* This chapter reproduces much of the content published in previous editions of this book, but has been extensively reformatted and updated.
Carpal tunnel syndrome (CTS) is the most common peripheral neuropathy treated by hand surgeons.
Most cases of CTS are idiopathic, but the pathology clearly involves increased pressure in the carpal tunnel.
The association of specific jobs with CTS is controversial.
Nonsurgical treatments, such as activity modification, orthosis use at night, and steroid injection, can be helpful in milder cases.
The most effective treatment is surgical decompression, by sectioning of the flexor retinaculum (transverse carpal ligament).
The evidence supporting one surgical approach over another is mixed; limited exposures may have less pain immediately postoperatively, but also seem to have a higher risk of incomplete release and possibly nerve injury.
The use of postoperative immobilization or therapy is not supported by high-quality evidence.
Compression of the median nerve at the wrist, or carpal tunnel syndrome (CTS), is the most common upper extremity compressive neuropathy. Recent data have indicated that CTS may affect as much as 3% of the population at any one time. Approximately half of these individuals will seek medical attention, and half of those will ultimately be treated with carpal tunnel release, making carpal tunnel surgery the most common hand surgical procedure. It is also one of the most common causes of lost work time in the United States. However, there is little consensus on the most appropriate nonsurgical treatment, the optimal surgical procedure or rehabilitation protocol, or even the most appropriate way to assess outcome. More significantly, in most cases, the etiology remains idiopathic, making preventive strategies entirely conjectural.
Although CTS is the most common condition surgically treated by hand surgeons, the term and our current understanding of the condition both date from the last half-century, as documented in several recent reviews. The earliest reports of median neuropathy were cases of median nerve entrapment after fracture of the radius. Early treatments included amputation and orthosis use. Although amputation is no longer favored as a treatment for CTS, orthosis use is still employed. Later, excision of prominent palmar callus was tried and, even later, median neurolysis. In 1933, Abbott and Saunders injected dye into cadaver carpal tunnels and noticed increased resistance to dye flow with wrist flexion. On the basis of this study, they condemned the wrist flexion (Cotton-Loder) position, which had been commonly used until that time for the treatment of Colles’ fracture. CTS after Colles’ fracture remains an important clinical problem.
Although post-traumatic CTS was the first type to be noticed, by the later 1800s, nontraumatic cases were also noted, although their true nature was not known. In 1880, Putnam described 37 patients with nocturnal paresthesias. He noted that “simply letting the arm hang out of the bed or shaking it about … [or the use of] prolonged rubbing would relieve the symptoms.” These are, of course, classic symptoms of CTS. Putnam recognized the median nerve as the source of the symptoms, but did not consider compression as an etiology. When compression was considered, the location selected was erroneous, with the brachial plexus being considered the site. Unfortunately for patients with CTS, this hypothesis of brachial plexopathy as an etiology became popular and resulted in a variety of misdirected surgery until the late 1940s.
In 1913, Marie and Foix described an autopsy finding of a large median pseudoneuroma just proximal to the carpal tunnel. They suggested that “perhaps in a case in which the diagnosis is made early enough … transection of the ligament could stop the development of these phenomena.” Unfortunately, this advice went unheeded for two decades, until Learmonth described two cases in which he divided the flexor retinaculum to treat a median neuropathy. Subsequently, Phalen et al. popularized the condition and emphasized the frequent intraoperative finding of synovial fibrosis without inflammation.
The carpal tunnel is an inelastic structure. Its floor is composed of the concave arch of the carpal bones. The hook of the hamate, the triquetrum, and the pisiform constitute the ulnar border, whereas the radial aspect includes the trapezium, the scaphoid, and the fascia over the flexor carpi radialis. The roof of the carpal tunnel is made up of the deep forearm fascia, the flexor retinaculum, and the distal aponeurosis of the thenar and hypothenar muscles. The flexor retinaculum extends from the scaphoid tuberosity and the trapezium to attach to the pisiform and the hook of the hamate. The contents of the carpal tunnel consist of the median nerve and nine flexor tendons. The flexor pollicis longus, four flexor digitorum profundus, and four flexor digitorum superficialis tendons pass through the carpal tunnel, with the nerve lying superficially and anteroradially in the tunnel ( Fig. 48-1 ).
At the distal edge of the carpal tunnel, the median nerve most commonly divides into six branches: two common digital nerves, three proper digital nerves, and the recurrent motor branch, which innervates the thenar musculature. The nerve gives off the palmar cutaneous branch proximally and radially, approximately 5 cm proximal to the wrist crease. This branch travels with the main nerve for 1.4 to 2.6 cm and then penetrates the antebrachial fascia between the palmaris longus and flexor carpi radialis. It emerges subcutaneously, usually proximal to the wrist crease, to innervate the palm.
Many variations in the branching pattern and anatomy of the median nerve have been described. These include high (i.e., proximal) division, persistence of the median artery, anomalous subretinacular muscle and tendon attachments, and proximal cross-connections with the ulnar nerve, most of which are rare. More commonly seen are proximal incursions of the lumbrical muscles, variations in the branching and courses of the recurrent motor branch ( Fig. 48-2 ) and palmar cutaneous branch, and distal cross-connections with the ulnar nerve. As discussed later, an awareness of these variations is essential when performing surgical release of the carpal tunnel.
The etiology of CTS is unknown in most cases; however, the pathogenesis appears to be clearly related to increased pressure within the carpal tunnel. Experimentally increasing carpal tunnel pressure produces characteristic symptoms of CTS in normal volunteers, and patients with CTS have increased resting CTS pressure, which can rise to as much as 1000 mm Hg with activity, compared with normal levels of 30 to 50 mm Hg with activity, and 10 mm Hg at rest.
Factors Influencing Pressure in the Carpal Canal
In general, any disorder that decreases the cross-sectional area of the carpal tunnel or that produces increased carpal tunnel canal volume content may result in increased pressure in the carpal canal. In the former category, fractures and carpal arthritis are common culprits. In the latter category, and occurring somewhat less frequently, are tumors and synovitis, increasing the size of the nerve or synovium. Most commonly, however, none of these factors is present, and although there is a statistical association of such conditions as diabetes and certain work activities involving forceful grip and wrist flexion with a diagnosis of CTS, a specific cause-and-effect relationship has not been definitively established.
Recent studies have suggested a role for the lumbricals, intermittent space-occupying structures within the carpal tunnel. The lumbrical muscles are distal to the carpal tunnel with the fingers held in extension, but they lie within the carpal canal when the fingers are actively flexed, especially with the wrist flexed. Clinically, this may be important, because sustained contraction of the finger flexors may increase the hydrostatic pressure that the median nerve experiences as a result of crowding in the carpal canal by the lumbricals. As noted in the subsequent chapter on therapy interventions, it may be appropriate to consider the possibility of lumbrical incursion when devising orthotic fabrication strategies.
Intracarpal tunnel pressure may be affected by external pressure to the palm. Cobb and associates. showed that when a 1-kg external force was applied over the flexor retinaculum, the pressure in the carpal canal increased to 103 mm Hg. Force over the thenar area increased the pressure to 75 mm Hg, and force over the hypothenar area increased the pressure to 37 mm Hg. These findings may be clinically significant. Tools, orthoses, and casts can all apply pressure to the palm and thereby elevate carpal tunnel pressure.
The classic clinical presentation of a patient with CTS is paresthesias in the thumb, index finger, and middle finger. Not uncommonly, patients also report a loss of dexterity, particularly with manipulation of small objects, such as coins, needles, or pins. These symptoms are characteristically experienced or exaggerated at night, or with characteristic activities, such as holding a book or newspaper or driving a car, and are worsened by repetitive forceful hand motion. As noted by Putnam 130 years ago, improvement in symptoms after shaking or straightening the affected hand is common.
Clinical findings in CTS include the limitation of paresthesias to the distribution of the distal median nerve (i.e., the tips and palmar aspects of the thumb, index finger, middle finger, and radial half of the ring finger, with sparing of the palm). These symptoms can often be reproduced with pressure over the median nerve, either constantly (Durkan’s sign ) or by tapping (Tinel’s sign), but these signs have only moderate sensitivity and specificity. The wrist flexion test, or Phalen’s test ( Fig. 48-3 ), is performed by flexing the wrist maximally for 60 seconds to determine whether this produces or exaggerates numbness and tingling in the affected fingers. Although it is commonly performed, this test has limited sensitivity and specificity. Thenar atrophy and loss of sensibility (two-point discrimination or Semmes-Weinstein pressure thresholds) can be observed, indicating more chronic median nerve compression. These findings, when present, are very specific, but they are not sensitive because they are only present in advanced cases.
Electrodiagnostic testing remains the best-known and most reliable study for confirmation of a diagnosis of CTS. Normal values vary, but general standards include abnormal values for distal motor latency of more than 4.5 msec and for distal sensory latency of more than 3.5 msec. A difference in median and ulnar latency of more than 1.0 msec is also considered significant. Electromyography that shows positive waves or fibrillations in the thenar musculature reflects greater severity and chronicity of nerve injury. Again, this finding is highly specific but is not sensitive because it is less frequently present. A recent meta-analysis of the literature on the use of electrophysiologic testing and its sensitivity and specificity found median nerve electromyography and nerve conduction velocity studies to be valid, reproducible, and highly sensitive and specific.
Ultimately, CTS remains a clinical syndrome, and as many as 15% of patients in some series have clinically evident and surgically relieved median nerve compression in the presence of normal electrodiagnostic results. Consequently, in some series, the best diagnostic predictors have not been laboratory or clinical tests but symptom questionnaires and even symptom diagrams. In some cases, response to specific therapies, such as steroid injection, has also been used to confirm a clinical diagnosis of CTS.
The differential diagnosis of CTS is vast. It may be confused with other peripheral neuropathies, such as pronator syndrome, diabetic neuropathy, or hand–arm vibration syndrome, as well as cervical radiculopathy and even central nervous system pathology. It may be distinguished from these other conditions, typically by the distribution of symptoms, which are limited to the median innervated fingers and spare the palm or dorsum of the hand (which are affected by more proximal neuropathies). The finding of split symptoms in the ring finger, with sparing of the ulnar side, is nearly pathognomonic of CTS.
The efficacy of nonsurgical management of mild CTS has been well documented. Conservative management is generally not an option for moderate to severe CTS, especially in patients who have signs of muscle atrophy or significant sensory impairment. The diagnosis of CTS should be confirmed with clinical testing, and potential proximal compression sites should be ruled out. Conservative management options that have been described for CTS include orthosis use, nonsteroidal anti-inflammatory drugs, injection of the carpal tunnel with a corticosteroid, tendon and nerve gliding exercises, vitamins, iontophoresis, ultrasound, and activity or job site modifications. The following sections describe various conservative treatment options, including those that may warrant further study and consideration based on current research.
The purpose of a steroid injection is ostensibly to decrease the mass of the thickened flexor tendon synovium by decreasing the inflammatory process, but this has never been documented, and it is generally agreed that there is little inflammation in most cases of CTS. Thus, although steroid injection has been shown to be effective, the mechanism of this effect is unknown. Studies show that injection and orthosis use initially provide relief in approximately 40% to 80% of patients, but by 18 months after injection, this number decreased to 22%. There is value in steroid injection, even if symptoms return, in that improvement in symptoms confirms the diagnosis, and a positive response may be indicative of a favorable outcome if surgery is required. However, poor relief after injection does not predict a poor surgical result.
Anti-Inflammatory Drugs and Vitamin B 6 Treatment
Nonsteroidal anti-inflammatory drugs have never been shown to be effective in a scientific study of CTS.
In 1976, Ellis et al. proposed that a deficiency in pyridoxine (vitamin B 6 ) may cause CTS. Although adding vitamin B 6 to the diet is still occasionally advocated in the literature and the press, no controlled, randomized, prospective study has shown vitamin B 6 to be clearly efficacious compared with other conservative treatments for CTS.
Acute CTS requires prompt recognition and early open carpal tunnel release. Wrist trauma, as well as infections and systemic rheumatologic and hematologic disease, can acutely increase carpal canal pressure, threatening median nerve viability. Thus, when CTS occurs in the context of acute injury, it is important to distinguish neurapraxia secondary to nerve contusion from acute CTS. In such cases, the surgeon should measure carpal tunnel pressure. If it is elevated, then carpal tunnel release should be considered on an urgent basis. If not, the surgeon may elect a period of observation or, depending on the severity of injury, the surgeon could still elect release to inspect the nerve and rule out complete or partial transection or the presence of intraneural hematoma. CTS developing after closed fracture reduction should also prompt exploration because entrapment of the nerve between fracture fragments and even nerve transection have been reported. CTS in the context of presumed or confirmed septic flexor tenosynovitis should also be acutely released to avoid septic necrosis of the nerve.
For idiopathic CTS, a relative indication for surgical release exists when clinical or electrodiagnostic evidence of denervation of the median nerve–innervated muscles is present. In the absence of significant clinical or electrodiagnostic denervation changes, failure to respond with a few weeks of nonsurgical treatment remains a reasonable guideline, as noted by a recent review of the available evidence. Particularly among patients who must continue in a manual labor occupation, prolonged nonsurgical treatment without symptomatic improvement does not appear to be fruitful.
Open Carpal Tunnel Release
For nearly a century, sectioning the flexor retinaculum has been the mainstay of treatment of CTS. Alternatives, such as synovectomy without transection, have had some success, as reported in studies with lower levels of evidence, but the strongest evidence has always supported release of the flexor retinaculum, especially open release. Division of the flexor retinaculum significantly increases the cross-sectional area of the carpal tunnel, with imaging studies showing a mean 24% increase in carpal canal volume. With the relatively high incidence of variations in the location of the median nerve and the ulnar neurovascular structures, open carpal tunnel release remains the preferred procedure of most surgeons for decompression of the median nerve at the wrist. This technique affords full inspection of the transverse carpal ligament and the contents of the carpal canal ( Fig. 48-4 ).
Local, regional, or general anesthesia may be used, but except in the anxious patient, local anesthesia provides adequate local pain relief for uncomplicated release. The use of local anesthetic mixed with epinephrine obviates the need for a tourniquet (another source of intraoperative pain), but may interfere with the surgeon’s ability to assess intraoperative nerve vascularity, which some consider a prognostic sign. A rapid return of blood flow on tourniquet release is associated with better outcomes. My preference is local infiltration of the tissues with lidocaine, with conscious sedation reserved for patients who request it. Although some have expressed a concern that local anesthetic infiltration can obscure tissue planes, and thus express a preference for axillary or intravenous regional block, I have never found this to be a problem. Unless epinephrine is used, tourniquet control is used to minimize bleeding and improve visibility, although there is no evidence comparing outcomes with and without tourniquet use. When tourniquet use may be relatively contraindicated, such as after axillary node excision or radiation, I do not hesitate to use lidocaine with epinephrine (both 1 : 100,000 and 1 : 200,000 work well) and forego tourniquet use. If epinephrine is used as an alternative to a tourniquet, it is important to wait several minutes for vasoconstriction to occur before making an incision.
In open release, a palmar incision is made ulnar to the depression between the thenar and hypothenar eminences, more or less in the line of the middle–ring web space, to minimize damage to branches of the palmar cutaneous branches of the median and ulnar nerves. Staying radial to the hook of the hamate minimizes risk to the ulnar neurovascular bundle. The incision is extended from Kaplan’s cardinal line along the axis of the third web space proximally ( Fig. 48-5 ). Extension proximal to the wrist crease is not required, but I prefer it, because it is easier to identify the median nerve at this level, release the distal antebrachial fascia, and then dissect from proximal to distal, in the direction of, and thus less likely to injure, the branches of the median nerve, whatever variation they may follow. Others prefer a distal to proximal dissection or even a direct palmar to dorsal approach. There is little evidence to support that any of these variations is better than the others.
In the subcutaneous tissues overlying the flexor retinaculum, an effort should be made to preserve the larger branches of the palmar cutaneous branch of the median nerve, the nerve of Henle, and cutaneous branches of the ulnar nerve. The superficial palmar fascia is divided in line with the skin incision. The underlying flexor retinaculum is divided longitudinally along its ulnar aspect. I prefer to avoid incision into the periosteum of the hamate because periosteal injury may predispose to pillar pain, but again there is no evidence to support this opinion. At the distal end of the incision, the superficial palmar arterial arch is identified in its bed of adipose tissue and is protected. This represents the distal limit of dissection. After hemostasis is achieved, the wound is irrigated and closed with nonabsorbable suture. Some prefer to close the palmar fascia as well.
Open decompression of the median nerve has stood the test of time, providing reliable symptom relief in patients with a multitude of associated diagnoses. The procedure allows for visualization of all anatomic variations, and if necessary, removal of space-occupying lesions, such as osteophytes or proliferative synovium. A number of high-quality studies and meta-analyses of open carpal tunnel release document patient satisfaction and symptom improvement rates ranging from 86% to 96%. Nocturnal pain improves to a greater extent than any other symptom. The resolution of symptoms and functional limitations tends to follow a temporal course, with nocturnal pain, tingling, and subjective numbness that improve within 6 weeks after surgery. However, two-point discrimination, if abnormal initially, may remain abnormal. Weakness and functional status improve more gradually. Grip and pinch strength worsen initially and usually return to preoperative levels after approximately 2 or 3 months, with maximum improvement at approximately 10 months. In a series of 44 patients with mean follow-up of 35 months after open carpal tunnel release, Osterman reported a 96% rate of patient satisfaction and symptom improvement, with 84% of patients returning to their preoperative jobs after surgery.
One of the advantages of open release is that technical complications are uncommon. The most common is incomplete release. Wide exposure should minimize the risk of inadvertent nerve or vascular injury. Deep wound infection is also uncommon, although superficial “stitch abscesses” occur. The principal postoperative problem is usually palmar tenderness, so-called pillar pain on either side of the cut retinaculum. Interestingly, this pain occurs after all types of carpal tunnel release. One study reported a 41% rate of postoperative allodynia over the thenar and hypothenar eminences at 1 month after surgery, decreasing to 25% at 3 months and 6% at 12 months. Postoperative palmar pain may contribute to delayed return to work, particularly among manual laborers and those receiving worker’s compensation.
Pillar pain is localized to the thenar or hypothenar areas and must be distinguished from palmar incisional or scar tenderness. The relevant anatomy appears to be the osseous and muscular attachments of the flexor retinaculum, the origin of the three thenar and three hypothenar muscles. Usually, this pain decreases over time and again appears to be independent of the type of release performed. A more rare complication is flexor tendon bowstringing, which may cause painful snapping of the tendons over the hook of the hamate with grip or, in extreme cases, painful adherence of the median nerve to the palmar subcutaneous tissues. I attempt to identify the potential for this problem intraoperatively by asking the patient to grip strongly after the release is completed; if the tendons and nerve displace out of the tunnel anteriorly, I reconstruct the retinaculum with a distally based flap ( Fig. 48-6 ). I also perform this procedure if the patient has a job that requires forceful pinch or grip with the wrist flexed and the work cannot be modified, such as dental hygienists.
For many years, traditional postoperative management after open carpal tunnel release included a period of immobilization ranging from several weeks to a month or more. More recently, early mobilization has been more commonly considered after open release, with a consequent reduction in postoperative morbidity and more rapid return to work. Based on several reports showing no benefit to postoperative immobilization, I now allow my patients to begin moving their fingers and wrists immediately after surgery, requesting only that they limit the weight that they lift and that they lift with the forearm supinated, so that the wrist will be extended during lifting. This is discussed in more detail later.
Endoscopic and Mini-Open Carpal Tunnel Release
A variety of less invasive techniques, such as endoscopic and mini-open carpal tunnel release, have been proposed to decrease the morbidity of carpal tunnel release by selectively transecting the flexor retinaculum through small incisions, often placed outside the palm or at least away from the high-contact midpalm of the hand. Despite these innovations, the incidence of pillar pain, by strict definition, is probably the same, regardless of technique. Proponents of alternative techniques state that less incisional tenderness is present, but most studies suggest that this difference is only significant in the first few weeks after surgery. Nonetheless, endoscopic and mini-open procedures do appear to be associated with a few weeks’ earlier return of grip and pinch strength and an earlier return to work. These benefits are to some extent outweighed by higher rates of technical complications in most series, with higher rates of incomplete release, median nerve injury, and (something exceedingly unusual with open release) injury to the ulnar neurovascular bundle. Although many surgeons criticize incisions for open procedures that cross the wrist, all endoscopic release techniques require an incision in the distal forearm. The one-portal releases employ only this proximal incision ( Fig. 48-7 ), whereas the two-portal techniques use this proximal incision and a small distal palmar incision. The one- and two-portal releases use a variety of specially designed devices to release the retinaculum.