A retrospective study of consecutive patients with acute TBAD undergoing TEVAR was conducted in Guangdong Provincial People's Hospital (Guangdong, China) between January 2010 to December 2017. Contrast-enhanced computed tomography angiography (CTA) was used to confirm the diagnosis of TBAD using Stanford classification criteria. Patients were excluded for the following criteria: (1) traumatic aortic dissection, (2) previous aortic surgery, (3) malignant tumor, (4) Marfan syndrome (confirmed preoperatively according to the Ghent criteria or revised Ghent criteria) (20, 21), (5) incomplete clinical data. The remaining 637 patients were included for a retrospective analysis (Figure 1). This research program was approved by the Ethics Committee of Guangdong Provincial People's Hospital with an informed consent waiver due to the retrospective study design.

Patients were treated with optimal medications and TEVAR as recommended by guidelines (2). The considerations for TEVAR in patients with uncomplicated TBAD included aortic diameter more than 40 mm, primary entry tear diameter more than 10 mm, a patent or partial thrombosed false lumen, and false lumen diameter more than 22 mm (22). The details of the standardized procedure for TEVAR had been described in our previous study (22). Briefly, the procedures were conducted in a cardiac catheterization room generally under local anesthesia. To cover the primary entry tear, stent-grafts were placed retrograde via femoral artery access using pre-closing method, and the diameter of the stent-graft was generally oversized by 5-10%. To obtain a 1.5-2 cm proximal landing zone, it was permitted to cover the origin of the left subclavian artery (LSA). The decision of reconstruction of the LSA was mainly determined by radiographic assessments of the vertebrobasilar circulation.

The information of survival and clinical manifestations were obtained through outpatient clinic interviews or telephone interviews. CTA was repeated at one, three, six, and 12 months following surgery, as well as annually thereafter. The follow-up imaging could be undertaken in any institution, and the results of reexaminations were assessed by two independent physicians. Baseline demographics, clinical characteristics, laboratory test results, imaging findings, medications, and outcomes were recorded on an electronic standardized form.

Acute TBAD was defined as a type B aortic dissection that occurred within 14 days of the onset of symptoms (2). Complicated TBAD was considered with the presence of any following: dissection with refractory pain, uncontrolled hypertension despite full medication, rapid aortic expansion, malperfusion syndromes, or signs of rupture (hemothorax, increasing periaortic, and mediastinal hematoma) (2).

Blood samples were collected from all patients following admission. Monocyte counts were measured by an automated hematological counter (XE-5000, Sysmex, Kobe, Japan). Levels of HDL-C were evaluated enzymatically from fasting venous blood, using an automated biochemical analyzing apparatus (UniCel DxC 800 Synchron, Beckman Coulter, Brea, CA, USA). MHR was derived by dividing monocyte counts (×10⁹/L) by HDL-C level (mmol/L). All laboratory tests were carried out in the study center's Laboratory Department using established measuring procedures (ISO 9000 Quality Management and Assurance Standards).

Outcomes were reported in accordance with the TBAD reporting standards (23). The primary outcomes of interest were in-hospital and long-term all-cause death. In-hospital death was defined as mortality occurring during the hospitalization after TEVAR. Long-term death was defined as mortality that occurred after discharge or more than 30 days after TEVAR. In-hospital major adverse clinical events (MACE), including death, stroke, spinal cord ischemia, limb or visceral ischemia, and re-intervention, were considered as the secondary outcomes of interest.

Continuous variables were presented as means ± standard deviation or median (quartiles 1 through 3) according to the distribution characteristics, and compared by student t-test or Mann-Whitney test. For categorical variables, data were expressed as percentages and performed by Chi-squared test or Fisher's exact test. Furthermore, we compared the level of MHR among different timing group based on time intervals from the symptom onset to admission according to the International Registry of Aortic dissection (IRAD) classification system for characterizing survival (24): ≤ 24 h, 2-7 days, and 8-14 days.

Cumulative survival curves were performed by Kaplan-Meier method, with log-rank tests used to discriminate between the curves of groups. The receiver operating characteristic curve (ROC) was constructed to evaluate the predictive validity of MHR for in-hospital and long-term all-cause mortality, and the area under curve (AUC) was compared by Delong's method. To assess the association between preoperative MHR and mortality, univariable and multivariable logistic regression models were built for in-hospital mortality, and cox regression models for long-term mortality. The candidate variables for the multivariable model were listed in Table 1. Variables with a P-value <0.1 in the univariable analysis were entered in the multivariable models using a forward stepwise approach. Odds ratio (OR) or hazard ratio (HR) and corresponding 95% confidence interval (CI) were reported. MHR was firstly entered as a continuous variable. Moreover, MHR was modeled as a categorical variable, with the optimal cutoff evaluated by ROC curve.

To adjust for baseline variations and diminish selection bias, a propensity score matching (PSM) analysis was conducted using a 1:1 nearest-neighbor matching with a caliper of 0.01. Propensity score was calculated for each patient based on a logistic regression model using variables listed in Table 1. The difference between groups after PSM were measured by standardized mean difference (SMD). A maximum SMD of 0.1, or even 0.15, is usually regarded as appropriate. Moreover, we categorized patients into three groups by the tertile of MHR to enhance clinical utility, and repeated the analyses. To investigate the consistency of the conclusion, subgroup analyses were performed by age (<65 vs. ≥65 years), gender (male vs. female), severity of disease (complicated vs. uncomplicated), anemia status (yes vs. no), and kidney function (eGFR <60 vs. eGFR ≥ 60 ml/min/1.73 m²). Subgroup analyses were not performed for in-hospital death because of the low mortality.

A two-tailed P-value <0.05 was considered statistically significant. All statistical analyses were carried out using the SPSS 23.0 (IBM SPSS 23 Inc) and R software (version 3.5.1).

Among 637 enrolled patients, the mean age was 54.0 ± 11.1 years, and 554 (87.0%) were male. The average BMI 24.8 ± 3.8 kg/m², and the most common comorbidity was hypertension (85.9%). The median (interquartile range) of preoperative MHR was 1.09 (0.77-1.44), ranging from 0.02 to 4.78. Compared with uncomplicated cases, MHR was significantly higher in patients with acute complicated TBAD (1.13 [0.82-1.46] vs. 0.93 [0.70-1.39], p = 0.007). Moreover, patients with MHR > 1.13 were more likely to have a greater percentage of complicated patients compared with those with MHR ≤ 1.13 (77.2 vs. 64.8%, P = 0.001, Table 1). Regarding different timing groups, MHR was significantly higher in the 8-14 days group compared with 2-7 days and ≤ 24 h group (1.19 [0.80-1.66] vs. 1.01 [0.83-1.45] vs. 0.96 [0.65-1.36], P = 0.001]). Subgroup analyses showed that MHR was still significantly higher in the 8-14 days group compared with other two timing groups in complicated cases (1.24 [0.85-1.76] vs. 1.15 [0.88-1.49] vs. 0.97 [0.68-1.38], P = 0.001), while the differences among three timing groups in uncomplicated cases were not statistically significant (1.13 [0.74-1.51] vs. 0.90 [0.73-1.36] vs. 0.92 [0.55-1.19], P = 0.180). Demographic and clinical characteristics were shown in Table 1. Patients were classified into two groups using a cutoff of 1.13 (sensitivity 78.8%; specificity 58.9%) calculated based on the ROC curve for long-term mortality. In the entire cohort, patients with MHR > 1.13 were more likely to be male, with a higher percentage of complicated patients or pleural effusion, and with a lower percentage of hyperlipidemia history or patent false lumen. In addition, patients with MHR > 1.13 had higher preoperative WBC counts, platelet counts, serum creatinine, alanine aminotransferase (ALT) as well as aspartate aminotransferase (AST), and lower levels of TC as well as low density lipoprotein cholesterol (LDL-C).

In total, 21 (3.3%) in-hospital deaths were recorded, with 16 patients presented with MHR > 1.13. The incidence of in-hospital death was significantly higher in patients with MHR > 1.13 than that with MHR ≤ 1.13 (5.5 vs. 1.4%, P = 0.004) (Figure 2A). During the hospitalization period, 21 (3.3%) patients had stroke, 7 (1.1%) patients had spinal cord ischemia, 14 (2.2%) patients had limb ischemia, 3 (0.5%) patients had visceral ischemia, and 9 (1.4%) patients had re-intervention. The rates of in-hospital MACE were significantly higher in patients with MHR > 1.13 than that with MHR ≤ 1.13 (12.8 vs. 7.2%, P = 0.019) (Figure 2A). Multivariable logistic analysis revealed that MHR (modeled as a continuous variable) was associated with a significantly increased risk of in-hospital mortality [odds ratio (OR) 2.11, 95% confidence interval (CI) 1.16-3.85, P = 0.015) (Table 2). Other independent predictors for in-hospital death were maximum aortic diameter, hemoglobin, AST, and calcium channel blockers (Supplementary Table 1). Though the cut-off valve of 1.3 was generated from follow-up deaths, we found a significant association between MHR > 1.13 and in-hospital mortality (OR 4.53, 95% CI 1.44-14.30, P = 0.010) (Table 2 and Supplementary Table 1).

Following a median follow-up of 48.1 months (interquartile range, 26.2-72.2 months), 52 deaths (8.4%) were documented after discharge, with 41 patients in MHR > 1.13 group. The incidence of long-term mortality was significantly higher in patients with MHR > 1.13 than that with MHR ≤ 1.13 (15.0 vs. 3.2%, P < 0.001) (Figure 2A). Multivariable Cox regression analysis indicated that MHR (modeled as a continuous variable) was independently associated with long-term mortality [hazard ratio (HR) 1.78, 95% CI 1.31-2.41, P < 0.001] (Table 2). Other independent predictors for long-term mortality included maximum aortic diameter, coronary artery disease history, and serum creatinine (Supplementary Table 2). As a categorical variable, MHR > 1.13 was associated with a significantly increased risk of long-term mortality (HR 4.16, 95% CI 2.13-8.10, P < 0.001) (Table 2 and Supplementary Table 2). Patients with MHR > 1.13 had a significantly greater cumulative long-term mortality risk than patients with MHR ≤ 1.13. (log-rank = 21.66, P < 0.001) (Figure 3A).

After PSM, 190 matched pairs were obtained. The baseline characteristics in the matched cohort were outlined in Table 1. In-hospital mortality rate remained significantly higher for patients with MHR > 1.13 in the matched cohort (6.3 vs. 1.6%, P = 0.018), while the rates of in-hospital MACE were comparable between two matched groups (13.2 vs. 9.5%, P = 0.257) (Supplementary Figure 1). Either as a continuous variable (OR 2.72, 95% CI 1.23-5.99, P = 0.013) or as a categorical variable (OR 7.12, 95% CI 1.52-33.45, P = 0.013), the association between MHR and in-hospital death persisted in the multivariable logistic analysis after matching (Table 2 and Supplementary Table 3). In survival analyses, long-term mortality rate remained significantly higher for patients with MHR >1.13 in the matched cohort (16.3 vs. 4.3%, P < 0.001) (Supplementary Figure 1). Multivariable Cox regression analyses showed that MHR, modeled as either a continuous variable (HR 1.87, 95% CI 1.27-2.76, P = 0.002) or a categorical variable (more than 1.13) (HR 4.02, 95% CI 1.83-8.83, P = 0.001), was an independent predictor for long-term mortality after matching (Table 2 and Supplementary Table 4). Patients with MHR > 1.13 had a significantly greater cumulative long-term mortality risk than patients with MHR ≤ 1.13 (log-rank = 13.37, P < 0.001) (Supplementary Figure 2).

The predictive value of monocyte count, HDL-C, and MHR for long-term mortality was determined using a ROC curve analysis. Figure 4 showed that MHR (AUC 0.694, 95% CI 0.656-0.730, P < 0.001) was superior to either monocyte count (AUC 0.615, 95% CI 0.575-0.654, P = 0.006) or HDL-C (AUC 0.635, 95% CI 0.596-0.674, P = 0.001) (compared with monocyte count and HDL-C: P = 0.008 and 0.176, respectively).

We categorized patients into three groups by the tertile of MHR to enhance clinical utility: <0.88, 0.88-1.33, and ≥1.33. The in-hospital mortality (0.9 vs. 1.9 vs. 7.1%, P = 0.001), in-hospital MACE (6.1 vs. 7.5 vs. 15.6%, P = 0.002), and long-term mortality (3.3 vs. 6.7 vs. 15.8%, P < 0.001) increased from the first to third MHR tertile (Figure 2B). Compared with the lowest MHR tertile, the highest MHR tertile was independently associated with in-hospital mortality (adjusted OR 12.43, 95% CI 2.17-71.11, P = 0.005) (Supplementary Table 1) and long-term mortality (adjusted HR 3.78, 95% CI 1.66-8.60, P = 0.002) (Supplementary Table 2). The cumulative long-term mortality risks among the three groups were significantly different (log-rank = 15.02, P = 0.001) (Figure 3B).

In subgroup analyses, we stratified the patients by age (<65 vs. ≥65 years), gender (male vs. female), severity of disease (complicated vs. uncomplicated), anemia status (yes vs. no), and kidney function (eGFR <60 vs. eGFR ≥ 60 ml/min/1.73 m²). Multivariable Cox regression analyses revealed that MHR, modeled as either a continuous or categorical (more than 1.13) variable, remained independently associated with long-term mortality in subgroups analyses (Table 3 and Supplementary Tables 5-14).