Treatment failure and complications are encountered in 1% to 25% of all carpal tunnel releases. Besides hematoma, infection, skin necrosis, and intraoperative iatrogenic injuries, persistence and recurrence should be included in this discussion. Persistence is often related to incomplete release. Similar symptoms recurring after a symptom-free interval of 6 months are considered recurrent and may be caused by intraneural or perineural scarring. Adequate diagnosis and treatment of these failures can be challenging. Operative release is the main treatment consisting of complete decompression of the median nerve. In some circumstances, coverage of the median nerve may be necessary.
- •
Treatment failure and complications are encountered in 1% to 25% of all carpal tunnel releases.
- •
Besides hematoma, infection, skin necrosis, and intraoperative iatrogenic injuries, persistence and recurrence should be included in this discussion. Persistence is often related to incomplete release.
- •
Operative release is the main treatment for persistence and recurrence consisting of complete decompression of the median nerve. In some circumstances, coverage of the median nerve may be necessary.
Introduction
Carpal tunnel syndrome (CTS) is the most common compressive peripheral neuropathy of the upper extremity. The annual incidence is approximately 7 cases per 10 000 humans. Carpal tunnel release (CTR) usually provides a good outcome with complete resolution of symptoms. However, treatment failure and complications are encountered in 1% to 25% of all CTR published reports, with a reoperation rate of up to 12%.
In 1966, Phalen published his experience of 212 CTRs, with only 2 patients requiring reoperations. He reported that one patient had incomplete severance of the distal portion of the transverse carpal ligament and the second had recurrence caused by scarring.
As described by Phalen and other investigators, the failure of CTR can be classified by the affected or injured structures, the presenting symptoms or complaints, and by the timing of their appearance. Adequate diagnosis and treatment of these failures can be challenging. The definitions and the clinical evidence are yet unclear. They are based only on retrospective studies or small series with anecdotal conclusions. It is also important to note that the terms recurrence and persistence are typically used interchangeably, which makes analysis of any existing data extremely challenging.
There is a paucity of data specifically addressing recurrent CTS, with many studies purposely excluding patients with recurring symptoms requiring treatment and, furthermore, lacking information regarding the incisions used or subsequent coverage of the median nerve. The authors present a short review that addresses the evaluation and treatment of failed CTR.
Complications and failed treatment of primary CTR
The authors distinguish complications based on their timing, namely, intraoperative, early postoperative, and late postoperative complications ( Table 1 ). Furthermore, symptoms after failed primary CTR can be classified as persistent, recurrent, or new.
| Intraoperative | Injuries to nerves, vessels, or tendons |
| Early postoperative | Infection, hematoma, skin necrosis, scar issues, incomplete release of transverse carpal ligament, persistent symptoms |
| Late postoperative | Complex regional pain syndrome and pillar pain, neuroma formation, scar issues, recurrent or new symptoms |
Typically, complications of CTR can arise because of iatrogenic injuries to nerves (palmar cutaneous branch of median nerve, median nerve, and ulnar nerve), vessels, or tendons (including adhesions to the adjacent tissues) as well as bowstringing of the flexor tendons. Pillar pain and complex regional pain syndrome are complications that are not completely understood. Infection, hematoma, skin necrosis, and scar hypertrophy are other additionally reported complications of CTR. Rare complications should also be mentioned, such as pisotriquetral subluxation.
The actual technique of primary CTR does not seem to have an obvious impact on the incidence of complications. Boeckstyns and Sorensen analyzed 54 publications that reported complications of 9516 endoscopic and 1203 open releases. They concluded that endoscopic release was comparable with open release. Although the rate of irreversible nerve damage was similar (0.3% vs 0.2%), they noted that reversible nerve problems were more common after endoscopic release. There was, however, a slightly higher risk of catastrophic complications with endoscopic release, such as complete transection of the median nerve. Tendon lesions were very rare (0.03%); the rate of other complications, such as reflex sympathetic dystrophy, hematoma, or wound problems, was about the same between endoscopic and open CTR.
The main focus of this review is to analyze and discuss the treatment of failed CTR. This treatment can be classified into 2 categories: persistence and recurrence. The clinician must be able to distinguish old from new symptoms, which forms a keystone in distinguishing persistence from recurrence. The rate of revision surgery for persistent or recurrent symptoms has been reported to be 3.1%.
Persistence or transient (less than 6 months) relief of symptoms was often related to incomplete release of the transverse carpal ligament, especially seen with endoscopic techniques. In one large series, 54% of all redo-CTRs were caused by incomplete transection of the transverse carpal ligament. Although incomplete release (usually distally, especially in CTR with transverse or short longitudinal incisions) was the most common cause, other causes of persistent or transient symptoms after CTR included tenosynovitis and fibrosis.
In a cadaveric study, Cobb and Cooney showed that carpal arch widening was not affected significantly by the incidence of complete versus incomplete release of the carpal tunnel. They also showed in a clinical study that the outcome of revision surgery was independent of complete versus incomplete release.
An incorrect primary diagnosis can also be a cause for an apparent postoperative persistence of CTS. Conditions like cervical radiculopathy; spinal cord lesions; proximal nerve compressions, such as the pronator syndrome; cervical disk disease; thoracic outlet syndrome; or systemic diseases, such as diabetes mellitus, thyroid disorders, and hemodialysis-related neurologic sequelae, may manifest with symptoms similar to that of CTS. Lesions around the carpal tunnel, like ganglia, arterial aneurysms, gouty tophi, amyloidosis, sarcoidosis, or fracture callus, may compress the median nerve producing symptoms suggestive of CTS. Secondary gain should also be considered in the differential diagnosis.
Recurrent symptoms have been defined as “a significant or complete relief of the patient’s original pre-operative symptoms which lasts for a definite period of time after the initial carpal tunnel release, but, eventually, similar symptoms develop again.” Jones and colleagues considered 6 months as the minimum symptom-free interval before attributing symptoms to be arising from a recurrence. Possible causes of recurrent symptoms include perineural or intraneural fibrosis (32% of all cases in a large series), scarring, and adherence of the median nerve to adjacent tissue as well as nerve traction with wrist motion caused by adherence, neuroma formation, median nerve subluxation from the carpal tunnel, tenosynovitis, or flexor retinaculum regrowth. Possible cofactors that have been described include poor hemostasis, excessive period of immobilization, and inappropriate physical therapy. As in patients with persistent symptoms, incomplete release can cause recurrence. Furthermore, less likely causes of recurrence must be kept in mind. For example, there is a case report of an aneurysm of an aberrant median artery causing recurrent symptoms.
New symptoms are usually distinct and can be even more distressing than the primary complaints of the CTS. There are multiple possible causes, including iatrogenic injuries to nerves, vessels, or tendons. Scarring can result in neuromas, fibrosis, and hypertrophic scars leading to neuropathic or pillar pain, dysesthesias, muscle weakness, and stiffness.
Symptoms and diagnosis of persistence and recurrence
In making the diagnosis and differentiating between persistence and recurrence of CTS, the clinician must perform a thorough history, carefully examine and make a record of the clinical presentation, review the operative report, and possibly use or review additional diagnostic studies. A thorough assessment begins with reviewing the symptoms and, if possible, the electrodiagnostic studies before the primary CTR was performed. Many patients with persistent or recurrent CTS complain of similar symptoms as in primary CTS. Numbness, pain and tingling at night, or even persistent numbness can be present. Physical examination findings, such as weakness of opposition of the thumb or thenar wasting, can also be observed because of long-standing median compression. In persistence, mainly caused by incomplete release (this can also be the cause in recurrence), the symptoms are similar or even unchanged as in primary CTS. In recurrence, they are typically slightly different and accompanied by scar hypersensitivity and oftentimes pain.
Standardized questionnaires for the assessment of severity, functional status, and quality of life in patients who have CTS may be of further value in the assessment after failure of resolution following surgery. These questionnaires have been shown to improve quality of care in the initial surgical management of CTS. The consistency of these questionnaires in patients with failed CTR is yet unclear.
The review of the operative report is another essential element in the treatment of patients with failed surgical management of CTS. It encompasses not only what steps were performed but can also offer important hints regarding the individual anatomy.
A thorough clinical examination of both upper extremities as well as the cervical spine follows. The essential components of a thorough evaluation include the inspection of the prior CTR incision and any other scars; sensibility of the median, ulnar, and radial nerves; static and moving 2-point discrimination; grip strength; and provocative testing. Many tests used for the diagnosis of primary CTS can also be used for recurrent or persistent CTS. Among them, the Phalen, Tinel, and Durkan tests are the ones most frequently used. De Smet published 5 cases with persistent symptoms and concluded that a negative Phalen test and a positive tourniquet (Gilliatt) test can indicate a distally incomplete release of the retinaculum.
All patients with a diagnosis of failed CTR should have further work-up consisting of electrodiagnostic testing to improve accurate diagnosis but also to serve as an objective follow-up parameter. An abnormal electromyogram (EMG) is usually present; however, a normal EMG should not exclude the possibility of persistent/recurrent compressive neuropathy. In addition, it is important to note that electrical changes in EMG can persist for an unpredictable period of time, even after successful CTR. If the clinical signs clearly indicate persistent or recurrent CTS, surgery may be indicated despite a normal EMG. The authors also recommend evaluating bony structures with either a conventional radiograph or even sophisticated imaging, such as an magnetic resonance imaging (MRI) or a computed tomography scan in the case of a posttraumatic presentation to rule out fracture callus or dislocation causing a compression of the median nerve. High-resolution ultrasound may also be be used as an optional resource. Tenosynovitis, any space occupying mass, fibrosis or nerve enhancement can be evaluated. However, it does not seem that MRI can reliably detect incomplete CTR. Furthermore, MRI is expensive and the reference standards are not clearly defined.
Therapy
As already mentioned, the adequate diagnosis and treatment of these failures can be challenging. It is, therefore, essential to make the correct diagnosis of recurrent or persistent CTS. A neuroma of the palmar cutaneous branch of the median nerve is treated with re-exploration of the wound, different from recurrent or persistent CTS. Also, some symptoms should be treated nonoperatively, such as in patients with complex regional pain syndrome.
The next step in management following the accurate diagnosis is the decision of how to treat patients. Nonoperative treatment, such as splinting, injection with local anesthesia or steroids, and activity modification, can be used as a temporizing measure or as a diagnostic tool. However, the data supporting this are lacking. Some investigators prefer to initially try nonsurgical treatment, especially in recurrent CTS. Craft and colleagues recommended microneurolysis and coverage with the hypothenar fat pad flap in recurrent symptoms but only after failed initial nonoperative treatment. In a recently published study, cortisone injection improved the symptoms in 23 out of 28 patients with recurrent CTS. Revision CTR using an extended incision, which crosses the wrist crease into the distal forearm, is the most recommended treatment of persistent or recurrent CTS. Endoscopic release can be a safe treatment in the hands of experienced surgeons but only in cases of late recurrence.
The timing of surgery depends on several factors. Most of the symptoms of numbness and night pain resolve within 24 hours after primary CTR. Jones and colleagues used chronic severe pain for more than 1 year as an indication for revision CTR. However, persistence of numbness for more than 1 year can be present and subsequently resolve in up to 40% patients. O’Malley and his colleagues recommended re-exploration of the carpal tunnel in patients with persistent symptoms causing nocturnal wakening or exacerbated by activities and in those with a positive Phalen test. All of their patients had hypesthesia in the distribution of the median nerve. If incomplete release of the transverse carpal ligament is suspected, a reoperation can be indicated earlier in the course. One large study had an average time of 996 days (range, 15 days to 12 years) between the first and second operation. There is one retrospective study showing no influence on outcome based on the time from the initial operation to the time of revision surgery. The psychological strain of ongoing symptoms on patients can also be used as a helpful guide in timing suitable surgery. However, one must be cautious in attributing too much significance to this because there is always the possibility of secondary gain, poor coping skills, or pain catastrophizing that may exert undue influence not only on the symptom complex but also on the decision-making algorithm of the surgeon, particularly if inexperienced.
In deciding the general operative approach, the main goal of treatment must be exploration of the carpal tunnel in its entirety to completely decompress the median nerve. All persisting fibers of the transverse carpal ligament must be released as well as the antibrachial fascia at the level of the wrist crease and well proximal to it. The exploration, potential external neurolysis, and epineurectomy of the median nerve should be started proximal (from the ulnar side) and distal from the areas that have been scarred to dissect from the known to the unknown anatomy. This procedure is usually done under loupe magnification but may be performed under microscopic magnification. If intrafascicular scarring is significant, internal neurolysis may be considered; however, there is not much data to support this.
Incision
Langloh and Linscheid published their results of 34 re-explorations caused by recurrence in 2053 CTRs. In this series, the carpal tunnel was approached via the previous incision. In 1992, Dellon and Chang presented their series of patients with recurrent CTS. They suggested using an alternative incision for approaching recurrent median nerve compression through unscarred soft tissue, expressing concerns that the median nerve could be intimately associated with or even found within the scar. They used a new incision, which was 1.0 to 1.5 cm ulnar to the prior incision and divided the most ulnar part of the transverse carpal ligament just adjacent to the hook of hamate. They had no complications with a follow-up period of 20 months. Similarly, Mathoulin and colleagues used a linear incision 1 cm away from the original scar along the axis of the fourth finger.
Coverage
Historically, many investigators recommended simple revision nerve decompression depending on the condition of the tissues found intraoperatively. More recently, there has been increasing discussion that may support the attempt to create a biologically friendly gliding area for the median nerve. In general, adequate soft tissue coverage of the median nerve is considered to be important. Locally, this is made difficult when there is a significant amount of scarring, which can decrease the ability of the nerve to glide freely. Besides, unlike the ulnar nerve at the elbow, transposition is not an option.
There are several local subcutaneous and muscle flaps that have been described in many articles. Varitimidis and colleagues described the use of autogenous saphenous vein wrapping of the median nerve in revision surgery after successful trials in rats. Some investigators indicate that free and pedicled flaps are a superior form of coverage in this setting. Circumferential coverage of the median nerve after neurolysis by a pedicled or free flap of soft tissue might improve the revascularization of the nerve, reduce the traction forces, and improve its gliding capabilities. Described flaps include the lateral arm, hemi-latissimus dorsi muscle, scapular, pedicled reverse radial forearm fascia, and the groin flap. The use of a pedicled omentum flap has also been described and can be an alternative in patients who have undergone multiple previous failed surgical procedures.
As discussed by Tollestrup and colleagues, some flaps may be too small to cover the carpal tunnel (palmaris brevis flap, synovial flap), whereas others have potentially high donor site morbidity (abductor digiti minimi [ADQ] flap, pedicled omentum flap) or are technically difficult and require extensive dissection (free flaps). Tham and colleagues also stated that motor loss and bulkiness of some flaps can be limiting. They promote the radial artery fascial flap as another option and reported their results in 6 patients. To reduce scaring and improve nerve gliding, postoperative mobilization should start as soon as possible. However, in the case of a pedicled or free flap, immobilization of the wrist for 10 to 14 days is recommended. Some investigators advocate the immobilization of the wrist for 2 to 3 weeks after surgery, even when a flap has not been used. Advances in technology now offer alternative coverage options, such as the Canaletto implant (Eurymed, Nimes, France), which is a semirigid silicone and polyethylene implant. In a recent study, the investigators showed good results comparable with other described techniques, avoiding the use of flaps and their associated morbidity.
The Authors’ Preferred Choice
If there is a clear reason for persistent or recurrent CTS, such as an incomplete release with adequate tissue left to cover the nerve, the authors proceed with the simple release of the transverse fibers and wound closure. However, if the coverage or surrounding soft tissue bed are in question, the authors prefer coverage with the hypothenar fat pad flap. This flap is technically simple to perform, can be accessed through the same incision, does not influence the function of the hand, provides good covering of the median nerve, and also provides a gliding surface. The surgical technique is eloquently described in many articles. The hypothenar fat pad flap receives its vascular supply from the superficial branches of the ulnar artery. If necessary, the deep branch of the ulnar artery near the deep motor branch of the ulnar nerve can be dissected to allow the flap to be mobilized more easily.
Description of the Revision CTR and Coverage with the Hypothenar Fat Pad Flap
Anatomic landmarks are marked out, including the pisiform, the hook of the hamate, Kaplan’s cardinal line, and the axis of the ring finger. Thereafter, an incision is made in the base of the palm radial to the axis of the ring finger, if the previous scar is indistinguishable. In cases when the old scar can be easily identified, it is used and is extended across the wrist crease in a Z-shaped fashion and into the distal forearm. After making the initial skin incision, deeper dissection is initially started in the distal forearm to encounter normal tissue identifying the median nerve; from here on out, the median nerve is protected at all times as old scar tissue is encountered. The palmar fascia is opened along the long axis of the limb, and then the transverse carpal ligament is identified. If there is no clear definition to the transverse carpal ligament, dissection is once again started proximally. A Freer elevator is placed deep to the confluence of the transverse carpal ligament and the distal forearm fascia and the dissection proceeds from proximal to distal. Any adhesions of the median nerve to the contents of the carpal tunnel are carefully identified and lysed. Special attention must be paid to adhesions of the median nerve to the undersurface of the transverse carpal ligament or its remnant. The median nerve is decompressed completely. It is often difficult to identify the distal extent of the transverse carpal ligament because of scar formation. The senior author recommends using the sentinel pad of fat as a reliable landmark for having reached the distal most portion of the carpal tunnel. This pad of fat harbors the superficial palmar arch and almost always signifies the distal most extent of the transverse carpal ligament. It is almost always present but may seem atrophic in the very elderly. It is almost always of a darker yellow color than the subcutaneous fat. A hypothenar fat pad flap is then fashioned by undermining the hypothenar skin, dissecting the fat pad, elevating it from its ulnar side but leaving it attached at the radial edge off the palmaris brevis as described by Strickland. An excellent fat pad graft can be acquired in this manner. Dissection in a radial to ulnar direction as the skin is undermined ensures that the fat pad remains well perfused as it receives its perfusion from the radial side. The width of the graft should be adequate to completely cover the median nerve and reach the undersurface of the radial leaf of the transverse carpal ligament. Once this width is determined, the fat pad is divided at its ulnar border and carefully elevated in a radial direction maintaining its radial pedicle. It is then sewn to the undersurface of the radial leaf of the transverse carpal ligament to cover the median nerve but should avoid any tension in the graft to minimize any chance of compression from it ( Fig. 1 ). In situations when the quality of the local tissues is suboptimal and scarring of the hypothenar fat pad is either suspected preoperatively or evident during surgery, the senior author prefers using the ADQ to cover the median nerve. It is important to alert patients and obtain their consent preoperatively regarding the possibility of using the ADQ. The bulk of the ADQ after transfer provides a well-perfused muscle cover for the nerve but can preclude direct wound closure ( Figs. 2–4 ). A split-thickness skin graft is then used for skin coverage over the transferred muscle. Great care must be taken to maintain the integrity of the vascular pedicle of the ADQ at its base. In most circumstances, however, the hypothenar fat pad flap is available and affective for coverage and use of the ADQ flap is extremely uncommon. After a routine closure, the patients’ upper limb is placed in a splint. Digital motion is encouraged from the day of surgery. After 1 week, the authors discontinue the splint and start a range-of-motion program for the wrist (see Fig. 3 ). If a split-thickness skin graft has been used, the authors continue immobilization for 10 to 12 days before starting a range-of-motion program.
