|
 |
|||||||||||||||||||
|
|
|||||||||||||||||||
A seminar on the Shoulder, held at the Saint-Antoine Hospital, Paris, provided an opportunity for the school to describe and explain a number of its treatment methods. Professor André Apoil, who chaired the seminar, particularly emphasised his department's experience in repairing rotator cuff lesions. He shares his views on shoulder disorders with us here.
M.O. During the seminar, there was a lot of discussion about large rotator cuff tears. What is your preferred treatment?
A.A. At the moment, the deltoid flap repair is our basic procedure for large rotator cuff tears involving only the supraspinatus and infraspinatus: In the best cases, it will allow us to restore the position of the humeral head; and in most cases, it will at least provide a sheet of living muscle between the humeral head and the coracoacromial arch. However, when muscles other than the supraspinatus and infraspinatus are involved, in other words, when the tear extends to the subscapularis anteriorly and the teres minor posteriorly, we feel that a deltoid flap repair is not the answer. We have not yet found the ideal treatment for this type of lesion, and we would probably wait as long as we could before operating, knowing that in the end we would probably have to insert a bipolar prosthesis. However, it is not really very logical to use a prosthetic joint replacement to treat what is essentially a muscle deficit problem.
A.A. It is not so much a tear of these muscles as degeneration. In other words, you find that not only have the tendons disappeared (the supraspinatus tendon has retracted to the glenoid, and the infraspinatus tendon is stuck to the coracoacromial arch), but, most of all, there is fatty degeneration of the belly of the muscles. So, in fact, you can't do any form of direct repair other than a deltoid flap (or a flap repair using some other of muscle).
A.A. Before we started doing the flaps, we used to do what we call an arthrolyse. The Americans borrowed the technique from us, and then called it "decompression and debridement". In practice, we would resect the tendon of the long head of biceps if it was still there, tidy up the edges of the tear, resect the coracoacromial ligament and do a minimal acromioplasty.
A.A. In two-thirds of cases there was a marked improvement in terms of pain relief, but there was little functional improvement in mobility: The ROM did not increase significantly, and the shoulder would tire very easily. The most serious drawback was that almost half the cases inevitably progressed to cuff-tear arthropathy within ten years.
A.A. Generally, it was slow; apart from industrial accidents, there seemed to be a good initial result, but after three years they would gradually deteriorate.
A.A. Definitely, provided there was only an isolated tear of the supraspinatus or infraspinatus, and the subscapularis anteriorly was sound.
A.A. That's right; we call them mini-flaps. In cases where we would have used an arthrolyse for tears 3 cm in diameter exposing the long biceps tendon, we now do a deltoid mini-flap so as not to have to resect the tendon. This patches the cuff with healthy living muscle and lets the tear heal. It is very simple to do, as you have no subluxation of the head with these lesions; and since the tear is relatively small, the sutures are not put in under tension.
A.A. Yes, but in that case you decrease the articular capacity of the glenohumeral joint and you decrease rotation; I don't think it's a good idea.
A.A. You have to resect them, but sparingly. In practice, you only have to freshen the edges, remove any fibrous tissue, and get back to tissue which is well vascularised rather than actually bleeding.
A.A. There are a number of explanations for the disappointing results obtained by authors who tried the technique. The operative technique may have varied somewhat, and postoperative care may not have followed the sequence that we recommend. Then again, some authors tried to repair very large tears which extended well beyond the supraspinatus and infraspinatus, and they had failures just as we have had with those kinds of tears. So I think the procedure became rather discredited partly because of poor technique, partly because of poor postoperative care, and partly because the indications were stretched to include very large tears, well outside the scope of what could be repaired. And, of course, deltoid has always been regarded as the most important muscle in the shoulder, and a lot of surgeons feel that it is sacrilege to touch it.
A.A. You have to take the anterior fascicle of the middle deltoid, and not go too far forwards or backwards; that is very important. You must not split the deltoid too far anteriorly, because there are branches of the thoracoacromial artery which supply the anterior part of the flap. You have to identify very clearly the edges of the cuff where the sutures will be placed; this is not always easy, especially posteriorly between the infraspinatus and the teres minor, where it is rather difficult to distinguish between the teres minor, which you bring up from its insertion on the greater tuberosity, and the infraspinatus, which is in a different plane and is retracted underneath the acromion; it is quite difficult to perform the dissection so that you follow the edge of the tear posteriorly. You have to place all the sutures before tying them, as you would when suturing the intestine, and then you have to tie the knots one after the other, distributing the tension evenly.
A.A. The standard reply is that, for a long time, there is no parallel between the anatomical situation and the clinical manifestations. In other words, there are very small tears which cause a lot of problems in tight shoulders; and there are very large tears which cause very little trouble in relatively loose shoulders. It depends on how much give the capsule has: Major cuff lesions can be tolerated in a very mobile shoulder where there is no capsulitis; while a small lesion in a shoulder which is a little tight, with inferior capsulitis, will be very poorly tolerated. This is why patients will put up with some types of large tear for a long time, and then one day the shoulder becomes stiff, either from disuse or because it has been immobilised after trauma, and the tear becomes a real problem.
A.A. With small tears which are causing problems, we do a decompression and debridement if the biceps tendon is not exposed. If it is exposed, there are two options: You can either resect it, if it is really very damaged; or you can do a mini-flap, if it looks as though it can be saved.
A.A. Yes, you could, but only if you can suture it in the line of the fibres, which isn't often the case.
A.A. Because tendon-to-bone repair doesn't work very well. Also, if you reinsert the torn part of the cuff onto the greater tuberosity, you create a discrepancy between the proximal part of the tear and the healthy part of the cuff which remains inserted on the greater tuberosity. Logically, you should divide the torn part of the cuff away from the healthy part both anteriorly and posteriorly, so as to mobilise only the damaged section, which can then be reinserted onto the greater tuberosity. In any case, one part of the cuff is going to be taut, alongside other muscles which will still be under normal tension. I feel that the cuff will not be able to function normally in that situation.
A.A. In practice, we only resected the coracoacromial ligament, removing its zone of insertion on the acromion, which meant that we were literally peeling the anterior undersurface of the acromion. We used the deltopectoral approach for the procedure, and even so we found it quite difficult to locate the acromial insertion; it was basically a resection of the coracoacromial ligament.
A.A. I do not think that there is such a thing as an isolated acromioplasty: It must be an acromioplasty combined with resection of the coracoacromial ligament. It's part of the procedure for relieving a mechanical impingement by enlarging the subacromial outlet.
A.A. Our routine practice is to use the superolateral approach for arthroplasty, detaching the supraspinatus; thus, when there is a rotator cuff tear, we finish the procedure with a deltoid flap, to effect a tight closure. In an arthroplasty patient with a known rotator cuff tear, we have a ready-made approach, and we do the arthroplasty through the tear. Then we finish by closing the joint tightly with a deltoid flap. But here again, the limitations of an unconstrained arthroplasty plus flap are the same as the limitations of a flap repair. To put it another way, you can only do this procedure if there is a tear which is either restricted to the supraspinatus (in which case it is just an extra stage in the procedure, to finish the suturing of the cuff), or if there is a tear of the supraspinatus and infraspinatus which does not involve the subscapularis anteriorly nor the teres minor posteriorly. Ideally, you would even have a little strip of infraspinatus behind, to stabilise the teres minor. In that situation the flap works very well, and the results are similar to the results you would have with an unconstrained prosthesis without a flap, where the cuff is undamaged. But as soon as you have a tear which extends anteriorly into the subscapularis (more than a third of subscapularis), or when you have a lesion which extends posteriorly to involve the teres minor, then a flap will no longer be any use, since the prosthesis will ride up and displace, and you will get a poor result.
A.A. No, you can't do that, because it would greatly reduce the capacity of the joint, which makes rehabilitation difficult; and you would have a prosthesis which does not restore mobility. The result would be poor because there would inevitably be excessive tightness and increased superior loading, and the cuff sutures would fail. A cuff suture will only hold in a flexible shoulder. If there is any capsulitis, or if you reduce the joint capacity by making capsular rotation flaps, which comes to the same thing, you will have problems in the cuff because free abduction will be impossible.
A.A. Exactly. But in nearly a quarter of cases it will actively recentre the head. It can be difficult to assess, because you're not going to ask patients who are doing well to have repeated X-rays over a number of years.
A.A. We prefer bipolar prostheses, which we are developing at the moment. They are useful in Stage II arthritis where the head has subluxed, because there a new glenoid will have been established, and therefore a new fulcrum for the head, which only has to receive the socket of a bipolar prosthesis. But it is also definitely a good solution in cases of large rotator cuff tears which cannot be repaired with a deltoid flap. Once the tear extends to more than the upper third of the subscapularis anteriorly and to the teres minor posteriorly, an unconstrained prosthesis with a flap ceases to be a viable proposition, and at that stage one could consider a bipolar prosthesis.
A.A. Yes, indeed - and this is why you must not compromise the integrity of the arch in any way; in other words, you have to preserve not just the coracoacromial ligament but all the osteophytes on the acromion; you mustn't touch the acromion. So, firstly, you need to preserve the entire arch, which is not easy; and, secondly, if the new fulcrum that has established itself is wrong, you may have to relocate the fulcrum by taking down the apex of the glenoid so that you can centre the head properly and lodge the socket securely under the coracoacromial arch.
You have to insert the prosthesis in a little more retroversion, so that the socket is well stabilised under the acromion and does not migrate upwards and forwards, which is a potential problem with this type of device. But, obviously, these prostheses give limited results; they would produce 80-90° abduction and good rotation.
A.A. I think you get the best clinical results with total prostheses, provided the patients have been properly selected. You can do shoulder arthroplasty without replacing the glenoid, but you need a total prosthesis to abolish pain, which after all is the main goal of surgery. Certainly, you can circumvent the problem of glenoid loosening by inserting only a humeral hemiarthroplasty, but the pain relief obtained will never be as good. There is virtual agreement on this point, throughout all the series.
A.A. They have a very poor purchase in the neck of the scapula. In fact, there is only one vertical fin, which goes into an extremely thin scapular neck, and the whole situation is very different from what you have in the acetabulum, where there is a solid hemisphere that goes into a hollow hemisphere, with good pressure and stress distribution. With the glenoid cavity, you always have uneven loading, for which it is very difficult to compensate; and the less mobile the shoulder, the greater the stress will be. If a shoulder is very mobile after surgery, the glenoid component is unlikely to loosen, or you will find non-progressive lucencies which are well tolerated; whereas if the shoulder is stiff, there is a great deal of stress on the glenoid, with a nutcracker effect in the upper part of the implant; if the lower part of the capsule does not open in abduction, there is excessive superior loading of the glenoid component, which tends to rock and work loose.
A.A. No, it has been designed quite differently. There were a lot of things I didn't like about the Neer. First, it is in too much valgus, which causes local stretching of and damage to the cuff; in fact, this stretching was what Neer originally wanted, to try to increase the efficiency of the cuff, but in practice it caused cuff tears in a third of cases. So the main objective of the prosthesis which we have been developing is to protect the cuff. The implant is in very little valgus and corresponds more closely to the normal anatomy of the upper end of the humerus.
A.A. 135°. Another advantage is that the head has a variable radius of curvature, which is longer at the apex; this means that the apex of the prosthesis has been effectively shaved off so that it will not damage the cuff. Ideally, the apex of the prosthetic component should be practically at the level of the greater tuberosity, certainly not extending beyond it, and the cuff above the prosthesis should really be under no tension at all. Also, we wanted to develop a prosthesis that would favour a combination of rolling and sliding. You have to have a combination of rolling and sliding to achieve a physiological range of movement in the shoulder joint, as the bearing surfaces have very little congruency, with the surface of the humeral head being much larger than the surface of the glenoid. In fact, the PA prosthesis is a compromise with a condylar type of joint, such as the knee, where typically there is rolling and sliding of the condyle over a more or less flat surface.
A.A. Yes, but it was not part of the design. In fact, if the surfaces are completely congruent, the mobility of the glenohumeral joint stops at the limit of the overall range of motion allowed by the bearing surfaces.
A.A. I find it difficult to answer that objectively. What I do know is that we have incomparably better results with these prostheses than what we used to achieve with other patterns.
A.A. To tell the truth, the superolateral approach certainly has an effect on the result. I think that preserving the subscapularis is crucial in prosthetic shoulder surgery; it is very important that the subscapularis is not cut: This vital stabilising restraint of the shoulder must be preserved. Similarly, the shape of the bearing surfaces protects the cuff and encourages its action. I think our results are really excellent, particularly in terms of mobility.
A.A. Absolutely. But we gained our experience from operating on degenerated rotator cuffs. Paul Dautry said a long time ago that the supraspinatus was not by any means the most important muscle in the shoulder; it was not the muscle that started the abduction process, as everyone was saying at the time. We have seen a fair number of shoulders in which the supraspinatus was missing (for whatever reason) and in which a tear was perfectly well tolerated and completely normal abduction was preserved, because the shoulder had remained flexible. It was this observation, combined with the fact that we wanted to be able to use the same approach for a deltoid flap procedure in case we experienced problems with a cuff repair, that gave us the idea of inserting prostheses via the superolateral approach, which was originally conceived as an approach for flap procedures. I have no scruples about cutting the supraspinatus; I cut it in the so-called avascular zone, about 5 mm from the insertion on the greater tuberosity; in this zone the proximal edge is vascularised by the muscle body, and the distal zone is vascularised from the periosteum, and there is every likelihood that the repair will heal well, while tendon-to-bone repairs are always dicey.
A.A. Yes, but experience has shown that if the subscapularis is cut, it does not heal well. You see it all the time in recurrent dislocations. When you see a patient years after surgery for instability where the deltopectoral approach was used, you find that if the subscapularis was cut, the muscle is useless and you end up with anterior/posterior instability; if you look at an arthro-CT or an MRI scan, you find there is practically no subscapularis left. When you do a revision by the deltopectoral route, you don't find very much anteriorly.
A.A. I think Jean Gosset was the first person to become interested in it, especially in the management of recurrent shoulder instability, for which he devised the rib graft technique. That goes back to the early 1950s.
A.A. He had always been interested in traumatology and bone surgery, amongst other areas. It went with his interest in mechanics, craft work, anything with a bit of DIY about it. And he had produced acrylic hip prostheses at the same time as Judet. He was very interested in tibial plateau fractures, and in recurrent shoulder dislocation. At the same time, his assistant Paul Dautry was interested in degenerative shoulders and in rotator cuff tears, and was tending to do decompression and debridement while everybody else was repairing the cuff. Dautry acquired a reputation for relieving impingement, and this way Saint-Antoine came to be known as a centre of shoulder surgery.
A.A. Yes, that's true - in fact, that's a bit of an understatement. He was a man who had lots of ideas, he was very innovative, and he was an excellent surgeon with a great interest in many subdisciplines, such as surgery of the hand, the shoulder, the stomach, the colon, and the uterus. He devised a lot of techniques. Combined synchronous abdominoperineal excision of the rectum was also a Jean Gosset technique. He had a very open mind, he was keenly interested in all sorts of things, he had very humanist views, and he taught me a great deal.
A.A. There was a trial of strength going on most if not all of the time. The relationship between these two very strong personalities was a complex one: They were bound to clash; at the same time, there was a lot of mutual appreciation; and, equally, an inability to get on with each other. Their ethical codes were very different, too, as the boss was a devout Catholic and Dautry was a Protestant. They "split up" after the student riots of May '68 - or perhaps we should say that they decided to go separate ways. At the time, Dautry was very strongly opposed to all the unrest and resisted it vehemently, while Jean Gosset was more inclined to try to understand it all.
A.A. It's certainly true that Dautry was a bit of a rebel in his concept of surgery, and he did question all the accepted ideas; but then he had his own vision of hospital medicine and of the hierarchy within a department. He felt that the riots were counterproductive - that they would achieve the opposite of what the students wanted. And there, I think, he had a point.
