TAAs are more common in males than females, with a male-to-female ratio of 2–4:1 and an overall incidence of approximately 5–10 per 100,000 person-years [14]. In two of the largest population-based epidemiological studies from Sweden [15] and Canada [16], 36–39% of individuals with AADs were female. This is consistent with the 37% of women who presented with AADs in a recent International Registry of Acute Aortic Dissection (IRAD) study [17].

Regarding the later age of AAD presentation in women, there is growing clinical and experimental evidence that the increased risk of AAD in peripartum and postmenopausal women is caused at least in part by withdrawal of estrogen and/or progesterone [23,33]. However, the role of sex hormones in TAD is almost entirely unknown, in part due to fact that mouse studies have focused on Fbn1^C1041G/+ Marfan mice, which develop aneurysms but rarely dissect [34,35,36]. Estrogen is believed to mediate the major sex differences in the cardiovascular system and has shown beneficial effects in reducing hypertrophy and ischemia-reperfusion injury in mice [37,38]. Notably, a recent study found that estrogen replacement significantly attenuated angiotensin II (Ang II) plus β-aminopropionitrile (BAPN)-induced ascending aortic enlargement [39]. On the other hand, the role of progesterone is controversial. Acute administration of progesterone increases endothelial-specific nitric oxide synthase production and decreases blood pressure, which may prevent cardiovascular diseases [40,41]. However, prolonged exposure to high progesterone levels is linked to premature coronary artery disease in young males [42]. It remains unclear how or if female or male sex hormones modulate aortic wall homeostasis to affect dissection risk.

Pregnancy leads to hemodynamic changes and rapid hormone alterations and is a risk factor for TAD. Aortic dissection in pregnancy is extremely rare, occurring in 0.2% to 1% of all female AAD cases [18,43,44], with an annual incidence of 4–12 per million maternities [43,44,45,46,47]. Kamel et al. [43] identified 36 pregnancy-related AAD cases, defined as 6 months prior to and 3 months after delivery, and 9 nonpregnancy-related AAD cases during an equivalent 9-month period exactly 1 year later, from a total of 14,999 AAD patients in 4,933,697 women with 6,566,826 pregnancies. Based on these data, pregnancy was associated with a five-fold increased risk for AAD [43]. Hypertension further increases the risk of aortic complications by up to ~three-fold during pregnancy [43]. However, the most significant risk factor for pregnancy-associated AAD are syndromic heritable thoracic aortic diseases (HTAD) such as Marfan syndrome (MFS). In these patients, the absolute increased risk for AAD is about 1000 times higher than the general population [43]. Sixty percent of women with pregnancy-associated AADs have at least one clinically identifiable risk factor for AAD, including a genetic syndrome associated with TAD (primarily MFS), a family history of TAD, or a bicuspid aortic valve (BAV) [44,48,49,50].

Similar to other adverse cardiovascular events, such as acute myocardial infarction, acute pulmonary thromboembolism, and cerebrovascular accidents, AAD has chronobiological patterns [58,59,60]. In 1997, Gallerani et al. [61] analyzed hour of symptom onset from 67 cases and identified that AADs has a circadian pattern, with a primary peak of onset at around 10 a.m. and a secondary peak at around 8 p.m. In 1999, Manfredini et al. [62] studied monthly data from 85 AAD patients and demonstrated a circannual pattern with a peak of maximal occurrence in January/winter/cold season. Since then, increasing numbers of large-cohort studies confirmed these findings of AAD onset patterns from different locations of the northern hemisphere [58,59,60,63,64,65]. Despite the lack of AAD data from the southern hemisphere, two recent studies revealed similar seasonal pattern in cervical artery dissection in Australia and the latest study showed consistent increased risk of cervical artery dissection in cooler months [66,67]. These findings together indicate circannual and circadian causes are important in the pathogenesis of AAD. Studies did not show any differences of in-hospital adverse clinical events or mortality with different seasons or different time-of-day onset.

Approximately 20% of families with TAD exhibit an autosomal dominant inheritance pattern, indicating Mendelian inheritance of a pathogenic variant conferring a highly penetrant risk for TAD [11]. In 2018, 11 out of 53 candidate genes were designated as causal for HTAD [13]. Cascade screening of families who are found to have mutations in any of these 11 genes is clinically indicated to prevent premature deaths of affected relatives due to TAD. The genetic risks for TAD extend from highly rare penetrant variants that trigger disease in almost all individuals carrying the alteration to common and lower penetrant variants more commonly found in the general population that confer a lower risk for disease, which has been extensively discussed in previous reviews [11,87,88,89] and will not be presented in this paper.

BAV is the most common adult congenital heart malformation with an overall prevalence of 1% but is three times more frequent in males (1.5%) than in females (0.5%) [90]. The prevalence of BAV is also significantly higher in European populations than in populations of African ancestry [91]. BAV is primarily inherited as an autosomal dominant trait with incomplete penetrance and variable expressivity [92]. BAV inheritance is best explained by a complex genetic architecture involving many different interacting genes [93].

BAV is enriched 5–10-fold in rare syndromic forms of HTAD such as Loeys–Dietz syndrome that do confer a high risk for AAD [98]. This may explain why the prevalence of BAV is increased in AAD cohorts compared to the general population (3.2% in IRAD [99]). However, more than 90% of BAV cases occur as isolated non-syndromic lesions, and in this group, the overall risk of AAD appears to be very low (0.4% over 15 years), even with significant aortic dilation [100,101]. Current clinical guidelines reflect these observations by setting surgical thresholds for aneurysms in non-syndromic BAV cases as high as 5.5–6.0 cm [102,103].

The incidence of aortic dissection ranges from 3 to 6 cases per 100,000 person-years globally [2,10]; however, data from large-scale registries have demonstrated that the frequency of dissections reported, demographics, management and outcomes, are variable depending on the ancestral populations studied and geographic location of data collection.

Overall, multicenter registries incorporating data from patients in different geographic regions with diverse ancestries has enabled research into associations between TAD incidence and other demographic and clinical characteristics, and geographic location, ancestral origin and self-reported race of patients with TAD. However, the collection and reporting of race, ethnicity and ancestry data are not consistent across studies, and interpretation of their role in dissection risk and clinical outcomes requires further investigation of genetic factors, dietary habits, stress burden, drug use, and other social determinants of health to delineate new risk factors and improve TAD management.

Hypertension is the most common comorbidity condition in patients with AADs with prevalence from 45% to 100% in previous observational studies [2,17,118,119]. In a recent autopsy study, Huynh et al. [120] found that about 84% cases died of aortic dissection showed left ventricular hypertrophy, a marker of preexisting hypertension. The first prospective population-based study identified 67.3% of AAD patients diagnosed with hypertension. In this study, the prevalence of hypertension increased from 42.9% in the first 3 years to 85.7% in the last 3 years of the 10-year study due to increased blood pressure screening in primary care, suggesting a high risk of poorly controlled hypertension prior to AAD events at the beginning of the study [2,121]. Importantly, the study showed that only 67.3% of the patients were on antihypertensive medication during the 5 years prior to dissection, and the proportion analyses of all the blood pressure data showed that 61.9% of subsequent blood pressure readings were >140/90 mmHg despite the majority of patients being treated with combined antihypertensive therapy [2]. Moreover, during the 5 years before aortic events, premorbid systolic blood pressure was significantly higher in patients with type A AAD that were immediately fatal than in those who survived to admission (151.2 ± 19.3 vs. 137.9 ± 17.9 mmHg; p < 0.001) [2]. 

Dyslipidemia has been well documented to advance disease progression of abdominal aortic aneurysm (AAA) and rupture, and statins treatment could significantly reduce rupture risk in AAA. Studies have revealed that dyslipidemia is also a common comorbidity in patients with TAD [119,122,125], however, only two recent studies found that it is associated with aortic enlargement and AADs. In a prospective 20-year follow-up study, Landenhed et al. [122] found that lower apolipoprotein A1 levels were associated with aortic dissection. In another prospective study, Yiu et al. [125] found that hyperlipidemia (p = 0.0321) was positively correlated with the growth rate of aortic arch aneurysm. Hypercholesterolemia promotes atherosclerosis of the aorta, which is initiated with oxidative LDL-overloaded macrophages in the intima and buildup by continuous fatty deposition. There is increasing evidence showing that SMC phenotypic modulation driven by cholesterol is a major contributor to atherosclerosis [126,127,128,129,130]. This modulation of SMCs may also contribute to thoracic aortic disease, and anti-hyperlipidemic therapy may have a protective effect against TAD and rupture, in addition to abdominal aortic disease. These findings suggest that dyslipidemia may be a significant risk factor for TAD, and the inhibition of cholesterol synthesis with statins may be beneficial in preventing AAD incidence in TAD patients, but further prospective randomized controlled trials are needed.

Aortitis refers to inflammation of the aorta and can be divided into infectious and non-infectious categories. Infectious aortitis is caused by specific organisms, either from intraluminal endocarditic vegetations or external adjacent infectious process, most commonly Salmonella, Staphylococcus, Strptococcus, Treponema pallidum, fungal, and mycobacterial [133,134,135,136]. Because diagnosing infected tissue is rarely possible at TAD onset, uncontrolled sepsis causes about 21–44% in-hospital mortality in patients with infectious aortitis despite of combined antimicrobial and surgical interventions [137]. In contrast, syphilitic aortitis usually occurs decades after primary infection, and the aorta enlarges due to inflammatory responses and fibrosis and leads to aortic aneurysm and rupture but rarely dissection [138,139,140]. In two recent case reports, ascending aortas were involved by the syphilitic process in all patients [139,140], indicating a strong association between syphilis infection and TAD.

Despite significant heterogeneity in aortitis definition and manifestation over-lapping, several studies attempted to investigate the aortic complication rates between different aortitis subtypes. Miller et al. [149] found similar ascending aortic complication rates between CIA and LVV patients in an 83-month follow-up study. Similarly, Clifford et al. [151] reported outcomes from 196 patients (66% classified as CIA) with histologically proven aortitis following surgical repair, 44% CIA and 40% LVV patients developed new vascular lesions and required re-intervention. Importantly, the availability of high-resolution vascular imaging has improved the identification of aortitis. Fluorine-18-fluorodeoxyglucose positron emission tomography/computed tomography (18F-FDG PET/CT) has a role in diagnosing aortitis, showing an increase in FGD uptake by inflammatory cells in the aorta [152,153,154]. Ferfar et al. [155] followed up 353 patients with aortitis for 52 months and found that AADs occurred in 5%, 2%, and 10% of patients with GCA, TAK, and CIA, respectively. These data indicate that patients with aortitis are at risk for future aortic complications, particularly in those with CIA. Additionally, other non-infectious diseases may also increase the risk of aortic diseases, including ankylosing spondylitis, Behçet’s disease, Cogan’s syndrome, granulomatosis with polyangitis, IgG4-related disease, relapsing polychondritis, rheumatoid arthritis, sarcoidosis, and systemic lupus erythematosus [135]. The pathogenesis and triggers of AADs in these aortitis subtypes remain largely unknown and require further studies.

OSA is the most common sleep disorder, which provokes the remodeling of respiratory and cardiovascular systems due to intermittent hypoxia caused by frequent obstruction of the upper respiratory tract during sleep. Symptoms of OSA include snoring, apneic attack, and daytime hypersomnia. Accordingly, the frequency of apneas, hypopneas, oxygen desaturation, and intermittent hypoxia and re-oxygenation are sensitive indicators for diagnosing and assessing the severity of OSA. The prevalence of OSA has been estimated at 22% in men and 17% in women, and the prevalence has increased over the past two decades, partially due to increased rates of obesity [156,157]. Importantly, cohort studies including the Sleep Heart Health Study have consistently demonstrated that over 50% of individuals with OSA have hypertension [158,159,160], a major risk factor for cardiovascular disorders, including AADs. Since Sampol et al. [161] observed a significant correlation between OSA and TAD in 2003, evidence from retrospective studies and meta-analysis support OSA as an independent risk factor for aortic root dilation, type A and B AAD [162,163,164], in the general population and in patients with genetic disorders, including MFS and Loeys–Dietz syndrome [165,166,167,168]. A prospective study followed up 44 MFS patients for 24–36 months, five of fifteen patients with OSA had aortic root replacement or aortic rupture while none of the 29 patients without OSA had aortic event within the follow-up period, further analysis showed a significantly decreased event-free survival in patients with OSA compared to those without OSA (p = 0.012) [166]. Gaisl et al. [169,170] reviewed two prospective studies and found that the prevalence of OSA was significantly higher in patients with TAA (apnea-hypopnea index ≥ 5/h) compared to a 2-to-1 matched control population (63% vs. 47%; odds ratio 1.87, 95% CI 1.05–3.34; p = 0.03), and the odds ratio was 3.25 (95% CI 1.65–6.42; p < 0.001) in patients with moderate and severe OSA (apnea-hypopnea index ≥ 15/h). After follow up the TAA cohort for an average of 3 years, Gaisl et al. [169] found a strong positive association of aortic root and ascending aortic enlargement with OSA. However, further prospective, randomized match-control trials are required to evaluate the causal relationship between OSA and TAD.

Additionally, OSA is a treatable respiratory disorder through the use of continuous positive airway pressure (CPAP) therapy. Several case reports have shown beneficial effects of CPAP in TAD patients in slowing the aortic expansion [167,171,172]. However, there has been no evidence that CPAP therapy for OSA can reduce the incidence of aortic diseases.

OSA-induced pressure differences, including increased intra-aortic blood pressure and decreased negative intrathoracic air pressure, have been widely considered mechanical stressors on the aorta. However, no significant difference was found when comparing the location of entry in patients with OSA between thoracic and abdominal aortic dissection subgroups [173], suggesting that mechanisms other than negative intrathoracic pressure involved in the pathogenesis of TAD. It is believed that intermittent hypoxia and re-oxygenation is the potential underline contributor of systemic vascular remodeling caused by OSA, leading to sympathetic nervous system activation and subsequent hypertension or increased systemic oxidative stress [174]. Interestingly, intermittent hypoxia and re-oxygenation induction significantly augmented aortic rupture death in BAPN plus Ang II-induced aortic dissection mouse model [175]. Hif-1α, a major transcriptional factor in response to hypoxia, was significantly upregulated and activated in the triple-treatment group when compared with BAPN alone or BAPN + Ang II group, which was consistent with exaggerated reactive oxygen species production both in the aortic tissue and explanted SMCs [175]. Additionally, administration of a specific HIF-1α inhibitor, KC7F2, delayed the aortic event and significantly decreased the mortality due to aortic rupture [175].

Taken together, OSA has been demonstrated as an independent risk factor for TAD in the general population, and CPAP should be used to treat OSA, even in the absence of a controlled clinical trial.

Fluoroquinolone (FQ) is one of the most commonly prescribed antibiotic classes mainly due to the broad spectrum and tolerance for long-term administration [176]. The common side effects of FQ are nausea, vomiting, peripheral neuropathy, dysglycemia, and arrhythmias, with more serious collagen-disruption associated complications including tendon rupture and retinal detachment [177,178,179]. In 2015, two independent observational studies [180,181] found that FQ was associated with an approximately 2- to 3-fold increased risk of TAD, which raised safety concerns about FQ administration. Additionally, two propensity score-matched studies provided further evidence that use of FQ within 60 days was associated with high risk of aortopathy [182,183].

To answer the question whether FQs are associated with aortic aneurysm and dissection in patients with specific infections or pre-existing aortic defect, Chen et al. [189] performed a retrospective study on FQ or amoxicillin exposure to patients with existing aortic disease and found that FQ was associated with a higher risk of aortic-related events and death. Consistent with this clinical finding, a preclinical study using high-fat diet treatment plus Ang II infusion-induced TAD mouse model found that ciprofloxacin exposure significantly increased the aortic incidences and augmented aortic wall disruption [190]. In advance to the two studies [184,185] raising concern about confounding in previous findings, Newton et al. [191] looked at all FQ-related indications in a population-based study and found that FQ was associated with about 20% increase of risk of AAA and iliac artery aneurysm, and all adults were more likely to receive intervention. Further stratified analyses found that adults of 35 years or older were at increased risk of arterial aneurysm formation [191]. Since only outpatients were studied, whether in-hospital patients with more severe infections will show a higher risk of aortic disease remains to be assessed [191]. In addition to these adult-only studies, Yu et al. [192] evaluated the safety of using FQ in 0–18 years old population in Taiwan and found that FQ exposure did not increase risk of collagen-associated adverse effects, including tendons rupture, retinal detachments, gastrointestinal tract perforation, aortic aneurysm or dissection.

Noting that patients with genetic variants were either completely excluded or presenting limited proportion in previous studies. A recently preclinical study found that ciprofloxacin accelerates aortic enlargement and promotes dissection and rupture in the Fbn1^C1041G/+ mouse model [193]. Taken together, in 2018 and 2019, the United States Food and Drug Administration and the European Medicines Agency requested that manufactures update FQ safety information related to the increased risk for TAD and its complications, and warned that the use of these drugs in patients at risk for aortic complications in clinic, including MFS and other related conditions, respectively [194].

Finally, several in vivo and in vitro studies have consistently found that FQ exposure regulated extracellular matrix remodeling-related enzyme expression or activity, including increasing matrix metalloproteinases (MMPs), decreasing tissue inhibitors of MMPs and lysyl oxidase [190,195,196]. However, knowledge gaps remain exist between these observations and the direct-action mechanism of FQ, which mainly targets two essential bacterial type II topoisomerase enzymes, DNA gyrase and DNA topoisomerase IV [197]. Interestingly, early studies found imbalanced cellular calcium homeostasis was involved in pharmacological effects of FQ [198,199]. Noting that calcium-channel blockers are effective antihypertensive drugs and are frequently administrated as an alternative to β-blockers; however, the use of calcium-channel blockers was associated with increased 30-day mortality after acute or elective abdominal or thoracoabdominal aortic aneurysm surgery [200]. When compared to other antihypertensive agents, Doyle et al. [201] found that calcium-channel blockers significantly increased the risk of AAD incidence in MFS patients, and the need for aortic surgery in patients with various forms of HTAD beyond MFS. It remains unclear if intracellular calcium imbalance-mediated cellular energy metabolism and calcium homeostasis are involved in FQ-associated disruption of aortic integrity.

Insufflation of cocaine powder is the most widely used method due to its longer duration time after each use, and cocaine users have shown about six-fold higher risk of all-reason mortality when compared to the general population [202]. Since 1987 [203], several case reports and retrospective studies have linked cocaine with AAD [204,205].

The common underline mechanism of cocaine-associated TAD and other cardiovascular diseases, such as coronary artery infarction, hemorrhagic or ischemic stroke, and atherosclerosis is cocaine-mediated sympathomimetic effect. Cocaine activates central sympathetic outflow directly [217] and blocks reuptake and breakdown of norepinephrine and dopamine at the presynaptic clefts, leading to constitutive activation of postsynaptic receptors in the peripheral sympathetic nerve [218,219]. Additionally, peripheral sympathetic nerve activation may stimulate adrenal release of catecholamines, which further activate the sympathetic nerve systemically via an endocrine effect. The excessive activated sympathetic nerve system and release of catecholamines mainly act on cardiomyocyte and SMC mainly via the β and α adrenergic receptors, respectively [217]. Stimulation of these receptors will cause the following effects related to TAD: (1) A rapid elevation of blood pressure; (2) increases of heart rate and stroke volume, which accelerate the shear stress over the aortic wall and cause intimal injury; and (3) SMC hypercontraction leads to vasospasm which may initiate thrombus formation in the microcirculation system, such as the vaso vasorum supplying blood to the aorta [220]. Importantly, increased oxidative stress and mitochondria dysfunction may trigger cardiotoxicity, and a similar response may occur in the hypercontractile SMCs [221,222]. Additionally, cocaine-induced endothelial injury characterized as decreased vasodilator nitric oxide in ex vivo tissues, and increased vasoconstrictor endothelin-1, prothrombotic factor, and even circulating endothelial cells in cocaine-use patients [223,224], suggesting a significant role of endothelial dysfunction in cocaine-caused cardiovascular incidents.

Several acquired factors have been reported to be associated with TAD, including: (1) intrinsic factor, weightlifting [225,226,227]; (2) extrinsic factors, (i) trauma [228]; (ii) iatrogenesis [229,230]; and (iii) misdiagnosis [9]. The latter two extrinsic factors are relevant to medical knowledge and skill trainings. Nevertheless, the awareness of these acquired TAD risk factors will be beneficial for disease management as well as lowering the incidence rate.

Diabetes is a well-established risk factor for coronary and cerebrovascular diseases. Surprisingly, the prevalence of diabetes with AAA ranged from 6 to 14% which is significantly lower than 17–36% in those without AAA [231], and diabetes was found independently associated with reduced AAA growth in a 3-year follow-up study [232]. In 2012, Prakash et al. first found an inverse association between diabetes and TAD hospitalization rate [233]. Similarly, Takagi et al. performed a meta-analysis which recruited 11 studies with a total of 47,827 patients with TAD and found that diabetes presented in 2.3–15.8% of TAD patients while the range was 8.9–35.9% in control cohorts (OR 0.43; 95% CI, 0.31–0.59; p < 0.00001) [234].

The underlying mechanism of hyperglycemia-associated beneficial effects in TAD is not fully understood. Given the fact that ACTA2 variants predispose to both TAD and early onset stroke and coronary artery disease [240], suggesting that SMCs are potentially the predominant cell type that involved in both large elastic artery enlargement and small muscular artery occlusion; however, it is unclear but interesting if a common signaling pathway is involved. In vitro study has shown significantly increased capacities of proliferation, adhesion, and migration in SMCs from patients with diabetes compared to nondiabetic SMCs, which is believed to promote plaque formation and the major cause of small artery occlusion [241]. Interestingly, our unpublished data show that SMC-specific Pdgfrb deficiency augments aortic root enlargement in mice with Acta2^−/− background. These results indicate a possible protective role of SMC proliferation in inhibiting aortic dilation under hyperglycemia condition. Moreover, hyperglycemia enhances collagen glycation and networking which counteracts proteolysis driven by MMPs, while glycated type I collagen further decreases MMPs and interleukin-6 generation in activated monocytes [232] which are another major type of cells that contribute to extracellular matrix degradation and remodeling involving in both atherosclerotic plaque and TAD progression.

Increasing aortic diameter is a demonstrated risk factor for TAD, easy assessment of this parameter by various imaging methods concludes a diameter of 5.0–5.5 cm as an inflection point for elective surgical repair based on predictive risk analyses of dissection, rupture, and death [1]. However, accumulating evidence shows that many patients with acute TAD have diameters <5.5 cm [6], indicating an urgent need of lowering this prophylactic indicator for surgical treatment of TAD patients without connective tissue disorders or aortitis.

Nevertheless, the above IRAD study also found that mortality was not associated with aortic diameter and did not differ significantly across the 2 to 10 cm subgroups, leading to the difficulty of setting up a single threshold aortic diameter for predicting dissection, rupture, and death risks in all patients [6]. It is notable that undiagnosed patients with inherited TAD and aortitis in the entire population make it even more difficult to set up a precise cutoff point as a surgical indicator because these patients have high risks of dissection and rupture at a small or even normal aortic size.

In addition to circumferential dilatation, the longitudinal enlargement of proximal aorta has been studied and considered a more reliable risk predictor for aortic dissection due to the fact that aortic diameter enlarges by 16.9% to 31.9%, while the aortic length increases by only 2.7% to 5.4% when dissection occurs [242,243,244]. Despite continuous elongation of the ascending aorta with aging [245,246,247,248], Krüger et al. [246,249,250] found that ascending aortic length was significantly increased in both dissected and pre-dissection aortas when compared with healthy controls, and their studies proposed a score including both ascending aortic diameter (4.5 to 5.4 cm) and length (12 cm) as indicators for preventative surgical repair.

Aortic curvature and angulation usually present with ascending aortic elongation due to limited space in the superior mediastinum. In a recent study, Della Corte et al. [252] confirmed aortic elongation in aneurysmal and dissected patients, they also found that the angle between the ascending axis and arch axis was significantly narrower than that in the control aortas and suggested that an ascending-arch angle <130° appears to be a highly sensitive independent predictor of aortic dissection. However, the specificity of the 130°-angle is low, and it requires further study to analyze the pre-dissection data to validate its clinical value.