Peripheral Neuropathy Evaluation

Length-dependent neuropathy: the dying-back geometry

The most common pattern, the axonal length-dependent polyneuropathy, is a consequence of cell biology rather than anatomy. A lumbar motor neuron must maintain an axon a metre long on a fixed transcriptional and mitochondrial budget; when axonal transport, energy supply, or metabolic homeostasis fails, the distal-most segment — the territory most expensive to supply — degenerates first. This dying-back process produces the cardinal clinical gradient: distal > proximal, and feet before hands. Symptoms reach the upper limb only once the lower-extremity deficit has ascended to roughly the knee, because the fingertips and the mid-calf lie at comparable distances from their respective somata. A stocking deficit that has just begun to glove tells you, without imaging, that the longest axons are failing synchronously by length.

The protocol operationalises this geometry. Because the abnormality is maximal distally and earliest in the longest nerves, the highest-yield comparison is distal lower limb against upper limb: the sural sensory response (or the medial plantar in subtle cases) against the median and ulnar sensory responses. An absent sural with a preserved median SNAP is the electrophysiological fingerprint of length-dependence — the long nerve has dropped out while the short one survives. Inverting that gradient (upper-limb SNAPs lost while the sural is spared) is a red flag that the process is not length-dependent and should redirect you toward an acquired demyelinating, ganglionopathic, or multifocal aetiology.

Multifocal neuropathy and mononeuritis multiplex

When individual named nerves are picked off in a stepwise, asymmetric, painful sequence, the lesion is not metabolic but ischaemic: occlusion of the vasa nervorum, the hallmark of vasculitic neuropathy. The pathology is axonal, but its distribution is the diagnosis. Because infarcts strike random watershed segments, the electrodiagnostic signature is asymmetry between homologous nerves and between nerves in the same limb — an absent superficial peroneal sensory response on one side with a normal one contralaterally, or a radial that is gone while the median is intact. A routine bilateral screen will under-call this; the protocol must be deliberately asymmetric, studying multiple individual nerves in several limbs and explicitly comparing left with right. Side-to-side amplitude ratios become the primary readout. Confirmation by combined nerve-and-muscle biopsy (e.g., superficial peroneal nerve with peroneus brevis) is frequently required, and recognising the pattern is urgent because vasculitis is treatable and progressive.

Small-fibre neuropathy: why the machine reads normal

The single most important conceptual limit of routine electrodiagnosis is that nerve conduction studies and needle EMG interrogate only large, myelinated fibres. The sensory and motor latencies and amplitudes are generated by the fastest, thickly myelinated Aβ and Aα populations. Pain and thermal sensation, and autonomic function, travel in thinly myelinated Aδ and unmyelinated C fibres, which contribute nothing measurable to a surface-recorded response. Consequently a patient with burning feet, allodynia, and autonomic symptoms from a pure small-fibre neuropathy will have a completely normal NCS and EMG— a result that must be read as “the large fibres are intact,” not “no neuropathy.” Reporting normal large-fibre studies as exclusionary here is a classic false-negative trap.

Small-fibre involvement requires dedicated modalities that bypass the large-fibre limitation. Quantitative sensory testing (QST) measures thermal and pain thresholds psychophysically. Skin punch biopsy with measurement of intraepidermal nerve fibre density (IENFD), immunostained for PGP 9.5, is the structural gold standard, demonstrating reduced density of the unmyelinated terminals at the distal calf relative to the thigh — itself a length-dependent readout. QSART (quantitative sudomotor axon reflex testing) assays the postganglionic sympathetic C-fibres that drive sweating. The teaching point is structural: the modality must match the fibre calibre, and a normal nerve conduction study never excludes a small-fibre process.

Diabetic neuropathy: the composite pattern

Diabetes is the most instructive teaching case because it disobeys the single-pattern model. The prototypical distal symmetric polyneuropathy is predominantly a length-dependent axonopathy, but chronic hyperglycaemia and the polyol/AGE pathways also injure myelin, producing a mixed axonal-demyelinating picture with conduction velocities slowed further than pure axon loss would predict — though rarely into the unequivocally acquired-demyelinating range. Crucially, the metabolic milieu renders nerves hypersusceptible to compression, so the baseline polyneuropathy is routinely decorated with superimposed entrapments: a disproportionately prolonged median distal latency at the wrist, or a focal ulnar conduction slowing across the elbow, layered on a diffuse background. Reading the median slowing as merely “part of the neuropathy” misses a surgically treatable carpal tunnel syndrome. And because small fibres are affected early — often before large-fibre studies turn abnormal — many symptomatic diabetics have normal NCS and require small-fibre testing. Diabetes is, in effect, every pattern at once.

Bayesian priors: letting the presentation set the protocol

Expert electrodiagnosis is the disciplined application of pre-test probability. A symmetric stocking-glove distribution in a diabetic or alcoholic patient carries a high prior for distal axonal polyneuropathy: a focused lower-limb-dominant screen confirms it efficiently. Stepwise, painful, asymmetric mononeuropathies shift the prior decisively toward vasculitis and demand an asymmetric multi-nerve protocol with biopsy planning. Burning feet with intact reflexes and a normal large-fibre study raise the prior for small-fibre neuropathy and should trigger IENFD rather than a repeat NCS. A subacute, areflexic, proximal-and-distal weakness with conduction block points to acquired demyelination (CIDP/AIDP), not a dying-back process. The protocol is a hypothesis test: choose the nerves and the modalities that maximise the likelihood ratio for the specific pattern your patient's history has already made probable.