The gold standard for treatment of symptomatic full thickness rotator cuff tears has historically been open rotator cuff repair as pioneered by Codman [46
]. Klepps et al and others have documented the validity and reproducibility of this procedure [13
]. Despite good results reported with open rotator cuff repair, significant morbidity and prolonged rehabilitation have been associated with the requisite deltoid take-down and repair [9
]. In response to reports of prolonged pain and rehabilitation after open rotator cuff repair, the arthroscopically assisted "mini-open" or "portal-extension" technique was popularized [23
]. In an effort to further decrease post-operative pain and rehabilitation time, Johnson described the first completely arthroscopic rotator cuff repair [62
]. Since the introduction of the all-arthroscopic rotator cuff repair technique, there has been considerable debate over the benefits of this procedure versus the "mini-open" technique. Several reports have documented good results after arthroscopic repair [33
]. Numerous reports have also touted the arthroscopically-assisted "mini-open" procedure (< 3 cm) for small and medium sized tears of the rotator cuff [23
The current study evaluated functional outcome in similar patient groups undergoing arthroscopically-assisted or completely arthroscopic rotator cuff repair. With the numbers available, there was no statistical difference between the two groups for any independent variable. (Table ). When data at the most recent follow-up was compared to pre-operatively for the whole group, there was a statistical improvement in 7 out of 9 clinical parameters. Although active internal rotation was improved compared to pre-operatively, the improvement did not meet statistical significance. Finally SF-12 scores were essentially unchanged from pre-operatively. Since the SF-12 measures well being, in addition to physical parameters, several parameters not-related to the patients' shoulder may have contributed to this lack of improvement [41
]. For both groups, the overall improvement observed in pain and function is comparable to reports by other authors [41
The amount of biceps pathology noted in our study was over 50%. We attributed this relatively high prevalence of biceps abnormalities to the strict criteria used in our evaluation. Any fraying of the long head of the biceps was considered abnormal. The strict criteria followed may have over-classified biceps abnormalities that did not correlate clinically.
In order to better analyze outcome, ANOVA was performed to analyze the outcome improvement between the 2 groups for the 9 measures used in the study. We found no statistical difference in improvement between the 2 groups for any variable. With the numbers available, we found no statistical difference in shoulder range of motion, pain, or functional outcome between an arthroscopically-assisted or completely arthroscopic technique.
Our analysis using the SF-36 outcome measures demonstrated no significant difference between pre and post operative scores, despite having significant improvement in SST, UCLA and Constant & Murley scores. This is in agreement with Gartsman et al who have used UCLA, Constant & Murley and SF-36 forms to evaluate patients after rotator cuff repair [41
There are several weaknesses to the current study. The data is limited to one surgeon and may not necessarily be applied to all surgeons who perform rotator cuff repairs with varying skill levels. The numbers in the current study are relatively small. With the numbers available, we did not achieve statistical power (power = 0.07). In order to statistically confirm that both mini-open and arthroscopic techniques have similar results with a power value of 0.8 and alpha value of 0.05, we would require 511 patients in each group assuming the current mean scores and standard deviation. Although the authors standardized the post-operative physical therapy regimen, we did not have the same therapist for all patients. This potential variability in post-operative treatment may have influenced the outcome in some patients.
MRI accuracy in the current study was 58%, with 42% of full thickness tears missed. Although the increased number of MRI misdiagnosed complete rotator cuff tears is a cause for concern, we do not believe that this weakness had any bearing on the indications, surgical intervention, nor outcome of the study cohort. Certainly, all patients who underwent surgical intervention failed at least 3 months of conservative treatment, regardless of whether the pre-operative MRI demonstrated a full thickness tear. Arguably, if post-operative magnetic resonance imaging were to be used to evaluate cuff integrity, the current imaging techniques at our institution would be called into question. However, when using the clinical criteria and post-operative measures currently used, we do not believe this weakness in the current study confounded any outcome variable.
Finally, we did not perform magnetic resonance imaging or ultrasonography on all patients at the most recent follow-up. Several authors have described the lack of integrity of rotator cuff repairs when analyzed with these modalities [47
]. Despite these reports, the lack of rotator cuff integrity may not correlate with clinical outcome [47
]. Currently the authors obtain magnetic resonance imaging of all patients' operated shoulders at yearly intervals. However, the current data indicates no significant difference in clinical outcome between the 2 groups. Such imaging data may be more pertinent in evaluating the technical aspects of repair in the 2 groups or as a component of outcome analysis at longer term follow-up.