PE symptoms, such as dyspnea, chest pain, haemoptysis, syncope, and arterial hypotension, are often poorly predicted. As the clinical manifestation is low in terms of uniqueness and sensitivity, especially for massive patients with unstable haemodynamics, the therapeutic window is very narrow. It has been reported that about 10% of patients with symptomatic PE die within 1 h of oneset . Therefore, rapid diagnosis is very important. CTPA, which has a sensitivity of 83% and a specificity of 96% , is the most common useful tool for rapid diagnosis and risk stratification.. CTPA provides a thrombosis clot location visualisation in the main pulmonary arteries and down to at least the segmental level. Acute PE was categorised as massive, submassive, and nor-massive PE. Massive PE  was defined as thrombosis clots in the main pulmonary artery as found by CTPA. In our study, 39 patients were diagnosed with CTPA, though 2 patients were diagnosed without CTPA because of continued CPR. If patients have severe haemodynamic instability or are unable to travel for CTPA, transthoracic echocardiography (TTE) should be considered a critical useful tool for diagnosis. TTE provides not only the location of the thrombus but also assesses the structure and function of the RV. In our study, all patients were diagnosed with PE by TTE.
The algorithm treatment for acute PE includes anticoagulation, systemic thrombolysis, catheter intervention and SPE..SPE was considered as a very dangerous treatment for PE due to its high operative mortality of 27.2% to 59% . However, Leacche et al  first reported that pulmonary embolectomy mortality was only 6%. The indication of SPE was extended and submassive PE with RVD accepted SPE treatment. The mortality rate of SPE has gradually decreased. Lehnert  reported SPE mortality rate of 6%, and more excellent results of SPE were reported [12,13,14]. Cho  reported that SPE had a lower cardiac mortality risk than thrombolysis in PE patients with haemodynamic stability. Increasingly good SPE results indicated that SPE should not only be a rescue therapy for patients when ST failed, but it should also be considered as the first-line PE treatment choice. However, there is still controversy regarding SPE outcomes. Park  reported a 14.8% surgical treatment mortality rate. Reza  recently reported 36 massive PE surgical treatments with 10 deaths and a mortality rate of 27.8%.
In our centre, there were 41 PE patients who underwent surgical embolectomy over 14 years, of whom 3 patients died postoperatively. The overall operative SE mortality rate was 7.32%. There were 39 PE patients who received SE as first-line treatment, of whom 1 patient died, and the mortality rate was 2.56% (1/39). The causes of death were operative haemorrhage in 2 patients and severe hypoxia in 1 patient. It was coincidental that both massive PE deaths were in patients who accepted ST as their first line of treatment but ST failed to work and then resorted SPE. ST seemed to increase the risk of operative haemorrhage and mortality of SPE. This finding was verified by the study of Aymard , who found that SPE had a lower early mortality rate (SPE: 3.6% versus ST 13.5%). whereas early mortality was 27% in those patients treated initially with thrombolysis and subsequently requiring SPE.. For massive PE patients with severe unstable haemodynamics or cardiac shock, if ECMO is unavailable and once ST treatment fails, patients would be faced with deteriorated haemodynamics and a high operative haemorrhage risk. In such case, SPE is a rescue treatment, and a higher SPE morbidity and mortality may occur. However,,SPE is the only choice of treatment when ST fails, if ECMO unvalaible, SPE should be carried out as rapidly as possible for sake of rescue life. A recent study  showed that patients treated with systemic thromnolysis have a higher cardiac mortality risk than SPE. Most PE patients without haemodynamic instability, who are afraid of surgical trauma and potential injury from cardiopulmonary bypass, are more likely to accept ST as a first-line treatment when there are no ST contraindications. Considering the favourable early and long-term outcomes in our study with no cases of chronic thromboembolic pulmonary hypertension, it is reasonable to consider SPE as an alternative therapy for submassive PE.
The most common risk factor for surgical treatment is preoperative cardiac arrest , which causes a mortality rate of up to 59%. Keeling  reported that the in-hospital mortality of patients with preoperative CPR was significantly higher (9/28,32.8%) than in those without CPR (16/186, 8.6%). Takahashi  reported that 73% of patients who received CPR for longer than 30 min died after pulmonary embolectomy. Recent reports showed that combining ECMO with SPE improved the outcome of massive PE surgery significantly, especially in massive PE with CPR . However, in developing countries, ECMO is not available in every cardiac centre. In this situation, further treatment for PE with CPR is very challenged. There were 3 patients who had preoperative cardiac arrest in this study, of whom 2 underwent SE as a first-line treatment even after continued CPR, which lasted more than 40 min, and achieved very good recovery without any brain damage. However, the other patient with successful 3 min of CPR accepted thrombolysis as first line treatment but failed, and resorted to SPE. This patient died of refractory surgical site bleeding. In the light of our experience, SPE seemed to be more effective in treatment PE with cardiac arrest when a surgical expertise team was available. SPE as an initial therapy can avoid refractory operative haemorrhage due to thrombolysis. In this study, 2 patients with long-duration CPRs achieved very good SPE recovery without any brain damage. The decisive factor was that cardiac arrest was witnessed by us on-site, and rapid effective CPR started on time and continued till to surgery. In fact, during the study period, there were several PE cases with successful CPR, who were transferred from another hospital to our centre; however, the pupillary light reflex disappeared and the pupils were dilated and fixed, indicating cerebral death. The patients were declined SPE because there was no possibility of cerebral resuscitation. For PE with long-duration CPR, further treatment selection depends on the assessment of cerebral resuscitation rather than CPR duration.
ECMO has been recommended as the standard of management in PE treatment centres . Acute massive PE with pre-operative significant haemodynamic instability with or without the need of CPR requires ECMO as an advanced life support. Postoperative sever hypoxemia and severe RVD may require ECMO support. In this study, 2 patients required ECMO support postoperatively. One patient had postoperative severe RVD who survived after 3 days of ECMO support. The other patient had severe hypoxaemia and died of multiple organ failure even after 1 week of ECMO support. Use of ECMO is associated with a high complication and the place of ECMO in the algorithm of PE treatment need to be further investigated .
Surgical techniques have improved over the past decades, and surgery under cardiopulmonary bypass with the”beating heart” technique is recommended as a standard protocol in new guidelines . Aortic cross-clamping causes ischemia–reperfusion myocardium injury and deteriorates RVD. In this study, there were 31 SPE surgeries performed without aortic cross-clamping. This may benefit RV function, and may be a reason for the low rate of postoperative ECMO support.
Massive lung haemorrhage was an SPE-related complication. There were 3 cases of massive lung haemorrhage postoperatively in the study. One patient died during the operation because of an uncontrolled large amount of bleeding in the lung. Two other patients survived. Massive lung haemorrhage is often due to injured pulmonary arterial vasculature. During surgery, in case of pulmonary massive haemorrhage, it is necessary to open the pleural cavity and locate the responsible pulmonary artery and repair it with a 6–0 polypropylene suture. In order to avoid damage to the infarcted smaller pulmonary artery, when removing the thrombus clot in the distal pulmonary arterial branches, it is forbidden to perform extraction without visualisation. It is difficult to extract smaller clots deep in the distal pulmonary arterial branches. We used heparin saline to vigorously irrigate the pulmonary arterial branches and aspirated them until flesh red blood flowed from the distal pulmonary arterial branch. The alternative option  was retrograde pulmonary venous perfusion with haparin saline, which was useful to remove a very small clot in the distal pulmonary arterial branches.