17.3 Martin-Gruber Anastomosis Patterns

Why Martin-Gruber anastomosis is a technical-considerations item

Martin-Gruber anastomosis (MGA) is outline VII.A.4.a. It is a median-to-ulnar motor crossover in the forearm, usually supplying ulnar-innervated hand muscles. Independent OpenExamPrep teaching for the National Registry Examination for Nerve Conduction Studies is the three recording-site patterns the outline names: hypothenar abductor digiti minimi (ADM), first dorsal interosseous (FDI), and thenar abductor pollicis brevis (APB). This is not an AAET atlas of every forearm communication. It does not cover Riche-Cannieu anastomosis or accessory fibular (peroneal) innervation of extensor digitorum brevis; those belong in the next anomaly chapter.

MGA is common in electrodiagnostic teaching series (often discussed around 15–30% of limbs, with a wide published range). Combinations of muscle targets are frequent. Those incidence figures are typical teaching, not unpublished official AAET cut scores. The exam cares whether you can recognize the pattern and not call it ulnar conduction block (CB).

Chapter 7 already listed anomalous innervation as a mimic of block. This section is that mimic in full: where the axons sit at each cathode, what each recording muscle shows, and how the millivolts add up.

Anatomy: axons leave the median trunk in the forearm

In a typical MGA, motor axons travel in the median nerve (often via anterior interosseous contributions in anatomic descriptions) through the elbow, then cross to the ulnar nerve in the forearm, then continue to hand muscles that laboratories think of as ulnar: ADM, FDI, and ulnar-innervated thenar muscles (adductor pollicis, deep head of flexor pollicis brevis) that sit under or beside an APB electrode.

Critical geography:

• At the wrist, those crossing axons are already in the ulnar nerve. Median wrist stimulation does not include them.
• At the median elbow, those axons are still in the median nerve. Median elbow stimulation does include them.
• At the ulnar below-elbow site (proximal to the crossover), those axons have not yet joined the ulnar nerve. Ulnar below-elbow stimulation misses them.
• At the ulnar wrist, the ulnar nerve contains both native ulnar axons and the MGA axons that joined in the forearm. Ulnar wrist stimulation includes them.

That geography produces the two screenshots every candidate must recognize:

1. Median elbow CMAP larger than median wrist CMAP when you record from a muscle that receives MGA fibers.
2. Ulnar wrist CMAP larger than ulnar below-elbow CMAP at the same muscle — a fake forearm conduction block, because axons appear to leave the ulnar trunk. They were never in the proximal ulnar nerve; they arrive from the median.

True ulnar neuropathy at the elbow drops amplitude across the elbow (above-elbow versus below-elbow). MGA drops amplitude in the forearm (below-elbow versus wrist) on ulnar stimulation. If you only compare wrist with above-elbow and skip the below-elbow site, MGA can be misread as an across-elbow drop. Put the below-elbow cathode on and the geography becomes obvious.

Type I / II / III by recording site

Papers number MGA subtypes slightly differently. Independent teaching uses the three outline targets. Frequencies below are typical electrodiagnostic teaching, not unpublished official AAET percentages.

A single forearm can carry more than one type. Finding FDI MGA does not prove ADM is spared. Check the muscle the question stem records from.

Type I — hypothenar ADM

Median nerve does not normally innervate ADM. Median wrist to ADM: tiny volume-conducted bump or essentially nothing. Median elbow to ADM: a definite CMAP if Type I fibers are present. Ulnar wrist to ADM: the full hypothenar CMAP. Ulnar below-elbow to ADM: smaller, missing the MGA contribution. The ulnar forearm block is those median-traveling axons. Initial positivity at ADM from median stimulation is volume from thenar until G1 is truly on ADM; do not confuse a thenar far-field with a huge Type I anastomosis.

Type II — FDI

FDI is ulnar-innervated via the deep motor branch. The same logic applies, and the extra millivolts are often larger than at ADM, which is why FDI is the demonstration muscle in many laboratories. Median elbow to FDI can look surprisingly big. Median wrist to FDI should be small (volume from thenar or lumbricals). If median wrist to FDI is huge, think co-stimulation of ulnar at the wrist or a misplaced G1, not routine MGA. The anastomosis is the elbow-minus-wrist extra CMAP, not the volume bump.

Type III — thenar APB

APB is median-innervated, so this pattern looks different. Median wrist already evokes a real APB CMAP. Median elbow evokes that CMAP plus MGA fibers that will cross to ulnar thenar muscles recorded under the same G1. Proximal median CMAP is larger than distal — the opposite of median forearm conduction block, which would shrink the proximal response. Ulnar wrist stimulation recording APB often shows a CMAP from adductor pollicis and deep FPB. Do not call that ulnar-thenar pickup proof the median is dead, and do not report median forearm block from a proximal CMAP that grew.

Worked numbers

Recording FDI, supramaximal stimuli, warm limb, G1 on FDI:

MGA contribution from the median side: 5.2 − 0.8 = 4.4 mV. Ulnar forearm drop: 14.0 − 9.6 = 4.4 mV. Above-elbow versus below-elbow: 9.6 − 9.4 = 0.2 mV (no meaningful across-elbow drop).

The missing ulnar axons at the below-elbow site are accounted for by median elbow stimulation. Duration of the ulnar below-elbow CMAP stays compact. Ulnar SNAP (digit V) is preserved. That is MGA, not neurapraxic ulnar CB in the forearm.

Type I ADM example. Median wrist to ADM 0.4 mV; median elbow to ADM 2.1 mV; ulnar wrist to ADM 10.5 mV; ulnar below-elbow to ADM 8.4 mV. Check: 2.1 − 0.4 = 1.7 mV; 10.5 − 8.4 = 2.1 mV. A few tenths can come from volume, a slightly submaximal proximal ulnar shock, or extra thenar pickup. The direction still matches: extra median-elbow millivolts, smaller ulnar below-elbow CMAP, full ulnar wrist CMAP.

Type III APB example. Median wrist to APB 7.0 mV; median elbow to APB 9.5 mV. Proximal is 2.5 mV larger, which would be an illegal negative block if you thought it was median forearm CB. Ulnar wrist to APB 2.6 mV (ulnar thenar under the electrode). Report anastomosis physiology, not a machine that added 2.5 mV of median axons between wrist and elbow.

Martin-Gruber versus true ulnar conduction block

True conduction block (chapter 7) is an abnormal proximal amplitude or area drop across a lesion with a relatively preserved distal CMAP, often with clinical weakness, and sometimes with temporal dispersion. MGA mimics a wrist-large / below-elbow-small ulnar CMAP without a lesion.

Protocol when you see a surprising ulnar forearm drop: stimulate the median at the elbow while recording the same muscle (FDI or ADM). If a CMAP appears that was absent or tiny from median wrist, you have demonstrated MGA. Then measure above-elbow ulnar to see whether a second, real drop exists at the elbow — MGA and UNE can coexist. Anastomosis does not grant immunity to cubital tunnel.

Do not use MGA as a blanket excuse to ignore a dispersed, area-collapsed ulnar CMAP in a weak, numb ulnar hand. Anastomosis explains accounted-for, compact millivolt shifts. It does not explain a 90% area collapse with a dead digit-V SNAP.

What this section deliberately does not cover

Riche-Cannieu anastomosis is a palmar communication, typically ulnar-to-median toward thenar muscles. It produces a different thenar map (ulnar wrist can drive median thenar; median wrist may look small). Accessory fibular (peroneal) innervation makes the extensor digitorum brevis CMAP grow with stimulation behind the lateral malleolus. Both are named anomalies. They are not VII.A.4.a and are left for the next chapter. Mixing them into MGA is an outline error.

Traps

• Calling ulnar forearm amplitude drop conduction block without a median-elbow check
• Expecting MGA to drop amplitude across the elbow (that is UNE geography)
• Forgetting Type III: proximal median APB CMAP larger than wrist is MGA, not a machine error and not ordinary median CB
• Treating FDI as an afterthought; it is the most common MGA recording site
• Inventing unpublished official AAET incidence percentages as if they were cut scores
• Dragging Riche-Cannieu or accessory fibular into this item
• Ignoring coexistence: MGA plus true UNE can both be present
• Using a volume-conducted median-wrist-to-FDI bump as if it were the anastomosis itself; the anastomosis is the elbow-minus-wrist extra CMAP