Can the LANSS scale be used to classify pain in chronic cancer pain trials?

Abstract

The Leeds Assessment of Neuropathic Symptoms and Signs (LANSS) scale was developed to differentiate pain of predominantly neuropathic or nociceptive origin. The aim of this study was to determine whether the LANSS scale was an appropriate tool to classify pain in a trial of patients with advanced cancer and chronic refractory pain. Clinician assessment of pain type (neuropathic or nociceptive) was used to determine the sensitivity and specificity of LANSS scores in 112 trial participants. Those classified as "mixed" or of uncertain aetiology were excluded. We undertook several analyses in an attempt to improve the LANSS scale and better diagnose pain type for our specific dataset. There was strong association between the LANSS score and a diagnosis of neuropathic versus nociceptive pain, p < 0.001. When the clinical assessment was compared with the LANSS scale, the overall accuracy was 94% (79/84). The 5 false negatives and no false positives resulted in a sensitivity of 0.86 (0.70, 0.95), specificity of 1 (0.93, 1), positive predictive value of 1 (0.88, 1) and negative predictive value of 0.91 (0.80, 0.97). The negative likelihood ratio was 0.14 (0, 0.32). The scale had good discriminant and construct validity. Reliability was assessed via internal consistency with Cronbach's α = 0.76, similar to that of the original validation study (α = 0.74). None of the new scales developed was better at differentiating pain type. The LANSS scale predicted well for pain type in a cancer population and is a useful tool for classifying pain in cancer pain trials.