Which variables are associated with pain intensity and treatment response in advanced cancer patients?--Implications for a future classification system for cancer pain

Background: This study is part of a research program to reach consensus on an international cancer pain classification system. A confirmative and explorative approach was applied to investigate which of the variables identified in the literature, by experts and patients that are associated with pain. Methods: Data from an international, multicentre, cross-sectional study of cancer patients treated with opioids were investigated. Dependent variables were: average pain, worst pain, and pain relief (11-point Numerical Rating Scales). Forty-six independent variables were chosen based upon previous studies. Bivariate analyses identified independent variables associated with at least one of the dependent ones; 21 were included in multivariate linear regression analyses. Results: Two thousand two hundred and seventy-eight patients were investigated; 52% males, mean age 62 years, mean Karnofsky Performance Status 59%, mean daily opioid oral equivalent dose 341 mg. Fifty-eight percent had breakthrough pain. Mean pain scores were: average pain 3.5, worst pain 5.3 and pain relief 74%. Variables most strongly associated with these three dependent variables were: breakthrough pain, psychological distress, sleep, and opioid dose. Conclusions: Breakthrough pain and psychological distress were confirmed as key variables of a future classification system. Candidate variables were: sleep, opioid dose, pain mechanism, use of non-opioids, pain localisation, cancer diagnosis, location of metastases, and addiction.