Risk factors for hypoxemia following surgical repair of acute type A aortic dissection

OBJECTIVES

To identify the risk factors for hypoxaemia following surgical repair of acute type A aortic dissection.

METHODS

This was a retrospective study of patients treated between October 2013 and December 2014 at the Department of Cardiovascular Surgery, Anzhen Hospital, China. Univariable and multivariable analyses were performed on the clinical data of 160 patients with acute type A dissection and who underwent ascending aortic and arch replacement under deep hypothermic circulatory arrest.

RESULTS

Hypoxaemia occurred in 30% (48/160) of patients (age: 49 ± 7.9 years; 40 males, 83.3%). The duration of ventilation and the lengths of intensive care unit and hospital stays were significantly longer in patients with hypoxemia (77.9 ± 56.0 vs 16.5 ± 11.5 h, P < 0.0001; 6.0 ± 2.3 vs 2.0 ± 1.2 days, P = 0.001; 18.1 ± 6.3 vs 13.5 ± 4.7 days, P = 0.0012; respectively). The difference in operative mortality was not statistically significant between the hypoxaemia and non-hypoxaemia groups (6.25% vs 3.57%, P = 0.351). The independent risk factors of postoperative hypoxaemia were time from symptom onset to surgery ≤72 h [odds ratio, 3.63; 95% confidence interval, 1.31–10.02; P = 0.013], preoperative PaO₂/FiO₂ ≤300 (odds ratio, 15.30; 95% CI, 5.52–42.43; P < 0.001), preoperative white blood cell count >15 000/μl (odds ratio, 9.79; 95% CI, 2.47–38.87; P = 0.001); and deep hypothermic circulatory arrest time >25 min (odds ratio, 3.26; 95% CI, 1.18–8.99; P = 0.023).

CONCLUSIONS

Time from symptom onset to surgery ≤72 h, preoperative PaO₂/FiO₂ ≤300, white blood cell count >15 000/μl and deep hypothermic circulatory arrest time >25 min were found to be independently associated with hypoxaemia after surgery for acute type A aortic dissection.

INTRODUCTION

The aorta is the largest artery in the human body and is submitted to important haemodynamic stresses. Therefore, the aortic wall is subject to diseases such as dissection or aneurysms. Aortic diseases represent an important proportion of arterial diseases, and threaten the health of hundreds of people in the world. The incidence of aortic dissection in China is high (5–10 per 100 000 individuals) [1, 2]. Acute aortic dissection is often complicated with severe systemic pathophysiological changes and is associated with a high mortality rate because of severe multi-system and multi-organ perioperative complications [2].

Postoperative hypoxaemia is a frequent complication of surgery for acute type A aortic dissection, with an incidence of 51% [3]. Postoperative hypoxaemia prolongs mechanical ventilation, which in turn leads to serious complications, and increases the healthcare costs. Long-term postoperative hypoxaemia causes functional impairments in lungs and other organs, and increases the mortality, which is an important issue. Many studies suggest risk factors for respiratory dysfunction following aortic dissection surgery or aortic surgery [4, 5]. As studies focusing on the postoperative period after acute dissection surgery are rare, conclusions on risk evaluation and prevention measures are lacking; therefore, additional studies are required.

The risk factors for hypoxaemia following surgical repair of acute dissection need to be explored to provide relevant theoretical bases for identifying patients at higher risk of postoperative hypoxaemia in order to make the best clinical decisions. Therefore, this study aimed to identify the risk factors for hypoxaemia following surgical repair of acute type A aortic dissection.

METHODS

Study design

This retrospective study analysed the patients with acute type A dissection who underwent Sun's procedure (arch replacement + elephant trunk stent graft) [6] between October 2013 and December 2014 at the Department of Cardiac Surgery of the Anzhen Hospital affiliated to the Capital Medical University, China. This study was approved by the Ethics Committee of Anzhen Hospital, China, affiliated to the Capital Medical University.

Patients

Consecutive patients diagnosed with acute (within the past 2 weeks) type A dissection were selected. Inclusion criteria were: (i) patients diagnosed with type A aortic dissection confirmed by enhanced computed tomographic scan and (ii) patients who underwent ascending aortic + arch replacement + elephant trunk stent graft. Patients were excluded if they suffered from any perioperative complications (such as cardiac failure, massive haemorrhage pneumothorax, endotracheal haemorrhage, atelectasis and pneumonia) which might result in hypoxaemia.

Grouping

Patients with a postoperative oxygenation index <200 on two or more consecutive time points (each point 24 h apart) were allocated to the hypoxaemia group and the other patients were placed in the non-hypoxaemia group.

Surgical procedure

All patients were tested before surgery (prothrombin time, plasma thromboplastin antecedent and platelets) and these tests had to show that the patient was without disseminated intravascular coagulation to perform surgery. All procedures were conducted by the same surgical team according to the Sun's procedure [6]. After entering the operation room, fluid pathways were established via peripheral venous puncture. The left radial artery and left foot dorsal artery were punctured to measure the upper and lower limb blood pressures. The induction of anaesthesia was achieved with 10–15 mg of etomidate, 5–10 µg/kg of fentanyl and muscle relaxants. Endotracheal intubation for mechanical ventilation was performed with an inspired oxygen concentration of 100%. Anaesthesia was maintained intraoperatively with fentanyl (total dose <20 µg/kg) and isoflurane; diazepam, thiopental, or propofol were administered intermittently. The Sun's procedure [7] was used: a median sternotomy was performed and cardiopulmonary bypass (CPB) was established through axillary artery cannulation and the right atrium, with the left heart venting from the right superior pulmonary vein. After clamping the ascending aorta, operations at the proximal end were done (for instance, aortic valvuloplasty or replacement and proximal graft anastomosis). When the patient was cooled to 20–26°C, the bypass was restricted to 5–10 ml/kg/min. The arch vessels (obscure artery, left common carotid artery, and left clavicle artery) were clamped separately, and selective cerebral perfusion was started. The arch's sub-branches were clamped separately, and the antegrade selective cerebral perfusion was started via the right axillary artery. Under deep hypothermic circulatory arrest, total arch replacement was conducted using a four-branched graft to re-establish the aortic arch. First, the anastomosis between the tetrafurcate graft trunk distal and the descending thoracic aorta was made (for patients with dissection, the stent graft was inserted into the distal segment and deployed), restoring the blood supply to the descending thoracic aorta via graft perfusion (single pump dual vessel perfusion was adopted). Then, anastomoses between the left carotid artery, left subclavian artery, innominate artery, and the corresponding graft branches were achieved. Perfusion was restored every time an anastomosis was completed; and finally, the tetrafurcate graft trunk proximal was anastomosed to the ascending aorta in an end-to-end fashion.

Data collection

Patient's age, gender, comorbidities (hypertension, diabetes and coronary disease), preoperative complications, preoperative cardiac function, surgical pattern, history of aorta surgery, operation time, CPB time, clamping and circulation times, minimal temperature, intraoperative and postoperative blood transfusions, postoperative complication, mechanical ventilation time, intensive care unit (ICU) stay, hospital stay and death were retrieved.

Oxygenation index monitoring: Patient's blood gas and fraction of inspired oxygen (FiO₂) were measured and recorded within 24 h preoperatively and at 0, 1, 2 and 3 days postoperatively; PO₂/FiO₂ was calculated.

Systemic inflammatory indices monitoring: Patient's white blood cell (WBC), C-reaction protein and procalcitonin (PCT) were monitored within 24 h preoperatively.

Statistical analysis

SPSS 16.0 (IBM, Armonk, NY, USA) was used for statistical analysis. Continuous data are expressed as mean ± standard deviation, while categorical data are expressed as frequencies. Differences between groups for each characteristic were tested for significance with Fisher's exact test for categorical variables and t-test for continuous variables. Multivariable binary logistic regression analysis was used to identify independent prognostic factors using variables that were significantly (P < 0.05) associated with hypoxaemia in univariable analyses. A P-value of <0.05 indicated a significant difference.

RESULTS

Patients' baseline data and perioperative condition

Between October 2013 and December 2014, 181 patients were diagnosed with acute type A dissection and underwent the Sun's procedure at Anzhen Hospital, affiliated to Capital Medical University, China. Of these 181 patients, 21 were excluded. Among the 160 patients, 86 underwent ascending aorta replacement + Sun (scope of surgery: resection and replacement of ascending aorta and aortic arch), 66 underwent Bentall + ascending aorta replacement + Sun (replacement of the aortic valve, aortic root, ascending aorta and aortic arch), six underwent Sun (replacement of the aortic arch), and two underwent wheat + Sun (replacement of the aortic valve, ascending aorta and aortic arch). Among these patients, there were four cases of combined bypass surgery and 68 cases of combined valve replacement surgery. The patients were then allocated to the hypoxaemia group (n = 48) and non-hypoxaemia group (n = 112). In the preoperative baseline data, time from symptom onset to surgery was significantly correlated with hypoxaemia (P < 0.001). Age and gender were similar between the two groups (both P > 0.05). Comorbidities (hypertension, diabetes and coronary disease) and complications were similar between the two groups (all P > 0.05) (Table 1). There was no case of complete thrombosis in the false lumen. Among the 41 cases with partial thrombosis, 15 cases were in the hypoxaemia group and 26 cases were in the non-hypoxaemia group (P = 0.325).

Preoperative WBC count, incidence of preoperative hypoxemia and preoperative PCT levels in the hypoxaemia group were all significantly higher than in the non-hypoxaemia group (13.82 ± 3.85 vs 10.51 ± 2.72, P < 0.001; 79.2% vs 22.3%, P < 0.001; and 32.98 ± 12.17 vs 12.23 ± 14.03, P < 0.001, respectively). Emergency surgery rate and deep hypothermic circulatory arrest time were both significantly higher in the hypoxaemia group than in the non-hypoxaemia group (89.6% vs 59.8%, P < 0.001 and 31.29 ± 11.66 vs 27.19 ± 10.48, P < 0.030, respectively) (Table 2).

For postoperative indices, mechanical ventilation time, ICU stay, and hospital stay were all significantly longer in the hypoxaemia group than in the non-hypoxaemia group (P < 0.001) (Table 3). The difference in mortality between the two groups was not statistically significant.

Multivariable analysis for risk factors of hypoxaemia following surgical repair for acute type A aortic dissection

All factors associated with hypoxaemia in univariable analyses were included in the multivariable model (Table 4). Time from symptom onset to surgery ≤72 h [odds ratio (OR) 4.355; 95% confidential interval (CI) 1.57–12.082; P = 0.005], preoperative PaO₂/FiO₂ ≤300 (OR 15.30; 95% CI 15.30–5.52; P < 0.001), preoperative WBC count >15 000/μl (OR 1.352; 95% CI 1.139–1.604; P = 0.001) and deep hypothermic circulatory arrest time >25 min (OR 3.26; 95% CI 1.001–1.095; P = 0.043) were independently associated with the occurrence of hypoxaemia.

DISCUSSION

Reports show that acute respiratory distress syndrome (ARDS) is the primary cause of hypoxaemia [8, 9]. One of the diagnostic criteria of ARDS is an oxygenation index (PaO₂/FiO₂) ≤200 mmHg. However, due to the application of sternotomy and CPB for surgery of type A aortic dissection, it was difficult to use typical pulmonary infiltration X-ray images as the diagnostic criterion, and no diagnoses of ARDS or acute lung injury were made. The diagnostic criteria of ARDS were adopted and postoperative hypoxaemia was defined as PaO₂/FiO₂ ≤200 mmHg [10, 11].

Patients undergoing cardiac surgeries are at higher risk for postoperative hypoxaemia [8]. Indeed, it is reported that 12.2–27.1% of the patients have postoperative hypoxaemia, while 51% of those undergoing aortic repair for acute aortic dissection may suffer from hypoxaemia. In these previous studies, patients with many different types of aortic diseases undergoing various types of surgeries were included. On the other hand, in this study, only one type of patients was included and they all underwent a single type of surgery. Therefore, this study has a more homogeneous population of patients. Risk factors for postoperative hypoxaemia include old age, obesity, history of smoking, history of cardiac surgery, history of emergency surgery, reduced left ventricular ejection fraction, chronic pulmonary disease, myocardial infarction, diabetes, cardiac and non-cardiac pulmonary oedema, pneumonia, excessive blood transfusion and long CPB duration [12–14]. In this study, the risk factors for postoperative hypoxaemia were identified as: time from symptom onset to surgery ≤72 h, preoperative PaO₂/FiO₂ ≤300, preoperative WBC count >15 000/μl and deep hypothermic circulatory arrest time >25 min.

The incidence of postoperative hypoxaemia increased in patients with time from symptom onset to surgery ≤72 h and was possibly associated with the following process: during the early stage of dissection, aortic false lumen is formed and multiple components in the blood are massively activated, inducing systemic inflammatory response syndrome and coagulation system disorder, which further result in the sequential multiple organ injuries (lung injury is one of the main manifestations). This result indicated that early surgery could improve and stabilize haemodynamics, ameliorating the oxygenation function.

Studies showed that patients with a preoperative PaO₂/FiO₂ ≤300 mmHg have a significantly increased frequency of postoperative hypoxaemia. The acute onset of aortic dissection results in inflammatory cascades, which increase the alveolarcapillary membrane permeability and pulmonary vascular resistance, causing preoperative hypoxaemia. Moreover, the preoperative inflammatory state significantly impacts the intra- and postoperative degrees of systemic inflammation, indirectly raising the incidence rate of postoperative hypoxaemia. This is consistent with the result of this study that preoperative PaO₂/FiO₂ ≤300 mmHg is a risk factor for hypoxaemia.

Reports indicated that preoperative inflammatory state severely affects intraoperative systemic inflammatory degree during aortic dissection surgery [3, 12]. Inflammatory cascades release pro-inflammatory cytokines, leading to alveolar accumulation and activation of neutrophils and macrophages. Activated neutrophils release toxic mediators and proteolytic enzymes, which increase the permeability of endothelial and epithelial cells and pulmonary vascular pressures, affect the alveolar surfactant function, impair the oxygenation function, and cause postoperative hypoxaemia. Peripheral WBC count is an important index for inflammatory responses; therefore, the result of this study that preoperative WBC count is significantly correlated with postoperative hypoxaemia is consistent with earlier results.

Hypoxaemia is a common complication following cardiac surgeries. It is reported that CPB can significantly affect the occurrence of postoperative hypoxaemia [15–17]. Compared with other cardiac surgeries, standard type A aortic dissection surgery has a higher risk of postoperative hypoxaemia [16, 17], because of longer time of CPB and deep hypothermic circulatory arrest. During cardiac surgeries, studies showed that sternotomy with CPB results in systematic inflammatory responses, activates the complements, thrombin, cytokines, endothelin, endotoxin, neutrophils, adhesion molecules, macrophage and a variety of inflammatory mediators, and weakens the immune responses, causing multi-organ dysfunction [18–20].

Because CPB is a non-physiological circulation, it significantly alters the perfusion of peripheral tissues. In particular, prolonged perfusion will damage the permeability of capillary membrane, resulting in hypoperfusion, tissue hypoxia and pulmonary complications [21, 22]. Although deep hypothermia and antegrade selective cerebral perfusion can maintain a good cerebral function during cardiac arrest, hypoperfusion of other tissues lead to extensive ischaemia-reperfusion injuries. Deep hypothermia is an important technique universally used during surgery for acute Stanford A aortic dissection. It can slow the cellular metabolism and reduce cell damages during the circulatory arrest. However, deep hypothermia is involved in the activation of platelet and coagulation factor enzymes, which may cause haemorrhage and require subsequent blood transfusion. Excessive blood transfusion may result in transfusion-related acute lung injury, which is manifested as increased intrapulmonary effusion and poor oxygenation. Furthermore, coagulation factor loss, platelet destruction and microthrombosis during the transfusion may all lead to lung injuries, affecting the oxygenation function and result in hypoxaemia [23, 24]. It was also found that prolonged deep hypothermic circulatory arrest time was associated with postoperative hypoxaemia.

Some of the factors identified in the present study are non-modifiable. Nevertheless, improving oxygenation, administering preoperative anti-inflammatory treatment, adjusting the intraoperative and early postoperative mechanical ventilation strategy, and strengthening the management of respiratory system should be useful to decrease the occurrence of hypoxaemia. Additional management studies are necessary to confirm these points.

This study has a few limitations. First, it is a retrospective study and some data, such as the length of the dissection were not available. Second, patients were selected from the Department of Cardiac Surgery of Anzhen Hospital, China, i.e. a single centre, and the sample size is small. Nevertheless, this study included a single type of patients who all underwent a single type of surgery, increasing the homogeneity of the study population. A multi-centre study with a larger sample size is required to confirm the risk factors for hypoxaemia following surgical repair of acute type A aortic dissection so as to provide the guidance of early clinical intervention and lower the incidence rate of hypoxaemia following the surgical repair of acute type A aortic dissection.

CONCLUSIONS

Time from symptom onset to surgery (≤72 h), preoperative PaO₂/FiO₂ (≤300), preoperative WBC and deep hypothermic circulatory arrest time are independently associated with postoperative hypoxaemia in patients with acute aortic dissection. Early discovery, early diagnosis, early surgery, preoperative oxygenation improvement, inflammation control and cardiac arrest time reduction all contribute to the improvement of clinical prognosis.

Funding

National Natural Science Foundation of China (81000135). The research special fund for public welfare industry of health (201402009).

Conflict of interest: none declared.

REFERENCES

Author notes