12.1 Median Neuropathy at the Wrist (Carpal Tunnel)
Key Takeaways
- Focal median neuropathy at the wrist is diagnosed by distal median motor and sensory slowing relative to ulnar and radial internal controls, not by a prolonged median latency in isolation.
- Comparative studies — median versus ulnar digit 4, median versus radial thumb, and 8 cm palmar mixed median versus ulnar — detect mild carpal tunnel when routine latencies are borderline.
- Typical teaching palmar median-ulnar latency differences on the order of 0.3–0.4 ms are lab comparison tools, not unpublished official AAET cutoffs.
- Low median SNAP or thenar CMAP plus thenar denervation with a normal pronator teres indicates axonal loss at the wrist; Martin-Gruber anastomosis can enlarge or positivize the elbow thenar CMAP.
- Coexisting polyneuropathy, a cold thumb, and stimulating distal to the flexor retinaculum are the common false-positive and false-negative traps.
Carpal tunnel syndrome is the most common upper-limb entrapment you will interpret on nerve-conduction studies. Electrode sites, distances, and machine setup were the job of the previous chapter. This section is case interpretation: deciding whether the median nerve is slowed across the wrist, whether the process is mainly demyelinating or has axonal loss, how comparative studies catch mild disease, and which technical and anatomic traps mimic or hide the diagnosis. The topic is the one listed as VI.B.5.a (Median at the wrist / Carpal Tunnel) on the AAET R.NCS.T content outline. Independent OpenExamPrep material here teaches localization logic; it does not publish secret AAET reference values.
Why the wrist is a focal median lesion
The median nerve enters the hand under the transverse carpal ligament. Digital sensory fibers to the thumb, index, middle finger, and the radial half of the ring finger, plus motor fibers to abductor pollicis brevis (APB), opponens pollicis, and the superficial head of flexor pollicis brevis, travel through that canal. The palmar cutaneous branch usually leaves the median nerve several centimeters proximal to the wrist crease and does not enter the carpal tunnel, so thenar-eminence skin sensation may be spared even when digits 1–3 tingle. That clinical split is a reminder that a “median problem” is not automatically a wrist problem — and that NCS must prove focal slowing at the wrist, not merely a slow median number.
Distal median slowing relative to ulnar and radial
A prolonged median motor distal latency to APB or a prolonged median sensory peak latency to digit 2 or 3 is suspicious, but it is not a localization by itself. Cold hands, long digits, a length-dependent polyneuropathy, and a more proximal median lesion can all delay those same waveforms. The interpretive move is to compare the median distal segment with an internal control that does not cross the carpal tunnel: typically ulnar motor and sensory fibers at the wrist, and superficial radial sensory fibers to the thumb.
In a straightforward demyelinating carpal tunnel:
- Median SNAP peak latency is delayed compared with the ulnar SNAP recorded from digit 5 (or compared with radial SNAP to the thumb).
- Median motor distal latency to APB is delayed compared with ulnar motor distal latency to abductor digiti minimi (ADM).
- Median forearm conduction velocity is usually preserved until the lesion is severe enough to drop the fastest axons.
- Ulnar across-wrist and radial sensory studies remain the patient’s own “normal.”
If both median and ulnar distal latencies are long, do not stamp carpal tunnel on the median number alone. That pattern is how a polyneuropathy presents. You can still support superimposed focal median-at-wrist disease, but only if comparative median-versus-ulnar or median-versus-radial differences remain after temperature and distance are controlled.
Comparative studies that carry the mild case
When routine median SNAP and motor distal latency are borderline, comparative techniques increase sensitivity by putting median and non-median fibers on the same finger or over the same 8 cm palmar distance.
| Comparison | What you stimulate and record | Teaching point |
|---|---|---|
| Median vs ulnar digit 4 | SNAPs from the ring finger; median wrist versus ulnar wrist | Same digit, two nerves. A longer median peak latency than ulnar supports focal median slowing rather than a long finger or a uniformly cool hand. |
| Median vs radial thumb | SNAPs from the thumb; median wrist versus radial forearm or wrist | Radial fibers do not traverse the carpal tunnel. |
| Palmar mixed median vs ulnar (8 cm) | Mixed-nerve action potentials: palm stimulation, wrist recording, 8 cm for both median and ulnar | Short segment across the retinaculum. Often the first study to show a latency gap when digit SNAPs are still almost normal. |
Typical lab teaching — not unpublished official AAET cutoffs — often treats a palmar mixed median-minus-ulnar peak-latency difference on the order of about 0.3–0.4 ms, and digit-4 or thumb peak-latency differences on the order of about 0.4–0.5 ms, as supportive when distance, cathode placement, and temperature are matched. Individual labs publish their own reference values. On the exam, the concept is relative median delay over a short wrist segment, not a memorized secret millisecond.
A second-lumbrical versus interosseous (2L-INT) motor comparison is another teaching tool when the thenar CMAP is tiny or absent: both motor points sit in the midpalm at similar distance, so a delayed lumbrical (median) versus interosseous (ulnar) latency still localizes to the wrist.
Demyelination across the wrist versus axonal loss
Demyelinating median neuropathy at the wrist shows focal latency prolongation and sometimes temporal dispersion of the SNAP. Distal CMAP and SNAP amplitudes can remain relatively preserved if axons are intact. Clinically the patient may have nocturnal paresthesias with little atrophy. EMG of APB may be normal or show only mild chronic reinnervation.
Axonal loss shows low SNAP and/or low thenar CMAP amplitude, and needle EMG of APB (and often opponens) may show fibrillation potentials and large, fast-firing motor unit potentials. Pronator teres, flexor carpi radialis, flexor pollicis longus, and pronator quadratus should remain normal in isolated carpal tunnel — those muscles are innervated proximal to the wrist. If they are denervated, the lesion is not confined to the tunnel.
Severity is a spectrum, not a single number:
| Teaching grade | Typical NCS/EMG picture |
|---|---|
| Very mild | Routine latencies near normal; only comparative studies (palmar mixed, digit 4, thumb) exceed the lab’s difference |
| Mild | Prolonged median SNAP latency; motor distal latency still normal; amplitudes preserved |
| Moderate | Sensory and motor distal latencies prolonged; amplitudes still reasonable |
| Severe | Low or absent median SNAP; thenar CMAP dropping; active or chronic denervation in APB |
| Extreme | Unobtainable median motor and sensory responses; thenar atrophy |
Always say what you compared. “Severe axonal median neuropathy at the wrist” is an interpretation. “Median SNAP 8 µV” without ulnar or radial context is only a measurement.
Case: demyelinating wrist lesion
A 52-year-old dental hygienist reports nocturnal tingling in the thumb, index, and middle fingers. Digit-3 median SNAP peak latency is 4.1 ms with a preserved amplitude; ulnar digit-5 SNAP peak latency is 3.0 ms. Palmar mixed studies at 8 cm show median 2.4 ms and ulnar 1.8 ms. Median motor distal latency to APB is 4.6 ms; APB CMAP is 7.2 mV. Ulnar motor study to ADM is normal. Needle EMG of APB shows a few large motor units without fibrillations; pronator teres is silent and normal.
Interpretation: focal demyelinating median neuropathy at the wrist, moderate by latency, limited axonal loss. The palmar 0.6 ms median-ulnar gap is a teaching-lab comparison, not an official AAET cutoff, and it is what makes the localization wrist rather than “median nerve somehow slow.”
Case: axonal loss still at the wrist
A 67-year-old has thenar atrophy. Median SNAP is absent, APB CMAP is 1.1 mV with a very long distal latency, and APB has fibrillation potentials. Pronator teres and flexor carpi radialis are normal. Ulnar SNAP and dorsal ulnar cutaneous SNAP are normal. This is still severe median neuropathy at the wrist with axonal loss, not a plexus lesion — the forearm median muscles and non-median sensory studies keep the lesion distal. If instead pronator teres and flexor carpi radialis were fibrillating and the palmar mixed comparison were unimpressive, you would stop calling it carpal tunnel and think median nerve in the forearm, lateral cord, or C6–C7 root.
Case: polyneuropathy plus extra median slowing
A 71-year-old with diabetes has bilateral stocking numbness. Median digit-2 SNAP peak latency is 4.4 ms (amplitude 6 µV); ulnar digit-5 SNAP is 4.0 ms (7 µV); radial thumb SNAP is 3.3 ms. After warming to 33°C, palmar mixed studies at 8 cm show median 2.7 ms and ulnar 2.1 ms (difference 0.6 ms). Median motor distal latency is 5.1 ms; ulnar motor distal latency is 3.4 ms. Teaching interpretation: a length-dependent polyneuropathy plus extra focal median slowing across the wrist. The shared distal delay is the neuropathy; the palmar difference is the wrist-localizing piece. Absolute median cutoffs without an ulnar or radial partner would have been non-specific.
Martin-Gruber anastomosis: preview, not the full chapter
A Martin-Gruber anastomosis (MGA) is a median-to-ulnar communication in the forearm. Those crossing fibers do not travel through the carpal tunnel. When you stimulate the median nerve at the elbow and record over thenar muscles, you may activate MGA axons that reach ulnar-innervated thenar or deep-head flexor pollicis brevis fibers, producing a thenar CMAP that is larger, faster, or initially positive compared with the wrist median CMAP. That pattern is not proof that the wrist is electrically healthy, and it is the opposite of conduction block (block would drop the proximal CMAP). Do not invent a “median forearm conduction block” from a wrist-smaller-than-elbow thenar waveform without considering MGA. Full anastomosis maps belong in a later chapter; here, remember MGA can contaminate thenar CMAP comparisons in a carpal-tunnel study.
Exam traps
Coexisting polyneuropathy. Diabetes and other length-dependent neuropathies delay median and ulnar (and often radial) distal responses. Absolute median distal latency loses specificity. Lean on short-segment comparative differences. Superimposed carpal tunnel is common; it is diagnosed by extra median slowing across the wrist, not by a slow median SNAP in isolation.
Low thumb temperature. Cool digits prolong SNAP and motor latencies and can manufacture a false median delay against an absolute cutoff. Warm the hand to the lab’s temperature standard (many labs use about 32°C skin temperature as a teaching target) and repeat the comparisons. A latency that was “abnormal” only while the thumb was cold was not demyelination.
Stimulating too distally. Median motor distal latency is a latency over a stated distance (commonly 8 cm to APB). If the cathode slides into the palm, two errors appear: the true distance is shorter than the documented 8 cm (latency looks too short — false-negative tunnel), and you may depolarize the recurrent thenar motor branch distal to the flexor retinaculum, bypassing the entrapment entirely. Recheck tape-measure distance, cathode-over-nerve, and that wrist stimulation is proximal to the tunnel. A reported “normal” 3.3 ms distal latency with the cathode sitting near the thenar crease is a technical miss until it is repeated at a true 8 cm.
Independent OpenExamPrep takeaway: name the segment (wrist), the pathophysiology (demyelinating slowing versus axonal loss), and the controls (ulnar, radial, temperature, distance). Electrode recipes without that interpretation do not localize the lesion.
A 48-year-old has suspected median neuropathy at the wrist. Digit-2 median SNAP peak latency is 3.6 ms and ulnar digit-5 SNAP peak latency is 3.4 ms. Which additional study is specifically designed to compare median and ulnar fibers across an 8 cm palmar mixed segment?
Which constellation most strongly indicates axonal loss at the wrist rather than isolated demyelinating median slowing across the carpal tunnel?
In a patient being studied for carpal tunnel, the thenar CMAP is larger and begins with a positive dip when the median nerve is stimulated at the elbow than when it is stimulated at the wrist. The best teaching interpretation is: