Two reports are published that represent important contributions to our understanding of the natural history of cerebrovascular disease, one addressing the haemorrhage and growth risk of small paraclinoid aneurysms and the other addressing arteriovenous malformation (AVM) features associated with seizure presentation.

Jeon and colleagues¹ evaluated a cohort of 524 patients possessing very small, unruptured paraclinoid aneurysms, revealing an exceedingly low rate of rupture (2 cases) and growth (17 cases) over 1675.5 aneurysm-years. Large natural history studies often group internal carotid artery (ICA) or internal carotid artery aneurysms except posterior communicating artery aneurysms (non-PCcomm). This can obscure the haemorrhage risk of paraclinoid aneurysms as a result of mixing them with ICA bifurcation and anterior choroidal artery aneurysms,1 ,2 the latter more akin to posterior communicating artery aneurysms. This more selective study allows for a direct analysis of very small paraclinoid aneurysms and reinforces their generally very low risk of rupture. However, like all natural history studies,2 ,3 selection bias is an important limitation; in this report, nearly half of the patients presenting with unruptured paraclinoid aneurysms were initially selected for treatment. Although reasons for early treatment were not provided and likely individualised to variables including patient preference, it is likely that a considerable subset of these aneurysms possessed concerning morphologic features. As a result, selection bias likely deflated the observed rupture risk. In particular, it is doubtful that blister aneurysms were observed, making the conclusion about ‘non-arterial branch-related’ aneurysms possessing a lower risk of rupture misleading. This may also reflect an obscurity in terminology, as some would specifically consider ‘non-arterial branch-related’ aneurysms to implicate a blister-type aneurysm. Another vagary in terminology is the overuse of the term ‘paraclinoid.’ Paraclinoid aneurysms include aneurysms of the clinoidal segment (extradural origin) and ophthalmic segment (intradural origin). Although large clinoidal aneurysms may extend intradurally, it is likely that the very small clinoidal aneurysms in this study were not intradural. This makes an assessment of their risk of subarachnoid haemorrhage moot and deflating to the more important cohort of intradural ophthalmic segment aneurysms. Despite these limitations, the authors are congratulated on this important, timely addition to the literature, coming in an era where paraclinoid aneurysms are fervently treated by endovascular means, even very small lesions at some institutions. Particularly when considering wide-necked lesions, added morbidity from stent-assisted or remodelling techniques or even off-label usage of flow diversion must be weighed against accumulating evidence from studies like these that suggest a very low risk of rupture for these aneurysms.

Although most natural history studies of AVMs focus on rupture risk,4 it is important to underscore that approximately 25–30% of these lesions will present with seizures,4 another source of neurological morbidity. Galletti and colleagues illustrated that seizure on presentation was associated with a superficial AVM location, particularly in the temporal lobe. While size was evaluated as a categorical variable, it only showed a slight trend towards significance. One might postulate that if size was evaluated as a continuous variable, it may have also been a significant risk factor for seizures. Increasing AVM size and greater seizure risk has been postulated as the cause of the discrepancy between old natural history studies of AVMs that purported small size as a risk factor for haemorrhage on presentation and more recent studies employing prospective follow-up periods that have not. As opposed to analyses over prospective follow-up periods, those based on haemorrhagic presentation can be obscured if small lesions are likely to remain quiescent unless they bleed and if large lesions are more likely to present with other symptoms, particularly seizures. By analogy, this carries over to the analysis of AVM epileptogenesis, where risk may be better assessed prospectively over a time period rather than only at presentation. Nevertheless, the report by Galletti and colleagues⁵ is an important confirmatory study. Natural history studies often associate superficial lesions and those with superficial venous drainage with a lower haemorrhage risk.4 However, a comprehensive analysis of AVM-associated morbidity must also add-in the risk of seizures. It would seem that from the perspective of AVM location, the latter seems to have an inverse relation to haemorrhage risk. Does this make all AVMs in all locations dangerous and worthy of treatment? We hope the ARUBA (A Randomized trial of Unruptured Brain Arteriovenous Malformations) study will begin to answer these important questions, from the perspective of haemorrhage-associated morbidity and epilepsy-associated morbidity from treatment and conservative management.