Purpose of the study: The objective of this study was to identify clinical and anatomic factors which could affect the outcome of tendon healing after arthroscopic repair of rotator cuff tears.

Material and methods: This prospective cohort study included 122 patients who underwent arthroscopic treatment between May 1999 and September 2002. One hundred twelve patients (114 shoulders) were reviewed (93.4%). Mean age at surgery was 61 years. An arthroscan (78% of patients) or magnetic resonance imaging (MRI) were performed six months postoperatively. Mean follow-up was 24 months (range 13–46 months).

Results: Rotator cuff tears healed completely in 64% of the shoulders (n=73) and partially in 7% (n=8). No healing was observed in 29% (n=33). The Constant score was 49.8 preoperatively and 82.4 at last follow-up (p< 0.0001). Cuff healing improved clinical outcome with a Constant score of 85 points versus 77, particularly for force, 14.5 points versus 10 without healing. Four factors were statistically predictive of tendon healing: duration from symptom onset to operation (24 months for healed tears versus 37 months, p< 0.05); age at cuff repair (81% healed tears in patients aged less than 50 years versus 50% for patients aged over 65 years, p< 0.002); sagittal extension of the rotator interval or extension to the upper third of the infrascapularis (45% healed tears versus 79% without anterior extension, p< 0.0001); fatty degeneration (69% of healed tears for Goutallier grade 0 versus 38% for grade 1 and 2, p< 0.01).

Conclusion: This study demonstrated that four factors can predict tendon healing: time to treatment, age at surgery, anterior extension of the tear, fatty degeneration.

Purpose: We hypothesised that the rate of tendon healing after arthroscopic repair of full-thickness tears of the supraspinatus is equivalent to that obtained with open techniques reported in the literature.

Material and methods: We studied prospectively a cohort of 65 patients with arthroscopically repaired full-thickness tears of the supraspinatus. The patients were reviewed a mean 19 months (12–43) after repair. At arthroscopy, patients were aged 59.5 years on average (28–79). Bone-tendon sutures were performed with resorbable thread and self-locking anchors positioned on the lateral aspect of the humerus. Repair was protected with an abduction brace for six weeks. Forty-one patients (63%) accepted an arthroscan performed six months to two years after arthroscopy to assess tendon healing.

Results: Ninety-four percent of the patients were satisfied with the outcome. The mean Constant score was 51.6±10.6 points preoperatively and 80.2±13.2 at last follow-up (p< 0.001). The arthroscan showed that the rotator cuff had healed in 70% of the cases (29/41). The supra-spinatus had not healed on the trochiter in eight cases (25%) and was partially healed in two (5%). The size of the residual tendon defect was less than the initial tear in all cases except one. The rate of patient satisfaction and function was not significantly different if the tendon had healed (Constant score 81.3/100, satisfaction 93%) or if there was a residual tendon defect (Constant score 77.5/100, satisfaction 92%). Shoulder force in patients with a healed tendon (6±1.9 kg) was better than in those with a tendon defect (4.5±2.8 kg), but the difference was not significant. Factors affecting tendon healing were age > 65 years (43% healing, p< 0.02), and wide tears.

Conclusion: Arthroscopic repair of isolated supraspinatus tears enables tendon healing in 70% of cases as demonstrated by arthroscan. This rate was equivalent to those reported in historical series of open repair. Patients aged over 65 years had significantly less satisfactory healing. The absence of tendon healing does not compromise functional and subjective outcome despite reduced force.