DOI: https://doi.org/10.21203/rs.3.rs-15942/v1
Objective: The study objective was to investigate the incidence and risk factors of continuous renal replacement treatment (CRRT) in patients undergoing emergency surgery for type A acute aortic dissection (TA-AAD) and evaluate the perioperative and long-term outcomes.
Methods: From January 2014 to December 2018, 712 consecutive patients were enrolled in the study. These patients were divided into two groups according to whether or not needed severe postoperative acute kidney injury (AKI) requiring CRRT: the CRRT group vs the control group. Univariate analysis and binary logistic regression analysis were used to analyze the risk factors of CRRT. Significant variables by univariate analysis were included in binary logistic regression analysis. To avoid the selection bias and confounders, baseline characteristics were matched for propensity scores. One-to-one pair matching was performed using nearest neighbor matching without replacement within 0.02 standard deviations of the logit of the propensity score as caliper width. Kaplan-Meier curves were generated to provide survival estimates at postoperative points in time. Differences between the 2 groups were determined by log-rank tests.
Results: Before propensity score matching, univariate analysis showed that there significant differences in age, preoperative hypertension, pericardial effusion, preoperative serum creatinine (sCr), intraoperative need for combined coronary artery bypass grafting (CABG) or mitral valve or tricuspid valve surgery, cardiopulmonary bypass (CPB) time, extracorporeal circulation assistant time, aortic cross-clamp time, drainage volume 24 hours after surgery and ventilator time between two groups. All were higher in the CRRT group (p<0.05). These risk factors were included in binary logistic regression. It showed that preoperative sCr (OR=1.008, 95% CI:1.002-1.014, P=0.005) and CPB time (OR=1.022, 95% CI:1.003-1.042, P=0.026) were independent risk factors for CRRT patients undergoing surgery for TA-AAD. And there were significant differences regarding 30-day mortality (P<0.001) and long-term overall cumulative survival (P<0.001) with up to a 6-year follow-up. After propensity scoring, 29 pairs (58 patients) were successfully matched. Among these patients, the analysis showed that CPB time was still significantly longer in the CRRT group (P = 0.004), and the 30-day mortality rate was also higher in this group (44.8% vs 10.3%; P = 0.003).
Conclusion: CRRT after TA-AAD is common and worsened short- and long- term mortality. The preoperative sCr and CPB time are independent risk factors for postoperative CRRT patients. Shorten the CPB time as much as possible is recommended to reduce the risk of CRRT after the operation.
Previous studies have shown a high occurrence of acute kidney injury (AKI) in patients undergoing cardiothoracic surgery procedures. It is related to increased mortality, morbidity, longer intensive care unit (ICU) stay time and reduces late survival[1–4]. Continuous renal replacement treatment (CRRT) is an effective treatment strategy for those patients with acute renal failure (ARF), especially those who require both circulatory and respiratory support.
Although mild to moderate acute kidney injuries has a high incidence[5], 2–15% of patients with AKI need CRRT after aortic surgery, and especially higher in those with TA-AAD surgery[1, 6, 7]. Despite continued progress in renal replacement therapy and intensive care in the past years, the short- and long-term mortality still remains high, with a rate ranging from 50% to over 80% in patients undergoing CRRT[8]. However, there is scarce literature on the incidence, risk factors and long-term outcomes for CRRT in patients with TA-AAD repair. Identification of risk factors for CRRT in patients with TA-AAD may lead to timely initiation of CRRT, and improve clinical outcomes. This retrospective study aims to identify the risk factors and 30-day and long-term outcomes for CRRT after surgical repair in patients with TA-AAD.
A total of 730 consecutive patients with TA-AAD in our hospital from January 2014 to December 2018 were retrospectively analyzed. Previous approval was obtained from the institutional research ethics committee, which waived the need for individual informed consent. Aortic dissection was diagnosed by computed tomography angiography at either our institution or the referring hospital. Patients with a history of chronic renal failure (CRF) or with intraoperative and postoperative death within 24 hours were excluded. Patients were divided into the control group (601 cases) and the CRRT group (111 cases) according to whether they received CRRT after surgery. The basic clinical data of the two groups are shown in Table 1.
Demographic variables included age, gender, body mass index (BMI), previous medical history (hypertension, diabetes, cardiac surgery, coronary artery disease, cerebrovascular disease), aortic dissection features (blood supply of renal artery) and pericardial effusion. Operation-related variables were the duration of cardiopulmonary bypass (CPB) and aortic cross-clamping, extracorporeal circulation assist, the duration of deep hypothermic circulatory arrest (DHCA). Laboratory variables included preoperative serum creatinine (sCr) and serum blood urea nitrogen (BUN) levels. Postoperative variables included drainage volume 24 hours after surgery, duration of mechanical ventilation, ICU and hospital stay, and 30-day mortality.
Criteria for the initiation and termination of CRRT after severe AKI is referred to the guidelines for the clinical practice of AKI from the global organization for the improvement of renal prognosis: Kidney Disease Improving Global Outcomes (KDIGO)[9]. When patients have severe AKI after surgery, that is, within 48 h after surgery, when the increase in sCr was greater than 26.5umol/L or the urine volume was less than 0.5 ml/kg/h lasting for 6 h and serum K + > 6.0 mmol/L or blood HCO3− <10 mmol/L were observed, CRRT may be considered. During treatment, the corresponding CRRT mode and treatment amount were set according to the individual situation of the patient. CRRT termination criteria: the increase in sCr within 48 hours after surgery is less than 26.5umol/L or the urine volume is more than 0.5 ml/kg/h lasting for 6 h, and serum K + < 6.0 mmol/L or HCO3− >10 mmol/L. Within 48 h after the last CRRT treatment, without CRRT support, the results of two sCr tests decreased by a range of > 50umol/L (the sampling interval is greater than 12 h) or urine volume > 0.5 ml/kg/h within 12 h, and CRRT treatment could be considered for termination when serum K + < 5.5 mmol/L and HCO3− >18 mmol/L in the latest blood gas test [10].
CRRT was performed in our department, using the 11.5f double-chamber dialysis catheter, the AV600S polysulfone membrane blood filter and the connection pipeline of blood filtration, infusion pump, and syringe pump. The internal jugular vein or femoral vein or subclavian vein was selected to place a single double-chamber blood filter catheter. The hemodynamic force is provided by the blood pump. 1000 ml heparin brine was pre-flushed before using the filter to empty the air bubbles in the filter and pipeline. We then placed the sterile collecting bag 30-50cm below the filter, and recorded the flow of liquid in and out every hour. In the early postoperative patients after aortic dissection, local anticoagulation of prefilter citrate was used to reduce bleeding. The replacement solution was 0.9% normal saline and 5% glucose solution, with a ratio of 3:1. Additionally, 250 ml 5% sodium bicarbonate was added for q4h or q6h to timely supplement the physiological needs and nutrients lost by blood filtration. The input method is pre or post-dilution method, which can balance the fluid and adjust the infusion speed according to the amount of filtrate and input. The filter is usually replaced after a blockage or when the filtrate drops, and the continuous veno-venous hemofiltration (CVVH) blood flow should be 100–150 ml/min.
The median sternal incision was used in all surgeries under general anesthesia and DHCA. All patients were treated with Terumo inlet membrane lung, no pre-rinse containing sugar was used in extracorporeal circulation, ultrafiltration and autologous blood recovery devices were routinely used. Extracorporeal circulation was established by a routine femoral artery or right axillary artery and right atrial intubation. When the nasopharyngeal temperature dropped to 34℃, the ascending aorta was blocked and cardiac arrest fluid was injected to complete the operation of the proximal end of the aorta. When the nasopharyngeal temperature dropped to 18 ~ 20℃ and the bladder temperature to 22 ~ 24℃, systemic circulatory was arrested, and the flow was reduced to 3 ~ 5 ml· kg− 1·min− 1 to complete the operation of the aortic arch and descending arch. All patients returned to the ICU for routine monitoring and treatment.
SPSS 25.0 software was used for statistical analysis. Univariate analysis was performed for each variable. Data were compared using the Student t test or non- parametric Wilcoxon Mann-Whitney U test for continuous variables and the chi-squared or Fisher's exact test for categorical variables before matching; The multivariate model included variables that were significant on univariate analysis. Baseline characteristics (Demographic variables, previous medical history, aortic dissection features and pericardial effusion) were matched for propensity scores. We performed one-to-one pair matching using nearest neighbor matching without replacement within 0.02 standard deviations of the logit of the propensity score as caliper width.
Kaplan-Meier curves were generated to provide survival estimates at postoperative points in time. Differences between the 2 groups were determined by log-rank tests. For all analyses, a probability value of less than 0.05 was considered statistically significant.
Three patients with CRF before surgery and 15 patients who died intraoperatively or within 24 hours after surgery were excluded, a total of 712 patients were included in the study. Among them, 111 cases (15.9%) underwent CRRT after thoracic aortic surgery. For the CRRT group, there were 84 cases (75.7%) of males, the medical history included hypertension in 91 cases (82%), diabetes in 1 case (0.9%), coronary heart disease in 9 cases (8.1%), cardiovascular surgery in 8 cases (7.2%), cerebrovascular accident in 4 cases (3.6%), and pericardial effusion in 13 cases (11.7%). The operation methods included aortic root or ascending aorta replacement in 49 cases (44.1%), total aortic arch replacement in 62 cases (55.9%), and aortic valve treatment in 31 cases (27.9%), coronary artery bypass grafting (CABG) or mitral or tricuspid valve operation in 17 cases (15.3%). 107 patients (96.4%) received emergency operations. The mean CPB duration was 273.4 ± 80.6 minutes, the duration of aortic cross-clamp was 186.0 ± 70.5 minutes, and the mean duration of DHCA was 30.2 ± 50.1 minutes, as shown in Table 1.
Compared with the CRRT group, preoperative age, hypertension, pericardial effusion, preoperative sCr and BUN, intraoperative need for combined CABG or mitral valve or tricuspid valve surgery, CPB time, extracorporeal circulation assistant time, aortic cross-clamp time, drainage volume 24 hours after surgery, ventilator time, 30-day mortality, ICU time between two groups,all above results were higher in CRRT group (P < 0.05). There was no significant difference in hospital stay between the two groups. (Table 1,2)
Binary logistic regression analysis showed that preoperative sCr (OR = 1.008, 95%CI: 1.002–1.014, P = 0.005) and CPB time (OR = 1.022, 95%CI: 1.003–1.042, P = 0.026) were independent risk factors for CRRT after thoracic aortic surgery. (Table 3)
After propensity scoring, 29 pairs (58 patients) were successfully matched. In these patients, the analysis showed that CPB time was significantly longer in CRRT group (P = 0.004), and the 30-day mortality rate was also higher in this group (44.8% vs 10.3%; P = 0.003). (Table 2)
52 patients in the CRRT group and 91 patients in the control group died during the hospitalization period. At last, a total of 59 CRRT patients and 510 non-CRRT patients were included in the follow-up. The median follow-up time was 29 months. 48 patients were lost from follow-up. There were 40 late deaths, including 21 related to cardiovascular events; 25 deaths (4.9%) occurred in the control group and 15 (25.4%) in the CRRT group (P < 0.001). The survival is shown in Fig. 1. The 6-year survival rate for the CRRT group was 55.0 ± 3.3%, and that for the control group was 76.8 ± 0.8% (P < 0.001).
|
Table 1 Baseline clinical characteristics of patient population and aortic dissection features before and after propensity score matching |
||||||
|
|
Before Matching |
After Matching |
||||
|
Characteristics |
Control(n=601) |
CRRT(n=111) |
P Value |
Control(n=29) |
CRRT(n=29) |
P Value |
|
Patient characteristics |
|
|
|
|
|
|
|
Demographics |
|
|
|
|
|
|
|
Age (year) |
52.0±13.0 |
55.4±14.0 |
0.011 |
54.6±13.2 |
54.8±13.2 |
0.937 |
|
Male (%) |
438(72.9) |
84(75.7) |
0.540 |
20(69.0) |
23(79.3) |
0.368 |
|
BMI (kg/m2) |
25.2±5.0 |
26.1±5.0 |
0.146 |
25.8±3.4 |
27.0±5.0 |
0.278 |
|
Medical history |
|
|
|
|
|
|
|
Hypertension (%) |
401(66.7) |
91(82.0) |
0.001 |
20(69.0) |
21(72.4) |
0.773 |
|
Diabetes mellitus (%) |
11(1.8) |
1(0.9) |
0.703 |
0 |
0 |
-- |
|
Previous cardiac surgery (%) |
28(4.7) |
8(7.2) |
0.260 |
0 |
0 |
-- |
|
Previous Coronary artery disease (%) |
18(3.0) |
9(8.1) |
0.025 |
1(3.4) |
2(6.9) |
1.000 |
|
Cerebrovascular disease (%) |
20(3.3) |
4(3.6) |
0.779 |
2(6.9) |
1(3.4) |
1.000 |
|
Pericardial effusion (%) |
15(2.5) |
13(11.7) |
<0.001 |
1(3.4) |
1(3.4) |
1.000 |
|
Preoperative laboratory data |
|
|
|
|
|
|
|
BUN (mmol/L) |
7.0±3.8 |
11.5±8.2 |
<0.001 |
7.8±3.0 |
6.6±1.9 |
0.063 |
|
sCr (μmol/L) |
91.0±48.4 |
176.0±156.5 |
<0.001 |
96.6±46.7 |
90.1±35.1 |
0.552 |
|
Aortic dissection features |
|
|
|
|
|
|
|
Blood supply of left renal artery (%) |
|
|
|
|
|
|
|
True lumen |
326（64.2） |
38（62.3） |
|
14（58.3） |
6（37.5） |
|
|
False lumen |
133（26.2） |
17（27.9） |
|
9（37.5） |
7（43.8） |
|
|
True lumen and false lumen |
49（9.6） |
6（9.8） |
0.978 |
1（4.2） |
3（18.8） |
0.272 |
|
Blood supply of right renal artery (%) |
|
|
|
|
|
|
|
True lumen |
390（76.8） |
41（67.2） |
|
18（75.0） |
11（68.8） |
|
|
False lumen |
91（17.9） |
13（21.3） |
|
4（16.7） |
2（12.5） |
|
|
True lumen and false lumen |
27（5.3） |
7（11.5） |
0.106 |
2（8.3） |
3（18.8） |
0.766 |
|
Notes: Data presented as n (%); mean ± standard deviation. Abbreviations: BMI, body mass index; BUN, blood urea nitrogen; sCr, serum creatinine |
||||||
|
Table 2 Operative data and outcome before and after propensity score matching |
||||||
|
|
Before Matching |
After Matching |
||||
|
Characteristics |
Control(n=601) |
CRRT(n=111) |
P Value |
Control(n=29) |
CRRT(n=29) |
P Value |
|
Operative procedures |
|
|
|
|
|
|
|
Total arch replacement (%) |
309(51.4) |
62(55.9) |
0.389 |
16(55.2) |
20(69.0) |
0.279 |
|
CABG/MVR/MVP/TVP (%) |
51(8.5) |
17(15.3) |
0.024 |
5(17.2) |
4(13.8) |
1.000 |
|
Aortic valve (%) |
185(30.8) |
31(27.9) |
0.548 |
9(31.0) |
7(24.1) |
0.557 |
|
CPB-related profiles |
|
|
|
|
|
|
|
CPB time (min) |
235.5±64.1 |
273.4±80.6 |
<0.001 |
236.0±48.8 |
279.3±59.8 |
0.004 |
|
Extracorporeal circulation assist time (min) |
54.3±31.0 |
65.9±40.4 |
0.001 |
50.3±30.7 |
60.4±24.5 |
0.174 |
|
Aortic cross-clamp time (min) |
165.3±55.1 |
186.0±70.5 |
0.005 |
174.6±48.2 |
191.8±50.9 |
0.192 |
|
DHCA time (min) |
28.0±12.7 |
30.2±50.1 |
0.117 |
26.8±10.4 |
31.3±13.9 |
0.173 |
|
Complications and short-time outcomes |
|
|
|
|
|
|
|
Drainage volume 24 hours after surgery (ml) |
695.0±604.1 |
1091.0±1013.9 |
0.001 |
971.6±1104.1 |
1011.5±681.9 |
0.873 |
|
Ventilation time (hour) |
32.0±50.1 |
56.8±61.8 |
0.009 |
37.6±41.6 |
79.9±81.9 |
0.058 |
|
30-day mortality (%) |
40(6.7) |
50(45.0) |
<0.001 |
3(10.3) |
13(44.8) |
0.003 |
|
ICU Stay time (day) |
6.5±11.5 |
9.4±6.6 |
0.020 |
5.9±5.7 |
9.3±6.9 |
0.055 |
|
Hospital stay time (day) |
22.3±11.6 |
22.5±17.0 |
0.892 |
19.4±10.0 |
21.3±14.3 |
0.560 |
Notes: Data presented as n (%); mean ± standard deviation.
Abbreviations: CABG, coronary artery bypass graft; MVR, mitral valve replacement;
MVP, mitral valvuloplasty; TVP, tricuspid valvuloplasty; CPB, cardiopulmonary bypass; DHCA, deep hypothermic circulatory arrest; ICU, intensive care unit.
| Variable | Odds Ratio | 95% CI | P value |
|---|---|---|---|
| Age | 1.052 | 0.984–1.052 | 0.312 |
| Hypertension | 2.027 | 0.645–6.376 | 0.227 |
| Previous Coronary artery disease | 0.319 | 0.021–4.866 | 0.412 |
| Pericardial effusion | 0.552 | 0.043–7.051 | 0.647 |
| BUN | 1.041 | 0.966–1.123 | 0.293 |
| sCr | 1.008 | 1.002–1.014 | 0.005 |
| CABG/MVR/MVP/TVP | 0.501 | 0.110–2.279 | 0.371 |
| CPB time | 1.022 | 1.003–1.042 | 0.026 |
| Extracorporeal circulation assist time | 0.997 | 0.980–1.014 | 0.721 |
| Aortic cross-clamp time | 0.987 | 0.967–1.008 | 0.229 |
| Drainage volume 24 hours after surgery | 1.000 | 1.000-1.001 | 0.314 |
| Ventilation time | 1.006 | 0.998–1.014 | 0.131 |
| Abbreviations: CI, confidence interval; BUN, blood urea nitrogen; CABG, coronary artery bypass graft; | |||
| MVR, mitral valve replacement; MVP, mitral valvuloplasty; TVP, tricuspid valvuloplasty; CPB, cardiopulmonary bypass. | |||
To the best of our knowledge, the study we report herein is the first one to investigate the risk factors and outcomes of emergent surgery for TA-AAD in postoperative CRRT patients. In this study, 111 patients (15.6%) suffered severe AKI after surgery and needed CRRT. The 30-day and long-term mortality rates were significantly higher in the CRRT group (P < 0.001). Our findings showed that in patients undergoing surgery for TA-AAD under the DHCA procedure, preoperative sCr and CPB time are independent risk factors for postoperative CRRT. To balance the selection bias and other confounders, baseline characteristics were matched for propensity scores. After matched, CPB time and 30-day mortality still showed significant differences between the two groups. Our Kaplan-Meier plots also revealed that the mortality of postoperative dialysis patients was significantly higher than that of the control group.
Consistent to our findings, Rice, R.D et al.[7] confirmed that the proportion of CRRT for such patients was 18.2%. In another study conducted by Estrera et al., 347 patients with typical type A aortic dissection were analyzed, 64 cases of which suffered from CRRT (18.4%) [6]. The incidence of postoperative CRRT in our cohort was comparable with these previous studies. CRRT procedure involved in the context of renal failure may lead to circulation instability, infection, thrombosis, and electrolyte imbalance, which will impair the postoperative recovery of patients, and even cause death[11], implying a low survival rate of the CRRT group as shown in our study. Elahi et al. suggested that early initiation of CRRT may reduce mortality and mortality in patients with severe AKI after cardiac surgery[12].
sCr level was identified as a risk factor for CRRT in our study, which shows a significant influence of elevations in the preoperative creatinine serum levels on postoperative renal function. Study from Wu’s group also suggested that sCr level is a risk factor for CRRT after surgical repair of TA-AAD[13]. They concluded that hypotension, nephrotoxins and inflammation during aortic dissection may lead to acute renal failure, followed by a substantial increase in serum creatinine. Actually, pre-operation elevated sCr concentration might also indicate structural kidney damage or hemodynamic derangements in acute aortic dissection patients, which further aggravated the development of post-operation AKI and the requirement of CRRT. Previous reports also showed the CPB time is an independent risk factor for postoperative AKI during routine cardiac surgery[14, 15]. Specifically for aortic surgery, Roh GU et al.[1] analyzed 98 patients who underwent graft replacement of the thoracic aorta and found that longer CPB time (> 180 minutes; OR, 7.50; P = 0.008) were independent risk factors for postoperative AKI. Recently another analysis of 113 cases of type A aortic dissection after surgery by Xu and his colleagues also found that CPB time (OR = 1.171; 95% CI: 1. 002–1.368; P = 0.047) was an independent risk factor for postoperative renal injury[16]. CRRT has emerged as the preferred dialysis modality for critically ill patients with acute kidney injury, particularly those with hemodynamic instability. From this perspective, AKI and CRRT may have common risk factors in patients with surgery for TA-AAD. Not surprisingly, CPB time is proved an independent risk factor for CRRT in our study. The mechanism of severe postoperative AKI and ARF caused by CPB is still unknown. Renal hypotension and hypoperfusion during CPB may be the main causes[17]. The CPB procedure causes damage to red blood cells, and the resulting cell debris forms tiny emboli that block the renal tubular vascular network, resulting in a decrease in the filtration area of renal tubules[18, 19], inflammatory mediators release causing tubular damage[20]. Besides, circulation is stopped during aortic dissection surgery, and ischemia-reperfusion injury may occur during the operation, ischemia-reperfusion injury may lead to massive apoptosis of renal endothelial cells[21].
Based on our findings, the mortality rate of the CRRT group is significantly higher than that of the control group on 30-day and six years postoperative outcomes. This may be related to the functional loss of kidney caused by AKI. Several studies have reported postoperative AKI is an independent predictor of in-hospital and long-term mortality after operation for TA-AAD[22, 23]. A report from Japan found that in patients who underwent surgery for TA-AAD, the overall 30-day and mid- to long-term mortality was 1.6% and 8.8%, respectively. The mortality and major adverse cardiovascular and cerebrovascular events correlated significantly with the severity of AKI, and the severity of AKI strongly influences patient outcomes. Furthermore, multivariate analysis showed that AKI stage 3 (the most severe stage) was an independent risk factor for mortality (hazard ratio 6.83, with 95% CI 2.52 to 18.52) after adjustment for important confounding factors[24]. Sasabuchi et al[2] also reported that stage 3 AKI defined by the KDIGO criteria had a significantly lower survival rate with long-term follow-up, with a median survival time of 58 months. The hazard ratios for patients with stages 1 and 2 AKI were not significantly higher than patients without AKI. Consequently, the effective control of renal injury is worthy of study to improve the prognosis of aortic dissection.
Like any retrospective study of patients with TA-AAD, this study has some limitations. First of all, this is a single-center retrospective study, and there may be other confounding factors that influence the results. Secondly, despite propensity score matching was used, further potential bias and con-founders in the context of patient selection and treatment cannot be completely excluded, and the statistical power was limited due to a small patient cohort (29 pairs). Finally, our surgical technique has evolved over the study period, and our findings may have been influenced by the involvement of different surgeons.
In conclusion, patients receiving CRRT treatment who underwent emergency surgery for TA-AAD had higher rates of perioperative mortality and postoperative morbidity. The preoperative serum creatinine and CPB time were independent risk factors for postoperative CRRT. Further prospective multicenter studies are needed to assess the prognostic significance of CRRT and establish the most effective strategies to prevent and treat postoperative CRRT and thereby improve patient outcomes.
CRRT: Continuous renal replacement treatment; TA-AAD: Type A acute aortic dissection; AKI: Acute kidney injury; sCr: Serum creatinine; CABG: Coronary artery bypass grafting; CPB: Cardiopulmonary bypass; ICU: Intensive care unit; ARF: Acute renal failure; CRF: Chronic renal failure; BMI: Body mass index; DHCA: Deep hypothermic circulatory arrest; BUN: Blood urea nitrogen; KDIGO: Kidney Disease Improving Global Outcomes; CVVH: Continuous veno-venous hemofiltration; MVR: Mitral valve replacement; MVP: Mitral valvuloplasty; TVP: Tricuspid valvuloplasty; CI: Confidence interval
Acknowledgements
Not applicable.
Funding
Not applicable.
Availability of data and materials
The datasets used or analyzed during the current study are available from the corresponding author on reasonable request.
Authors’ contributions
DJW, MG, ZGW designed the study; ZGW, TC, CC collected the data; ZGW, LCL analyzed the data; ZGW, QYZ analyzed and interpreted the results; DJW support and encourage the study; ZGW wrote this article; All the authors have read and reviewed this manuscript.
Ethics approval and consent to participate
The study protocol was approved by the ethics committee at Nanjing Drum Tower Hospital, and all experimental methods were performed in accordance with the relevant guidelines and regulations.
Consent for publication
Consent was obtained from the patients or their relatives.
Competing interests
The authors have declared that no interest.