Abstract
Despite the apparent success of continuous passive motion after soft-tissue procedures or joint replacements, its effect after repair of the rotator cuff is still unknown. The purpose of this prospective, randomized outcome study was to compare the results of continuous passive motion with those of manual passive range-of-motion exercises after repair of the rotator cuff.Thirty-one patients (thirty-two rotator cuffs) were randomly assigned to one of two types of postoperative management: continuous passive motion (seventeen patients) or manual passive range-of-motion exercises (fifteen patients). There were seventeen women and fourteen men, and the mean age was sixty-three years (range, thirty to eighty years). The patients were followed for a mean of twenty-two months (range, six to forty-five months). Five tears of the rotator cuff were small, eighteen were medium, and nine were large.All of the operations were performed by one surgeon. The patients who were managed with continuous passive motion used the device for the first four weeks postoperatively. The patients who were managed with manual passive range-of-motion exercises were assisted by a trained relative, friend, or home-care nurse. After the four-week period, the two groups were managed similarly for two to five months.According to the Shoulder Pain and Disability Index, a valid and reliable self-administered questionnaire, the treatment was extremely successful in both groups. The overall score was excellent for twenty-seven shoulders (84 per cent), good for two (6 per cent), fair for two (7 per cent), and poor for one (3 per cent). With the numbers available, we could detect no significant differences (p > 0.05) between the two groups with respect to the score according to the Index, pain (according to a visual-analog scale), range of motion, or isometric strength.Manual passive range-of-motion exercises were more cost-effective than continuous passive motion. The limited number of physical-therapy visits associated with the manual passive range-of-motion exercises in the present study appeared to be more cost-effective than a traditional physical-therapy schedule of three visits per week.Postoperative therapy with continuous passive motion or manual passive range-of-motion exercises appears to yield favorable results after repair of a small, medium, or large tear of the rotator cuff.
The functional outcome after repair of the rotator cuff depends on many variables. Appropriate postoperative management is critical for maximizing function in a timely and efficient manner. Typically, the postoperative care is individualized to the patient and depends on the status of the deltoid muscle, the size of the tear, the ability to move the shoulder without injuring the tissues, and the patient's postoperative support system15. Other factors that have an effect on therapy and the final outcome are the duration between the original injury and the operation, the age of the patient, and the limitation of the active range of motion preoperatively. Supervised therapeutic exercises, starting with passive motion, are generally begun soon after the repair and are continued for approximately three months16.
The use of early passive range-of-motion exercises is no longer controversial, as the exercises can be performed with protection of the repair and prevention of deleterious adhesions that can limit function2,18. Continuous passive motion, performed in a reciprocal fashion by a mechanical device, is one form of passive motion that facilitates beneficial changes in articular cartilage, synovial membrane, joint capsule, ligament, and tendon12,25,26. Although much of the clinical research on continuous passive motion has focused on patients who have had an operative procedure on the knee3,4, its use after procedures on the upper extremity is becoming more frequent12. Better healing of repaired flexor tendons of the hand after continuous passive motion compared with that after intermittent passive motion1,9, and the successful outcome associated with its use at other joints3,12,26, have prompted clinicians to use this mode of therapy in an attempt to improve the biological response after repair of the rotator cuff. Continuous passive motion has been used after repair of the rotator cuff when no other mode of passive motion was available, and it also has been used to assist in the relief of pain, to improve healing, and to restore a passive range of motion5,18. However, we are not aware of any prospective study documenting its effectiveness in a well defined population of patients who had a repair of the rotator cuff. The purpose of the present prospective, randomized study was to compare the functional outcome of continuous passive motion with that of a more traditional program of manual passive range-of-motion exercises for patients who had a repair of the rotator cuff.
*No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. No funds were received in support of this study.
†DeRosa Physical Therapy, 77 West Forest Avenue, Suite 303, Flagstaff, Arizona 86001.
‡Departments of Orthopaedics (T. W. and R. J.) and Statistics (J. H.), University of Florida, Gainesville, Florida 32610.
Patients
The study included thirty-one patients (seventeen women and fourteen men) who had had repair of a total of thirty-two tears of the rotator cuff between 1991 and 1994. Nine of the tears were large (greater than three but less than five centimeters), eighteen were medium (one to three centimeters), and five were small (less than one centimeter). One patient had bilateral repair; the tear on the dominant side was repaired first, after which the patient was randomly assigned to management with continuous passive motion, and the repair on the non-dominant side was performed six months later, after which she was randomly assigned to management with manual passive range-of-motion exercises. The mean age at the time of the repair was sixty-three years (range, thirty to eighty years) (Table I). The third-party payer was Medicare or a private insurance company, or both, for twenty-nine patients; the remaining two patients were receiving Workers' Compensation. Patients who had a massive (greater than five-centimeter) or irreparable tear of the rotator cuff; preoperative evidence of pathological instability of the shoulder, a rheumatological disorder, reflex sympathetic dystrophy, fracture, or glenohumeral arthritis; or concomitant adhesive capsulitis were excluded from the study. Those who had had a previous operation on the shoulder were excluded as well. The study was approved by the Institutional Review Board at the University of Florida, and informed consent was obtained from all patients. Randomization was carried out with use of a table of random numbers.
Operative Technique
All repairs were performed in an open fashion by one of us (T. W.). The attachment of the deltoid to the anterior and lateral aspects of the acromion was exposed with use of a saber incision. The deltoid was split for two centimeters between its anterior and lateral heads, and the two portions were elevated as fascial-periosteal sleeves off of the anterior aspect of the acromion. An anterior-inferior acromioplasty was performed, and the bursa was debrided. The rotator cuff was mobilized by first placing stay sutures in the torn ends and then applying traction to the torn ends to retract the cuff. The subacromial and subdeltoid bursae were excised. All scar tissue between the torn tendon and the overlying acromion was released by blunt dissection with use of Darrach retractors or, when scar tissue was thick, by sharp dissection with Mayo scissors. Once this had been done, the cuff could be stretched to its original insertion, where it was repaired back to bone with use of osseous tunnels. (No suture anchors were used.) The subscapularis was not transposed in any patient. Patients who were found to have a massive defect were eliminated from the study. The fascial-periosteal sleeves of the deltoid then were carefully repaired to themselves. (Drill-holes through the acromion for attachment of the deltoid were not used routinely.)
Postoperative Management
All patients were managed initially with a sling. Pain was managed with the parenteral administration of morphine (two to four milligrams every three to four hours) as needed for one day, and then Vicodin (hydrocodone bitartrate and acetaminophen; one or two tablets every four hours) was taken orally as needed. No specific protocol for pain medication was followed postoperatively, but all patients had stopped taking narcotics by two to four weeks postoperatively.
While hospitalized, each patient was managed with passive range-of-motion exercises once a day for a mean of two days (one, two, or three days). The patient was discharged from the hospital when he or she was comfortable enough to return home and initiate the home-exercise program. Each patient was instructed to remove the sling three or four times a day to perform gentle passive pendulum exercises. Before the patients were discharged from the hospital, they were randomly assigned to management with continuous passive motion or management with manual passive range-of-motion exercises. Continuous passive motion was not started while the patient was in the hospital because it would have been difficult to arrange for and set up a bed-mounted continuous-passive-motion device, the hospital stay was relatively short, and the availability of bed-mounted continuous-passive-motion devices was limited. The patients were instructed in the proper use of a home continuous-passive-motion device (Thera-kinetics, Mount Laurel, New Jersey) for elevation and external rotation of the shoulder. The patients were told to use the device for four hours a day (three or four periods of continuous motion, each lasting for one to one and a half hours). The actual duration of continuous passive motion averaged three hours a day according to the patients' reports in a diary. The patients were instructed to progressively increase the range of the continuous passive motion but to maintain a comfortable arc.
Manual passive range-of-motion exercises were carried out by a relative, friend, or home-care nurse who was trained to perform them. Elevation and external rotation exercises—three sets of ten to fifteen repetitions—were performed individually three times a day. The total duration of the exercises averaged forty-five minutes a day according to the patients' reports. If the trained assistant was not available for a session, the passive external-rotation exercises were performed with the use of a cane or a pendulum. For the exercises with the cane, the patient lay supine, with the involved shoulder supported by a towel roll to prevent hyperextension of the shoulder; held the cane in the hand on the involved side; and moved the cane with the other hand so as to externally rotate the involved shoulder.
After the first four postoperative weeks, use of the continuous-passive-motion device was stopped and both groups were managed similarly: passive external-rotation exercises were performed with a cane, and passive elevation exercises were performed by an assistant or with the use of pulleys. Active rotation of the shoulder was initiated at six weeks; active-assisted elevation, at eight weeks; active elevation, at ten to twelve weeks; resistive rotation, at ten to twelve weeks; and resistive elevation, at twelve to fourteen weeks. Functional use and progressive strengthening began during the third or fourth postoperative month and were encouraged for as long as one year.
Evaluation of the Patients
The data were collected by three occupational therapists (two of whom were certified hand therapists), one physical therapist who was a certified hand therapist, and one occupational-therapy assistant. These examiners were not blinded with regard to the treatment groups, and there was crossover in terms of which therapist examined a patient at a given time-point.
Outcome Measures: Pain, Impairment, and Disability
During the first four weeks postoperatively, the patients assessed the pain with use of a visual-analog scale19 once a week and recorded the amount of time for which the passive exercises were performed each day. Active and passive external rotation and elevation were measured with a goniometer (North Coast Medical, San Jose, California) at two weeks (except for active elevation), six weeks, and twelve weeks; at six and twelve months; and at the most recent follow-up evaluation. Isometric strength in external rotation and in elevation were measured with a handheld dynamometer (Spark Instruments and Academics, Coralville, Iowa) at six months, at twelve months, and at the most recent follow-up evaluation. The patient completed the Shoulder Pain and Disability Index23 at the most recent follow-up evaluation. The range of motion and strength at each time-point were measured three times, and the mean was determined. The mean duration of follow-up (and standard deviation) was 22 ± 9.8 months (range, six to forty-five months).
A pilot study was performed on twelve normal shoulders to assess the within-examiner and the between-examiner reliability of the measurements of range of motion with the goniometer and the measurement of strength with the handheld dynamometer. Intraclass correlation coefficients (model 3,127) were calculated to assess the within-examiner and between-examiner reliability for the five therapists who examined the patients. The within-examiner reliability was relatively good, with the intraclass correlation coefficients ranging from 0.72 to 0.97 for the measurements of range of motion and strength. The intraclass correlation coefficients for the between-examiner reliability, however, were fairly low (range, 0.11 to 0.56). Still, the between-examiner reliability and the error square root of the mean square, estimates of the standard deviation in the unit of measurement, may be clinically acceptable (range, 4.5 to 9.7), with the possible exceptions of the between-examiner reliability of the measurement of active external rotation (13.0), active elevation (18.3), and strength in external rotation (30.1).
Statistical Methods
The range-of-motion profiles for the patients were found to be relatively linear when considered on a logarithmic time scale and a logarithmic range-of-motion scale. Linear regression24 was thus used to estimate these slopes, which were inversely related to the rate at which the range-of-motion profile leveled off over time. Therefore, a smaller slope for the range of motion would level off sooner than a larger slope.
Because the scores according to the Shoulder Pain and Disability Index23 were excessively skewed in both groups, the scores for the two groups were compared with use of the non-parametric Fisher exact test and the Wilcoxon rank-sum test24.
The t test for independent samples24 and the Wilcoxon rank-sum test were used to compare the two groups with respect to the simple mean values and the rank sums, respectively, of all other dependent variables.
Analysis of covariance6 also was used to compare the mean values (except for the score according to the Index23) in the two groups while controlling separately for the baseline range of motion (the passive range of motion at two weeks and the active range of motion at twelve weeks), isometric strength (at four months), and the score on the visual-analog scale (at one week). In addition, analysis of covariance was used to compare the mean values for the two groups, controlling separately for gender, age (less than sixty-five years or more than sixty-five years), and size of the tear (small or medium [one group] or large).
A retrospective power analysis was performed with use of the given variability and sample sizes to determine detectable differences between the two groups. Detectable differences were reported as the per cent difference in the mean response between the two groups. The reported percentages were calculated with use of a power of 80 per cent and a type-I error rate of 5 per cent.
Clinically, both groups did well across all outcome measures. Overall, there were no significant differences, before or after adjustments for baseline levels, between the two groups with respect to the simple mean values for active and passive motion or strength. The patients who were managed with continuous passive motion had much less pain during the first postoperative week.
Pain and Disability
The scores according to the Shoulder Pain and Disability Index23 were analyzed in two different ways. First, the total score was arbitrarily divided into quartiles: 0 to 25 points indicated an excellent result, 26 to 50 points indicated a good result, 51 to 75 points indicated a fair result, and 76 to 100 points indicated a poor result. There were fifteen excellent results, one good result, one fair result, and no poor results in the group that had been managed with continuous passive motion, and there were twelve excellent results, one good result, one fair result, and one poor result in the group that had been managed with manual passive range-of-motion exercises (Fig. 1). With the numbers available for study, we could not detect a significant difference between the two groups with respect to the scores (p = 0.853, Fisher exact test). The raw scores also were analyzed statistically and, again, we could detect no significant difference between the two groups (p = 0.16, Wilcoxon rank-sum test).
Pain
The data from the visual-analog scales indicated that pain had decreased in both groups from the first to the fourth week postoperatively. With the numbers available, we could detect no difference between the two groups with respect to the slopes for the scores with use of the Wilcoxon rank-sum test (p = 0.22) or the t test (p = 0.15). In addition, analysis of covariance demonstrated no significant difference between the slopes for the mean scores in the two groups (p = 0.92). However, the group that was managed with continuous passive motion had significantly less pain during the first postoperative week than the group that was managed with manual passive range-of-motion exercises (p = 0.046, Wilcoxon rank-sum test) (Fig. 2).
Range of Motion
Predicted time-courses for the change in the range of motion were derived with use of analysis of covariance, which controls for the initial (baseline) measurement. The time-points thereafter were the actual time-points for the respective patients in the group, and a line was drawn that best fit the data. The line is interpolated and is not an extrapolation of the actual data, and it allows for interpretation of the progress over time. With the numbers available, we could detect no significant difference between the two groups with respect to the baseline measurements of external rotation and elevation (p > 0.20, t test) (Fig. 3). We also could detect no difference between the passive external rotation or elevation in the two groups at twelve or twenty-four months, after adjusting for the baseline passive range of motion at two weeks (p > 0.15, analysis of covariance). Similarly, we found no differences in the active range of motion in the two groups at twelve and twenty-four months, after adjusting for the baseline values at twelve weeks (p > 0.20, analysis of covariance).
With use of the logarithmically transformed slopes for the range of motion, the rates at which the two groups regained passive and active external rotation and elevation were compared. With the numbers available, we could detect no differences between the two rates when they were compared parametrically with use of the t test and non-parametrically with use of the Wilcoxon rank-sum test (p > 0.20). Adjusting for the baseline range of motion yielded the same results (p > 0.15, analysis of covariance). The median rate of recovery (expressed as degrees per month) was similar for the two groups at three, six, and twelve months (Fig. 4). The similarities in the medians at each time-point demonstrated that the patients in both groups regained motion at the same rate. The decreasing pattern also demonstrated that the recovery of motion leveled off with time in both groups. The power analysis denoted that, given the variability in the data, a 115 per cent increase compared with the mean rate of recovery of active external rotation in the group managed with manual passive range-of-motion exercises could have been detected with 80 per cent power. Similarly, a 124 per cent increase compared with the mean rate of recovery of active elevation in that group could have been detected with 80 per cent power.
Strength
The isometric strength was measured with the handheld dynamometer during external rotation and elevation in scaption16. We compared the two groups with respect to the values at six months, the values at twelve months, and the difference between the six and twelve-month values. The comparisons were made after adjusting for the baseline strength at four months. With the numbers available for study, we could detect no significant difference between the two groups with respect to the strength in external rotation (p = 0.20, analysis of covariance). An 18 per cent increase compared with the mean response in the group managed with manual passive range-of-motion exercises could have been detected with 80 per cent power. There was a marginal difference between the two groups with respect to the strength in elevation at six months (p = 0.06, analysis of covariance), but we detected no difference in the values for the two groups at twelve months or between the differences between the values at six and twelve months. A plot of the adjusted mean values for each group showed that strength continued to improve for one year (Fig. 5).
Age, Gender, and Size of Tear
The data were further subdivided and analyzed for the potential effect of age. With the numbers available, we could detect no significant differences (p = 0.07 to 0.89), across all dependent variables, between the two groups when controlling for age (less than sixty-five years old [fifteen patients] and more than sixty-five years old [seventeen patients]). We also could detect no significant difference (p = 0.08 to 0.97) between the two groups when controlling for gender, with one exception: women indicated significantly less pain (p = 0.029) on the visual-analog scale at one week than men did. We could detect no significant differences (p = 0.07 to 0.92) between the two groups when controlling for the size of the tear (small or medium and large).
Outpatient Visits for Postoperative Therapy
During the first three months postoperatively, the group that was managed with continuous passive motion had a median of nine outpatient visits for physical therapy and the group that was managed with manual passive range-of-motion exercises had a median of ten visits. During the visits, the therapist measured the range of motion and the strength and advanced the postoperative program as dictated by the protocol or modified the program depending on the clinical progress of the patient. Three of the visits during the three-month period were mandatory to allow for the collection of data and to continue instruction in the exercise program. Any visits other than the three mandatory ones were considered clinically necessary by the therapist to monitor the patient more closely, to adjust the program, or to reinforce the patient's home-exercise program. By the third or fourth postoperative month, all patients were using the arm on the involved side for functional activities and were performing progressive strengthening exercises at home. Any outpatient visits after four months were for follow-up and for modification of the exercise or strengthening program. No patient was seen regularly on an outpatient basis for formal therapy after four months.
Complications
One patient who was managed with continuous passive motion had an infection four months postoperatively; the clinical result was excellent after débridement and repeat repair of the cuff. One patient who was managed with manual passive range-of-motion exercises had reflex sympathetic dystrophy, and one had occult glenohumeral instability; the latter patient had a fair result at the time of the latest follow-up.
To our knowledge, this is the first prospective, randomized study to compare the functional outcome after continuous passive motion with that after a program of manual passive range-of-motion exercises for patients who had had repair of the rotator cuff. Our primary finding is that continuous passive motion is a safe technique that results in little disability and an excellent or good outcome after a repair of a small, medium, or large tear of the rotator cuff. Continuous passive motion does not, however, provide a better outcome than a program of manual passive range-of-motion exercises, which is more cost-effective.
It appears, from the results of the present study, that forty-five minutes of manual passive range-of-motion exercises a day, performed outside of the clinical setting, is adequate for a good functional outcome after repair of a small, medium, or large tear of the rotator cuff. The two groups had very similar improvements in the range of motion and strength. For example, the improvements in the active and passive ranges of motion were greater at earlier time-points (three and six months) than at later time-points (twelve months) in both groups; however, we could detect no significant differences between the groups with the numbers available. We adjusted the data for the first (baseline) measurement (at two weeks for the passive range of motion and at twelve weeks for the active range of motion) as a covariate to eliminate the effect of the amount of motion that the patient had before the operation, as this can vary tremendously and may confound the results. We did consider the influence, either beneficial or detrimental, that passive range-of-motion exercises may have on active range of motion and strength, but, again, we could detect no significant differences with respect to any measurement (with the possible exception of strength in elevation at six months, which approached significance [p = 0.06]) between the two groups at any time postoperatively. The group that was managed with continuous passive motion had significantly less pain during the first postoperative week (p = 0.046), but we detected no differences thereafter.
We were concerned that the between-examiner reliability was poor for the measurement of the active range of motion and of strength in external rotation. In an attempt to overcome the influence of any problems with between-examiner reliability, we made an effort to have the same therapist perform all of the measurements for a given patient. (The within-examiner reliability was very good.) However, this could not be strictly controlled. Given the power to detect only very large differences between groups with regard to active external rotation and elevation, the non-significant results in the present study need to be verified. However, there was good power to detect a fairly small difference in the strength in external rotation in the two groups, which supports the conclusion of the study. The measurement of pain and disability (the Shoulder Pain and Disability Index23) had good reliability.
More importantly, there was little functional limitation or disability in either group, regardless of age, gender, or size of the tear. The three complications subsequently resolved or had little effect on function. The one poor result, in a patient who was managed with manual passive range-of-motion exercises, was thought be due to limited function of the deltoid both before and after the operation.
The scores according to the Shoulder Pain and Disability Index were equally good for the two groups. We chose this scale because it is a unique shoulder-specific index that has demonstrated consistency, test-retest reliability, and criterion and construct validity, primarily in a population of older men22,23. Therefore, it might be safer to generalize the results of this study to male patients who are beyond the fourth or fifth decade of life than to female patients.
As both groups had good results, we were interested in determining which postoperative protocol was the most cost-effective. After repair of the rotator cuff, the greatest investment in terms of both time and dollars is during the first three months. Although there was only a small difference between the two groups with respect to the number of outpatient visits, the cost to the third-party payer for continuous-passive-motion therapy would have been substantially higher if the patients had been charged for the device. This additional cost may not be warranted after repair of a small, medium, or large tear in a patient who is forty years old or more and has a relative, friend, or home-care nurse who can be trained to perform manual passive range-of-motion exercises.
Before initiating the study, we arbitrarily chose four weeks as the duration of continuous passive motion, as the first four postoperative weeks is the period when passive motion is the focus of therapy and active motion is encouraged. (Passive range-of-motion exercises are continued after the four-week period on an as-needed basis.) We believe that, although the repair should be protected throughout this period, passive motion is needed to prevent the formation of restrictive adhesions, which can occur during the biologically active fibroblastic proliferation phase of wound-healing13. We were surprised to find that the passive range of motion was better (although not significantly so) in the group that was managed with manual passive range-of-motion exercises, even though the exercises with the continuous-passive-motion device were performed for a longer period each day than the manual exercises were. Another reason that we chose a duration of four weeks for continuous passive motion was to try to facilitate greater tendinous force capabilities, which have been noted in healing tendons that are subjected to prolonged periods of motion17,28. The long-term use of continuous passive motion may have contributed to a trend (although not a significant one) toward improvements in active elevation and strength in the group treated with that method.
When no one is available to perform manual passive range-of-motion exercises for the patient, it is possible that other types of passive range-of-motion exercises, performed at the same frequency and for the same duration as were used in the present study, could yield similar results. Performance of passive external-rotation and elevation exercises with use of a cane or with use of the uninvolved extremity in a gravity-minimized plane has been shown to result in little activity of the shoulder muscles. Any exercises that require the arm to move against gravity, such as those done with pulleys, will result in more muscle activity14. Despite the anecdotal evidence that exercises with pulleys are safe and effective, the question still remains as to how much muscle activity facilitates a successful outcome. When passive range-of-motion exercises cannot be performed in an appropriate manner, more frequent visits for physical therapy or the use of a continuous-passive-motion device, or both, may be warranted. Although the efficacy of continuous passive motion has not been proved, we would seriously consider its use, with an arc of motion that protects the repair, when the passive range-of-motion exercises must be performed for a long duration. This is the case when a patient has a preexisting frozen shoulder or when restrictive adhesions are likely to form, such as after a repair of a massive tear or after a repeat repair. We excluded patients who had a massive tear or adhesive capsulitis, or both, in order to better represent the typical patient who has a non-complicated tear of the rotator cuff. Future studies are needed to assess the effectiveness of continuous passive motion for a patient who has a more complex tear.
It is difficult to compare the findings of the present study with those of previous reports, as we know of no other prospective analyses of postoperative protocols after repair of the rotator cuff. Still, we believe that the results in both of our groups compare favorably with those in other outcome studies7,10,11. In a review in which the results were presented as the average of the percentages derived from several series, Cofield reported an 85 per cent rate of excellent or good results; 87 per cent of the patients had relief of pain, and 77 per cent were satisfied with the outcome. We did not measure our patients' satisfaction with the results, but the scores according to the Shoulder Pain and Disability Index23 suggest that function was more than adequate for the patients' needs.
Other investigators1,9 have found that the outcome after repair of digital tendons was better when continuous passive motion had been used. The quicker healing and greater strength of repaired tendons and ligaments associated with continuous passive motion has been attributed to increased blood flow and metabolic activities of cells8,26. The healing capability of the rotator cuff, specifically that of the supraspinatus, may be enhanced by continuous passive motion, as the supraspinatus tendon has a tenuous blood supply21. However, this potential for improved vascular and cellular activity resulted in no significant clinical advantage in our study. Craig conducted a preliminary study comparing immediate continuous passive motion for two to three days with immobilization for three to four days after operative reconstruction of the shoulder and found no adverse effects, a shorter hospital stay, and less postoperative pain for the patients who were managed with continuous passive motion. It is difficult to compare the results of our study with those of Craig's investigation. We compared two therapeutic modes of early passive range-of-motion exercises in a well defined group of patients who had had repair of the rotator cuff. In contrast, Craig compared continuous passive motion with immobilization in a series of patients who had had repair of the rotator cuff, acromioplasty, excision of the coracoacromial ligament, or total shoulder replacement. Even though Craig found a beneficial response after forty-eight to seventy-two hours, we chose a four-week period of continuous passive motion, as we sought to influence the remodeling phase of the wound-healing process. In both the study by Craig and the present study, continuous passive motion was more effective in diminishing pain during the first postoperative week and, empirically (in our study), it seemed to help minimize involuntary co-contractions during the exercises. This effect during the first week may be clinically relevant in terms of decreased use of narcotics or increased independence, or both, during this time, but those data were not collected. In our study, however, continuous passive motion had no significant influence on the range of motion, strength, or disability. We would, however, consider its use if pain and co-contractions interfered with the rehabilitation process. Reyes et al., who reported retrospectively the results of continuous passive motion after repair of the rotator cuff, found no important differences in the duration of the hospital stay, the intake of pain medication, and the days on which range-of-motion landmarks were achieved between patients who had and those who had not been managed with continuous passive motion. The patients who were managed with continuous passive motion in that study had fewer visits for physical therapy, but the total cost for the therapy was greater21.
It seems unlikely that the functional outcome after a repair of the rotator cuff could be much better than it was in the present study. However, future studies that document the results of more visits for therapy or the use of continuous passive motion by patients who have had a repeat repair or a repair of a massive tear may be worthwhile. Conversely, the assessment of the outcomes of repairs of the rotator cuff in patients who had even fewer visits for outpatient therapy than the patients in the present study may reveal an even more cost-effective postoperative approach.
Although no significant differences in the clinical outcome were noted between the two groups in our study, as stated we would still consider the use of continuous passive motion when restrictive adhesions are a concern, as when a patient has had a repeat repair of the rotator cuff, concomitant adhesive capsulitis, or a repair of a massive tear. Continuous passive motion also may be beneficial for a patient who lives alone and does not have a relative, friend, or home-care nurse who can be trained to perform manual passive range-of-motion exercises. In addition, continuous passive motion may be helpful for patients who have severe postoperative pain or problems with muscular co-contractions during passive range-of-motion exercises.
Limitations of the Study
It is difficult to assess the effects of two different interventions during the first four weeks after repair of the rotator cuff and then to analyze the effects of those treatments at twenty months. Many variables can influence the outcome, such as the consistency of range-of-motion and strengthening exercises performed at home after formal therapeutic visits have been discontinued. There was no formal documentation of the compliance of the patients or of modifications to the program after the first four weeks. These variables may have confounded our results.
In conclusion, continuous passive motion and manual passive range-of-motion exercises contributed positively to the range of motion, strength, function, and relief of pain; however, we could detect no significant difference in the clinical outcome between the two groups.
Bunker, T. D.; Potter, B.; and Barton, N. J.: Continuous passive motion following flexor tendon repair. J. Hand Surg.,14-B: 406-411, 1989.14-B406
1989
Cofield, R. H.: Current concepts review. Rotator cuff disease of the shoulder. J. Bone and Joint Surg.,67-A: 974-979, July 1985.67-A974
1985
Coutts, R. D.; Toth, C.; and Kaita, J. H.: The role of continuous passive motion in the rehabilitation of the total knee patient. In Total Knee Arthroplasty: A Comprehensive Approach, pp. 126-132. Edited by D. S. Hungerford, K. A. Krackow, and R. V. Kenna. Baltimore, Williams and Wilkins, 1984.
Coutts, R. D.; Craig, E. V.; Mooney, V.; Osterman, A. L.; and Salter, R. B.: Symposium: the use of continuous passive motion in the rehabilitation of orthopaedic problems. Contemp. Orthop.,16: 75-106, 1988.1675
1988
Craig, E. V.: Continuous passive motion in the rehabilitation of the surgically reconstructed shoulder. A preliminary report. Orthop. Trans.,10: 219, 1986.10219
1986
Fleiss, J. L.: The Design and Analysis of Clinical Experiments. New York, Wiley, 1986.
Gazielly, D. F.; Gleyze, P.; and Montagnon, C.: Functional and anatomical results after rotator cuff repair. Clin. Orthop.,304: 43-53, 1994.30443
1994
[PubMed]
Gelberman, R. H.; Amiel, D.; Gonsalves, M.; Woo, S.; and Akeson, W. H.: The influence of protected passive mobilization on the healing of flexor tendons: a biochemical and microangiographic study. Hand,13: 120-128, 1981.13120
1981
[PubMed]
Gelberman, R. H.; Nunley, J. A., II; Osterman, A. L.; Breen, T. F.; Dimick, M. P.; and Woo, S. L-Y.: Influences of the protected passive mobilization interval on flexor tendon healing. A prospective randomized clinical study. Clin. Orthop.,264: 189-196, 1991.264189
1991
[PubMed]
Grana, W. A.; Teague, B.; King, M.; and Reeves, R. B.: An analysis of rotator cuff repair. Am. J. Sports Med.,22: 585-588, 1994.22585
1994
[PubMed]
Harryman, D. T., II; Mack, L. A.; Wang, K. Y.; Jackins, S. E.; Richardson, M. L.; and Matsen, F. A., III: Repairs of the rotator cuff. Correlation of functional results with integrity of the cuff. J. Bone and Joint Surg.,73-A: 982-989, Aug. 1991.73-A982
1991
LaStayo, P. C.: Continuous passive motion for the upper extremity. In Rehabilitation of the Hand: Surgery and Therapy, edited by J. M. Hunter, E. J. Mackin, and A. D. Callahan. Ed. 4, vol. 2, pp. 1545-1560. St. Louis, C. V. Mosby, 1995.
Liu, S. L.; Sen Yang, R.-S.; Al-Shaikh, R.; and Lane, J. M.: Collagen in tendon, ligament, and bone healing. A current review. Clin. Orthop.,318: 265-278, 1995.318265
1995
[PubMed]
McCann, P. D.; Wootten, M. E.; Kadaba, M. P.; and Bigliani, L. U.: A kinematic and electromyographic study of shoulder rehabilitation exercises. Clin. Orthop.,288: 179-188, 1993.288179
1993
[PubMed]
Marks, P. H.; Warner, J. J.; and Irrgang, J. J.: Rotator cuff disorders of the shoulder. J. Hand Ther.,7: 90-98, 1994.790
1994
[PubMed]
Matsen, F. A., III, and Arntz, C. T.: Rotator cuff tendon failure. In The Shoulder, edited by C. A. Rockwood, Jr., and F. A. Matsen, III. Vol. 2, pp. 647-677. Philadelphia, W. B. Saunders, 1990.
Michna, H., and Hartmann, G.: Adaptation of tendon collagen to exercise. Internat. Orthop.,13: 161-165, 1989.13161
1989
Gristina, A. G.; Craig, E. V.; Neviaser, R. J.; and Norris, T. R.: Symposium: management of rotator cuff problems. Contemp. Orthop.,20: 621-646, 1990.20621
1990
Price, D. D., and Harkins, S. W.: Combined use of experimental pain and visual analogue scales in producing standardized measurement of clinical pain. Clin. J. Pain,3: 1-8, 1987.31
1987
Rathbun, J. B., and Macnab, I.: The microvascular pattern of the rotator cuff. J. Bone and Joint Surg.,52-B(3): 540-553, 1970.52-B(3)540
1970
Reyes, A. M.; Pati, A. B.; and Gartsman, G. M.: Effects of shoulder CPM vs no CPM on rotator cuff repair rehabilitation [abstract]. Phys. Ther.,72 (Supplement): 99, 1992.72 (Supplement)99
1992
Richards, R. R., and Beaton, D. E.: Measuring shoulder function: a cross-sectional comparison of five different questionnaires [abstract]. J. Shoulder and Elbow Surg.,4-1(Part 2): 61, 1995.4-1(Part 2)61
1995
Roach, K. E.; Budiman-Mak, E.; Songsiridej, N.; and Lertratanakul, Y.: Development of a Shoulder Pain and Disability Index. Arthrit. Care and Res.,4: 143-149, 1991.4143
1991
Rosner, B.: Fundamentals of Biostatistics. Ed. 3. Boston, PWS-Kent, 1990.
Salter, R. B.: Continuous Passive Motion (CPM): A Biological Concept for the Healing and Regeneration of Articular Cartilage, Ligaments and Tendons: From Its Origination to Research to Clinical Applications. Baltimore, Williams and Wilkins, 1993.
Salter, R. B.; Hamilton, H. W.; Wedge, J. H.; Tile, M.; Torode, I. P.; O'Driscoll, S. W.; Murnaghan, J. J.; and Saringer, J. H.: Clinical application of basic research on continuous passive motion for disorders and injuries of synovial joints: a preliminary report of a feasibility study. J. Orthop. Res.,1: 325-342, 1984.1325
1984
[PubMed]
Shrout, P. E., and Fleiss, J. L.: Intraclass correlations: uses in assessing rater reliability. Psychol. Bull.,86: 420-428, 1979.86420
1979
[PubMed]
Woo, S. L.; Gomez, M. A.; Woo, Y. K.; and Akeson, W. H.: Mechanical properties of tendons and ligaments. II. The relationships of immobilization and exercise on tissue remodeling. Biorheology,19: 397-408, 1982.19397
1982
[PubMed]