Electrodiagnostic features of hereditary neuropathy with liability to pressure palsies

Subjects.

The records of the Oregon Health Sciences University Neurophysiology Laboratory, Portland, OR, were searched for all cases of HNPP, CIDP, or diabetic polyneuropathy studied between June 1995 and June 1998. All HNPP cases required confirmatory chromosome 17 deletion analysis and exclusion of other causes of neuropathy by the treating neurologist. Three additional genetically confirmed HNPP cases, two from the neurophysiology laboratory at the University of Washington Seattle and one from the Northern Navajo Medical Center, were added to the cohort.

The coauthors used identical nerve conduction study methods. Inclusion criteria for CIDP required clinical evidence of symmetric proximal and distal weakness of arms and legs associated with hyporeflexia or areflexia, compatible electrodiagnostic findings, and exclusion of alternative causes. As a method of data quality control, cases of diabetic polyneuropathy were subdivided into those with exclusive or predominant upper symptoms and those with lower extremity symptoms as reason for the evaluation. These subgroups were compared both with each other and with CIDP and HNPP.

Nerve conduction studies.

Limb temperature was maintained at or above 32.0 °C for the hands and 31.0 °C for the feet. In the few studies wherein limb temperature was recorded below these limits, nerve conduction velocities and distal motor latencies (DMLs) were corrected by 2 m/second or 0.2 msec per degree centigrade, respectively. Motor nerve conduction studies of the median, ulnar, peroneal, and tibial nerves were performed using conventional methods, recording at the abductor pollicis brevis, abductor digiti minimi, extensor digitorum brevis, and abductor hallucis muscles, respectively. The distal stimulation distance was 8 cm.

The following nerve segments were used for calculating motor nerve conduction velocities (MNCV): wrist to elbow for the median nerve, wrist to below elbow for the ulnar nerve, ankle to fibula head for the peroneal nerve, and ankle to popliteal fossa for the tibial nerve. At the distal sites, F-wave stimulations were performed by reversing the polarity of the stimulating electrode. Sensory responses were obtained antidromically using ring electrodes for the median and ulnar nerves on the second and fifth digits, respectively, and using bar electrodes at the ankle for the sural nerve and at the web space of the thumb for the radial nerve. Sensory nerve stimulation was performed 14 cm proximally, and conduction velocity was determined from the onset latency. All results were expressed as the percentage of the lower or upper limit of the laboratory normative values for age.

In a similar way, F-wave latencies were corrected for both age and height. If any responses were unrecordable, they were excluded from the analysis. A terminal latency index (TLI) was calculated for each motor nerve using the following formula:

Statistical analysis.

Statistical analysis was performed by analysis of variance using a commercially available software statistics package (Statview®, SAS Institute, Cary, NC). The English language literature was searched for HNPP studies published with sufficient information to permit comparison with the data obtained in the current study. Normative values of the publishing authors were used, and if the lower limit of normal was not provided, it was assumed to be 2 SDs below normal.

Subjects.

The current study included nine HNPP subjects from eight kinships, ages 17 to 63 years, all symptomatic. There were 22 subjects with CIDP and 49 with diabetic polyneuropathy. Of the diabetic subjects, 23 had presenting symptoms exclusively or predominantly in the upper extremity, and 26 had symptoms in the lower extremity. Because no mean differences were found between these two diabetic subgroups in any of the electrophysiologic parameters (data not shown), results for the diabetic nerves were pooled.

Electrophysiology.

The outstanding characteristic of HNPP was abnormally slow sensory nerve conduction velocity (SNCV). This was found in 27 of 29 (93%) HNPP nerves, with the remainder exactly at the lower limit of normal (table), which was approximately three times more frequently than in CIDP or diabetic polyneuropathy. The frequency of a reduced sensory nerve action potential amplitude was 41% (12/29) in HNPP, indicating that the conduction slowing did not merely reflect the secondary consequences of axonal loss, but rather was “demyelinating” in character.

In a separate attempt to limit the potential confounding contribution by entrapment neuropathy, we excluded median sensory responses from the analysis. The frequency of abnormal slowing remained largely unchanged for HNPP and CIDP, whereas that of diabetic nerves dropped to 10% (SNCV-ES; see table). Therefore, the prominent SNCV slowing observed in HNPP was not caused by entrapment.

To assess better the magnitude of electrophysiologic differences between HNPP, CIDP, and diabetes, scatterplot analyses were performed (figure, A and B). Mean SNCVs for HNPP were 85.6% ± 10.6% of the lower limit of normal and significantly slower than for CIDP (114.3% ± 20.1%; p < 0.0001) or diabetes (108.1% ± 14.8%; p < 0.0001). Excluding the carpal tunnel site from the analysis did not alter this observation: HNPP 87.9% ± 10.0%, CIDP 121.1% ± 19.2% (p < 0.0001), and diabetes 111.5% ± 12.5% (p < 0.0001). In the only large series in the literature amenable to comparative post hoc analysis (consisting of 32 subjects),8 mean SNCV of HNPP was not different from that of our smaller cohort (83.5% ± 18.2% and 84.8% ± 16.2% of the lower limit of normal with and without the median response, respectively) (see figure, A and B).

• Download figure
• Open in new tab
• Download powerpoint

Figure. Nerve conduction box plots of nerve responses in hereditary neuropathy with liability to pressure palsies (HNPP), chronic inflammatory demyelinating polyneuropathy (CIDP), and diabetic polyneuropathy (DM) for (A) sensory nerve conduction velocity (SNCV), (B) SNCV excluding median nerve responses, (C) motor nerve conduction velocity (MNCV), (D) distal motor latency (DML) with (open) and without (shaded) median nerve responses, (E) terminal latency indices, and (F) F-wave latency. Nerve conduction responses are expressed as a percentage of the relevant upper or lower limit of normal with the 10th, 25th, 50th, 75th, and 90th percentile distributions indicated and the responses outside these limits plotted as separate circles. The HNPP mean and SDs for this cohort (indicated by letter a) and those of Pareyson et al.8 (indicated by letter b) and Gouider et al.9 (indicated by letter c) are represented on the right side of the panels.

In contrast to sensory responses, HNPP motor nerve conduction slowing was much less frequent and minor. Motor nerves were abnormally slow in 31% (10/32) of the nerves studied, which was less frequent than in either CIDP or diabetes (see table). Scatterplot analysis demonstrated that motor nerve slowing in HNPP was minimal in degree (figure, C), and mean MNCV remained within normal limits (106.4% ± 12.9%) and similar to that of diabetes, whereas that of CIDP was slowed (86.2% ± 23.2%; p < 0.001).

The HNPP DML was prolonged in 78% (25/32) of the nerves, which was more frequent than in CIDP or diabetes. This difference with diabetes became even more evident when the carpal tunnel entrapment site was removed from analysis (see table). Mean DMLs were more prolonged in HNPP, even without median nerve data (a potential entrapment site) in the analysis (118.5% ± 31.0% of the upper limits of normal), than in CIDP (103.2% ± 31.6%; p < 0.05) or diabetes (86.3% ± 18.3%; p < 0.0001) (figure, D).

Determination of the terminal latency index is a method of evaluating for differential motor slowing between distal and proximal nerve segments. The mean terminal latency indices were lower in HNPP than in the other two conditions for all (p < 0.05 versus p < 0.001) but the tibial nerves (differences not significant and limited by small sample size), which is indicative of distally accented conduction slowing in the condition (figure, E). The frequency of prolongation of F-wave latencies was similar in all three conditions (see table) and not helpful in discrimination, although it was of greater magnitude in CIDP (p < 0.001; figure, F). The results reported here are similar to those obtained by post hoc analysis of published cohorts8,9 (figure, C through F).

Sensory nerve conduction velocity slowing.

HNPP was first reported in the English literature more than four decades ago,3 and abnormally slow sural nerve conduction was described in one of the earliest detailed descriptions.5 Moreover, prolonged sensory nerve latencies were demonstrated in recent data.10 Yet attention to widespread sensory polyneuropathy has not been emphasized, nor have electrophysiologic comparisons been made with other conditions in which entrapment neuropathy is more commonly seen.

To confirm whether SNCV nerve slowing as shown here is a consistent feature of HNPP, published data of other cohorts was analyzed. Because the frequency of abnormal sensory responses has not been reported in large series, data were pooled from pathologically or genetically diagnosed cases in the 12 reports obtained by a MEDLINE literature search (involving 53 subjects1,5,6,13-21 providing sufficient information for analysis). Abnormally slow sensory responses were reported in 82% (60/73) of ulnar and sural sensory nerves tested (median data ignored because it spans a typically affected entrapment site), and the mean SNCV of HNPP in the only published cohort8 permitting calculation was similar to that obtained here (see figure, A and B).

In summary, diffuse SNCV slowing is a common feature of HNPP, indicating a background of dysmyelinative polyneuropathy independent of entrapment neuropathy. It thus provides a useful marker of the condition for the clinical electromyographer. The diffuse distribution of the SNCV slowing in HNPP was seen to be in contrast with CIDP and diabetes in which it was focal and much less frequent.

Motor nerve conduction slowing.

In contrast to widespread SNCV slowing, slowing of MNCV in HNPP was less common and tended to be minor. Notably, this occurred in the face of frequent and relatively marked prolongation of DMLs, indicating that there is a motor as well as a sensory polyneuropathy. This distally accented motor slowing confirms an observation demonstrated previously.10 Evidence that this too is a feature of an HNPP polyneuropathy can be seen in the figure, C, where derived data from other cohorts are shown.8,9 Such a pattern of distal motor conduction slowing (most clearly demonstrated by indices of low terminal latencies) is an unusual electrodiagnostic finding that also is characteristic of IgM monoclonal gammopathy directed against myelin-associated glycoprotein or sulfated glucuronyl paragloboside.22 Like peripheral myelin protein-22, myelin-associated glycoprotein is believed to be important for myelin compaction, and a defect in the Schwann cell–axon interaction by these adhesion molecules has been suggested to underlie both disorders.23

The reason why conduction slowing in HNPP is preferentially apparent for sensory rather than motor nerves likely is simply because of the method used to determine velocity determination: Because the slowing in HNPP nerves is predominantly distal, motor nerve conduction is necessarily calculated using more proximal sites of stimulation. The prolonged motor latencies support this contention. The mechanism of the reduced conduction velocities and prolonged F-wave latencies in HNPP likely is secondary to the segmental demyelination seen histopathogically,1,5,23 although a minor role for reduced axonal diameter5,14 also has been postulated. The features described for HNPP polyneuropathy are independent of entrapment neuropathy. Therefore, clinical differentiation from other conditions such as diabetes are easier if the carpal tunnel entrapment site is excluded from analysis.

Motor nerve conduction slowing in CIDP is more frequent and more severe than SNCV slowing,24 as seen in the table and figure. This is the opposite of what is found for HNPP. One possible reason for this contrast is the preferential loss from analysis of recordable sensory responses as compared with motor responses in CIDP. Sensory responses are more susceptible to phase cancellation from temporal dispersion, whereas the background slowing in HNPP is not focal, similarly affecting all nerve fascicles.

HNPP has a distinctive background sensorimotor polyneuropathy on which entrapment neuropathies are superimposed. The HNPP phenotype occurring without the typical 1.5-Mb chromosome 17 deletion has been described both sporadically10,25 and in kinships26 and has clinical and electrophysiologic characteristics identical to those carrying the deletion. The findings in our cohort are thus likely to be applicable to HNPP cases without the deletion. The following electrodiagnostic profile of HNPP should raise suspicion of the disorder: 1) conduction velocity slowing in most sensory nerves; 2) relatively less frequent and more minor MNCV slowing; 3) prolonged DMLs (or low terminal motor latency indices); and 4) prolonged F-wave latencies.