EVALUATION OF THE RESULTS FROM REOPERATIONS ON PATIENTS WITH ROTATOR CUFF LESIONS

Poor clinical results have been reported in up to 25% of the cases repaired(5-7), and the following factors may be associated with such occurrences: presence of large and extensive lesions, quality of the tendon to be sutured, fatty degeneration of the muscle, surgical technique used, surgical damage caused to the origin of the deltoid muscle and inadequate postoperative rehabilitation(1,8-10).

There is little information in the literature on asessments of reoperated cases and their long-term evolution. These are situations that are difficult to resolve, and the results are generally inferior to the primary repair, given that although pain relief may be achieved, improvement of limb function is less likely(1,5,9,11). DeOrio and Cofield(12) reported that 58% of the results from attempted second repair of lesions using an open approach were poor, with little pain relief or improvement of mobility. Neviaser and Neviaser(5) reported an improvement of pain in 92% of the cases reoperated via an open approach, with an increase in mean elevation from 92° to 137°. Ma et al(6) found that 55% of the results from second repair using open surgery were satisfactory; Lo and Burkhart(11) obtained four excellent and five good results (64% satisfactory) out of 14 patients who underwent a new surgical procedure using an arthroscopic route.

The aim of this study was to clinically and functionally assess patients with recurrence of rotator cuff lesions who underwent a new surgical procedure by means of either an open or an arthroscopic approach.

SAMPLE AND METHODS

Between December 1990 and July 2007, the Shoulder and Elbow Surgery Group of the Department of Orthopedics and Traumatology of Santa Casa de São Paulo, Fernandinho Simonsen Wing, surgically treated 30 patients (30 shoulders) who presented symptomatic recurrence of rotator cuff lesions.

The inclusion criteria were that these were cases that were operated using either an open or an arthroscopic approach because of recurrence of symptomatic rotator cuff lesions, with a minimum postoperative follow-up of 24 months after the second surgery. To make the diagnoses, anamnesis, physical examination and complementary tests were used to demonstrate the lesion and other abnormalities that were possibly associated, including the use of AP, axillary and lateral radiographs of the scapula and magnetic resonance imaging (MRI) (Figure 1). Patients who did not fit within the criteria established above were excluded.

Out of the 30 patients who underwent a new surgical procedure, 18 were male (60%) and 12 were female (40%). The patients' mean age was 58 years, with a range from 33 to 76 years. The dominant limb was affected in 25 cases (83.3%) (Table 1). In 17 cases (56.7%), the first procedure had been done in our service, while 13 (43.3%) underwent the initial treatment in another place. In 20 patients (66.7%), the access route in the first surgery was open, while it was arthroscopic in 10 (33.3%). With regard to the initial size of the lesion among the patients operated in our service, according to the classification of Hawkins et al(13), five patients (29.4%) had extensive lesions, 10 had large lesions (58.8%) and two had medium-sized lesions (11.8%). There were no cases of small lesions in this study. In relation to the 13 patients (43.3%) operated in other services, we did not have any information regarding the size of the initial lesion.

The mean time taken for the symptoms to restart was 21 months, with a range from zero to 228 months. Five patients who had undergone their first operation at

Figure 1 - Example of radiograph (A) and magnetic resonance image (B) of the right shoulder, showing tear in the tendon of the spinal supraspinatus muscle.

another service were unable to give precise information regarding the time when their symptoms returned. Histories of trauma were associated with symptom recurrence in 10 patients (33.3%) and of these, seven (70%) reported falls to the ground and three (30%) reported having made unusually large physical effort. The time interval between the first surgery and the revision was on average 27 months, with a range from one to 230 months (Table 1).

An open surgical access route was used for 11 patients (36.6%), while arthroscopic surgery was used for 19 patients (63.4%) (Table 1). All the patients underwent their operations in the "deckchair" position, under general anesthesia in association with anesthetic block of the brachial plexus.

In the cases that were operated using an open route, we used an anterior deltopectoral access. In 10 of these cases (90.9%), the lesion was repaired using transosseous stitches, while one case underwent fixation using anchors (9.1%). In the cases treated arthroscopically, anterior, posterior and lateral portals were constructed, with accessory portals when necessary for repairing the lesion using anchors (Table 1).

With regard to the sizes of the lesions at the time

of the reoperation, according to the classification of Hawkins et al(13), we found nine extensive lesions (30%), six large lesions (20%), 12 medium-sized lesions (40%) and three small lesions (10%). Associated procedures were performed in 16 cases (53.3%): acromioplasty in 14 cases; tenotomy on the tendon of the long head of the brachial biceps muscle in three cases, among which tenodesis was performed in one case; resection of the lateral extremity of the clavicle in six cases; and revision of this procedure in one case. The mean number of anchors installed, when used, was 2.2, with a range from one to five (Table 1).

During the postoperative period, the patients were immobilized using slings for six weeks, with release for free active exercises for the peripheral joints. After four weeks, passive lateral rotation was started (with a physiotherapist) and pendular movements without load-bearing. During the postoperative follow-up, the patients were reassessed using the criteria proposed by the University of California at Los Angeles (UCLA)(14). The joint range of motion was measured in accordance with the criteria of the American Academy of Orthopaedic Surgeons (AAOS)(15).

For the statistical analysis, we applied the Mann- Whitney test with the aim of investigating possible differences between the categories of the variables of sex, side affected, trauma, type of surgery (open or arthroscopic), treatment for lesions of the long head of the brachial biceps muscle and lesion size. We used the SPSS software (Statistical Package for Social Sciences), version 17.0, to obtain the results. We took the significance level to be 5% (0.050) when applying the statistical tests.

RESULTS

The mean postoperative follow-up was 61 months, with a range from 24 to 152 months. Among the 30 patients who underwent a new surgical procedure, we found that the mean UCLA score was 28, ranging from 12 to 35. The results were considered to excellent in 36.7% of the cases; good in 33.3%; fair in 13.3%; and poor in 16.7% (Table 1).

The mean range of motion in the postoperative assessment was 133° for elevation, ranging from 60° to 150°; 51° for external rotation, ranging from 30° to 90°; and T12° for internal rotation, ranging from the gluteus to T7.

There were no statistically significant correlations for age, sex, presence of trauma in the first or second surgery and length of time with symptoms, in relation to the results and treatment for the brachial biceps (p > 0.050). On the other hand, the size of the lesion, both in the initial surgery and in the second procedure, showed a statistically significant relationship (p = 0.049) with the results, given that among the nine cases with unsatisfactory results, the lesions were predominantly extensive or large, both in the first and in the second surgery (Tables 1 and 2).

Our study also showed that the results were predominantly unsatisfactory in the cases in which the access route was open. This was statistically significant in comparison with the operations performed using the open and arthroscopic routes (p = 0.001) (Table 2). Out of these nine unsatisfactory cases, seven (77.8%) underwent treatment via the open route in the second operation.

Complications were observed in five cases (16.7%): one case that evolved with adhesive capsulitis and four with recurrence of symptoms (Table 1)

DISCUSSION

Before thinking of repairing recurrences of rotator cuff lesions, it needs to be defined what constitutes a failure. Imaging examinations alone are not criteria for indicating a second surgical procedure(2,4,16), since the lesion is not always compatible with the patient's functioning and complaints(17). Using magnetic resonance examinations on asymptomatic volunteers, Sher et al(18) showed that rotator cuff lesions could be found in individuals with good shoulder functioning and absence of symptoms. Thus, an anamnesis and complete physical examination are of paramount importance for diagnosing recurrences of these lesions.

The great majority of authors have shown in their studies that pain relief is achieved through the reoperation, although functional improvement is less likely to occur. Nonetheless, pain relief is the primary aim of the reoperation(1,5,6,9,11). In our study, 86.7% of the patients reported that they had achieved improvement of their pain, and 76.7% said that they had recovered a functional level similar to that of their contralateral limb, or only presented small limitations in the affected shoulder.

A variety of factors have been cited as causes of failure of the initial repair, thereby directly or indirectly compromising the functional result. These could also compromise the result from a second procedure if they continue to be present(2,6,9,19). Among these factors, the following can be cited:

1) Inadequate subacromial decompression, which is one of the main causes of failure after the initial repair(2). Among our sample, at the time of the second procedure, 14 patients (46.7%) also underwent acromioplasty because they presented signs of impact;

2) Size of the lesion in the first repair: some studies have cited this as the most common cause of failure, with rates ranging from 67% to 97% in some series(5,12). Studies have shown that patients who have small or medium-sized lesions at the time of the first procedure tend to evolve with better results from the reoperation than do those who initially presented large or extensive lesions(6,9,19). In our sample, out of the nine patients with unsatisfactory results, six presented large or extensive lesions in the first procedure, while it was not possible to identify the size of the lesion in two cases, because they were operated at other services (Table 2).

3) Compromising of the deltoid muscle, which occurs in cases in which lateral or complete acromionectomy is performed, thereby modifying the lever arm of the deltoid muscle and resulting in difficulty in achieving elevation(2). In assessing our unsatisfactory results, we saw that 66.7% underwent the first and second procedures via an open route. Two procedures using an open approach may compromise a greater number of fibers of the deltoid muscle, thus impairing the result;

4) Quality of the suture performed: in cases with loss of function in the presence of an adequate deltoid muscle, this leads towards envisaging that repair failure may have occurred(2). Factors such as inadequate mobilization of the cuff, freeing of adherences, fatty degeneration of the tendon and inadequate fixation of the tendon to the bone or under tension are possible causes that might compromise the result from the repair, thus leading to recurrence of the lesion(2,6,9,19). Three of our patients evolved with new tears after the reoperation (Table 1);

5) Inadequate rehabilitation: postoperative rehabilitation may compromise the procedure that has been carried out if it is not done correctly(2). Delays in starting passive and active mobilization of the limb may progress to stiffness, thereby impairing the result, and this may also occur if mobilization and strength gain are started too early(2). In our study, we did not identify any case in which the rehabilitation could be indicated as the cause of impaired results;

6) State of the long head of the brachial biceps muscle: it has been suggested in some studies that a non- -functional biceps or a lesion in the biceps might contribute towards failure in the initial surgery because of the depressive effect of the humeral head, especially in large and extensive lesions. Thus, the use of tenotomy or tenodesis on these types of lesions is questionable( 5,6). In our sample, we did not find any relationship with unsatisfactory results, among the patients who underwent tenotomy or tenodesis of the long head of the biceps (p = 0.05).

With regard to lesion size, the literature(16,20) has demonstrated that recurrent lesions are smaller than the lesions treated initially. This was also seen in our study, given that out of the 17 patients whose first and second operations were both done in our service, eight cases presented smaller lesions than the initial ones, five cases had lesions of the same size and only four cases had lesions that were bigger than the primary lesions.

Regarding the route used in the second operation, the literature provides little information about cases treated arthroscopically(11). Most of the published papers have reported that the open route was used(1,2,5,6,9,12,19). Lo and Burkhart et al(11) cited the following as advantages of performing the second procedure arthroscopically: a) it allows complete evaluation of the glenohumeral joint and the acromial space; b) there is minimal aggression to the deltoid muscle; c) it allows better viewing of the rotator cuff and the advantage of better mobilization of the lesion and viewing the tension in the repair and the edges of the repair, especially in lesions of U or L shape; d) lastly, arthroscopic surgery presents lower incidence of postoperative stiffness. In that study, the authors concluded that arthroscopic revision was a technically more difficult procedure, but that it might lead to improvement not only of pain but also of limb function. We agree with their conclusion: we had better results from arthroscopic surgery than from open repair of lesions (p = 0.001) (Figure 2).

CONCLUSION

Surgical treatment via open and arthroscopic routes to treat renewed tearing of the rotator cuff tends to present worse results than in the first operation. In the present study, we found that 70% of the results were excellent and good. The presence of extensive lesions in the reoperation tends to evolve towards a greater number of unsatisfactory results (p = 0.049). In our study, we obtained better results from surgery performed arthroscopically than from open surgery (p = 0.001).

Figure 2 - Same case as in Figure 1, showing a lateral view of the lesion in the tendon of the spinal supraspinatus muscle (A) and after suturing (B).

semitendinosus and gracilis tendons(2).

The items that contribute towards satisfactory evolution of the procedure are: appropriate choice of surgical technique for each patient; the condition of the secondary restrictors (meniscus and ligaments), the postoperative analgesia used and safe early rehabilitation. Improvements and innovations in ACL reconstruction techniques have made it possible to achieve satisfactory results with regard to control over instability and early return to sports(3). Among the most frequent complications from ACL reconstruction, pain on the anterior face of the knee and residual muscle strength deficit seem to be connected with the choice of donor source(4-6).

The aim of this study was to analyze the impact that residual pain has on functional results, two years after arthroscopic ACL reconstruction, comparing the use of grafts from the central third of the patellar ligament with the use of grafts from the flexor tendons of the semitendinosus and gracilis muscles.

METHODS

In this series, 129 patients who underwent ACL ligament reconstruction were evaluated descriptively and retrospectively. Sixty-eight patients were benefited through using the patellar tendon as the graft (patellar group) and 61 received grafts from the flexor tendons of the semitendinosus and gracilis muscles (flexor group). The patients were evaluated after an average follow-up of 28 months. The series was composed of 81 men and 48 women, and their average age was 27.8 years.

All the patients were operated by the same surgeon, Dr. David Dejour, and were reassessed by the same examiner, Dr. Wilson Vasconcelos. The choice of donor source for the graft was made according to the sport practiced by the patient (Table 1). The inclusion criteria were that these were patients operated in the years 2005 and 2006, aged between 16 and 45 years, who did not present any associated ligament injury and did not have any injury in the contralateral knee. Patients were excluded if they had been followed up for less than 18 months or if they presented an advanced chondral lesion diagnosed at the time of the surgery. Associated meniscal movement was considered to be an exclusion criterion.

During the reassessment, the following data were

gathered: anamnesis; surgical characteristics (status of the joint cartilage, associated movements and complications); subjective information on the pain (presence and intensity on a numerical scale from 0 to 10 and topographic location on a tracing) (Figure 1); objective information on the pain (walking on knees) (Figure 2); information on touch sensitivity (topographic location and intensity (Figure 1) and scores (subjective IKFC, subjective IKDC, femoropatellar LILOAS and the SF-36 quality-of-life score).

The data gathered were collated using Excel (Microsoft) and were analyzed with the aid of the Statview ® software. The statistical significance level was taken to be 5%. Qualitative variables were assessed according to their frequency distribution and quantitative variables were assessed according to their means, medians and standard deviations.

Figure 2 - Knee walking test.

RESULTS

Overall in this series, 26% of the patients reported the presence of anterior pain; 45%, abnormalities regarding walking on the knees; and 47%, sensitivity alterations. The mean pain intensity gauged by the patients was 2.9 out of 10. The mean scores were: 90.4 out of 100 on the femoropatellar scale; 84/100 on the subjective IKDC; and 84/100 on the overall SF-36, while the physical SF-36 was 83/100 and the mental SF-36 was 80/100. For the patients with complains of residual pain, the femoropatellar score was 84, versus 93 for those without complaints of pain (p < 0.005). Likewise, the IKDC went from 84 to 78 (p < 0.05) in cases with pain and the physical SF-36 went from 83 to 78 (p = 0.01).

The frequency of anterior pain for the patellar tendon group was 34%, while it was 27% for the flexor group (p = 0.03). On the other hand, the pain intensity classified by the patients was 2.1/10 in the patellar group, versus 3.3/10 for the flexor group (p = 0.004).

The most common location for pain in the patellar group was on the patellar tendon (37%), whereas for the flexor group, the most common location was close to the tibial tunnel (51%), followed by on the patellar tendon (26%). The mean femoropatellar score was 90.5% for the patellar group, versus 91.5% for the flexor group. The mean score for the subjective IKDC was 84.1 for the patellar group and 83 for the flexor group. The mean SF-36 score was 83/100 for the patellar group versus 80/100 for the flexor group. There were no statistically significant differences in these three scores.

Walking on the knees produced better results in the flexor group, given that 68% of these patients presented normal results in the test, versus only 35% in the patellar group (p = 0.01). This task was considered to be impossible by 10% of the patients in the patellar group, versus none of the patients in the flexor group (Figure 3).

With regard to sensitivity problems, hypoesthesia was found to be present in 68% of the patients in the patellar group, versus 32% in the flexor group (p < 0.001). Regarding topographical location, there was a difference: the patellar group presented a deficit at the side of the scar, while the flexor group showed abnormalities in the middle lower part of the scar (Figure 4).

Among the patients who complained about sensitivity abnormalities, 40% presented pain. On the other hand, among the patients without such complaints, only 18% presented pain (Figure 5). Out of the 26% with pain, 17% presented an association with sensitivity problems. Abnormalities with regard to walking on the knees were only present in 28% of the patients without complaints of pain; they were present in 48% of the patients with pain alone; they were present in 56% of the patients with hypoesthesia alone; and they were present in 82% of the patients with both pain and hypoesthesia.

Figure 4 - Topographic location of hypoesthesia.

DISCUSSION

Analysis on the frequency of anterior pain leads to a variety of interpretations, in view of the subjective nature of such pain. In most studies, these data are gathered according to the authors' analyses, thereby giving rise to considerable differences. In our series, in order to avoid the observer's influence regarding the frequency and intensity of the pain, the criterion adopted was a positive response from the patient to the question: "Do you have any pain in the front of your knee?"

Thus, even pains of lesser intensity were recorded, which increased the sensitivity of the study to this question. In a general manner, and without taking into consideration the technique used, when pain was present, it was at a low level. In our series, the frequency of pain was greater in the patellar group than in the flexor group, but this difference was not statistically significant (p = 0.03). The primary cause of this pain is believed to be correlated with graft harvesting, given the topographical locations of the pains, which were more anterior for the patellar group and more medial for the flexor group. A large portion of the studies have reported that there was no difference in the frequency of anterior pain, in comparisons between the two groups(7-12).

Between our two groups, there was a statistically significant difference in relation to walking on the knees. The flexor group showed better performance, which should be taken into consideration at the time of choosing the graft to be used in the reconstruction. For patients who, during their professional or sports activities support themselves with their knees on the ground, the flexor group will be a better choice. The literature confirms that techniques using the flexors are superior in this respect(7,9,10).

With regard to sensitivity abnormalities, reconstruction using the flexors is significantly less iatrogenic than with the use of the patellar tendon (0.001). There was a close correlation with the presence of pain and hypoesthesia. It was observed that, among the patients without sensitivity problems, only 18% presented pain. On the other hand, among the patients with hypoesthesia, 40% presented pain, which denotes imprinting of hypoesthesia over the residual pain. It is known that sensitivity problems originate from lesions of the infrapatellar branches of the medial saphenous nerve(13,14), and thus, minimizing the aggression during graft harvesting will benefit such patients.

In our assessment of the two groups, one with reconstruction using the central third of the patellar ligament and the other with flexor tendons from the semitendinosus and gracilis muscles, with a minimum of two years of follow-up, there was no difference in terms of patient satisfaction, from the SF-36 questionnaire, subjective IKDC evaluation, femoropatellar score and pain analysis using a numerical scale. The literature also shows that neither group was better regarding patient satisfaction(7,15-17).

In our study, the pain levels reported by the patients were low, as also found by Aglietti et al(18).

However, even though this pain was low, it significantly decreased the subjective IKDC, femoropatellar and SF-36 scores. Thus, the result was related to the presence of such pain. Seeking to achieve lower morbidity in graft harvesting will have a direct repercussion on the patient satisfaction rate.

CONCLUSION

The presence of anterior pain in ACL reconstructions, even if at low levels, has a deleterious effect on the final result over the medium term. The cause of this pain is multifactorial, but there is an important neuropathic contribution relating to the morbidity caused through graft harvesting. Because of the influence of this harvesting on the presence of sensitivity abnormalities and on walking on the knees, the choice of graft should take into account the patient's professional and sports activities.