Predictive Value of D-dimer on Mortality in Patients with Acute Aortic Dissection

Objective To clarify the relation between D-dimer and in-hospital mortality of acute aortic dissection, a meta-analysis was performed by summarizing all relevant studies. Methods All related studies were retrieved and identied in PubMed and Embase databases. Precise data was extracted from standard articles, such as sample size, odds ratio, and 95% condence interval. Then pooled odds ratio (OR) accompanying 95% condence interval (CI) were calculated using random model. Study heterogeneity examined by Q text and I 2 statistic. Sensitivity analysis was performed to assess the stability of the results. Publication bias was assessed by Egger’s test. Ten studies (1954 patients) that met the inclusion criteria were included in this review. The results suggested a link between D-dimer and in-hospital mortality of acute aortic dissection (OR=1.17 95%CI=1.08-1.27). With higher of cutoff value of D-dimer, the closer for the mortality of AAD, with ORs (95% CIs) ranging from 1.13(1.09–1.16) to 4.12 (1.56–10.93). The relationship was also found in six Type A AAD studies without heterogeneity (OR=1.13 95%CI=1.08-1.18). According to sensitivity analysis, the link was stable after exclusion of one study at a time. Publication bias was nd among studies (P=0.02). The result of this meta-analysis indicated that D-dimer maybe a predictor in-hospital mortality of acute aortic dissection patients. What’s more, the higher of cutoff value of D-dimer, the stronger for the predictive ability.


Introduction
Acute aortic dissection (AAD) is a life-threatening disease requiring an early diagnosis and effective treatment to prevent from death. As estimated, the annual incidence of aortic dissection was about 3/100000 [1]. The in-hospital mortality rate of type A AAD and type B AAD were 24.4% and 10.7% [2], which remains a paramount concern for clinicians even after successful surgery. However, AAD inpatients always have unexpected adverse events, and transferred to a danger situation in several minutes, thus easily delaying the rescue opportunity. Given the atypical presentations, and time dependent mortality, sensitive and quicker predictive tools to identify patients at increased risk of death are paramount to take rst-aid measures for clinicians.
Imaging equipment, such as computed tomography angiography (CTA), trans-esophageal echocardiography (TEE), or magnetic resonance angiography (MRA) is recognized as the powerful weapon for AAD diagnosis and treatment [3] . However, the use of these devices is based on typical symptoms, doctor's experience and high suspicious patients. As a result, high risk patients without atypical symptoms might be neglected. Besides, these imaging devices are invasive and expensive. The above severe condition indicated an insu cient medical supply in helping to identify patients in danger at an early time-point.
Biomarkers, such as D-dimer, CRP, cardiac troponin T and so on, were widely studied to predict in-hospital mortality of AAD [4,5] . D-dimer, a degradation product of cross-linked brin, is reported highly elevated in AAD patients [6] . Its role as a "rule out" diagnosis biomarker has been widely studied in previous study [7] . The European Society of Cardiology guidelines introduced the use of d-dimer as a secondary lter in patients with suspect aortic dissection [8] . In recent years, it was proposed to use in prediction mortality of AAD patients [9] . However, heterogeneous results were shown for the relationship between D-dimer and in-hospital mortality of AAD. It should be noted that these studies both conducted with small sample size and conservative conclusions. We conducted this meta-analysis to summarize all the related articles to clarify the predict value of D-dimer for in-hospital mortality of AAD patients with a relatively large sample size.

Materials And Methods
This meta-analysis was performed in accordance with the standards set forth by the statement from the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) and the guidelines for performing a Meta-analysis Of Observational Studies in Epidemiology (MOOSE) [10][11] .

Literature and search strategy
Combinations and derivatives of D-dimer, brin, fragment, acute aortic dissection were used as terms for retrieving in PubMed and Embase. In addition, the references of included articles were also retrieved. The literature search was limited to English language. If more than one article was published on the same cohort, only the recently published study was included. The literature search is updated till Dec.31, 2019.

Inclusion criteria and data extraction
Studies included in the system review met the following inclusion criteria: (1) evaluation of the predict value of D-dimer for inhospital mortality of AAD patients; (2) written in English; (3) reporting ORs/RRs/HRs with 95% con dence intervals (CIs) or relevant data to calculate them. Case report, system review, conference abstracts and other researches were excluded based on the guidelines. Relevant data were extracted from the quali ed articles using a standardized form. Two authors independently assessed the articles for compliance with the inclusion/exclusion criteria and entered the data in a prede ned database. Discrepancy was resolved through discussion and repetitive analysis.

Statistical analysis
Random effects model was chosen to calculate pooled ORs/RRs with 95% CIs for D-dimer. Heterogeneity was assessed by the Q-test and the I 2 statistic. The signi cance for the Q-test was de ned as P < 0.10. The I 2 statistic represents the amount of total variation attributed to heterogeneity. Low, moderate and high degrees of heterogeneity correspond to I 2 values of 25%, 50% and 75%, respectively. Sensitivity analysis after exclusion of each study at a time was performed to assess the stability of the results. Publication bias was assessed by funnel plot, and Egger's regression test was applied (P < 0.05 was considered statistically signi cant). Statistical analysis was conducted using STATA version 11 (Stata Corp LP, College Station, TX, USA).

Search strategy of included studies
A search owchart of relevant studies presented as Fig. 1. A total of 162 and 482 records were identi ed respectively in PubMed and Embase for potentially screening of title and abstracts. Remarkably, 248 articles were duplicated. 330 were excluded after screening of title and abstracts. 42 were not related to the topic after reading. 5 were systematic review. As a result, 19 studies were considered for inclusion. After reading, 4 were excluded due to Chinese writing, 1 for duplicated samples and 4 didn't provide relative data independently. Finally, 10 studies (1954 patients) that met the inclusion criteria were included in this review. A majority of the included patients were men. Six were from China. The D-dimer was measured once admission or within 48 h before any surgical procedure. Adjustment OR and 95%CI were extracted in the included studies. Detailed characteristics of included studies were displayed in Table 1-1 and 1-2. Characteristics of studies included in the meta-analysis of the association between D-dimer and in-hospital mortality of AAD  Morphologic class, extent of dissection, and false lumen status, and differences between patient groups NA represent there is not relative data in studies.

Meta-analysis Results
The results suggested that D-dimer was associated with AAD mortality (OR = 1.17 95%CI = 1.08-1.27) (Fig. 2). According to the value of D-dimer, the included studies were divided into three groups. Group one included three studies with a D-dimer cutoff > 5.0 ug/mL. The forest map suggested a link between D-dimer and in-hospital mortality of AAD (OR = 2.92 95%CI = 1.79-4.77), no heterogeneity between studies (I 2 = 0.0%, P = 0.557). In Group two, the relationship remained for ve studies with continuous D-dimer value (OR = 1.13 95%CI = 1.10-1.16). Group three included two studies with D-dimer > 8.0 ug/mL, and the pooled OR also indicated a predictive value of D-dimer to AAD mortality (OR = 4.12 95%CI = 1.56-10.93). The association remained for six studies which explored the relation between D-dimer and mortality of Type A AAD patients (OR = 1.13 95%CI = 1.08-1.18) (Fig. 3).

Discussion
To our knowledge, this is the rst meta-analysis to explore the relationship between D-dimer and in-hospital mortality of acute aortic dissection. After pooling the relevant studies, a link between D-dimer and in-hospital mortality was found. Much more than this, the predicted value was more signi cant with higher D-dimer. It may be an effective predictive biomarker for clinicians to identify patients at high risk of death.
As we know, the role of D-dimer used as an initial diagnosis marker was widely studied [11][12][13] . However, the evidence of predicting role of D-dimer for AAD in-hospital mortality is not only scarce but controversial [14][15] .  [16] . A similar relationship was observed in the study conducted by Ohlmann et al. They found that D-dimer may reach a high level of above 20ug/mL in some AAD patients three hours after the symptoms onset, and the concentration was high in dead than survival patients [17] . In the present meta-analysis, we combined all the eligible studies together and con rmed the link between D-dimer and in-hospital mortality of AAD. What's more, we found a stronger predictive power with higher D-dimer cutoff value.
The mechanism that D-dimer can predicted mortality of AAD remains unclear. According to previous researches, dissection anatomical extent was assumed to associate with the amount of coagulation and brinolytic activation. Tissue factor of dissected aorta triggered coagulation cascade, the activation of extrinsic pathway can be counteracted by endogenous brinolytic activity. While, the serum D-dimer, as the degradation product of cross-linked brin, may rise due to endogenous brinolytic activity [18] . An increased plasma D-dimer may be a manifestation of severe anatomical degree of dissection and greater mortality. D-dimer might be high when ascending aorta is involved for re ecting the higher frequency of unfavorable outcomes. According to our study, the association was obvious and stable in type A AAD. It is better to separate the type of AAD when using D-dimer to predict the in-hospital mortality of AAD patients. This meta-analysis had several strengths. Firstly, we included both case-control and prospective studies to clarify the relationship between D-dimer and in-hospital mortality of AAD. Secondly, a large number of participants and cases guaranteed the su cient statistical power to draw the credible conclusions. Third, in the current meta-analysis, the covariate adjusted OR and 95% CI were extracted from each study, which increased the accuracy of the summary estimation. However, several limitations should also be considered. Firstly, the included studies were mainly conducted in China. Therefore, the nding might be limited to apply to other countries. Secondly, the included studies adjusted for different confounders might have in uenced our results. Lastly, publication bias was found among studies.

Conclusions
The result of this meta-analysis indicated that D-dimer maybe a predictor in-hospital mortality of acute aortic dissection patients. What's more, the higher of cutoff value of D-dimer, the stronger for the predictive ability.

Declarations
Ethics approval and consent to participate The study was approved by the ethics committee of the Changhai Hospital, Naval Military Medical University.

Consent for publication
The authors thank the patients for informed consent to publication of data.

Availability of data and materials
All the data and materials can be available.