Novel Insights of Aortic Dissections

Aortic dissections (AD) are a relatively uncommon but disastrous aortic condition. They occur when a tear in the innermost layer of theaortic wall allows blood to flow between the layers of the aortic wall. Depending on the location of the tear, ADs are classified as eitherType A or Type B. This classification is not only important regarding the prognosis for the afflicted patient, but even more for theimmediate treatment a patient should receive. Type A dissections require acute surgical repair because of the inherent risk ofcompromising perfusion of the brain. Type B dissections, however, can be medically managed, if they are not complicated by intolerablepain, or affected organ perfusion. Though ADs are well understood, several paucities surrounding the epidemiology and treatmentpersists.

The aims of this thesis were: i) to examine the accuracy of the Danish National Patient Register (DNPR) regarding the ICD-10 diagnostic codes for ADs, ii) to present population-based incidence and mortality rates for AD, as well as comorbidities related to ADs, iii) to report long-term outcomes after Thoracic endovascular aortic repair throughout a 20-year period at Odense University Hospital (OUH), and iv) to evaluate the potential mitigating effect of BB use on the risk of AD development.

In study I, patients who had been admitted under a diagnosis of AD during the period from 1996-2016 had their medical records
reviewed by vascular surgeons and the diagnosis validated if present. These validated diagnoses were then merged with register data from the DNPR. Then we computed positive predictive values (PPV) for the registered diagnostic codes in the DNPR stratified by four time periods: 1996-2000, 2001-2005, 2006-2010 and 2011-2016. We found that the PPV for the specific aortic diagnostic codes were high at 90.3% and 88.1% for Type A and B, respectively, for the entire study period, whilst the unspecific diagnostic code had a lower PPV at 63.5%. We found that the pooled PPV improved significantly over time during the study period.

Study II was based on the same validated AD cases as Study I, and the same period. These cases were merged with register data from the DNPR to enrich our cohort with comorbidities, potential death, and migration dates. Additionally, a matched cohort of hypertensive Danish citizens was used as a control group. Kaplan-Meier plots displayed the differences in mortality between Type A and B ADs and revealed that both types had a comparable 5-year mortality. However, when omitting 30-day deaths from the 5-year survival analyses, Type B cases had a significantly lower mortality rate than Type A. Additionally, the study examined incidence rates regarding AD and found that the overall incidence increased significantly over time and was 5.4/100,000 persons years at the end of the study period.

In study III, 256 patients treated with thoracic endovascular aortic repair (TEVAR) were included during a 20-year period at OUH from 1999-2019. Patients were stratified as either acute aortic syndrome (AAS), chronic condition, or traumatic rupture. Patients were followed with computed tomography scans three months post intervention. Hereafter, patients were followed with annually CT scans. AAS patients had significantly higher 30-day mortality compared to both chronic and traumatic patients. Reintervention was seen in 29 (11.7%) patients. The reintervention rate was consistent throughout the study period, though most common within the first five years.

Study IV was a nested case-control study with the source population being all Danish citizens in the period from 1996-2016 suffering from hypertension. Cases were validated AD Type A or B, and controls were matched 1:10 on age, sex and their defined daily dosage (DDD) history of antihypertensive drugs. After matching, 1657 cases remained with corresponding 10 controls per case. Crude OR between never and ever users of BBs was 2.46 (95% CI: 2.21-2.74), adjusted OR was 2.14 (95% CI: 1.90-2.40). Dose response analysis showed a non-significant trend toward increasing AD risk OR 1.000035 (95% CI: 0.99-1.00).

In summary, we found that when delving into the data recorded in the DNPR, the most reliable diagnostic codes were the specific ICD10 coding for either Type A or B, whereas the unspecific diagnostic codes are relatively unreliable and should be avoided when using register data on AD for research. Concurrently, our findings show that incidence rates for ADs have been increasing over the past two decades, which has been reported in other studies, but never on a population-based scale of validated cases of AD. We found that both AD type A and B cases had similar mortality rates. However, when we omitted the cases that died within 30-days of diagnosis, type A cases had a significantly better mortality, this might be explained by Type A cases having the highest mortality rate in the acute phase, but after intervention, their risk is reduced. On the other hand, risk of death from AD Type B continues after diagnosis because relatively
few are surgically repaired compared to Type A. This could explain their higher mortality rates. Additionally, it could be considered that patients with Type B are most compliant with their antihypertensive treatment in the early period from the diagnosis, and therefore increase their mortality when compliance decreases. We observed TEVAR mortality rates which were comparable to other reported rates. Endoleak was seen mostly in patients treated for aortic aneurysms. Furthermore, endoleaks were observed throughout the whole study period, indicating that lifelong follow-up is needed for patients undergoing TEVAR intervention. Though not statistically significant, we demonstrated that BBs show a slight dose response tendency towards increased risk of AD. Contrary to our hypothesis, we found that the use of BBs increased risk of AD with an adjusted OR of 2.14 comparing ever and never users. This result is surprising as BBs are thought to protect against development of ADs and are part of the primary antihypertensive management in the acute phase according to international guidelines.

In conclusion, this Ph.D. thesis has evaluated the validity of data from the DNPR for research regarding ADs. Furthermore, the validated database of ADs has provided the first of its kind population-based incidence rates for AD, which we found had been increasing significantly over a twenty-one-year period. Our single center experiences with TEVAR intervention showed that while TEVAR is an intervention form that can be used for a variety of aortic conditions, patients will have to undergo regular scans to identify potential endoleaks. Surprisingly, our study examining the potential mitigating effect from BBs on risk of AD, found the opposite to be the case: BBs may increase the risk of dissection.