INTRODUCTION: Quantitative Sensory Testing (QST) is a valuable tool for understanding chronic low back pain (cLBP) through the lens of central sensitization and pain mechanisms. By measuring reactions to regulated sensory stimuli, QST can identify changes in excitatory and inhibitory pain mechanisms, providing insights into central pain processing changes associated with chronic conditions. This study employs Latent Class Analysis (LCA) to phenotype cLBP patients and generate latent classes based on their responses to QST measures.
METHODS:The analysis was conducted using the Longitudinal Clinical Cohort for Comprehensive Deep Phenotyping of Chronic Low-Back Pain (cLBP) Adults Study (COMEBACK) from the University of California, San Francisco (UCSF) Core Center for Patient-centric, Mechanistic Phenotyping in Chronic Low Back Pain (REACH) Mechanistic Research Center of NIH HEAL BACPAC. LCA using StepMix package v2.2.1 in python to identify unobserved subgroups within the population based on several key variables, including control summation, test summation, control threshold, test threshold, and lingering pain sensations at 15 and 30 seconds at both the control and test sites. In this study, the control site was the shoulder, and the test site was the back. The optimal number of latent classes was determined via 10-fold cross validation with 100 bootstraps to optimize AIC (Akaike Information Criterion), BIC (Bayesian Information Criterion), and entropy. One-way ANOVA with Bonferroni correction was used to evaluate differences key variables between the classes in further detail.
RESULTS: A total of 350 patients were included in the analysis and five different latent classes were determined to be the optimal number of classes with class sizes of 24, 21, 168, 79, and 58, respectively. ANOVA findings with Bonferroni adjustments revealed statistically significant differences between classes (p < 0.05) in key areas such as temporal summation of pain, and other related QST measures specifically related to lingering pain.
DISCUSSION: This study used LCA to identify five distinct latent classes among patients with chronic low back pain (cLBP) based on their responses to QST. The classes were indicated by differences in sensory processing and pain modulation mechanisms, highlighting the variability in how patients experience and manage pain. For instance, certain classes showed higher pain sensitivity and reduced pain modulation, indicating central sensitization, while others demonstrated better pain regulation through more effective inhibitory mechanisms. These findings suggest that not all patients with cLBP are alike in their pain experience and provide evidence for the presence of multiple pain processing mechanisms within the population. The identification of these latent classes offers significant potential for the development of more personalized treatment strategies. These findings could inform future research on the relationship between pain mechanisms and psychosocial factors, such as anxiety, depression, and emotional awareness, which may further influence the progression and management of chronic pain. Overall, this study contributes to a deeper understanding of the heterogeneity in cLBP.
Keywords: Chronic low back pain (cLBP), Latent Class Analysis (LCA), Quantitative Sensory Testing (QST)