Abstract
Introduction Late onset epilepsy (LOE, onset >50y) is often attributed to occult cerebrovascular disease (CVD), and is associated with a nearly three–fold increased stroke risk. The basis of epileptogenecity associated with occult CVD is obscure, especially in the absence of overt cortical structural involvement by CVD, but is likely to involve disruption of cortico–subcortical circuits. We investigated haemodynamic measures, blood brain barrier (BBB) permeability, and cerebral and white matter lesion volume in a pilot clinical imaging study of patients with LOE and age–matched healthy controls (HC). Methods 16 LOE patients without clinical stroke or TIA, and 14 HC, underwent a 3T MRI scan protocol which included arterial spin labelling (ASL), high–resolution T1–weighted structural image (grey matter volume), fluid attenuated inversion recovery (FLAIR) (white matter lesion [WML] volume) and dynamic contrast enhanced (DCE) (BBB permeability) sequences. ASL enabled quantitation of cerebral blood flow (CBF), arrival time (tA–time for labelled blood to reach tissue of interest) and BOLD (blood oxygenation level dependent) signal. A vasodilator stimulus (hypercapnia) enabled measurement of cerebrovascular reactivity (CVR), which was calculated as % change in CBF, tA and BOLD divided by % change in end tidal CO2 (ETCO2). Results No difference was seen between the groups in baseline ETCO2 values, or the change induced by hypercapnia. Baseline CBF was lower in LOE patients compared to HC, but this difference was not significant. Baseline tA showed a trend towards being longer in the patient group (1483 vs 1363 ms, p=0.09), potentially suggesting lower blood velocity. Induction of hypercapnia caused CBF and BOLD to increase in both groups and tA to decrease, but in each case the response was greater in the control group. A non–significant trend was observed towards greater changes in HC than LOE patients in respect of CVR_CBF, CVR_tA, and CVR_BOLD. Voxel–based analysis however suggests significant regional differences in tA. Grey matter volume was significantly reduced in the patient group compared to the control group (p=0.02). WML volume was significantly higher in the patient group (p<0.05). Kinetic modelling revealed grey matter contrast leakage (ktrans) to be significantly greater in the patient group (p=0.01). ktrans showed significant correlation with WML volume in respect of both white matter and grey matter leakage (p=0.02). Conclusions Our findings indicate differences in both structural and haemodynamic measures of CVD between LOE patients and HC, even in the absence of clinical CVD. We observed trends towards longer baseline tA and reduced tA in response to hypercapnia in LOE patients, and significant regional tA differences. The relevance of these changes in tA is presently unclear, but may be indicative of haemodynamic changes in response to otherwise occult CVD. Significantly greater WML volume, likely to be predominantly due to cerebral small vessel disease, correlated with increased BBB permeability, itself likely to be important in epileptogenesis. Further study will be required to confirm and extend these preliminary findings and to delineate potential mechanisms.