Phenotypic Stratification of Patients With Painful Neuropathy May Predict Response to Therapy

diabetic neuropathy
The researchers’ objective was to compare the classification of patients with that reported in a recent study by identifying 3 clusters of patients with different sensory profiles.

A transetiologic, phenotype-based approach was associated with response to treatment and may improve the management of neuropathic pain. These findings were published in the journal Pain.

Patients (n=628) with probable or definite neuropathic pain were recruited from the Pain Center at Ambroise Paré Hospital in France. All patients were assessed by a standard clinical neurologic examination. They then rated their pain during the past 24 hours on an 11-point numerical scale and completed the Neuropathic Pain Symptom Inventory (NPSI).

A clustering approach was used to stratify patients by pain phenotypes. Researchers used data from a previous study of patients (n=97) with neuropathic pain who were randomly assigned to receive botulinum toxin A (BTX-A) or placebo to assess whether patient clusters would be more or less responsive to BTX-A treatment.

The internal and BTX-A cohorts comprised patients who had a mean age of 56.4 (±15.2) and 52.3 (±16.1) years, 51.9% and 50.5% were women, their duration of pain was 64.8 (±74.7) and 63 (±67) months, and their pain intensity was 6.4 (±1.7) and 6.4 (±1.6) points, respectively.

Among the internal cohort, 3 major clusters were identified. Patients in cluster 1 (31.4%) had pinpointed pain with above-average paresthesia or dysesthesia and below-average evoked pain. Patients in cluster 2 (30.4%) had evoked pain with above-average pain that was provoked by brushing, cold, pressure, or electric shocks, and below-average deep pain and paresthesia or dysesthesia. Patients in cluster 3 (38.2%) comprised patients who had deep pain that felt like squeezing and above-average pressure with below-average paresthesia or dysesthesia.

Patients in the BTX-A cohort were stratified into 3 patient clusters. Patients in group 2 had a significant group (F, 6.71; P =.013) and group by time interaction (F, 2.24; P =.009). Similarly, patients in cluster 3 had both a significant group (F, 4.41; P =.042) and group by time interaction (F, 2.01; P =.021). BTX-A therapy did not have a significant effect compared with placebo among patients in cluster 1 (F, 0.35; P =.56).

This study was based on post hoc analyses and requires validation in a prospective cohort.

These findings suggest that a transetiologic, phenotype-based approach may have the potential to improve treatment selection for patients with painful neuropathy.


Bouhassira D, Branders S, Attal N, et al. Stratification of patients based on the Neuropathic Pain Symptom Inventory: development and validation of a new algorithm. Pain. 2021;162(4):1038-1046. doi:10.1097/j.pain.0000000000002130