Local steroid injections are widely used in the management of carpal tunnel syndrome (CTS) [1,2,3,4,5,6,7,8]. The management of CTS with steroid injections was first reported by Kopell [9] in 1958. Since then numerous cases have been reported. Steroid injections can provide a symptoms relief and a diagnosis aid for CTS. Such injections can also act as a prognostic factor. For example, a good response to a steroid injection will result in a favorable surgical outcome; however, the opposite may not be true [3].

Despite this, steroid injections can cause complications such as ischemia, skin depigmentation and atrophy. The median nerve injury (MNI) is the most serious complication associated with a local steroid injection for CTS among them [10,11,12,13].

The incidence of MNI during carpal tunnel injections is unclear. Also the incidence seems not common although the MNI are underreported probably. Bland [14] reviewed the published studies of steroid injection and documented a pool of combined series with over 3,000 injections without any instance of median nerve injury which means the risk may be less than 0.1% in competent hands.

For this study a PubMed database search was performed from 1950 to 2012 using the search terms median nerve, median nerve injury, injection injury, steroid injections, management of carpal tunnel syndrome, and different structures relevant to this review including palmaris longus tendon, flexor carpi radialis tendon and flexor digitorum superficialis.

Median nerve injuries result from direct needle injuries [1,11,12], breakdown of the blood-nerve barrier [15], neurotoxicity of the injected steroids [16], and pressure effect of the steroid [17]. A median nerve injury may lead to a formation of a granuloma. Both, myelinated and unmyelinated nerve fibers, are affected. There is also a direct neurotoxic effect on the axon and the Schwann cells [15].

Typically, the median nerve is just dorsal and radial to the palmaris longus tendon at the carpal tunnel level. If the needle inserted radial to the palmaris longus tendon, the nerve can be injured. However, the patients with CTS are more vulnerable to needle injury than healthy subjects even if the needle is inserted at correct position, because the median nerve is swollen and/or flatten around the wrist crease. Anatomic variation such as a median nerve in an abnormal location, a bifid median nerve and anomalous muscle also may affect the procedure [18,19,20,21].

Several agents were reported as potential causes for median nerve injuries, including local analgesics and preservatives [11], celestone [17] and methylprednisolone in polyethylene glycol mixed with bupivacaine hydrochloride containing methylparaben preservative [11]. Other members of the polyethylene glycol family and buffers and other compounds which were found in some steroids are neurotoxic [22,23]. Also an injection of bupivacaine administered in the intrafascicular plane has been reported as a cause of axons degeneration [24]. Owen [25] postulated the preservatives used in local anesthetics may in combination cause a flocculation of steroid preparations.

Mackinnon et al. [16] experimented with different steroids in an animal study and reported as followed:

The degree of nerve damage depends on the agent which was used and the site in which the agent was injected [16,26,27]. Gentili et al. [26] performed experimental studies on Wistar rats and found that the site of injection was the most important factor for the degree prediction of nerve damage. For injections in the intrafascicular plane, the degree varied depending on the agents which were used. In some instances pathological changes could be seen in less than 30 minutes after injection [15].

Shooting pain at the time of the injection along with numbness, paresthesia and other sensory deficits and distortions in the median nerve supply area will be presented in a patient with MNI. Other reported symptoms include motor weakness, thenar atrophy, allodynia, hyperesthesia and hypoalgesia. The patient may also show positive Phalen's and Tinel's signs [1,11,13,15,17]. A transient exacerbation of CTS symptoms following the steroid injection is present in many patients. A median nerve damage should be considered if the symptoms persist for more than 48 hours [1,2,8].

B-mode ultrasonography helps in lesion localization. An usual finding for a postiatrogenic nerve injury is an enlarged hypoechoic nerve. A sonoelastography can evaluate the tissue stiffness and the elasticity of the peripheral nerve. A with sonoelastography observed change from stiff to soft tissue correlates with the peripheral nerve recovery over the time. Nerve stimulation studies are also important for an evaluation of the regenerative capacity and the state of the nerve [13,28]. Park et al. [13] reported serial changes of ultrasonography, sonoelastographic and electrodiagnostic studies to evaluate a median sensory neuropathy after a carpal tunnel steroid injection. A swelling of the median nerve and a focal heterogenous hyperechoic area lateral to the median nerve were observed in the proximal carpal tunnel 3 days after injection. The appearance of the hyperechoic area was red under sonoelastographic imaging, which indicated hard tissue. A conservative treatment was initiated. On day 14 after injection, a prolonged onset latency of compound muscle action potential of median nerve was revealed. Six months after injection, symptoms such as pain or paresthesia were cured. The hyperechoic area was no longer visible and the cross-sectional area of the median nerve had decreased. Under sonoelastography, the area appeared yellow-green, which indicated an intermediate tissue hardness. Also a shorter onset latency of compound muscle action potential was revealed.

Some physicians suggest an immediate surgical intervention of carpal tunnel release if symptoms of acute median nerve compression were present after steroid injection. Other physicians believe conservative management alone should be appropriate for an incomplete nerve injury [1,11].

The management recommended in general is in continuity similar to that of other nerve lesions. A nonsurgical management consists of activity modification, splinting physiotherapy, medication and local steroid injections [13,14]. If no spontaneous resolution or no sign of recovery can be observed for over 3 months, then a surgery may be considered. Other indications for a surgery include a severe neuromotor functional loss, a severe debilitating pain or if the injected agent is known to be extremely neurotoxic [11,15,26,29,30].

A surgical intervention such as neurolysis, resection or graft repair is recommended for complete lesions. Also a gentle debridement of the crystal from the nerve may be performed [12]. An appropriate graft availability, difficulties in obtaining the autologous grafts, unsatisfactory cable grafts and rejection of homografts can be issues which are associated with the resection. Therefore, a resection is recommended only for cases in which the nerve function is lost completely and a self-recovery cannot be expected, when a neuroma has replaced the nerve or when the nerve action potential is through the neuroma negative or with the possibility to obtain a good length of the dissected nerve for an opposing of the resected ends. Intraoperative electrophysiological techniques may improve the quality of the nerve repair [31]. With a regenerative capacity, the surgeon should thus consider that the recovery occurring after surgery might be due to the natural course of nerve regeneration [29].

Carpal tunnel injections should be used at a minimum. However, the optimal management of CTS is still controversy. Phalen [2] reported a thickened flexor synovium in most of CTS cases and prescribed steroids to reduce the inflammation of flexor tenosynovitis. He suggested up to 3 steroid injections should be attempted in patients with CTS before a surgery should be considered. Marshall et al. [32] reported 2 steroid injections compared with 1 injection did not provide further clinical improvement. They concluded also local steroid injections did not improve the clinical outcome significantly when compared to either anti-inflammatory drugs and splinting after 8 weeks or helium-neon laser treatment after 6 months. Other investigators suggest steroid injections for the CTS management should be used in cases in which temporary relief is required, such as pregnancy, change in occupation or myxedema [1,10,17]. Some surgeons believe steroid injections provide only symptomatic relief and the medial nerve degeneration continues until the surgery becomes inevitable [14]. Ly-Pen et al. [33,34] evaluated the effect of steroid injections and surgery in patients with CTS over a 2 years period. At the 1-year follow-up, little differences in the symptomatic relief was seen between the two treatment options. However, the surgery had an additional long-term benefit at the 2-year follow-up. In another recent study, the authors found that while both local steroid injection and decompressive surgery were clinically effective in reducing symptoms of CTS, the neurologic parameters of the nerve conduction study were improved only in patients with surgery at 12-months follow-up [35].

If the steroid injection is required, careful attention to anatomic landmarks and an appropriate needle positioning is vital for nerve injury prevention. Most widely accepted approach is in which the needle is inserted just ulnar to the palmaris longus tendon at the wrist crease [3,11,36]. Other routes of injections based on anatomic landmarks have been described [11,12,17,36,37,38,39]. Recently, ultrasound-guided carpal tunnel injections have been raised in several studies [18,19,40]. Ultrasound provides real-time anatomic image, therefore it could improve the accuracy of the intervention (Fig. 1). A randomized prospective study compared the efficacy of ultrasound-guided and blind steroid injection in patients with CTS. Although both methods were effective in reducing the symptoms of CTS, earlier onset and better improvement of symptom relief were detected in the ultrasound-guided group [40]. However, more studies are needed to clarify the safety issue that image-guided injection could significantly reduce the incidence of MNI in patients with CTS. Several recommended methods for preventing MNI are summarized in Table 1.

To choice a least neurotoxic agent is also an important factor. Frederick et al. [12] used betamethasone, although McConnell and Bush [1] recommend the use of a soluble dexamethasone preparation. Also the use of a blunt beveled 22-gauge needle may diminish the risk of nerve penetration and compromise [41]. An injection administered into the skin can cause pigment changes and fat atrophy in the subcutaneous layer. Hence, the steroid should be injected only into the carpal tunnel.

During the CTS management, the patient should not be heavily sedated and should be encouraged to inform the interventionist if he/she experiences numbness/paresthesia during the procedure immediately. In this case the needle should be repositioned.

When a steroid injection is given for CTS treatment, pain may be caused due to transient inflammatory changes. Such changes may be confused with a median nerve injury. Therefore the follow-up with the patient is important. If the pain symptom continues 48 hours post injection, a median nerve injury should be suspected [1,11,13,17].

A MNI due to a carpal tunnel injection is a preventable complication. Injections should be used at a minimum. If an injection is needed, the appropriate needle positioning is vital to prevent a nerve injury. If a nerve injury is suspected, an immediate and optimal management can reduce the neurological sequels and maximize the recovery.