Aortic dissection (AD) is a life-threatening emergency with an overall mortality of 27.4% [1]. It is well established that the outcome of AD is associated with multiple factors, such as age, gender, type of AD, aortic diameter and presenting blood pressure [2‐4]. It has also been frequently proposed that individual ABO blood groups affect serum cholesterol concentration, inflammation and hemostatic status [5‐7], all of which could independently alter the course of AD [8‐10]. Furthermore, previous studies have associated ABO blood group with other cardiovascular diseases such as atherosclerosis and coronary heart disease [11], the former being a risk factor for and contributor to the pathogenesis of AD [12]. Nonetheless, the relationship between ABO blood group and AD prognosis has attracted little focus. In this study, we investigate the association of ABO blood group and the in-hospital mortality of AD.

Baseline characteristics of the study population are presented in Table 1 and quantitative data (age, SBP and DBP) normality are presented in Additional file 1: Table S1. Eight hundred seventy-seven AD patients were included from January 2015 to July 2019. Male gender, non-O blood group, TBAD, higher SBP or DBP, and receipt of surgery or stent-grafting were associated with significantly lower in-hospital mortality. The most common cause of in-hospital death was ADR, followed by arrhythmia and septic shock (Additional file 1: Table S2).

After being selected by the combination of univariate Cox proportional hazards models and a multivariate stepwise logistic regression model, non-O blood group, TBAD, higher SBP and surgical intervention remained associated with decreased in-hospital all-cause mortality (Additional file 1: Tables S3, S4). In the unadjusted Cox proportional hazards model, non-O blood group presented a significant 33.2% decrease in the risk of in-hospital mortality compared with blood group O (HR = 0.668, 95% CI 0.473–0.944). In Model 2 (adjusted for age and sex) and Model 3 (adjusted for AD type, SBP and surgery), the protective effect was still observed (HR = 0.662, 95% CI 0.468–0.935 and HR = 0.641 95% CI 0.453–0.906 respectively) (Table 2, Fig. 2). The same method was applied in analyzing the in-hospital mortality caused by ADR and showed that the non-O blood group was significantly associated with lower risk of ARD-related death irrespective of adjustment (Additional file 1: Tables S5–7).

For further analysis, patients were divided into 2 groups based on their type of AD (Tables 3, 4, Fig. 3). In the TAAD group, risk of in-hospital mortality in patients with non-O blood type was 34.3% lower than those with blood type O (HR = 0.657, 95% CI 0.440–0.975). After dividing these patients into surgery and non-surgery subgroups, the non-O blood groups remained associated with lower risk of early mortality (HR = 0.579, 95% CI 0.377–0.889), notably caused by ADR (Additional file 1: Tables S8, S9), in TAAD patients without surgery. However, in the TAAD surgery subgroup, different blood types presented no significant difference in in-hospital mortality risk (HR = 1.169, 95% CI 0.414–3.305). In the TBAD group, no association was found between ABO blood group and mortality of AD, regardless of the involvement of a stent-graft procedure.

Interestingly, when analyzing individual ABO blood groups, patients with blood type AB seemed to have the best prognosis (Fig. 4), followed closely by blood group B, A, and then O. However, statistically significant lower mortality was only observed in blood type B as compared with blood type O (HR = 0.575, 95% CI 0.361–0.916) (Table 5). The difference in mortality was neither significant between blood type AB and O (HR = 0.526, 95% CI 0.212–1.306), nor between blood type A and O (HR = 0.799, 95% CI 0.525–1.216).

The main finding of this study is that patients with a non-O blood type have considerably lower risk of early mortality than those with blood type O. In particular, this trend is exhibited by patients who were diagnosed with TAAD but did not receive standard surgical intervention. Despite the advances in surgical skills and improved prognosis of AD patients over the past two decades [14], AD remains a medical catastrophe, making it essential to identify patients who have a lower chance of survival and require additional attention. Previous studies have proposed several variables, including age, gender and certain medical conditions, as predictors of AD prognosis [15‐19]. In this study, we find male gender, TBAD, higher presenting SBP and DBP and surgical intervention are related to better outcome, basically in line with the results of previous research. In addition, we sought to establish a relationship between AD mortality and ABO blood group, the latter a readily accessible and yet potentially important clinical feature. To be noted, this relationship has also been investigated in two other studies. Nikola et al. designed a case–control study involving 115 patients with type III AD and revealed no significant difference in mortality between various blood groups [20]. A more recent study concerning the surgical prognosis of AD, by Nozohoor et al., included a larger population from The Nordic Consortium for Acute Type A Aortic Dissection database [21]. The authors also found no association between individual ABO blood types and the surgical outcome of AD, except that blood group A was related to poorer long-term prognosis. Although patients with blood type O have a lower level of serum von Willebrand factor (vWF) [22], and hence a stronger tendency to bleed during and after surgery, the authors suggested that the impact of other surgical complications more frequently seen in non-O blood group, such as thromboembolic events [23], outweigh that of hemorrhaging in the long run. While Nozohoor’s study is in many ways comparable to ours, the former was a multi-center study and the latter was a single-center one. More importantly, Nozohoor and colleagues were more concerned with the postoperative outcome of the subjects, whereas a major portion of our patients did not receive surgery or stent-grafts. It is in these non-surgically managed patients that we observed the opposite trend, with non-O blood group being related to lower in-hospital mortality of TAAD. Furthermore, we found non-O blood group also significantly decreases the risk of mortality caused by ADR, the major cause of death for AD patients [24, 25], in TAAD patients without surgery. However, how ABO blood group is related to AD mortality or ADR has not been investigated, and more research is needed to elucidate the mechanism.

Our findings in the non-surgery subgroup also reflect the deeper goal of the study. Although implicated, some of our patients failed to receive surgical intervention due to realistic issues, such as patient preference and their financial status, and it is of equal importance to note that many patients died before surgery was available, due in a large part to lack of sufficient blood stored for transfusion. This finding should drive management decision-making. As treatment strategies are individualized based on age, comorbidities, and factors that might discourage surgery, physicians should take into account other factors that could render their patients at increased risk of death without operation, such as ABO blood group in this case. Furthermore, the significance of blood unit preparation needs to be addressed. Since blood type O is associated with a higher early mortality rate, possibly due to risk of bleeding complications, sufficient storage and supply of O-type blood is essential, especially in areas populated by a higher percentage of people with blood type O, such as South America [26].

Limitations of the study need to be acknowledged. First, this is a single-center study, and our subjects may not be representative of all AD patients. In addition, we did not study how ABO blood group could affect the outcome of AD, and further research can be designed to investigate the molecular mechanism of this process.

To summarize, non-O blood group is associated with lower risk of in-hospital mortality for patients with TAAD who do not receive surgical intervention. More attention must be paid to blood type O TAAD patients for whom early surgical intervention and adequate supply of O-type blood should be considered. More research is needed to better understand the mechanism of relationship between blood groups and AD.