With interest we read the review article by Gramegna et al. about the involvement of cerebral vessels in mitochondrial encephalopathy, lactic acidosis, and stroke-like episode (MELAS) syndrome [1]. Among 46 MELAS patients reported in 23 papers dilatation of the major cerebral vessels was found in 37% (in 88% of them during a stroke-like lesion (SLL)) and stenosis was reported in 24% of the cases [1]. It was concluded that dilatation or stenosis occurs during a SLL in MELAS [1]. The study is appealing but raises the followin1g comments and concerns. 

The main shortcoming of the study is its design. To assess the reagibility of intracerebral arteries during a SLL by means of magnetic resonance angiography (MRA) or digital subtraction angiography (DSA) a prospective design would be more appropriate. 

A further shortcoming is that the current medication the 46 patients were taking at the time of the MR investigations was not considered. Since several drugs may influence the arterial diameter, it is crucial to know to which degree the arterial diameter was influenced by the current medication. Another factor that may influence the diameter of cerebral arteries during a SLL is the heteroplasmy rate and the mtDNA copy number, factors which were not considered in the review. 

We do not agree with the statement that major vessel tone abnormalities could contribute to the pathophysiology of a stroke-like lesion (SLL) [1]. Though the vascular hypothesis of SLL development has been repeatedly propagated [2], there are more arguments against than in favour of this hypothesis. The strongest argument against the vascular hypothesis is that a SLL is not confined to a vascular territory. Hyperperfusion within the territory of an acute SLL, as documented on perfusion-weighted imaging (PWI) or HMPAO single photon emission tomography (SPECT) [3], most likely a compensatory mechanism for the locally disturbed metabolism. It is well established that the oxygen-extraction is reduced within the territory of a SLL on oxygen extraction fraction (OEF) MRI and that there is hypometabolism on fluor-deoxy glucose positron emission tomography (PET) suggesting that the oxygen and metabolite supply is increased trying to counterbalance the impaired energy output within the area of a SLL [4]. A further argument against the vascular hypothesis is that the SLLs are dynamic lesions that usually expand over time, to regress during the chronic stage of a SLL. This dynamic nature cannot be explained by impaired perfusion. 

Furthermore, it is well established that MELAS syndrome is a multisystem disorder affecting every organ or tissue to varying degrees, including the arteries [5]. Arteriopathy in MELAS may manifest as atherosclerosis of large or small arteries (macro-/micro-angiopathy), as spasms, dilatation, ectasia, aneurysm formation, or even dissection [6]. Notably, not only the intra- and extra-cranial arteries but all arteries may be affected. Importantly, arteriopathy has to be delineated from vascular reactions within a SLL. Arteriopathy in MELAS can be primary due to the underlying genetic defect or secondary due to diabetes, arterial hypertension, or hyperlipidemia, frequent phenotypic manifestations of mitochondrial disorders (MIDs), including MELAS [6]. Thus, arteriopathy can be present in MELAS or other MIDs even in the absence of a SLL.

Overall, the review has several limitations which should be addressed before drawing final conclusions. Factors influencing the arterial diameter, such as drugs or heteroplasmy rates should be considered. The metabolic hypothesis to explain the development of a SLL should be discussed.