WE PRESENT A simple risk prediction formula for arteriovenous malformation hemorrhage. Natural history studies have shown an annual risk of hemorrhage of 2 to 4% for patients with brain arteriovenous malformations. Although decision analysis programs and biostatistical models are available to predict long-term risks of hemorrhage, we hypothesized that there was varying knowledge regarding the use of such programs within the general neurosurgical community. To obtain information on the current use of risk data, we performed a survey of neurosurgeons at national meetings in 1988 and 1994. Neurosurgeons were asked to define the risk for arteriovenous malformation hemorrhage in the young adult patient over a 20- to 30-year period, given a 3 or 4% annual risk of hemorrhage. A wide range of answers was obtained (1–100% risk), and many different methods of calculation were used. The use of the multiplicative law of probability formula requires only knowledge of patient age and annual hemorrhage risk. Risk of hemorrhage = 1 − (risk of no hemorrhage)expected years of remaining life. The assumptions pertaining to this multiplicative formula include a constant yearly risk of hemorrhage and the independent behavior of all years of observation. We calculated the predictions of risk of hemorrhage across all age groups, as modified by published survival data. We think the use of this formula is justified by published natural history data across different ages and populations and that it is a simple and reasonable alternative to other methods of calculation.
Department of Neurological Surgery (DK, MRMcL), Presbyterian University Hospital, University of Pittsburgh, Pittsburgh, Pennsylvania, and Division of Neurosurgery (JRWK), Research Consulting Unit, British Columbia's Children's Hospital, University of British Columbia, Vancouver, British Columbia, Canada
Reprint requests: Douglas Kondziolka, M.D., M.Sc., F.R.C.S.(C), Associate Professor of Neurological Surgery and Radiation Oncology, Suite B-400, Presbyterian University Hospital, University of Pittsburgh Medical Center, 200 Lothrop Street, Pittsburgh, PA 15213-2582.
Received, January 12, 1995. Accepted, May 26, 1995.