The use of electrodiagnostic, data-based clustering may be able to distinguish electrodiagnostic feature patterns among those with chronic inflammatory demyelinating polyneuropathy (CIDP), according to a study recently published in the Journal of Neurology, Neurosurgery & Psychiatry. Poor treatment response may not be predicted by reduced distally evoked compound muscle action potentials (CMAPs).

This retrospective study included electrodiagnostic and clinical data from 56 individuals with confirmed CIDP at 2 teaching hospitals. Standard surface electrode recording with electromyography technology and percutaneous supramaximal stimulation were performed for nerve conduction studies. Individuals were clustered into subgroups of similar electrodiagnostic features through the use of a hierarchical agglomerative clustering algorithm.

Related Articles

Predictors of long-term outcome were analyzed using a stepwise logistic regression analysis. A chi-squared or Fisher’s exact test was used to compare categorical variables, while a Mann-Whitney U test was used to compare continuous variables.

Continue Reading

The clustering algorithm resulted in two clusters, one of which was distinguished by significant conduction slowing and coexistent reduced distally evoked CMAP amplitudes.

Distal-acquired demyelinating symmetric polyneuropathy significantly overrepresented this cluster compared with the other (70% vs 26.1%; P =.042).

This cluster was associated with 100% successful long-term treatment outcome, compared with 63% in the other cluster (P =.023). Poor long-term outcome was predicted by initial disability (odds ratio [OR] 6.1; 95% CI, 2.4-25.4), distal CMAP duration (OR 0.96; 95% CI, 0.91-0.99), and F-wave latency (OR 0.93; 95% CI, 0.86-0.98).

Limitations to this study include a retrospective and uncontrolled design, a lack of information on arm function in the modified Rankin scale, potential type I error inflation, and the potential for information loss due to a combination of values from multiple nerves.

The study researchers conclude that “clinical subtypes and long-term treatment outcomes were significantly different between the two [electrodiagnostic] data-based patient clusters, indicating the phenotypic and prognostic implications of the unsupervised approach.”


Baek S-H, Hong Y-H, Choi S-J, et al. Electrodiagnostic data-driven clustering identifies a prognostically different subgroup of patients with chronic inflammatory demyelinating polyneuropathy [published online March 23, 2019]. J Neurol Neurosurg Psychiatry. doi:10.1136/jnnp-2018-319758