A Comprehensive Review of Topography and Axon Counts in Upper-Extremity Peripheral Nerves: A Guide for Neurotization

Abstract

Purpose

Currently, no comprehensive database detailing topography and axon counts exists. This study aims to review the axon counts and topography of the major peripheral motor nerves of the upper extremity to allow for optimal surgical planning for peripheral nerve reconstruction via neurotization.

Methods

Peer-reviewed journal articles were identified through PubMed, ScienceDirect, Google Scholar, and CENTRAL. Studies were included for review based upon the identification of the described topography or axon count of any upper-extremity peripheral motor nerve. Animal research, laboratory studies, and unpublished studies were excluded from our review. A total of 43 studies were identified, and 38 of these met the inclusion criteria. Statistical analysis was performed to determine axon count averages for all upper extremity motor nerves identified in the included studies.

Results

Thirty-eight studies were reviewed, giving insights into the topography and axon counts of the major peripheral nerves of the upper extremity, including the brachial plexus and its terminal branches as well as common donor nerves such as the spinal accessory nerve and intercostal nerves. Studies showed considerable variability in reported axon counts.

Conclusions

Existing data were relatively weak and included several case reports and series. Taking this into consideration, we posit that there is a need for further studies of upper-extremity nerve axon counts that include large study populations and more consistent methods of nerve specimen analysis.

Clinical relevance

Understanding the topographical anatomy of donor and recipient nerves, as well as appropriately matching the motor axon counts for each donor and recipient, is helpful in upper-extremity nerve reconstruction.

Peripheral nerve injuries range from mild neurapraxia, with potential for spontaneous functional and sensory recovery to neurotmesis, associated with devastatingly debilitating functional and sensory deficits that require surgical intervention. Studies have documented a wide array of nerve transfer techniques for the restoration of both sensory and motor function in the setting of these more severe peripheral nerve injuries. Although a meticulous surgical technique is always helpful in nerve reconstruction, successful nerve transfer and grafting also depends on understanding the topographical anatomy of the donor and recipient nerves, as well as appropriately matching the motor axon counts for each donor and recipient. For example, a hypothetical motor nerve containing five axons should not be used as a donor for a recipient nerve innervating a muscle that normally requires 10 axons of input. Donor-to-recipient nerve axon count ratios below 0.7:1 have been associated with a decreased likelihood of a successful outcome.¹

Upper-extremity peripheral nerve topography and axon counts have been described in various literature sources, ranging from reference textbooks to journal articles.1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41 With regard to existing literature, the description of axon counts in upper-extremity nerves are largely described in case studies detailing specific nerve transfers. Textbooks have provided descriptions of specific nerve topography, but not in the context of nerve transfer studies.³⁹ No singular, comprehensive database of upper-extremity peripheral nerve topography and axon counts exists to date. Having access to such a database would give the surgeon a simple resource for planning and implementing successful nerve transfers, including specific nerve combinations as well as an ideal level of donor harvest. This study aims to identify the axon counts of the major peripheral motor nerves of the upper extremity and review nerve topography as it relates to axon count to allow for optimal surgical planning for peripheral nerve reconstruction via neurotization.

Methods

Peer-reviewed journal articles pertaining to upper-extremity nerve axon counts and topography were identified through PubMed, ScienceDirect, Google Scholar, and CENTRAL. Studies were included for review based upon the identification of the described axon count of any upper-extremity peripheral motor nerve. Animal research, laboratory studies, and unpublished studies were excluded from our review. A total of 43 studies were identified, and 38 of these met the inclusion criteria. Statistical analysis was performed to determine axon count averages for all upper-extremity motor nerves identified in the included studies (Supplemental Fig. S1, available online on the Journal’s website at https://www.jhsgo.org). All figures were created using biorender.com.

Upper extremity nerve topography

An understanding of relative axon counts is necessary when designing appropriate nerve transfers and grafts. Prior to nerve selection, however, cognizance of nerve topography is essential to ensuring that the coaptation between donor and recipient nerve is created properly and efficiently. Although it has been well documented that motor axons will preferentially regenerate down motor pathways after repair of nerves with both motor and sensory components, knowledge of nerve topography plays an important role in cases of size mismatch of a mixed nerve in which the appropriate fascicle must be chosen for coaptation to improve clinical outcomes.^39,40

For example, suturing a nerve graft to a recipient suprascapular nerve (SSN) may fail to restore motor function if the graft is erroneously connected to donor sensory fascicles from C5. Prior to delving into the axonal counts of the various nerves of the upper extremity, we will briefly review the normal topography of the brachial plexus and its terminal branches.

Topography of the brachial plexus roots

Within the C5 nerve root, fascicles destined for the SSN run through the cranioventral quadrant. The ventrocaudal quadrant contains lateral cord (LC) contributions to the anterior division of the brachial plexus as well as musculocutaneous nerve (MCN) fascicles. The dorsal aspect of C5 is composed of axons destined for the posterior cord (PC), with roughly 20% of the rostrodorsal axons contributing directly to the axillary nerve (AN) (Fig. 1).2, 3, 4

Figure 1.

Brachial plexus nerve root topography.

The C6 spinal nerve contains fascicles of the MCN, lateral pectoral nerve (LPN), lateral root of the median nerve (LRMN), radial nerve (RN), and AN. The ventral portion of the nerve is composed primarily of MCN fascicles, whereas fascicles of the LPN and LRMN occupy the ventrocaudal and dorsocaudal quadrants, respectively. Radial nerve and AN fascicles occupy the dorsal half of the spinal nerve, with the AN fascicles located in the rostral-most part of the dorsal quadrant. The rest of the dorsal quadrant was occupied by the RN fascicles.²

The C7 spinal nerve consisted of the fascicles of LRMN, LPN, ulnar nerve (UN), and RN. The ventral half of C7 is mostly composed of LRMN fascicles, but a small portion of the ventrocaudal quadrant contains LPN fascicles. The dorsal half of C7 contains RN and UN fascicles with an orientation such that the RN fascicles are found more cranially, and the UN fascicles occupy the dorsocaudal quadrant.²

Most fascicles destined for the medial root of the median nerve (MRMN) are contained within the C8 spinal root, with roughly the entire ventral half consisting of MRMN fascicles. The dorsal half of C8 contains contributions to the UN, PC, medial cutaneous nerve of the arm, and medial cutaneous nerve of the forearm. The PC fascicles occupy most of the dorsal half of the spinal nerve. The UN, medial cutaneous nerve of the arm, and medial cutaneous nerve of the forearm fascicles are found caudally and are abutted on either side between the PC and MRMN fascicles.²

The T1 spinal nerve consisted of fascicles of the medial cord (MC) on the ventral aspect. The UN fascicles are located dorsally, with PC fascicles flanking the UN fascicles in the rostrodorsal and dorsocaudal quadrants.² Figure 1 summarizes the topography of the brachial plexus spinal nerve roots.

MCN topography

The MCN is a continuation of the LC and contains fascicles that will diverge to supply the motor nerves to the biceps brachialis and brachialis musculature as well as convey sensory input from the lateral antebrachial cutaneous nerve (LACN). As the nerve courses distally, the motor fascicles are arranged from lateral to medial in the order of nerve branch to biceps brachialis, nerve branch to brachialis, and LACN fascicle, shown in Figure 2 below.5, 6, 7

Figure 2.

Musculocutaneous nerve topography.

AN topography

The AN is a terminal branch of the PC responsible for innervation of the deltoid and teres minor muscles; there has been documentation of anatomic variation in which the long head of the triceps also receives innervation from the AN.^8,41 Approximately 3.5 cm proximal to passing through the quadrangular space, the AN divides into an inferomedial branch, carrying fascicles to teres minor, and a superolateral branch containing fascicles to the deltoid muscle.⁸ Figure 3 illustrates this described topography.

Figure 3.

Axillary nerve topography.

RN topography

The RN is a terminal offshoot of the PC. Most proximally, the RN contains fascicles of the three heads of the triceps brachii, extensor carpi radialis brevis, extensor carpi radialis longus, brachioradialis, and posterior interosseous nerve (PIN).⁹ The branch to the long head of the triceps is the most medial fascicle, branching away early in the course of the nerve. Following its takeoff, the posteromedial aspect of the RN at this level is occupied by PIN fascicles and the fascicle of the medial head of the triceps. The anterolateral aspect of the RN contains the fascicle of the lateral head of the triceps. Medial to the lateral head fascicle lies the superficial branch of the RN (SBRN), along with posterior cutaneous nerve fascicles oriented anteriorly and posteriorly, respectively. The fascicles of extensor carpi radialis longus (ECRL), extensor carpi radialis brevis (ECRB), and brachioradialis run through the center of the nerve. As the branches to the triceps divide from the RN proper along with the posterior cutaneous nerve of the arm, the ECRL/ECRB fascicles take a more lateral position, and the PIN fascicles become more posteriorly oriented; likewise, the fascicle to brachioradialis shifts anteriorly, moving the sensory fascicle of SBRN medially. Proximal to the elbow, the ERCL and brachioradialis fascicles will split from the RN toward their targets. Following the ECRL and brachioradialis takeoff, the RN continues into the radial tunnel. At this level, the ECRB fascicle continues laterally, whereas the SBRN courses more anteriorly; the supinator fascicle separates from the PIN, running medially, with the remainder of the PIN fascicles staying posterior within the RN. Continuing distally, the SBRN will shift laterally prior to dividing away from the PIN, which itself continues distally toward its targets (Fig. 4).

Figure 4.

Radial nerve topography.

Median nerve topography

The median nerve is formed from contributions of the medial and LCs, with most motor axons coming from the MC, and the majority of sensory axons derived from the LC. As the nerve coalesces in the upper arm and runs along the brachial artery, the motor fascicles to pronator teres, flexor carpi radialis, palmaris longus, flexor digitorum superficialis, and the anterior interosseous bundle maintain their medial orientation within the nerve; the sensory component remains lateral and contains the recurrent thenar motor fascicle. As the nerve continues distally toward the proximal forearm, the sensory component intervenes between the anterior interosseous axons and other motor fascicles. The pronator teres fascicle is on the most anterior aspect of the median nerve, followed medially by the flexor carpi radialis, palmaris longus, and flexor digitorum superficialis fascicles. The thenar motor component remains within the sensory component at this level. The median nerve eventually diverges into its continuation and the branching anterior interosseous nerve (AIN). The AIN contains three fascicles: flexor pollicis longus, flexor digitorum profundus, and pronator quadratus. The flexor digitorum profundus and flexor pollicis longus fascicles occupy the anterior AIN in medial and lateral orientation, respectively. The pronator quadratus fascicle occupies the posterior AIN. Within the continuation of the median nerve, the recurrent thenar fascicle remains the only motor component, terminating in the thenar musculature of the hand, as shown in Figure 5.^10,11

Figure 5.

Median nerve topography.

UN topography

The UN is the distal continuation of the MC. It mainly contains fibers from C8 and T1 but can sometimes carry fibers from C7 via a communicating branch from the LC. It is a mixed nerve with motor innervation to the flexor carpi ulnaris, the medial half of flexor digitorum profundus, and most of the intrinsic muscles of the hand, except the thenar muscles and lateral two lumbricals. Its sensory component innervates the anterior and dorsal aspects of the small finger and half of the ring finger, medial palmar skin, and medial aspect of the dorsum of the hand.

After arising from the brachial plexus, the UN traverses medially in the upper arm. At this level, the motor fascicles to the flexor carpi ulnaris (FCU) are located in the anterolateral portion of the nerve. The motor fascicles to the flexor digitorum profundus (FDP) are also located in the lateral portion, but posterior to the motor fascicles to the FCU. The sensory fascicles to the dorsal cutaneous nerve are located more centrally in the posterior portion of the nerve. Most of the fascicles at this level are mixed motor and sensory fascicles to the terminal cutaneous nerves and intrinsic muscles of the hand; these mixed fascicles are located in the medial aspect of the nerve, extending anteriorly and posteriorly.

In the proximal forearm, the UN has sensory fascicles in both the lateral and medial portions of the nerve with motor fascicles in the central portion of the nerve. In the distal forearm, sensory fascicles run laterally, and motor fascicles run medially, with a separate dorsal cutaneous sensory branch nerve running medial to the UN.

At the level of Guyon canal, sensory fascicles throughout the whole nerve innervate the ulnar aspect of the small finger posteromedially and palmar cutaneous nerve distribution anteromedially. Additionally, a branch comes off the UN medially, containing motor fascicles to the deep muscles of the hand (Fig. 6).

Figure 6.

Ulnar nerve topography.

Upper extremity axon counts

With an understanding of normal topography of the brachial plexus and its terminal nerve branches, the next step in choosing appropriate nerves for neurotization is to consider each nerve’s average axon counts. The surgeon should be familiar with both the donor and recipient nerve axon counts to avoid coapting nerves that would not have the appropriate ratio to restore function. Schreiber found that there was a significant correlation between clinical outcomes of transfers and axon counts when attempting to restore elbow function.¹ Per Schreiber’s findings, an appropriate axon count ratio of the donor nerve to recipient nerve was 0.7:1 and appeared to contribute to improved outcomes. We will briefly review the axon counts of the upper-extremity nerves so that the surgeon may have a database for planning appropriate nerve coaptations.

Brachial plexus nerve roots

In a case study harvesting plexus nerve roots from nine human heart-beating organ donors, motor and sensory axons were cataloged and found to have a sensory axon: motor axon ratio of at least 9:1. The mean total number of ventral rami axons present at the level of respective neuroforamina C5 to T1 was found to be 350 ± 43 × 10³. The relative axon count of C5 was 38.210 ± 9.388 ×10³. The relative axon count of C6 was 75.86 ± 9.86 × 10³. The relative axon count of C7 was 84.01 ± 10.58 × 10³. The relative axon count of C8 was 90.38 ± 12.53 × 10³. The relative axon count of T1 was 61.421± 22.815 ×10³ (Table 1).¹²

Table 1.

Brachial Plexus Nerve Root Axon Counts

Article	Location	Axon Counts
Gesslbauer et al12	C5–T1	350 ± 43 × 103
	C5	38.210 ± 9.388 × 103
	C6	75.86 ± 9.86 × 103
	C7	84.01 ± 10.58 × 103
	C8	90.38 ± 12.53 × 103
	T1	61.421 ± 22.815 × 103

Upper, middle, and lower trunks

Axon counts of the upper, middle, and lower trunks are not well described in the literature. There is a need for nerve sampling and axon histomorphometry of the trunks at their specific levels between the nerve roots and cords (Table 2).

Table 2.

Brachial Plexus Trunk Axon Counts

Article	Trunk	Axon Counts
Gesslbauer et al12	Upper (C5 + C6 average)	19.14293 ± 9.46386 × 103
	Middle (C7)	84.01 ± 10.58 × 103
	Lower (C8 + T1 average)	30.75569 ± 22.82753 × 103

Spinal accessory nerve

Several studies describe the axon count of the spinal accessory nerve from cadaveric dissection via histomorphometry. Two of these studies obtained nerve biopsies at multiple points throughout the course of the nerve, and then, either reported the axon count at each distinct level or reported an average of the axon counts based on each level.^13,14 The remaining two studies simply provide a relative axon count found in a single nerve sampling.^12,15

Placheta et al¹³ performed a histomorphometric study of the spinal accessory nerve at multiple landmarks along its course, including just distal to the skull base, caudoposterior to the sternocleidomastoid muscle, proximal to the trapezius muscle, and at the level of the donor nerve harvest, which was a small distal branch that exited the main nerve trunk at a 90-degree angle medially. The axon counts of these four samples were averaged and reported to be 1.400 × 10³. Similarly, a case report by Vathana et al¹⁴ analyzed the spinal accessory nerve at multiple targets along bony landmarks (T1 spinous process, acromioclavicular joint, posterolateral corner of the acromion, and angle at the superior medial border of the scapula) when harvesting the nerve from a posterior approach and reported its average axon counts to be 1.328 × 10³ at the proximal aspect of the nerve, 1.021 × 10³ at the midway point of the nerve, and 0.817 × 10³ at the terminal aspect of the nerve. Proximal, midway, and terminal definitions were relative to the length of nerve they harvested via the posterior approach.

In a case report describing a posterior approach technique for spinal accessory-SSN transfer, the spinal accessory nerve was identified at a single point with the cadaveric shoulder in a prone position using the trapezius muscle splitting technique. Histomorphometric evaluation showed that the nerve contained a mean number of 1.602 × 10³ myelinated axons.¹⁵ In another study evaluating the spinal accessory nerve at a single point via cadaveric dissection, the mean axon count was 4.239 ± 2.031 ×10³ with 23.24% ± 6.01 of those axons being motor.¹² A summary of these axon counts is listed in Table 3.

Table 3.

Spinal Accessory Nerve Axon Counts

Article	Location	Axon Counts
Placheta et al13	Four locations along nerve course	1.400 × 103
Vathana et al14	Proximal	1.328 × 103
	Midpoint	1.021 × 103
	Terminal	0.817 × 103
Pruksakorn et al15	Single point	1.602 × 103
Gesslbauer et al12	Single point	4.239 ± 2.031 × 103

Dorsal scapular nerve

In a cadaveric study on the utility of the dorsal scapular nerve as a donor for brachial plexus reconstruction, 12 brachial plexus dissections were performed to reveal an average axon count of the dorsal scapular nerve as 0.8418 ± 0.198 × 10³ (Table 4).¹⁶ More studies are needed to provide a more accurate estimation.

Table 4.

Dorsal Scapular Nerve Axon Counts

Article	Axon Counts
Saltzman et al16	0.8418 ± 0.198 × 103

Suprascapular nerve

Regarding topography, the SSN is a monofascicular motor nerve that largely originates from the dorsal and ventral aspects of C5. Regarding axon counts, two studies were identified, which quantified the average number of myelinated axons in the nerve specimen. Both studies were cadaveric anatomical studies. Pruksakorn et al¹⁵ reported that the SSN contained an average number of 6.004 × 10³ axons, with an average number of 12 fascicles when studied as the recipient of a nerve transfer from a donor spinal accessory nerve. Saltzman et al¹⁶ studied the SSN as a potential recipient for neurotization via the levator scapulae motor nerve and found that the nerve contained 3.361 ± 627 × 10³ axons (Table 5).

Table 5.

Suprascapular Nerve Axon Counts

Article	Axon Counts
Purksakorn et al15	6.004 × 103
Saltzman et al16	3.361 ± 627 × 103

Thoracodorsal nerve

Existing literature about the average axon counts of the thoracodorsal nerve describes a wide range of axons. Only one study separates the axon counts of the thoracodorsal nerve into lateral and medial branches. This study analyzed the thoracodorsal nerve as a potential donor for the treatment of winged scapula and found that the lateral branch contained a relative axon count of 1.853 × 10³, and the medial branch contained a relative axon count of 0.974 × 10³.¹⁷ Another study analyzing potential donor nerves in brachial plexus reinnervation found that the thoracodorsal nerve contained an average axon count range of 1.530 × 10³ to 2.470 × 10³.¹⁸ Finally, Gesslbauer cataloged the nerves of the upper extremity and reported that the thoracodorsal nerve had 6.904 ± 3.070 × 10³ axons with 15% of those axons being motor (Table 6).¹²

Table 6.

Thoracodorsal Nerve Axon Counts

Article	Location	Axon Counts
Raksakulkiat et al17	Lateral branch	1.853 × 103
	Medial branch	0.974 × 103
Samardzic et al18	Average range	1.530 × 103 to 2.470 × 103
Gesslbauer et al12	Single point	6.904 ± 3.070 × 103

Long thoracic nerve

Regarding topography, the long thoracic nerve is a monofascicular motor nerve to the serratus anterior. Only one study was found that quantified the axon count of the long thoracic nerve in its analysis as a potential donor for brachial plexus reconstruction. This study reports the axon count of the long thoracic nerve to be 1.746 ± 332 × 10³ (Table 7).¹⁶

Table 7.

Long Thoracic Nerve Axon Counts

Article	Axon Counts
Saltzman et al16	1.746 ± 332 × 103

Medial pectoral nerve

Two studies were found, which described the axon count of the medial pectoral nerve. The first was a case study analyzing upper extremity nerve histology from 15 autopsies and 13 surgical patients with total or partial spinal root avulsions. It found that the relative axon count of the medial pectoral nerve ranged from 1.170 × 10³ to 2.140 × 10³.¹⁸ In another case study describing upper-extremity nerve dissections in 43 cadavers and five living patients, the medial pectoral nerve was found to contain an average of 0.500 × 10³ axons (Table 8).¹⁹ Between these two studies, the axon count of the nerve varies greatly and clearly demonstrates a need for more analyses of the nerve to provide a more accurate estimation.

Table 8.

Medial Pectoral Nerve Axon Counts

Article	Axon Counts
Samardzic et al18	1.170 × 103 to 2.140 × 103
Norkus et al19	0.500 × 103

Lateral pectoral nerve

Regarding topography, the LPN arises from the LC of the brachial plexus and is composed of branches of C5, C6, and C7. It primarily supplies motor innervation to the pectoralis major muscle. Because of a communicating branch to the medial pectoral nerve, some LPN motor fibers innervate the pectoralis minor. According to a description of LPN transfer to the spinal accessory nerve, the relative axon count of the LPN was 2.637 ± 0.839 × 10³ (Table 9).²⁰

Table 9.

Lateral Pectoral Nerve Axon Counts

Article	Axon Counts
Maldonado et al20	2.637 ± 0.839 × 103

Lateral, medial, and PCs

The lateral, medial, and PCs were analyzed in one study, Gesslbauer’s description of axonal components of the upper-extremity nerves.¹² All three cords were harvested from nine fresh cadavers. The LC was found to contain 61.734 ± 15.659 × 10³ axons with 4.6% of those being motor axons. The MC contained 98.512 ± 23.433 × 10³ axons with 5.2% of those being motor axons. The PC contained 92.694 ± 12.276 × 10³ axons with 8.9% of those being motor axons (Table 10).¹²

Table 10.

Brachial Plexus Cord Axon Counts

Article	Cord	Axon Counts
Gesslbauer et al12	Lateral	61.734 ± 15.659 × 103
	Medial	98.512 ± 23.433 × 103
	Posterior	92.694 ± 12.276 × 103

Musculocutaneous nerve

The MCN was analyzed in four studies, and the noted axon counts varied greatly. In a cadaveric study of 21 dissections, the axon count numbers in 12 nerves were found to be 14.044 × 10³ in the MCN proximal to the biceps brachii branch and 6.877 × 10³ in the LACN branch.²¹ Another anatomic study found that average axon counts were 6.061 × 10³.²² A study by Schreiber et al¹ looking at axon counts for donor and recipient nerves used in brachial plexus transfers used 10 cadaveric studies and found that mean recipient axon count was 1.826 × 10³ for the biceps brachii branch of the MCN and 1.840 × 10³ for the brachialis branch of the MCN. Finally, Gesslbauer et al¹² found that the MCN axon counts were 25.904 ± 5.916 ×10³ with 6.41 ± 1.33% of those axons being motor (Table 11).

Table 11.

Musculocutaneous Nerve Axon Counts

Article	Location	Axon Counts
Brandt et al21	Proximal to biceps brachii	14.044 × 103
	LACN	6.877 × 103
Bonnel et al22	Average	6.061 × 103
Schreiber et al1	Biceps brachii branch	1.826 × 103
	Brachialis branch	1.840 × 103
Gesslbauer et al12	Single point	25.904 ± 5.916 × 103

Radial nerve

There were three studies looking at axon counts of the RN. Khair et al²³ dissected 10 fresh-frozen cadavers and identified the RN branches to the long, medial, and lateral heads of the triceps, which had a median count of 2.302 × 10³, 2.198 × 10³, and 1.462 × 10³ axons, respectively. Nair et al²⁴ analyzed nerves in 14 formalin-fixed adult cadavers and found that the mean number of axons was 1.015 ± 0.76651 × 10³ in the RN and 0.38146 ± 0.22370 × 10³ in the SBRN. Finally, Gesslbauer et al¹² analyzed RN axon counts as various points along its course, finding that the average counts were 65.696 ± 10.526 × 10³ at its origin, 41.496 ± 9.141 × 10³ at the midhumerus, and 28.349 ± 4.872 × 10³ at the lateral epicondyle. The percentage of axons consisting of motor nerves was 6.74 ± 1.01% at the origin, 7.13 ±1.35 at the midhumerus, and 8.83 ± 2.15 at the lateral epicondyle. Table 12 summarizes the aforementioned literature. Additionally, Moldanado et al report the average axon count to be 2.637 ± 0.839 × 10³ (Table 9).

Table 12.

Radial Nerve Axon Counts

Article	Location	Axon Counts
Khair et al23	Long head of triceps	2.302 × 103
	Medial head of triceps	2.198 × 103
	Lateral head of triceps	1.462 × 103
Nair et al24	Main branch	1.015 ± 0.76651 × 103
	Superficial branch	0.38146 ± 0.22370 × 103
Gesslbauer et al12	Origin	65.696 ± 10.526 × 103
	Midhumerus	41.496 ± 9.141 × 103
	Lateral epicondyle	28.349 ± 4.872 × 103

Axillary nerve

Few studies exist that analyze axon counts in the AN. In a study involving 10 fresh-frozen cadavers, Khair et al²³ found that the main AN trunk had 7.887 × 10³ axons and that the anterior, posterior, and teres minor branches had 4.052 × 10³, 1.242 × 10³, and 1.161 × 10³ axons, respectively. Meanwhile, Gesslbauer et al¹² found that the AN had 22.498 ± 2.911 ×10³ axons in the main trunk with 9.44 ± 1.27% of those axons being motor (Table 13).

Table 13.

Axillary Nerve Axon Counts

Article	Location	Axon Counts
Khair et al23	Main trunk	7.887 × 103
	Anterior branch	4.052 × 103
	Posterior branch	1.242 × 103
	Teres Minor branch	1.161 × 103
	Superficial branch	0.38146 ± 0.22370 × 103
Gesslbauer et al12	Main trunk	22.498 ± 2.911 × 103

Median nerve

Existing literature about the axon count of the median nerve demonstrates that this nerve contains the second greatest number of fascicles after the RN. There are reports that the number of fascicles of the median nerve also does not vary significantly, in contrast to the majority of the other upper-extremity nerve fascicle counts, but when comparing two major analyses, Mioton found that the average number of median nerve fascicles was 20.43 ± 7.29, whereas Schreiber et al¹ found that the median nerve contained eight fascicles.²⁵ Thus, there is still significant variability of fascicle counts among existing studies. There is also a significant variation in reported axon counts. For example, Bonnel et al²² report a median nerve axon count of 18.288 × 10³, whereas Schreiber et al¹ report an axon count of 1.860 × 10³. The most comprehensive upper-extremity axon count analysis by Gesslbauer analyzed the median nerve at various points in the arm and found that its origin contained 60.523 ± 6.280 × 10³ axons with 5.6% of those being motor, the nerve at the midforearm contained 41. 034 ± 6.455 × 10³ axons with 1.3% of those being motor, and the nerve at the wrist contained 42.811 ± 8.423 × 10³ axons with 1.1% of those being motor (Table 14).¹²

Table 14.

Median Nerve Axon Counts

Article	Location	Axon Counts
Bonnel et al22	Single point	18.288 × 103
Schreiber et al1	Single point	1.860 × 103
Gesslbauer et al12	Origin	60.523 ± 6.280 × 103
	Midforearm	41.034 ± 6.455 × 103
	Wrist	42.811 ± 8.423 × 103

Ulnar nerve

Only a handful of studies have evaluated the number of axons of the UN. Bonnel et al²² performed 100 dissections and found the average number of axons to be 16.412 × 10³ per nerve. Schreiber et al¹ performed 10 cadaveric dissections and identified an average of 7.9 fascicles within the UN with a mean axon count of 1.318 × 10³. More recent work by Gesslbauer and his colleagues expanded on that work by harvesting the UN at the midforearm and wrist and found that at the total axon count at the root of the brachial plexus was 40.379 ± 5.638 × 10³, with the percent of motor axons being 6.66 ± 0.7912 of the total. At the level of the midforearm, the total axon count was 30.915 ± 2.869 × 10³, and the percent of motor of the total was 3.88 ± 0.74. Finally, at the wrist the number of motor axons was 26.610 ± 3.133 × 10³, with the percent of motor of them being motor 4.66 ± 1.08 (Table 15).

Table 15.

Ulnar Nerve Axon Counts

Article	Location	Axon Counts
Schreiber et al1	Single Point	1.318 × 103
Gesslbauer et al12	Origin	40.379 ± 5.638 × 103
	Midforearm	30.915 ± 2.869 × 103
	Wrist	26.610 ± 3.133 × 103

Anterior interosseous nerve

Multiple studies have described the topography of the AIN at the proximal forearm extending distally to the pronator quadratus and hand. Winston et al²⁶ demonstrated a consistent topographic location of AIN in the dorsocaudal portion of the median nerve with an average axon count of 2.6612 × 10³ at the proximal forearm in six specimens. Similarly, Sananpanich et al²⁷ showed an axon count of 2.903± 1.049 × 10³ in the proximal forearm with a brachialis to AIN transfer. Ziaziaris et al²⁸ showed axon count of 1.760 ± 0.638 × 10³ at the level of the midforearm. More distal at the pronator quadratus, multiple separate studies demonstrated an axon count of 0.506 × 10³, 0.606 ± 0.254 × 10³ axons, 0.670 ± 0.183 × 10³ axons, 0.600 ± 0.250 × 10³ axons, and 0.912 ± 0.088 × 10³ axons.29, 30, 31, 32, 33 In contrast, Namazi et al³⁴ found AIN fascicle numbers of 45.542 × 10³ (41.787–50.467 ± 18351 × 10³) proximally and 25.288 × 10³ (20.949–30.012 × 10³) distally (Table 16).

Table 16.

Anterior Interosseous Nerve Axon Counts

Article	Location	Axon Counts
Winston et al26	Proximal forearm	2.6612 × 103
Sananpanich et al27		2.903 ± 1.049 × 103
Namazi et al34		45.542 × 103
Ziaziariset et al28	Midforearm	1.760 ± 0.638 × 103
Sukegawa et al29	Distal pronator quadratus	0.506 × 103
Frank et al30		0.606 ± 0.254 × 103
Bertelli et al31		0.670 ± 0.183 × 103
Schenck et al32		0.600 ± 0.250 × 103
Wood et al33		0.912 ± 0.088 × 103
Namazi et al34		25.288 × 103

Posterior interosseous nerve

Understanding of the PIN remains critical for nerve transfers at the distal upper extremity. The PIN nerve branches have been described in cadaveric dissection as single primary nerve branches prior to reaching the motor end plate. In their hallmark article, Cheah et al³⁵ described the number of axons as each motor branch reaches the target muscle. The largest number of axons was found at the supinator muscle 0.602± 0.292 × 10³, and the smallest number of axons was found at the extensor indicis proprius (EIP) 0.267 ± 0.173 × 10³.³⁵ More recent histomorphologic data from 2022 found the average number of fascicles of the PIN to be 11, with an average of 5.872 × 10³ axons prior to its branching.³⁶ Two papers have described the number of axons of the PIN as they reach the extensor digitorum communis (EDC) with similar results: 0.560 ± 0.385 × 10³ versus EDC was 0.401 ± 0.190 × 10³ (Table 17).³⁷

Table 17.

Posterior Interosseous Nerve Axon Counts

Article	Location	Axon Counts
Cheah et al35	Supinator	0.602 ± 0.292 × 103
	EDC	0.481 × 103
	EDM	0.329 × 103
	ECU	0.543 × 103
	APL	0.591 × 103
	EPB	0.305 × 103
	EPL	0.436 × 103
	EIP	0.267 × 103
Martins et al36	Origin	5.872 × 103
Melamed et al37	EDC	0.401 ± 0.190 × 103

Intercostal nerves

Although not within the upper extremity, the intercostal nerves are a common choice for musculocutaneous neurotization to restore elbow flexion. The axon counts of the intercostal nerves described in current literature vary widely in comparison to nerves of the upper extremity. The most detailed description of the intercostal nerve axon counts by Malungpaishrope et al³⁸ evaluated 10 fresh cadavers to determine optimal intercostal to AN transfer to restore deltoid function. The axon counts of the third, fourth, and fifth intercostal nerves from the costochondral junction to the midaxillary line were recorded. The average numbers of myelinated nerve fibers of the third, fourth, and fifth intercostal nerves were 0.742 × 10³, 0.830 × 10³, and 1.353 × 10³, respectively. In another analysis of upper-extremity nerves from 15 autopsies and surgical results from 13 operated patients with total or partial spinal root avulsions by Samardzic et al,¹⁸ the average axon counts of the second through fifth intercostal nerves were found to be 0.520 to 0.720 × 10³. Schreiber et al¹ deduced that a single intercostal nerve was composed of 30% motor axons, making the mean axon count of a single intercostal nerve 0.375 × 10³ (Table 18). Variabilities in axon counts were again demonstrated in existing literature.

Table 18.

Intercostal Nerve Axon Counts

Article	Location	Axon Counts
Malungpaishrope et al38	Third intercostal	0.742 × 103
	Fourth intercostal	0.830 × 103
	Fifth intercostal	1.353 × 103
Samardzic et al18	Second to fifth intercostal	0.520 to 0.720 × 103
Schreiber et al1	Single intercostal	0.375 × 103

Discussion

In this review of 38 published studies drawing from several major databases, including PubMed, ScienceDirect, Google Scholar, and CENTRAL, we aimed to investigate the average axon counts of human upper-extremity nerves to provide a concise database for surgeons performing reconstructive neurotization. A review of nerve topography was included in this study to aid surgeons in ensuring appropriate coaptation between donor and recipient nerves, especially when specific fascicular targets are involved.

Overall, there is wide variability between the scale of existing studies, ranging from comprehensive catalogs of upper-extremity nerves to smaller case reports that sample a limited area of one nerve. Based on these differing study methods and sites sampled, there exists a considerable amount of variability in reported average axon counts of the upper-extremity nerves. We posit that the heterogeneity of outcomes between studies necessitates a need for further research in which the upper-extremity nerves are sampled in a more controlled manner at specific points along their course, among a larger population of similar fresh cadaver specimens. Schreiber et al¹ had arguably the most comprehensive description of axon counts and provided both a review of existing literature’s mean axon counts while also performing dissections on 10 cadavers and reporting their own mean axon count results. This was, however, the strongest study in existence on the topic. Only Gesselbauer and Schreiber documented the percentage composition of motor axons in their specimens; with such wide variation in overall axon counts between studies, further investigation into the precise motor composition of mixed motor-sensory nerves is needed and would enhance our understanding of upper-extremity anatomy. No other systematic reviews, randomized control trials, cohort studies, or case control studies existed on the topic, and no meta-analysis of axon counts of the upper-extremity nerves has been performed. Most of the existing literature on the topic was relatively weak and included case reports or case series.

One major aim of this review was to obtain a single summary estimate of each upper-extremity nerve’s average axon count. In doing so, we might have been able to improve the precision of these axon count averages and draw larger conclusions about the optimization of clinical outcomes of nerve transfers. However, given the variability of reported outcomes and the weakness of existing study designs, we can only provide a generalized overview of the upper-extremity nerve topography and list reported axon counts in existing literature. Averaging these reported values to provide the reader with a single average of each nerve’s axon count would be misleading.

Although existing data were weak because of limitations in the study population, methods, and design, there were some strengths in our review of the data. For example, although we were unable to provide more succinct and accurate average axon counts for each upper-extremity nerve, we were still able to provide an axon count range. In some cases, like the median nerve, we were able to report trends of axon counts as the nerve traversed from its proximal to distal points. We believed that for surgical planning purposes, there is merit in knowing an average axon count range and how this number changes along the course of the nerve.

Further studies are needed to sample different points along the course of more upper-extremity nerves to assess trends in motor/sensory axon ratios. Thus, although most existing literature sampled a small study population and yielded differing results, most of these studies were similar in design in terms of the number of cadavers involved, methods of dissection and histomorphometry, and reported outcomes.

Despite weaknesses in existing data, this review serves as a comprehensive database of upper-extremity nerve topography and average axon counts to facilitate surgical planning and improve clinical outcomes of nerve transfers. Furthermore, it demonstrates the variation in reported axon counts at the same level and highlights the need for further, larger studies to standardize counts and account for axonal loss from origin to distal muscle innervation.

Conflicts of Interest

No benefits in any form have been received or will be received related directly to this article.

Supplementary Data

Supplemental Figure 1.

Diagram demonstrating the article selection process. A total of 43 studies were identified, and 38 met inclusion criteria. Statistical analysis was performed to determine axon count averages for all upper-extremity motor nerves identified in the included studies.