Hand Tendon Transfers 

Updated: Oct 31, 2018
Author: Steffen Baumeister, MD; Chief Editor: Joseph A Molnar, MD, PhD, FACS 



Paralysis of the upper extremity produces major functional impairment. The ability to perform activities of daily life can be severely compromised, especially in bilateral paralysis. When muscle-tendon units remain functional in an extremity, consider sacrificing one function to restore another by transferring the working unit to a new location. Restoring something as simple as a pinch grip can create major improvement in the function of the extremity.

History of the Procedure

Tendon transfers have been used in upper extremity reconstruction for well over a century. Early on, the technique was used for reconstruction following obstetric brachial plexus palsy or paralysis secondary to polio. When hand surgery evolved as a subspecialty, transfer techniques expanded. The middle part of the 20th century saw the development of transfers for multiple peripheral nerve paralyses, including median, ulnar, and radial nerve palsies. Multiple surgeons contributed to the field, including Bunnell, Boyes, Brand, Burkhalter, Goldner, Littler, Moberg, Omer, Phalen, Riordan, and Zancolli.[1, 2, 3, 4, 5]

In the latter part of the century, microvascular techniques were developed that added free muscle transfers as a possible tool for paralysis reconstruction. While newer techniques of functional electrical stimulation and microvascular transfer have added new dimensions to reconstruction, tendon transfer remains a primary tool in upper extremity paralysis management.


Loss of hand function secondary to neuromuscular paralysis, tendon loss, or muscle loss can dramatically reduce a person's ability to perform normal activities of daily life.[6] The prospect of restoring function through transfer of a muscle-tendon unit to perform one of the lost functions is enthusiastically accepted by the patient and rewarding to the surgeon. Tendon transfers may also be useful in restoring function in patients with spastic disorders due to cerebral palsy or stroke. (For more information, visit Medscape’s Stroke/Cerebrovascular Disease Resource Center.


Motor dysfunction or paralysis in the upper extremities is due to traumatic or nontraumatic causes.

Direct trauma

Most commonly, trauma causes a direct, primary injury to the peripheral nervous system or the muscles and tendons. Penetrating injuries can result in transection of the median, ulnar, or radial nerves or more centrally in the brachial plexus. Blunt trauma less often affects the more peripheral nerves but is often the cause of brachial plexus or nerve root injuries. Any of the motor functions in the upper extremity may be compromised by injury to the brachial plexus or nerve roots.

Direct trauma is also the leading cause of spinal cord injuries resulting in upper extremity paralysis. Paralysis from a cord injury often results in bilateral loss of function, although the losses may be different between sides.

Indirect trauma

Rupture of the extensor pollicis longus (EPL) tendon is a common complication secondary to a distal radial fracture. Anatomically, the EPL tendon runs through the third extensor compartment and around the Lister tubercles. In this site, the EPL is prone to injury as a result of bony malalignment or iatrogenic injury after osteosynthesis, commonly after screw insertion from the palmar side. The EPL usually ruptures several weeks after the initial injury.


Paralysis of neurologic origin may also occur from causes other than trauma, but this tends to be less responsive to surgical solutions. These causes include stroke and neurologic diseases such as multiple sclerosis or cerebral palsy.


Specific loss of tendon or muscle substance can arise from rheumatologic causes, including gout and pseudogout. The tendons are usually destroyed by pannus formation, chronic inflammatory processes, uric acid, or calcium pyrophosphate depositions. These factors cause enough destruction of tendon substance that tendon transfers are required to replace lost function. Direct suture repair is not usually possible. Rheumatoid arthritis may be responsible for tendon rupture of any of the hand and wrist tendons, but it commonly leads to rupture of extensor tendons of the fingers or thumb.


Any interruption of the neurologic-muscular-tendinous chain results in paralysis. As outlined above, this may be due to tendon or muscle destruction, but it usually results from denervation. Therefore, physiologic considerations in upper limb paralysis first include the status of innervation. A scale from 1-5 is used to classify nerve injuries. The classification (and Sneddon classification when applicable), with the anatomical correction (definition) and recovery prognosis, is as follows:

  • Grade 1

    • Seddon classification - Neurapraxia

    • Axonal continuity preserved

    • Fast and complete recovery

  • Grade 2

    • Seddon classification - Axonotmesis

    • Axon injury but endoneurial sheath intact; no fibrosis or structural injury

    • Slow but complete recovery

  • Grade 3

    • Axon injury plus endoneurial sheath destroyed; perineurium intact; variable fibrosis

    • Slow and variable recovery

  • Grade 4

    • Axon injury plus fascicular structure of nerve destroyed/perineurium destroyed; complete fibrosis

    • No recovery

  • Grade 5

    • Sneddon classification - Neurotmesis

    • Loss of continuity of nerve trunk

    • No recovery

Classification of nerve injury according to Sunderland

Paralysis may be temporary, as in grade 1 or 2 injury, or permanent, as in grade 4 or 5 nerve injury. In grade 1 and 2 injuries, no intervention is needed and full recovery is expected.[7]

Denervated muscle usually loses its ability to accept reinnervation or respond to low-amplitude electrical stimulation after 9-18 months. Once the muscle has reached a point that it no longer accepts reinnervation reconstructive efforts, tendon transfers or assistive devices are required. In addition, as denervated muscle becomes fibrotic and contracted, the fibrosis may have a negative effect on the suppleness of muscles and joints. Thus, early treatment while awaiting motor reconstruction should include therapy to maintain supple joints, which is imperative for any tendon transfer. For these reasons, it is important in the acute and subacute care after an injury that produced paralysis to identify the grade 4 and 5 injuries as early as possible so that appropriate intervention can be attempted.

Spinal cord paralysis presents a special consideration because the paralysis may be either upper or lower motor neuron in origin. In paralysis with an intact lower motor neuron, the muscle remains sensitive to low-amplitude stimulation. Because the muscle remains innervated, it remains indefinitely receptive to reinnervation from a nerve transfer. If the paralysis includes lower motor neuron injury, early nerve transfer or long-term tendon transfer is the only reconstructive option other than traditional assistive devices.



The following must be addressed in the patient history:

  • The etiology of paralysis (ie, traumatic [eg, disruption of nerve, muscle, tendon] vs nontraumatic [eg, neurological/rheumatological])

  • The exact mechanism and timing of any injury

  • Previous treatments

  • Patient expectations and compliance

Traumatic and nontraumatic

  • Traumatic: High-energy wounds (eg, gunshot wounds) may produce nerve injuries of widely varying degrees. Some of these can be expected to recover spontaneously. Others are grade 4 and 5 injuries that may span long segments. Central gunshot wounds tend to produce unpredictable patterns of both upper and lower motor neuron injury. Conversely, low-energy injuries (eg, sharp penetrating wounds) produce disruption of the peripheral nerve in short segments, making these appropriate for successful nerve repair. For blunt injuries, the grade of injury is difficult to determine initially.

  • Nontraumatic: A history of rheumatoid arthritis, other synovial arthritis, or gout/pseudogout raises the suggestion of tendon rupture. Rupture secondary to synovial processes almost always requires tendon transfers or grafts. Direct repair is not usually possible because of segmental loss of tendon substance. In otherwise healthy individuals, a recent history of viral illness suggests paralysis of viral origin, such as anterior interosseous nerve palsy. Limited observation may be indicated in these patients.

Patient expectations and compliance

An important part of the history includes determining the expectations of the patient regarding long-term function. The patient's expectations need to be appropriate to the options available. A patient with overly ambitious expectations may be quite unhappy with more modest outcomes. In addition, tendon transfers require a lengthy recovery period and should not be performed in patients who will not or cannot adequately comply with therapy.

Physical examination

The physical examination should be thorough, with particular concern to the status of joints, the status of sensation, soft tissue coverage, which muscles are functioning, and which functions are missing. The authors' preference is to record sensation by dermatome, note soft tissue coverage, and record the suppleness of all involved joints. The authors proceed from proximal to distal with a manual muscle examination.

This starts with examination of the shoulder, with attention to the function of the deltoid (anterior, middle, posterior), latissimus, and pectoralis, and an assessment of shoulder subluxation. Next, test the triceps and biceps group, then the brachioradialis (BR). The examination should assess functions of pronation, supination, wrist extension, and flexion. With wrist extension, the authors attempt to determine whether extension is produced by only the extensor carpi radialis longus (ECRL) or by both the ECRL and the extensor carpi radialis brevis (ECRB). Differentiating between extension produced by the ECRL or both the ECRL and the ECRB may be difficult and ultimately requires examination in the operating room. Because both the flexor carpi radialis and the flexor carpi ulnaris (FCU) may be useful for transfers, determine the function of both.

Next, the function of finger and thumb flexion and extension is checked. This is followed by evaluation of the intrinsic muscle function of the hand. Of particular concern are opposition of the thumb and the presence of "clawing" of any fingers, as shown below.

Ulnar clawing produced by loss of intrinsics to th Ulnar clawing produced by loss of intrinsics to the little and ring fingers, characterized by hyperextension of the metacarpophalangeal joints and flexion of the interphalangeal joints.

After identifying lost function, the examiner should assess that joints are supple and passively unrestricted in movement. Only under this premise is tendon transfer possible.

Finally, the donor muscles and tendons must be checked for presence and function. The palmaris longus tendon can be felt in wrist flexion with opposition of the thumb and little finger. The extensor indicis proprius tendon usually lies ulnarly of the extensor digitorum communis (EDC)–2 tendon. It is present if the index finger can be extended with complete flexion of the other fingers.

A thorough examination may also help to determine the etiology of the paralysis. This is especially important in nontraumatic cases, in which differentiation between nerve paralysis and tendon rupture is important. Testing for a natural tenodesis effect helps establish the cause (see below). For finger paralysis, the wrist is extended and flexed. Tenodesis movement of the fingers implies a neurologic problem, while loss of tenodesis movement implies tendon rupture.

Natural tenodesis is demonstrated by flexing and e Natural tenodesis is demonstrated by flexing and extending the wrist with the hand relaxed. The effect shows extension of the fingers when the wrist is flexed and flexion of the fingers when the wrist is extended. If the tendons are not intact, this effect is lost.


Assessing the degree of nerve injury and the chance of recovery is important for adequate early treatment. An electrodiagnostic evaluation of peripheral nerve injury is used for this purpose.

Nerve conduction studies, both sensory and motor, are useful early after an injury to distinguish neurapraxia from axonotmesis or neurotmesis. The hallmark of a neurapraxic injury is a slowing or complete block of nerve conduction along a segment of nerve. The absence of a nerve conduction response in a distal nerve segment indicates a loss of axons (axonotmesis or neurotmesis). Serial nerve conduction studies show a progressive diminution in the amplitude of the response after 48-72 hours as wallerian degeneration of the axons occurs.

Two to 3 weeks after an axonotmetic or neurotmetic injury, the results of electromyography (EMG) become abnormal. Spontaneous activity, including fibrillations, fasciculations, and positive sharp waves, develops in the affected muscles. The results of EMG performed after a neurapraxic injury usually remain normal and can help distinguish a neurapraxic injury from the more severe axonotmetic and neurotmetic injury. EMG can help detect regeneration only at the stage during which nerve fibers begin to reinnervate their target muscles. In clinical practice, serial EMG at 6- to 12-week intervals helps identify recovering muscle. If evidence of reinnervation is not present at 3 months, especially in muscles close to the level of injury, spontaneous recovery is unlikely.

Clinically, nerve regeneration can be assessed by an advancing Tinel sign. Evidence of muscle reinnervation revealed by EMG may occur several weeks before clinically apparent voluntary muscle contraction is revealed by physical examination. Over time, muscle regeneration and function is classified into the following grades of muscle strength:

  • Grade 5 - Normal motor power

  • Grade 4++ - Able to overcome gravity and significant resistance, but strength not quite normal

  • Grade 4+ - Able to overcome gravity and moderate resistance

  • Grade 4 - Able to overcome gravity and mild resistance

  • Grade 3 - Able to overcome gravity but not resistance

  • Grade 2 - Movement in the plane of the support extremity, unable to overcome gravity

  • Grade 1 - Flicker movements only

  • Grade 0 - Total paralysis


When considering tendon transfers, the primary indication for surgery is the need to replace or reconstruct an absent function. To reconstruct a function, certain other requirements must be met. Joints that are required to move must be supple. A transferred tendon trying to move a stiff joint is doomed to failure. If joints are not sufficiently supple, they must be loosened, preferably by hand therapy.

If hand therapy fails, surgical release of the joints may be required prior to tendon transfer. Use passive motion exercises to maintain joint flexibility after surgical release. Allow sufficient time to allow all surgically induced tissue trauma adequately heal. The author usually waits at least 3 months, but preferably 6 months, before proceeding with transfers.

If sensation in the hand is absent, strongly consider restoring at least protective sensation prior to or at the time of tendon transfer. Even with normal motor function, an anesthetic hand is difficult to use; in a hand with transferred tendons, it may be impossible.

To restore functional motion to a hand or forearm, a suitable donor must be available. Generally, a suitable donor is a normally innervated muscle that is providing a redundant function. Redundancy is not absolutely necessary provided that the function to be sacrificed is not of vital importance to the overall function of the extremity. Except in unusual circumstances, a reinnervated muscle is considered a poor choice for a donor.

The following are functions in the upper extremity and the tendons commonly used for reconstruction:

  • Intrinsic balance

    • Flexor digitorum superficialis (FDS) for dynamic Zancolli lasso procedure (dynamic)[5]

    • Palmar plate for Zancolli capsulorrhaphies procedure (static)

  • Thumb opposition

    • Extensor indicis proprius

    • Flexor digitorum superficialis

    • Abductor digiti minimi

  • Thumb flexion

    • Pronator teres

    • Brachioradialis

    • Flexor digitorum superficialis

  • Thumb extension

    • Brachioradialis

    • Extensor indicis proprius

    • Palmaris longus

  • Finger flexion

    • Brachioradialis

    • Extensor carpi radialis longus

    • Adjacent profundus

  • Finger extension

    • Brachioradialis

    • Flexor carpi ulnaris

    • Flexor carpi radialis

    • Adjacent finger extensor

    • Extensor indicis proprius

  • Wrist extension

    • Brachioradialis

    • Pronator teres

  • Wrist flexion - Rarely reconstructed

  • Elbow extension

    • Posterior deltoid

    • Biceps

  • Elbow flexion

    • Pectoralis major

    • Triceps

    • Latissimus

    • Forearm flexor mass (Steindler)

The following are donor muscles and common recipients:

  • Brachioradialis

    • Extensor carpi radialis brevis

    • Flexor digitorum profundus

    • Flexor pollicis longus[8]

    • Extensor digitorum communis

    • Extensor pollicis longus

  • Extensor carpi radialis longus - Flexor digitorum profundus

  • Pronator teres

    • Extensor carpi radialis brevis

    • Extensor pollicis longus

  • Flexor carpi ulnaris - Extensor digitorum communis

  • Flexor carpi radialis

    • Extensor digitorum communis

    • Extensor pollicis longus

  • Extensor indicis proprius

    • Extensor pollicis longus

    • Opponens pollicis

  • Palmaris longus - Extensor pollicis longus

  • Flexor digitorum superficialis

    • Opponens/abductor pollicis[9]

    • Flexor digitorum profundus

    • A-1 pulley tenodesis

  • Abductor digiti minimi - Opponens pollicis

  • Biceps

    • Reroute/reinsert

    • Triceps

  • Triceps - Biceps

  • Posterior deltoid - Triceps

  • Pectoralis major - Biceps

  • Latissimus - Biceps


The timing of a tendon transfer after an injury depends on the likelihood of spontaneous reinnervation and nerve recovery. If nerve repairs or nerve transfers were performed initially, then sufficient time should be allowed to determine the outcome of the initial treatment before considering tendon transfers.[10] Keep in mind that axons regenerate at a rate of approximately 1 mm/d.

If one cannot determine from the initial injury whether the nerve was interrupted (neurotmesis) and if the clinical examination reveals a loss of motor or sensory function, determining if adequate recovery is likely is mandatory before considering tendon transfer as a reconstructive option. EMG performed immediately and then again at 6 weeks helps to determine which functions may be expected to recover. Lack of evidence of innervation at 6 weeks should prompt exploration and repair if possible. Once sufficient time has elapsed to allow for spontaneous or repaired recovery, consider reconstruction for missing functions.

Of note, some hand surgeons advocate early tendon transfers, particularly in patients with radial nerve palsies, even if recovery is still possible. In 1974, Burkhalter reported that the indications are (1) the transfer can act as a substitute during regrowth of the nerve, which will thereby reduce the time of external splinting and improve early function; (2) the transfer can act as a helper and add power to normal reinnervated muscle function; and (3) the transfer can act as a substitute when, statistically, the recovery after neurorrhaphy or nerve repair is poor.[1]

Relevant Anatomy

Select donor muscles that provide a synergistic action to the function to be restored. Wrist extensors and finger flexors are an example of a synergistic group. In normal activity, wrist extensors contract when finger flexion is initiated. This action stabilizes the wrist, maximizing the effect of the finger flexors. This makes wrist extensors good choices for transfer to provide finger flexion. The wrist extensor normally contracts simultaneously with finger flexors; this makes postoperative motor reeducation simpler.

Other synergistic muscle groups include (1) wrist flexors and thumb/finger extensors and (2) the extensor indicis proprius (EIP) and extensor pollicis longus (EPL).

On the contrary, muscles that normally relax while another is active are more difficult for a patient to learn to control. These are considered nonsynergistic, or out of phase. An example would be the use of a wrist flexor to provide finger flexion or a wrist extensor to provide finger extension.

Muscles that have no direct functional relationship to a particular function may provide excellent donors. This is true of the BR, which is often chosen to replace finger or thumb flexion, wrist extension, or finger or thumb extension. The strength, excursion, and usual redundancy of the brachioradialis (BR) make it a frequent choice as a donor for multiple functions. The pronator teres (PT) is a good replacement for wrist extension.

In addition to considering appropriate donors, pay attention to matching, as closely as possible, a donor's excursion and work capacity with the function to be replaced. Work capacity generally corresponds to the cross-sectional area of the muscle, while excursion is related to muscle fiber length.

Transferring a muscle-tendon unit that does not need to be extended by a graft is desirable, but not absolutely necessary. In cases in which an appropriate donor cannot reach the desired recipient, an interposed tendon graft is acceptable.

Donors should be used to provide only one function and, when possible, should cross only one joint. Transfers that cross 2 joints and are expected to produce motion in both add a level of demand to the restoration of function that may not be achievable. If traversing more than one joint is necessary, appropriate stabilization of one of the joints is desirable to maximize function. For instance, triceps function is desirable when using the BR transfer in which stabilizing the elbow in extension improves the transfer function. Similarly, expecting a transferred muscle-tendon unit to provide more than one function produces disappointing results.


The only absolute contraindication to tendon transfer is a lack of appropriate donors. The availability of muscle-tendon units with less than grade 5 strength is a relative contraindication. Similarly, if only muscles that have been denervated and then reinnervated are available, this is also a relative contraindication. Transfers planned in individuals with progressive neuromuscular diseases should be carefully considered before proceeding because the underlying disease process may affect the transferred unit. Lastly, satisfactory results are difficult to achieve in transfers performed to produce motion in less-than-supple joints.



Laboratory Studies

See the list below:

  • The workup of candidates for tendon transfers consists mostly of the physical examination to determine available muscles for transfer and function to be restored.

  • The only laboratory studies needed are those that ensure the safety of the patient during surgery.

Imaging Studies

See the list below:

  • Plain radiographs of joints that are stiff may be necessary to ensure that obtaining a supple joint is possible. Severe arthritic conditions in joints involved with the function to be restored represent a relative contraindication to proceeding with the procedure.

Other Tests

See the list below:

  • When loss of function is secondary to neurologic or peripheral nerve injury, nerve conduction studies and electromyography (EMG) can help differentiate permanent injury from recovering injury. EMG denervation patterns are evident early, to be followed by evidence of reinnervation.



Medical Therapy

Certain medical therapies may be effective for neurologic diseases and may lead to long-term recovery from the paralysis. In spinal cord injuries, activity-dependent therapies are thought to improve outcomes by improving the level of injury. In all cases, while awaiting stabilization of the injury or recovery of function, maintaining supple joints is necessary. Functional recovery or future reconstructive efforts will not be successful if joints are stiff.

Supple joints can be secured most easily by not letting them stiffen initially. Early on, patients should be referred to a hand therapist, with a specific request to maintain supple joints with active and passive motion programs. Obviously, most of the efforts involve passive methods and most require active patient participation with a home therapy program.

Alternatives to surgical reconstruction consist of modified tools, splints, or prostheses (see below).

Assist devices used in upper limb paralysis. Assist devices used in upper limb paralysis.

A study by Van Heest et al indicated that in children with upper extremity cerebral palsy, tendon transfer provides more functional improvement than does botulinum toxin injection or regular, ongoing therapy. The study involved 39 children with upper extremity cerebral palsy who underwent one of the three treatments, with the surgery group undergoing transfer of the flexor carpi ulnaris to the extensor carpi radialis brevis, along with release of the pronator teres and rerouting of the extensor pollicis longus with adductor pollicis release. At 12-month follow-up, the surgical patients showed greater improvement than the others as measured using the Shriners Hospital for Children Upper Extremity Evaluation dynamic positional analysis score, as well as the Canadian Occupational Performance Measure score for satisfaction and the Pediatric Quality of Life Inventory cerebral palsy module domain of movement score.[11]

Surgical Therapy

Non-nerve-related transfers

These include treatment for rheumatological conditions (eg, gout, pseudogout) and traumatic injury to muscles and tendons.

Tendon transfers for non–nerve-related loss of function tend to be some of the most simple and straightforward. The most frequently encountered of these is replacement of ruptured tendons as treatment for arthritic conditions. Rupture of extensor tendons of fingers or thumb is commonly associated with rheumatoid arthritis. (For more information, visit Medscape’s Rheumatoid Arthritis Resource Center.)

The extensor pollicis longus (EPL) tendon is at risk of rupture as a complication of distal radius fracture. Before planning early transfers for this category of injury, the surgeon should be sure that the loss of function is indeed from disruption of the tendon and not from an isolated nerve palsy such as an anterior or posterior interosseus nerve palsy. Loss of a tenodesis effect usually confirms that the problem resides with the tendon (see below).

Natural tenodesis is demonstrated by flexing and e Natural tenodesis is demonstrated by flexing and extending the wrist with the hand relaxed. The effect shows extension of the fingers when the wrist is flexed and flexion of the fingers when the wrist is extended. If the tendons are not intact, this effect is lost.

Replacement of the EPL function is usually performed with a transfer of the extensor indicis proprius (EIP) (see below). A tendon graft to replace extensor function may be an option and equal results of EIP transfer and intercalated free tendon graft have been demonstrated by Schaller et al.[12] However, if the rupture is of longstanding duration, a transfer is the most reasonable option.

Extensor indicis proprius transfer to the extensor Extensor indicis proprius transfer to the extensor pollicis longus.

While a tendon graft to replace extensor function may be an option if the rupture is not of longstanding duration, these patients often do not seek treatment until months after the rupture. Under these circumstances, a transfer is the most reasonable option.

The EIP has redundant function with the extensor digitorum communis (EDC) and, thus, is expendable. Before transfer of the EIP, especially in a patient with rheumatoid arthritis, the surgeon should be confident that the EDC has not been affected by the disease and is functioning normally.

For finger extensor tendons, a side-by-side transfer to an adjacent extensor tendon may provide adequate function and is simple to perform (see below). The authors' practice is to inspect the functioning tendons throughout their course in the dorsal compartments. This allows recognition and correction of any potential problem that may lead to further rupture of tendons. Correction of bony problems may require work on the carpus, distal ulna, or distal radius.

Tendons ruptured secondary to rheumatoid arthritis Tendons ruptured secondary to rheumatoid arthritis. Repair by transfer of long extensor digitorum communis (EDC) to the ring EDC and the index EDC to the little EDC. The fourth dorsal compartment has been opened to allow inspection of the compartment and repair of any problem causing the ruptures.

While extensor tendons tend to be more commonly involved with spontaneous ruptures, flexor tendons may also suffer a similar fate, although not as frequently. Tendon grafts are often the best solution when trying to reconstruct flexor function. When tendon grafts are impractical or impossible, transfers may be used. Spontaneous rupture of flexor tendons occurs most often in the palm or carpal canal. This precludes the use of wrist extensors for reconstruction unless the transfer is extended with a graft. Fortunately, surgeons can usually sacrifice a superficialis tendon to transfer to an adjacent profundus to restore flexion of the entire digit.

Median nerve paralysis

High median nerve lesions lead to loss of pronation, profundus function of the index finger, superficialis function to all fingers, flexor function of the thumb, and opposition.[13, 14] Profundus function to the ring finger may also be weak. Various muscles may be transferred to restore finger and thumb flexion (see Table 1).[15] The authors prefer the brachioradialis (BR) for thumb flexion.

Table 1. Recommended Transfers for a High Median Nerve Palsy (Open Table in a new window)




Flexor pollicis longus

Extensor carpi radialis longus

Flexor digitorum profundus

EIP opponensplasty

The extensor carpi radialis longus (ECRL) can be used to restore index and long profundus function. Specific restoration of superficialis function is not necessary. An EIP opponensplasty provides excellent and natural opposition (see below). If loss of pronation is a problem, the biceps tendon can be rerouted at its insertion to achieve modest improvement of function. Usually, this is not necessary.

Opponensplasty using the extensor indicis proprius Opponensplasty using the extensor indicis proprius for treatment of low median nerve palsy.

Anterior interosseous palsy leads to loss of thumb flexion, loss of index profundus function, and weakness in long finger profundus function. Opposition and pronation remain intact. An ECRL transfer to the profundus restores index profundus function, and a BR to flexor pollicis longus (FPL) transfer restores thumb flexion (see Table 2).

Table 2. Tendon Transfers for an Anterior Interosseous Nerve Palsy (Open Table in a new window)




Flexor pollicis longus

Extensor carpi radialis longus

Flexor digitorum profundus

In low median nerve lesions, the only motor loss is opposition. As in the high lesion, an EIP opponensplasty adequately restores the lost opposition.[16]

Ulnar nerve paralysis

High ulnar nerve lesions leave deficits in ulnar intrinsic function of the hand, flexor carpi ulnaris (FCU), and profundus function to the little and ring fingers.[13] If the flexor carpi radialis remains functional, the FCU does not need to be restored. Transfer of the BR to the flexor digitorum profundus (FDP) or a side-to-side profundus transfer (see Table 3) can restore profundus function. Loss of the ulnar-innervated intrinsic functions leads to clawing of the little and ring fingers, as shown below.

Ulnar clawing produced by loss of intrinsics to th Ulnar clawing produced by loss of intrinsics to the little and ring fingers, characterized by hyperextension of the metacarpophalangeal joints and flexion of the interphalangeal joints.

Table 3. Tendon Transfers for a High Ulnar Nerve Palsy (Open Table in a new window)




Flexor digitorum profundus

Flexor digitorum superficialis

Adductor pollicis

Flexor digitorum superficialis lasso ring and little

Blocking hyperextension of the metacarpophalangeal (MP) joints of the little and ring fingers can allow the extrinsic extensor to extend the interphalangeal (IP) joints of the little and ring fingers. The preferred transfer is a lasso procedure, which involves tying a slip of the superficialis tendon around the A-1 or A-2 pulley of the flexor tendon sheath, as shown below.

Zancolli-type lasso of A-1 pulley by the superfici Zancolli-type lasso of A-1 pulley by the superficialis.

In lower ulnar nerve lesions, restoration efforts are essentially the same as for high lesions except the profundus does not need attention.

In both high and low lesions, restoring an adductor of the thumb is occasionally necessary or desirable. This is accomplished with a superficialis transfer, usually from the long finger in high lesions and from the ring finger in low lesions, to the distal thumb metacarpal.

Radial nerve paralysis

High radial nerve lesions produce loss of finger, thumb, and wrist extension.[17] Supination may be weak but is usually adequately provided by the biceps function.[13] Of particular note in high lesions is the loss of BR and ECRL function. Loss of the BR eliminates an excellent donor, and loss of the ECRL imposes the need for wrist extension restoration.

Historically, various transfer combinations have been attempted. These may be chosen according to the experience and preference of the surgeon.[18] The authors prefer pronator teres (PT) to extensor carpi radialis brevis (ECRB) transfer for wrist extension, FCU to EDC transfer for finger extension, and a palmaris longus to a rerouted EPL transfer for thumb extension (see Table 4). If the palmaris is absent, then the ring superficialis can be used for the thumb extensor. However, other alternatives are available, and preferences vary from surgeon to surgeon. In 2006, Ropars et al recommended using the FCR instead of the FCU for restoration of finger extension.[19]

Table 4. Tendon Transfers for a High Radial Nerve Palsy (Open Table in a new window)



Pronator teres

Extensor carpi radialis brevis

Flexor carpi ulnaris

Extensor digitorum communis

Palmaris longus

Extensor pollicis longus (rerouted)

For lower lesions involving only the posterior interosseus nerve, ECRL function is usually adequate to provide wrist extension. If wrist extension is accompanied by severe radial deviation, side-to-side connection of the ECRL and ECRB may achieve a more centralized extension. Finger extension is again restored with FCU transfer to the EDC, and thumb extension can be restored with a BR transfer to the EPL (see Table 5). Some surgeons recommend the restoration of thumb abduction by a tenodesis of the abductor pollicis longus to the BR.[19]

Table 5. Tendon Transfers for a Low Radial Nerve Palsy or Posterior Interosseous Nerve Palsy (Open Table in a new window)




Extensor pollicis longus

Flexor carpi ulnaris

Extensor digitorum communis

Mixed peripheral nerve paralysis/brachial plexus paralysis/spinal cord paralysis

With mixed paralysis, such as may occur with multiple nerve involvement or more central lesions such as those involving the brachial plexus[20] or spinal cord, the authors believe drawing a table indicating which functions are present, which are absent, and which are in need of reconstruction is most convenient. The table also lists available muscles along with their function. A glance at the table, once filled out, shows which muscle may be transferred to provide the various functions needed (see Table 6).

Table 6. Muscles Available for Transfer (Open Table in a new window)






Muscles Available

Elbow flexion












Elbow extension




Forearm pronation



Pronator teres




Pronator quadratus

Forearm supination




Wrist flexion



Flexor carpi radialis




Flexor carpi ulnaris




Palmaris longus

Wrist extension



Extensor carpi radialis longus




Extensor carpi radialis brevis




Extensor carpi ulnaris

Finger flexion



Flexor digitorum profundus




Flexor digitorum superficialis

Finger extension



Extensor digitorum communis




Extensor indicis proprius




Abductor digiti minimi

Thumb flexion



Flexor pollicis longus

Thumb extension



Extensor pollicis longus

Thumb opposition



Abductor pollicis brevis

Thumb adduction



Abductor pollicis

Name and other procedures:

Remember that with more extensive loss of function, restoration of all desirable motions with active transfers may not be possible. Under circumstances that allow for minimal active function, the judicious use of tenodeses helps restore some useful function to the limb even though that function may be modest. A good example of this is found in the patient with only BR function below the elbow. In this case, restoration of flexion and extension of the thumb, finger, and wrist is desirable. However, with only one muscle available for transfer, active restoration of all of these functions is not possible.

A reasonable compromise consists of transferring the BR to the ECRB to provide active wrist extension. Then, perform tenodeses of the FPL, FDP, EPL, and EDC. These tenodeses augment a natural tenodesis with wrist extension and flexion. The BR provides active wrist extension, producing a pinching motion of the fingers and thumb. Release is provided by wrist flexion initiated by gravity. The entire reconstruction is performed in 2 procedures. The BR transfer and extensor tenodesis can be performed in one session, and the flexor tenodesis can be performed at another.

With fewer active motors available, those that are retained are called upon to produce more of a mass movement than the isolated individual digit movements that normally occur. For example, the transferred BR should extend the index, long, ring, and little fingers. While such a requirement is not ideal, the dearth of other choices may make it reasonable. In such cases, balancing the extrinsic tendon is necessary to provide optimum motion to each digit. This is performed via a side-to-side connection of the profundus or extensor tendons, as appropriate (see below). Tendons are positioned and connected to produce uniform motion in all digits from a single motor.

Side-by-side adjustments of the profundus or exten Side-by-side adjustments of the profundus or extensor tendons for extrinsic balance.

Also possible is performing one transfer to overpower a present but weak function. This is common with transfers to the FPL in the presence of weak or absent EPL function and paralysis of thumb intrinsics. This produces a hyperflexion of the IP joint of the thumb unless it is stabilized. The IP joint can be stabilized either by arthrodesis or a stabilizing transfer. The authors prefer a stabilizing transfer, in this case a split FPL-to-EPL transfer (see below).

Split flexor pollicis longus transfer for stabiliz Split flexor pollicis longus transfer for stabilization of the interphalangeal joint.

With significant paralysis, one can be tempted to perform a wrist arthrodesis to "free up" wrist flexors and/or extensors for transfer. While this procedure may occasionally be useful, experience has shown that if wrist motion can be preserved and made active, better function is obtained by letting the wrist motion augment active transfer with a tenodesis effect.

When considering transfer of the ECRL, ensure that the ECRB is of adequate strength to provide wrist extension. Clinically verifying that the ECRB is active or estimating its strength is difficult, even when one is sure that it is functioning. The best way to evaluate ECRB function is by performing a diagnostic procedure in the operating room.

Prior to providing a general or regional anesthetic but with the patient under local anesthesia, the ECRB is exposed distally and isolated from the ECRL. A suture is placed in the ECRB and attached to a strain gauge, weight, or elastic. The goal is to ensure the ECRB is capable of lifting 5 kg. The patient is requested to extend the wrist. If the ECRB is able to lift 5 kg, it is assumed to be adequate, allowing the ECRL to be used as a donor.

Reconstructions that require both flexor and extensor transfers must be staged. The first stage can be either flexor or extensor, although some experienced surgeons have recommended performing extensor transfer first because rehabilitation may be easier. However, the most dramatic improvement of function is from flexor transfer, leading some surgeons to start with flexors to maintain patient enthusiasm for the efforts.

A frequent problem in brachial plexus injury is loss of elbow flexion. This is often associated with severe paralysis of the entire upper extremity, resulting in few local donors. If the pectoralis major or latissimus is functioning, one of these may be transferred to the biceps distally to produce elbow flexion. In addition, forearm muscle function may remain intact with loss of proximal muscles. If the wrist and finger flexors remain strong, the patient may be a candidate for a Steindler flexorplasty. In this procedure, the flexor mass is detached at the medial epicondyle and moved proximally to attach to the distal humerus.

When desirable, elbow extension can be reconstructed with a posterior deltoid transfer to the triceps. This is a commonly required transfer in spinal cord injury.

Intraoperative Details

Once the appropriate transfers have been chosen, obtain surgical exposure to allow access to both the donor and recipient sites. Incisions tend to be longitudinal and centered to allow for appropriate exposure.

The donor should be harvested to maximize the length of the tendon, occasionally even extending the tendon with periosteum. The PT is often harvested with a strip of periosteum to allow it to reach its donor without requiring extension with a graft. One exception to this maximizing effort is when harvesting a flexor digitorum superficialis (FDS) tendon. A distal stump is left to help stabilize the proximal IP joint.

The donor muscle should also be completely freed from fascial attachments. This requires an extensive dissection about the muscle. Exercise care when performing this portion of the operation to avoid harming the muscle's nerve and vascular supply. Freeing fascial soft tissue attachments increases the overall excursion that can be obtained. Conversely, if the muscle is inadequately mobilized, poor excursion, thus poor function, is the result of the transfer.

It is vital for the transferred tendon to obtain the adequate tension during inset.[21] The 3 options are as follows:

  • The best method is a neuromuscular stimulator, which is useful to evaluate the excursion after dissection and, ultimately, the function of the transfer. The stimulator should provide a biphasic current of approximately 20 mA at 20 Hz with variable pulse width. Pulse widths of 100-200 milliseconds applied to the neuromuscular junction generally maximize the contraction of the muscle (see below). Active excursions of 5 cm or more are preferred.

    The neuromuscular stimulator is useful for intraop The neuromuscular stimulator is useful for intraoperative evaluation of tendon excursion and transfer function.
  • Most surgeons probably use the clinical assessment and the passive movements of the involved joints to judge the tension.

  • Measurements and markings of passive tension in the transferred muscle (eg, every 5 cm) can be used to find the correct tension after transfer. This method is commonly used in free muscle transfers.

Once the donor tendon is dissected, it is routed to the recipient. Routing is usually performed using the most direct route. When possible, routing the tendon around the forearm, rather than through the interosseus membrane, is better. Routing through the membrane often produces adhesions, minimizing active function. Otherwise, the most direct route should be taken. The authors prefer to use a Pulvertaft weave to connect a donor to a recipient (see below).

Pulvertaft weave used to connect tendon grafts. Pulvertaft weave used to connect tendon grafts.

A neuromuscular stimulator is useful in setting the tension of the transfer. Prior to completing the Pulvertaft weave, a preliminary attachment is made by attaching the donor to the recipient, with the donor in the midpoint of its total excursion and the recipient joint in neutral or the midpoint of its total functional arc. The donor is then maximally stimulated. This demonstrates the postoperative function that can be expected.

Final adjustments of tension are made based on the function observed during this stimulation, as shown below. When using a stimulator, the extremity should be warm and adequately perfused. Avoid too much tension because excessive tension on a muscle reduces its active function. Obviously, tension that is too loose will not produce the desired active motion.

Test excursion and transfer function using a neuro Test excursion and transfer function using a neuromuscular stimulator.

The choice of whether to perform an end-to-end or end-to-side connection should be based on the difference in expected function. If return of functional activity in the recipient is possible, choose an end-to-side connection.

The posterior deltoid does not reach the triceps tendon and is extended with a graft from the tibialis anterior or fascia lata. The authors prefer grafting by disinserting the central third of the triceps from the olecranon and reflecting it proximally to attach to the deltoid. The medial and lateral thirds of the triceps are reconstituted or repaired. The posterior deltoid is elevated with periosteum and split from the middle and anterior deltoid. Care is taken to avoid injury to the traversing nerve supply in the proximal third of the deltoid. The connection between the deltoid and triceps is reinforced with Dacron tape.

Postoperative Details

In general, joints are immobilized in the position of the function that is being reconstructed. Details of the immobilization, such as the degree and length of time, vary. Interestingly, studies are increasingly undertaken to compare early active motion regimens with passive motion regimens or cast immobilization. These studies show promising results.[22] Rath, for example, showed similar outcomes in a reduced recovery time by immediate active mobilization of opposition tendon transfer.[23]

Transfers for flexion

With flexor tendon transfers, the patient is placed in an immobilizing splint appropriate to protect tendon connections. This places the wrist in 20-30° of flexion and MP joints at 60-90° of flexion. IP joints can be allowed to extend. The patient is left in the initial dressing until the first postoperative visit.

Transfers for extension

With extensor tendon transfers, the patient is splinted with the wrist in 30° of extension and the MP joint extended. IP joints are allowed to flex. As with flexor transfers, the initial dressing is left in place until the first postoperative visit, which should be within 1 week.

For split FPL transfers to stabilize the IP joint of the thumb, the IP joint is splinted in extension. If not precluded by other concomitant transfers, splinting the wrist in 20° of flexion following this procedure also may be useful.

When performing a posterior deltoid transfer to the triceps, the elbow is splinted fully extended and the arm is abducted 30°. Abduction is usually maintained with an axillary wedge.

Elbow flexion reconstruction is splinted with the elbow flexed 90-100°.


Follow up with patients within 1 week. The patient is usually referred to a hand therapist at the first postoperative visit. The importance of a good, experienced hand therapist in the ultimate outcome of the procedure cannot be overemphasized.

Extensor transfers are generally managed by simple immobilization with the wrist and MP joints extended for approximately 4 weeks. After 4 weeks, the immobilization is discontinued and active and passive exercises are begun. At this point, the therapist initiates motor reeducation protocols. These include cognitive exercises, biofeedback, and task-oriented activity. Motor reeducation continues from 1-3 months.

The extensor indicis transfer to reconstruct the EPL presents a different scenario because it is performed using a Pulvertaft suturing technique, which is more stable than any other suture technique. An early dynamic motion protocol (reversed Washington regimen) can be applied for 3 weeks, with increasing active flexion (ie, 45°, then 60°, then unlimited flexion for the first, second, and third wk, respectively), which shortens the total rehabilitation time.[24]

Flexor transfer may be treated with a protocol of 4 weeks of immobilization (as with extensor transfers) or with an early motion protocol. Early motion usually consists of passive motion, with active motion efforts delayed until approximately 4 weeks postoperatively. During the fourth postoperative week, efforts at motor reeducation are started.

Posterior deltoid transfers present a special case. These patients are immobilized in extension and abduction for 4 weeks. After the fourth week, the patient is placed in an adjustable Bledsoe splint, and, at weekly intervals, flexion is increased by 15°.


Complications of tendon transfer surgery are typical of any surgery in the upper extremity. These include the relatively small potential for wound infections, blood loss or hematoma, injury to nerves, or injury to tendons. Because a tendon connection has been made, rupture of the tendon connection can occur. Most of the connections are performed with Pulvertaft weaves. Because these are quite strong, ruptured connections are infrequent. However, even a Pulvertaft weave can be pulled apart as a result of improper splinting or overactivity.

Of more concern is the possibility that the transfer does not function as desired because of too-loose or too-tight tension. The use of intraoperative stimulation to evaluate the transfer and tension helps to minimize disappointing postoperative results.

Function may also be reduced by strong tendon adhesions that reduce excursion. Appropriate aggressive therapy reduces the occurrence of clinically important adhesions, but some individuals may not be able to overcome strong adhesion formation. Use of anti-inflammatory drugs in the postoperative period may also reduce adhesion formation. Tenolysis after 3-6 months may be necessary to maximize function. After tenolysis, aggressive use of active and passive motion protocols is necessary.

Undesirable consequences of donor sacrifice are another concern. One example is loss of wrist extension after transfer of the ECRL. Careful planning and appropriate preoperative and intraoperative evaluation reduce unexpected donor consequences.

Despite appropriate planning, execution, and postoperative care, some patients are not able to develop satisfactory motor reeducation. These patients may continue to benefit from a tenodesis effect of the transfers, but they do not achieve its maximum benefit.

Outcome and Prognosis

Depending on the specifics of each transfer, such as the amount and quality of intact function, the choice of available donors, the suppleness of joints, and the presence of intact sensation, outcomes can vary tremendously. When function is lost from ruptured tendons or neuromuscular paralysis, tendon transfers generally provide enough improvement in function to warrant their recommendation (see images below).

Opponensplasty using the extensor indicis proprius Opponensplasty using the extensor indicis proprius for treatment of low median nerve palsy.
Flexor reconstruction in brachial plexus injury us Flexor reconstruction in brachial plexus injury using a brachioradialis to flexor digitorum profundus transfer and a pronator teres to flexor pollicis longus transfer along with extensor tenodesis for release.
Flexor reconstruction in a tetraplegic patient usi Flexor reconstruction in a tetraplegic patient using a extensor carpi radialis longus to flexor digitorum profundus transfer and a pronator teres to flexor pollicis longus transfer along with a split flexor pollicis longus transfer to stabilize the interphalangeal joint.

Continued strengthening and improved motor reeducation may result in functional gains and continued improvement for up to a year after the surgery. However, in most patients, function changes little after 6 months.

Outcomes are substantially influenced by postoperative therapy and by motivation and expectations. Without good therapy and patient motivation, good results are difficult to obtain.

Future and Controversies

Tendon transfers are very useful in restoring function but, at best, are a compromise. Loss of donor function always occurs, and the restored function is rarely as complete or as strong as natural function. Better nerve injury management in some cases may make transfers unnecessary. Better management of joints in the postinjury period may improve restored function, even when transfers must be relied upon.

In patients with spinal cord injuries or in patients with primary upper motor neuron paralysis, an implantable electronic neuroprosthesis restores some hand grasp and pinch function.[25] In tetraplegic patients, this system, which does not primarily rely on tendon transfers, has substantially improved patient quality of life and independence. In some tetraplegic patients in whom no donor is available, it may be the only hope of restoring upper extremity function. When first introduced, the system required insertion through large incisions, which could be changed to a percutaneous electrode placement technique.[26] While the neuroprosthesis in these patients becomes the primary motion system, tendon transfers may still be desirable to maximize function with and without the use of the prosthesis.

In high peripheral nerve lesions, the time necessary for nerve regeneration, even in a perfect nerve repair, may exceed the window of opportunity during which the target muscle is receptive to reinnervation. Current laboratory work suggests stem cells may be used as "baby sitters" during the time of nerve regeneration. In this scheme, motoneuron stem cells are placed in the distal peripheral nerve adjacent to the target muscle. The nerve injury can be repaired using current methods of nerve repair. The stem cells have been shown to innervate the target muscle. If these cells could be engineered by inserting an apoptosis gene to die on command, one could simply wait the appropriate amount of time for the regenerating nerve to reach the vicinity of the muscle, then eliminate the stem cell population. The regenerating nerve may even be able to form a synapse with the stem cell, thus not requiring its elimination. This scheme, used successfully, would eliminate some need for tendon transfer.

The same idea of prolonging the period a muscle remains receptive to reinnervation has prompted the use of direct current stimulation of muscle. Preliminary studies suggest that long-term stimulation of denervated muscle preserves its mass and receptiveness.

Free muscle transfers or island transfers may also be useful in restoring function in the upper extremity. For the last 2-3 decades, reports have described gracilis transfer to restore flexion or extension in the hand. The use of bipolar transfers, such as the latissimus, has also been reported to restore flexion or extension of fingers.