Addiction is a feared consequence of long-term opioid treatment of chronic pain patients. The ICD-10 and DSM-IV diagnostic addiction criteria may not be appropriate in these patients. Therefore Portenoy's criteria (PC) were launched. The aim was to estimate the prevalence of addiction, to investigate whether PC were applicable and to compare these criteria with the ICD-10 criteria. The study was cross-sectional and included 253 patients with chronic pain at a tertiary pain centre. Patients were screened for addiction by a physician and a nurse. The addiction prevalence was 14.4% according to ICD-10 and 19.3% according to PC. A significant difference between the prevalence of addiction according to ICD-10 and to PC was found. The inter-rater reliability was 0.95 for ICD-10 and 0.93 for PC. The sensitivity of PC was 0.85 and the specificity was 0.96. According to PC patients classified as addicted were treated with significantly higher opioid doses, drank more alcohol, smoked more tobacco, used benzodiazepines and had higher levels of depression. According to ICD-10 patients classified as addicted used higher doses of opioids, drank more alcohol and had higher scores of anxiety and depression. High opioid doses, concomitant use of alcohol and younger age were risk factors. The risk profile for PC was different to ICD-10 by adding risk factors as concomitant use of benzodiazepines, having depression and low educational level. PC seems to be appropriate for diagnosing addiction in opioid treated pain patients and seems to be more sensitive and specific than ICD-10 criteria.
European Journal of Pain, 2010, Vol 14, Issue 10, p. 1014-20