At the end of the 19th century, physicians first realized that transferring tendons could restore function to an extremity. The crippling results of the polio epidemic in Europe contributed to the advancement of tendon transfers. In addition, as anesthesia and aseptic techniques developed, the skills and technical acumen of surgeons improved.
With the contributions from such masters as Drs Leo Mayer,  Sterling Bunnell,  Guy Pulvertaft,  and Joseph Boyes,  tendon transfer surgery expanded not only to those with polio and cerebral palsy, but also to those who required reconstructive surgery for traumatic injuries that were incurred during World War I. The fundamentals of tendon transfer surgery were discovered, and the field of reconstructive tendon surgery was established. (See the image below.)
Since the tendon transfers that were performed during World War I, reconstructive tendon surgery has advanced significantly. As understanding of hand biomechanics and tissue healing continue to evolve, tendon transfer surgery will expand to new applications and dimensions. [5, 6, 7, 8, 9, 10]
For patient education resources, see Ruptured Tendon.
At the palm, the flexor digitorum superficialis (FDS) tendon lies volar to the profundus tendon. It then splits at the level of the proximal phalanx and reunites dorsal to the profundus tendon to insert in the middle phalanx. The flexor digitorum profundus (FDP) perforates the superficialis tendon to insert at the distal phalanx.
The relationship of flexor tendons to the wrist joint, metacarpophalangeal (MCP) joint, and interphalangeal (IP) joint is maintained by a retinacular or pulley system that prevents the bowstringing effect. For more information about the relevant anatomy, see Hand Anatomy, Flexor Tendon Anatomy, and Wrist Joint Anatomy.
The basic concept to remember in tendon transfer surgery, as advocated by Brand, is achieving balance in the extremity.  Balance surpasses strength; one must strive to achieve equality in the distribution of the forces, relocation, and replacement of tendons. If a tendon transfer cannot be performed, other options include free muscle transfer, intercalary tendon graft, tenodesis, tendon lengthening, and arthrodesis.
The function of the recipient muscle/tendon unit is more important than that of the donor unit
The optimal donor tendon must be expendable and functioning, and a wrist flexor and extensor must be preserved; in 1946, Zachary reported that the palmaris longus (PL) was not sufficient as a sole wrist flexor  ; a study evaluating whether the palmaris longus and plantaris tendon meet criteria for successful grafting concluded that these structures serve as useful grafts less often than previously thought; a second donor site should be chosen prior to surgery in case the primarily selected tendon cannot be used  ; muscles that have been reinnervated should be avoided
Synergy must be maintained; in phase transfers, include wrist extensors and digital flexors
The transferred tendon unit has one function; if a tendon is transferred to multiple functioning tendons, the force is distributed, resulting in subsequent weakness in the transfer
Maintain a straight line of transfer; better results are achieved by avoiding multiple angulations in the direction of pull in the transfer
The patient should have conscious voluntary control of the muscles and tendons that are involved in the transfer
Any joint contractures should be corrected preoperatively; maximize passive motion of all joints before the transfer procedure; postoperative active motion cannot surpass intraoperative passive motion
The donor tendon should have adequate strength, work capacity, and amplitude
In 1919, Steindler advocated achieving soft-tissue equilibrium, in which edema is resolved, joints are supple, and scars are soft, before proceeding with tendon transfer surgery  ; in a 1988 article, Brand discussed the mechanical properties of the peritendinous scar that determine the final success of a tendon transfer 
The absence of sensibility affects the postoperative use and the effectiveness of the tendon transfer
Timing of Tendon Transfer
Approximately 9-12 months after a nerve repair, the maximum nerve regeneration has occurred. (Nerve regeneration occurs at a rate of approximately 1 mm/day.) Kallio et al reported that better results are obtained with neurorrhaphy if the gap is less than 5 cm.  In a 1970 article, Brown discussed the factors that contribute to a poor prognosis due to nerve repair, including a gap greater than 4 cm, a large wound, extensive scarring, and skin loss.  Nerve grafting should be considered if undue tension on the direct neurorrhaphy exists.
Muscles have several mechanical variables, including the following:
Muscle force, or strength, is the potential for creating tension; it is a measure of the pressure that is exerted by a contracting muscle. Strength is proportional to the transverse cross-sectional area of a muscle, but it is independent of length. The strength of a selected donor tissue depends on the force of the antagonist muscle. In a 1974 article, Omer reported that when a tendon is transferred, the muscle loses approximately one grade of strength on the Highet scale, which specifies grades up to 5. 
The muscles with the most potential force for cross-sectional area include, in decreasing order, the following:
Flexor carpi ulnaris (FCU) 
Pronator teres (PT)
Extensor carpi radialis longus (ECRL)
Extensor carpi ulnaris (ECU)
Flexor carpi radialis (FCR)
Work capacity is defined as the ability to exert a force over a certain distance. It is directly proportional to the muscle mass and depends on the cross-sectional area and fiber length. 
Amplitude, or potential excursion, is proportional to fiber length. In the image below, x equals the excursion length with traction minus the resting length; y equals the resting length minus the length at full contraction; and amplitude is equal to x plus y. Usually, the two measurements are equal.
Excursion can be divided into three types: potential, required, and available. The required excursion is determined more by the joints than by the muscles. The muscles with the greatest amplitude include the following, in decreasing order:
FDP – 7 cm
FDS – 6.5 cm
Digital extensors and extensor pollicis longus (EPL) – 5 cm
Wrist flexors and extensors – 3-4 cm
Brachioradialis – 3 cm
Amplitude can be augmented by various means, such as freeing muscle from fascial attachments or transferring a monoarticular muscle to a multiarticular muscle (ie, transferring the FCR to the extensor digitorum communis [EDC]). Volar flexion of the wrist increases amplitude by 2.5 cm via the tenodesis effect (see the image below).
Planning Tendon Transfer
- Address which muscles are functional; list the functioning muscles and assess their force
- Decide which muscles are available; the chosen muscles should be expendable and functioning; muscles that have previously lost innervation and now function again are poor choices for use in tendon transfers; a flexor and an extensor must be retained for each digit and for the wrist; the surgeon must strive to maintain balance and avoid creating a new functional deficit
- Assess the needed functions; Brand argued that each tendon transfer should be tailored to the patient's individual needs 
- Match the available muscles with function; address the biomechanical properties of amplitude, force, direction, and muscle integrity
- If the needed function cannot be provided with a tendon transfer, alternative procedures can be addressed, such as arthrodesis, tenodesis, capsulodesis, and pulley release
- Protect the tendon transfer postoperatively, with no tension on the transfer; for example, if a tendon passes volarly, the wrist is splinted in palmar flexion
The first three steps in the above list are also referred to as the three-column theory or principle for tendon transfers. Drawing out these steps in three columns facilitates the decision-making process.
Radial Nerve Paralysis
Tendon transfers for radial nerve paralysis have the best and most predictable results.  Muscles that are innervated by the radial nerve include the following:
Extensor carpi radialis brevis (ECRB)
Abductor pollicis longus (APL)
Extensor pollicis brevis (EPB)
Extensor indicis proprius (EIP)
Extensor digiti minimi (EDM)
Most injuries to the radial nerve occur distal to the triceps innervation, thus sparing elbow extension. These injuries are divided into proximal (or radial nerve proper) injuries and distal (or posterior interosseous nerve [PIN]) injuries. Proximal nerve injury causes loss of wrist, digit, and thumb extension. Distal injury near the supinator may spare the ECRL and perhaps the ECRB, thereby resulting in radial deviation of the wrist and weakness of wrist extension. According to Spinner, the superficial radial nerve can innervate the ECRB in about 58% of cases. 
Various treatment options exist for radial nerve palsy associated with humeral fractures. Indications for early exploration include open fractures, requirement of open reduction, associated vascular injuries, multiple traumatic injuries, and a deficit developing after closed reduction or initiation of treatment (ie, lactate [LAC] levels). After 6-8 weeks, the absence of an advancing Tinel sign can be an indication for exploration.  Zachary advocated waiting up to 12-16 months before exploring the radial nerve. 
Seddon's method for determining how long to wait before exploring the nerve involves measuring the distance from the fracture site to the brachioradialis innervation point (2 cm proximal to the lateral epicondyle) and adding 30 to this number.  The total number is then the number of days to wait before exploring the radial nerve.
The radial nerve is mostly a motor nerve, and reinnervation is usually apparent within 4-6 months after an uncomplicated neurorrhaphy. During the prolonged recovery time, an end PT–to–side ECRB transfer can be performed at the time of nerve repair to provide wrist extension as an internal splint. The tendon transfer is performed in an end-to-side manner so that if reinnervation occurs, the continuity of the reinnervated ECRB is not lost.
Low-profile dynamic splints can be worn during the day, with night splints maintaining the digits and wrist in extension.  All joints must maintain full passive range of motion, including the first web space.
During World War I, Sir Robert Jones developed a set of tendon transfers for radial nerve paralysis, which formed the basis for reconstructive tendon transfer surgery.  The transfer included the PT to the ECRL and ECRB; the FCU to the EDC III-V; and the FCR to the EIP, EDC III, and EPL. Many modifications have been made to this plan, primarily maintaining a wrist flexor.
For wrist extension, the most common transfer used is that of the PT (universal donor) to the ECRB (central wrist extensor), positioned superficially to the brachioradialis and ECRL. As discussed, this can be performed initially as an internal splint at the time of nerve repair in staged tendon transfers. Including the radius periosteum can lengthen the PT. In a 1989 report, Tubiana et al recommended freeing the PT proximally to improve excursion.  The tendon transfer should not create a new deformity or decrease function.
A study comparing whether externally stabilizing the wrist or externally stabilizing the elbow after transfer procedures would similarly improve the ability to activate the transferred brachioradialis and resulting pinch force found that even after multiple transfers, a strong ECRB has adequate strength to extend the wrist. Both supported conditions found significantly increased maximum effort pinch force magnitude and brachioradialis activation, concluding that the addition of wrist stabilization had no significant effect compared with elbow stabilization alone. 
Several transfers can be used to restore digital extension. The FCR, FCU, or FDS (long finger) can be transferred to the EDC. The FCU has twice the force of the FCR but less excursion. Brand argued that the FCU is the prime ulnar stabilizer of the wrist.  Also, because digital extension does not require a significant amount of force, the FCR tendon transfer is favored over the FCU. Brand discussed achieving an improved straight line of pull with an end-to-end transfer of the FCR to the EDC that runs superficial to the dorsal retinaculum. 
The Boyes transfer uses the ring-finger FDS transfer to the EPL and EIP and the long-finger FDS transfer to the long and ring fingers' EDC and EDM. The FDS travels through the interosseous membrane. This transfer can be used if independent digital extension is required. Wrist motion is preserved for the tenodesis effect, thereby providing an additional 2.5 cm of tendon excursion to augment wrist amplitude in flexion.
The disadvantages of this tendon transfer, however, include a possible weakened grip, an out-of-phase transfer (transferring a flexor to an extensor), increased difficulty in the patient learning—requiring motoric reeducation—and loss of independent flexion of the donor finger.
The transfers for thumb extension include the PL over the first dorsal compartment to the EPL, which is released from the third dorsal compartment (EPL rerouting). The EPB can be added to the EPL for additional metacarpal extension. In approximately 20% of the population, no PL is present. The long FDS can then be used either through the interosseous membrane or radially around the wrist to attach to the EPL.
The APL is the major thumb metacarpal extensor; if APL function is not restored, a flexion adduction contracture of the thumb results. This can be avoided with tenodesis of the APL around the brachioradialis insertion (see the image below) or with transfer of the FCR to the EPB and APL and transfer of the FDS to the EPL and digit extensors.
Distal radial nerve paralysis involving the PIN may preserve the ECRL, resulting in wrist radial deviation. The ECRL can be transferred to the ECRB or ECU to resolve the radial deviation (see the image below). The FCU should not be used for digital extension in PIN paralysis because the radial deviation increases.
One year before presentation to the authors, a right hand–dominant 30-year-old man was treated for a gunshot wound to the left upper extremity. Injuries included a humeral fracture and brachial artery injury. External fixation and repair of the brachial artery were performed (see the first image below). The patient presented approximately 1 year after the initial injury with the inability to extend the wrist, MCP joints, and thumb on the affected side (see the second image below).
The median- and ulnar-innervated muscles functioned well and were available for tendon transfer. The patient was taken to the operating room, where the radial nerve was explored through a lateral arm incision (see the first image below). No transection of the radial nerve was found. Significant axonal injury was the likely reason for this patient's left-upper-extremity dysfunction (see the second image below).
The following tendon transfers were then performed:
PL to EPL (see the first image below)
FCU to EDC (see the second image below)
PT to ECRB and ECRL (see the third image below)
The technique of a Pulvertaft weave was used for the transfers, with 3-0 polypropylene. The periosteum was harvested with the PT for additional length. The wrist was positioned in 30° of extension, the MCP joints in 30° of flexion, the digital interphalangeal (DIP) joints in extension, and the thumb in abduction and flexion. Postoperatively, the patient did well and regained wrist, MCP joint, and thumb extension (see the image below).
Median Nerve Paralysis
Proximal or high median nerve injury results in loss of opposition, as well as thumb IP joint and index finger DIP joint flexion. For thumb flexion, the brachioradialis can be transferred to the flexor pollicis longus (FPL). This transfer weakens with elbow flexion and requires brachioradialis mobilization to increase the excursion.
If the long-finger FDP is innervated by the ulnar nerve, then the long-finger FDP can be transferred side-to-side to the index-finger FDP. If the long-finger FDP is innervated by the median nerve, use the ring-finger FDP side-to-side transfer to the long and index fingers' FDP. If significant radial-side strength is needed, the ECRL can be transferred to the index-finger FDP.
Thumb opposition is a combination of trapeziometacarpal and MCP joint abduction, flexion, and pronation (see the image below). Position takes precedence over force for intact opposition. Proper positioning includes having the nail plates of the thumb and long finger in the same plane.
Distal median nerve injuries involve loss of opposition because of paralysis of the abductor pollicis brevis (APB), opponens pollicis, and superficial head of the flexor pollicis brevis (FPB). The APB appears to be the most important muscle in opposition. If a first web space contracture is present preoperatively, this condition must be corrected before performing the tendon transfer (see the image below).
Steindler performed the first opponensplasty in 1919 by transferring the radial slip of FPL to the dorsal base of the thumb proximal phalanx.  Several more options have since been developed for opposition transfers. In these transfers, the APB insertion is approached at approximately a 45° angle from the pisiform. If abduction and flexion are created, pronation occurs passively.
Aguirre and Caplan in 1956 and Burkhalter et al in 1973 described transferring the EIP to the APB, which is the most common transfer to restore opposition (see the image below). [36, 37] The advantages of this transfer include no requirement for a pulley or tendon graft, no loss of grasp force, and avoidance of dissection in scarred tissue. The disadvantage is that the length of the EIP is just enough to transfer to the APB. When the EIP is mobilized, the extensor hood overlying the index finger should be repaired to prevent an extension lag.
Bunnell in 1924, Royle in 1938, and Thompson in 1942 described transferring the ring finger FDS to the APB (see the first image below). [2, 38, 39] The ring-finger FDS is divided at its insertion and passed around the ulnar border of the palmar aponeurosis. A pulley can be created from the FCU or PL (see the second image below). Note: This transfer cannot be used in a high median nerve injury, because the ring-finger FDS is paralyzed.
In the Camitz procedure, also advocated by Bunnell, the PL is transferred in a subcutaneous tunnel to the APB (see the images below).
Harvesting the distal palmar fascia with the PL increases length. This transfer is beneficial for thenar paralysis that is secondary to chronic carpal tunnel syndrome. Because the PL is near the median nerve, this transfer is not optimal for traumatic injury to the distal median nerve. The PL transfer provides abduction, not flexion or pronation, so opposition is not the final result. Foucher et al attached the PL to the EPB or dorsal capsule to obtain some opposition. 
In the Huber procedure, the abductor digiti minimi (ADM) is transferred to the APB. This tendon transfer is especially useful for thumb hypoplasia because it creates thenar bulk. The neurovascular bundle, which is found on the dorsoradial aspect of the ADM, is preserved. If the pisiform origin is preserved, a tendon graft may be required, but better blood supply is retained (see the image below). The ADM muscle is mobilized to the pisohamate origin and is rotated 180° so that the superficial side becomes deep and radial through the subcutaneous tunnel.
Ulnar Nerve Paralysis
Tendon transfers for ulnar nerve injuries have less predictable results than those for radial nerve injuries.  Distal power pinch involves the adductor pollicis and the first dorsal interosseous muscles for index finger abduction. Approximately 33-50% of grip strength is lost in ulnar nerve paralysis.
The intrinsic muscles are the primary flexors of the MCP joints and these muscles also extend the IP joints. With intrinsic muscle paralysis, the deformity manifests as hyperextension of the MCP joints and IP flexion (see the image below). The primary functions lost in ulnar nerve paralysis include thumb power pinch, index abduction, and a claw deformity. 
The tendon transfer for low ulnar nerve injury with weak pinch involves the long- or ring-finger FDS to the adductor pollicis insertion, with no tendon graft required. When a high ulnar nerve injury is present, the long-finger FDS can be split so that the radial side transfers to the adductor pollicis and the ulnar side loops around the A2 pulley of the ring and small fingers.
The ECRB adductorplasty is another option. The ECRB or ECRL with tendon graft traverses through the second or third intermetacarpal space volarly across the palm to the adductor pollicis insertion. This transfer doubles pinch strength. Transferring the EIP from the ulnar to the radial aspect of the index MCP joint can strengthen the first dorsal interosseous muscle. Another option is transferring an accessory APL with tendon graft to the first dorsal interosseous or index-finger lateral band.
Combined Median and Ulnar Nerve Paralysis
Low median and ulnar nerve paralysis is the most common type of injury and affects the intrinsic muscles. High paralysis also involves the wrist and digital flexors. Transferring the ECRL to the FDP and the brachioradialis to the FPL restores digital flexion. Transferring the EIP to the APB is the opposition transfer. An accessory APL with tendon graft to the index finger lateral band provides pinch.
For a claw deformity, several tendon transfer choices exist. The intrinsic minus position is hyperextension of the MCP joints and extension loss at the IP joints. The goal is to provide flexion at the MCP joints, thereby allowing the EDC to extend the IP joints. All the tendon transfers for a claw deformity pass volar to the MCP joint.
With solely an ulnar nerve injury, the long-finger FDS remains innervated and can be transferred to the A1 or A2 pulley, proximal phalanx (Littler method), or radial lateral bands of the ring and small fingers while traveling volar to the transverse metacarpal ligament. In 1922, Stiles and Forrester-Brown described splitting the ring- or long-finger FDS and transferring it to the EDC. 
With combined median and ulnar nerve injuries, the static tenodesis described by Riordan in 1953 can be used.  ECRL, ECRB, or ECU static tenodesis with tendon graft transfers to the radial aspect of the lateral bands of the paralyzed fingers may be used. In a 1973 article, Parkes advocated another form of static tenodesis, which takes a free tendon graft from the radial lateral bands to the deep transverse metacarpal ligament. 
The Zancolli lasso procedure creates a functional dynamic tenodesis in which each FDS is looped around its corresponding A2 pulley to provide flexion of the MCP joints (see the images below). No change in grip strength occurs. This procedure is good for diffuse paralysis or if limited donor tendons are available.
Other options for dynamic tenodesis include a free tendon graft looped through the extensor retinaculum, passing volar to the transverse metacarpal ligament, and inserting onto the radial lateral bands, as described by Fowler in 1949.  The long-finger FDS can also be split into four slips and passed through the lumbrical canal to insert on the radial lateral band. Static blocks incorporate volar plate advancement over the MCP joint, allowing 20° of flexion (capsulodesis), as described by Zancolli in 1957.  In 1924, Bunnell discussed flexor pulley release, which allows bowstringing and flexion. 
Combined Median and Radial Nerve Paralysis
A combined median and radial nerve paralysis injury is usually addressed with a two-stage procedure. Initially, a wrist arthrodesis is performed with transfer of the FCU to the EPL and EDC to provide digit and thumb extension. The second stage involves the Huber procedure for thumb opposition (transfer of the ADM to the APB), thumb interphalangeal arthrodesis, and side-to-side distal forearm FDP tenodesis.