手術條件下的靜脈輸液治療: 從基礎醫學研究到臨床醫學應用

全文

(1)國立臺灣師範大學生命科學系博士論文手術條件下的靜脈輸液治療: 從基礎醫學研究到臨床醫學應用 Intravenous Fluid Therapy for Surgical Conditions: From Bench to Bedside. 研究生：吳峻宇指導教授：鄭劍廷中華民國 105 年 12 月 15 日.

(2) I.

(3) 誌謝身為臨床醫師，處理病患時常常有無法在臨床上驗證的問題，這時往往需要在基礎科學領域上精進學習。承蒙老師們的指導，在我的主要研究領域－靜脈輸液治療上，有更深入的收穫，也影響了我在臨床上執行輸液治療的決策模式。在師範大學生命科學所的期間，特別感謝指導教授鄭劍廷老師與實驗室的夥伴們，每周的實驗室會議更在思考上，時時給我許多啟發。另還要感謝吳忠信老師、林豊益老師、李冠群老師、賴韻如老師、郭奇芊老師，在課業上許多的教導，讓我對於生命科學有更廣且更深的認識。臺大醫院麻醉部不但是我醫學訓練的起點與工作的場所，麻醉部裡的老師們對於我的研究的支持，更是不可或缺的支柱，深深地感謝!鄭雅蓉老師是我所有研究上的啟蒙導師，在住院醫師期間內訓練我完成研究上包含研究設計、臨床收案、統計分析，以及論文撰寫與發表的所有細節。在博士班期間的研究，老師也每每給予建議。另外，若不是葉育彰老師的教導，我不會知道微循環的重要性，也無法在這個領域得到任何進展。詹光政老師，在肝移植手術研究，以及其他麻醉學問的討論上，都給予我許多協助。臨床研究收案上，李宗勳老師、林佩玲老師、鄭孝良學長，也提供了許多的協助。研究助理劉家珍、吳宗錦與蔡靜影，幫忙了臨床研究和基礎研究上許多的繁務，若無她們協助，博士修業不可能順利。在博士班的學習期間，對於靜脈輸液治療這個領域有些進展，但博士班的結業，只是一個新的起點，未來仍有無邊無盡的問題需要探討與解決。希望能更深入的探討這個問題，也盡力協助團隊所有成員，在台灣的生命科學領域，做出一點點的貢獻! 最後要謝謝我的父母、兄長和妻子，從高雄到台北求學與工作已經超過十五年，你們一路來的支持與包容，是我所有工作與研究最大的動力。謝謝大家!. 2016 年 12 月吳峻宇謹誌. II.

(4) 中文摘要靜脈輸液可視為一種藥物治療，為改善低血容狀態下身體循環的第一線治療方式，因此可算是所有外科病人最常接受到的治療項目。適當的靜脈輸液應包含三個層面：正確的輸液種類、正確的輸液時機以及正確的輸液總量，這三個層面只要有其中一項沒有達到，靜脈輸液便可能無法發揮療效，甚至導致傷害。靜脈輸液治療的目的，在於改善低血容狀態(例如:出血性休克)下的循環狀態。目前文獻已經證明，微循環(小於 100 微米小血管)的改善，比傳統大循環系統指標(如:平均動脈壓、心跳數)，與預後有更高的關聯性。不同腹部臟器微循環，例如:肝、腎、腸組織，不但對低血容有不同的耐受度，也是病變發展成多器官衰竭的關鍵因子。治療低血溶狀態時，唯有改善腹部臟器微循環才可真正預防不可逆的器官傷害。因此本論文的第一部分前半段，乃是利用雷射斑駁影像技術，發展出可即時同步觀察在出血性休克時，腹部臟器微循環變化的大鼠模型。我們發現雖然微循環血氧濃度在各腹部臟器中，有等量下降的狀態，但以腸黏膜微循環血流量對於出血性休克有最差的耐受性，相對之下，肝腎與周邊肌肉組織的微循環血流量則有較佳的耐受性。而本論文的第一部分後半段，則希望進一步探討臨床上常使用的靜脈液體製品包含晶體溶液(如:0.9%食鹽水)、膠體溶液(如:澱粉製品與明膠製品)以及高張溶液(如:3%食鹽水)，在出血性休克狀態下，對腹部臟器微循環的療效差異。以不同種類靜脈輸液治療，不僅可能會造成不同程度的循環改善，也可能伴隨不同程度的缺血再灌流症狀而造成不等量的氧化壓力。在腹部臟器中，以腎臟有最明顯的缺血再灌流效應。臨床上，以治療膠體溶液治療重症病人的低血溶狀態，被發現有較高機率在治療後發生急性腎損傷，這個現象可能與缺血再灌流與氧化壓力有關。因為氧化壓力產物有極短的半衰期，因此我們利用活體自由基偵測技術，精確定量大鼠在出血性休克下與靜脈輸液治療後，腎組織活體自由基產生的變化。我們發現，以靜脈輸液治療出血性休克，晶體溶液無法改 III.

(5) 善腸黏膜微循環血流量，只有膠體溶液與高張溶液可有效改善腸黏膜微循環血流量。但是膠體溶液，包含澱粉製品以及明膠製品，相較於晶體溶液與高張溶液，會產生極大量的腎臟自由基，此現象極有可能與臨床文獻觀察到膠體溶液造成的急性腎損傷相關。總結本論文第一部分的前半與後半段，我們發現到腸黏膜微循環血流量對於出血性休克狀態下有最明顯的損傷，臨床的靜脈輸液治療僅有膠體溶液與高張溶液可以有效改善腸黏膜微循環血流量。但是使用膠體溶液治療時，有可能造成缺血性再灌流症狀而產生大量腎臟氧化壓力而導致急性腎損傷。臨床上，如何精確判斷適當的靜脈輸液量時機。目前臨床上最適合的輸液指標，當屬動態輸液反應性指標。這類指標可反映出在正壓呼吸下，心臟與肺部間相互影響下的變異率。體容越低的狀態下，心臟受到呼吸的吸氣期與吐氣期的擠壓影響變越大，反映在同一個呼吸週期下的脈搏壓、心輸出量的變異率變越大。因此當動態輸液反應性指標反映出變異率高過閾值時，給予靜脈輸液便有極高機率可提升心輸出量(此現象稱之為「輸液反應性」)，進而改善循環。但這類指標過去多半是在血管張力正常的病人中被應用，而臨床上有某一些特殊病理狀態會造成血管張力的異常，例如:肝硬化的患者，因為體內血管擴張物質無法被正常代謝，而有周邊血管阻力大幅下降的特徵。但是動態輸液反應性指標，卻從未在肝硬化病人身上證實有效。本論文的第二部分，探討我們在肝硬化病人，驗證目前臨床上最常見的三種動態輸液反應性指標，包括:脈搏壓變異率、心搏量變異率以及指端血容積變異率，是否在肝硬化病人也適用。結果我們發現，這三個指標在肝硬化病人身上，雖然較其他外科病人有較為下降的準確度，仍保有臨床實用的偵測度，足以幫助臨床醫師，判斷肝硬化病人是否可藉由靜脈輸液而改善心輸出量與循環。本論文的第三部分，希望探討如何判斷正確的靜脈輸液總量，而此議題針對不同的手術器官，可能存在不同的標準。因為若器官沒有缺血風險存在，則即使在靜脈輸液可提升心輸出量而改善循環的時機下，過量的輸液仍可能導致組 IV.

(6) 織水腫而造成傷害。以腦部手術而言，此類的臨床判斷充滿挑戰性。因為腦部組織佔總體心輸出量的比例極高，尤其在手術過程中可能有更高的血流量需求，但同時過多的輸液也可能因腦組織水腫而造成傷害。前述之動態輸液反應性指標，研究證實存在著一個「灰色地帶」，在此區間的下界限值與上界限值，分別代表了對輸液反應性高特異度與高敏感度的狀態，亦即手術中輸液若以下界限值與上界限值為目標，分別代表著儘量輸液增加心輸出量與謹慎(限制性)輸液以避免過多體液的輸液策略。依此特性，我們在腦部切除手術患者，驗證究竟腦部手術的靜脈輸液量較適合在動態輸液反應性指標灰色地帶的上界或下界。研究結果發現，靜脈輸液控制在下界的患者，有較短的加護病房停留時間、較少的術後神經學症狀、較佳的出院日常生活功能性評分、較少的術中乳酸堆積，以及較少的術後血清神經專一性蛋白表現。證實接受腦部手術患者，術中輸液量應以儘量增加心輸出量為目標，才是較佳的輸液策略。靜脈輸液治療在手術狀態下的重要性，在現代醫學的進步下，不但沒有減少，反而更加重要。總結本論文的發現，我們從基礎醫學的動物模型探討腹部臟器微循環在低血容下的病理變化，進一步證實高張溶液與膠體溶液對腸黏膜微循環的效果，並且發現膠體溶液對於腎臟缺血再灌流的現象可能與急性腎損傷相關。進一步到臨床應用上，證實動態輸液反應性指標此類參數，在肝硬化此類血管張力特殊變化的病人，仍然足夠精確幫助臨床醫師判斷靜脈輸液的時機點。最後，在腦部手術中，證實靜脈輸液量已儘量增加心輸出量的目標的策略，對於預後有較佳的效果。對於靜脈輸液治療中，「正確的輸液種類、正確的輸液時機以及正確的輸液總量」這三個層面，未來有更進一步機制的探討，將有機會更加增進手術患者的臨床照護。我們的研究從此觀點出發，未來，也將在此層面做出更深入的探討。關鍵詞: 腹部臟器微循環、膠體溶液、高張溶液、氧化壓力、肝硬化、動態輸液反應性指標、腦部手術。. V.

(7) Abstract Intravenous fluid therapy is considered the first-line drug treatment for hypovolemia, which is common in surgical patients. Appropriate intravenous fluid therapy comprises three aspects: the correct type of fluid, correct timing of fluid infusion, and correct amount of fluid infusion. If even one of these aspects is not achieved, intravenous fluid infusion may become injurious instead of being therapeutic. The aim of intravenous infusion therapy is to improve circulation during hypovolemia or hemorrhaging. Recent studies have demonstrated that microcirculatory (blood vessels <100 µm in diameter) improvement is more strongly associated with favorable outcomes than are traditional macrocirculatory indicators (e.g., mean arterial pressure and heart rate). Splanchnic microcirculatory impairment during hypovolemia may not only vary among organs (e.g., the liver, kidneys, and intestine) but is also associated with the development of multiple organ dysfunction syndrome. Only those fluid infusion therapies that improve splanchnic microcirculation can actually prevent irreversible organ damage during resuscitation from hypovolemic shock. Therefore, the first part of the first section of this dissertation addresses the development of a method for the real-time simultaneous observation of splanchnic microcirculatory changes during hemorrhagic shock by using laser speckle contrast imaging and tissue oxygen saturation in a rat model. We found that although the tissue oxygen saturation decreased homogeneously in multiple splanchnic organs, the microcirculatory blood flow was more vulnerable to hemorrhaging in the intestinal mucosa than in the liver, kidneys, and skeletal muscles. In the second part of the first section of this dissertation, we describe an exploration of the microcirculatory therapeutic effects of several common clinical intravenous fluid products, including crystalloid solutions (e.g., 0.9% saline), colloidal solutions (e.g., starch and gelatin products), and hypertonic solutions (e.g., 3% saline) during resuscitation from hemorrhagic shock. The use of different types of resuscitation fluids not only may result in variations in the extent of microcirculatory improvement but may also be associated with varying degrees of ischemia-reperfusion injury, expressed as oxidative stress. Among the abdominal organs, the kidneys are the most vulnerable to ischemia-reperfusion injury. In the clinical scenario, resuscitation from shock in critically ill patients by using colloidal fluids is associated with an increased risk of acute renal injury after treatment. This phenomenon is closely related to ischemia-reperfusion injury and oxidative stress. Because the products of oxidative stress have a very short half-life, we used an in vivo method to accurately quantitatively measure the amount of renal reactive oxygen species in rats during fluid resuscitation from hemorrhagic shock. We found that fluid resuscitation with a crystalloid solution could not restore microcirculatory blood flow in the intestinal VI.

(8) mucosa. Instead, only colloidal and hypertonic solutions improved the microcirculatory blood flow in the intestinal mucosa. However, colloidal solutions, including starch and gelatin products, produce a considerably larger amount of renal reactive oxygen species than do crystalloid and hypertonic solutions. This phenomenon may be associated with acute renal injury after colloidal fluid resuscitation, as reported in the clinical literature. In the first section of this dissertation, we conclude that the microcirculatory blood flow in the intestinal mucosa is most vulnerable to hemorrhagic shock. Furthermore, both colloidal and hypertonic solutions can restore the intestinal microcirculatory blood flow during hemorrhagic shock. However, fluid resuscitation with colloidal solutions may result in significant ischemic-reperfusion injury, which is indicated by the production of a large number of renal reactive oxygen species. The accurate determination of the time for administering an intravenous infusion is clinically challenging. At present, the most precise indicators of the time for administering fluid infusion may be dynamic fluid responsiveness parameters. These indicators reflect the interaction between the heart and lungs during mechanical ventilation. The lower the blood volume, the greater the variations in hemodynamic parameters during the respiratory cycle are. Therefore, when the dynamic fluid responsiveness parameters are higher than their threshold values, a fluid challenge is likely to considerably increase the cardiac output (this phenomenon is called “fluid responsiveness”), thereby improving the circulation. However, the threshold values of the dynamic fluid responsiveness parameters are based on patients with normal vascular tones. Certain patient groups are characterized by an abnormal vascular tone, for instance, patients with live cirrhosis have reduced vascular resistance because vasodilation-inducing substances cannot be metabolized as a result of end-stage liver disease. In the second section of this dissertation, we describe an investigation on the accuracy of three dynamic fluid responsiveness parameters, namely pulse pressure variation, stroke volume variation, and the plethysmographic variability index, in patients with liver cirrhosis. We found that although these three indicators were less accurate for predicting fluid responsiveness in common surgical patients, they were sufficiently precise to predict fluid responsiveness in patients with liver cirrhosis. Therefore, clinicians can determine the correct time for fluid administration to patients with liver cirrhosis using these indicators. In the final section of this dissertation, we clarify how to determine the correct amount of intravenous fluid for infusion, particularly with regard to vital organs. If a vital organ does not have the risk of ischemia and although intravenous infusion can improve cardiac output, excessive fluid infusion may lead to tissue edema and injury. For brain surgery, such a clinical judgment is challenging because the brain tissue has. VII.

(9) high metabolic and perfusion demands, particularly during surgery. However, excessive fluid infusion may also result in brain edema, poor outcomes, and neuronal injury. Dynamic fluid responsiveness parameters are characterized with a “gray zone.” In this interval, the lower and upper limit cutoffs respectively represent a high specificity and high sensitivity of fluid responsiveness. In an intraoperative fluid strategy targeting the upper cutoff of the gray zone, the fluid challenge highly specifically increases cardiac output without the risk of tissue edema, but the sensitivity is low; hence, a restrictive fluid balance may exist. By contrast, in a fluid strategy targeting the lower cutoff of the gray zone, a higher intraoperative cardiac output may be achieved after the fluid challenge, but the risk of tissue edema simultaneously increases. With regard to brain surgery, we investigated whether different amounts of fluid infused during intraoperative fluid challenges favored the upper or lower cutoffs in the gray zone. The results revealed that patients receiving a fluid strategy targeting the lower cutoff of the gray zone had a shorter intensive care unit stay, fewer postoperative neurological events, a superior discharge functional status, less intraoperative lactate accumulation, and lower postoperative serum neuronal injury protein expression than did patients receiving a fluid strategy targeting the upper cutoff of the gray zone. We confirmed that for patients undergoing brain surgery, the intraoperative fluid strategy should aim to maintain and increase cardiac output as much as possible. The importance of intravenous fluid therapy to surgical conditions is being increasingly emphasized in modern medicine. In summary, we investigated the pathological microcirculatory changes in multiple splanchnic organs during hemorrhagic shock in a rat model and confirmed the therapeutic effects of hypertonic saline and colloidal solutions on the restoration of intestinal mucosal microcirculatory blood flow. In addition, in vivo induction of renal reactive oxygen species by colloid solutions is reported for the first time, and this phenomenon may be associated with acute kidney injury in critically ill patients after fluid resuscitation with colloidal solutions. For further clinical applications, we observed that dynamic fluid responsiveness parameters enable clinicians to accurately determine the timing of fluid infusion not only for general surgical patients but also for patients with liver cirrhosis with altered vascular tones. Finally, we confirmed that an intraoperative fluid strategy targeting the elevation of cardiac output may be more beneficial than a restrictive strategy to patients undergoing brain tumor resection surgery. As long as the fluid therapy is based on the concept of administering the correct type of fluid at the correct time and in the correct amount, opportunities to further improve the clinical care of surgical patients are numerous. Our research begins from this perspective, and we would like to further. VIII.

(10) investigate fluid therapy. Key words: splanchnic microcirculation、colloid solution、hypertonic solution、 oxidative stress、liver cirrhosis、dynamic fluid responsiveness parameter、brain surgery. IX.

(11) Abbreviations: ANOVA BIS BNP CL CVP FR GDFT GEL GFAP HES HTS LSCI MAP MCLA NS NSE PI PPV POD PVI PU ROI ROS SV SVI SVR SVRI SVV. Analysis of variance bispectral index B-type natriuretic peptide Chemiluminescence Central venous pressure Fluid responsiveness Goal-directed fluid therapy Succinylated gelatin Glial fibrillary acidic protein hydroxyethyl starch Hypertonic saline Laser speckle contrast imaging Mean arterial pressure 2-methyl-6-[4-methoxyphenyl]-3,7-dihydroimidazo[1,2-a]-pyrazin-3-one hydrochloride 0.9% saline Neuron-specific enolase Perfusion index Pulse pressure variation Postoperative day Plethysmographic variability index Perfusion unit Region of interest Reactive oxygen species Stroke volume Stroke volume index Systemic vascular resistance Systemic vascular resistance index Stroke volume variation. X.

(12) 目錄. 目錄國立師範大學博士學位論文口試委員會審定書 ................................................................ ..I 誌謝 ..........................................................................................................................................III 中文摘要 ..................................................................................................................................III Abstract ................................................................................................................................... IV Abbreviations: ........................................................................................................................... X 目錄 ......................................................................................................................................... XI Chapter 1、Introduction............................................................................................................. I Section 1：Introduction of perioperative fluid therapy for hypovolemia and hemorrhagic shock.......................................................................................................................................1 Section 2：Effects of intravenous fluid therapy on the splanchnic microcirculation ............ 2 Section 3：Hemodynamic assessment of fluid status in liver cirrhosis patients ...................7 Section 4： Considerations of fluid strategy for patients undergoing brain surgery .......... 10 Chapter 2 Materials and Methods.............................................................................................12 Section 1：Splanchnic microcirculatory changes during hemorrhagic shock and resuscitation in a rat model ...................................................................................................12 Section 2： Predicting stroke volume and arterial pressure fluid responsiveness in liver cirrhosis patients by using dynamic preload variables .........................................................22 Section 3：Comparison of two stroke volume variation-based goal-directed fluid therapies for supratentorial brain tumour resection ............................................................................. 26 Chapter 3 Results...................................................................................................................... 34 Section1：Splanchnic microcirculatory changes during hemorrhagic shock and resuscitation in a rat model ...................................................................................................34 Section 2：Predicting stroke volume and arterial pressure fluid responsiveness in liver cirrhosis patients by using dynamic preload variables .........................................................41 Section 3：Comparison of two stroke volume variation-based goal-directed fluid therapies for supratentorial brain tumour resection ............................................................................. 46 XI.

(13) Chapter 4. Discussion ...............................................................................................................50 Section1：Splanchnic microcirculatory changes during hemorrhagic shock and resuscitation in a rat model ...................................................................................................50 Section 2：Predicting stroke volume and arterial pressure fluid responsiveness in liver cirrhosis patients by using dynamic preload variables .........................................................66 Section 3：Comparison of two stroke volume variation-based goal-directed fluid therapies for supratentorial brain tumour resection ............................................................................. 72 References ................................................................................................................................ 78 Figures ...................................................................................................................................... 90 Figure 1、Animal model of splanchnic microcirculation. ...................................................90 Figure 2. The timeline of animal study I: the protocol of hemorrhagic shock and saline resuscitation.. ........................................................................................................................91 Figure 3. Timeline of the protocols for animal study II- the experiment for assessing splanchnic organ microcirculation and renal reactive oxygen species formation ................92 Figure 4. CONSORT flow diagram of the randomized controlled trial enrolment of orthotopic liver transplantation.............................................................................................93 Figure 5. Protocols for low SVV and high SVV goal-directed fluid therapies during supratentorial brain tumor resection.. ...................................................................................94 Figure 6. Laser speckle contrast imaging of the microcirculatory blood flow intensities during hemorrhagic shock and saline resuscitation. ............................................................. 95 Figure 7. Percent changes of microcirculatory blood flow intensity and tissue oxygen saturation. .............................................................................................................................96 Figure 8. Example of laser speckle contrast imaging of the microcirculatory blood flow intensity. ...............................................................................................................................97 Figure 9. Percent changes in the microcirculatory blood flow intensity at T1 and T2 compared with T0. ................................................................................................................98 Figure 10. Comparison of the amount of reperfusion-induced in vivo renal reactive oxygen species formation after fluid resuscitation. ...........................................................................99 Figure 11. Flowchart for orthotopic liver transplantation patient recruitment. ..................100. XII.

(14) Figure 12. Receiver operating characteristic curves describing the ability of pulse pressure variation (PPV), stroke volume variation (SVV), and the plethysmographic variability index (PVI). ........................................................................................................................101 Figure 13. CONSORT flow diagram..................................................................................102 Figure 14. Perioperative changes in serum neuronal biomarker levels. .............................103 Tables .....................................................................................................................................105 Table 1. Macrocirculatory changes secondary to hemorrhagic shock ................................105 Table 2. Splanchnic organ microcirculatory blood flow intensity changes secondary to hemorrhagic shock ............................................................................................................. 106 Table 3. Splanchnic organ tissue oxygen saturation changes secondary to hemorrhagic shock...................................................................................................................................107 Table 4. Microcirculatory blood flow intensity changes after hemorrhagic shock and fluid resuscitation in part I experiment .......................................................................................108 Table 5. Macrocirculatory changes after hemorrhagic shock and fluid resuscitation in part I experiment ..........................................................................................................................110 Table 6. Arterial blood gas analysis in part I experiment ...................................................111 Table 7. Orthotopic liver transplantation patients’ characteristics .....................................113 Table 8. Changes in hemodynamic parameters before and after fluid loadings in stroke volume fluid responders and non-responders .....................................................................114 Table 9. Changes in hemodynamic parameters before and after fluid loadings in mean arterial pressure fluid responders and non-responders .......................................................116 Table 10. Agreements between three dynamic preload variables before and after fluid challenges ...........................................................................................................................118 Table 11. Patient characteristics .........................................................................................119 Table 12. Intraoperative profiles ........................................................................................120 Table 13. Postoperative outcomes and neurological events ...............................................123 附錄 ....................................................................................................................................124. XIII.

(15) Chapter 1、Introduction Section 1 、 Introduction of perioperative fluid therapy for hypovolemia and hemorrhagic shock 1.1 Introduction of perioperative fluid management Fluid therapy is often the first-line treatment in the perioperative medicine because hypovolemia and hemorrhaging are the most common etiology of hemodynamic instability in surgical patients. To optimize cardiac preload, it is keen to anesthesiologist to administrate fluid with the right kind of fluid in appropriate amounts at the right time (Chappell et al., 2008). Regarding the right kind of fluid to be administrate in hypovolemic or hemorrhaging patients, there is a paradigm shift from macrocirculation to microcirculation, particularly in the splanchnic organs, which is more relevant to patient outcomes and the development of multiple organ failure syndrome (De Backer et al., 2013). The pathophysiological changes in splanchnic microcirculation during hemorrhaging and effects of fluid resuscitation on the recovery of splanchnic microcirculation remain less investigated. Regarding the appropriate amount of fluid requirement and right timing of fluid administration, there is more and more evidence indicating that perioperative fluid therapy based on a predefined hemodynamic goal (goal-directed fluid therapy; GDFT) is more favorable to conventional fluid strategy and it also improves postoperative outcomes (Rollins and Lobo, 2015, Pearse et al., 2014, Ebm et al., 2014, Benes et al., 2014). Perioperative GDFT is proved beneficial to patients undergoing major abdominal surgery and it is usually conducted based on cardiac output monitoring or dynamic preload parameters, which are currently the most precise indicators predicting fluid responsiveness after fluid therapy (Rollins and Lobo, 2015, Benes et al., 2014, Corcoran et al., 2012). Because the dynamic fluid responsiveness parameter is based on the cyclic changes in pulse pressure, stroke volume or plethysmographic amplitude 1.

(16) from heart-lung interaction (Michard, 2005), its accuracy is less clarified in special patients with altered vascular tones such as liver cirrhosis and the effect of GDFT on other surgical patient group, such as patients undergoing brain surgery remains unknown.. 2.

(17) Section 2、Effects of intravenous fluid therapy on the splanchnic microcirculation 2.1 Splanchnic microcirculatory changes during hemorrhagic shock Hemorrhagic shock is one of the major etiologies of mortality in traumatic injury. Both macrocirculatory and microcirculatory dysfunctions have been characterized in the acute phase of hemorrhagic shock. For instance, Dubin and colleagues reported that macrocirculation, sublingual and intestinal microcirculation were dissociated during hemorrhaging; the mean arterial pressure (MAP) and arterial pH were significantly changed only at the final stage of bleeding. By comparison, microcirculation decreased at the first stage of bleeding (Dubin et al., 2009). This indicates that the stabilization of macrocirculatory hemodynamic parameters, such as the MAP, is likely to occur at the expense of splanchnic microcirculatory perfusion during hemorrhagic shock. This phenomenon may occur because blood flow and tissue oxygenation of nonvital organs decrease to maintain the circulation required by vital organs during hemorrhaging. The splanchnic microcirculatory impairment is also associated with the development of multiple organ dysfunction syndrome (De Backer et al., 2013). However, the acute changes of microcirculation blood flow and tissue oxygen saturation in multiple splanchnic organs secondary to hemorrhagic shock are insufficiently clear. Understanding the microcirculatory changes of splanchnic organs during a hemorrhage and resuscitation may provide crucial information for further research and treatment to prevent multiple organ dysfunction syndrome (Boldt and Ince, 2010).. 3.

(18) 2.2 Effects of intravenous fluid therapy on splanchnic microcirculation and reperfusion-induced reactive oxygen species formation Intravenous fluid is mandatory in treatment of hypovolemia and hemorrhaging. However, even when a sufficient amount of fluid is administered for restoring macrocirculatory hemodynamic stability, splanchnic organ injury may persist. This may be because different types of resuscitation fluids may affect the recovery of microcirculatory blood flow and reperfusion-induced reactive oxygen species (ROS) formation in a varied extents (Boldt and Ince, 2010, Rushing and Britt, 2008). During resuscitation of shock, outcomes in organ function is more strongly correlated with microcirculatory improvement than macrocirculatory improvement (Pranskunas et al., 2013). Accordingly, numerous clinical investigations have been conducted to investigate the microcirculatory effects of different types of resuscitation fluid, including crystalloids, hypertonic saline (HTS), and synthetic colloids, by observing sublingual microcirculation (Bezemer et al., 2012, Boldt and Ince, 2010). However, because splanchnic microcirculation is compromised during hypovolemia, which participates in the development of multiple-organ dysfunction syndrome (Ceppa et al., 2003), and the splanchnic microcirculatory response to fluid challenge may become dissociated from the sublingual microcirculatory response (Edul et al., 2014), the effects of different types of fluid on the recovery of splanchnic microcirculation during resuscitation from hemorrhagic shock remain unclear. Therefore, a technology to simultaneously monitor multiple splanchnic organs, including liver, kidney and intestine, is desirable. In addition to microcirculatory change, reperfusion injury after fluid resuscitation is another factor influencing organ dysfunction. The kidney is one of the most vulnerable splanchnic organs targeted in reperfusion-mediated oxidative tissue injury (Devarajan, 2006). ROS formation is an early indicator of reperfusion-induced 4.

(19) oxidative stress and may be detectable in the acute phase of fluid resuscitation. Excess ROS formation is associated with systemic inflammation and can initiate cell death (Rushing and Britt, 2008); in addition, it is closely correlated to renal injury (Walker et al., 2001). The amounts of ROS formation after reperfusion may depend on the type of resuscitation fluid used (Rushing and Britt, 2008), and fluid resuscitation using synthetic colloids is relevant to acute kidney injury (Perel et al., 2013, Meybohm et al., 2013, Thomas-Rueddel et al., 2012). However, renal ROS formation during the acute phase of fluid resuscitation using synthetic colloids is less thoroughly investigated compared with that using other types of resuscitation fluids. Therefore, we also aimed to investigate the extent of renal ROS formation of different types of resuscitation fluid during acute management of hemorrhagic shock by using in vivo renal ROS detection technique.. 5.

(20) 2.3 Technology to monitor multiple splanchnic microcirculatory changes during hemorrhaging and fluid resuscitation Laser speckle contrast imaging (LSCI) is an increasingly prevalent technique for monitoring microcirculatory blood flow. Because it enables full-field imaging in near real time with multiple regions of interest, it is suitable for investigating microcirculatory changes among multiple organs (Ding et al., 2014, Boas and Dunn, 2010). LSCI in combination with tissue oxygen saturation measurements offers a comprehensive understanding of acute microcirculatory changes among multiple organs. The additional advantages of using LSCI to monitor splanchnic microcirculatory changes during hemorrhaging and fluid resuscitation are, firstly, a larger ROI can be set, avoiding the inaccurate results that can occur when choosing an inappropriately small region for monitoring. Previous studies have reported that LSCI reduces intersite and interindividual variabilities and may provide a comparable or even improved reproducibility of microcirculation in comparison with other techniques such as laser Doppler flowmetry or sidestream dark-field imaging (Sturesson et al., 2013, Tew et al., 2011, Rousseau et al., 2011, Roustit et al., 2010); secondly, the microcirculatory blood flow of viscera is likely obstructed by the physical attachment of the probes used in other techniques. LSCI is noninvasive, eliminating artifact contact (Draijer et al., 2009). In our animal studies, we firstly investigated the heterogeneity of microcirculatory responses to hemorrhagic shock among multiple splanchnic organs and the gracilis muscle by using both LSCI and tissue oxygen saturation measurements. Secondly, we aimed to investigate the effects of different resuscitation fluids on the restoration of microcirculation in multiple splanchnic organs, using the LSCI technique and determine the renal reperfusion injury-induced reactive oxygen species (ROS) formation by using an in vivo ROS detection technique. 6.

(21) Section 3、Hemodynamic assessment of fluid status in liver cirrhosis patients 3.1 Hemodynamic alterations in liver cirrhosis patients Liver cirrhosis patients present unique hemodynamic characteristics such as hyperdynamic circulation, low systemic vascular resistance (SVR), high arterial compliance and peripheral vasodilation (Henriksen et al., 2006), which often interfere with hemodynamic monitoring. Because the fluid challenge remains one of the most common treatments for improving blood flow and maintaining arterial pressure. Therefore, it is highly clinical relevant for a caregiver to predict whether a fluid challenge will elicit ‘fluid responsiveness’ in stroke volume (SV) and arterial pressure (Marik et al., 2011). Compared with conventional static fluid variables, dynamic fluid variables, which are derived from SV or pulse pressure fluctuations secondary to changes in intrathoracic pressure during mechanical ventilation, have been consistently reported to be superior predictors of SV fluid responsiveness in a lot of surgical settings (Yin and Ho, 2012, Marik et al., 2009).. 7.

(22) 3. 2 Predicting stroke volume and arterial pressure fluid responsiveness in liver cirrhosis patients by dynamic fluid variables Dynamic fluid variables are either pressure-based [(pulse pressure variation (PPV)], flow-based [SV variation (SVV)] or volume-based [plethysmographic variability index (PVI)]. For a given ventilator-induced change in SV, these three types of fluid variables may vary according to vascular compliance at the measurement level. However, the predictive ability of each dynamic preload variable is highly dependent on the reflection of normal arterial system compliance (Monnet et al., 2013, Henriksen et al., 2001). Therefore, the accuracy of PPV, SVV and PVI to predict SV fluid responsiveness in liver cirrhosis patients remains unclear. In addition to SV optimization, arterial blood pressure optimization is also a critical component of hemodynamic stability, and differentiating arterial vasodilatation from hypovolemia as the cause of hypotension is crucial; hence, a variable closely representing vascular tone is desirable. SVR is the variable most commonly used to describe vascular tone but SVR may not facilitate treating liver cirrhosis patients because this patient group often presents with a low SVR. The perfusion index, which is calculated as the ratio of pulsatile to nonpulsatile blood volumes and automatically calculated using a pulse oximeter, is a measure of the vascular tone in peripheral tissue (Lima et al., 2002) and has been reported to predict the incidence of spinal anesthesia-induced hypotension during caesarean section in parturient women (Toyama et al., 2013), which is another population characterised by a low SVR. Recently, in patients undergoing mechanical ventilation, including critically ill patients (Monge Garcia et al., 2011), and in patients undergoing hepatic surgery (Vos et al., 2013), the PPV/SVV ratio has been proposed to reflect dynamic arterial elastance, thus facilitating the prediction of arterial pressure responsiveness to fluid, when such patients experience a fluid challenge. To date, no study has 8.

(23) investigated the ability of these variables of vascular tone to predict arterial pressure fluid responsiveness in liver cirrhosis patients. In our first clinical study, we investigated the ability of dynamic fluid variables, namely, PPV, SVV and PVI, to predict SV fluid responsiveness and assessed the ability of vascular tone variables, namely, the SVR index (SVRI), perfusion index and PPV/SVV ratio, to predict arterial pressure responsiveness to fluid in liver cirrhosis patients. This may deliver important information regarding the implication of these variables to hemodynamic assessment of other patients with altered vascular tones (low SVR), such as sepsis or pediatric patients.. 9.

(24) Section 4、Considerations of fluid strategy for patients undergoing brain surgery Intraoperative goal-directed fluid therapy (GDFT) during brain surgery has not been investigated. During GDFT, repeat fluid boluses to achieve the near-maximal stroke volume showed improved outcomes compared with the conventional fluid regimen (Corcoran et al., 2012). However, recent studies have shown neutral or detrimental outcomes of GDFT compared with protocols with restrictive fluid balance (Phan et al., 2014, Srinivasa et al., 2013, McKenny et al., 2013, Brandstrup et al., 2012). Although fluid administration increases cerebral blood flow, even when blood pressure is within the range of autoregulation (Meng et al., 2015), an elevated net fluid balance is associated with an increase in serum brain natriuretic peptide (BNP) and a decrease in serum osmolality, and results in neuronal injury (Ertmer and Van Aken, 2014) and poor clinical outcomes (Kissoon et al., 2015). Optimal volume status for brain surgery remains uncertain. Another GDFT protocol based on dynamic parameters such as stroke volume variation (SVV), reduces the postsurgical morbidity and length of intensive care unit (ICU) stay (Benes et al., 2014). SVV is a useful predictor of fluid responsiveness for several settings (Slagt et al., 2014), including during brain surgery (Berkenstadt et al., 2001) and while in the prone position (Biais et al., 2010). However, there is a “grey zone” of two cut-offs within which the SVV validity is inconclusive (Cannesson et al., 10.

(25) 2011). Accordingly, fluid boluses based on the lower SVV cut-off exclude fluid responsiveness at near certainty, resulting in a higher cardiac output and a risk of positive net fluid balance, whereas those based on the higher SVV cut-off include fluid responsiveness at near certainty, resulting in a restrictive fluid balance. In this study, we hypothesized that fluid boluses based on the two SVV cut-offs in the grey zone during GDFT were feasible and result in different postoperative outcomes and varying severity of neuronal injury in patients undergoing elective supratentorial brain tumor resection.. 11.

(26) Chapter 2 Materials and Methods Section 1：Splanchnic microcirculatory changes during hemorrhagic shock and resuscitation in a rat model 1.1 Animal study I: Microcirculatory changes in multiple splanchnic organs during hemorrhagic shock and saline resuscitation. Experimental animals A total of 72 male Wistar rats were used (body weight 250 ± 50 g; Biolasco Taiwan Co., Taipei, Taiwan) in this experiment. The rats were raised on a 12-h light–dark cycle and had free access to water and food. Before conducting the experiments, all experimental procedures were approved by the Institutional Animal Care and Use Committee of National Taiwan University and performed in accordance with its guidelines.. Anesthesia and surgical procedure The rats were anesthetized and prepared as described in our previous study (Yeh et al., 2012). A tracheostomy was performed, and then a 14G catheter (Surflo; Terumo Corporation, Laguna, Philippines) was inserted into the trachea to maintain spontaneous breathing. The anesthesia was maintained using 1.2% isoflurane through the tracheal tube. Subcutaneous atropine 0.05 mg/kg in 10 mL/kg of saline was injected for replacement of water evaporization from surgical open wound and to prevent airway secretion. The body temperature was continuously monitored rectally, and a warmer pad was applied to maintain the rectal temperature between 36 and 37 °C. Polyethylene catheters (PE-50; Intramedic 7411, Clay Adams, Parsippany, NJ, USA) were inserted into the right common carotid artery and external jugular vein. The right common carotid artery catheter was used to continuously monitor the MAP and heart rate (HR), and to withdraw blood to establish the state of hemorrhagic 12.

(27) shock. The external jugular vein catheter was used to infuse resuscitation fluid. A long midline laparotomy was performed to expose splanchnic organs including the liver, left kidney, and a segment of the terminal ileum (approximately 6 to 10 cm proximal to the ileocecal valve). A 2-cm section was incised on the antimesenteric aspect of the intestinal lumen by using a high-frequency desiccator (Aaron 900; Bovie Aaron Medical, St. Petersburg, FL, USA) to carefully expose the opposing mucosa for Examination of the microcirculatory blood flow. Nearby intestinal serosal muscular layer (at the midline of antimesenteric aspect) and the central Peyer's patch (identified by visualize the lymph nodes) were also identified for examination of the microcirculatory blood flow. Moreover, the right gracilis musclewas exposed for measuring microcirculatory changes relative to those of the splanchnic organs (Fig 1). The exposed viscera and tissue were kept moist hourly with saline (37 °C 0.5 mL of aerosolized 0.9% saline).. Hemorrhagic shock and fluid resuscitation protocol After completion of the surgical procedures, the animals were allowed to stabilize for 30 min before the baseline hemodynamic measurements were performed (baseline condition was considered stable when all measurement values remained at 10% for 15 min; T0). After the baseline hemodynamic measurements were collected, the concentration of inhaled isoflurane was decreased to 0.7% to prevent over anesthesia for hemorrhaging animal without further surgical stimulation, and the rats were randomly assigned to either a sham operation (S) group, hemorrhagic shock (H) or fluid resuscitation with 0.9% saline (R) group. In the H group, hemorrhagic shock was initiated through controlled blood withdrawal via the right carotid arterial cannula (3 times of 10 mL/kg per 5 min; total blood loss of 30 mL/kg during 15 min). Further macrocirculatory and microcirculatory monitoring were measured according to the 13.

(28) time points shown in Fig. 2. In the S group, the rats received the same surgical preparation without undergoing blood withdrawal. The R group was resuscitated by a total of 30 mL/kg of 0.9% saline after hemorrhagic shock for 30 min.. Evaluation of splanchnic microcirculatory blood flow and oxygen saturation changes secondary to hemorrhagic shock A full-field laser perfusion imager (MoorFLPI; Moor Instruments, Ltd., Devon, UK) was used from the baseline (T0) to continuously quantify microcirculatory blood flow intensity among multiple splanchnic organs. The imager uses LSCI, which exploits the random speckle pattern that is generated when tissue is illuminated by a laser light. The random speckle pattern changes when blood cells move within the region of interest (ROI). When the level of movement is high (high flow), the changing pattern becomes more blurred, and the contrast in that region decreases accordingly. Therefore, low contrast represents high flow and high contrast represents low flow. The contrast image is processed to produce a 16-color coded image that correlates with blood flow in the tissue (e.g., blue is defined as low flow and red as high flow). The microcirculatory blood flow intensity of each ROI was recorded as a perfusion unit (PU), which is related to the product of the average speed and concentration of the red blood cells moving in the tissue sample volume (i.e., blood cell flux or perfusion). The images were recorded and analyzed in real time by using the MoorFLPI Version 3.0 software (Moor Instruments, Ltd.). Six separate ROIs were established on the liver, left kidney, mucosa, serosal muscular layer, Peyer's patch, and right gracilis muscle. Details of selecting and setting ROIs were as following: 1.. The left liver is identified by the hepatic fissure and falciform ligament. The hepatic ROI was set 1.5 cm left to the hepatic fissure.. 2.. The left kidney is easily identified by direct visualization. 14.

(29) 3.. The selected terminal ileum is about 6 to 10 cm proximal to the ileocecal valve in each animal. The length of section was fixed in each animal (2-cm; performed on the antimesenteric aspect of the intestinal lumen). In addition, the choice of specific interested region of intestinal tissue was the same for each of the rats as the below: First, the ROI of mucosa was set to cover as much of the mucosa as possible. The advantage of using a high-frequency desiccator to expose the intestine is that it is able to confine the surgical trauma to small and restricted area. Accordingly, the ROI of the mucosa was carefully selected not to cover the outmost margin. The choice of intestinal serosal muscular layer was at the midline of antimesenteric aspect (one to 2 cm away from the mucosa). The Peyer’s patch was identified by visualizing the clusters of lymph nodes.. 4.. This multiple organ model is involved with some stretching forces to fully expose the terminal ileum. In addition, the mesenteric perfusion has a gray zone that the distal portion of supplied tissue is more susceptible to stretch-induced decrease of blood flow. Therefore, the region of interests were selected in the central portion of exposed tissue and we were carefully to avoid covering the ROIs over the outmost margin of each tissue.. 5.. The skin (2 cm incision) to expose the gracilis muscle was about 3 cm lateral to the midline of abdomen (the midline could be identified by the xiphoid bone and anus).. 6.. The ROI was set up as a square with the length approximately 1.0 to 1.5 cm to cover as much of the exposed tissues as possible. The tissue oxygen saturation was measured using a white light spectroscopy. needle probe (moorVMS-LDF2; Moor Instruments, Ltd.) with a white LED for illumination, which emitted light with wavelengths between 450 and 700 nm (Liu et al., 2011). To minimize sampling bias, tissue oxygen saturation was measured and 15.

(30) recorded as an average of 5 points in each exposed target organ. The microcirculation, including blood flow intensities and tissue oxygen saturation between the 2 groups, were compared at the following time points: 0 min (baseline; T0), 75/105 min (1 h after shock or resuscitation; T1), 135/165 min (2 h after shock or resuscitation; T2), and 195/225 min (3 h after shock or resuscitation; T3) (Fig. 2). Microcirculatory blood flow intensity was recorded as an arbitrary PU, and the tissue oxygen saturation was recorded as a percentage (%). Because different organs may have different baseline states of both macrocirculatory and microcirculatory blood flow intensity and tissue oxygen saturation, the percent changes of the MAP, microcirculatory blood flow intensity, and tissue oxygen saturation at T1, T2, and T3 were compared to the T0 baseline values. The total 72 rats in the S, H and R groups were equally assigned to the following 3 subgroups: (1) measurements of changes in microcirculatory blood flow intensity; (2) measurements of tissue oxygen saturation in the intestinal mucosa, serosal muscular layer, Peyer's patch, and right gracilis muscle; and (3) measurements of tissue oxygen saturation in the liver and kidney.. 16.

(31) 1.2 Animal study II: Effects of different types of fluid resuscitation for hemorrhagic shock on splanchnic organ microcirculation and renal reactive oxygen species formation. Experimental animals A total of 96 male Wistar rats (body weight = 250 ± 30 g; Biolasco Taiwan Co., Taipei, Taiwan) were used in the second animal experiment. The rats were raised on a 12-h light–dark cycle and had a free access to water and food. All experimental procedures were approved by the Institutional Animal Care and Use Committee of National Taiwan University and were conducted in accordance with its guidelines.. Part I - changes of microcirculatory blood flow intensity in splanchnic organs Anesthesia and surgical procedure The anesthesia and surgical protocol were the same as those in the above mentioned animal study I. Accordingly, a tracheostomy was performed, and then a 14G catheter (Surflo; Terumo Corporation, Laguna, Philippines) was inserted into the trachea to maintain spontaneous breathing. Anesthesia was maintained using 1.2 % inhaled isoflurane through the tracheal tube. Subcutaneous atropine (0.05 mg/kg in 10 mL/kg of saline) was injected to limit the rate of moisture evaporation from the surgical open wound and to prevent airway secretion. The body temperature of the rats was continuously monitored rectally, and a warmer pad was applied to maintain it between 36 and 37 °C. Polyethylene catheters (PE-50; Intramedic 7411, Clay Adams, Parsippany, NJ, USA) were inserted into the right common carotid artery and external jugular vein. The catheter in the right common carotid artery was used to continuously monitor the MAP and HR, and to withdraw blood for inducing hemorrhagic shock. The external jugular vein catheter was used to infuse resuscitation fluid. A midline laparotomy was performed to exteriorize splanchnic organs, including 17.

(32) the liver, left kidney, and a segment of the terminal ileum (approximately 6–10 cm in length, proximal to the ileocecal valve). A 2-cm section was made on the antimesenteric aspect of the intestinal lumen by using a high-frequency desiccator (Aaron 900; Bovie Aaron Medical, St. Petersburg, FL, USA) for carefully exposing the opposing mucosa to study microcirculation. Furthermore, the nearby intestinal serosal muscular layer (at the midline of the antimesenteric aspect) and the central Peyer’s patch (identified by visualizing the lymph nodes) were identified to study microcirculation. Moreover, the right gracilis muscle was exposed for monitoring microcirculatory changes relative to those in the splanchnic organs. The exposed viscera and tissue were kept moist by treating them every hour with 0.5 mL of prewarmed to 37 °C aerosolized 0.9% saline. Evaluation of microcirculatory blood flow The microcirculatory blood flow intensity was assessed using the LSCI technique and the setup was similar to that employed in the animal study I. A full-field laser perfusion imager (MoorFLPI; Moor Instruments, Ltd., Devon, UK) was used to continuously record the intensity of microcirculatory blood flow in the splanchnic organs, starting from the baseline. The imager uses LSCI, in which a random speckle pattern is generated when a tissue is illuminated by a laser light. The random speckle pattern changes when blood cells move within the ROIs. When the level of movement is high (high flow), the changing pattern becomes more blurred, and the contrast in that particular region declines accordingly. Therefore, low contrast is associated with high flow, and high contrast is associated with low flow. The contrast image is processed to produce a 16-color-coded image that is associated with the blood flow in the tissue (e.g., blue indicates low flow and red, high flow). The microcirculatory blood flow intensity of each ROI was recorded as a PU, which is equal to the product of the average speed and concentration of red blood cells moving in the tissue sample 18.

(33) volume (i.e., blood cell flux or perfusion). The images were recorded and analyzed in real time using the MoorFLPI Version 3.0 software (Moor Instruments, Ltd.). Six individual ROIs were defined on the liver, left kidney, intestinal mucosa, serosal muscular layer, Peyer’s patch, and right gracilis muscle (identical to the animal study I). Microcirculatory blood flow intensity was recorded as an arbitrary PU.. Hemorrhagic shock and fluid resuscitation protocols A total 60 rats were randomly assigned to the following six groups (n = 10 in each group): (1) the sham group, in which the rats received all surgical procedures except blood withdrawal and fluid resuscitation; (2) the control group, in which the rats underwent hemorrhaging but no fluid resuscitation; (3) the NS group, in which the rats were resuscitated with 0.9 % saline; (4) the HTS group, in which the rats were resuscitated with 3 % HTS; (5) the GEL group, in which the rats were resuscitated with 4 % succinylated GEL (Gelofusine®, B. Braun, Taiwan); and (6) the HES group, in which the rats were resuscitated with 6 % HES 130/0.4 (Voluven®, Fresenius-Kabi, Bad Homberg, Germany). The timeline of part I of this experiment is shown in Figure 3a. After completion of surgery, the rats were allowed to hemodynamically stabilize for 30 minutes before the baseline hemodynamic measurements were recorded (baseline condition was considered stable when all measurement values remained at 10 % for 15 minutes; defined as 0’ minutes; T0). At T0, the concentration of isoflurane was lowered to 0.7 % to prevent overanesthesia in the rats without further surgical stimulation, and hemorrhagic shock was then induced through controlled blood withdrawal via a right carotid arterial cannula with total blood loss of 30 mL/kg from 15’ to 60’ (T1) minutes. Fluid resuscitation was administrated from 60’ to 90’ minutes, and the volume of fluid resuscitation was identical to that of blood withdrawn. Two hours 19.

(34) after fluid resuscitation, part I of the experiment was considered complete (T2; at 210’ minutes in the rats receiving fluid resuscitation and at 180’ min in the rats in the sham and control groups; Fig. 3a). Hemodynamic measurements of both macrocirculatory and microcirculatory variables were compared among the groups at T0, T1, and T2 (Fig. 3a). Because different organs may have different baseline states of both macrocirculatory and microcirculatory conditions, the percent changes in the microcirculatory blood flow intensity at T1 and T2 were compared with the T0 baseline values. The first 0.2 mL of blood withdrawn at 15’ minutes and the final 0.2 mL of blood withdrawn at T1 were analyzed using arterial blood gas analysis as the baseline status and shock status, respectively. At the end of the measurement (T2), arterial blood gas analysis was repeated to evaluate the effects of fluid resuscitation.. Part II. In vivo renal reactive oxygen species formation For measuring in vivo kidney ROS activity during hemorrhagic shock and fluid resuscitation, 36 rats underwent the same hemorrhagic shock and fluid resuscitation (N = 6, in each group) protocols as did the rats used for investigating microcirculatory changes in the splanchnic organs with two modifications. First, the rats were anesthetized with a subcutaneous urethane injection (1.2 g/kg) instead of volatile anesthesia because in vivo measurement was conducted in a closed box that did not allow the entry of a volatile anesthesia breathing circuit. In addition, volatile anesthesia may suppress ROS formation because of its strong antioxidant properties. 20.

(35) (De Wolf and Hendrickx, 2013). Second, pilot experiments showed that rats receiving anesthesia with subcutaneous urethane could survive only if the volume of blood loss was lowered. Therefore, hemorrhagic shock was induced by withdrawing a lesser amount of blood (20 mL/kg; Fig. 3b). The method for detecting chemiluminescence (CL) from the organ surface after intrarenal arterial infusion of a superoxide anion probe, 2-methyl-6-[4methoxyphenyl]-3,7-dihydroimidazo-[1,2-a]-pyrazin-3-one hydrochloride (MCLA) (TCI-Ace; Tokyo Kasei Kogyo Co. Ltd., Tokyo, Japan) (Chien et al., 2001, Yang et al., 2014), was adapted to demonstrate ROS production in kidneys subjected to ischemia-reperfusion injury. For excluding photon emission from sources other than the kidney, the rats were housed in a dark box with a shielded plate. Only the renal window was left unshielded and was positioned under a reflector, which reflected the photons from the exposed kidney surface onto the detector area. A single dose of N,N9-dimethyldiacridium (1 mM in 0.1 mL) (lucigenin; Sigma) or a continuous infusion of MCLA (0.2 mg/mL per h) was administered through an intrarenal arterial catheter. The lucigenin- or MCLA-enhanced CL signal from the renal surface was measured continuously during administration by using a CL analyzer (CLD-110, Tohoku Electronic Industrial Co., Sendai, Japan). The ROS response was directly evaluated from the renal surface through intrarenal arterial infusion of MCLA (0.2. 21.

(36) mg/mL/h) (Chien et al., 2001). At T2, the changes in ROS formations from baseline were compared.. 22.

(37) Section 2：Predicting stroke volume and arterial pressure fluid responsiveness in liver cirrhosis patients by using dynamic preload variables Patients Before conducting this clinical study, this study was approved by the Research Ethics Committee of National Taiwan University Hospital, Taipei, Taiwan, (Chairperson: Professor Hong-Nerng Ho) on 4 October 2013 (201309024RIND). After obtaining written informed consent from all patients, adult recipients of a living donor orthotopic liver transplant were consecutively enrolled from November 2013 to April 2015. The exclusion criteria were a history of pulmonary resection, chronic respiratory insufficiency, left ventricular ejection fraction (LVEF) of <60% or pulmonary hypertension, and arrhythmia (NCT01971333). Anesthesia General anesthesia induction is performed by using intravenous fentanyl (1.2 µg kg−1), etomidate (0.3 mg kg−1), and cisatracurium (0.15 mg kg−1). Anesthesia was then maintained in a standard manner with desflurane in an air/oxygen mixture and intravenous infusions of fentanyl and cisatracurium. During surgery, the anesthetic depth was maintained by keeping the bispectral index (BIS) between 40 and 60. For maintenance of normocapnia, mechanical ventilation was set to a tidal volume of 8 mL kg−1, a respiratory rate of 10–15 min−1, and a positive end-expiratory pressure of 5. 23.

(38) cmH2O. The respiratory settings were not changed throughout the measurements. Hemodynamic assessment Each patient received standard monitoring by using a Philips Intellivue MP 70 monitor (Philips, Suresnes, France). After general anesthesia, a triple-lumen 5.5-Fr catheter (Arrow central venous catheter; Teleflex Life Sciences Ltd., Athlone, Ireland) was placed into the right internal jugular vein. A 4-Fr thermistor-tipped arterial catheter (Pulsiocath thermodilution catheter; Pulsion Medical Systems, Munich, Germany) was inserted into the right femoral artery, advanced to the abdominal aorta, and connected to the PiCCOplus™ system monitor (Version 7.0; Pulsion Medical Systems). The PiCCO system is using the transpulmonary thermodilution technique to generate the cardiac index, SV index (SVI; SV divided by the body surface area for normalisation), and SVRI, which are comparable to those generated using the pulmonary artery catheter (Della Rocca et al., 2002a, Hofer et al., 2005, Della Rocca et al., 2002b).Continuous SVI measurement was initiated after the system was initially calibrated by injecting 20 mL of ice-cold normal saline (< 8°C ) into the central venous catheter (transpulmonary thermodilution) three times, and further calibration was performed hourly. The cardiac index was calculated as follows: SVI × heart rate. The Masimo Radical 7 co-oximeter (version 7.8; Masimo Corp., Irvine, CA, USA). 24.

(39) uses a finger clip to measure arterial oxygen saturation concentration noninvasively on the basis of transcutaneous multiwavelength analysis. The Masimo sensor was attached to the right index finger according to the manufacturer’s instructions and was on the side opposite from the intravenous catheter. Dynamic preload variables SVV and PPV were calculated according to the following formula: SVV (%) = [(SVmax − SVmin)/SVmean] × 100 and PPV (%) = [(PPmax − PPmin)/PPmean] × 100. Using the PiCCOplus™ system, we measured SVmax and SVmin (or PPmax and PPmin) as the mean values of the four extreme values of SV (or PP) during measurement from the femoral artery for 30 s. The SVmean (or PPmean) was recorded as the average value during this period. The PVI, which is based on the perfusion index, was obtained from the Masimo Radical 7 SET co-oximetery. The PVI was calculated as [(perfusion indexmax − perfusion indexmin)/ perfusion indexmax] × 100 over a period sufficient to include multiple respiratory cycles. Vascular tone variables Vascular tone variables including the SVRI, perfusion index, and PPV/SVV ratio were measured. The SVRI was calculated as (MAP – CVP) × 80/cardiac index. The perfusion index was calculated as the ratio of pulsatile to non-pulsatile blood volumes. 25.

(40) and was automatically calculated using the Masimo Radical 7 SET co-oximetery.. Fluid challenge criteria, timing, and the definition of stroke volume and arterial pressure fluid responders After the PiCCO monitor was set up, the patients were allowed to stabilize for 15 min. Subsequently, they received a fluid challenge of 10 mL kg−1 0.9% saline within 15 min during the surgical preparation before skin incision. An additional saline challenge was administered 2 h after the surgical incision. During the dissection phase, each patient received basal 0.9% saline infusion at a rate of 5 mL kg−1 h−1. Hemodynamic data, namely the HR, MAP, CVP, SVI, PPV, SVV, PPV/SVV ratio, PVI, perfusion index, and SVRI, were recorded before (baseline) and 5 min after each fluid challenge. The averages of the hemodynamic measurements for 1 min before and after the fluid challenge were used for statistical analysis. During the measurement of hemodynamic variables, surgical and anesthetic teams stopped surgery for 1 min without any stimulation to the patient and without clamping or unclamping major vessels. Patients were considered SV responders if their SVI increased by at least 15% after the fluid challenge. Patients were considered arterial pressure responders if their mean arterial pressure increased by at least 15% after the fluid challenge.. 26.

(41) Section 3: Comparison of two stroke volume variation-based goal-directed fluid therapies for supratentorial brain tumor resection Study design and participants In this single-institution, randomized, single-blind, open-label trial with two parallel arms, we compared the low SVV and high SVV GDFT strategies in patients undergoing elective supratentorial brain tumor resection. This trial was approved by the Research Ethics Committee of National Taiwan University Hospital and was registered at clinicaltrials.gov with the identifier NCT02113358. We enrolled patients older than 20 years who had undergone elective craniotomy for supratentorial brain tumor resection during May 2014–July 2015. Patients who met any of the following criteria were excluded from the trial: age > 70 years; recurrent or metastatic tumors; body mass index (BMI) < 18.5 kg/m2 or > 27.0 kg/m2; a history of cardiac dysfunction, such as coronary artery diseases, NYHA class II or higher heart failure, and arrhythmia; renal insufficiency with an estimated glomerular filtration rate of < 60 mL/min/1.73 m2; and chronic obstructive pulmonary disease (Fig. 4). All surgery was performed by five neurosurgeons who specialized in oncological neurosurgery and were unaware of the study purpose. All patients provided written informed consent on the day before surgery, which was performed by an investigator who was unaware of the randomization result. On. 27.

(42) arrival at the operation theatre, patients were allocated to the study arms in a 1:1 ratio according to a predefined block randomization list, which was prepared by an independent statistician before the trial, with blocks randomly sized between four and six patients. Only the attending anesthesiologist and research nurse were aware of patient allocation; furthermore, the surgical team and patients were not informed of the group allocation. Anesthesia General anesthesia was induced through total intravenous anesthesia (TIVA) comprising a combination of target-controlled infusion of propofol (with an initial plasma concentration of 4 mg/mL, using the Schnider model for induction, and maintaining a plasma concentration of 4 ± 1 mg/mL) (Passot et al., 2005), fentanyl, and cisatracurium. Patients were ventilated in a volume-controlled mode with a tidal volume of 8 mL kg−1, and air–oxygen ratio of 1:1, and a total flow rate of 1 L/min. A respiratory rate was set to maintain PaCO2 between 25 and 30 mmHg to improve the operating conditions according to the surgeon’s request (Gelb et al., 2008). For intraoperative and postoperative analgesia, each patient received a scalp nerve block containing 10–20 mL of 0.5% levobupivacaine before head clamp placement and skin incision. In addition, intraoperative neurophysiological monitoring techniques were used to enhance resection safety.. 28.

(43) Hemodynamic monitoring Each patient fasted from midnight and received standard intraoperative monitoring, including electrocardiography, noninvasive blood pressure measurement, pulse oximetry, capnography, and nasopharyngeal temperature measurement, by employing a Philips Intellivue MP70 monitor (Philips, Suresnes, France). After anaesthesia induction, a 20-G radial arterial line was inserted and connected to the fourthgeneration Vigileo/Flotrac system (Edwards Lifesciences, Irvine, CA, USA) to obtain the SVV and cardiac index. The Vigileo/Flotrac system analyses the pressure waveform 100 times s−1 over 20 s, capturing 2,000 data points for analysis and performing calculations by using data obtained in the most recent 20 s. The cardiac index values were determined by assessing the arterial tree impedance as the cardiac output per body surface area. The SVV was calculated as the variation in beat-to-beat SV from the mean value obtained during the most recent 20 s: SVV = (SVmax − SVmin/SVmean). A large-bore intravenous catheter (16-G) in forearm or a double-lumen 5.5-Fr catheter (Arrow central venous catheter; Teleflex Life Sciences Ltd., Athlone, Ireland) was inserted into the femoral vein for fluid infusion. During surgery, the surgeons had no access to the information generated by the Vigileo/Flotrac system. Hemodynamic variables, including heart rate, mean arterial pressure, SVV, and cardiac index, were recorded at baseline (i.e., before anesthesia for heart rate and. 29.

(44) mean arterial pressure; before surgery for SVV and cardiac index) and every 15 min during surgery. The baseline and average values of the total measurements were analyzed and compared between the two study groups. Goal-directed fluid therapy protocols and fluid bolus indications Before conducting this study, a grey zone was created for the fourth-generation Vigileo/Flotrac-derived SVV to predict a 10% increase in the cardiac index values of 124 fluid boluses [using 250 mL of tetrastarch (Voluven; Fresenius Kabi, Uppsala, Sweden) within 15 min] from 45 patients who underwent supratentorial brain tumor resection through retrospective analysis of a prospectively maintained database. In summary, the Youden index for each bootstrapped sample from 1,000 replications of the pilot study population was calculated, and the median value and 95% confidence interval of these 1,000 optimal cut-offs were obtained (Cannesson et al., 2011). The grey zone resampling results showed that the median cut-off was 13% with a 95% confidence interval of the distribution of optimal cut-offs ranging between 9% and 17%. Therefore, the threshold SVV values of ≥ 10% and ≥ 18% for the low SVV and high SVV groups, respectively, were selected to initiate fluid boluses. Moreover, because the prone position increases intrathoracic pressure, the SVV threshold for surgery in the prone position was 5% higher than that for surgery in the supine position (Biais et al., 2010) (15% for the low SVV group and 23% for the high SVV. 30.

(45) group). Figure 5 presents the GDFT protocols for the low SVV and high SVV groups. The GDFT hemodynamic goal was evaluated every 15 min. The difference between the two fluid strategies used in this study is the employment of fluid bolus using 250 mL of tetrastarch within 15 min on the basis of the various SVV threshold values. The attending anesthesiologist was allowed to infuse supplementary colloid fluid bolus to prevent inadvertent hypovolemia if a ≥ 10 % decrease in cardiac index concurrent with a ≥ 5% increase in the dynamic parameter from baseline states was observed because of its highly possible association with a ≥ 5% decrease in estimated blood volume.(Kungys et al., 2009, Pizov et al., 2012) The SVV-based protocol was temporarily suspended when blood transfusion was required. Intravenous ephedrine (4–8 mg per bolus) was administered at the discretion of the attending anesthesiologist to maintain the cardiac index when patient SVV was within the threshold limit; dopamine was infused if the total ephedrine dose was > 40 mg (Fig. 5). Standard measures were applied to each patient to maintain oxygenation (SpO2 ≥ 98%), hemoglobin (>9.0 dg/L), blood glucose (<140 mg/dL), and core temperature (37°C, by using a fluid warmer and the Bair Hugger forced-air patient warming system; Augustine Medical Inc., Eden Prairie, MN, USA). Each patient was. 31.

(46) maintained at a cardiac index of ≥ 2.5 L/min m2, a mean arterial pressure of ≥ 70 mmHg, and a baseline level of > 85%. Intravenous nicardipine, labetalol, or esmolol was administered at the discretion of the attending anesthesiologist to maintain a systolic blood pressure of < 140 mmHg and a heart rate of < 100 beats/min. Each patient received a basal 0.9% saline infusion fluid at a rate of 1 mL/kg. The attending anesthesiologist could administer additional saline to replace urinary loss resulting from mannitol (20%; 1 g/kg; rapid infusion within 15 min) administration at the discretion of the attending surgeon (Fig. 5). Postoperative care and outcome measurements After surgery, all patients were immediately transferred to the same neurosurgical ICU and received identical postoperative care. The attending neurosurgeons and anesthesiologist determined whether the patient could be extubated in the operating theatre; if the tracheal tube could not be extubated, a weaning program was initiated in the ICU as soon as possible by respiratory therapists who were independent of this study. Prolonged mechanical ventilation was defined as a failure to extubate the tracheal tube within 4 h after arrival at the ICU. After extubation, resumption of oral fluid was encouraged, and postoperative nausea and vomiting were treated through intravenous ondansetron. The patients were observed daily by surgeons who were unaware of the study. 32.