Profiling migraine patients according to clinical and psychophysical characteristics: a cluster analysis approach

Abstract

Aim: This study aims to profile migraine patients according clinical and psychophysical characteristics. Method: In this observational study, two cohorts of migraine patients(episodic/chronic) were included. Cohort-1: ictal/perictal phase; Cohort-2: interictal phase. The following variables were assessed: headache frequency; disability; cervical active range of motion(AROM) in flexion, extension, right/left lateral flexion, right/left rotation; pressure-pain threshold(PPT) over: temporalis, two cervical areas(C1/C4 vertebral segments), and two distal pain-free areas(hand/leg). Cluster analysis was performed using the K-means algorithm. Differences across clusters were investigated. Results: Cohort-1: 100 patients were included, and two clusters were identified. Cluster-1.1 (19%), Cluster-1.2 (81%). Cluster 1.1 had a higher percentage of men (P = .037) and higher disability (P = .003) compared to Clusters 1.2. Cluster 1.2 had reduced AROM in flexion, extension, and left/right lateral flexion (P < .037), and lower PPT value in all areas (P < .001) compared to Cluster 1.1. Cohort-2: 98 patients were included and three clusters were identified. Cluster-2.1(18%), Cluster-2.2(45%), and Cluster-2.3(37%). Cluster-2.1 had a higher percentage of men compared to clusters-2.2 and 2.3 (P = .009). Cluster-2.3 had higher headache frequency, and disability compared to Cluster-2.2 (P < .006), and higher disability compared to Cluster-2.1 (P = .010). Cluster-2.3 had reduced AROM in all directions compared to Clusters-2.1 and 2.2 (P < .029). Clusters-2.2 and 2.3 have lower PPT values in all areas compared to Cluster-1.1 (P