Risks of history of diabetes mellitus, hypertension, and other factors related to radiation-induced changes following Gamma Knife surgery for cerebral arteriovenous malformations

Abstract

Diabetes mellitus (DM) and hypertension may be associated with complications following fractionated radiotherapy. To date no studies have determined the risk of radiation toxicity in patients with DM or hypertension who have undergone Gamma Knife surgery (GKS) for brain arteriovenous malformations (AVMs). The goal of the present study was to determine associations between DM or hypertension and other factors in the development of radiotoxicity, as measured by radiation-induced changes (RICs) on MR images following radiosurgery for AVM. Using univariate methods and multivariate logistic regression, the authors compared the RIC status in patients 18 years of age and older with these patients' history of, or medication use for, DM or hypertension; tobacco use; patient age and sex; AVM volume; Spetzler-Martin AVM severity scale (Grades I and II vs Grades III-V); AVM surgery, AVM embolization, or hemorrhage prior to radiosurgery; AVM location; number of draining veins; and radiosurgery margin dose. Radiation-induced changes occurred in 38% of 539 adults within a mean (± standard deviation) of 12 ± 10 months after radiosurgery, as observed during a median follow-up time of 55 months. Among patients in whom RICs occurred, 34% had headaches, neurological deficits, or new-onset seizures. Larger RICs were associated with worse symptoms. According to a univariate analysis, DM (3% of patients), larger AVM volume, worse Spetzler-Martin grade, lack of AVM surgery prior to radiosurgery, lack of hemorrhage prior to radiosurgery, and smaller margin dose of radiation had significant associations with the presence of RICs. Hypertension (20%), patient sex, tobacco use, number of draining veins, superficial or deep location of the lesion, and AVM embolization prior to radiosurgery had no association with the presence of RICs. According to a multivariate analysis, larger AVM volume, worse Spetzler-Martin grade, and no AVM surgery prior to radiosurgery predicted the occurrence of an RIC. Diabetes mellitus had borderline significance. Vascular factors such as hypertension, patient sex, and tobacco use did not convey additional risks of radiotoxicity, but DM remained a possible cardiovascular risk factor in the development of RICs.