Manuscript

Interposition grafting of upper-extremity nerve injuries is often required to perform a tension-free coaptation. Classically, the sural nerve has been the predominant source of nerve autograft, especially when a larger amount of graft material is required. Although sural donor site morbidity is minimal, it does result in diminished sensation at the lateral foot and a visible scar. For upper-extremity reconstruction, sural graft harvest involves a second operative extremity, awkward intra-operative positioning, and/or position changes that may lengthen operative time; the latter are especially true in obese patients [13].

Upper-extremity sources for nerve autograft have several potential advantages, which include confining donor morbidity to the affected limb, limiting additional incisions, and simplicity of harvest. Several upper-extremity nerve grafts have been described, including the medial antebrachial cutaneous nerve (MABC) [15, 31], lateral antebrachial cutaneous nerve (LABC) [15, 28–30, 45], median nerve fascicular branch to the third webspace (TWM) [5, 38, 41], palmar-cutaneous branch of median nerve (PCM) [1], dorsal-cutaneous branch of ulnar nerve (DCU) [3, 19], and posterior interosseous nerve (PIN) [22].

A summary of the literature on donor options for nerve autograft, including specific advantages and disadvantages, is shown in Table 1. Most studies have focused on specific graft options or reconstruction of specific nerve injuries. The results following nerve grafting are well described [9, 26, 27, 36, 40, 44, 51]. The purposes of this study were to review our experience using nerve autografts in upper-extremity nerve reconstruction and develop recommendations for donor site selection with an emphasis on alternatives to sural nerve.

Table 1 Characteristics of nerve autograft donors based on published data
Full size table

Materials and Methods

A retrospective case series study was performed of all consecutive patients undergoing nerve grafting procedures for upper extremity nerve injuries over an 11-year period (2001–2012). Pediatric patients (<18 years) were excluded. Patient consent and study approval from the Institutional Review Board were obtained.

Eligible patients were identified from a prospectively maintained database of nerve surgeries performed by the senior author (SEM). Data on patient demographics, mechanism and timing of injury, nerve donor, nerve recipient, gap length, graft length, cable number, and operative site were collected. Descriptive statistics (means, standard deviations, frequencies, percentages) were performed for all variables, and an algorithm for upper-extremity nerve reconstruction was created.

Results

Patients and Nerve Injury

Eighty-six patients received 109 nerve grafts for upper extremity nerve reconstruction over the study period. Mean patient age at time of surgery was 42.9 ± 18.3 years, and 57 % were male. Mechanism of nerve injury was documented in 80 patients and included sharp laceration (37 %), iatrogenic lacerations (24 %), traction (9 %), neoplasm (7 %), gunshot (6 %), fracture (3 %), crush (2 %), and other blunt trauma (1 %). Nerve repairs were performed at an average of 13.3 ± 17.2 months after primary injury, and 48 patients (56 %) had undergone one or more prior attempts at nerve repair. In 14 patients (16 %), multiple nerves were injured but only 6 patients (7 %) required repair of multiple nerves with a nerve graft.

There were 51 median (59 %), 26 ulnar (30 %), 14 digital (13 %), 13 radial (16 %), and 3 musculocutaneous (4 %) nerve injuries repaired with nerve graft. Pure sensory nerve injuries were most common (n = 47, 53 %), followed by mixed (n = 32, 36 %) and pure motor nerve injuries (n = 10, 11 %).

Nerve Graft Donors

In total, 99 nerve autografts (71 from upper extremity, 28 from lower extremity) and 10 acellular nerve allografts (ANA) were used. The type and characteristics of nerve grafts utilized in this cohort are summarized in Table 2. Figure 1 demonstrates which nerve grafts were utilized by nerve injured.

Table 2 Type and characteristics of 109 nerve grafts utilized to reconstruct upper-extremity nerve injuries
Full size table
Fig. 1
figure 1

Upper-extremity graft use by injured nerve. Histogram representing the frequency of use of each donor nerve and the nerve repaired with that donor. Includes ANAs used to facilitate and end-to-side repair of nerve graft donors and repair of other sensory nerves not specifically listed.

Full size image

Multiple upper-extremity nerve autograft donors were utilized, including the MABC, TWM, LABC, PCM, and DCU. By using an upper-extremity donor, a second operative limb was avoided in 58 patients (67 %) and a second incision was avoided in 26 patients (30 %). Lower-extremity autograft donors included the sural nerve and obturator nerve. Six patients (7 %) required bilateral sural nerve harvest. The frequency of sural nerve graft use declined from 40 % (n = 17/43) to 11 % (n = 7/64) between the first and second half of the study period, with a corresponding increase in upper-extremity nerve donor utilization (Fig. 2). The sural was used primarily to repair complex injuries involving multiple nerves.

Fig. 2
figure 2

Utilization of different nerve grafts over time. Histogram comparing the use of each donor nerve in the first half of the study period to the second half. Sural nerve use decreased as use of upper extremity donors increased

Full size image

Where possible, the distal cut end of the donor nerve was repaired in an end-to-side fashion to an adjacent, normal sensory nerve. This was achieved in 25 of 99 autografts, specifically following MABC (n = 13), TWM (n = 7), and PCM (n = 3) harvest. In two cases where a direct end-to-side repair was not possible, ANA was utilized to bridge the nerve gap.

An algorithm for nerve graft selection is proposed in Fig. 3.

Fig. 3
figure 3

Algorithm for sural alternatives. Flow diagram representing our algorithm for choosing donor nerves. ANA acellular nerve allograft, DCU dorsal cutaneous branch of ulnar nerve, LABC lateral antebrachial cutaneous nerve, MABC medial antebrachial cutaneous nerve, PCM palmar cutaneous branch of median nerve, TWM third webspace branch of median nerve

Full size image

Discussion

Nerve grafting remains an important reconstructive technique for acute management of upper-extremity nerve injuries. Our algorithm for selection of nerve graft material has evolved over the last decade owing to several factors. First, a growing understanding of upper-extremity nerve anatomy and internal topography now allows alternative reconstructions with nerve transfers and has expanded the available options and harvestable length of upper-extremity nerve donors. Second, the ongoing drive to minimize additional incisions, maximize ease of harvest, and limit donor morbidity has led us away from using the sural nerve when possible. Third, with the increasing prevalence of obesity in North America, the added difficulty and operative time required for sural nerve harvest is not insignificant. As a result, we now use the sural nerve only for situations requiring large amounts of graft material, such as multiple major nerve transection injuries.

For distal nerve injuries in the forearm, the MABC remains our first choice for repair of median and ulnar nerve injuries because of its caliber, available length, and proximity. Donor morbidity from harvesting the MABC can be minimized by performing an end-to-side anastamosis of the distal cut end of the MABC to the adjacent median nerve. Recent studies show that while motor axons require injury for end-to-side sprouting, sensory axons collaterally sprout spontaneously without need for additional axonal injury [35, 46]. Restoring some degree of innervation to the donor nerve territory may reduce forearm anesthesia or hypersensitivity caused by sprouting of adjacent sensory nerves after MABC harvest [11, 13]. This concept can also be applied to harvest of other nerves, such as the TWM.

Utilizing non-critical portions of the injured nerve also helps to minimize donor morbidity, avoid additional incisions, and minimize operative time. For example, the DCU becomes a useful donor in more proximal ulnar nerve injuries, as do the PCM and TWM in median nerve injuries. In this study, using DCU and TWM for graft material avoided a second incision in 3 of 6 patients and 10 of 14 patients, respectively. Owing to its small caliber and short length, we tend to reserve the PCM as a secondary source of graft material in distal median nerve injuries where the TWM alone was insufficient. Addition of the PCM in this review eliminated the need for a second incision in four of eight patients.

Along the same lines, our preference for radial sensory nerve reconstruction for short gap injuries in the forearm is the LABC because of its proximity, ease of harvest, and the likelihood that it is also injured in distal forearm injuries. Although the radial sensory nerve does not provide critical sensation to the hand, we prefer to reconstruct it because of the propensity for hyperalgesia in this region secondary to collateral sprouting [11]. Alternatively, in high radial nerve injuries, we will crush and proximally transpose the injured radial sensory nerve to avoid formation of a painful neuroma when reconstruction is not feasible. The distal radial sensory nerve can then be transferred end-to-side to the normal median nerve for recovery of sensation. We strongly advise against harvesting an uninjured radial sensory nerve as graft material.

Our preference for digital nerve reconstruction depends on the specific nerve injured. Critical sensory nerves supplying the ulnar and radial borders of the hand should be reconstructed with nerve autograft. Multiple donor options exist; however, we prefer the MABC and LABC owing to their size match and limited donor morbidity. The distal PIN is described as a graft source for digital nerve reconstruction [22] but it is small in diameter and leaves a visible scar on the dorsal forearm. Therefore, we tend not to use this graft.

We reserve the use of ANA for non-critical digital nerve injuries less than 3 cm in length. While a comprehensive discussion on the advantages and disadvantages of ANA is outside the scope of this manuscript, evidence supporting good outcomes following ANA reconstruction of proximal nerve injuries or large nerve gaps is lacking [4, 8]. In fact, we will utilize long ANA segments specifically when we wish to encourage incomplete nerve regeneration, as in the management of painful neuromas. As such, our personal practice is to use autogenous nerve for critical sensory or motor nerve repair.

The vast majority of autograft donors in this study were sensory nerves. In four cases, we utilized the obturator motor nerve for reconstruction of motor nerve injuries based on the hypothesis that motor and sensory Schwann cell specificity would markedly improve regeneration [21]. However, subsequent experimental work revealed that nerve architecture and endoneurial tube size are likely more important factors in facilitating nerve regeneration [23, 34]. Our current indications for harvesting a motor nerve as a graft source are therefore limited to critical motor nerve reconstruction not amenable to motor transfer, such as the deep ulnar motor branch in the hand and the spinal accessory nerve.

The advent of distal nerve transfers has limited our requirements for long nerve grafts. Our improved understanding of nerve injury and regeneration now cautions us against the use of long grafts (>6 cm) whenever possible [42]. A state of Schwann cell senescence following denervation was recently described and correlated with failures of axonal regeneration through long nerve grafts [42]. Irreversible arrest of proliferation, altered gene expression, and changes in the secretory profile characterize the senescent state [2, 6, 7, 10, 12, 14, 18, 24, 32, 33, 37, 39, 43, 47, 48]. Compared with short grafts, long and large diameter nerve grafts expose distal Schwann cells to prolonged ischemia-related oxidative stress and prolonged denervation, which may induce Schwann cell senescence [42, 49, 50]. This phenomenon may explain poor regeneration across long grafts and grafts without immediate distal repair [17, 20, 25].

Based on this knowledge, we prefer to acutely manage single nerve injuries in the upper arm with distal nerve or tendon transfers rather than direct repair with nerve grafting (Table 3). Increased experience with nerve transfer and the use of MABC as graft material has reduced the need for sural nerve graft. We now only address a proximal single nerve injury if there is neuropathic pain. In these situations, rather than directly repair the nerve with a nerve graft, we address the injury with neuroma excision, and crushing the nerve proximal to the injury to create a neurotmetic injury and thus, “reset” the regenerative process. Finally, the nerve end is transposed into an adjacent muscle for pain control. As a result, even if nerve grafts are needed to complete a portion of reconstruction, donors from the upper extremity typically suffice. Sural nerve grafting is reserved for multiple major nerve transection injuries, where distal nerve or tendon transfers are not possible.

Table 3 “Recipes” for high-level nerve injuries as alternatives to sural graft
Full size table

In summary, upper-extremity nerve reconstruction can be successfully accomplished using nerve donors from the ipsilateral limb in the majority of cases. Judicious selection of nerve donors can reduce patient morbidity and operative time but necessitates a thorough understanding of nerve anatomy and topography. While the sural nerve remains a useful donor for extensive, multiple nerve injury reconstruction, a growing body of scientific and clinical evidence suggests that in situations requiring large amounts of nerve autograft, alternative reconstructions such as nerve transfers may provide better results [16, 36]. In cases of distal injury requiring interposition grafting, we recommend use of alternative graft material including TWM and DCU for median and ulnar nerve injuries respectively, as well as MABC and LABC.