J Gastrointest Surg DOI 10.1007/s11605-015-2855-7
ORIGINAL ARTICLE
A Simple Method to Evaluate Whether Pancreas Texture Can Be Used to Predict Pancreatic Fistula Risk After Pancreatoduodenectomy Samet Yardimci 1 & Yalçın Burak Kara 1 & Davut Tuney 2 & Wafi Attaallah 1 & Mustafa Umit Ugurlu 1 & Ender Dulundu 1 & Şevket Cumhur Yegen 1
Received: 7 April 2015 / Accepted: 4 May 2015 # 2015 The Society for Surgery of the Alimentary Tract
Abstract Introduction Soft pancreas is one of the most important risk factor for postoperative pancreatic fistula after pancreatoduodenectomy. The aim of this study is to investigate whether pancreatic attenuation index utilized to assess the pancreatic texture with computed tomography can be used to predict the risk of developing a clinically relevant postoperative pancreatic fistula after pancreatoduodenectomy. Methods We reviewed 76 consecutive patients undergoing pancreatoduodenectomy between 2012 and 2014. The pancreatic attenuation index is found by dividing the pancreas density by the spleen density achieved with non-enhanced computed tomography. The independent predictors of clinically relevant postoperative pancreatic fistula were investigated. Results Clinically relevant postoperative pancreatic fistula occurred in 13 patients (17.1 %). The group of patients with postoperative pancreatic fistula is compared with the group of patients without postoperative pancreatic fistula in terms of age, gender, body mass index, the American Society of Anesthesiologists (ASA) score, smoking, alcohol consumption, medical comorbidities, preoperative biliary drainage, type of anastomosis, and pancreatic duct size and pancreatic attenuation index. Univariate analyses have shown a significant difference in relation to chronic obstructive pulmonary disease and pancreatic attenuation index. The multivariate analyses showed that only pancreatic attenuation index was associated with a high postoperative pancreatic fistula rate (P=0.012). Conclusion A preoperative non-contrast computed tomography scan evaluating pancreatic attenuation index could help to predict the occurrence of clinically significant postoperative pancreatic fistula after pancreatoduodenectomy. Keywords Pancreatoduodenectomy . Computed tomography . Pancreaticojejunostomy . Pancreatic fistula
* Samet Yardimci [email protected] 1
Department of General Surgery, Marmara University Pendik Education and Research Hospital, Mimar Sinan C. Marmara Universitesi Pendik EAH Genel Cerrahi Klinigi, Ust Kaynarca, Pendik, Istanbul, Turkey
2
Department of Radiology, Marmara University Pendik Education and Research Hospital, Istanbul, Turkey
Introduction Pancreatoduodenectomy (PD) is the standard surgical treatment used for resection of most malignant and benign neoplasms in the periampullary region. PD has been performed in high-volume centers with a less than 4 % mortality rate over the last decades.1,2 In contrast, postoperative morbidity remains high (30–50 %) even in the high-volume centers.3–6 One of the most important reasons of morbidity after PD is postoperative pancreatic fistula (POPF), which occurs in 5– 30 % of cases.1,6–8 Although the conservative treatment of POPF usually is successful,9 it results in prolonged hospital stay and increased cost and readmissions, and it can also lead to more severe complications with a mortality rate, ranging from 3 to 15 %.2,9,10
J Gastrointest Surg
Identifying patients with a high risk for POPF is critical to further decrease morbidity and mortality after PD. One of the most important risk factors for POPF is soft pancreas. Pancreatic texture is mainly influenced by two factors. One of them is fatty infiltration of the pancreas and the other is parenchymal fibrosis. Absence of parenchymal fibrosis11,12 and fatty infiltration of the pancreatic parenchyma were reported as independent risk factors for POPF.11,13,14 Although soft pancreas is the most widely accepted risk factor for developing a POPF following PD,15–17 there is no absolutely objective method to evaluate the softness of the pancreas. So far, only few radiological studies have addressed the issue of whether the texture of pancreas can be reliably detected and quantified.14,18–21 Though significant results were reported in these studies, these methods have not been practiced widely because they are difficult to apply and calculate. Pancreatic attenuation index (PAI), defined as the ratio of the pancreatic attenuation to splenic attenuation, is an easy method to evaluate pancreatic texture. The aim of this study is to investigate whether a simple method (PAI) with using CT to assess the pancreatic texture can predict the risk of developing a clinically relevant POPF after PD.
Surgical Technique Standard PD with pancreatojejunostomy was performed. An end-to-side hepaticojejunostomy was performed with interrupted sutures approximately 10 cm distally. Approximately 40 cm distal to the biliary anastomosis, an antecolic, side-to-side gastrojejunostomy was performed. Two closedsuction drains were routinely placed in proximity to the pancreatic and biliary anastomosis.
Perioperative Care All patients were managed according to a standard postoperative clinical pathway. Somatostatin analogs were not used preoperatively. The nasogastric tube was removed on the first postoperative day. A clear liquid diet was started on the second postoperative day and advanced to a regular diet as tolerated. The amylase concentration in the drain fluid was measured on the 1st, 3rd, and 5th postoperative day. All patients with symptoms or findings concerning for POPF underwent a contrast-enhanced CT. The patients were discharged from the hospital when solid diet was tolerated, and postoperative pain was controlled with oral analgesics.
Methods CT Evaluation Patients and Clinical Data Collection Using a prospectively maintained database, we identified 76 consecutive patients undergoing PD between September 2012 and August 2014. Demographic data, body mass index, preoperative American Society of Anesthesiologists (ASA) score, medical comorbidities, preoperative biliary drainage, type of anastomosis (duct to mucosa vs. dunking), perioperative data, histopathological evaluation, and clinical follow-up were reviewed.
Definitions of Pancreatic Anastomotic Failure POPF, pancreatic anastomotic leak, and pancreatic fistula were considered synonyms for the purpose of this study. POPF was defined and classified in accordance with the International Study Group of Pancreatic Surgery (ISGPS) classification.22 The ISGPS definition of POPF provides three levels of severity. Whereas grades B and C are clinically relevant as they require changes in the postoperative management and further diagnostic and therapeutic interventions, grade A is not. For the purpose of this study, we combined grades B and C in the clinically relevant POPF group, while grade A and no POPF were combined in the no POPF group.
The dual-phase pancreatic CT protocol includes an unenhanced scan followed by dual-phase contrast-enhanced scans through the abdomen. All CT attenuation values were independently measured by one expert radiologist (DT) on the electronically stored CT images without knowledge of the clinical information. CT images taken for diagnostic and staging purposes within 30 days before PD were included in this analysis. CT attenuation values on unenhanced images were measured. CT attenuation values of the pancreatic parenchyma were quantified using Hounsfield unit thresholds, placed on a region of interest (ROI) in six points in two segments of the pancreas (body and tail). The mean value for the two segments was computed for each patient. ROIs in the body were placed over the superior mesenteric artery in an area unaffected by the tumor and in the tail about two thirds distal from the body–tail transition. The largest possible spherical ROI was placed making every effort to avoid the pancreatic duct and extrapancreatic structures. PAI was calculated by dividing the pancreatic density by the splenic density. The spleen was chosen as a reference because it does not contain any fat, and its density does not vary with body mass index (BMI). Additionally, pancreatic duct size was measured by the same radiologist, using the same CT images.
J Gastrointest Surg
Statistical Analyses All data were analyzed using SPSS 17.0 statistical software package (SPSS, Inc., Chicago, IL, USA). Background clinical data were analyzed using the t test or Mann–Whitney U test for continuous data and Fisher’s exact test or the chi-squared test for categorical data. All tests were two-sided and P values below 0.05 were considered statistically significant. A multivariate analysis was carried out using a model of logistic regression. Receiver operating curve (ROC) analysis was used to find the most sensitive cutoff value of the PAI.
Results Patient Characteristics Between September 2012 and August 2014, 76 consecutive patients underwent PD in the General Surgery Clinic at Marmara University School of Medicine Pendik Training and Research Hospital. The median age of the patients was 61 (30–82)years, and 45 (59 %) patients were male. The median BMI was 26 (17–43). The majority (n=44; 58 %) had an ASA score of 2. The comorbidities were determined in the patients as the following: diabetes mellitus in 23 (30 %), hypertension in 34 (45 %), chronic obstructive pulmonary disease (COPD) in 33 (43 %), and chronic cardiovascular disease (CCVD) in 26 (34 %) patients. Smoking was determined in 3 (49 %) while alcohol consumption was determined in only 2 (3 %) patients. Among those, 44 (58 %) patients were preoperatively diagnosed as head of pancreas tumors, whereas 32 (42 %) patients were diagnosed as other periampullary tumors. History of pancreatitis was determined in 9 (12 %) patients. Preoperative biliary drainage was performed for 47 (62 %) patients (26 (34 %) with internal and 21 (28 %) with external drainage). According to CT scan, the diameter of the main pancreatic duct was ≤3 mm in 48 (63 %) patients, while it was >3 mm in 28 (37 %) patients (Table 1). Perioperative Findings and Postoperative Follow-up Outcomes The type of anastomosis was pancreatic duct-to-jejunal anastomosis in 53 (70 %) patients and Bdunking^ procedure in 23 (30 %) patients. POPF was determined in the patients as the following: grade A in 17 (22 %), grade B in 9 (12 %), grade C in 4 (5 %), and no PAI in 46 (61 %) patients. Mortality related to POPF was seen in 2 patients (3 %), while it was seen in 1 (1 %) patient independent of POPF. Blood transfusion was needed in 40 (53 %) patients. Definitive pathology reports showed ductal adenocarcinoma in 28 (37 %), cystic neoplasm in 3 (4 %), neuroendocrine tumor in 5 (7 %), chronic pancreatitis in 1 (1 %), and other
Table 1
Demographics and preoperative findings of the study cohort
All patients
N=76
Median age Gender Male Female Body mass index ASA score 1 2 3 4 Comorbidity DM Hypertension COPD CCVD Smoking Alcohol consumption
61 (30–82) 45 (59 %) 31 (41 %) 26 (17–43) 8 (11 %) 44 (58 %) 23 (30 %) 1 (1 %) 23 (30 %) 34 (45 %) 33 (43 %) 26 (34 %) 37 (49 %) 2 (3 %)
History of pancreatitis 9 (12 %) Preoperative biliary drainage Internal drainage 26 (34 %) External drainage 21 (28 %) Preoperative diagnosis Head of pancreas tumor 44 (58 %) Other periampullary tumors 32 (42 %) Diameter of the main pancreatic duct according to CT scan (mm) >3 28 (37 %) ≤3 48 (63 %) ASA American Society of Anesthesiologists, DM diabetes mellitus, COPD chronic obstructive pulmonary disease, CCVD chronic cardiovascular disease
periampullary tumors in 37 (49 %), and common bile duct stone appeared like a tumor in 2 (3.2 %) patients (Table 2).
Comparing the Groups With or Without POPF Univariate analyses have shown significant differences between the group of patients with POPF (n=13) and the group of patients without POPF (n=63) in relation to COPD and PAI (Table 3). However, there was no significant difference between the two groups in relation to age, gender, body mass index, preoperative ASA score, smoking, alcohol consumption, other medical comorbidities, preoperative biliary drainage, and type of anastomosis (duct to mucosa vs. dunking). Interestingly, the multivariate analyses showed that only PAI was associated with a high POPF rate (P=0.012) (Table 3). This means that the PAI which is a measure of the texture of the pancreas is an independent predicting factor for POPF.
J Gastrointest Surg Table 2
Perioperative findings and postoperative outcomes
Type of anastomosis Duct to mucosa Dunking Pancreatic fistula Grade A Grade B Grade C None Other complications Gastrointestinal bleeding Delayed gastric emptying Other Mortality PAF related PAF independent Transfusion Pathologic diagnosis Ductal adenocarcinoma Cystic neoplasm Neuroendocrine tumor Chronic pancreatitis Other periampullary tumors Common bile duct stone
53 (70 %) 23 (30 %) 17 (22 %) 9 (12 %) 4 (5 %) 46 (61 %) 8 (11 %) 2 (3 %) 4 (5 %) 2 (3 %) 1 (1 %) 40 (53 %) 28 (37 %) 3 (4 %) 5 (7 %) 1 (1 %) 37 (49 %) 2 (3 %)
Also, these results have shown that PAI is a more powerful predicting factor than the other known predicting factors. Outcomes According to a Proposed PAI Cutoff Value The median PAI according to CT scan measures was 0.67 (0.01–1.18) for the study cohort. In an attempt to find a cutoff value for predicting the rate of POPF, the study population was divided into two groups according to the median number (0.67) of PAI for all cohorts. The patients with PAI lower than 0.67 (n=37) were compared with the patients with PAI ≥0.67 (n=39). We found that the POPF rate was significantly higher in the group with PAI
ORIGINAL ARTICLE
A Simple Method to Evaluate Whether Pancreas Texture Can Be Used to Predict Pancreatic Fistula Risk After Pancreatoduodenectomy Samet Yardimci 1 & Yalçın Burak Kara 1 & Davut Tuney 2 & Wafi Attaallah 1 & Mustafa Umit Ugurlu 1 & Ender Dulundu 1 & Şevket Cumhur Yegen 1
Received: 7 April 2015 / Accepted: 4 May 2015 # 2015 The Society for Surgery of the Alimentary Tract
Abstract Introduction Soft pancreas is one of the most important risk factor for postoperative pancreatic fistula after pancreatoduodenectomy. The aim of this study is to investigate whether pancreatic attenuation index utilized to assess the pancreatic texture with computed tomography can be used to predict the risk of developing a clinically relevant postoperative pancreatic fistula after pancreatoduodenectomy. Methods We reviewed 76 consecutive patients undergoing pancreatoduodenectomy between 2012 and 2014. The pancreatic attenuation index is found by dividing the pancreas density by the spleen density achieved with non-enhanced computed tomography. The independent predictors of clinically relevant postoperative pancreatic fistula were investigated. Results Clinically relevant postoperative pancreatic fistula occurred in 13 patients (17.1 %). The group of patients with postoperative pancreatic fistula is compared with the group of patients without postoperative pancreatic fistula in terms of age, gender, body mass index, the American Society of Anesthesiologists (ASA) score, smoking, alcohol consumption, medical comorbidities, preoperative biliary drainage, type of anastomosis, and pancreatic duct size and pancreatic attenuation index. Univariate analyses have shown a significant difference in relation to chronic obstructive pulmonary disease and pancreatic attenuation index. The multivariate analyses showed that only pancreatic attenuation index was associated with a high postoperative pancreatic fistula rate (P=0.012). Conclusion A preoperative non-contrast computed tomography scan evaluating pancreatic attenuation index could help to predict the occurrence of clinically significant postoperative pancreatic fistula after pancreatoduodenectomy. Keywords Pancreatoduodenectomy . Computed tomography . Pancreaticojejunostomy . Pancreatic fistula
* Samet Yardimci [email protected] 1
Department of General Surgery, Marmara University Pendik Education and Research Hospital, Mimar Sinan C. Marmara Universitesi Pendik EAH Genel Cerrahi Klinigi, Ust Kaynarca, Pendik, Istanbul, Turkey
2
Department of Radiology, Marmara University Pendik Education and Research Hospital, Istanbul, Turkey
Introduction Pancreatoduodenectomy (PD) is the standard surgical treatment used for resection of most malignant and benign neoplasms in the periampullary region. PD has been performed in high-volume centers with a less than 4 % mortality rate over the last decades.1,2 In contrast, postoperative morbidity remains high (30–50 %) even in the high-volume centers.3–6 One of the most important reasons of morbidity after PD is postoperative pancreatic fistula (POPF), which occurs in 5– 30 % of cases.1,6–8 Although the conservative treatment of POPF usually is successful,9 it results in prolonged hospital stay and increased cost and readmissions, and it can also lead to more severe complications with a mortality rate, ranging from 3 to 15 %.2,9,10
J Gastrointest Surg
Identifying patients with a high risk for POPF is critical to further decrease morbidity and mortality after PD. One of the most important risk factors for POPF is soft pancreas. Pancreatic texture is mainly influenced by two factors. One of them is fatty infiltration of the pancreas and the other is parenchymal fibrosis. Absence of parenchymal fibrosis11,12 and fatty infiltration of the pancreatic parenchyma were reported as independent risk factors for POPF.11,13,14 Although soft pancreas is the most widely accepted risk factor for developing a POPF following PD,15–17 there is no absolutely objective method to evaluate the softness of the pancreas. So far, only few radiological studies have addressed the issue of whether the texture of pancreas can be reliably detected and quantified.14,18–21 Though significant results were reported in these studies, these methods have not been practiced widely because they are difficult to apply and calculate. Pancreatic attenuation index (PAI), defined as the ratio of the pancreatic attenuation to splenic attenuation, is an easy method to evaluate pancreatic texture. The aim of this study is to investigate whether a simple method (PAI) with using CT to assess the pancreatic texture can predict the risk of developing a clinically relevant POPF after PD.
Surgical Technique Standard PD with pancreatojejunostomy was performed. An end-to-side hepaticojejunostomy was performed with interrupted sutures approximately 10 cm distally. Approximately 40 cm distal to the biliary anastomosis, an antecolic, side-to-side gastrojejunostomy was performed. Two closedsuction drains were routinely placed in proximity to the pancreatic and biliary anastomosis.
Perioperative Care All patients were managed according to a standard postoperative clinical pathway. Somatostatin analogs were not used preoperatively. The nasogastric tube was removed on the first postoperative day. A clear liquid diet was started on the second postoperative day and advanced to a regular diet as tolerated. The amylase concentration in the drain fluid was measured on the 1st, 3rd, and 5th postoperative day. All patients with symptoms or findings concerning for POPF underwent a contrast-enhanced CT. The patients were discharged from the hospital when solid diet was tolerated, and postoperative pain was controlled with oral analgesics.
Methods CT Evaluation Patients and Clinical Data Collection Using a prospectively maintained database, we identified 76 consecutive patients undergoing PD between September 2012 and August 2014. Demographic data, body mass index, preoperative American Society of Anesthesiologists (ASA) score, medical comorbidities, preoperative biliary drainage, type of anastomosis (duct to mucosa vs. dunking), perioperative data, histopathological evaluation, and clinical follow-up were reviewed.
Definitions of Pancreatic Anastomotic Failure POPF, pancreatic anastomotic leak, and pancreatic fistula were considered synonyms for the purpose of this study. POPF was defined and classified in accordance with the International Study Group of Pancreatic Surgery (ISGPS) classification.22 The ISGPS definition of POPF provides three levels of severity. Whereas grades B and C are clinically relevant as they require changes in the postoperative management and further diagnostic and therapeutic interventions, grade A is not. For the purpose of this study, we combined grades B and C in the clinically relevant POPF group, while grade A and no POPF were combined in the no POPF group.
The dual-phase pancreatic CT protocol includes an unenhanced scan followed by dual-phase contrast-enhanced scans through the abdomen. All CT attenuation values were independently measured by one expert radiologist (DT) on the electronically stored CT images without knowledge of the clinical information. CT images taken for diagnostic and staging purposes within 30 days before PD were included in this analysis. CT attenuation values on unenhanced images were measured. CT attenuation values of the pancreatic parenchyma were quantified using Hounsfield unit thresholds, placed on a region of interest (ROI) in six points in two segments of the pancreas (body and tail). The mean value for the two segments was computed for each patient. ROIs in the body were placed over the superior mesenteric artery in an area unaffected by the tumor and in the tail about two thirds distal from the body–tail transition. The largest possible spherical ROI was placed making every effort to avoid the pancreatic duct and extrapancreatic structures. PAI was calculated by dividing the pancreatic density by the splenic density. The spleen was chosen as a reference because it does not contain any fat, and its density does not vary with body mass index (BMI). Additionally, pancreatic duct size was measured by the same radiologist, using the same CT images.
J Gastrointest Surg
Statistical Analyses All data were analyzed using SPSS 17.0 statistical software package (SPSS, Inc., Chicago, IL, USA). Background clinical data were analyzed using the t test or Mann–Whitney U test for continuous data and Fisher’s exact test or the chi-squared test for categorical data. All tests were two-sided and P values below 0.05 were considered statistically significant. A multivariate analysis was carried out using a model of logistic regression. Receiver operating curve (ROC) analysis was used to find the most sensitive cutoff value of the PAI.
Results Patient Characteristics Between September 2012 and August 2014, 76 consecutive patients underwent PD in the General Surgery Clinic at Marmara University School of Medicine Pendik Training and Research Hospital. The median age of the patients was 61 (30–82)years, and 45 (59 %) patients were male. The median BMI was 26 (17–43). The majority (n=44; 58 %) had an ASA score of 2. The comorbidities were determined in the patients as the following: diabetes mellitus in 23 (30 %), hypertension in 34 (45 %), chronic obstructive pulmonary disease (COPD) in 33 (43 %), and chronic cardiovascular disease (CCVD) in 26 (34 %) patients. Smoking was determined in 3 (49 %) while alcohol consumption was determined in only 2 (3 %) patients. Among those, 44 (58 %) patients were preoperatively diagnosed as head of pancreas tumors, whereas 32 (42 %) patients were diagnosed as other periampullary tumors. History of pancreatitis was determined in 9 (12 %) patients. Preoperative biliary drainage was performed for 47 (62 %) patients (26 (34 %) with internal and 21 (28 %) with external drainage). According to CT scan, the diameter of the main pancreatic duct was ≤3 mm in 48 (63 %) patients, while it was >3 mm in 28 (37 %) patients (Table 1). Perioperative Findings and Postoperative Follow-up Outcomes The type of anastomosis was pancreatic duct-to-jejunal anastomosis in 53 (70 %) patients and Bdunking^ procedure in 23 (30 %) patients. POPF was determined in the patients as the following: grade A in 17 (22 %), grade B in 9 (12 %), grade C in 4 (5 %), and no PAI in 46 (61 %) patients. Mortality related to POPF was seen in 2 patients (3 %), while it was seen in 1 (1 %) patient independent of POPF. Blood transfusion was needed in 40 (53 %) patients. Definitive pathology reports showed ductal adenocarcinoma in 28 (37 %), cystic neoplasm in 3 (4 %), neuroendocrine tumor in 5 (7 %), chronic pancreatitis in 1 (1 %), and other
Table 1
Demographics and preoperative findings of the study cohort
All patients
N=76
Median age Gender Male Female Body mass index ASA score 1 2 3 4 Comorbidity DM Hypertension COPD CCVD Smoking Alcohol consumption
61 (30–82) 45 (59 %) 31 (41 %) 26 (17–43) 8 (11 %) 44 (58 %) 23 (30 %) 1 (1 %) 23 (30 %) 34 (45 %) 33 (43 %) 26 (34 %) 37 (49 %) 2 (3 %)
History of pancreatitis 9 (12 %) Preoperative biliary drainage Internal drainage 26 (34 %) External drainage 21 (28 %) Preoperative diagnosis Head of pancreas tumor 44 (58 %) Other periampullary tumors 32 (42 %) Diameter of the main pancreatic duct according to CT scan (mm) >3 28 (37 %) ≤3 48 (63 %) ASA American Society of Anesthesiologists, DM diabetes mellitus, COPD chronic obstructive pulmonary disease, CCVD chronic cardiovascular disease
periampullary tumors in 37 (49 %), and common bile duct stone appeared like a tumor in 2 (3.2 %) patients (Table 2).
Comparing the Groups With or Without POPF Univariate analyses have shown significant differences between the group of patients with POPF (n=13) and the group of patients without POPF (n=63) in relation to COPD and PAI (Table 3). However, there was no significant difference between the two groups in relation to age, gender, body mass index, preoperative ASA score, smoking, alcohol consumption, other medical comorbidities, preoperative biliary drainage, and type of anastomosis (duct to mucosa vs. dunking). Interestingly, the multivariate analyses showed that only PAI was associated with a high POPF rate (P=0.012) (Table 3). This means that the PAI which is a measure of the texture of the pancreas is an independent predicting factor for POPF.
J Gastrointest Surg Table 2
Perioperative findings and postoperative outcomes
Type of anastomosis Duct to mucosa Dunking Pancreatic fistula Grade A Grade B Grade C None Other complications Gastrointestinal bleeding Delayed gastric emptying Other Mortality PAF related PAF independent Transfusion Pathologic diagnosis Ductal adenocarcinoma Cystic neoplasm Neuroendocrine tumor Chronic pancreatitis Other periampullary tumors Common bile duct stone
53 (70 %) 23 (30 %) 17 (22 %) 9 (12 %) 4 (5 %) 46 (61 %) 8 (11 %) 2 (3 %) 4 (5 %) 2 (3 %) 1 (1 %) 40 (53 %) 28 (37 %) 3 (4 %) 5 (7 %) 1 (1 %) 37 (49 %) 2 (3 %)
Also, these results have shown that PAI is a more powerful predicting factor than the other known predicting factors. Outcomes According to a Proposed PAI Cutoff Value The median PAI according to CT scan measures was 0.67 (0.01–1.18) for the study cohort. In an attempt to find a cutoff value for predicting the rate of POPF, the study population was divided into two groups according to the median number (0.67) of PAI for all cohorts. The patients with PAI lower than 0.67 (n=37) were compared with the patients with PAI ≥0.67 (n=39). We found that the POPF rate was significantly higher in the group with PAI
