Patellar fixation in MPFL reconstruction: a network meta-analysis comparing bone tunnel and suture anchor techniques

Abstract

Although medial patellofemoral ligament (MPFL) reconstruction has become the cornerstone procedure for recurrent patellar instability (RPI), the optimal patellar-side fixation technique remains debated. Direct head-to-head comparisons are scarce, and traditional pairwise meta-analyses cannot simultaneously rank multiple competing techniques. This network meta-analysis (NMA) aimed to compare available evidence on clinical outcomes and complications across patellar-side fixation techniques. This systematic review and network meta-analysis was performed in accordance with the PRISMA-NMA statement and prospectively registered in PROSPERO (CRD420261366055). A systematic search of Web of Science, PubMed, Embase, Scopus, Google Scholar, and Cochrane CENTRAL was conducted through April 2026. Eleven studies (5 RCTs and 6 observational studies) involving 1,362 knees were included. Five patellar-side fixation techniques for MPFL reconstruction were evaluated: single-patellar-tunnel, double-suture-anchor, semi-patellar-tunnel, double-patellar-tunnel, and triple-patellar-tunnel. Primary outcomes were RPI and postoperative Kujala score at final follow-up (not change from baseline); secondary outcomes were revision rate and complications. A frequentist NMA was performed using a random-effects model, with interventions ranked by SUCRA. A 0.5 continuity correction was applied for zero-event cells. Heterogeneity was assessed using the Q test and I 2 statistics, and consistency between direct and indirect evidence was evaluated via node-splitting models. For RPI, semi-patellar-tunnel ranked highest (SUCRA 81.8%), although no pairwise comparison reached statistical significance (e.g., double-suture-anchor vs. semi-patellar-tunnel: RR 1.88, 95% CI 0.93–3.83). For Kujala score, semi-patellar-tunnel ranked highest (SUCRA 98.5%); certain pairwise comparisons were statistically significant (e.g., semi-patellar-tunnel vs. single-patellar-tunnel: MD 6.14, 95% CI 4.12–8.16), but the observed mean differences lay below the minimal clinically important difference. For revision rate, semi-patellar-tunnel ranked highest (SUCRA 76.0%), followed by double-suture-anchor; the most informative pairwise contrast (double-suture-anchor vs. semi-patellar-tunnel) yielded an RR of 1.83 (95% CI 0.76–4.42). For surgical complications, semi-patellar-tunnel ranked highest (SUCRA 68.0%), with triple-patellar-tunnel ranking second. All 95% CIs for binary outcomes remained wide and included the null value, reflecting limited statistical precision rather than evidence of equivalence. These SUCRA-based rankings should be interpreted with caution: SUCRA values reflect cumulative ranking probability rather than direct evidence of superiority, and the probability of being the single best treatment was substantially lower than SUCRA for all outcomes (e.g., RPI: 45.7%; Kujala: 94.1%; revision: 33.5%; complications: 25.3%). Sensitivity analyses confirmed the robustness of the findings for RPI (probability of being best ranged 35.3–50.3% across leave-one-out iterations), whereas Kujala score rankings demonstrated moderate sensitivity to individual study exclusion (range 64.5–97.6%). The certainty of evidence was rated as very low for all outcomes per the GRADE framework. Semi-patellar-tunnel fixation ranked highest for RPI and Kujala score in this network meta-analysis, but binary outcome comparisons were not statistically significant and the certainty of evidence was very low. These findings should be considered hypothesis-generating and require confirmation in adequately powered comparative studies.