Consecutive patients diagnosed with AAD were recruited from January in 2012 to December in 2021 for the cohort (8, 9) of acute aortic dissection patients in the First Affiliated Hospital of Shantou University Medical College in Shantou, Second Affiliated Hospital of Xi'an Medical College. AAD was defined as a separation of the layers of the aortic wall due to an intimal tear within two weeks from initial clinical presentation. Computed tomographic angiography was employed to definitively diagnose aortic dissection, and its categorization was determined using the Stanford classification system. Specifically, Stanford type A aortic dissection is that includes the ascending aorta, regardless of intimal tear's site and distal extension of the aortic lesion. And Stanford type B is referred to involved the descending thoracic aorta distal to the left subclavian artery and its distal end. Within the two-center aortic dissection cohort study, a total of 2,413 patients were identified as having AAD. 14 patients had an unclear type of aortic dissection, 263 patients were lost to follow-up, and a further 155 patients had incomplete serum fibrinogen data, as illustrated in Figure 1. Ultimately, the study included 1981 patients diagnosed with AAD. Ethical approval for the study was obtained from the ethics committee of the First Affiliated Hospital of Shantou University Medical College. This cohort study adhered to the Declaration of Helsinki, and was constructed based on the STROBE cohort reporting guidelines.

In-hospital data was collected from medical records. In order to reduce selected bias risk, we exclusively collected PFL on the first day of the hospital admission and excluded patients diagnosed with aortic penetrating ulcer or chronic aortic dissection. Age, gender, type of AAD, surgery for type A AAD or endovascular stent-graft implantation for type B AAD, and history of hypertension, diabetes mellitus and coronary heart disease(CHD), international normalized ratio of prothrombin time (PT-INR) and plasma D-dimer concentration were collected. After discharge, follow-up was performed, by clinical interview or telephone, at 1, 3, 6, and 12 months, and every year after that. The primary outcome was 30-day all-cause mortality and long-term all-cause mortality.

Literature searches were performed in PubMed Library up to May 2024 by using various combinations of keywords including “Aortic Dissection”, “Aortic Dissections”, “Dissection, Aortic”, “Aortic Dissecting Aneurysm”, “Aneurysm, Aortic Dissecting”, “Aortic Dissecting Aneurysms”, “Dissecting Aneurysm, Aortic”, “Dissecting Aneurysm Aorta”, “Aneurysm Aorta, Dissecting”, “Aneurysm Aorta, Dissecting”, “Aorta, Dissecting Aneurysm”, “Dissecting Aneurysm Aortas”, “Aneurysm, Dissecting”, “Dissecting Aneurysms”, “Dissecting Aneurysm”, “Fibrinogen”, “Blood Coagulation Factor I”, “Coagulation Factor I”, “Factor I, Coagulation”, “Factor I”, “gamma-Fibrinogen” and “gamma Fibrinogen”. In addition, we manually searched the references of the retrieved literature in order to avoid missing potential eligible studies. The outcome for this analysis was AAD mortality. The articles were included if they: (1) were described in English, (2) compared the association of PFL and mortality in a cohort study, and (3) reported relative risks with 95% CIs, or the data to calculate them. Studies where only the abstract was available were excluded. Finally, name of first author, year of publication, age, the content to which the subject belonged, sample size, OR (adjusted OR if provided) and 95%CI were extracted.

Patients were classified as survivors or non-survivors, with those lost to follow-up excluded from the analysis. Continuous variables were presented as mean/SD or median/interquartile range depending on the results of the Kolmogorov–Smirnov normality test, while categorical data were presented as numbers and proportions. Mann–Whitney U tests and chi-square tests were used to compare baseline characteristics between the two groups for continuous and categorical variables, respectively. The Cox proportional hazards model was used to calculate the hazard ratio and 95% CI for the association between PFL and AAD mortality risk. Three multivariate Cox models, adjusted for potential confounders, were established to determine whether PFL was an independent predictor of mortality. X-tile analysis (version 3.6.1) and Area Under the Receiver Operating Characteristic Curve(AUC) were used to identify the optimal PFL cut-off value for predicting AAD mortality risk (10). Kaplan–Meier survival curves, stratified by PFL cut-off and quartile-based categorization, were constructed using the log-rank test. Subgroup analyses of interaction were performed for age (>60 vs. ≤60), gender (female vs. male), hypertension (yes vs. no), diabetes (yes vs. no), and aortic dissection type (A vs. B). Statistical significance was considered at a threshold of p < 0.05. In the meta-analysis, the Newcastle-Ottawa Scale (nine-star system) was used to assess the quality of the included studies (11). The I² statistic was used to quantify heterogeneity, with a fixed-effect model applied if I² ≤ 50% and a random-effect model used otherwise. In addition, a sensitivity analysis was performed using a one-study-removed analysis. All analyses were conducted using R (version 4.3.3).

This study included 1981 AAD patients, aged 62 on average, with 440 females (22.2%). The median follow-up was 21.6 months (interquartile range 8.6–45.4 months). For the two endpoint events, 294 patients died within 30 days, indicating a 14.8% 30-day mortality rate, and 487 patients died during the long-term follow-up, showing a 24.6% long-term mortality rate. The non-survivor group had a significantly higher average age and a larger proportion of Stanford type A dissection cases compared to the survivor group. PFL in the non-survivor group was also markedly lower (3.28, 2.16–4.28 g/L vs. 3.72, 2.59–4.73 g/L in the survivor group, p < 0.001) (Table 1).

Fibrinogen levels on admission are associated with all-cause mortality in patients with AAD.

The optimal cutoff points determined by X-tile analysis and AUC curve were both 2.0 g/L (Supplementary Figures S1, S2). Patients were divided into higher (>2.0 g/L) and lower (≤2.0 g/L) PFL groups. Kaplan–Meier survival curve analysis showed that fibrinogen levels ≤2.0 g/L were associated with higher risk of all-cause mortality in AAD patients (p < 0.001, Figure 2). Using multivariate Cox proportional hazard analysis, three distinct models were constructed. In the unadjusted Cox proportional hazards model 1, low fibrinogen was a risk factor for long-term mortality (HR = 2.923, 95% CI: 2.353–3.632, p < 0.001). Model 2, adjusting for age and sex, low fibrinogen was a risk factor for long-term mortality (HR = 3.11, 95% CI: 2.499–3.870, p < 0.001). Moreover, in Model 3, which incorporated additional adjustments for region, hypertension, diabetes, CHD, type of AD, surgical operation, stent-grafting, and plasma D-dimer levels. For every 1 g/L increase in PFL, the 30—day mortality rate decreased by 18.9%, and the long—term mortality rate decreased by 11.5%.Compared with PFL > 2 g/L, the risk of 30-day all-cause mortality for PFL ≤ 2 g/L was (HR = 2.429, 95% CI: 1.879–3.141, p < 0.001), and the effect on long-term mortality was HR = 2.017, 95% CI: 1.591–2.556, p < 0.001 (Table 2).

It showed the subgroup analysis of interaction between PFL and the risk of mortality in patients with aortic dissection (Figure 3). PFL was analyzed as a continuous variable, and the results indicated that low PFL was associated with an increased risk of all-cause mortality in patients with aortic dissection. No significant interaction was observed in the analysis of long-term mortality. However, a significant interaction between dissection type and PFL was noted in the analysis of 30-day mortality (P for interaction = 0.049). The hazard ratio (HR) for 30-day all-cause mortality was more pronounced in patients with type B dissection compared with those with type A dissection [HR: 0.64 (95% CI: 0.54–0.76) vs. HR: 0.79 (95% CI: 0.71–0.89)].

Supplementary Figure S2 showed the search strategy. According to the inclusion criteria, five studies involving 989 patients were eventually enrolled, with a total of 2,970 participants ultimately being included in the meta-analysis after including the participants of the present study (6, 7, 12–14). Table 3 describes the main features of the selected studies. Most of the studies were given more than five stars according to the Newcastle control study (Table 3) (15). To reduce heterogeneity among studies, we used 30-day mortality as the endpoint for the meta-analysis. The datasets were heterogeneous (I² = 47.9%). Therefore, fixed effects models were used for analyses. Meta-analysis of the included studies and our study revealed a significant association between lower PFL and increased all-cause mortality [OR = 3.30, 95% CI: 2.58–4.23, P = 0.0739; I² = 47.9%]. Meta-analysis of the included studies and our study revealed a significant association between lower PFL and increased 30-day mortality in type A and type B AAD [OR = 3.30, 95% CI: 2.58–4.23, P = 0.0739; I² = 47.9%]. Simultaneously, similar associations were observed in Stanford type A in for long-term mortality [OR = 3.62, 95% CI: 2.23–5.87, P = 0.0438; I² = 56.2%]. We could observe negative correlation trend in Stanford type B AAD patients, but it was not significant difference [OR = 1.45, 95% CI: 0.84–2.4] (Figure 4).

The strength of our study is that this is a two-center study. People from different regions were included.