- Email: [email protected]
1026 Predicting Organ Space Surgical Site Infection with a Nomogram
endoscopic assisted transgastric approach. This Video demonstrates our technique using an endoscope and two 5mm ports placed through the anterior gastric wall 1022 Laparoscopic Assisted Transanal Sigmoidectomy (LATAS): A Bridge to NOTES? Joep Knol, Eric J. Dozois
SSAT Abstracts
The combination of standard laparoscopy and specimen extraction through a natural orifice has the potential to decrease wound related complications. We describe a new technical approach to laparoscopic sigmoidectomy in which the specimen is extracted transanally using a TEM-like proctoscope. Laparoscopic assisted transanal sigmoidectomy (LATAS) could be considered an alternative minimally invasive approach in patients with benign disease of the sigmoid colon. A 44-year-old male with recurrent diverticular disease of the sigmoid colon was consented to undergo the LATAS procedure. Operating time was 92 minutes, no perioperative or postoperative complications occured. 1023
1025
High Resolution Manometry Has Less Interobserver Variability Than Conventional Manometry Attila Dubecz, Renato Salvador, Marek Polomsky, Oliver Gellersen, Daniel Raymond, Carolyn E. Jones, Thomas J. Watson, Jeffrey H. Peters
Murine Functional Liver Mass Is Reduced Following Partial Small Bowel Resection Zhaohua Qiu, Shannon W. Longshore, Brad W. Warner, David A. Rudnick Liver mass is regulated in precise proportion to body mass in healthy animals, and this ratio (liver:body weight) is specifically restored by regeneration following acute injury. Despite extensive analysis of the molecular mechanisms that regulate hepatic regeneration, the specific signals that initiate and terminate the regenerative response and those that determine the ideal liver:body mass ratio have not been elucidated. Previous analyses have suggested the mechanisms involved in such regulation are likely to include signals derived from the bowel via the portal circulation. Therefore, in the studies reported here we investigated the effect of partial small bowel resection on regulation of liver mass in a murine surgical model. Methods: Two month old, male C57BL/6J wildtype mice were subjected to partial small bowel resection or sham surgery (transection without resection) followed by primary reanastamosis, as we have previously described. These animals were allowed to recover, after which they were sacrificed at serial time points for determination of liver:body mass ratio and analysis of liver tissue. Results: In mice subjected to removal of ~50% of proximal small intestine, liver:body weight ratio was decreased 10% at 48 hrs and 20% at 72 hrs after surgery compared to sham operated animals (p<0.01). Similar results were seen in animals in which a comparable amount of distal small bowel was resected. Functional liver mass was also reduced based on analyses of total hepatic protein (reduced 30% versus sham, p<0.01) and total hepatic alanine transaminase (ALT) activity (reduced 50% versus sham, p<0.001) in the remnant organ 72 hours after small bowel resection. Morphometric analysis also showed reduction in hepatocellular size in animals subjected to small bowel resection (reduced 15% versus sham, p<0.02). Protein immunoblot analysis demonstrated increased expression of LC3-II, a marker of autophagic induction, and an increase in the ratio of Bax:Bcl-2 protein expression, an indicator of activation of pro-apoptotic signaling, in the livers of bowel-resected animals (p<0.01 versus sham). Summary and Conclusions: The studies reported here show that total and functional liver mass is reduced following partial small bowel resection, and this reduction is associated with induction of pro-autophagic and pro-apoptotic signaling events. These data have important implications with respect to the pathogenesis and management of liver disease seen in the setting of partial small bowel resection, for example in neonates who undergo substantial small bowel resection for necrotizing enterocolitis.
Background: Previous reports have questioned the validity of conventional solid state manometry, citing low interobserver agreement. The development of high-resolution manometry (HRM) catheters and graphic visualization software may offer advantages over conventional methods including less interobserver variability. The aim of this study was to compare interobserver variability between High Resolution (HRM) and Conventional Manometry (CM). Methods: 50 healthy volunteers underwent duplicate high resolution and conventional manometry studies performed on the same day in each subject. Studies were interpreted by 2 groups of three esophageal motility fellows (Group I:. 2005-6 vs. Group II: - 20078), Group I during the introduction of HRM into the diagnostic laboratory. All 6 individuals separately analyzed each of the 50 HRM and CM studies. Eleven outcome variables were measured for each technique:e including; lower esophageal sphincter (LES) resting pressure, residual pressure, abdominal length, total length, esophageal body contraction amplitudes in each of three segments and the percent simultaneous, failed and peristaltic contractions. Intraclass Correlation Coefficients (ICC) were calculated for the two groups of three and all six fellows separately. ICC > 0.75 was taken as good/excellent reproducibility. Results: Interobserver variability in CM was poor (ICC=0.07951-0.58124)in for every measured parameter with the exception of measurements of esophageal body pressures. Interobserver variability in HRM was excellent for measurements of LES resting and residual pressures, percent of failed contractions and esophageal body pressures at 5 and 10 cm but remained poor for measurement of LES lengths and % simultaneous waves (ICC=0.94406, 0.83196, 0.79124, 0.89381, 0.90312, and 0.21608, 0.2486, 0.18469 respectively). When values for the groups were averaged, interobserver agreement was superior in all measurements for HRM compared to CM and for Group II compared to Group I. Conclusion: Interobserver variability in the interpretation of high resolution manometry data is less than that of conventional manometry and improves with time suggesting an institutional learning curve. Further standardization of HRM analysis is needed. 1024 The National Mortality Burden and Significant Factors Associated with Open and Laparoscopic Cholecystectomy: 1997-2006 James P. Dolan, Brian S. Diggs, Brett C. Sheppard, John G. Hunter
1026
PURPOSE: To determine the mortality rate, and significant associated factors, related to laparoscopic (LC) and open cholecystectomies (OC) over a 10-year period in the United States. METHODS: Using the Nationwide Inpatient Sample (NIS) of approximately seven million inpatient records per year, we extracted and analyzed data for both LC and OC between 1997 and 2006. Cholecystectomies performed as part of another primary procedure were excluded. Using procedure-specific codes, we calculated annual national volumes for both open and laparoscopic cholecystectomies for the time period under review and the associated in-hospital mortality rate following both of these procedures. Using logistic regression modeling, we then analyzed selected patient (age, gender, admission type, diagnosis) and institutional (hospital location and teaching status) characteristics to determine if a significant association existed between these factors and in-hospital mortality. RESULTS: There was an 8.9% increase in the volume of LC and a corresponding decrease in open procedures over the 10 years under review. In 2006, 19% of cholecystectomies were still performed using an open approach and the associated mortality remained significantly higher than that seen with LC (Table 1). Overall, after adjusting for patient and hospital characteristics, the mortality for OC was higher than that for LC (Odds Ratio 4.99 (95% CI 4.77-5.22), p<0.001). Age (> 61 years), male gender, non-elective admission, a primary diagnosis other than cholelithiasis and hospital location were all independently associated with increased mortality. The average mortality rate associated with conversion from LC to OC was found to be 0.7%. CONCLUSIONS: These data indicate an increase in the proportion LCs performed over the years under study with a decrease in the proportion of OCs. However, OCs remains associated with a significant mortality burden when compared with the laparoscopic approach. Table 1. National volumes for cholecystectomy and associated mortality from 1997 to 2006.
Purpose: The American College of Surgeons National Surgical Quality Improvement Program(ACS-NSQIP) created a prediction model for surgical site infection(SSI). This ASCNSQIP model is for superficial, deep and organ space SSI(OrgSSI). Most SSI are superficial and cause less severe morbidity than OrgSSI (abdominal or pelvic abscess). We hypothesized that OrgSSI are a unique type of SSI and therefore they are associated with a unique set of risk factors. The aim of our study was to create a predictive model for OrgSSI after small bowel, colon and rectal surgeries. Methods: The 2006 ACS-NSQIP sample(N=12,373) was used to identify the CPT codes for small bowel, colon and rectal laparoscopic(Lx) or open surgical procedures. The following variables were used to build a predictive model of OrgSSI within 30 days post-op: age, gender, BMI, ASA class, smoking, diabetes, steroid use, 30 days previous radiotherapy or surgery, preoperative creatinine and albumin, Lx, wound class, perioperative transfusion, operative time and surgical site. Patients on chronic mechanical ventilation, dialysis, wound infection or sepsis preoperatively were excluded. Results: Our OrgSSI model achieved a C-index of 0.65 when validated in 2007 NSQIP patients(N=7,802). The point value for each variable in the nomogram indicates its relative weight in predicting OrgSSI. A risk calculator designed with our model is available at: http://simpal.com/rcc/ links/orgssi.html Conclusion: This novel and validated nomogram is useful to predict OrgSSI associated with small bowel, colon and rectal surgical procedures. Risk stratification and risk modification are potential uses of this nomogram.
SSAT Abstracts
Predicting Organ Space Surgical Site Infection with a Nomogram Luiz F. Lobato, Brian Wells, Elizabeth Wick, Kevin T. Pronty, Pokala R. Kiran, Feza H. Remzi, Jon D. Vogel
A-884
without drainage. The clinical manifestations of SIAC were: fever (77%), abdominal pain (27%), and nausea/vomiting (12%). 11(22%) patients required two drainage procedures to manage the same collection. No perioperative deaths could be directly attributed to SIAC or percutaneous drainage. On multivariate analysis, a concomitant bilioenteric anastomosis, HAIP placement and right or extended right hepatic hepatectomy were independently associated with SIAC (p<0.05). On univariate analysis, older age, lower pre-operative albumin, longer operative time, bilioenteric anastomosis and right or extended right hepatic lobectomy were associated with biliary fistula. All factors including temperature, white blood cell count, estimated blood loss, Pringle and operative time were similar between patients with infected SIAC compared to those with non infected SIAC. Conclusion: SIAC are uncommon after hepatectomy and effectively managed with percutaneous drainage. One third of patients with SIAC have non-infected/non-bilious collections and these are not easily discerned from infected SIAC based on standard clinical factors. SIAC are most likely to develop after right or extended right hepatic lobectomy, bilioenteric anastomosis and with HAIP placement. 1029
Despite of advances in modern surgical and intensive care treatment mortality seems to remain high in patients with liver cirrhosis undergoing nonhepatic general surgery. In the few existing articles mortality was reported as high as 70% in patients with poor liver function (high CHILD- or MELD-score). Since data are scarce we analyzed our recent experience with cirrhotic patients undergoing emergent or elective nonhepatic general surgery since 2001 in a German University hospital. Methods: Since 2001 138 nonhepatic general surgical procedures (99 intraabdominal, 39 abdominal wall) were performed in patients with proved liver cirrhosis. Liver cirrhosis was preoperatively classified according to the CHILD- (41 CHILD A; 59 B, 38 C) and the MELD-score (MELD median 13). Sixty-nine (50%) of the patients underwent emergent operations. Most abdominal wall operations were for hernias. Intraabdominal operations consisted of GI-tract procedures (n=51), cholecystectomies (n=15) and various others (n=33). Perioperative data were gained by retrospective analysis. The electronic hospital charts included all data for classification of cirrhosis (CHILD, MELD). Results: Overall mortality of all 138 patients was 28% (10% elective surgery, 45% emergent surgery; p<0.001). Mortality was higher after intraabdominal than after abdominal wall operations (35% vs. 8%; p=0.001) or in patients requiring transfusions (47% vs. 6% without transfusions; p<0.001). Mortality increased with the CHILD-score: 10% (Child A), 17% (B), 63% (C; p<0.001) and the MELD-score (12% MELD 6-11; 18% MELD 12-17; 69% MELD > 17; p<0.001). Patients requiring surgery for bleeding or for perforation/ peritonitis (n=34) showed a higher mortaliy than patients without these indications (56% vs. 18; p<0.01). In multivariate risk factor analysis the CHILD-/MELD-scores and transfusions (all p<0.01) but not the indication for surgery or location (intraabdominal vs. abdominal wall) were independent risk factors for a lethal outcome. Conclusions: Our results demonstrate that perioperative mortality remains high in patients with liver cirrhosis undergoing general surgery. Patients with poor liver function and/or need for blood transfusions even had a very high mortality. In our experience both the CHILD- and the MELD-score significantly predicted postoperative mortality. Both scores may, therefore, be applied to preoperatively predict outcome and modify management in patients with liver cirrhosis.
Nomogram predictor of OrgSSI 1027 Impact of KIT and PDGFRA Gene Mutations On Prognosis of Patients with Gastrointestinal Stromal Tumors After Complete Primary Tumor Resection Andrea Frilling, Ying Hou, Florian Grabellus, Frank Weber, Christoph E. Broelsch Background: Although the importance of KIT and PDGFRA mutations in the oncogenesis of gastrointestinal stromal tumors (GIST) is well established, their prognostic role remains controversial. The aim of our study was to investigate the impact of KIT and PDGFRA gene mutations on the prognosis of patients with GIST after complete primary tumor resection. Methods: Genomic DNA from tumor tissue from 184 patients with primary GIST was submitted to mutational analysis. Exons 9, 11, 13, and 17 of the KIT gene and exons 12, 18 of the PDGFRA gene were sequenced. In addition to the mutational status, several clinical and pathological parameters were analyzed and correlated to the risk of recurrence and long-term disease-free survival (DFS). Results: Somatic mutations were detected in 162 tumors (88.0%). Age (p=0.019), clinical stage (p<0.001), mitotic count (p<0.001), and tumor size (p=0.009) were of prognostic relevance on both univariate and multivariate analysis. Five-year DFS was 41.9%. While the presence of a KIT or PDGFRA mutation per se was not associated with tumor recurrence and/or disease- free survival, exon 11 deletion and hemizygous mutation status were highly predictive for poor survival (p=0.014 and p= 0.0004, respectively). Conclusions: KIT exon 11 deletions and somatic loss of the wild-type KIT identified patients with poor prognosis. Age, clinical stage, tumor size, and mitotic count were standard clinicopathologic features that significantly influenced the prognosis. Mutation type of the mitogen receptor c-kit has a potential for predicting the course of the disease and might contribute to management individualization of GIST patients.
1030 Major Complications After Laparoscopic Cholecystectomy:a Simple Risk Score Melissa M. Murphy, Shimul A. Shah, Jessica P. Simons, Nicholas Csikesz, Sing Chau Ng, Zheng Zhou, Jennifer F. Tseng Laparoscopic cholecystectomy (LC) has become extensively used in the US; however, reported morbidity varies widely. A reliable method to determine complication risk might be useful to optimize care. In this study, we developed an integer-based risk score to determine the likelihood of major complications following LC. Methods: Using the Nationwide Inpatient Sample 1998-2006, patient-discharges for LC were identified. Previously validated major complications including acute MI, aspiration pneumonia, procedure-related laceration, postoperative infection, DVT, PE, hemorrhage, and reoperation were assessed. Using preoperative covariates including patient demographics, comorbidities, surgical indication, and hospital characteristics, we used logistic regression and bootstrap methods to generate an integer score for predicting in-hospital complication rates. A randomly-selected 80% of cohort was used to create the risk score, with testing in the remaining 20% validation-set. Results: 561,923 patient-discharges were identified with an overall complication rate of 6.5%. Predictive characteristics incorporated in the model included: age, sex, Charlson comorbidity score, surgical indication, hospital type, and admission type. Integer values were assigned, and used to calculate an additive score. 3 groups were assembled to stratify risk, with a 4-fold gradient for in-hospital complications ranging from 3.2 to 13.5%. In both derivation and validation sets, the score discriminated well, with respective c-statistics of .657 and .656. Conclusion: An integer-based risk score can be used to predict complications following LC, and may assist in preoperative risk stratification and patient counseling. These findings may be useful to determine which patients should be referred to specialized centers and/or providers for treatment of gallbladder disease.
1028 Management of Symptomatic Intraabdominal Collections After 672 Hepatic Resections Pedro Mastrodomenico, Vin Yael, Sofocleous T. Constantinos, Mithat Gonen, Peter J. Allen, Ronald P. DeMatteo, Yuman Fong, Leslie H. Blumgart, William R. Jarnagin, Michael D'Angelica Introduction: Before 1980 patients with symptomatic intraabdominal collections (SIAC) after hepatectomy required surgical drainage. Image guided percutaneous drainage has improved in the last two decades allowing nonsurgical management. We sought to determine the utility of percutaneous drainage in the management of patients with SIAC and analyze factors potentially associated with SIAC and biliary fistula. Patients and Methods: From a prospective database 672 patients who underwent hepatic resection between 2004 to 2006 were identified. Patients requiring drainage for SIAC were recorded and outcome after drainage was analyzed. Patient demographics, laboratory reports, intraoperative data and their association with SIAC and biliary fistula were investigated using chi-square and t-tests as well as multiple logistic regression. Results: Six hundred seventy-two underwent hepatic resection from 20042006; 48(7.1%) developed SIAC; 16 (33%) seroma/hematomas, 22 (46%) biliary fistulas and 10 (21%) abscesses. All patients underwent percutaneous drainage only. A small subhepatic collection was not possible to drain in one patient; however the patient's symptoms resolved
A-885
SSAT Abstracts
SSAT Abstracts
Mortality After Nonhepatic General Surgery in Patients with Concomitant Liver Cirrhosis: An Analysis of 138 Operations in the 2000s Using Child- and MELD Scores Frank Makowiec, Dimitri Mariaskin, Hans-Christian Spangenberg, Ulrich T. Hopt
SSAT Abstracts
The combination of standard laparoscopy and specimen extraction through a natural orifice has the potential to decrease wound related complications. We describe a new technical approach to laparoscopic sigmoidectomy in which the specimen is extracted transanally using a TEM-like proctoscope. Laparoscopic assisted transanal sigmoidectomy (LATAS) could be considered an alternative minimally invasive approach in patients with benign disease of the sigmoid colon. A 44-year-old male with recurrent diverticular disease of the sigmoid colon was consented to undergo the LATAS procedure. Operating time was 92 minutes, no perioperative or postoperative complications occured. 1023
1025
High Resolution Manometry Has Less Interobserver Variability Than Conventional Manometry Attila Dubecz, Renato Salvador, Marek Polomsky, Oliver Gellersen, Daniel Raymond, Carolyn E. Jones, Thomas J. Watson, Jeffrey H. Peters
Murine Functional Liver Mass Is Reduced Following Partial Small Bowel Resection Zhaohua Qiu, Shannon W. Longshore, Brad W. Warner, David A. Rudnick Liver mass is regulated in precise proportion to body mass in healthy animals, and this ratio (liver:body weight) is specifically restored by regeneration following acute injury. Despite extensive analysis of the molecular mechanisms that regulate hepatic regeneration, the specific signals that initiate and terminate the regenerative response and those that determine the ideal liver:body mass ratio have not been elucidated. Previous analyses have suggested the mechanisms involved in such regulation are likely to include signals derived from the bowel via the portal circulation. Therefore, in the studies reported here we investigated the effect of partial small bowel resection on regulation of liver mass in a murine surgical model. Methods: Two month old, male C57BL/6J wildtype mice were subjected to partial small bowel resection or sham surgery (transection without resection) followed by primary reanastamosis, as we have previously described. These animals were allowed to recover, after which they were sacrificed at serial time points for determination of liver:body mass ratio and analysis of liver tissue. Results: In mice subjected to removal of ~50% of proximal small intestine, liver:body weight ratio was decreased 10% at 48 hrs and 20% at 72 hrs after surgery compared to sham operated animals (p<0.01). Similar results were seen in animals in which a comparable amount of distal small bowel was resected. Functional liver mass was also reduced based on analyses of total hepatic protein (reduced 30% versus sham, p<0.01) and total hepatic alanine transaminase (ALT) activity (reduced 50% versus sham, p<0.001) in the remnant organ 72 hours after small bowel resection. Morphometric analysis also showed reduction in hepatocellular size in animals subjected to small bowel resection (reduced 15% versus sham, p<0.02). Protein immunoblot analysis demonstrated increased expression of LC3-II, a marker of autophagic induction, and an increase in the ratio of Bax:Bcl-2 protein expression, an indicator of activation of pro-apoptotic signaling, in the livers of bowel-resected animals (p<0.01 versus sham). Summary and Conclusions: The studies reported here show that total and functional liver mass is reduced following partial small bowel resection, and this reduction is associated with induction of pro-autophagic and pro-apoptotic signaling events. These data have important implications with respect to the pathogenesis and management of liver disease seen in the setting of partial small bowel resection, for example in neonates who undergo substantial small bowel resection for necrotizing enterocolitis.
Background: Previous reports have questioned the validity of conventional solid state manometry, citing low interobserver agreement. The development of high-resolution manometry (HRM) catheters and graphic visualization software may offer advantages over conventional methods including less interobserver variability. The aim of this study was to compare interobserver variability between High Resolution (HRM) and Conventional Manometry (CM). Methods: 50 healthy volunteers underwent duplicate high resolution and conventional manometry studies performed on the same day in each subject. Studies were interpreted by 2 groups of three esophageal motility fellows (Group I:. 2005-6 vs. Group II: - 20078), Group I during the introduction of HRM into the diagnostic laboratory. All 6 individuals separately analyzed each of the 50 HRM and CM studies. Eleven outcome variables were measured for each technique:e including; lower esophageal sphincter (LES) resting pressure, residual pressure, abdominal length, total length, esophageal body contraction amplitudes in each of three segments and the percent simultaneous, failed and peristaltic contractions. Intraclass Correlation Coefficients (ICC) were calculated for the two groups of three and all six fellows separately. ICC > 0.75 was taken as good/excellent reproducibility. Results: Interobserver variability in CM was poor (ICC=0.07951-0.58124)in for every measured parameter with the exception of measurements of esophageal body pressures. Interobserver variability in HRM was excellent for measurements of LES resting and residual pressures, percent of failed contractions and esophageal body pressures at 5 and 10 cm but remained poor for measurement of LES lengths and % simultaneous waves (ICC=0.94406, 0.83196, 0.79124, 0.89381, 0.90312, and 0.21608, 0.2486, 0.18469 respectively). When values for the groups were averaged, interobserver agreement was superior in all measurements for HRM compared to CM and for Group II compared to Group I. Conclusion: Interobserver variability in the interpretation of high resolution manometry data is less than that of conventional manometry and improves with time suggesting an institutional learning curve. Further standardization of HRM analysis is needed. 1024 The National Mortality Burden and Significant Factors Associated with Open and Laparoscopic Cholecystectomy: 1997-2006 James P. Dolan, Brian S. Diggs, Brett C. Sheppard, John G. Hunter
1026
PURPOSE: To determine the mortality rate, and significant associated factors, related to laparoscopic (LC) and open cholecystectomies (OC) over a 10-year period in the United States. METHODS: Using the Nationwide Inpatient Sample (NIS) of approximately seven million inpatient records per year, we extracted and analyzed data for both LC and OC between 1997 and 2006. Cholecystectomies performed as part of another primary procedure were excluded. Using procedure-specific codes, we calculated annual national volumes for both open and laparoscopic cholecystectomies for the time period under review and the associated in-hospital mortality rate following both of these procedures. Using logistic regression modeling, we then analyzed selected patient (age, gender, admission type, diagnosis) and institutional (hospital location and teaching status) characteristics to determine if a significant association existed between these factors and in-hospital mortality. RESULTS: There was an 8.9% increase in the volume of LC and a corresponding decrease in open procedures over the 10 years under review. In 2006, 19% of cholecystectomies were still performed using an open approach and the associated mortality remained significantly higher than that seen with LC (Table 1). Overall, after adjusting for patient and hospital characteristics, the mortality for OC was higher than that for LC (Odds Ratio 4.99 (95% CI 4.77-5.22), p<0.001). Age (> 61 years), male gender, non-elective admission, a primary diagnosis other than cholelithiasis and hospital location were all independently associated with increased mortality. The average mortality rate associated with conversion from LC to OC was found to be 0.7%. CONCLUSIONS: These data indicate an increase in the proportion LCs performed over the years under study with a decrease in the proportion of OCs. However, OCs remains associated with a significant mortality burden when compared with the laparoscopic approach. Table 1. National volumes for cholecystectomy and associated mortality from 1997 to 2006.
Purpose: The American College of Surgeons National Surgical Quality Improvement Program(ACS-NSQIP) created a prediction model for surgical site infection(SSI). This ASCNSQIP model is for superficial, deep and organ space SSI(OrgSSI). Most SSI are superficial and cause less severe morbidity than OrgSSI (abdominal or pelvic abscess). We hypothesized that OrgSSI are a unique type of SSI and therefore they are associated with a unique set of risk factors. The aim of our study was to create a predictive model for OrgSSI after small bowel, colon and rectal surgeries. Methods: The 2006 ACS-NSQIP sample(N=12,373) was used to identify the CPT codes for small bowel, colon and rectal laparoscopic(Lx) or open surgical procedures. The following variables were used to build a predictive model of OrgSSI within 30 days post-op: age, gender, BMI, ASA class, smoking, diabetes, steroid use, 30 days previous radiotherapy or surgery, preoperative creatinine and albumin, Lx, wound class, perioperative transfusion, operative time and surgical site. Patients on chronic mechanical ventilation, dialysis, wound infection or sepsis preoperatively were excluded. Results: Our OrgSSI model achieved a C-index of 0.65 when validated in 2007 NSQIP patients(N=7,802). The point value for each variable in the nomogram indicates its relative weight in predicting OrgSSI. A risk calculator designed with our model is available at: http://simpal.com/rcc/ links/orgssi.html Conclusion: This novel and validated nomogram is useful to predict OrgSSI associated with small bowel, colon and rectal surgical procedures. Risk stratification and risk modification are potential uses of this nomogram.
SSAT Abstracts
Predicting Organ Space Surgical Site Infection with a Nomogram Luiz F. Lobato, Brian Wells, Elizabeth Wick, Kevin T. Pronty, Pokala R. Kiran, Feza H. Remzi, Jon D. Vogel
A-884
without drainage. The clinical manifestations of SIAC were: fever (77%), abdominal pain (27%), and nausea/vomiting (12%). 11(22%) patients required two drainage procedures to manage the same collection. No perioperative deaths could be directly attributed to SIAC or percutaneous drainage. On multivariate analysis, a concomitant bilioenteric anastomosis, HAIP placement and right or extended right hepatic hepatectomy were independently associated with SIAC (p<0.05). On univariate analysis, older age, lower pre-operative albumin, longer operative time, bilioenteric anastomosis and right or extended right hepatic lobectomy were associated with biliary fistula. All factors including temperature, white blood cell count, estimated blood loss, Pringle and operative time were similar between patients with infected SIAC compared to those with non infected SIAC. Conclusion: SIAC are uncommon after hepatectomy and effectively managed with percutaneous drainage. One third of patients with SIAC have non-infected/non-bilious collections and these are not easily discerned from infected SIAC based on standard clinical factors. SIAC are most likely to develop after right or extended right hepatic lobectomy, bilioenteric anastomosis and with HAIP placement. 1029
Despite of advances in modern surgical and intensive care treatment mortality seems to remain high in patients with liver cirrhosis undergoing nonhepatic general surgery. In the few existing articles mortality was reported as high as 70% in patients with poor liver function (high CHILD- or MELD-score). Since data are scarce we analyzed our recent experience with cirrhotic patients undergoing emergent or elective nonhepatic general surgery since 2001 in a German University hospital. Methods: Since 2001 138 nonhepatic general surgical procedures (99 intraabdominal, 39 abdominal wall) were performed in patients with proved liver cirrhosis. Liver cirrhosis was preoperatively classified according to the CHILD- (41 CHILD A; 59 B, 38 C) and the MELD-score (MELD median 13). Sixty-nine (50%) of the patients underwent emergent operations. Most abdominal wall operations were for hernias. Intraabdominal operations consisted of GI-tract procedures (n=51), cholecystectomies (n=15) and various others (n=33). Perioperative data were gained by retrospective analysis. The electronic hospital charts included all data for classification of cirrhosis (CHILD, MELD). Results: Overall mortality of all 138 patients was 28% (10% elective surgery, 45% emergent surgery; p<0.001). Mortality was higher after intraabdominal than after abdominal wall operations (35% vs. 8%; p=0.001) or in patients requiring transfusions (47% vs. 6% without transfusions; p<0.001). Mortality increased with the CHILD-score: 10% (Child A), 17% (B), 63% (C; p<0.001) and the MELD-score (12% MELD 6-11; 18% MELD 12-17; 69% MELD > 17; p<0.001). Patients requiring surgery for bleeding or for perforation/ peritonitis (n=34) showed a higher mortaliy than patients without these indications (56% vs. 18; p<0.01). In multivariate risk factor analysis the CHILD-/MELD-scores and transfusions (all p<0.01) but not the indication for surgery or location (intraabdominal vs. abdominal wall) were independent risk factors for a lethal outcome. Conclusions: Our results demonstrate that perioperative mortality remains high in patients with liver cirrhosis undergoing general surgery. Patients with poor liver function and/or need for blood transfusions even had a very high mortality. In our experience both the CHILD- and the MELD-score significantly predicted postoperative mortality. Both scores may, therefore, be applied to preoperatively predict outcome and modify management in patients with liver cirrhosis.
Nomogram predictor of OrgSSI 1027 Impact of KIT and PDGFRA Gene Mutations On Prognosis of Patients with Gastrointestinal Stromal Tumors After Complete Primary Tumor Resection Andrea Frilling, Ying Hou, Florian Grabellus, Frank Weber, Christoph E. Broelsch Background: Although the importance of KIT and PDGFRA mutations in the oncogenesis of gastrointestinal stromal tumors (GIST) is well established, their prognostic role remains controversial. The aim of our study was to investigate the impact of KIT and PDGFRA gene mutations on the prognosis of patients with GIST after complete primary tumor resection. Methods: Genomic DNA from tumor tissue from 184 patients with primary GIST was submitted to mutational analysis. Exons 9, 11, 13, and 17 of the KIT gene and exons 12, 18 of the PDGFRA gene were sequenced. In addition to the mutational status, several clinical and pathological parameters were analyzed and correlated to the risk of recurrence and long-term disease-free survival (DFS). Results: Somatic mutations were detected in 162 tumors (88.0%). Age (p=0.019), clinical stage (p<0.001), mitotic count (p<0.001), and tumor size (p=0.009) were of prognostic relevance on both univariate and multivariate analysis. Five-year DFS was 41.9%. While the presence of a KIT or PDGFRA mutation per se was not associated with tumor recurrence and/or disease- free survival, exon 11 deletion and hemizygous mutation status were highly predictive for poor survival (p=0.014 and p= 0.0004, respectively). Conclusions: KIT exon 11 deletions and somatic loss of the wild-type KIT identified patients with poor prognosis. Age, clinical stage, tumor size, and mitotic count were standard clinicopathologic features that significantly influenced the prognosis. Mutation type of the mitogen receptor c-kit has a potential for predicting the course of the disease and might contribute to management individualization of GIST patients.
1030 Major Complications After Laparoscopic Cholecystectomy:a Simple Risk Score Melissa M. Murphy, Shimul A. Shah, Jessica P. Simons, Nicholas Csikesz, Sing Chau Ng, Zheng Zhou, Jennifer F. Tseng Laparoscopic cholecystectomy (LC) has become extensively used in the US; however, reported morbidity varies widely. A reliable method to determine complication risk might be useful to optimize care. In this study, we developed an integer-based risk score to determine the likelihood of major complications following LC. Methods: Using the Nationwide Inpatient Sample 1998-2006, patient-discharges for LC were identified. Previously validated major complications including acute MI, aspiration pneumonia, procedure-related laceration, postoperative infection, DVT, PE, hemorrhage, and reoperation were assessed. Using preoperative covariates including patient demographics, comorbidities, surgical indication, and hospital characteristics, we used logistic regression and bootstrap methods to generate an integer score for predicting in-hospital complication rates. A randomly-selected 80% of cohort was used to create the risk score, with testing in the remaining 20% validation-set. Results: 561,923 patient-discharges were identified with an overall complication rate of 6.5%. Predictive characteristics incorporated in the model included: age, sex, Charlson comorbidity score, surgical indication, hospital type, and admission type. Integer values were assigned, and used to calculate an additive score. 3 groups were assembled to stratify risk, with a 4-fold gradient for in-hospital complications ranging from 3.2 to 13.5%. In both derivation and validation sets, the score discriminated well, with respective c-statistics of .657 and .656. Conclusion: An integer-based risk score can be used to predict complications following LC, and may assist in preoperative risk stratification and patient counseling. These findings may be useful to determine which patients should be referred to specialized centers and/or providers for treatment of gallbladder disease.
1028 Management of Symptomatic Intraabdominal Collections After 672 Hepatic Resections Pedro Mastrodomenico, Vin Yael, Sofocleous T. Constantinos, Mithat Gonen, Peter J. Allen, Ronald P. DeMatteo, Yuman Fong, Leslie H. Blumgart, William R. Jarnagin, Michael D'Angelica Introduction: Before 1980 patients with symptomatic intraabdominal collections (SIAC) after hepatectomy required surgical drainage. Image guided percutaneous drainage has improved in the last two decades allowing nonsurgical management. We sought to determine the utility of percutaneous drainage in the management of patients with SIAC and analyze factors potentially associated with SIAC and biliary fistula. Patients and Methods: From a prospective database 672 patients who underwent hepatic resection between 2004 to 2006 were identified. Patients requiring drainage for SIAC were recorded and outcome after drainage was analyzed. Patient demographics, laboratory reports, intraoperative data and their association with SIAC and biliary fistula were investigated using chi-square and t-tests as well as multiple logistic regression. Results: Six hundred seventy-two underwent hepatic resection from 20042006; 48(7.1%) developed SIAC; 16 (33%) seroma/hematomas, 22 (46%) biliary fistulas and 10 (21%) abscesses. All patients underwent percutaneous drainage only. A small subhepatic collection was not possible to drain in one patient; however the patient's symptoms resolved
A-885
SSAT Abstracts
SSAT Abstracts
Mortality After Nonhepatic General Surgery in Patients with Concomitant Liver Cirrhosis: An Analysis of 138 Operations in the 2000s Using Child- and MELD Scores Frank Makowiec, Dimitri Mariaskin, Hans-Christian Spangenberg, Ulrich T. Hopt