Low back related leg pain: development and preliminary validation of a new classification system

Abstract

Background summary. Leg pain is a common complaint in relation to low back pain (LBP), present in up to 65% of all patients with LBP. Radiating leg pain is an important predictor for chronicity of LBP and an indicator of the severity of the disorder. Consequently patients with back and leg pain account for a disproportionate amount of the costs of medical care and disability compensation caused by LBP. As many structures in the lower quarter are capable of evoking a similar pattern of pain, the primary pathology causing low back related leg pain is often difficult to differentiate resulting in inappropriate investigations and treatment.A number of published classification systems for LBP based on symptom response and pathoanatomy incorporate radiating leg pain. While these systems may be suitable and successful for LBP, this approach may be insufficient for the more complex pain conditions related to nerve injury. Failure to correctly classify subjects into homogenous subgroups could be one reason why randomized controlled trials often report conflicting results for conservative treatment of patients with sciatica. Mechanism based classification of patients is crucial in that it not only provides a better understanding of the presenting complaint, but also directs treatments which interact with specific neurophysiologic pain mechanisms. Therefore we developed a new classification system for low back related leg pain on the basis of pathomechanisms extending the ideas originally proposed by Elvey and Hall, with four distinct subgroups. These hierarchical categories are: Neuropathic Sensitization (NS) comprising major features of neuropathic pain with sensory sensitization; Denervation (D) arising from significant axonal compromise without major features of neuropathic pain; Peripheral Nerve Sensitization (PNS) arising from nerve trunk inflammation without clinical evidence of significant denervation; and Musculoskeletal (M) referred from non-neural structures such as the disc or facet joints.Objective. The objective of the research project was to investigate aspects of validity of the classification system: Inter-rater reliability, concurrent validity, and predictive validity.Methods. As there is no gold standard against which the classification system can be tested, three studies were undertaken to investigate inter-rater reliability, concurrent validity and predictive validity. Inter-rater reliability was tested consecutively by six blinded experienced clinicians in a subset of 40 patients. Concurrent validity was investigated by implementing a comprehensive quantitative sensory test (QST) protocol as a criterion measure within a cross-sectional cohort design. The aim was to explore whether sensory profiles of the different diagnostic groups would correlate with the putative pathomechanisms in each group. To investigate predictive validity we conducted a longitudinal cohort study to explore whether treatment response differs between diagnostic groups. Outcome measures were pain intensity, degree of disability and patient global perceived change as well as QST parameters related to small nerve fibre function and pain sensitivity.Results. Seventy-seven patients with unilateral low back related leg pain and 18 healthy age and sex matched controls were included. The proportion of patients in each classification were 26% in Group NS, 36% in Group D, 12% in Group PNS and 26% in Group M. Interrater reliability was good (percentage agreement = 80; κ = 0.72; 95% CI = 0.57 to 0.86). QST revealed signs of pain hypersensitivity in Group NS compared to a healthy control group. Significant sensory deficits occurred only in Group D and were most pronounced over the affected foot. QST parameters in Groups PNS and M were not significantly different to healthy controls. The longitudinal cohort study showed that the proportion of patients responding to treatment was greater in PNS (55.6%) than the other three groups (NS 10%; D 14.3 % and M 10%; Fishers exact test p=0.031). The mean magnitude of improvement was greater in Group PNS in all outcome measures compared to the other groups (p ≤ 0.052). Additionally, Group PNS showed a greater improvement pre to post treatment of C fibre function, decreased sensitivity to cold pain and decreasedtemporal summation compared to the other groups (all p<0.05). In contrast, GroupNS exhibited loss of C fibre function and increased pressure pain sensitivityfollowing treatment (all p<0.05).Discussion. Patients with low back related leg pain could be reliably classified into four groups by experienced musculoskeletal physiotherapists using a new classification system. Reliability is an essential requirement for any classification system to be valid. Significant differences in QST thresholds were found in groups NS and D when comparing them to a group of healthy controls, indicating that in these two groups sensory dysfunctions were most pronounced. These dysfunctions match findings from the clinical examination: Sensory deficits in Group D and pain hypersensitivity in Group NS. These findings support concurrent validity. The findings of the longitudinal study indicate that the new classification system identifies a subgroup of patients (PNS) more likely to respond to neural mobilisation intervention. This result contributes to the evidence for predictive validity of the classification system and supports the concept that correct categorization of subjects will improve treatment outcomes.Conclusion. The findings of our research project support validity of the new classification system and its use as a reliable and feasible tool to improve diagnosis and treatment outcome in patients with low back related leg pain.