This survey indicates that most centers in Taiwan employ DLT for OLV. Arterial catheter and BIS are the common perioperative monitoring systems used in Taiwanese centers during thoracic surgery. 50% of these centers consider IVPCA for postoperative pain control. During OLV, most thoracic anesthesiologists recommended high oxygen fraction supplement (FiO2>80%) and ventilated the patients with a tidal volume of 6-8 ml/kg/PBW and a PEEP of 2-6 cmH2O using the dual-controlled mode. A PAP less than 30 cmH2O is considered the threshold to avoid barotrauma. Most thoracic anesthesiologists try to maintain relatively normal levels of CO2 and SpO2 during OLV. Manual hand squeezing method is more often used for lung recruitment at the end of operation. The Taiwanese thoracic anesthesiologists urge for an international practice guideline for protective lung ventilation during OLV.
Two other nationwide surveys were reported by the Italian and Taiwanese groups [1, 10]. These two retrospective studies found that more than 90% of the Italian and Taiwanese centers used a DLT as their first choice for intraoperative OLV (90-96%) [1, 10]. Consistent with these previous reports, our study also found that only 6.2% of the Taiwanese thoracic anesthesiologists would prefer to use a bronchial blocker for selective lung ventilation during thoracic surgery. A total of 39% of the Italian centers recommended epidural analgesia for postoperative pain management . Our study found that 43.8% of the participated centers performed loco-regional blocks for postoperative pain control. According to the American Pain Society clinical practice guidelines, thoracic epidural analgesia is considered as the most effective route for thoracic pain control and should be routinely considered for management of surgical pain after thoracotomy . In the recent two decades, minimally invasive video-assisted thoracoscopic surgery (VATS), which is associated with minimal tissue injury, have been widely adapted by the thoracic surgeons. Therefore, less invasive loco-regional techniques are the more favorable approaches than epidural analgesia for VATS . Since majority of the thoracic surgeries are currently underwent using the minimally invasive techniques in Taiwan , the fact that more than half of these centers used parenteral analgesic techniques instead of epidural analgesia as the first-line analgesia method is therefore reasonable. Nevertheless, loco-regional block techniques, such as paravertebral block, intercostal block and serratus anterior plane block, for perioperative pain control after VATS or other minimally invasive thoracic procedures should be vigorously promoted in the Taiwan medical institutes in order to enhance more effective postoperative pain relief, shift toward opioid-free analgesia, and prevent the development of chronic pain syndromes . Furthermore, direct comparisons of the anesthesia management for thoracic surgery between these studies might be inappropriate, as the standards of anesthesia care have changed after the introduction of enhanced recovery after surgery (ERAS) protocols  and other clinical pathways .
The main objective of this survey was to determine the strategies of ventilatory support during and after OLV. Compared with the Italian study reported 6 years ago, more Taiwanese thoracic anesthesiologists ventilated the patients using the dual-controlled ventilatory modes (PRVC or PCV-VG mode) during OLV, which might be a reflection of increased availability of these novel ventilatory modes in clinical anesthesia. These dual-controlled modes deliver the preset tidal volumes with lowest optimal airway pressure, which may theoretically reduce the risk of barotrauma . Although several clinical studies have suggested that dual-controlled modes enhanced oxygenation parameters with improved respiratory mechanics during OLV in general population and elderly [18-20], large-scale clinical trials are needed to confirm the overall pulmonary protective outcomes of the dual-controlled ventilatory modes during OLV and at the lung recruitment phase.
Low tidal volume (6-8 ml/kg PBW) is one of the hallmark parameters for intraoperative lung protective ventilation during non-thoracic surgery . However, the application of an “optimally low” tidal volume during OLV is not standardized. Our survey and other retrospective database analysis suggest that there are a considerably large proportion of patients continue to receive the similar range of tidal volume (6-8 ml/kg PBW) during two-lung and one-lung ventilation . However, the level of tidal volume has been shown to be inversely related to the incidence of respiratory complications and major postoperative morbidity . Furthermore, the Italian and Japanese anesthesiologists recommend a lower tidal volume (4-6 ml/kg PBW) for OLV [1, 22]. Nevertheless, opinions from the expert anesthesiologists highlight that protective ventilation in thoracic anesthesia is not simply synonymous of a low tidal volume, but also involves the appropriate application of PEEP, alveolar recruitment and other ventilatory settings during OLV [23, 24]. Most recently, a double-blind, randomized controlled trial conducted at the Samsung Medical Center (Seoul, Korea) demonstrated that driving pressure-guided ventilation (median ∆P of 9 cmH2O) during OLV significantly reduced the incidence of postoperative pulmonary complications compared with the conventional protective ventilation (tidal volume 6 ml/kg PBW, PEEP 5 cmH2O and recruitment) in thoracic surgery . PEEP is another important element in practicing intraoperative lung protective ventilation. This survey found that most of the thoracic anesthesiologists in Taiwan apply a PEEP level of 4-6 cmH2O during OLV, which is comparable with mean levels (4.2±1.6 cmH2O) reported in a large retrospective analysis of the US database . The authors concluded that low tidal volume failed to reduce postoperative pulmonary complications without application of adequate PEEP . A previous study also indicated that individualized PEEP determined by a PEEP decrement trial significantly increased oxygenation and lung mechanics than the standardized PEEP (5 cmH2O) . However, the appropriate PEEP levels for OLV are yet to be determined by the ongoing clinical trials (Table 2). Our study also found that most Taiwanese thoracic anesthesiologists currently use the bag squeezing maneuver to recruit of the collapsed non-dependent lung. Although the stepwise recruitment methods have been shown to reduce the incidence of postoperative pulmonary complications in comparison to bag squeezing maneuver in abdominal surgery , the evidence for re-expansion methods for the non-dependent lung after OLV requires further investigation. In fact, the ongoing Prothor and iPROVE-OLV trials are analyzing the lung protective effects of high PEEP, recruitment maneuver, and postoperative high-flow nasal cannulas for thoracic surgeries requiring OLV (Table 2) .
Collapse of non-dependent lung and atelectasis of dependent lung during OLV increases intrapulmonary shunt and leads to the development of intraoperative hypoxemia . Therefore, higher oxygen fractions are more commonly supplemented during lung separation procedures than the non-thoracic surgeries [22, 30]. However, oxygen therapy in clinical anesthesia is considered as a two-edged sword and excessive oxygen supplement should be avoided to prevent the potential oxygen toxicity , as potentially preventable hyperoxemia is considered as a SpO2 greater than 98%, despite a FiO2 of more than 0.21 . An observational study found that higher FiO2 during OLV was associated with significantly higher incidence of postoperative pulmonary complications (OR 1.30; 95% CI 1.04-1.65) . High quality-controlled studies are thus essential to compare the clinical outcomes of low versus high fractions of oxygen used for OLV.
Current clinical practice guidelines recommend that a tidal volume of 6-8 ml/kg predicted body weight and an optimal PEEP of 5 cmH2O are the most important ventilatory indices to guide intraoperative lung protection during mechanical ventilation in general population and obese patients . However, consensus on the individual ventilatory parameters that are considered as lung protective during OLV is still lacking. Our results found that the Taiwanese thoracic anesthesiologists concern that there is currently no common consensus in the intraoperative lung protective ventilation during thoracic surgery, particularly at the OLV phase. These anesthesiologists have diverse agreement to recommend the most important ventilator-derived parameters (i.e. tidal volume, PAP, PEEP and ∆P) for the guidance of lung protection during OLV (degrees of agreement: Vt >PAP >PEEP >FiO2 >∆P >mode >recruitment; Fig. 2). In fact, a number of prospective randomized controlled trials are currently undertaking, including several international multicenter studies, to determine the strategy for lung protective ventilation during thoracic surgery (Table 2).
There are a number of limitations with this study. First, the case scenario of lung tumor proposed in section 2 of the questionnaire specified that the lung resection was performed with VATS. Therefore, the data collected in the study may not be applicable to anesthesia and ventilatory care for patients receiving open thoracotomies. Secondly, this study analyzed the opinions of anesthesiologists in perioperative care for general patients. Patients with other underlying disease, such as chronic obstructive pulmonary disease, may need individualized ventilatory support strategy for thoracic surgery. Thirdly, all responses were based on the expert opinion or clinical experience of the participating thoracic anesthesiologists and could be subject to respondent bias. The optimal ventilatory settings or indices during OLV (e.g. Vt, PAP, PEEP and FiO2) suggested by the anesthesiologists could be arbitrary or not evidence based. Fourthly, this study primarily aimed to analyze the expert opinions of thoracic anesthesiologists on the anesthesia care and ventilatory support during thoracic surgery requiring OLV. Other unexpected perioperative adverse events such as surgical-related injury, severe bleeding, unstable hemodynamics, hypothermia, delirium or drug-responses  that could influence the general outcomes of thoracic surgeries were not taken into consideration in this report. Lastly, these results might not be completely representative of the expert opinions of all the Taiwanese thoracic anesthesiologists, as some medical centers and other regional hospitals were not included in our study.