Marc REVOL, Alain CORMERAIS, Isabelle LAFFONT, Jean-Paul PEDELUCQ, Olivier DIZIEN, Jean-Marie SERVANT

Instructional Course Lecture

Friday 27th February 2004

Upper limb surgery for tetraplegia - 8th International Conference. Christchurch, New Zealand, February 24-27 2004




What is a surgical "complication" ?

It's a problem that occurs as a result of surgery. So, limits between "complications" and "bad or disappointing results" are not clear.

Moreover, the negative consequences of complications or bad results are the same for the surgeon, because they make other patients and medical teams become reluctant to surgery.


That was the first "complication of complications" in the beginning of upper limb surgery for tetraplegia, and that remains true for any surgeon who wants to start this surgery.


I have chosen to divide my talk in 5 parts :

  • First, general complications
  • Second, skin complications,
  • Third, tendons
  • Then, joints
  • Lastly, the complications of Functional Electrical Stimulation.



Let’s begin with general complications.


They can be related to indication, anaesthesia, or surgery.




Professor Eduardo ZANCOLLI has detailed the causes of unsatisfactory results in tetraplegia and, as many other things in hand surgery, he inspired our present work.


Bad indications will certainly lead to complications and bad results.

Because they depend mainly on the surgeon, they can be avoided.

Preoperative evaluation and selection of patients is of utmost importance in order to avoid complications.

The patient's general condition includes : pressure sores, problems with urinating and defecating, psychological condition, motivation, and neurovegetatives syndromes. It is important that the patient can sit before the hands are operated on.

The tetraplegic patient's ability to carry out basic daily activities, and consequently his capacity to function independently, must always be evaluated.

Before beginning reconstructive surgery on the upper limbs, the spared muscles must have recovered their maximum possible function. Only those with a strength of MRC 4 or 5 can be used for transfers.

As severe spasticity constitutes a contra-indication to surgery, minor spasticity does not constitute a contra-indication.

Elbow, forearm, wrist and fingers joints must be as flexible as possible. Any defect in passive range of motion will be worsened postoperatively.

After Möberg's work everybody here knows the importance of the sensory function of the hand and its consequences in tetraplegia.

Obviously, the surgeon himself must be experienced and trained in upper limb surgery for tetraplegia. Experience does not avoid the risk of complications, but certainly lowers it, especially in the selection of patients and in the operative planning. For exemple, Zancolli does not recommend initiating a two-stage procedure with the reconstruction of the flexor phase. We strongly agree with him since my first tetraplegic patient was referred to us fifteen years ago because he had underwent such a procedure and was simply unable to open his hand.

Even if the surgeon is experienced, it does not help without a very experienced rehabilitation team. Without a trained and experienced rehabilitation center, surgery is certainly dangerous and must be absolutely avoided.



Whatever the anaesthetic technique used for upper limb surgery, it has it's own risks of complications. I don't think convenient to develop here all the possible risks, but I want to emphasize one of them.

In tetraplegic patient, you must be careful to look for a possible tracheal stenosis that could lead : 1/ to a difficult intubation, and 2/ to perform a tracheotomy after extubation due to a postoperative tracheal edema at the level of the stenosis, as it happened once in our experience.



As for any surgery, general complications are always possible, such as :

  • infections (especially pin-tract infections, with long lasting exposed K-wires),
  • haematoma (in this case after a deltoid-to-triceps transfer, and here after a triceps-to-biceps transfer).
  • Thromboembolism is possible but rare in tetraplegia in our experience. I would be happy to get more informations on this puzzling topic.



Pin migration with K-wires is classical in upper limb surgery, and it was observed in tetraplegia by authors such as Möberg and Smith.

Reflex Sympathetic Dystrophy can follow any upper limb surgery. We have seen one case out of 241 operations. Its treatment is equally difficult in tetraplegic patient as in other ones.

All the upper limb surgeons are aware of the dangers of the tourniquet. Most complications are directly related either to the duration of ischemia, or to the direct pressure generated by the cuff. Postoperative swelling and oedema is a direct consequence of tourniquet ischemia. For didactic purpose, I'll present them in the next point, related to skin complications.



Postoperative oedema begins as soon as the tourniquet is deflated, and may occasionally make the skin closure very difficult.



In this case, patient underwent a surgical revision two days after the initial operation, because he experienced an excruciating pain. Removal of the cast did not improve the pain, and sensory troubles of the fingers began to appear.



Surgical revision did not find any significant haematoma, …



even in the palm where lassos had been performed.



But we found skin of the distal forearm under great tension due to oedema.



Removal of the stitches at the ulnar side of the wrist, where ECU had been transferred to indirect lassos, resulted in this opening of the wound.



Even though skin closure was partial, some tension remained.



Spontaneous healing of the wound occurred in 3 weeks, but a Reflex Sympathetic Dystrophy occurred…



… and functional result was bad in this case, which illustrates relations between tourniquet, oedema, pain, and RSD.



An other postoperative complication remains somewhat mysterious to me : blisters.



In three or four cases, when I have removed cast and dressing for some reason a few days after surgery, I found that kind of blisters.



Similar to those observed in superficial burns or in epidermolysis, they can be very impressive, but their healing is very simple since dermis is intact. I don't exactly know their mechanism.



Necrosis of the full thickness of the skin can be minor, or major as in this case where it happened under a cast on an insensitive area of the forearm.



Unfortunately, not only necrosis concerned the skin, but also the underlying tendon transfer performed 4 weeks before.



A surgical revision was made to remove all the dead tissues, tendinous transfer included.



Healing spontaneously occurred in 2 months. Obviously wrist was not able to extend actively [without BR-to-ECRB transfer].



An other operation was performed for a tendon graft between stumps…



… of BR proximally,….



… and ECRB distally.



Skin tension on the radial side of the forearm was lowered by incision on the ulnar side, covered with a full thickness skin graft.



Postoperative functional result was poor.



Even if you suture skin with subcutaneous absorbable stitches and intraderma running suture, wound healing may occasionally result in a conspicuous scar, either enlarged or hypertrophic.



Let's move on now to the core of the topic : complications of tendon transfers.



Basically there are 6 complications in tendon transfers, namely :

  • Ineffective and/or insufficient results
  • Slackening or stretching or elongation
  • Rupture or breakage
  • Infection
  • Adhesions
  • And nerve or vascular compression.



Each complication can theoretically occur in any tendon transfer or tenodesis, so that at least sixty or seventy cases are possible.



Actually some complications are more frequent than others, depending on the operation.



Let's start with biceps-to-triceps.



Insufficient results have been reported by some authors, the last one is Arvid Ekeskär, who describes difficulties in activating the biceps as elbow extensor, which often takes months before patients can do it with ease. This drawback, when added to the risk of loosing too much strength in elbow flexion, might result in a subjective functional loss.

In the same paper, Ejeskär describes elongation of the transfer, and thus weakening of elbow extension after biceps-to-triceps. Yves Allieu has also mentioned that stretching of the transfer is the main complication. …

…and Ejeskär reports 6 total ruptures of the tendon anastomosis out of 28 arms treated by biceps-to-triceps transfer.

Finally, Ejeskär reports 2 cases of radial nerve entrapment after using the lateral route. Both patients underwent a neurolysis which helped one of them ; the other did not improve completely despite secondary medial rerouting of the biceps.

One ulnar nerve compression after medial routing of the biceps was reported by Michael Keith 3 years ago in Bologna.

In our personal 41-case experience with biceps-to-triceps transfer, the only one complication we have encountered so far is one bad functional result which was clearly related to poor-motivated patient in an inexperienced rehabilitation center.

Although we always use a tunnel superficial to the ulnar nerve, we have not observed any nerve compression so far.



Deltoid-to-triceps is certainly the most widely used method to restore active elbow extension, and many technical methods have been described to bridge the gap between deltoid and triceps tendons.

No bad functional result seems to be reported with deltoid.



But there are a lot of data about elongation that occurs in the two junctions of this transfer.

Three years ago, Alastair Rothwell concluded from his work that tendon stretch is inevitable, and that a 12% average increase in length occurred in 86% of cases during the 0-60° mobilisation.

Jan Friden stated that a significant tendon elongation of 23 mm was observed in his 7-patient series, but concluded that the long fibers of the posterior deltoid render it a very forgiving transfer because of its tremendous excursion as compared to the triceps one.

After completing the deltoid-to-triceps transfer, Ejeskär always places a suture of 3-0 stainless steel wire on both sides of each junction. In order to reduce by 50% elongation that occurs in the two junctions, he has described a postoperative protocol with arm-support and extra-care to avoid a too early and too important elbow flexion.

Rupture of deltoid-to-triceps transfer was observed by some surgeons, among them Jacques TEISSIER has two cases. Sophie TOUCHAIS described one case of olecranon fracture after a deltoid transfer into this bone.

Treatment of these ruptures is particularly difficult when synthetic material had been used.

In the same way, infection of the synthetic material used for deltoid-to-triceps transfer is particularly severe. We observed two cases out of 31.




In both cases, surgical excision of the entire material was necessary, and a biceps-to-triceps transfer was successfully performed 6 months afterwards.



Although I did not find reports about ineffectiveness of BR-to-ECRB transfer, the bad reputation of the BR to be a poorly educable muscle is well known and detailed by Caroline LECLERCQ in the book she edited two years ago.

 Incidentally, I want to emphasize that the main blood supply of the BrachioRadialis and Extensor Carpi Radialis muscles comes from the radial recurrent artery. When you perform a biceps-to-triceps transfer, it is of utmost importance to avoid any damage to the recurrent radial vessels.

Actually the only complication I found following BR-to-ECRB transfer seems to be the rupture of the transfer. Caroline LECLERCQ writes she experienced one rupture out of 65 BR transfers.

In our personal series we have also observed one case of rupture out of 5 BR-to-ECRB transfers.



Three months after the transfer, far after immobilization, it was obvious that the active extension of the wrist was very poor.



Surgical revision showed a stretching of the transfer.



A tendon graft was performed, whose distal end was put at the dorsal side of the retinaculum extensorum, in order to avoid suture entrapment , and to increase the moment of the wrist extensor as Möberg described it in 1978.



Functional result was pretty good, with an eighty degrees (80°) of active range of motion of the wrist.



In order to prevent rupture of BR-to-ECRB transfer in group 1, we have suggested the following procedure.



BR is transferred only to ECRB, in a classical way,…



… then ECRL is sutured along to the transfer…



… which is thus strongly reinforced.



Slackening and rupture are possible after EDC tenodesis.



But the usual problems of this tenodesis are its insufficient results, as emphasized by Antonio LANDI 3 years ago.



Whatever the structure with which you choose to perform EDC tenodesis :

either the distal radius,…



… or the extensor retinaculum of the wrist, …



…MP extension is produced by the flexion of the wrist. This MP extension is best when the flexion of the wrist is active, that is to say in group 5 [in group 6, there is no need to restore MP extension].Usually, the limitation of MP extension in groups 2, 3 and 4 is due to the lack of motors for active wrist flexion.



Moreover, the shortening of fingers flexors when they are restored results in a limitation of the flexion of the wrist by reciprocal tenodesis effect, thus impairing MP extension.



(clinical examples)



BR-to-EDC transfer was unable to produce any active finger extension in 6 cases out of 30 in the series reported by Ottonello and Leclercq. According to them, failure of the transfer seemed directly related to the lack of active FCR in groups 2, 3 and 4. In those cases, they recommended a passive EDC tenodesis to the radius, thus preserving BR for restoring another function.



In our personal series, 3 unsuccessful revisions were performed out of 35 BR-to-EDC transfers.



Ottonello reported one case of suture breakage, secondarily repaired with an unsatisfactory result,

and one case of adhesions at the suture site, which improved after tenolysis.



I have heard and read about cases of slackening and/or rupture of lassos, but, except accident in the first postoperative 6 weeks, and except for the uncommon case where FDS tendons are proximally fixed to the radius, I have problems understanding how secondary slackening and/or rupture of lassos can occur distally on tendons which originate proximally from a paralysed muscle.

Actually, the main problem with passive lassos in tetraplegia is a failure rate which was estimated to be 50% by Antonio LANDI in his series of 15 cases.

Some authors like Jacques TEISSIER  recommend neither the passive nor activated forms of Zancolli's lasso for the treatment of supple claw.

When the flexor superficialis is paralysed, the lasso technique could be efficient only if the visco-elasticity of the muscle is preserved. That is to say, only if the injured segment of the spinal cord is short or intermediate-size, as emphasized by Bertrand COULET and Yves ALLIEU.

However result is unpredictable, and the procedure is inefficient in some cases, especially when there is MP joint laxity.



We performed one surgical revision in such a case.



In this group 5 hand, FCR was transferred to the FDS in order to activate the lassos.

Intraoperative settings seemed to be correct,…

… but finally, claw deformity recurred, except on the 3rd finger.



An other patient of our series experienced two surgical revisions. This group 2 hand had been reinforced in two stages :

  1. EDC tenodesis, CM arthrodesis, 4 lassos
  1. BR-to-FDP and FPL transfer.

Result evidenced a swan-neck deformity…



…which have seemed to be related to an intrinsic cause, because Finochietto's test was positive.



A distal intrinsic release was performed on the 3rd and 4th fingers. Intraoperative correction was good,…



… but the deformity relapsed and worsened postoperatively.



Obviously it was a mistake, and the swan-neck had an extrinsic cause, related to the permanent MP flexion.

A FDS spasticity was thought to be responsible, and a surgical revision was indicated to cut them.



Lassos of 3rd and 4th fingers were dissected, and incidentally proved to be intact.



FDS were cut at this level,…



… and FDP were passively checked.



Final result was poor on 3rd and 4th fingers, and surgical revisions had been unuseful.



Adhesions in the palm are almost constant after lasso operation, and usually they don't impair FDS but FDP course.



That is why we systematically perform a blind tenolysis by pulling on each FDP tendon before performing a transfer to FDP when lassos have been made in the first operative stage.



I did not find any paper related to the complications of other intrinsic palliative than lassos in tetraplegia. But it is certain that complications exist with these operations.



As stated by Vincent HENTZ and Caroline LECLERCQ in their recent book, tenodesis of the deep digital flexor tendons to the radius is seldom performed today, because when its setting is correct to favor grasp, many patients are unhappy with this clawed hand.



Active flexion of the fingers is usually restored either by BR or by ECRL transfer to FDP. Although ineffectiveness and slackening of the transfer are possible, I could not find a pertinent reference about such clinical cases.



In regard to rupture, Jacques TEISSIER reported two cases to me, and described a stress rupture which occurred during a ping-pong party, 5 years after the ECRL-to-FDP transfer.

Concerning adhesions, …



… we have seen that adhesions in the palm are almost constant after lasso operation, and usually they impair FDP course.



That is why we systematically perform a tenolysis by pulling on each FDP tendon before performing a transfer to FDP when lassos have been made in the first operative stage.



Moreover, we are soon going to see that when FDP and FPL are activated by two different motors, like BR and ECRL, sutures, are very close one from each other; so that adhesions between them reduce their independence, and make the flexion of the thumb occurs along with the flexion of fingers.



There are three main FPL tenodesis procedures, described by Möberg, and by Allieu.

Complications I found with Moberg and Moberg-Brand FPL tenodesis are related to its progressive slackening, and to the difficult positioning of the thumb and index finger relative to each other.

If laxity of the tenodesis becomes significant Caroline LECLERCQ mentions it is possible to tighten the original tenodesis by exposing the distal insertion of the FPL, and re-anchoring the tendon at its new tension to the distal phalanx.



Motors used to restore thumb flexion by a transfer to the FPL are : BR, ECRL, ECRB, supernumerary ECR, and PT. Many operations mean many complications can occur.



We have observed some insufficient results, specially in PT-to-FPL transfer. I have found three reasons to explain this bad result :



   1. Even lengthened with periosteum, PT tendon is usually too short to reach FPL tendon, and a tendon graft is often required, thus increasing adhesions.



   2. Even covered by the fibers of FPL muscle, the bare area of radius where the periosteum was removed can produce adhesions to the transfer.



   3. Sutures between FPL and its motor on one hand, and FDP and its motor on the other hand, are very close one from each other; so that adhesions between them reduce their independence, and make the flexion of the thumb occurs along with the flexion of fingers.



Stretching of the suture is possible, particularly with side-to-side suture between ECRB and FPL.



45 year old tetraplegic female, group 5…



Initial side-to-side suture between FPL and ECRB.



Functional result was good, except for the pinch, whose strength was very weak.



Surgical revision revealed a stretching between ECRB and FPL, without any rupture.



Treatment was a section of FPL, whose distal end was sutured to the ECRL-to-FDP transfer.



Final result was pretty good.



6 years ago, in Cleveland, we have presented our cases of rupture related to the side-to-side suture between FPL and ECRB;



One patient experienced 2 surgical revisions.



The first one revealed a rupture of ECRB, whose proximal end was still I, continuity with FPL.



Treatment was a Y-shaped tendinous graft.



Unfortunately, soon after the removal of the 6-week immobilisation, there was a recurrence of the tendinous rupture.



The second surgical revision revealed a rupture of the distal end of the tendinous graft.



Treatment was a new tendon graft, whose distal end was put at the dorsal side of the retinaculum extensorum.

FPL was sutured to the ECRL-to-FDP transfer;


In contrast to the left hand, final result was poor on the right hand :


  • Active extension of the wrist was seventy degrees (70°) with a fair grasp and a fair key-grip
  • But flexion of the wrist was impossible, even passively, and severely impaired the opening of the hand.



I have previously said that …



….when FDP and FPL are activated by two different motors, like BR and ECRL, …



… sutures are very close one from each other,…



… that promotes adhesions between the transfers.


Other complications are possible, among them :
Pronator transfer of the biceps
Thumb extension and abduction.
Teissier lent me this case of rupture of EPL which occurred 2 years after the tenodesis.



Move on now to the joints…



(MP of the fingers have been previously seen)



The only complication I have found related to the elbow joint was emphasized by Yves ALLIEU, who warned us of the risk of instability in the shoulder joint if posterior deltoid is transferred to the triceps when the pectoralis major is paralysed.

I am glad I have followed his advice so far,

but other authors like Arvid EJESKAR disagree with it, and don’t give consideration as to wether the pectoralis major is paralysed or not.

Concerning Buntine’s clavicular advancement of the medial third of the deltoid muscle, I have asked Jacques TEISSIER, who has performed this procedure, with no complications.



I have previously said that fractures of olecranon had been observed after deltoid-to-triceps transfer, when synthetic ligament was inserted into the bone.



Actually, the main problem at the elbow joint is the preoperative flexion contracture, which must be solved before any transfer to the triceps. In our experience so far, rehabilitation and static orthoses have always achieved successfully this goal, without anterior capsulor release nor tendon lengthening of the brachialis muscle.






The main problem of the carpo-metacarpal arthrodesis of the thumb is the postoperative pain, which is frequently reported. This pain usually impairs the transfers of the patient from the bed to chair or vice versa. Fortunately the pain spontaneously disappears, usually after a few months.

Position of arthrodesis is very important. One surgical revision was performed in our series.



It was due to a bad initial setting of the arthrodesis, with thirty degrees (30°) antepulsion instead of forty-five (45°).



Situation was corrected by an intermetacarpal arthrodesis.



Permanent MP flexion of the thumb can prevent the pinch to be effective.



In this group 5 patient who had previously underwent carpo-metacarpal arthrodesis, thumb was always in the palm,…



… and the pinch was improved by preventing MP flexion.



A surgical EPL tenodesis was performed, proximal to the MP level

Postoperative result was good.



Postoperative result was good.



Thumb MP can also dislocate dorsally, as in this case Jacques Teissier lent me.




Interphalangeal arthrodesis results in the shortening of the thumb, which can impair the pinch when thumb was preoperatively short.



Difference between the pinch obtained by an interphalangeal arthrodesis and a carpo-metacarpal arthrodesis is shown in this group 4 patient.



A rupture of the split distal tenodesis is possible, as described by Caroline LECLERCQ. In order to avoid this rupture, the IP joint may be fixed with a Kirschner-wire for a period of 4 to 8 weeks.



Interphalangeal hyperextensibility of the thumb can impair the pinch.



In this group 1 patient, …



… a split distal FPL tenodesis had been previously performed, but pinch was innefective due to the hyperextensibility of the IP joint.



An IP arthrodesis was performed…



It proved to be useful intra-operatively…



… and postoperatively.



The device for Functional Electric Stimulation is not available any more nowadays, which is a genuine complication, both for the tetraplegic patients who are potential indications for FES,

….and for the patients who have been implanted and are prone to the well known complications of this method.



In conclusion, most surgical complications are avoidable, mainly by good indications and good techniques.

But many complications remain unavoidable risks, and the medical and surgical teams must be aware of them, in order to manage them at best.

Because the good news is that the autonomy of patients was eventually improved in all the cases of our series, even after complications.