Abstract
BackgroundEndoscopic repair is an emerging surgical modality for proximal hamstring ruptures, however the effect on return to sport (RTS) and sport performance outcomes compared to open repair remains unclear.
MethodsMEDLINE, Embase and CENTRAL were searched for studies reporting sport-related outcomes of open and/or endoscopic repair of proximal hamstring ruptures. The proportion of patients who returned to preinjury or any level of sport, time to RTS, and sport-related patient-reported outcome measures were abstracted. For each repair type, a meta-analysis of proportions was conducted to pool the proportion of patients who returned to sport using a random effects model.
ResultsThirty-eight studies published between January 2004 and March 2025 were reviewed, including 1,684 patients (1,698 repairs, 34 studies) treated by open repair and 116 patients (117 repairs, 5 studies) by endoscopic repair. Mean follow-up time was approximately 35 months (range 6–156 months) in the open group and 32 months (range 12–77 months) in the endoscopic group. Among open repairs, the pooled rate of return to preinjury level of sport (24 studies, n = 1175) was 80.7% (95% CI, 71.4–88.6%, I2 = 93.0%). Furthermore, the pooled rate of return to any level of sport (16 studies, n = 879) was 95.0% (95% CI, 89.6–98.7%, I2 = 85.2%). Among endoscopic repairs, the pooled rate of return to preinjury level of sport was 92.3% (95% CI, 50.1–100%, I2 = 60.4%). No study on endoscopic repair reported the proportion who returned to sport at any level. Re-ruptures were reported in 43 patients (3.35%, n = 1,283, 22 studies) in the open group and 1 patient (1.6%, n = 62, 3 studies) in the endoscopic group. Reoperations were reported in 38 patients (2.8%, n = 1,369, 24 studies) in the open group and 1 patient (0.9%, n = 116, 5 studies) in the endoscopic group.
ConclusionReturn to sport after proximal hamstring repair is high across surgical techniques. While endoscopic repair may achieve comparable rates of return to preinjury level of sport, confidence in this estimate is limited by a lack of available data.