预缺血与间断缺血肝切除随机对照研究.pdf

ORIGINAL ARTICLE Ischemic preconditioning prior to intermittent Pringle maneuver in liver resections Anders Winbladh Bergthor Bjo rnsson Lena Trulsson Karsten Offenbartl Per Gullstrand Per Sandstro m Published online 7 June 2011 Japanese Society of Hepato Biliary Pancreatic Surgery and Springer 2011 Abstract Background Purpose Continuous infl ow vascular occlu sion during liver resections causes less severe ischemia and reperfusion injury IRI if it is preceded by ischemic pre conditioning IP or if intermittent infl ow occlusion is used during the resection No previous clinical trial has studied the effects of adding IP to intermittent infl ow occlusion MethodsConsecutive patients n 32 with suspicion of malignant liver disease had liver resections minimum 2 segments performed with infl ow occlusion intermittent clamping in a manner of 15 min of ischemia and 5 min of reperfusion repetitively 15 5 Half of the patients were randomized to receive IP 10 min of ischemia and 10 min of reperfusion before parenchymal transection 10 10 The patients were stratifi ed according to volume of resection and none had chronic liver disease The patients were followed for 5 days with microdialysis lD ResultsAll patients completed the study and there were no deaths No differences were seen between the groups regardingdemographicsorperioperativeparameters bleeding duration of ischemia resection volume com plications and serum laboratory tests There were no differences in alanine aminotransferase ALT aspartate aminotransferase AST bilirubin or prothrombin time PT INR levels but lD revealed lower levels of lactate pyruvate and glucose in the IP group having major liver resections analysis of variance ANOVA Nitrite and nitrate levels in lD decreased postoperatively but no dif ferences were seen between the groups In one patient an elevated lD glycerol curve was seen before the diagnosis of a stroke was made ConclusionsIP before intermittent vascular occlusion does not reduce the serum parameters used to assess IRI IP seems to improve aerobic glucose metabolism as the levels of glu cose pyruvate and lactate locally in the liver were reduced compared to controls in patients having 3 segments resec ted lD may be used to monitor metabolism locally KeywordsLiver Microdialysis Surgery Cancer Lactate Ischemia Introduction Clamping of the portal pedicle i e the Pringle maneuver PM is used to reduce bleeding during liver transection 1 2 The liver suffers from ischemia and reperfusion injury IRI when the portal pedicle is clamped 3 5 Intermittent clamping in a manner of 15 min of ischemia and 5 min of reperfusion repetitively 15 5 has been shown to reduce the IRI compared to continuous PM 6 7 In 1986 it was shown that ischemic preconditioning IP reduced ATP depletion in the heart during the subsequent ischemic insult 8 IP also protects the liver from IRI if it is employed before continuous PM 2 9 11 IP in liver surgery is typically performed as 10 min of PM followed by 10 15 min of reperfusion prior to continuous PM There are two studies comparing 15 5 intermittent vascular Previously presented at the Swedish Surgical Week 2010 August 22 as a 10 min oral presentation and poster and at the A EHPBA meeting in Cape Town South Africa April 14 2011 as an oral presentation A Winbladh 10 10 before a 15 5 PM reduce the IRI or change the glucose metabolism compared to the 15 5 PM alone in surgical liver resections The IRI was evaluated using postoperative serum anal yses of bilirubin aminotransferases and PT INR Glucose metabolism was evaluated by the blood levels of glucose and lactate as well as by the recovered levels of glucose pyruvate lactate and glycerol in the microdialysate This prospective clinical trial has been registered at the website http www controlled ISRCTN29593599 Subjects materials and methods The study protocol was approved by the Regional Ethics Committee at the University of Linko ping University Hospital Linko ping Sweden Study design and patient selection This prospective single blinded clinical trial was conducted between January 16 2008 and June 8 2009 at the Department of Surgery University Hospital Linko ping Sweden Patients intended for resection of two or more segments of the liver due to metastatic liver disease or suspicion of T1 2 gallbladder cancer were asked to par ticipate in the study No patients intended for multiorgan resection or portal ligature or those with a suspicion of chronic liver disease were included The defi nitive decision to include a patient in the study was taken after the laparotomy and the patient was con sidered to have resectable disease Before randomization patients were stratifi ed according to the size of the resec tion Sixteen patients had a resection of 2 3 segments and another 16 patients had a resection volume of 3 segments Closed envelopes were prepared from a random table generated at the website http www random org Half of the patients in each of the strata were randomized to have IP prior to the PM making a control group n 8 within each stratum Fig 1 The patients were blinded to their group allocation whereas the surgeons were not Power In this trial we introduce a novel method of assessing postoperative liver metabolism after IRI No reliable cal culations of power could thus be made instead we looked at clinical signifi cance IP takes 20 min to perform and the benefi t from this maneuver must justify the extra time spent in the operating room OR It was therefore decided that if no differences were seen in 16 patients having IP i e 5 h in the OR the results would not be of clinical interest Operative procedure Anesthesia was induced with fentanyl together with either propofol or sodium thiopental Rocuronium was used for relaxation and anesthesia was then maintained with sevo fl urane To minimize bleeding during the liver transection an intravenous i v infusion of nitroglycerine 0 1 mg mL was given to keep a low central venous pressure CVP Five patients had patient controlled analgesia with i v Fig 1 Schematic illustration of the randomization The study population was set at 32 subjects When the size of the resection was decided an envelope of that stratum was opened to reveal ischemic preconditioning IP or no IP 160J Hepatobiliary Pancreat Sci 2012 19 159 170 123 morphine pumps postoperatively and 27 patients had epi dural catheters activated postoperatively All operations started between 8 30 and 9 30 a m and all patients were operated through a right subcostal incision angled and extended to the sternum Resectability and the size of the resection was decided after the abdomen was palpated the liver mobilized and an intraoperative ultrasound had been performed Defi nitive inclusion stratifi cation 2 3 or 3 segments and randomization IP or no IP were made at this point A cotton tape was placed around the hepa toduodenal ligament In all patients the liver was transected with the Cavitron Ultrasonic Surgical Aspirator CUSA Excel Integra Lyon France device Blood samples were taken preoperatively after ische mia at wound closure at 8 p m on the day of the opera tion and then on postoperative days PODs 1 4 at 7 a m All blood and serum samples lactate AST ALT bilirubin PT INR albumin hemoglobin Hb and WBC were analyzed at the Department of Clinical Chemistry Uni versity Hospital Linko ping using the standard methods of this department Bedside analyses of capillary blood glu cose levels were performed using the Accu Check Inform II Roche Mannheim Germany Just after the abdomen had been opened a liver biopsy was taken and immediately frozen in liquid nitrogen It was then freeze dried and stored at 20 C until analysis Microdialysis The lD catheter consists of two concentric plastic tubes one inside the other The tubes communicate at the distal end of the catheter where there is a semi permeable membrane The outer tube is perfused slowly with a physiologic solution and equilibrates with the interstitial fl uid before it is recovered through the inner tube Fig 2a Two 0 9 mm thick lD catheters CMA 63 CMA Solna Sweden with 30 mm long semi permeable membranes and a pore size of 20 kD were tunnelated through the skin and the abdominal wall They were introduced close to each other in the anticipated residual liver and secured to the capsule of Glisson with a 4 0 absorbable suture Both lD catheters were anchored to the skin with adhesive tape Fig 2b Perfusion of the lD catheters with Ringers ace tate Braun Stockholm Sweden was kept at 2 0 lL min during the surgical procedure and the postoperative inten sive care unit ICU stay and then decreased to 0 3 lL min in the ward using CMA 107 micropumps CMA A steady state was considered to have been achieved after 20 min of lD catheter perfusion The micro vials were collected every 30 min in the OR every hour in the ICU and every fourth hour in the ward The catheters remained in place until the afternoon of POD 4 Fig 3 One of the micro vials was immediately analyzed in the bedside analyzer ISCUS CMA for lactate glucose glycerol and pyruvate levels The other micro vial was frozen at 20 C for later analysis of NOx see below levels Classifi cation of complications and clinical status The most severe complication was registered for each patient according to the Dindo Clavien classifi cation Dindo et al 18 The modifi ed early warning score Fig 2 Schematic illustration of the microdialysis catheter where diffusion of molecules over the semi permeable 20 kDa membrane is driven by concentration gradients The micropump perfuses the catheter and the dialysate is recovered in the micro vial a presented with kind permission of CMA microdialysis Schematic illustration showing the positioning of the catheter tips within the liver parenchyma The micropumps are kept in a waist belt b J Hepatobiliary Pancreat Sci 2012 19 159 170161 123 MEWS 19 was used to assess the clinical status of the patients every fourth hour MEWS includes the following parameters pulse blood pressure respiratory frequency temperature urinary output and patient consciousness Nitrite nitrate in microdialysate The instructions in the commercial Nitrite Nitrate Fluoro metric Assay Kit Cayman Chemical Company Ann Arbor MI USA were followed when the sum of NO2and NO3 was analyzed in the lD samples The fl uorometric intensity was measured with the Victor 3 1420 Multilabel Plate Reader V instrument PerkinElmer Va sby Sweden using an excitation wavelength of 355 nm and emission wavelength of 460 nm with a fi lter tolerance of 20 nm The included standard curve was used when calculating the results given in lM The in vitro lD recovery of NO2 NO3 was tested and found to be 70 10 at the perfu sion rate of 1 0 lL min using a 30 mm 20 kDa lD catheter Glycogen Glycogen in the liver tissue was quantifi ed with the Bio vision Glycogen Assay kit BioVision Research Products Mountain View CA USA Freeze dried tissue was put in iced Eppendorf tubes mortared and extracted into sterile water Enzymes were then inactivated by boiling for 5 min and thereafter the tubes were centrifuged at 13 000 rpm for 5 min The supernatant was put in 96 well plates As rec ommended in the assay kit the glucose levels in the extracts corresponding to 50 lg freeze dried tissues were measured with and without hydrolyzing the glycogen by glucoamylase After reaction with OxiRed the absorbance was fi nally colorimetrically measured in an enzyme linked immunoassay ELISA reader at 550 nm The glycogen level was calculated against a standard curve included in the kit and the result is shown as lg glycogen per mg freeze dried tissue Statistics Patients having IP n 16 were compared with the con trols n 16 and then the patients having IP n 8 were compared to the controls n 8 within the same stratum The results are given for these stratifi ed groups unless otherwise indicated in the text A secondary com parison was made between minor 2 3 segments n 16 and major 3 segments n 16 resections to control for confounding changes depending on resection size The microdialysate samples were retrieved at the time points specifi ed above The lD samples were then grouped in 7 different phases baseline sample 1 peroperatively samples 2 7 postoperatively the day of surgery samples 8 24 POD 1 samples 25 30 POD 2 samples 31 36 POD 3 samples 37 42 and POD 4 samples 43 46 Individuals were nested in the IP and control groups respectively and assigned as the random factor in an ANOVA A Tukey post hoc analysis was performed if signifi cant statistical differences were found between the groups in the ANOVA Non parametric values are given as ranges medians andwereanalyzedwiththeMann WhitneyU test parametric values are given as means SEM except in Table 1 where mean SD is used and were analyzed usingtheStudent st test STASTICA8 0software StatSoft Tulsa OK USA was used for all statistical calculations A p value of 0 05 was considered statisti cally signifi cant Results All randomized patients completed the study according to intention to treat and there was no mortality within the study All patients n 79 having liver surgery at the department of surgery Linko ping during the study period were retrospectively reviewed Two patients had declined participation in the study and 45 did not meet the study criteria Demographics There were no differences between the whole IP group n 16 and the controls n 16 or between the strati fi ed groups regarding demographics sex age and neoad juvant chemotherapy or intraoperative data bleeding transfusions resection volume duration of the ischemia or the glycogen levels see Table 1 Subjects having the large volume 3 segments n 16 resections proved to have had a larger liver volume resected mean SD 700 349 vs 184 105 ml p 0 001 a greater transectional area mean SD 122 29 vs 93 37 cm2 p 0 05 and Fig 3 Treatment protocol of the IP group and the control group The black arrows show when the parenchymal transection started Microdialysate lD was collected at the given intervals min The fi rst sample 10 min was discarded and the second 30 min was used as baseline data IP ischemic preconditioning ICU intensive care unit 162J Hepatobiliary Pancreat Sci 2012 19 159 170 123 a longer duration of the operation mean SD 328 95 vs 251 50 min p 0 05 than those having 2 3 segments resected n 16 Two patients in the control group had non insulin dependent diabetes mellitus Seven patients received insulin i v perioperatively 4 in the IP group and 3 in the control group No insulin whatsoever was administered in the study after the day of surgery as glucose levels remained lower than 12 mM in all patients during the study period The glucose levels in the blood showed no differ ences between the groups Tables 2 3 The overall med ian length of stay LOS was 9 days with no differences between the groups see Table 1 Patients having major resections had longer LOS than those having small resec tions 9 vs 10 days p 0 014 Complications and diagnoses Complications occurred in 12 patients 37 5 and these were scored according to Dindo Clavien 18 No differ ences were seen between the groups There was one stroke due to carotid artery stenosis grade 4 one reoperation due to wound dehiscence grade 3b one biliary leakage one abscess and 4 pleural effusions all drained percutaneously grade 3a and 4 antibiotic or conservatively treated infections grade 2 or 1 Table 1 Twenty six patients were diagnosed with metastases 23 colorectal 1 anal 1 ovarian and 1 malignant melanoma 4 patients had benign lesions 1 hemangioma 1 cyst 1 gall bladder polyp and 1 with chronic cholecystitis 1 patient had gallbladder cancer and 1 had hepatocellular cancer Table 1 Demographic and perioperative data 2 3 segmentsp 3 segmentsp IPControlsIPControls Sex Males 4 50 5 62 0 724 50 5 62 0 72 Age years 67 15 62 9 0 4460 13 66 10 0 27 Preop glycogen levels lg mg fdt 40 6 0 42 9 8 0 7441 7 4 37 7 7 0 41 Preop chemotherapy 4 50 3 38 0 727 88 6 75 0 72 Units of insulin given total range 51 0 47 24 0 20 0 7284 0 60 60 0 60 0 72 LOS days 7 13 9 6 11 6 0 518 13 10 7 51 11 0 51 Blood loss mL 100 1700 359 50 2400 300 1 0100 1500 625 200 1800 750 0 38 Units of blood transfused total range 6 0 4 4 0 2 0 9610 0 5 7 0 3 0 88 Units of plasma transfused total range 9 0 7 7 0 5 0 9613 0 6 18 0 7 0 57 Resected volume mL 55 340 190 60 400 165 0 57225 1150 698 300 1200 550 0 96 Transected area cm2 91 36 95 40 0 84132 33 113 23 0 20 Bleeding cm2 mL cm2 1 0 11 5 2 0 5 15 5 9 0 850 9 25 4 4 2 3 21 6 6 0 44 Duration of ischemia min 44 10 44 13 0 9835 11 44 8 0 0 08 Transection time min 64 22 77 40 0 4662 32 60 11 0 86 Duration of surgery min 242 57 261 42 0 52292 60 359 112 0 21 Bleeding min transection mL min 1 8 19 5 7 0 9 20 6 5 0 882 6 36 9 2 3 9 42 11 2 0 65 Bleeding min total mL min 0 5 5 2 1 8 0 6 2 9 1 1 0 450 4 5 2 2 0 0 9 4 7 3 6 0 76 Complications n 3 38 2 25 0 724 50 3 38 0 96 Diagnosis Metastases n 7 88 4 50 0 288 100 7 88 1 0 Benign n 1 12 3 38 00 Carcinoma HCC n 0001 12 Gallbladder cancer n 01 12 00 Steatosis moderate n 1 12 003 38 0 20 Portal fi brosis moderate n 1 12 000 Results are given as means SD for parametric data whereas non parametric data are given as ranges medians Student s t test and the Mann Whitney U test were used respectively and the Fisher exact test was used for categorical variables The IP group n 8 and controls n 8 are compared within each stratum 2 3 segments and 3 segments resected respectively IP ischemic preconditioning fdt freeze dried tissue LOS length of stay HCC hepatocellular carcinoma Preop preoperative J Hepatobiliary Pancreat Sci 2012 19 159 170163 123 HCC in a non fi brotic liver No patient was preoperatively suspected of having parenchymal liver disease Histological examination of the parenchyma revealed that 4 patients had grade 2 10 29 moderate macrovesicular steatosis post chemotherapy and one patient had moderate portal fi brosis according to Ludwig and Batts grading see Table 1 The rest had no or mild fatty changes or portal fi brosis