Clinical recovery and structural integrity after tendon-bone interface versus intra-articular platelet-rich plasma during arthroscopic rotator cuff repair: a three-group retrospective cohort study

Platelet-rich plasma (PRP) is used as a biologic adjunct during arthroscopic rotator cuff repair (ARCR), but anatomical delivery may influence its local effect. We hypothesized that tendon-bone interface delivery would enhance early functional recovery and structural healing by concentrating PRP at the repair footprint, whereas intra-articular delivery would mainly modulate postoperative inflammation. Because these effects were expected to act early, we anticipated that functional differences would diminish as healing and rehabilitation progressed. This retrospective three-group cohort study included 365 of 478 screened adults who underwent ARCR between January 2021 and May 2024: tendon-bone interface PRP (TBI-PRP; n = 122), intra-articular PRP (IA-PRP; n = 119), and control ( n = 124). The primary outcome was change in the University of California, Los Angeles (UCLA) Shoulder Rating Scale through 12 months. Secondary outcomes included Constant-Murley and American Shoulder and Elbow Surgeons (ASES) scores, pain visual analog scale (VAS) pain scores, active range of motion, and MRI structural integrity classified by Sugaya type. Linear mixed-effects models assessed longitudinal clinical outcomes. The primary UCLA outcome and all eight secondary clinical outcomes improved over time (all time effects P < 0.001). UCLA and Constant-Murley scores showed no group or time-by-group effects. ASES scores demonstrated a group effect ( P = 0.013) and time-by-group interaction ( P = 0.034). At 3 months, TBI-PRP exceeded control by 3.0 points (95% CI 1.34–4.66); at 6 months, TBI-PRP and IA-PRP exceeded control by 2.5 points (95% CI 1.08–3.92) and 1.6 points (95% CI 0.16–3.04), respectively. No pairwise difference was significant at 12 months. These small, transient ASES differences had uncertain clinical importance. Pain VAS and range-of-motion measures showed no significant group or interaction effects. MRI-defined retear rates were 6.6% for TBI-PRP, 11.8% for IA-PRP, and 12.9% for control ( P = 0.223). The primary UCLA trajectory did not differ significantly among groups through 12 months. Constant-Murley scores did not differ significantly; ASES differences were transient; pain and range-of-motion trajectories did not differ significantly; and the lower TBI-PRP retear rate was not statistically significant. This direct three-group comparison adds evidence on PRP delivery location without establishing a preferred anatomical target.