Patient-specific guides and osteosynthesis versus CAD/CAM occlusal splints in orthognathic surgery: a systematic review and meta-analysis of accuracy, operative time, and complications.
BACKGROUND Orthognathic surgery increasingly relies on digital transfer workflows, most commonly CAD/CAM occlusal splints or patient-specific implants (PSI). However, prior meta-analyses were limited by methodological heterogeneity, sparse randomized evidence, and insufficient axis-specific analyses, leaving the magnitude and clinical relevance of any accuracy advantage unclear. This systematic review and meta-analysis aimed to compare PSI and CAD/CAM splint workflows in orthognathic surgery. METHODS This study followed PRISMA 2020 guidelines and was registered in PROSPERO (CRD420251039754). PubMed, Embase, and Cochrane were searched from inception to August 23, 2025. Comparative clinical studies evaluating PSI versus CAD/CAM splints in orthognathic surgery were included. Outcomes included linear accuracy, angular accuracy, operative time, and postoperative complications. The primary linear accuracy outcomes were axis-specific translational deviations across the mediolateral, anteroposterior, and superoinferior directions. Angular accuracy was assessed across pitch, roll, and yaw. Randomized trials were assessed using RoB 2.0, and non-randomized studies using the Newcastle-Ottawa Scale. RESULTS Thirty studies (12 RCTs and 18 cohort studies; 1317 patients) were included; 21 provided sufficient quantitative data for meta-analysis and 9 were synthesized narratively. Compared with CAD/CAM splints, PSI was associated with smaller axis-specific linear deviations across all three translational axes (all P < 0.01); a supportive combined linear summary showed the same direction but was not treated as a primary outcome. A possible advantage in angular accuracy was suggested (MD - 0.54°, 95% CI - 1.07 to - 0.01), but this was based on only six maxillary comparisons and should be interpreted cautiously. No statistically significant difference in postoperative complications was detected (RR = 1.01, 95% CI 0.64-1.60, P = 0.959). For operative time, current evidence did not demonstrate a statistically significant difference between workflows; however, few studies and substantial heterogeneity precluded firm conclusions. CONCLUSION Compared with CAD/CAM splints, PSI workflows were associated with smaller axis-specific translational deviations, while the available angular evidence suggests possible but limited rotational advantages. Current evidence remains inconclusive regarding operative time and does not demonstrate a significant difference in complication rates. PSI may be considered selectively in cases requiring high transfer precision; further studies should assess long-term stability and patient-reported outcomes.