point. The mean difference of bone levels between EHC and IHC was 0.25 mm (95% CI: −0.18 to 0.69) at implant placement. Mean difference between IHC and EHC was −0.01 mm (95% CI: −0.34 to 0.36) at 3 months, 0.13 mm (95% CI: −0.48 to 0.22) at 6 months and 0.11 mm (95% CI: −0.45 to 0.25) at 12 months. All the implants showed good periodontal health at the 1-year-in-function visit, with no statistically significant differences between groups. At 12 months mean (± standard deviation) PI was 2 (± 1.5) for the EHC and 1.85 (± 1.58) for the IHC group (P = 0.57) with a mean difference between the two groups of 0.15 (95% CI: −0.56 to 0.85). Mean PD was 2.23 mm (± 0.52) for the EHC and 2.10 mm (± 0.39) for the IHC group (P = 0.39), with a mean difference between the two groups of 0.12 mm (95% CI: −0.08 to 0.33). At 12 months 41.4% of EHC and 43.6% of IHC implants presented no BoP (mean difference: −2.2%, 95% CI: −24.0 to 19.3; P = 0.51). No significant effect of centres over all outcomes was identified (P = 0.71 for MBL, P = 0.14 for PI, P = 0.14 for PD and P = 0.20 for BoP). Conclusions: On the basis of the present trial the two types of implant connections were clinically reliable. After 12 months in function, both implants provided good clinical outcomes, without statistically significant differences between the two groups. Keywords: dental implants, immediate loading, implant abutment connection Conflict of interest statement: The authors declare that this study was partly funded by Sweden & Martina.