Kocher’s Method (c. 1870)

Kocher is considered a pioneer in shoulder relocation. Here is Kocher’s original paper, translated from German by Dr Neil Cunningham and Dr Leo Kretzenberger. Admittedly, it’s a long read, but Kocher makes several profound observations regarding analgesia (which in his day consisted of chloroform) and traction-countertraction methods. To download a PDF of this paper, click here. For a review of Kocher’s key points, click here.


Kocher, Theodor: “Eine neue Reductionsmethode für Schulterverrenkung“, Berliner Klinische Wochenschrift (Berlin Weekly Clinics), vol 7, issue 9, 28 Feb 1870, 101-105.

A new method of reducing dislocated shoulders, By Dr Theodor Kocher, Lecturer in surgery, Berne.


It might be held against me if I attribute the title of a new treatment to a method of setting (reducing) dislocated shoulders whose first action totally coincides with a method already known. Considering however that Schinzinger himself admits a certain violence and painfulness of his method – for the following deals with a modification of that method – and also that eminent luminaries not only point out the dangers of considerable injury, but also lament the many deficiencies of the technique, it seems justified to me to direct attention to suggestions which are based on the results of experimental studies and attempt to remove these problems.

Hardly anybody who has done experimental studies would want to exchange such new views on the obstacles to reduction in dislocations as have resulted from the studies done by Roser, Schinzinger, Streubel, Busch and others and are clearly explained in Volkmann’s paper (Pitha and Billroth’s Surgery), for the old notion. All the same, the almost exclusive importance for the difficulties in setting the hip joint of capsular tears and of the non-ruptured parts of the capsule must not directly be transferred to all joints. Even if Busch states that the muscular tension has no bearing on the position of the arm in shoulder dislocations since cutting the muscles does not change that position at all, that does not mean that muscles do not play any part in the obstacles of the reduction. Streubel entirely correctly appreciated the importance of muscles when he said that those which already are put in tension by the dislocation of the humeral head will strongly resist any method of reduction by which they are meant to be stretched even further.[EXPAND Read Comments]

Muscle spasm: the role of muscular tension is noted with reference to resisting forced reduction and he quotes Streubel those which already are put in tension by the dislocation of the joint head will strongly resist any method of reduction by which they are meant to be stretched even further.” [/EXPAND]

Therefore, one cannot draw immediate conclusions from the corpse to the living body, all the more so since it is well known how much the reducing particularly of dislocated shoulders is helped by administering chloroform. [EXPAND Read Comments]

Again, regarding muscle spasm, the effect of active muscular spasm on position is noted, including the difficulty of inferring position during dislocation from findings during dissection or anaesthesia (include paralysis in modern joint operations) – “one cannot draw immediate conclusions from the corpse to the living body, all the more so since it is well known how much the reducing particularly of dislocated shoulders is helped by administering chloroform.”[/EXPAND]

Incidentally, even with corpses an influence of the muscles cannot be mistaken if the difficulties in reduction are compared between corpses still in the state of rigor mortis and corpses already relaxed. Notwithstanding all this, the statement remains valid that even without relaxation of the muscles dislocation can be set by relaxing the remainders of the capsule still under tension and opening (widening) of the capsular tear, while a mere relaxation of the muscles without any regard to the obstacles in the capsule are not producing any results.


In regard to the deltoid, another particular feature merits mention which – to the best of my knowledge – has not been paid attention yet in anatomical research. If one studies the deltoid more closely at a living persons or a corpse with subcoracoid dislocation, one finds an unseizable, roundish (muscle) strand stretching from the point projecting most laterally of the edge of the acromion towards the insertion at the upper arm. This strand is under more tension than all other parts of the muscle. During preparation, it appears as a roundish tendon running within the muscle close to its interior surface which can be isolated perfectly. It is so distinct that one could mistake it for the displaced biceps tendon in the living body unless one follows it to its origins, and it was this mistake that first arose my attention. It follows from the above that the deltoid whose tone was previously (e.g. by A. Cooper) seen as the main obstacle for reducing the dislocation, is primarily involved with its fibrous (tendinous) components and only secondarily with its muscle fibres.

I have performed, in part together with my friend Dr Otz, experiments with corpses in such a way that a cut was made from the axilla towards the head of the joint (humeral head), parallel to its anatomical neck, opening the capsule more or less to the extent that was found by Malgaigne in his sections of fresh cases of subcoracoid dislocation, i.e. from the lower edge of the subscapularis to the origin of the triceps. The arm was then forcefully lifted to the vertical and pushed downward. At lowering, a complete subcoracoid dislocation was to be found. The head of the joint stands beneath the coracoid process, only separated from this by the subscapularis muscle. The upper arm is not twisted around its axis. Either the anatomical neck or, more often, a point of the greater tuberosity stands at the anterior rim of the socket, in a way that a plane laid across the socket will just touch the exterior surface of this tuberosity.


As far as the muscles are concerned, pectoralis major, latissimus dorsi and teres major are relaxed. The subscapularis, lifted, runs over the anterior circumference of the humeral head and is not taut (tense), in corpses still in the state of rigor mortis, it is torn at its lower edge. Supraspinatus and infraspinatus as well as teres minor are drawn across the socket, the latter across its lower rim; only the supraspinatus shows a larger degree of tension. The long biceps tendon is taut and has taken a forward and inward direction from its origin at the upper rim of the socket. The combined tendons of the short head of the biceps and the coracobrachialis run under tension across the anterior circumference of the anatomical neck or outward and downward from the lesser tuberosity – depending on the state of the humeral head. The deltoid is tense, particularly the tendon running in its acromial part mentioned above, and it is unmistakable that it contributes to the fixation of the humeral head towards the anterior rim of the socket by its tension and in conjunction with the capsule. It is the circumstances of the capsule which determine the position of the humeral head. The capsular tear has the artificially produced extent mentioned above. The whole posterior wall of the capsule lies closely at the socket, covering it completely, it is moderately tense and covered by the respective muscles and tendons. The lower edge of the capsular tear, running from the position of the insertion of the triceps at the socket inward and slightly forward, is under strong tension. What is particularly tense, however, is the unharmed rest of the anterior upper wall of the capsule which is attached to the upper edge of the tendon of the subscapularis and is situated between that tendon and the tendon of the biceps. This part of the capsule suffers a twist in a way as if the upper arm had been rotated outward, and it is thus twisted into a strand running diagonally to the part of the capsule, from the upper corner of the socket insertion to the lower corner of the humerus insertion, which continues a short distance underneath the tendon of the subscapularis and partially fuses with the upper edge. This strand is twisted to such an extent that in my first merely tactual examinations I mistook it for the tendon of the biceps and therefore attributed much too much importance to that tendon. These characteristics of the anterior upper part of the capsule, which have not been appreciated sufficiently to date, I found to be constant, and it can be assumed that they will show in this way as a general rule, because the anterior upper wall of the capsule is protected from direct impact of the humeral head by the coracoid process, and from being involved in a continuing rupture of the capsule that started below by the tendon of the subscapularis.


Pitha appears to go a bit too far in “Pitha and Billroth’s Surgery” by maintaining that a complete anterior dislocation of the head of the humerus was hardly conceivable without the tuberosity being ripped off in larger or smaller parts. Incomplete dislocations of the shoulder joint are extremely rare. However, not only does no insertion of a tendon rip off in an ordinary anterior dislocation produced in a corpse, there are also a sufficient number of autopsy reports in fresh and slightly older cases where no muscle was found to be ruptured and no tuberosity ripped off. A. Cooper reports two such cases in his “Principles of Surgery”. Jarise in Lille described such a case (Schmidt’s Jahrbuch ‘yearbook’ vol. 122). Malgaise also mentions only small lateral tears at the edges of the subscapularis and teres major as a rule in subcoracoid dislocations. It was only in intracoracoid dislocations that he found in most cases the greater trochanter major ripped off, or the posterior part of the humeral head shattered, or almost the entire capsule ripped off. Since, according to Reuss’s observations, the greater tuberosity, as a general rule, does only heal imperfectly, while remaining dysfunctions are rather rare after shoulder dislocations, this is enough to conclude that such dislocations often occur in an uncomplicated way. This question of the frequency in which larger parts of the head are ripped off is of importance for the mechanism used in the method of reduction.

In the ordinary form of the dislocation underneath the coracoid process (subcoracoid/ true anterior), the main deviation of the head from the socket is inward, the second most important forward, and finally the least important one downward. The first and most correct indication for the purpose of reduction is therefore to lead the humeral head in a lateral direction. Once this has been done to a sufficient extent, the two further indications are met very easily by lifting a point of the surface of the humerus joint to the rim of the socket, thus causing the head to slide from this point backward and upward by itself, driven by the forces striving to make it approach the bottom of the socket.

How do the different methods try to meet the above indications? One can distinguish two large groups depending on whether the humeral head is lead in a lateral direction in a curve, or simply pushed outward. The elevation method as well as Schinzinger’s and Gordon’s methods belong to the first group, while Richet’s (respectively M. Langenbeck’s) and A. Cooper’s methods belong to the second. [EXPAND Read Comments]

Contemporary techniques in 1870: Kocher clearly notes that the first action (manoeuvre 1 – external rotation with humerus adducted) has previously been described by Schinzinger. He notes multiple techniques in use at the time, and includes comment on the similarity of some of these – elevation method, Schinzinger’s and Gordon’s, Richet’s (respectively M. Langenbeck’s), A. Cooper’s abduction method, Lücke (add chloroform to Richet), Hamilton, Mothe. [/EXPAND]

The method of direct pressure according to Richet and M. Langenbeck appears to be the simplest and mildest one. The fingers which have been pushed as high as possible inward from the head press it outside, while the scapula is moved inward. The tension of the capsule is only taken into account by lifting the arm slightly outward. In fresh cases when the head can be properly grasped around by the fingers, i.e. particularly in cases of primary axillary dislocations, this method is successful; in all other cases, which means in the vast majority of cases, it allows far too little force to move the humeral head. My distinguished teacher Prof Lücke uses it with chloroform anaesthesia.


A. Cooper’s abduction method, in which the surgeon’s heel is wedged into the axilla of the patient lying on the bed and the surgeon pulls at a sling running over the elbow joint of the dislocated arm with his whole body weight, appears to pay no regard to the circumstances of the capsule. [EXPAND Read Comments]

Traction: Kocher dismisses traction/counter traction in 1870 as anatomically inferior – “A. Cooper’s abduction method, in which the surgeon’s heel is wedged into the axilla of the patient lying on the bed and the surgeon pulls at a sling running over the elbow joint of the dislocated arm with his whole body weight, appears to pay no regard to the circumstances of the capsule.” [/EXPAND]

At least the head reaches a more favourable position in respect to the capsular tear as the consequence of a rotary movement of the scapula (with its outer angle downward and its lower one inward and upward); this head position corresponding to the one reached by lifting the arm slightly at the side of the body. The pull at the arm must be seen as the counter-hold allowing a forceful pressing of the heel in the axilla, respectively a pressure from the inside outward on the humeral head. The disadvantage of this forceful and efficient method is that blood vessels and nerves in the axilla are squeezed and pain is caused. Combined with the rotation, it can be applied in cases where the milder methods of rotation fail, as will be described below.

Among the methods of the first group, the elevation is still possibly the most commonly used method of reduction. The arm is abducted either horizontally according to Hamilton, or better up to the vertical according to Mothe et al. The fixation of the scapula is easier with the vertical elevation. The muscles tensed by the dislocation are most completely relaxed with these methods; only the long tendon of the biceps stays tense and pectoralis major as well as latissimus dorsi are tensed. The humeral head turns around a fulcrum at the anterior lower rim of the socket around a place of the greater tuberosity, thus the surface of the joint, respectively the head proper is given an outward and downward rotation curve. At the highest elevation, the outer surface of the humerus is opposite the socket, the surface of the humerus joint does not touch the rim of the socket. If the capsule were relaxed now, why should the head not snap into place by applying light pressure from below and lowering the arm? It is well known that usually very considerable force is needed to achieve this. Experimenting with a corpse, one can easily satisfy oneself of the fact that the view – still very commonly held – that the capsule were best lifted off by elevation, is wrong. Although the posterior wall of the capsule is relaxed, it still lies completely against the socket until it is pushed upward by the head brought upward by forceful pull or pressure. In addition to this, however, the parts of the capsule stretching from the upper and lower rim of the socket remain in strong tension at complete elevation, thus fixating the part of the greater tuberosity now resting on the rim of the socket as strongly to it as is the case with the arm lowered. So the main advantage of the elevation method, i.e. of leading the humerus head part of the joint outward and downward, is not so much the relaxation of the muscles – for those relaxed muscles others are tensed – and neither the widening of the capsular tear and the relaxation of the remainders of the capsule, but rather the fact that the method of moving by direct pressure from below upward can be applied with much more force than would be possible in the given dislocated position from inward, and also that this pressure can be supported by pulling the arm. Thus this method remains a forceful and injuring one. Incidentally, a simultaneous inward rotation during the elevation brings the surface of the head closer to the socket.


According to Gordon’s method, as in the elevation, the humeral head is lead in a curve downward and outward around a sagittal axis leading through the fulcrum on the rim of the socket, but only after the upper arm was lifted forward in the sagittal plane until it was perpendicular to the anterior surface of the body. This latter manipulation relaxes the upper capsular wall and counteracts its torsion in a most helpful way; however it does tense the lower posterior wall of the capsule strongly and this fixates the head against the anterior rim of the socket. Thus the reduction cannot be achieved at all, or only by force.

The external rotation as suggested by Schinzinger is definitely a very big improvement. With this, the head makes an outward and forward curve around a vertical axis leading through the fulcrum on the rim of the socket. This method has the advantage before all others that the tension of the muscles and the remainders of the capsule is not only avoided – this could also be said of M. Langenbeck’s method – but is even used for the reduction. If one performs the manipulation with a corpse that has been given a subcoracoid dislocation in the manner described above, the following becomes apparent: The first action of pressing the upper arm towards the body has as consequences a stronger tension of the deltoid muscle, slight outward movement of the humerus head and stronger fixation of the greater tuberosity or of the rear circumference of the anatomical neck toward the anterior rim of the socket by tensing the upper wall of the capsule even more strongly. It appears to me that Schinzinger did not lay enough stress on this; for now a fixed fulcrum has been gained around which the head can be externally rotated. During this rotation, the joint surface of the humerus appears, depending in the position of the head, completely or partially under the short biceps head with coracobrachialis outwards, until it (the humeral joint surface) faces more or less forward. The upper wall of the capsule, respectively the part of it that starts at the upper corner of the socket and is twisted into a strand, stays strongly tensed with this manoeuvre – one can satisfy oneself of this without any doubt – it is twisted even more that it was before and describes the surface of a cone whose apex is at the upper corner of the socket. As a consequence of this action, the rear wall of the capsule which is linked to the strand, is also forced to lift off the socket; and there is no other method by which this would happen in such an efficient way; i.e. by which the capsular tear would be made to gape that much. B. Bruns is absolutely wrong to deny this to Schinzinger. At the same time, the rear wall of the capsule with the supraspinatus, and of course infraspinatus and teres minor, is relaxed. There are cases where the humeral head snaps in at this moment; however as a general rule this does not happen, but the inward rotation rather brings back the humeral head to its position underneath the coracoid process. In order to avoid that, Schinzinger stipulates that at the end of the external rotation an assistant apply pressure from the front on the humeral head, while the inward rotation is performed. But even so the reduction fails as often as it succeeds. The frequent failure is easily explained: The point of the humerus lying on the rim of the socket at the end of the external rotation stretching is not part of the joint surface, but part of the rear circumference of the greater tuberosity. It is the outer surface of the same tuberosity that is turned toward the capsular tear. [EXPAND Read Comments]

Throughout this paper the anatomy is described in wonderful detail, as above:

At the end of the external rotation “The point of the humerus lying on the rim of the socket at the end of the external rotation stretching is not part of the joint surface, but part of the rear circumference of the greater tuberosity. It is the outer surface of the same tuberosity that is turned toward the capsular tear.” [/EXPAND]

Since, as explained above, the upper wall of the capsule is even more strongly tensed and rotated that it was in the original dislocation position – which, as Schinzinger himself notes, explains a slight upward movement of the humeral head at the rim of the socket – the humeral head has not become any more movable on the rim of the socket by the upward rotation at all. Schinzinger tries to remedy that drawback by forcing the outward rotation. Sometimes this ruptures the strand at the upper anterior wall of the capsule, thus enabling the backward movement of the humerus head. But it is just this intensification of the outward rotation that makes the method forceful and painful at the least – as the author himself states.


I set myself the task to find a manipulation in my experiments at corpses which, starting from the unforced external rotation, brings a point of the humerus joint surface onto the rim of the socket, relaxing and untwisting that abovementioned strand at the upper wall of the capsule. By having turned the arm in an outward direction, so much has already been achieved toward the approximation of the head to the socket and the widening of the capsular tear, that it would be a pity to forfeit these advantages. That task has now been fulfilled. If one lifts the externally rotated arm forward in the sagittal plane, one can see how that strand relaxes completely, and how the humeral head describes a curve upward and backward around a cross axis running through the fulcrum of the greater tuberosity onto the rim of the socket with its joint surface, until a part of the humeral joint surface has been lifted onto the anterior rim of the socket. If the arm is rotated inward now, the head will unfailingly slide into the socket. The upward rotation of the part of the joint around a cross axis is provided by the tension of the part of the capsule starting at the lower rim of the socket resulting from lifting the arm. This tension does not result in a renewed fixation of the head toward the rim of the socket now, since the respective part of the capsule exercises its pull in the same plane in which a movability of the head is given, respectively in the plane of the socket. This is the reason why the lifting of the arm in the sagittal plane from the dislocation position up to the horizontal according to Gordon does not remotely have the same effect, since with this procedure the lower rear wall of the capsule is tensed in a direction that must result in a strong fixation of the humeral head toward the anterior rim of the socket. This method: Pressing the arm bent at the elbow towards the body, turning outward until resistance is felt, lifting of the outwardly rotated upper arm in the sagittal plane as far as possible, and finally slowly turning it inward, was demonstrated to the Cantonal Association of Physicians in Berne at a corpse and has already been tried with living patients several times e.g. in the hospital ward of my distinguished teacher Prof. Lücke. [EXPAND Read Comments]

This method is designed for the subcoracoid dislocation: The dislocations that Kocher and Dr Otz produced were all subcoracoid (pure anterior dislocations) via an artificially produced anterior capsule tear. [/EXPAND]

Among 11 cases of shoulder dislocation underneath the coracoid process which I treated during the last three years, I tried Schinzinger’s method 8 times. Three times it was successful following Schinzinger’s instruction to the point, two times I applied the reduction according to Cooper’s method after Schinzinger’s had failed, and in the most recent three cases I immediately succeeded with the new method.

The advantages of the suggested method of reduction as compared to Schinzinger’s method consist in the facts that it is not painful, since the external rotation does not have to be forced any more, that it does not cause any further injuries, that it can be performed without a single assistant, also without chloroform, and finally that it succeeds in cases where Schinzinger’s method fails. B. Bruns wanted to restrict the latter to cases of incomplete dislocation where a part of the joint surface is still situated on the rim of the socket. If by this he is referring to cases where the anatomical neck is situated on the rim of the socket, i.e. complete dislocation with the minutest deviation inward (genuinely incomplete dislocations being extremely rare, as we know), his indication is correct. The suggested modification extends the applicability of the method to cases where a part of the upper end of the humerus can still gain a fulcrum for rotation at the anterior rim of the socket.


With this, the contraindications to my method are suggested: If the joint capsule is almost completely ripped off, particularly if upper and anterior wall of the capsule are missing; if the whole mass of the greater tuberosity is not broken off, respectively the posterior circumference of the humeral head is not shattered; finally if the anterior rim of the socket has been avulsed (and not just the glenoid rim) – in those cases the mechanism for rotation is disturbed, the head will turn around its own axis and will not be lifted out laterally. [EXPAND Read Comments]

Contraindications to Kocher’s method are clearly stated:

“If the joint capsule is almost completely ripped off, particularly if upper and anterior wall of the capsule are missing”

“if the whole mass of the greater tuberosity is not broken off, respectively the posterior circumference of the humeral head is not shattered” (manoeuvre 2 and 3 require an intact greater tuberosity)

“finally if the anterior rim of the socket has been avulsed (and not just the glenoid rim) – in those cases the mechanism for rotation is disturbed, the head will turn around its own axis and will not be lifted out laterally.” [/EXPAND]

The more therefore the head has departed from the coracoid process toward the interior, the more marked the crepitus when rotation is attempted, the less can be expected from the method. A case of shoulder dislocation where, according to Prof. Lücke’s diagnosis, the anterior rim of the socket had been avulsed, and which I had the opportunity to observe at the hospital in Berne, could not be reduced following the suggested method. It was only direct pressure applied with the arm elevated under anaesthesia that brought the fragment and the head into their places.

In cases where the head does not have a fulcrum at the socket rim any more, the suggested procedure can be combined with Cooper’s method to the best advantage: The heel pushes the head in an outward direction, while an upward rotation combined with pulling the arm widens the gap of the capsular tear. Then the surgeon sitting next to the patient rises, thus bringing the arm forward and upward into the sagittal plane as far as possible; finally an inward rotation is performed.

There is no other method that would meet the requirement of approaching the head to the socket in all three spatial dimensions as well as the one suggested by me. The external rotation brings the head not only laterally, but at the same time forward, in a sense removing it slightly from the socket. The subsequent elevation however makes up for this by bringing the head backward; at the same time it is moved in an upward direction toward its normal position by the elevation.

Since with the suggested method, chloroform is detrimental rather than useful, this paper did not deal with the pendulum and the rubber hoses methods, because the specificity of those two methods is also based on its effect on the musculature.[EXPAND Read Comments]

Kocher notes that drugs don’t assist with relocation using this method – “Since with the suggested method, chloroform is detrimental rather than useful, this paper did not deal with the pendulum and the rubber hoses methods…” [/EXPAND]


Of course the mentioned procedure does not apply to old cases. The attached figures are intended to visualise the suggested mode of reduction. Figures1, 2, 3 represent lateral views, figures 4, 5, 6 anterior views. Besides the socket, the coracoid process and the top end of the humerus which are represented in all figures, figures 1 and 4 also show the scapula with the acromion and a part of the clavicula.

Apart from the osseous parts, only the end of the subscapularis muscle with its insertion at the lesser tuberosity and the upper anterior wall of the capsule, positioned between the upper edge of the subscapularis tendon and the biceps tendon, are depicted. The lower edge of the capsular tear starting from the lower rim of the socket, is only suggested by a ring. The whole of the rear wall of the capsule between the mentioned rest of the capsule above and below was omitted from the drawing for the sake of clarity.

Figures 1 and 4 depict the subcoracoid dislocation, with the humeral head positioned beneath the coracoid process, only separated from it by the subscapularis. The latter covers the upper part of the joint surface of the humerus. A strongly tensed strand runs from the upper corner of the socket toward the upper edge of the subscapularis tendon close to its origin. It runs diagonally across the oblong rectangular part of the anterior upper wall of the capsule and has been created by a rotation of the latter which tends to remove the upper corner of the origin of the socket from the lower one of the origin of the humerus. The part more distinctly drawn corresponds to the interior surface, the one less distinctly drawn to the exterior surface of the respective part of the capsule. There is only one point of the outer part of the rear surface of the greater tuberosity at the anterior rim of the socket. The lesser tuberosity faces forward, the greater tuberosity outward.

Figures 2 and 5 show the circumstances at the end of the external rotation with the arm pressed to the body. A small piece of the greater tuberosity protrudes toward the back in the direction of the anterior rim of the socket; i.e. it is opposite the socket. The main mass of the humeral head is positioned outward from the plane of the socket, but forward at the same time. The lesser tuberosity faces forward, the surface of the joint forward. The strand of the upper capsule wall is tensed as much as in the previous figures, and even more twisted. In figure 5, only the outward surface of the capsule part is visible. The subscapularis is tensed.

Figures 3 and 6 represent the arm in the elevation toward the front in the sagittal plane. It can be seen how the head, fixated by the tension of the lowest intact part of the capsule, turns on the anterior rim of the socket, thus moving a place of the joint surface up onto the latter. At the same time, the upper anterior part of the capsule is twisted open and relaxed, so that no particular strand is forming out of it. In figure 3, a part of the interior surface of the capsule part is visible; in figure 6 only its exterior surface.


Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>