A Prospective Cohort Study of Patients Treated Surgically or Non-Surgically for Full-thickness Rotator Cuff Tears

OBJECTIVES: The objectives of this project are: (1) to compare the efficacy of surgical versus non-surgical management of full-thickness rotator cuff tears, and (2) to detect variables that predict success within each treatment group. METHODS: Adult patients presenting at MedSport, University of Michigan, with full thickness rotator cuff tears diagnosed by MRI or ultrasound and no history of shoulder surgery were recruited for inclusion in the study. Consenting subjects were given baseline written questionnaires which included a Functional Comorbidity Index, a Charlson Comorbidity Index, demographics form, WORC score, ASES score, VR-12, modified shoulder activity level, and patient satisfaction forms, as well as an 100 point visual analogue pain scale. Patients then proceeded with planned treatment, either surgical or non-surgical, as recommended by their treating surgeon and decided upon themselves. Written questionnaires were mailed to participants at the following intervals after baseline: 4 weeks, 8 weeks, 16 weeks, 32, weeks, 48 weeks, and 64 weeks. Electronic medical records of subjects were monitored for treatment received. We described all patient demographic characteristics, and performed logistic regression for variables associated with treatment allocation and longitudinal modeling with generalized estimating equations for treatment effects across the follow-up period. We also used Student’ t-tests and Wilcoxon rank-sum tests where appropriate, to explore differences in treatment effects between the groups for all outcome measures at all time points. RESULTS: A total of 184 patients were included with 101 allocated to surgery and 73 to non-surgical treatment. Those allocated to surgery were younger (OR=0.87, P<0.0001), had a lower BMI (OR=0.83, P=0.007), less likely to be smokers (OR=0.04, P=0.002), more likely to have a known traumatic injury (OR=2.40, P=0.002), and had a lower comorbidity score (OR=0.74, P=0.096). Table 1 contains the results of all clinical outcomes measures used. At baseline the surgery group tended to be in more pain and have a higher, worse, WORC score. Both the surgical group and non-surgical group tended to improve on all outcome measures across the follow up period with the surgery group improving at a faster rate and to a greater degree after 8 or 16 weeks post baseline (see figure 1). Several variables predicted improvements across time in the surgical group including: older age, non-diabetic, shorter duration of symptoms, being male, without a known traumatic injury and a lower BMI. For the non-surgical group the following variables predicted improved outcomes across time: shorter duration of symptoms, higher baseline activity score, non-smoker, non-worker’s compensation case, nonsmoker, younger in age and lower BMI. These predictors varied by outcome measure. CONCLUSION: Patients with rotator cuff tears who undergo surgical or nonsurgical treatment tend to improve, with surgical patients improving to a greater degree. There appear to be several predictors of improved of treatment allocation and outcomes that may help us to tailor our treatments to individuals