Introduction and historical aspects

Amyotrophic lateral sclerosis (ALS) is a neurodegenerative disorder in which there is preferential loss of both the upper and lower motor neurons.1 At clinical onset, manifestations appear rather focal, resulting in bulbar, cervical or lumbar phenotypes. It has been suggested that neuronal loss spreads contiguously from the site of origin, becoming diffuse and resulting in complex motor deficits.2 Among the intrinsic hand muscles, wasting predominantly affects the ‘lateral (thenar) hand’ involving both median innervated muscles (abductor pollicis brevis (APB) and opponens pollicis) and ulnar innervated muscles (first dorsal interosseous (FDI), adductor pollicis and flexor pollicis brevis), with relative sparing of the hypothenar muscles (the abductor digiti minimi (ADM)) (figure 1). This peculiar pattern of dissociated atrophy of the intrinsic hand muscles was termed ‘split hand’ by Dr Asa Wilbourn.3 4

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

(A) The ‘split hand’ syndrome of amyotrophic lateral sclerosis. There is clear wasting of the first dorsal interosseous (FDI) and thenar complex but sparing of the hypothenar muscle (black arrows). FDI, thenar and hypothenar muscles are innervated by C8–T1, and FDI and hypothenar muscles by the ulnar nerve. The dissociated hand muscle involvement cannot be anatomically explained. (B) Hand muscle atrophy in C8 radiculopathy. Note the marked atrophy of the FDI as well as the hypothenar muscle (white arrows).

During the 1990s, Wilbourn routinely compared the amplitudes of compound muscle action potentials (CMAPs) and extent of denervation in the thenar muscles, FDI and ADM in patients with ALS. It was noted that the thenar muscles are frequently and substantially more denervated than the hypothenar muscles in ALS. Sometimes the FDI was even more severely denervated than the other muscles, and could be involved before the thenar muscles. As the muscles on the lateral aspect of the hand were preferentially affected, compared with those on the medial aspect of the hand, the disassociation was labelled ‘split hand’.3–5

At the Third International Symposium on ALS/motor neuron disease (MND), Wilbourn presented in abstract form ‘dissociation of muscle wasting of ulnar innervated hand muscles with motor neuron disease’,6 with a subsequent abstract ‘dissociated wasting of medial and lateral hand muscles with motor neuron disease’.5 The same year Eisen introduced the term ‘thenar hand’ in ALS, emphasising the preferential wasting of the lateral hand in ALS.7 But the term ‘split hand’ was first coined in 1996.3 These early reports prompted one of the authors (SK) to prospectively examine motor neuron loss in the APB and ADM in ALS using motor unit number estimates. Motor unit loss was significantly greater in APB than ADM, and a simple comparison of CMAP amplitude showed significantly reduced APB/ADM ratios in ALS patients compared with normal subjects and neurological controls.8 Wilbourn responded to this publication with a letter to the editor entitled ‘The split hand syndrome’.4

Over the past decade, the split hand has become increasingly recognised as a useful clinical sign of ALS.9 The thenar complex muscles (APB and FDI) constituting the split hand are innervated through the same spinal segments (C8 and T1) but the FDI and ADM, which are differentially affected, share ulnar innervation. Therefore, the mechanisms underlying the split hand are complex and incompletely understood. However, recent evidence suggests that both cortical and spinal/peripheral mechanisms must be involved in the split hand of ALS.10–13

Sensitivity, specificity and diagnostic utility

Wilbourn stressed that not every ALS patient with hand wasting showed the split hand phenomenon and sometimes muscle wasting could be more or less equal between the lateral and medial hand muscles, especially when the disease was advanced. Very rarely were the hypothenar muscles more severely affected than the muscles of the lateral hand (reversed split hand).4 Furthermore, the split hand was not completely specific to ALS, and Wilbourn encountered it in several other conditions, including Werdnig–Hoffmann disease, ‘benign’ focal motor neuron disease and remote poliomyelitis. However, “the demonstration of a split hand during nerve conduction studies on a patient referred to the EMG laboratory with suspected ALS was considered to be a very ominous finding and subsequent needle electrode examination almost never demonstrated any alternative diagnosis”.4

The split hand has been rarely reported in patients with non-ALS related ‘pure lower motor neuron disease’—namely, autosomal dominant distal spinal muscular atrophy, spinocerebellar ataxia type 3 (Machado–Joseph disease) and juvenile muscular atrophy,14 as well as spinal and bulbar muscular atrophy (SBMA).9 There are only a few causes of painless hand wasting in the absence of sensory deficit, as is typical of ALS. These include thenar wasting from a cervical rib and band,15 spondylotic cord compression,15–18 compression of the deep motor branch of the ulnar nerve19 20 and carpal tunnel syndrome in the elderly, in whom there may be moderate or marked thenar wasting with little if any sensory deficit,21 and a C8 root lesion. All of these conditions have characteristic clinical and electrophysiological features, and the split hand phenomenon has not been reported in any of these conditions (see figure 1).

A multicentre Japanese study systematically examined the sensitivity and specificity of the split hand in ALS (The Tokyo Metropolitan Neuromuscular Electrodiagnosis Study group).9 There were 77 patients with ALS, 171 normal subjects and 199 patients with other disorders affecting lower motor neurons/axons, such as spinal muscular atrophy, SBMA and axonal polyneuropathies. The split hand was strictly defined as APB/ADM CMAP amplitude ratio <0.6 or FDI/ADM ratio <0.9 (mean−3SD of the normal control data). Based on the electrophysiological criteria, split hand was found in 52% of ALS patients, in 5% of normal subjects and in 13% of disease controls. Among the disease controls, two patients with SBMA showed split hand; in ALS patients, the split is found with moderate sensitivity (52%) and high specificity (95% vs normal controls and 87% vs disease controls). A recent study has described the use of a novel neurophysiological measure of the split hand, ‘the split hand index’.22 This is the ratio of the CMAP amplitude of the APB times the CMAP amplitude of the FDI divided by the CMAP amplitude of the ADM. This study needs confirmation but appears to be a robust measure of split hand in ALS.

Spinal and peripheral aspects of split hand syndrome

Wilbourn thought that the lesion responsible for the ALS split hand was at the level of the cervical anterior horn cell, a view shared by others.4 9 14 In ALS, there is a greater reduction of the estimated number of motor units in the APB compared with the ADM.8 Also, the CMAP ratio of the APB/ADM is reversed in ALS—that is, <1 instead of >1, as is normal.8 Both of these observations indicate preferential involvement of the APB (lateral hand) compared with the ADM (medial hand), typical of split hand.

It has been hypothesised that, in humans, use of the APB muscle is greater than that of the ADM, which could render the former muscles (involved in split hand syndrome) subject to greater oxidative stress.8 23 Normal ageing, associated with cumulative oxidative stress, is also associated with a tendency of greater reduction of both the FDI and APB CMAPs compared with the ADM.23

Several nerve excitability studies in ALS have shown two types of abnormal peripheral motor axonal ionic conductances: increased persistent sodium currents and reduced potassium currents, both of which increase axonal excitability, presumably leading to generation of fasciculations of peripheral origin.11 12 Nodal persistent sodium currents are major determinants of axonal excitability. It has been suggested that the split hand may be explained by differences in membrane properties among motor axons innervating the APB, FDI and ADM. A recent study has investigated differences in axonal excitability indices in normal subjects.13 Nodal persistent sodium currents estimated by strength–duration time constant and latent addition were prominently greater for APB and FDI motor axons compared with ADM axons (figure 2). These findings suggest that motor axons with physiologically greater persistent sodium currents—namely ALS-like—may be more vulnerable and more readily die in ALS, and this may contribute to the development of split hand.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Peripheral aspects of the split hand syndrome. Nodal persistent sodium currents estimated by the strength–duration time constant (SDTC). (A) Abductor pollicis brevis (APB), first dorsal interosseous (FDI) and abductor digiti minimi (ADM) motor axons of normal subjects. APB/FDI axons have a longer SDTC than ADM axons, suggesting more prominent persistent sodium currents (modified from Bae and colleagues).13 (B) APB motor axons of amyotrophic lateral sclerosis patients and normal controls. Amyotrophic lateral sclerosis axons have a longer SDTC than normal axons, suggesting more prominent persistent sodium currents (modified from Kanai and colleagues).11 CMAP, compound muscle action potential.

Cortical aspects of split hand syndrome

When analysing the physiological basis of the split hand syndrome, simple comparison of the APB with the ADM CMAPs have major limitations as it is the thenar complex (both the median and ulnar innervated muscles) that are preferentially involved, not just the APB.

The thenar complex together with the FDI is responsible for precision or ‘pincer grip’, which is the main clinical deficit of the split hand syndrome. Precision hand movements depend on the ability to individuate (fractionate) movements of particular fingers from a more fundamental multidigit grasping plan, a capacity provided by motor synergies between intrinsic and extrinsic hand muscles that appears relatively late, both in evolution and in individual development.24 Lawrence and Kuypers25 showed that after the bilateral pyramidotomy at the brainstem level, monkeys could reach for an object, but the ability of precision grip was permanently impaired, suggesting the importance of the corticospinal tract in the control of dexterous finger movements. This would support the concept that the split hand deficit, and maybe ALS, are primarily cortical in origin.26 27

The evolutionary origins and variations of the precision grip, in which an object is held between the thumb and other digits, are poorly understood.28–31 This is surprising because the neural basis of this grasp pattern, including the motor cortex and pyramidal tract input, have received extensive study.32 33 Precision grip has been demonstrated in infants as young as 3–5 months old.34 However, grip force is exaggerated.34 In the normal elderly, precision grip tends to be more forceful and requires a greater degree of visual feedback.32 35 36

The ratio of the amplitude to the motor evoked potential (MEP) evoked by transcranial magnetic stimulation to the CMAP evoked by peripheral nerve stimulation is a measure of the motor cortical versus the spinal motor neuron pool input to a muscle. Because the MEP/CMAP ratio takes account of the lower motor neuron contribution, or disease, it is a useful indicator of abnormality originating in the motor cortex. In normal subjects, when TMS stimulation is maximal, the cortical/peripheral (MEP/CMAP) ratio is significantly larger for the thenar complex but not for the hypothenar complex.37–39 This is because the thenar complex is composed of both median nerve and ulnar nerve components. Median nerve peripheral stimulation at the wrist only activates the ‘median’ thenar complex but cortical stimulation activates both the ‘median and ulnar’ thenar complex.

Recording specifically from the thenar complex (ABP) and hypothenar complex (ADM), there is a significant reduction in the thenar MEP/CMAP ratio in ALS but this does not occur in the hypothenar complex (figure 3). This indicates that the corticomotoneuronal input to the thenar complex is preferentially affected in ALS and that corticomotoneuronal disease/dysfunction may be the prime determinant for the split hand in ALS.10 Similar studies have not compared the MEP/CMAP ration in ALS recording from the FDI compared with the ADM. Such a study would not show a significant difference because only one nerve (the ulnar) is involved. However, corticomotoneuronal input, determined by the size of the MEP in response to transcranial magnetic stimulation, has been measured in the FDI in response to various hand grip movements, and was greatest when a pincer grip was employed.40 It has recently been shown that there is concordance for the side of onset of ALS and handedness.41 This is not so for lower limb onset. This may in part reflect activity dependence for the upper limb in which dominance is prominent. However, neuronal connectivity and cortical excitability in relation to handedness and reflected by the ‘split hand‘ phenomenon consistently observed in ALS, may be more important.41

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Cortical aspects of the split hand syndrome. (A) Cortical representation of the thenar hand (abductor pollicis brevis and first dorsal interosseous) used in precision grip is large compared with the hypothenar complex, and has a larger corticomotneuronal input. The motor evoked potential (MEP) is larger than the compound muscle action potential (CMAP) because both median and ulnar components are activated by transcranial magnetic stimulation. The hypothenar MEP is often smaller than the CMAP. (B) There is reversal of the normal situation. Cortical representation ‘shrinks’ and the thenar MEP is smaller than the thenar CMAP. The MEP/CMAP ratio of the hypothenar remains normal until more advanced disease. FDI, first dorsal interosseous.

Grasping and manipulating objects is extremely important for everyday functional activities. Precision grip, the act of grasping objects between the thumb and index finger, is a complex sequential action that requires precise application of forces large enough to prevent objects from slipping but not too large to break fragile objects or cause fatigue.42 Impaired precision, especially pincer grip, is a likely surrogate for split hand. Formal study or measurement of finger fractionation has not been studied in ALS but is impaired in several upper motor neuron disorders. After a capsular stroke, especially if posterior, there is impairment of precision grip, even 2 or more years after the insult, despite otherwise good hand function.43 There is a condensed organisation of corticofugal projections and the density of pyramidal fibres from the primary motor cortex in this subsector of the internal capsule. In writer's cramp the cortical silent period measured during a pincer grip is significantly shorter in patients than in controls, indicating a task specific impairment of motor cortical excitation and inhibition.44 Impaired precision grip has been documented in Huntington's disease and also in pre-manifest Huntington's disease.45 46