Pancreaticoduodenectomy (PD) is the well-known

Transcription

1 Meta-analysis Hepatobiliary & Pancreatic Diseases International Preoperative diabetes as a protective factor for pancreatic fistula after pancreaticoduodenectomy: a meta-analysis Xiang Xia, Chen Huang, Gang Cen and Zheng-Jun Qiu Shanghai, China BACKGROUND: The role of diabetes mellitus (DM) in pancreatic fistula (PF) or clinical relevant PF (CR-PF) after pancreaticoduodenectomy (PD) or pylorus-preserving pancreaticoduodenectomy (PPPD) is unclear. We conducted a meta-analysis to investigate the relationship between DM and PF or CR-PF. DATA SOURCES: Embase, MEDLINE and Cochrane databases were searched systematically for relevant articles from January 2005 to June The selected studies that examined clinical risk factors of PF or CR-PF were included. We created pooled estimates for our outcomes using the random-effects model. RESULTS: Sixteen observational clinical studies were included. Pooling of PF rates from ten studies revealed that DM was associated with a decreased risk of PF (P=0.01). CR-PF rates from 8 studies showed no significant difference between DM and control group (P=0.14). CONCLUSIONS: DM is not a risk factor for PF in patients undergoing PD or PPPD. On the contrary, patients without DM are at a higher risk of PF because the pancreases in these patients have more fatty tissue and the pancreas is soft. (Hepatobiliary Pancreat Dis Int 2015;14: ) KEY WORDS: pancreatic fistula; diabetes; postoperative complication Author Affiliations: Department of General Surgery, Shanghai Jiaotong University Affiliated First People's Hospital, Shanghai , China (Xia X, Huang C, Cen G and Qiu ZJ) Corresponding Author: Zheng-Jun Qiu, MD, PhD, Department of General Surgery, Shanghai Jiaotong University Affiliated First People's Hospital, 100 Haining Road, Shanghai , China (Tel: ; Fax: ; qiuzjdoctor@sina.com) 2015, Hepatobiliary Pancreat Dis Int. All rights reserved. doi: /S (15) Published online January 2, Introduction Pancreaticoduodenectomy (PD) is the well-known standard resection for various benign and malignant diseases of the periampullary region. [1, 2] The advances in operative techniques and postoperative management have greatly improved the safety of PD worldwide, the patients' perioperative mortality rates are significantly decreased and the quality of patients' life after [3, 4] PD was significantly improved. However, this complicated surgical procedure historically regarded as high operative mortality and morbidity, [5, 6] with 30%-40% of patients experiencing one or more postoperative complications. [7] Pancreatic fistula (PF) is the most common complication after PD, with a frequency ranging from 2% to 20%. PF is associated with a longer hospital stay and more costs. [6, 8] In addition, PF could be related to some potentially life-threatening complications such as intraabdominal hemorrhage and abscess, possibly leading to death. There are several risk factors for PF. The quality of the pancreatic parenchyma, such as soft pancreatic texture, non-dilated principal pancreatic duct or a thick pancreatic stump has been considered as the main risk [9, 10] Obesity factor. or high BMI is another risk factor causing a "fatty pancreas". [11, 12] A number of techniques have also been proposed to reduce PF following PD, such as the use of the somatostain analogue octreotide, [13] reinforcing anastomosis with fibrin glue, [14] and placement [15, 16] of a pancreatic duct stent. However, the role of diabetes mellitus (DM) in this setting is unclear. The correlation between DM and pancreatic ductal adenocarcinoma is well established. [17] Meanwhile, perioperative DM, especially if poorly controlled, has been shown to significantly impact postoperative complications after different operations. [18] In pancreatic resection, some studies showed that DM is a significant risk factor for PF after PD. [19-21] On the contrary, others revealed that patients with DM did not 132 Hepatobiliary Pancreat Dis Int,Vol 14,No 2 April 15,2015

2 Diabetes for pancreatic fistula show a greater rate of PF as compared with non-dm patients. [22-25] Because of the controversial results of these studies, we performed a meta-analysis focusing on the relationship between DM and PF in patients undergoing conventional PD or pylorus-preserving pancreaticoduodenectomy (PPPD). Methods This meta-analysis was conducted according to the "meta-analysis of observational studies in epidemiology: a proposal for reporting". [26] Search strategy Embase, MEDLINE and Cochrane databases were searched to identify relevant articles published from January 2005 to June The search terms included "pancreaticoduodenectomy" or "pylorus-preserving pancreaticoduodenectomy" or "whipple" or "pancreatic surgery" or "pancreatic resection"; "pancreatic fistula" or "pancreatic leak" or "anastomotic leak"; "diabetes" or "diabetes mellitus". Then the reference lists of relevant articles were manually searched to find other potentially eligible studies. No language restrictions were applied and our search was limited to humans only. All searches were performed by two independent reviewers (XX and HC). Study selection The titles and abstracts of the search results were screened for eligible studies. All cohort studies (prospective and retrospective) and all case-control studies that investigated DM directly influencing the occurrence of PF after PD or PPPD were included. The following were excluded: abstracts only, comments, letters, expert opinions, reviews without original data, case reports, studies with distal pancreatectomy or pancreatic enucleation and studies without a control group. Both investigators (XX and HC) assessed all articles by these above criteria and scored on a standardized form. Disagreement was resolved by consensus; when this failed, a third investigator adjudicated. We contacted the authors to get missing data when important information in the published articles was not adequate. Definition of PF and DM The primary outcome evaluation was the incidence of PF, as described by the International Study Group on Pancreatic Fistula (ISGPF). [6] Output via an operative placed drain (or a subsequently placed, percutaneous drain) of any measurable volume of drain fluid on or after postoperative day 3, with an amylase content greater than 3 times of the upper normal serum value. Then three different grades were defined according to the patients' clinical symptom: 1. Grade A has no clinical impact and requires little intervention in management; 2. Grade B needs some intervention such as partial or total parenteral or enteral nutrition, repositioning the drainage, antibiotics and somatostain; 3. Grade C requires aggressive clinical intervention and sometimes reoperation. As the definition of the ISGPF has been commonly used for PF evaluation, any article that failed to meet this PF criterion would be eliminated. The American Diabetes Association (ADA) has issued comprehensive diagnostic criteria for DM. [27] However, the definitions of DM in the published literature do not always follow the ADA criteria. In addition to ADA, various organizations (e.g., World Health Organization, International Diabetes Association) have their own diagnostic guidelines. Second, due to the retrospective data of most our articles, the diagnosis of DM is mainly derived from medical history and self-reporting. Therefore, caution is required to compare results between studies that have different criteria for DM. Data extraction The following data were extracted by two independent reviewers (CG and QZJ): first author, year of publication, study design, sample size, operative techniques, DM rate, presence or absence of PF and grade if possible, and primary outcomes. Inconsistencies were resolved through discussion until consensus was made, or the third reviewer (HC) would participate in this discussion. Because patients with grade B and C PF were generally considered to show clinical symptoms differed from grade A, we defined grade B and C PF as clinical relevant PF (CR-PF) and thus categorized selected studies into two groups according to their study outcomes. In the event of multiple publications from the same cohort or overlapping series of patients, data included only once from the most recent article to avoid double counting of patients. Qualitative assessment of observational studies was based on the Newcastle-Ottawa Scale. [28] "Good" was defined as a total score of 7-9; "Fair", 4-6; and "Poor", less than 4. Data analysis Meta-analyses were performed for studies which provided comparative data on the outcomes of patients who underwent PD or PPPD with or without DM. The odds ratio (OR) was chosen as an effect measure for those dichotomous parameters, which were reported along with the corresponding 95% confidence intervals (CIs) and the P<0.05 level was utilized to specify statistical significance. Heterogeneity was evaluated by inspection of the Hepatobiliary Pancreat Dis Int,Vol 14,No 2 April 15,

3 Hepatobiliary & Pancreatic Diseases International forest and I 2 statistic; an I 2 value of more than 50% was considered as high statistical heterogeneity. [29] Initially, a fixed-effects model was used to synthesize all data. However, random-effects model of DerSimonian and Laird [30] was applied if there was evidence of heterogeneity among included studies. Clinical heterogeneity could be explained by different outcome parameters, variability of interventions and perioperative management. Funnel plots were constructed to evaluate potential publication bias, [31] based on the complication PF. All statistical analyses were performed using Review Manager 5 software (The Cochrane Collaboration, Oxford, UK). for PF containing grade A, B and C (Table 1); [11, 12, 32-37] six articles combined grade B and C as CR-PF and compared CR-PF patients to non-pf and grade A patients for Results Study selection Of 450 articles we searched, 412 were excluded after scanning titles and abstracts, leaving 38 for full-text evaluation. Of these articles, 16 were finally included in this metaanalysis (Fig. 1). Among the 16 articles, two groups were identified. Eight articles assessed the predictive risk factors Fig. 1. Study flow chart. Studies Year Country Study type Table 1. Characteristics of 10 studies evaluating diabetes-related PF in pancreatectomy No. of Rate of Rate of patients DM (%) PF (%) Histology Operation Reconstruction Gaujoux et al [11] 2010 France Retrospective Benign (29.0%)/Malignant (71.0%) PD PG Good Mathur et al [12] 2007 USA Retrospective Benign (41.2%)/Malignant (58.8%) PD, PPPD PJ Good Wellner et al [32] 2010 Germany Retrospective * Benign (24.2%)/Malignant (75.8%) PD, PPPD PJ/PG # Fair Ausania et al [33] 2012 UK Retrospective * Malignant (100.0%) PD PJ (98.4%)/ PG (1.6%) Fair Uemura et al [34] 2012 Japan Retrospective Benign/Malignant # PD, PPPD PG Fair Roberts et al [35] 2013 UK Retrospective Malignant (100.0%) PD, PPPD PJ Good Akamatsu et al [36] 2010 Japan Retrospective Benign (13.9%)/Malignant (86.1%) PD, PPPD PJ Good Lermite et al [37] 2007 France Retrospective * Benign (34.4%)/Malignant (65.6%) PD PG Good Kawai et al [42] 2009 Japan Retrospective Benign (22.1%)/Malignant (77.9%) PD, PPPD PJ Good Tranchart et al [43] 2012 France Retrospective Benign (33.9%)/Malignant (66.1%) PD PG Good *: prospective maintained database; #: percentage not given. PD: pancreaticoduodenectomy; PPPD: pylorus-preserving pancreaticoduodenectomy; PG: pancreaticogastrostomy; PJ: pancreaticojejunostomy; NOS: Newcastle-Ottawa Scale. NOS score Studies Year Country Study type Table 2. Characteristics of 8 studies evaluating diabetes-related CR-PF in pancreatectomy Rate of No. of DM patients (%) Rate of CR-PF (%) Histology Operation Reconstruction NOS score Sachs et al [16] 2013 USA Retrospective * Benign/Malignant # PD PJ Good Hiyoshi et al [22] 2013 Japan Retrospective Benign (66.5%)/Malignant (33.5%) PD, PPPD PJ Good Dong et al [38] 2011 China Retrospective Benign (16.0%)/Malignant (84.0%) PD PJ Fair Hashimoto et al [39] 2011 USA Retrospective * Benign (32.5%)/Malignant (67.5%) PD, PPPD PJ Good Frozanpor et al [40] 2012 Sweden Retrospective * Benign (37.4%)/Malignant (62.6%) PD PJ Good Park et al [41] 2012 Korea Retrospective Benign (16.0%)/Malignant (84.0%) PD, PPPD PJ/PG # Fair Kawai et al [42] 2009 Japan Retrospective Benign (22.1%)/Malignant (77.9%) PD, PPPD PJ Good Tranchart et al [43] 2012 France Retrospective Benign (33.9%)/Malignant (66.1%) PD PG Good *: prospective maintained database; #: percentage not given. PD: pancreaticoduodenectomy; PPPD: pylorus-preserving pancreaticoduodenectomy; PG: pancreaticogastrostomy; PJ: pancreaticojejunostomy; NOS: Newcastle-Ottawa Scale. 134 Hepatobiliary Pancreat Dis Int,Vol 14,No 2 April 15,2015

4 Diabetes for pancreatic fistula risk factors analysis (Table 2). [16, 22, 38-41] The remaining 2 articles listed patients' data in each PF grade and thus [42, 43] could be categorized into two groups. The indications for PD included benign and malignant pancreatic and periampullary diseases, such as pancreatic adenocarcinoma, malignant ampulloma, chronic pancreatitis, benign ampulloma, etc. The standard operation was PD with or without pylorus-preservation, reconstructed by pancreaticogastrostomy (PG) or pancreaticojejunostomy (PJ). All identified articles followed an observational design, described "fair" or "good" in the Newcastle-Ottawa Scale. Outcomes DM and PF Ten studies comprising 1251 patients (sample size range: ) reported PF risk estimates of cohorts with DM (Table 1). The rate of DM ranged from 7.0% to 63.1%, whereas the rate of PF varied from 10.7% to 50.0%. On univariate analysis, Mathur and colleagues [12] found that patients with DM had a significantly decreased hazard of postoperative PF following PD or PPPD. The other 9 articles, including 3 prospectively [32, 33, 37] maintained databases, did not find a significant association between DM and PF in patients who underwent PD or PPPD. A meta-analysis was subsequently conducted to investigate DM as a protective factor for PF. Pooled analyses of PF by preoperative DM revealed a significant association between DM and non-pf (Fig. 2). Overall, PF occurred in 58 of 303 (19.1%) patients with diabetes and 276 of 948 (29.1%) patients without diabetes with an odds ratio of 0.64 (95% confidence interval 0.45 to 0.90, P=0.01). DM and CR-PF Eight studies evaluated whether preoperative DM was associated with the higher incidence of CR-PF (Table 2). A total of 1781 patients (sample size range: ) were analyzed, including 395 patients with diabetes and 1386 without diabetes. The rate of DM ranged Fig. 2. Meta-analysis for aggregated data in observational clinical studies comparing effect of diabetes or non-diabetes on pancreatic fistula after pancreaticoduodenectomy or pylorus-preserving pancreaticoduodenectomy. Pooled odds ratios with 95% confidence intervals were calculated using the fixed-effects model. Fig. 3. Meta-analysis for aggregated data in observational clinical studies comparing effect of diabetes or no diabetes on clinical relevant pancreatic fistula after pancreaticoduodenectomy or pylorus-preserving pancreaticoduodenectomy. Pooled odds ratios with 95% confidence intervals were calculated using the fixed-effects model. Hepatobiliary Pancreat Dis Int,Vol 14,No 2 April 15,

5 Hepatobiliary & Pancreatic Diseases International from 7.8% to 32.0%, while the rate of PF varied from 13.3% to 28.0%. Univariate (P=0.015) and multivariate analyses (P=0.003) showed that patients without DM were more likely to develop CR-PF following PD. [16] The other 7 articles, including 2 prospective maintained databases, reported that there was no significant dif- [39, 40] ference between non-dm patients and DM patients with regard to CR-PF. We compared the association between DM and CR-PF by meta-analyses and found that there was no significant difference in CR-PF between the DM and control group (Fig. 3). Pooled analysis displayed the incidence of CR-PF was 11.6% in the 395 patients with DM, compared with 15.0% in the 1386 patients without DM with an odds ratio of 0.77 (95% confidence interval 0.54 to 1.09, P=0.14). Discussion DM is considered to be one of the major and growing public health burdens in the world. The relationship between DM and pancreatic ductal adenocarcinoma (PDAC) is well-established. [17] Additionally, approximately 20%-80% of PDAC patients have either insufficient glucose tolerance or definite DM at the time of diagnosis, which appears most frequently within only 1 or 2 years before diagnosis. [44] DM has also been identified as a risk factor for a number of postoperative adverse events after various types of operations such as cardiac, vascular and general surgeries. [45-47] Because of the close association between DM and PDAC, the impact of preoperative DM on postoperative course of patients undergoing PD or PPPD may be of great importance. Whether preoperative DM has a significant impact on PF is not clear. In 2001, Srivastava et al [48] analyzed 120 patients who underwent PD and had pancreaticoenteric anastomotic leakage. Multivariate logistic regression revealed that patients with preoperative DM had a significantly higher incidence of PF than those without DM (28% vs 8.4%, respectively; P<0.05). In 2004, Lin et al [49] analyzed 1891 patients with PF. The diagnosis was either a radiologically proved anastomotic leak or the continued drainage of amylase-rich fluid on or after postoperative day 10. Univariate analysis showed that the incidence of PF in patients with preoperative DM was significantly lower than that in those without DM (7.7% vs 12.0%, P<0.05). Unlike previous studies, by pooling diabetes rates from sixteen studies, we were able to establish a large cohort to perform a statistical review and meta-analysis. Interestingly, we found that DM did not significantly increase the risk of PF or CR-PF. On the contrary, a significant trend was observed toward decreased PF in patients with DM (Fig. 2). Although there was no significant difference, patients without DM were more likely to develop CR-PF (Fig. 3). This interesting result could be explained by two reasons: 1. Pancreases in patients with DM had significantly less fat or more fibrosis (hard texture) compared with those in patients without DM. [12, 19, 39] Both soft or fatty pancreases had been identified as risk factors for PF. [12, 50, 51] 2. Several articles analyzed the risk factors for PF in patients undergoing PD or distal pancreatectomy. However, the rate of PF in these two different operative techniques are different, PF occurs more frequently following distal pancreatectomy than that following PD. [3, 52, 53] As a result, this heterogeneity in PF rates may impose an unintentional impact on identifying DM as a risk factor for PF or CR-PF. More importantly, identifying the association between DM and PF or CR-PF may help perioperative therapeutic considerations and improve surgical outcomes. Our study has several limitations. First, we did not differentiate DM type (type I, type II and other types), the diagnosis of DM was not standardized. When we categorized the patients, those with mild abnormalities in glucose tolerance were not diagnosed with DM. [42] Second, this meta-analysis was limited by the availability of observational studies, which may be more prone to confounding factors and bias. [54] Nevertheless, rigorous inclusion and exclusion criteria as well as the uniform ISGPF definition were applied and study quality was appraised. In our systematic review process, many articles were excluded because of poor definition of PF or inappropriate study design. Third, substantial heterogeneity between studies may preclude a pooled analysis. The preformulated hypothesis and comprehensive search can help to minimize the publication bias. Significant heterogeneity was not observed when we used funnel plots and the Chi-square test and inconsistency (I 2 ) statistics. Fourth, as literature search stops on June 2013, works published in the last 9 months are not considered in the meta-analysis. However, after a careful search by our surgical team, few articles about such issue were found in these months. Thus, the articles included in our manuscript were not "out of date". In conclusion, investigations of postoperative PF in patients with preoperative DM who underwent PD or PPPD are rare. The present meta-analysis suggested that preoperative DM is not a risk factor of PF after PD or PPPD. On the contrary, patients without DM might carry a higher risk of PF which may result from fatty or soft pancreas. However, the overall results of the present meta-analysis do not permit any definite conclusions because of the heterogeneity in the definition of DM among different studies. More evidence from prospective randomized clinical trials is needed. 136 Hepatobiliary Pancreat Dis Int,Vol 14,No 2 April 15,2015

6 Diabetes for pancreatic fistula Contributors: XX proposed the study. CG performed research and wrote the first draft. HC collected and analyzed the data. All authors contributed to the design and interpretation of the study and to further drafts. QZJ is the guarantor. Funding: This study was supported by grants from the National Natural Science Foundation of China ( , and ), Shanghai Municipal Human Resources and Social Security Bureau ( and 13PJD024) and Shanghai Municipal Health Bureau (13Y068). Ethical approval: Not needed. Competing interest: No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. References 1 Burcharth F, Olsen SD, Trillingsgaard J, Federspiel B, Moesgaard F, Struckmann JR. Pancreaticoduodenectomy for periampullary cancer in patients more than 70 years of age. Hepatogastroenterology 2001;48: Büchler MW, Wagner M, Schmied BM, Uhl W, Friess H, Z'graggen K. Changes in morbidity after pancreatic resection: toward the end of completion pancreatectomy. Arch Surg 2003;138: Yeo CJ, Cameron JL, Sohn TA, Lillemoe KD, Pitt HA, Talamini MA, et al. Six hundred fifty consecutive pancreaticoduodenectomies in the 1990s: pathology, complications, and outcomes. Ann Surg 1997;226: Bassi C, Dervenis C, Butturini G, Fingerhut A, Yeo C, Izbicki J, et al. Postoperative pancreatic fistula: an international study group (ISGPF) definition. Surgery 2005;138: Gouma DJ, van Geenen RC, van Gulik TM, de Haan RJ, de Wit LT, Busch OR, et al. Rates of complications and death after pancreaticoduodenectomy: risk factors and the impact of hospital volume. Ann Surg 2000;232: Okamoto A, Tsuruta K. Fistulation method: simple and safe pancreaticojejunostomy after pancreatoduodenectomy. Surgery 2000;127: Bassi C, Falconi M, Salvia R, Mascetta G, Molinari E, Pederzoli P. Management of complications after pancreaticoduodenectomy in a high volume centre: results on 150 consecutive patients. Dig Surg 2001;18: Kim JH, Li G, Baek NH, Hwang JC, Hong J, Yoo BM, et al. Surgical outcomes of distal pancreatectomy. Hepatogastroenterology 2013;60: de Castro SM, Busch OR, van Gulik TM, Obertop H, Gouma DJ. Incidence and management of pancreatic leakage after pancreatoduodenectomy. Br J Surg 2005;92: Sugimoto M, Gotohda N, Kato Y, Takahashi S, Kinoshita T, Shibasaki H, et al. Risk factor analysis and prevention of postoperative pancreatic fistula after distal pancreatectomy with stapler use. J Hepatobiliary Pancreat Sci 2013;20: Gaujoux S, Cortes A, Couvelard A, Noullet S, Clavel L, Rebours V, et al. Fatty pancreas and increased body mass index are risk factors of pancreatic fistula after pancreaticoduodenectomy. Surgery 2010;148: Mathur A, Pitt HA, Marine M, Saxena R, Schmidt CM, Howard TJ, et al. Fatty pancreas: a factor in postoperative pancreatic fistula. Ann Surg 2007;246: Graham JA, Johnson LB, Haddad N, Al-Kawas F, Carroll J, Jha R, et al. A prospective study of prophylactic long-acting octreotide in high-risk patients undergoing pancreaticoduodenectomy. Am J Surg 2011;201: Lillemoe KD, Cameron JL, Kim MP, Campbell KA, Sauter PK, Coleman JA, et al. Does fibrin glue sealant decrease the rate of pancreatic fistula after pancreaticoduodenectomy? Results of a prospective randomized trial. J Gastrointest Surg 2004;8: Pessaux P, Sauvanet A, Mariette C, Paye F, Muscari F, Cunha AS, et al. External pancreatic duct stent decreases pancreatic fistula rate after pancreaticoduodenectomy: prospective multicenter randomized trial. Ann Surg 2011;253: Sachs TE, Pratt WB, Kent TS, Callery MP, Vollmer CM Jr. The pancreaticojejunal anastomotic stent: friend or foe? Surgery 2013;153: Ben Q, Xu M, Ning X, Liu J, Hong S, Huang W, et al. Diabetes mellitus and risk of pancreatic cancer: A meta-analysis of cohort studies. Eur J Cancer 2011;47: Raghavan SR, Ballehaninna UK, Chamberlain RS. The impact of perioperative blood glucose levels on pancreatic cancer prognosis and surgical outcomes: an evidence-based review. Pancreas 2013;42: Chu CK, Mazo AE, Sarmiento JM, Staley CA, Adsay NV, Umpierrez GE, et al. Impact of diabetes mellitus on perioperative outcomes after resection for pancreatic adenocarcinoma. J Am Coll Surg 2010;210: Addeo P, Delpero JR, Paye F, Oussoultzoglou E, Fuchshuber PR, Sauvanet A, et al. Pancreatic fistula after a pancreaticoduodenectomy for ductal adenocarcinoma and its association with morbidity: a multicentre study of the French Surgical Association. HPB (Oxford) 2014;16: Subhedar PD, Patel SH, Kneuertz PJ, Maithel SK, Staley CA, Sarmiento JM, et al. Risk factors for pancreatic fistula after stapled gland transection. Am Surg 2011;77: Hiyoshi M, Chijiiwa K, Fujii Y, Imamura N, Nagano M, Ohuchida J. Usefulness of drain amylase, serum C-reactive protein levels and body temperature to predict postoperative pancreatic fistula after pancreaticoduodenectomy. World J Surg 2013;37: Malleo G, Mazzarella F, Malpaga A, Marchegiani G, Salvia R, Bassi C, et al. Diabetes mellitus does not impact on clinically relevant pancreatic fistula after partial pancreatic resection for ductal adenocarcinoma. Surgery 2013;153: Nakata B, Ishikawa T, Amano R, Kimura K, Hirakawa K. Impact of preoperative diabetes mellitus on clinical outcome after pancreatectomy. Int J Surg 2013;11: Yamamoto Y, Sakamoto Y, Nara S, Esaki M, Shimada K, Kosuge T. A preoperative predictive scoring system for postoperative pancreatic fistula after pancreaticoduodenectomy. World J Surg 2011;35: Stroup DF, Berlin JA, Morton SC, Olkin I, Williamson GD, Rennie D, et al. Meta-analysis of observational studies in epidemiology: a proposal for reporting. Meta-analysis Of Observational Studies in Epidemiology (MOOSE) group. JAMA 2000;283: American Diabetes Association. Diagnosis and classification of diabetes mellitus. Diabetes Care 2012;35:S Saunders LD, Soomro GM, Buckingham J, Jamtvedt G, Raina P. Assessing the methodological quality of nonrandomized intervention studies. West J Nurs Res 2003;25: Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring Hepatobiliary Pancreat Dis Int,Vol 14,No 2 April 15,

7 Hepatobiliary & Pancreatic Diseases International inconsistency in meta-analyses. BMJ 2003;327: DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials 1986;7: Sterne JA, Egger M, Smith GD. Systematic reviews in health care: Investigating and dealing with publication and other biases in meta-analysis. BMJ 2001;323: Wellner UF, Kayser G, Lapshyn H, Sick O, Makowiec F, Höppner J, et al. A simple scoring system based on clinical factors related to pancreatic texture predicts postoperative pancreatic fistula preoperatively. HPB (Oxford) 2010;12: Ausania F, Snowden CP, Prentis JM, Holmes LR, Jaques BC, White SA, et al. Effects of low cardiopulmonary reserve on pancreatic leak following pancreaticoduodenectomy. Br J Surg 2012;99: Uemura K, Murakami Y, Sudo T, Hashimoto Y, Nakashima A, Yamaoka E, et al. Elevation of urine trypsinogen 2 is an independent risk factor for pancreatic fistula after pancreaticoduodenectomy. Pancreas 2012;41: Roberts KJ, Storey R, Hodson J, Smith AM, Morris-Stiff G. Pre-operative prediction of pancreatic fistula: is it possible? Pancreatology 2013;13: Akamatsu N, Sugawara Y, Komagome M, Shin N, Cho N, Ishida T, et al. Risk factors for postoperative pancreatic fistula after pancreaticoduodenectomy: the significance of the ratio of the main pancreatic duct to the pancreas body as a predictor of leakage. J Hepatobiliary Pancreat Sci 2010;17: Lermite E, Pessaux P, Brehant O, Teyssedou C, Pelletier I, Etienne S, et al. Risk factors of pancreatic fistula and delayed gastric emptying after pancreaticoduodenectomy with pancreaticogastrostomy. J Am Coll Surg 2007;204: Dong X, Zhang B, Kang MX, Chen Y, Guo QQ, Wu YL. Analysis of pancreatic fistula according to the International Study Group on Pancreatic Fistula classification scheme for 294 patients who underwent pancreaticoduodenectomy in a single center. Pancreas 2011;40: Hashimoto Y, Sclabas GM, Takahashi N, Kirihara Y, Smyrk TC, Huebner M, et al. Dual-phase computed tomography for assessment of pancreatic fibrosis and anastomotic failure risk following pancreatoduodenectomy. J Gastrointest Surg 2011; 15: Frozanpor F, Loizou L, Ansorge C, Segersvärd R, Lundell L, Albiin N. Preoperative pancreas CT/MRI characteristics predict fistula rate after pancreaticoduodenectomy. World J Surg 2012;36: Park CM, Park JS, Cho ES, Kim JK, Yu JS, Yoon DS. The effect of visceral fat mass on pancreatic fistula after pancreaticoduodenectomy. J Invest Surg 2012;25: Kawai M, Tani M, Hirono S, Ina S, Miyazawa M, Yamaue H. How do we predict the clinically relevant pancreatic fistula after pancreaticoduodenectomy?--an analysis in 244 consecutive patients. World J Surg 2009;33: Tranchart H, Gaujoux S, Rebours V, Vullierme MP, Dokmak S, Levy P, et al. Preoperative CT scan helps to predict the occurrence of severe pancreatic fistula after pancreaticoduodenectomy. Ann Surg 2012;256: Chari ST, Leibson CL, Rabe KG, Timmons LJ, Ransom J, de Andrade M, et al. Pancreatic cancer-associated diabetes mellitus: prevalence and temporal association with diagnosis of cancer. Gastroenterology 2008;134: Wright CD, Kucharczuk JC, O'Brien SM, Grab JD, Allen MS; Society of Thoracic Surgeons General Thoracic Surgery Database. Predictors of major morbidity and mortality after esophagectomy for esophageal cancer: a Society of Thoracic Surgeons General Thoracic Surgery Database risk adjustment model. J Thorac Cardiovasc Surg 2009;137: Halkos ME, Puskas JD, Lattouf OM, Kilgo P, Kerendi F, Song HK, et al. Elevated preoperative hemoglobin A1c level is predictive of adverse events after coronary artery bypass surgery. J Thorac Cardiovasc Surg 2008;136: Little SA, Jarnagin WR, DeMatteo RP, Blumgart LH, Fong Y. Diabetes is associated with increased perioperative mortality but equivalent long-term outcome after hepatic resection for colorectal cancer. J Gastrointest Surg 2002;6: Srivastava S, Sikora SS, Pandey CM, Kumar A, Saxena R, Kapoor VK. Determinants of pancreaticoenteric anastomotic leak following pancreaticoduodenectomy. ANZ J Surg 2001; 71: Lin JW, Cameron JL, Yeo CJ, Riall TS, Lillemoe KD. Risk factors and outcomes in postpancreaticoduodenectomy pancreaticocutaneous fistula. J Gastrointest Surg 2004;8: Lee SE, Jang JY, Lim CS, Kang MJ, Kim SH, Kim MA, et al. Measurement of pancreatic fat by magnetic resonance imaging: predicting the occurrence of pancreatic fistula after pancreatoduodenectomy. Ann Surg 2010;251: Rosso E, Casnedi S, Pessaux P, Oussoultzoglou E, Panaro F, Mahfud M, et al. The role of "fatty pancreas" and of BMI in the occurrence of pancreatic fistula after pancreaticoduodenectomy. J Gastrointest Surg 2009;13: Cameron JL, Pitt HA, Yeo CJ, Lillemoe KD, Kaufman HS, Coleman J. One hundred and forty-five consecutive pancreaticoduodenectomies without mortality. Ann Surg 1993;217: Assifi MM, Lindenmeyer J, Leiby BE, Grunwald Z, Rosato EL, Kennedy EP, et al. Surgical Apgar score predicts perioperative morbidity in patients undergoing pancreaticoduodenectomy at a high-volume center. J Gastrointest Surg 2012;16: Egger M, Schneider M, Davey Smith G. Spurious precision? Meta-analysis of observational studies. BMJ 1998;316: Received March 7, 2014 Accepted after revision October 20, Hepatobiliary Pancreat Dis Int,Vol 14,No 2 April 15,2015