This study reviews the published English language literature on the incidence of acute AD, with a focus on the epidemiology and treatment of acute TBAD. Hospital episode statistics (HES) data were available from six countries (UK, Sweden, Germany, USA, mainland China and Taiwan). Four population based epidemiological studies (USA, Iceland, UK and Sweden) and two large registries (IRAD and Sino-RAD) have published data [24
]. There are a number of smaller observational studies of varying population size and quality. Several of the studies report TBAD treatment with no data on incidence. Many studies reported a composite figure for both TAAD and TBAD. Table 1
provides a summary of the study findings from the selected papers.
The range of reported acute AD incidence was 2 to 15 per 100,000 person years. The reported incidence for acute TBAD was 0.5 to 6.3 per 100,000 person years [1
]. In IRAD, the incidence of TBAD is consistently one third that of TAAD [24
]. A ratio of 2.6:1 between TAAD and TBAD was observed in studies by Clouse et al. (Olstead County, USA) and Acosta and Gottsäter (Malmö, Sweden) [20
A significant study limitation is that the type of aortic dissection, ascending or descending, is not captured in ICD codes. This means that for medically managed patients, it not possible to separate TAAD from TBAD. In operated patients the inability of procedural codes to differentiate the urgency of treatment (“emergent”, “urgent” or “elective”) may also lead to incorrect allocation of treatment type.
Population based prospective epidemiological studies should provide the best data on AD incidence. However, the absolute numbers of TBAD cases identified are small in the four reported studies; OxVasc (n
= 15), Olmstead County (n
= 6), Icelandic population (n
= 52) and Malmö Diet and Cancer study (n
= 30) [1
]. The variation in acute AD incidence between populations is highly influenced by age, as shown by Howard et al. in the OxVasc data, as shown in Figure 6
]. Consistently across studies, older age is associated with a higher AD incidence. In the study by Landenhead et al., the reported acute AD incidence in the Malmö population—with a median age of 70 years—is 15 per 100,000 person years (6.3 per 100,000 person years for acute TBAD) [23
]. Acosta and Gottsäter report a higher incidence of 26 per 100,000 person years in this same Malmö population [28
]. Mody et al. used US Medicare data to give an acute AD hospitalisation rate of 10 per 100,000 person years in a cohort of over 30,000 people with median age of 77 years [10
]. Howard et al. and Olsson et al. have published the relative risk for AD by age group [1
A gender difference, a higher TAAD prevalence in males, was observed in a large German dataset for the open repair of acute TAAD (GERAADA) [39
]. In the studies identified here, the proportion of females with acute AD varied from 19% to 52%. The lowest female proportions are reported in mainland China (19–28%) [13
]. Women were observed to present with acute AD at an older age than men; the mean age difference was 10 years in the OxVasc study (United Kingdom) and 5 years in the study by Smedberg et al. (Sweden) [1
]. There was also a higher proportion of out of hospital deaths observed in women in both these studies.
Out of hospital deaths are a potential cause of the under-reporting of acute AD incidence. In the study by Dias et al. (São Paulo, Brazil), 68% of AD cases were identified as out of hospital deaths [16
]. The proportion of out of hospital deaths reported in other studies is lower; Olsson et al.—22%, Melvinsdottir et al.—21%, Acosta and Gottsäter TAAD—78%/43% and TBAD—21%/22%, Smedberg et al.—29% and Mészáros et al.—21% [8
]. The proportion of TBAD deaths is less than TAAD deaths, but remains significant, especially if the calculated AD incidence figure is extrapolated to be applied to a larger population. Ascertainment of pre-hospital deaths is becoming more difficult, as in many countries post-mortem examinations are now rarely performed. Takeuchi et al. have suggested the use of CT to make the diagnosis of acute AD following an out of hospital sudden death [40
Disease specific registry data and the observational studies are inevitably subject to selection bias, typically with more complete entry and more aggressive surgical treatment in large specialist units. This may explain the high treatment rates with TEVAR in the studies by Li and Xiong from mainland China [33
]. Bottle et al. have reported wide variation in UK aortic management, between 7.6% and 31.5% intervention rates, across different regions [2
]. This variation may be explained in part by the absence of formal aortic dissection pathways, resulting in patients with TBAD being managed in different ways by a cardiologists, vascular surgeons and cardiac surgeons.
INSTEAD provided limited level one (randomized trial) evidence that thoracic aortic stenting in the subacute phase (4–12 weeks) is safe and reduces the risk of late aneurysmal degeneration [5
]. TEVAR in this trial was associated with increased early risk of retrograde aortic dissection, stroke (n
= 1) and spinal cord ischaemia (n
= 2). In a meta-analysis by Chen et al., the relative risk of retrograde dissection following TEVAR is 5.33 times higher than when treating a degenerative aneurysm (95% CI 2.70–10.51) [41
]. In a meta-analysis by Hossack et al., the risk of stroke was increased by 50% (OR 1.56, 95% CI 1.17–2.08, p
= 0.002) [42
]. The ADSORB study showed the benefit of TEVAR in reducing the rate of aortic dilatation after acute AD. [4
] Hossack et al.’s meta-analysis supports intervention with TEVAR for acute AD as both late all-cause mortality (HR 1.54, 95% CI 1.27–1.86, p
< 0.001) and aorta related mortality (HR 2.71, 95% CI 1.49–4.94, p
= 0.001) improved with TEVAR, in addition to the best medical therapy [42
]. Follow up has been short in most studies and overall the evidence base for TEVAR in acute AD is weak. It is therefore striking that the rate of TEVAR is increasing fast, with 76% of patients receiving TEVAR in one study [33
]. The more consistent range of TEVAR usage in acute AD is 15%–30% [1
Whilst aortopathy is an important cause of acute AD in younger patients, the absolute number of people affected is small. Many of the studies presented excluded people aged < 50 years. AD incidence is linked to rates of smoking, hypertension and other atherosclerotic risk factors. In mainland China, prevalence of hypertension is high, rates of hypertension treatment are low, and 2015 WHO data report smoking rates of 47.6%, as compared to 19% in the USA [43
]. Smoking demographics and access to healthcare, for instance for the monitoring and control of hypertension, may be important influencers of AD incidence between countries and over time. A meta-analysis to investigate these risk factors would require new studies with standardised end points.
Coding errors and omissions are well reported in routinely collected hospital data [45
]. There is also the risk of over counting; patients with AD are frequently readmitted, for instance for “subacute” endovascular intervention, and can easily be counted twice in hospital episode datasets. Some studies look back 2 years for previous coding of an AD diagnosis. For a chronic disease this time period may not be sufficient to discriminate acute and chronic dissection. The limitations of HES data have been widely discussed elsewhere and this discussion will not be repeated here [46
]. Specifically for AD, a study by Banerjee et al. investigating acute AD in pregnancy in the UK, recently demonstrated the difficulties of making comparisons between data sets [48
]. In this study the annual incidence of acute AD between 2003–2011 was reported as 1.23 per 100,000 maternities in NHS hospital episode statistics (HES), but as 0.80 per 100,000 maternities in the UK Obstetric Surveillance System (UKOSS), a national system to record rare disorders in pregnancy.
Other than when the method of data collection changes, hospital episode statistics can provide an accurate indication of trends over time. Whilst Mody et al. showed no increase in US Medicare admissions for acute AD for the years 2000–2011 [10
], consistently across other studies, hospitalisation for acute AD has been shown to be increasing worldwide. The actual incidence of AD is less certain, as recognised in the 2015 European Society of Cardiology (ESC) guidelines [49
]. Acute AD incidence was reported to be increasing between 1987 to 2013 in three studies [8
]. However, more recent studies suggest incidences to be stabilising, or even falling [18
]. A falling incidence would be consistent with data showing that in many countries cardiovascular risk factors are being better addressed.
In conclusion, there remains uncertainty regarding the incidence of acute TBAD. Changing incidence could be attributed to changes in population demographics and changing exposure to cardio-vascular risk factors. It equally may reflect changes in diagnosis and coding of cases (case ascertainment). The treatment for acute TBAD remains predominantly medical, with control of hypertension and serial cross-sectional imaging. There is increasing use of thoracic aortic stent grafts worldwide, despite a limited evidence base for this change of practice. For future studies, specific ICD-10 codes of acute, chronic, ascending and descending aortic dissection would greatly improve the ability to compare data on acute TBAD incidence and treatment between countries. Similarly, TEVAR codes should include whether the procedure is urgent or elective, as is the case for open aortic replacement.