PurposeTo compare the clinical and functional outcomes of arthroscopic rotator cuff repair over a period of 2 years using three postoperative rehabilitation modalities: aquatic therapy, land-based therapy, and self-rehabilitation therapy. The null hypothesis was that aquatic therapy would provide no difference in Constant score compared to land-based therapy and self-rehabilitation therapy.MethodsA prospective study was performed on subjects scheduled for arthroscopic rotator cuff repair between 2012 and 2017 that complied with the following criteria: (i) small to medium sized symptomatic supraspinatus and/or infraspinatus tendon tears, (ii) low to moderate tendon retraction according to Patte, and (iii) fatty infiltration stage ≤2. Patients were allocated to perform either aquatic therapy, land-based therapy, or self-rehabilitation therapy for 2-4 months. Independent observers blinded to the study design collected Constant score, SSV, and patient satisfaction at 2 months, 3 months, 6 months, 1 year and 2 years.Study designLevel III, cohort studyResultsAt 2 months follow-up, patients performing aquatic therapy had significantly higher Constant scores (p < 0.001) and SSV (p < 0.001) compared to those performing land-based therapy or self-rehabilitation therapy. At 3 months follow-up, patients performing aquatic therapy had significantly higher Constant scores (p < 0.001), and SSV (p < 0.001), both of which exceeded the respective minimal clinically important differences (MCIDs) of 10.4 and 12. Patients performing aquatic therapy continued to have significantly higher Constant scores and SSV at 6 months, 1 year, and 2 years.ConclusionAquatic therapy has a positive effect on clinical outcomes at 3 months after surgery, but yields no relevant improvements on function or satisfaction at 1 to 2 years follow-up.