The effect of incomplete patient follow-up on the reported results of AVM radiosurgery.

Background: The reported efficacy of AVM radiosurgery--80-85% 2-year obliteration rate--is based exclusively on the results of follow-up arteriography in a small percentage of treated patients; it is therefore inaccurate. We examined the effect of incomplete follow-up on the results of AVM radiosurgery.

Methods: We reviewed the results of AVM radiosurgery in 82 patients after a minimum of 24 months of follow-up. Patients were not preselected to undergo arteriography on the basis of any other imaging study. Data were analyzed using the Kaplan-Meier method and stratified by size of AVM. Results were compared with those obtained from the same data using the reporting techniques described in the literature.

Results: When data analysis was limited to patients who had follow-up arteriography, the 2-, 3-, and 4-year obliteration rates were 37%, 73%, and 84% after a minimum 24-month follow-up. Using Kaplan-Meier analysis the 2-, 3-, and 4-year obliteration rates were 32%, 55%, and 55% (95% CI = +/-18%), respectively. The 2-year obliteration rate was 43% for AVMs <30 mm in diameter and 16% for AVMs >30 mm in diameter, respectively.

Conclusion: If data analysis is limited to the patients who undergo follow-up arteriography, the obliteration rate of AVM radiosurgery is overestimated. The actual 2-year obliteration rate if all data is considered is in the range of 40% rather than the commonly reported 80%. Therefore, treated patients are exposed to the risk of intracerebral hemorrhage for a longer period than previously appreciated. Compulsive long-term follow-up is required to document the true AVM obliteration rate after treatment by radiosurgery.