Fifteen of the studies included were case–control (15–22) and case-series studies (12, 13, 23–27) and one study was a cohort study (28). The demographic and clinical characteristics of subjects included in these 16 studies are summarized in Tables 1, 2. The aggregate number of patients included in this review was 459 (range, 3–126). All studies included sCeAD adult patients with single vessel, bilateral or multivessel involvement, and they were compared to healthy controls, other than two studies without a control group (12, 24). Fifteen studies reported ages ranged from 18–70, but most of the patients were young adults with a mean age of 41.5 ± 9.3 years. The time since the event/stroke was mostly not reported. All studies evaluated possible underlying connective tissue disorders by analysis of skin samples obtained from biopsy, in which histologic and/or ultrastructural changes were searched for, as well as by genetic analysis for possible gene involvement. Three studies investigated families with sCeAD patients, for a total of 11 families with 19 sCeAD patients (24–26).

Table 1 summarizes the findings from ten studies (12, 13, 16, 19–21, 24–28) that evaluated the connective tissue morphology showed ultrastructural aberrations, in the absence of clinical manifestations, such as alteration of the collagen and elastic fiber networks (12), significantly thinner dermis (19), smaller collagen fibrils diameter in the skin (20, 27), abnormal collagen fibril structures with faint and absent banding pattern (21, 24, 27), aberrations in numerous composite fibrils within mid-dermal collagen bundles with enlarged diameters of composite fibrils (13, 24) (very similar to the aberrations seen in patients affected by EDS-II or III (12, 16)) and elastic fiber abnormalities with mini calcifications and fragmentation (16). All of these studies were consistent in showing aberrations of the connective tissue, in a cumulative number of 149 out of 295 (50.5%) patients.

Besides the morphological evaluation, molecular analyses that were performed in 10 out of the 15 studies did not detect any causative mutations in genes encoding for collagen type I, III, V, and elastin and sCeAD (15, 17, 18, 22–26, 28). Table 2 summarized the findings of genetic analyses from skin biopsy studies.

Notwithstanding, the linkage analysis of a large family with sCeAD-associated connective tissue alterations suggested the presence of a candidate locus on chromosome 15q2 or on chromosome 10q26 (26), while other studies revealed copy number variants (CNV) enrichment in genes involved in extracellular matrix organization (COL5A2, COL3A1, SNTA1), collagen fibril organization (COL5A2, COL3A1), smooth cell functions (MHY11) and possibly in genes involved in transforming growth factor beta (TGF)-beta receptor signaling pathway (COL3A1, DUPS22) (22, 27). Finally, Mayer-Suess and co-workers, based on the analysis of the extracellular matrix proteins by the proteomics approach, showed different expressions of 25 proteins, 13 of which were linked to connective tissue disorders, in patients with recurrent sCeAD (28).

Overall, although the results of the genetic analyses included in this systematic review do not allow for definitive conclusions, they seem to support the hypothesis of a potential involvement of “connective genes” in sCeAD pathogenesis.

Our review showed that alterations of the collagen and elastic fiber networks are frequent findings in patients with sCeAD as opposed to what is observed in subjects without sCeAD. Such dermal connective tissue aberrations include irregular (flower-like) contours with variable diameters of collagen fibrils, and fragmented, moth-eaten-like appearance and microcalcifications of elastic fibers. These alterations are similar to those that may be found in classic HCTD such as in Ehlers–Danlos syndrome type IV or pseudoxanthoma elasticum, a further argument in favor of the “connective hypothesis” of sCeAD.

Over the last decades, it has been repeatedly emphasized that classical HCTD, such as EDS type IV and Marfan syndrome, or other more recently identified nosographic entities, such as the Loeys-Dietz syndrome (LDS) comprise sCeAD in their phenotypic spectrum. However, diagnostic criteria for HCTD are met in only 1%–5% of sCeADs patients (10, 29). This implicates that there might be other unrecognized connective tissue abnormalities predisposing the vessel wall to dissection. The results of the studies included in the present review support this hypothesis. First, a family history of CeAD in some cases indirectly suggests that genetic factors might be operant in the pathogenesis of the disease. This hypothesis was further supported by the observation that dermal ultrastructural connective abnormalities not fulfilling the diagnostic criteria for known HCTD (26) aggregate in familial groups in some cases, where they follow an autosomal-dominant pattern of inheritance.

Second, besides searching for morphological aberrations, molecular analyses were also conducted in some of the studies included in our systematic review. The majority of these studies focused on the genes of collagen type I (COL1A1) (25), type III (COL3A1) (15, 17, 25), type V (COL5A1, COL5A2) (23, 25) as well as elastin (ELN) (23), but they were unable to identify any causative mutations in sCeAD. Although collagen type III was particularly an area of interest as the mutation in this gene is related to EDS type IV (the “vascular” subtype of EDS), none of these studies identified causative mutations in this gene among patients with sCeAD other than the missense mutation G157S in two patients from the same family. The major limitation of these studies is inherent in their nature of small case series, which makes them not fully adequate for determining the prevalence of such causative mutations among sCeAD patients.

Similarly, most of the studies in which a genetic linkage analysis was performed gave negative results, with the only exception of one study suggesting a linkage between connective tissue alterations and mutations in genes involved in extracellular matrix and collagen fibril organization, especially in patients with recurrent sCeAD, in whom a greater pathogenic impact of structural vessel wall alterations is assumed.

More advanced methods such as whole-exome and whole-genome sequencing have been recently used to investigate patients with sCeAD. One study performed a genome-wide linkage analysis and found suggestive linkage between CAD-associated connective tissue alterations and mutation of the locus on chromosome 15q2 and chromosome 10q26 (26). This approach identified several promising candidate genes such as CSPG2, LOXL1, and FGFR2, all already known to be involved in aortic dissections and aneurysms formation.

Another study searched for rare genetic deletions and duplications that predispose to sCeAD based on the analysis of CNV (22). This study showed significant CNV enrichments in genes involved in extracellular matrix organization (COL5A2, COL3A1, SNTA1) and collagen fibril organization (COL5A2, COL3A1). Interestingly, none of these rare CNV enrichments were found in more than one patient, which indicated that the underlying genetic variation is a complex and heterogeneous process that cannot be explained with simple monogenetic variations. Furthermore, one small study performed whole-exome sequencing and CNV analysis in patients with independently occurring dissections in both the aorta and cervical arteries, and similarly identified pathogenic CNV in COL3A1, COL5A2, and MYH11 genes, which all involved in arterial connective tissue functions (27).

Finally, a recent study used a cutting-edge quantitative proteomics approach to identify extracellular protein aberrations in sCeAD patients and identified 25 proteins expressed differently only in patients with recurrent sCeAD (28). The Authors also identified two main protein clusters; (1) desmosome-associated cluster with four proteins and (2) collagen and elastin cluster with eight proteins, suggesting a more complex pathophysiology of sCeAD and possibly more evident aberrations at the proteome level. These results suggest underlying pathological genetic variants, particularly in patients with recurrent and multiple territory dissections.

The literature suggests further arguments in favor of the “connective hypothesis” in the pathogenesis of sCeAD.

First, echocardiographic studies assessing cardiac manifestations of connective tissue disorder showed that valvular abnormalities such as mitral valve prolapse, mitral valve dystrophy, aortic valve dystrophy as well as the enlarged diameter of the aortic root were observed more frequently in sCeAD patients (56%) compared to controls (15%) (30). One study, in particular, showed that aortic root dilatation, one of the classical signs of HCTD, particularly seen in Marfan syndrome or Ehlers-Danlos syndrome, was strongly associated with sCeAD.

Second, studies investigating arterial wall biomechanical proprieties in sCeAD patients found altered arterial distensibility. Guillon and co-workers, in an ultrasound study focusing on common carotid artery diameter and diameter changes during the cardiac cycle showed that the diameter changes were more pronounced in sCeAD, reflecting less elastic properties of the arterial wall (31). Similarly, in another study Calvet and co-workers assessed the elastic properties of the carotid wall in sCeAD patients using noninvasive high-resolution echo tracking systems and showed that sCeAD patients had higher circumferential wall stress and stiffness (32). Besides, this study also showed that Young’s modulus, a direct measure of elasticity proprieties of the vessel, was significantly higher (indicating stiffer arterial wall), and associated with an up to the 8-fold increased risk of sCeAD. Lucas and co-workers assessed endothelial function and reactivity using high-resolution ultrasonography and showed that the brachial artery flow-mediated vasodilatation was significantly reduced in patients with sCeAD compared to controls, indicating impaired endothelial-dependent vasodilation (33).

Third, weakened arterial structure caused by connective aberrations may also predispose to vascular deformities, such as tortuosity. Higher arterial tortuosity was associated with connective tissue diseases, particularly Marfan syndrome and Loeys-Dietz syndrome (34, 35). In a study using magnetic resonance angiography Giossi and co-workers showed that tortuosity indexes in sCeAD patients were significantly higher compared to those of matched controls (36). These findings are indirectly supported by the observation that aortic dissection in patients affected by Marfan syndrome correlates with arterial tortuosity of the aorta (37).

Fourth, Brandt and co-workers showed that the ultrastructural connective tissue abnormalities in collagen and elastin fibrils of patients with sCeAD were related to disease recurrence (16). They also showed that these aberrations were more commonly seen in male patients but were not associated with age and vascular risk factors.

Fifth, indirect evidence in favor of the connective hypothesis also comes from studies investigating clinically detectable connective signs in sCeAD patients. Dittrich and co-workers assessed the clinical connective tissue phenotype of sCeAD patients using a standardized examination containing 25 clinical items that mainly included signs found in Marfan and Ehlers-Danlos syndromes (38). Presumably due to the fact that the sample size was relatively small to detect small clinical changes, they found no significant difference in clinically detectable connective tissue abnormalities in patients with sCeAD compared with the non-sCeAD group. Conversely, another more recent study with a larger sample size assessed clinically detectable connective tissue abnormalities with a more extensive standardized examination protocol including 68 items found in HCTD (11) and showed a higher prevalence of connective tissue abnormalities in sCeAD patients, further supporting the connective hypothesis. Interestingly, the low prevalence of these clinically detectable signs in patients with traumatic CeAD contrary to sCeAD further supports this hypothesis. Similarly to previous studies discussed in this review, these patients were not diagnosed with a definite HCTD, implicating that sCeAD may represent a multifactorial disease that is the result of the combination of triggering factors in settings of an underlying subclinical connective tissue disorder.