Incidence of acute type A aortic dissection in emergency departments

Due to the symptoms, patients with acute type A aortic dissection are first seen by the ambulance service and diagnosed at the emergency department. How often an aortic dissection occurs in an emergency department per year has been studied. The incidence in the emergency department may be used as a quality marker of differential diagnostics of acute chest pain. A multi-institutional retrospective study with the municipal Berlin hospital chain Vivantes and its Department of Pathology and the Charité - University Medicine Berlin was performed. From the Berlin Hospital Society, the annual numbers of publicly insured emergency patients were obtained. Between 2006 and 2016, 631 aortic dissections were identified. The total number of patients treated in the emergency departments (n = 12,790,577) was used to calculate the “emergency department incidence.” The autopsy data from six clinics allowed an estimate on how many acute type A aortic dissections remained undetected. Across all Berlin hospitals, the emergency department incidence of acute type A aortic dissection was 5.24 cases in 100,000 patients per year. In tertiary referral hospitals and, particularly, in university hospitals the respective incidences were markedly higher (6.7 and 12.4, respectively). Based on the autopsy results, about 50% of the acute type A aortic dissection may remain undetected, which would double the reported incidences. Among different hospital types the emergency department incidences of acute type A aortic dissection vary between 5.93/100,000 and 24.92/100,000. Aortic dissection; Incidence; Emergency Department; Epidemiology


Results
emergency department incidence und pathology data. With 12,734,308 ED patients from 2006 to 2016 in Berlin and 631 ATAADs, an ED incidence of 4.96/100,000 was calculated. During the observation period from 2006 to 2016, all these 631 clinical cases of ATAAD were transferred to and treated at the German Heart Center Berlin or the cardiac surgery department of Charité-University Medicine Berlin. Of these, 410 (65%) were men and 221 (35%) were women. The average age was 61.4 ± 13.8 years. The youngest patient was 23 and the oldest patient was 90 years old. The mean body mass index (BMI) was 27.2 ± 4.9 kg/m 2 and ranged from 16.3 to 55.56 kg/m 2 . In 425 (67.8%) cases, a pre-existing arterial hypertension was detected. Eighty-one (12.8%) patients had coronary artery disease. In 43 (6.8%) patients, type 2 diabetes mellitus was detected. Pre-existing aortic diseases such as aortic aneurysm or ectasia of the aorta were found in 86 (13.6%) of the patients. Fifteen (2.4%) patients had a Stanford type B dissection in their clinical history. One hundred ninety (30.1%) had a history of nicotine abuse, and 32 (5.1%) patients were alcohol addicted. Arteriosclerosis was found in 51 (8.1%) of the patients. A hereditary connective tissue disorder (Marfan or Ehlers-Danlos) was found in 18 (2.9%) patients. 66 (10.5%) patients were known to have hypothyroidism. There were only three (0.5%) cases of pregnancy and five (0.8%) cases of cocaine abuse.
The Charité and the Vivantes clinics play a major role in the care of critically ill patients. We assigned all patients to the respective clinics in Berlin. We divided them into three groups. The first group includes all Charité patients, the ED group patients from Vivantes clinics formed the second group. The third and last group includes all other EDs in Berlin. Detailed data including the distribution of patient characteristics among different types of hospitals are shown in Table 1. The Charité and the Vivantes clinics provided us with their own ED patient numbers. With our data, it was possible to calculate the exact incidence in the ED for the respective observation periods.
In Table 2 the cases of ATAAD, the population and the incidence are compared for the congruent period from 2010-2016. The first column contains all patients in the Berlin region.
The following columns compare the Charité clinics, the Vivantes clinics and all other hospitals in terms of cases, the emergency department incidence and the extrapolated data.The Charité has 131 ATAAD cases and an ED population of 1,050,994. This results in an incidence of 12.46/100,000. At the Vivantes hospitals, 153 cases of ATAAD were found among 2,112,610 ED patients resulting in an ED incidence of 7.24/100,000. The remaining Berlin hospitals had 5,195,208 ED patients and 154 ATAAD cases. The ED incidence was 2.96/100,000. ( Table 2). The information of the pathology department of six Vivantes clinics was used to estimate how many ATAADs www.nature.com/scientificreports www.nature.com/scientificreports/ went undetected. The ATAAD was not known prior to the autopsy. Compared with the data from the German Heart Institute and the Vivantes clinics, this suggests that only one of every two ATAAD cases was detected. The details are shown in Table 3. Extrapolating this assumption to all hospitals, the incidence of ATAAD in the EDs all over Berlin would increase to 10.48/100,000, to 14.48/100,000 at Vivantes, to 24.92/100,000 at the Charité, and to 5.93/100,000 at all other hospitals in Berlin ( Table 2).

Discussion
The literature review has shown that the data on the incidence of ATAAD vary substantially. The study characteristics are summarized in Table 4. It is important to provide an accurate estimation of the incidence and also to consider its development over the years. This may serve as a quality indicator for emergency services, statistical offices, public health offices, legal medicine, and other healthcare providers. The total population of the 14 studies included 1,733,937,352 person years. The indicated incidences of all population-based studies were very heterogeneous and ranged from 2.1 to 16.3/100,000 inhabitants 1-12 .
A retrospective study from Rogers et al. 2011 roughly estimated an ED incidence for acute aortic dissection in the United States. They assumed 10,000 cases of AAD annually and 100,000,000 ED visits in the same time period. This led to an ED incidence of 10/100,000 patients 20 .
Our study confirmed this assumption using exact numbers of ED patients and ATAAD cases.
Another study analyzed the incidence in clinical and forensic autopsies in the regions of Berlin and Brandenburg. The analysis provided 1150.5 cases of ATAAD per 100,000 autopsies 2 . In a retrospective study from Mollo et al. in 1983, autopsy reports were analyzed from 1932 to 1981. They could show a relative frequency of 1/168 or an incidence of 595/100,000 autopsies 21 . The incidence rates in the population-based studies were related to AADs without distinction between type A and type B. For this reason, we calculated the incidence for the ATAADs from the given data. A population-based incidence of 2.3/100,000 for all population-based studies was obtained. Due to the high preclinical mortality and missing forensic and clinical autopsy data, the number of unreported cases can safely be assumed to be very high. 26% of the patients who received surgical treatment due to a type A dissection died within the clinic. The mortality of the conservatively treated patients was 58% 13 . The ED incidence in the Berlin area varied over the years, but no sustained increase occurred. Across all 83 hospitals in Berlin, an ED incidence of 4.96/100,000 was calculated for the period from 2006 to 2016. The German Heart Center is responsible for the surgical care of nearly 90% of acute type A aortic dissections in the Berlin region. Looking separately at the largest hospital companies, the incidence at the Vivantes clinics was 7.24/100,000, and at the Charité hospitals it was 12.46/100,000 emergency patients. Hospitals with a higher level of care and a simultaneous "Chest Pain Unit" qualification are increasingly being approached by the ambulance  www.nature.com/scientificreports www.nature.com/scientificreports/ service for the symptom of chest pain. With greater expertise and better occupation in the ED, these institutions generate larger numbers of cases and have a higher incidence of ATAAD in the ED. The study showed higher prevalence with lower incidence of ATAAD in the Vivantes clinics compared to the Charité clinics. This is due to the different number of EDs (Vivantes n = 9, Charité n = 3). The autopsy results and these data led to the conclusion that probably only one in two ATAADs was diagnosed. Following this assumption for the time period from 2010 to 2016, the incidence for all Vivantes clinics increased from 7.24/100,000 to 14.48/100,000. Extrapolating this assumption, the incidence in the Charité increased from 12.46/100,000 to 24.92/100,000. There is only one study that assessed the incidence in EDs in the United States of America. It provides a rough estimateof 10 AAD cases of any type among 100,000 ED patients 20 . We calculated an ED incidence of 5.24 ATAAD/100,000 for Berlin. Under the assumption that only one out of two cases are actually detected, the incidence increases to 10.48/100,000, which is almost identical to the United States of America. This incidence is higher than the average extrapolated population-based incidence in Berlin. In relation to the Berlin population during the study period (37,843,130 person years) and 1262 extrapolated cases of ATAAD, the population-based incidence was 3.3/100,000. The examination of the section material shows a very high number of unreported cases. In all deceased patients undergoing autopsy, the ATAAD was not known prior to the autopsy. Whether or not ATAAD was causal for death is not known, but, considering the high in-hospital mortality of untreated ATAAD we can assume that it was the cause of death in the majority of cases 22 . To reduce this number, it would be necessary to take a closer look at the deceased patients of each hospital. A study by Modelmog et al. could show that the autopsy findings are not in line with the diagnosis given on the death certificate in 45% of the cases 23 . Burton et al. from 2003 pointed out that an autopsy can reveal important and unsuspected diagnoses. The discrepancy between the clinical diagnosis and the findings of the autopsy has decreased over time. But, it remains so high that a continuation of the autopsies is indicated 24 .Brinkmann et al. provides data on the development of clinicopathologic and forensic autopsy rates between 1994 and 1999. Especially in Berlin, the autopsy rate decreased by 3.9% within 5 years, from 15.3% in 1994 to 11.4% in 1999 25 .
The study periods vary between Charité (2010-2016), Vivantes and Berlin (2006)(2007)(2008)(2009)(2010)(2011)(2012)(2013)(2014)(2015)(2016). Unfortunately, it was not possible to gather data on the missing years to get equal study periods. The ED population for 2016 was missing and had to be estimated based on the evolution of patient numbers. As a result, the numbers could be under-or overestimated. The autopsy data was provided only from the Vivantes clinics. The overall autopsy rate was 5.95%. This is not enough to provide reliable numbers on undetected cases of ATAAD. Therefore, the extrapolation of these undetected numbers to all Berlin hospitals should be considered with care. For a more detailed assessment, autopsy data from all hospitals in Berlin and the Charité would be necessary. The new data on incidence in EDs should sensitize the hospitals and ambulance services to this acute disease. The goal in preclinical care should be the establishment of a standard operating procedure for the symptom complex of chest pain in which the aortic dissection detection risk score is firmly anchored. There, a special focus should be on patients who are delivered by the ambulance or the rescue service.  conclusion Our results underline a higher incidence in the EDs compared with the population-based incidence in the general population of Berlin. The ED incidence in the period from 2010 to 2016 varies between 2.96/100,000 and 12.46/100,000 among different types of hospitals. The population-based incidence for the entire Berlin population from 2010 to 2016 is 3.45/100,000. Specialized clinics show significantly higher case numbers. The ATAAD is a severe disease and it is one of the four differential diagnoses of acute chest pain. Lower than expected incidences may mean that ATAAD remains undetected in a considerable number of patients. The pathology department information showed, that only one in two cases of ATAAD is detected. Under this assumption, the ED incidences and the population-based incidence would be twice as large.

Material and Methods
Study design. The databases PubMed und Google Scholar were used to search for studies on the subject of incidence of aortic dissection. Incidences in the fields of population, ED, pathology autopsies, and forensic medicine were considered. "Acute aortic dissection type A incidence" and "acute aortic dissection at autopsy" were used as search terms. The search on Pubmed and Google Scholar yielded 23,056 studies. The abstracts were screened, and studies were excluded if they dealt with other topics, full-text was not available, or language was different than English or German. Based on these criteria, 23,020 studies were excluded. Twenty-two of the remaining 34 full-text studies fulfilled all criteria for eligibility and met all criteria. The bibliographies of the included studies were explored for further relevant studies. In an additional search, six more relevant studies were identified. Fourteen articles that deal with autopsy-, population-, forensic medicine incidence, and ED incidence of ATAAD are summarized in Table 4, which includes the following information: kind of incidence, author, date, study type, population size and incidence. The study from Kurz et al. was considered twice, for a population based incidence and a forensic medicine incidence of ATAAD 2 .

Data availability
The datasets used and analyzed during the current study are available from the corresponding author on reasonable request.