Phantom pain has been given considerable attention in literature. Phantom pain reduces quality of life, and patients suffering from phantom pain make heavy use of the medical system. Many risk factors have been identified for phantom pain in univariate analyses, including phantom sensations, stump pain, pain prior to the amputation, cause of amputation, prosthesis use, and years elapsed since amputation. Multivariate analyses are lacking in the literature and, therefore, no estimation of an overall risk for phantom pain can be made. The aim of this study was to analyze risk factors in a multivariate analysis in 536 subjects (19% upper limb amputees and 81% lower limb amputees). These subjects filled out a questionnaire in which the following items were assessed; side, date, level, and reason of amputation, pre-amputation pain, presence or absence of phantom pain, phantom sensations and or stump pain, and prosthesis use. The prevalence of phantom pain was 72% (95% CI: 68 to 76%) for the total group, 41% (95% CI: 31 to 51%) in upper limb amputees and 80% (95% CI: 76 to 83%) in lower limb amputees. The most important risk factors for phantom pain were "bilateral amputation" and "lower limb amputation." The risk for phantom pain ranged from 0.33 for a 10-year-old patient with a distal upper limb amputation to 0.99 for a subject of 80 years with a bilateral lower limb amputation of which one side is an above knee amputation.