Body Mass Index Is a Risk Factor of Pancreatic Fistula after Pancreaticoduodenectomy MITSUGI SHIMODA, M.D., PH.D., MASATO KATOH, M.D., PH.D., ISO YUKIHIRO, M.D., PH.D., JUNJI KITA, M.D., PH.D., TOKIHIKO SAWADA, M.D., PH.D., KEIICHI KUBOTA, M.D., PH.D.
From the Second Department of Surgery, Dokkyo Medical University, Tochigi, Japan Pancreatic fistula (PF) after pancreaticoduodenectomy (PD) is still a severe complication and a challenging problem. The common risk factors are the soft pancreas and small pancreatic duct of the remnant pancreas. Those two risk factors were recognized during surgery. On the other hand, a preoperatively determined risk factor of PF is unclarified. We conducted a retrospective analysis of 203 patients consecutively treated by PD from April 2000 to October 2010. PF was defined according to the criteria of the International Study Group of Pancreatic Fistula. Clinical and preand intraoperative data were compared between PF and non-PF patients. The recommended cutoff value of body mass index (BMI) as 20 kg/m2 was defined by receiver operating characteristic curve analysis. PF occurred in 53 (26.1%) of 203 patients. In univariate analysis, BMI and soft remnant pancreas were found to be risk factors of PF (P 5 0.027, P 5 0.005). In multivariate analysis, BMI and soft pancreas were also risk factors of PF (P 5 0.040, P 5 0.005). Patients with PF had a significantly longer hospital stay than non-PF patients (P 5 0.005). High BMI and soft pancreas were significant risk factors for PF.
(PD) is performed for various malignant and benign diseases in highP volume centers and the current mortality rate is less ANCREATICODUODENECTOMY
than 4 per cent.1, 2 A decrease in the operative mortality rate is considered to be attributed to good patient selection, improved surgical technique, effective postoperative patient care, and early treatment for complications. On the other hand, the postoperative morbidity rate after PD was still in the high range of 30 to 40 per cent or more.1 A pancreatic fistula (PF) is the most severe complication after PD with a morbidity rate greater than 25 per cent in high-risk patients.2 Conservative treatment sometimes can be successful; however, PF induces prolonged hospital stay and increased medical cost and deteriorates the patient’s mental condition. PF sometimes leads to more advanced complications such as abdominal bleeding or peptic perforation, which is associated with mortality rates ranging from 3 to 15 per cent.2–4 Several preventive measures or treatments have been reported. The results of a randomized trial evaluating a reconstruction technique or use of octreotide treatment has been reported not to be effective for preventing PF.5 The risk factors of PF also have been evaluated, but soft
Address correspondence and reprint requests to Mitsugi Shimoda, M.D., Ph.D., Second Department of Surgery, Dokkyo Medical University, 880 Kitakobayashi, Mibu, Tochigi 321-0293, Japan. E-mail: [email protected]
pancreatic tissue is the only factor accepted worldwide. This risk factor is not confirmed preoperatively. Obesity is the one of the most important risk factors for in the abdominal surgery field and is closely associated with severe complications such as deep vein thrombosis or surgical site infection.6 In recent years, body mass index (BMI) has been widely accepted as an index of obesity, and it is also assumed to be an index of postoperative complications. Elevated BMI has also been one of the risk factors attributed to PF after distal pancreatectomy,6 but the influence of BMI on the occurrence of PF after PD has not been clarified in previous retrospective reports7; the influence of high BMI on the occurrence after PD has not been clarified in previous retrospective reports. In this retrospective study of 203 consecutive PDs, we evaluated the influence of preoperative factors, operative technique, and commonly accepted risk factors on the occurrence of PF.
From April 2000 to October 2010, 203 consecutive patients underwent PD at the Second Department of Surgery, Dokkyo Medical University. All clinical, pathological, and biochemical data were collected and were analyzed retrospectively. In our analysis, we compared the PF group (PF G) with the non-PF group (N-PF G).
BMI IS RISK FACTOR OF PANCREATIC FISTULA AFTER PD
Shimoda et al.
Preoperative Evaluation for Resection
A comparison of category data was conducted using the chi-square test. For continuous data, the MannWhitney U test was applied when the normality hypothesis within each group was not rejected. Odds ratios with 95 per cent confidence interval were calculated using univariate and comparative analyses. Univariate, multivariate, and comparative analyses were performed using logistic regression analysis. For all tests, statistical significance was defined by P < 0.05. The data were analyzed with the SPSS statistical software (Statistical Package for the Social Sciences, Chicago, IL).
Preoperative evaluation included physical examination, routine laboratory tests, and contrast computed tomography (CT). To be considered for resection, patients had to fulfill objective radiographic criteria for resectability as follows: 1) the absence of metastatic lesion; and 2) no evidence of tumor involvement of the supramesenteric artery or celiac axis. We performed biliary drainage for patients with jaundice through the percutaneous transhepatic or nasobiliary route. PD and subtotal stomach-preserving pancreaticoduodenectomy (SSPPD)8 were performed with en bloc resection of the lymph nodes. From April 2000 to August 2008, reconstruction of pancreaticojejunostomy (PJ) after PD was conducted using the Kakita method with an external pancreatic duct stent without duct-to-mucosa anastomosis.9 All stitches were absorbable sutures (4-0 Tycron using Kakita methods and 5-0 Tycron using external stent placement suture). From August 2008 to October 2010, PJ after SSPPD was performed using the Kakita method with an external pancreatic duct stent and duct to mucosa anastomosis. All stitches were monofilament nonabsorbable sutures (4-0 proline using the Kakita methods and or 5-0 or 6-0 Prolene using duct–to-mucosa sutures).10 After the procedure, multiple Dupple drains were placed at PJ, hepaticojejunostomy, and Winslow foramen. Amylase and bilirubin assay and cultures in drainage fluid were performed routinely on 1, 3, and 7 postoperative days. When amylase or bilirubin levels were normalized and the amount of fluid discharge became less than 100 mL/day approximately, and negative cultures were obtained, these drains were removed. CT was taken in case of postoperative complications. Prognostic Factors
The prognostic factors for the two groups included gender, age, operating time, blood loss, hospitalization, malignant or benign disease, diabetes mellitus, soft pancreas, preoperative albumin, C-reactive protein, HbA1c, triacylglycerol (TG), total cholesterol (TC), BMI, tumor makers carcinoembryonic antigen and CA1-9), PJ anastomosis (external tube stent vs. duct-to-mucosa anastomosis), gastrojejunostomy (Billroth II vs. Roux-en-Y), and gastrojejunostomy route (antecolic or retrocolic). The recommended cutoff value of BMI was defined by receiver operating characteristic (ROC). To determine the optimal cutoff point, the Youden index method was used.11, 12 PF was defined and graded (A, B, and C) according to the criteria of the International Study Group of Pancreatic fistula (ISGPF).13
The pre-, post-, and operative characteristics of the 203 patients are listed in Table 1. The original disease was malignant in 86.2 per cent of all patients, among whom PF occurred in 23.2 per cent. The patients in the PF G were more frequently associated with high BMI (greater than 20 kg/m2) (P 4 0.019) and prolonged hospital stay (P 4 0.005). Soft pancreas was significantly associated with PF after PD (P 4 0.0043). Surgical technique such as PJ anastomosis, gastrojejunostomy, and gastrojejunostomy route were not risk factors for PF. There were no differences in intraoperative conditions such as duration of surgery and blood loss between the two groups (Table 1). ROC curve revealed the most prominent point disclosed two parameters such as sensitivity (0.827) and 1-specificity (0.673) (Fig. 1). Because these two parameters indicated a recommended cutoff value as 20.055, we defined a cutoff value of BMI as 20 kg/m2. The results of uni- and multivariate analysis are listed in Table 2. The result of univariate analysis indicated BMI (greater than 20 kg/m2) and soft pancreas to be significant risk factors for PF (P 4 0.027, P 4 0.005). Multivariate analysis showed that both BMI and soft pancreas were also risk factors of PF (P 4 0.040, P 4 0.005). Discussion
Our retrospective study demonstrated that BMI and soft pancreas were independent risk factors of PF. In Western series, overweight (BMI more than 25 kg/m2) and obesity (BMI more than 30 kg/m2) are based essentially on criteria derived from studies. However, this criterion in Asian populations has yet not been determined. BMI may differ between Asian and European patients.14 Undergoing surgery is considered high risk because of increased cardiopulmonary activity, deep vein thrombosis, and surgical site infection.6 Also, obesity has been
THE AMERICAN SURGEON
TABLE 1. Characteristics of Patients According to Pancreas Fistula Characteristics
Overall (n 4 203)
PF G (n 4 53)
N-PF G (n 4 150)
Gender (male/female) Age Operating time (minutes) Blood loss (mL) Hospitalization With OORx/without Indication (malignant/benign) Soft pancreas (yes or no) Albumin level (less than 3.5 g/dL/3.5 g/dL or greater) C-reactive protein level (less than 0.3 mg/dL/0.3 mg/dL or greater) Body mass index (20 kg/m2 or greater/less than 20 kg/m2) Diabetes mellitus or not PJ anastomosis (ETS/DD) GJ (B-II vs. R-Y) GJR (antecolic or retrocolic)
91/112 69.7 (36–85) 520 (350–971) 555 (200–3390) 54 (19–238) 38/165 175/28 85/118 107/96
30/23 69.7 (48–82) 500 (350–971) 598 (255–2330) 62 (23–238) 5/48 47/6 31/22 27/26
91/59 69.5 (36–859) 520 (358–960) 536 (200–3390) 53 (19–185) 33/117 128/22 54/96 80/70
0.605 0.722 0.823 0.866 0.005 0.053 0.545 0.0043 0.961
53/150 162/41 32/171 56/147
11/42 44/9 8/45 17/36
42/108 118/32 24/126 39/111
0.257 0.497 0.876 0.555
PF G 4 pancreatic fistula group; N-PF G 4 nonpancreatic fistula group; OORx 4 other organ resection; PJ 4 pancreaticojejunostomy; ETS 4 external tube stent; DD 4 duct–to-mucosa anastomosis; GJ 4 gastrojejunostomy; B-II 4 Billroth II; R-Y 4 Roux–en-Y; GJR 4 gastrojejunostomy route.
FIG. 1. Receiver operating characteristic curve shows the recommended point of cutoff value for body mass index. Arrow indicates the recommended point.
described as a predictive risk factor for anastomotic leakage and easy bleeding during rectal operations.15, 16 In our study, the cutoff value of BMI (less than 20 kg/m2 or greater than 20 kg/m2) in Japanese patients with PF after PD was slightly lower than in Western series. PF after PD remains a challenging problem, even in high-volume centers. In this study, we have experienced a 23.2 per cent rate of PF among the patients who have undergone PD. This figure is even to the rate reported by previous reports.3 PF is defined according to the ISGPF criteria,13 which considers amylase-rich drainage fluid from postoperative Day 3, when we considered PF from the highest fluid amylase level at postoperative Days 3,
5, and 7. Identically, using the ISGPF definition, previous studies reported the rate of PF to be more than 25 per cent.17 Several reports described possible risk factors of PF; however, soft consistency of the remnant pancreas was the only one accepted widely and applied clinically.18–20 Other risk factors including gender, age, heart and vascular complications,19 small pancreatic duct (less than 3 mm),18 malignant site,21 amount of bleeding,21 and low volume per surgeon21 have been considered; however, these factors seemed less convincing to explain the risk factor of PF. Several reports suggested that age is a risk factor for intra-abdominal morbidity
BMI IS RISK FACTOR OF PANCREATIC FISTULA AFTER PD
Shimoda et al.
TABLE 2. Predictive Factors of Pancreas Fistula: Univariate and Multivariate Analysis Univariate
With OORx Malignant disease Soft pancreas Albumin C-reactive protein HbA1c Triacylglycerol Total cholesterol Carcinoembryonic antigen CA19-9 Body mass index (20 kg /2 or greater) Diabetes mellitus
0.369 1.346 2.505 1.125 1.029 0.822 0.999 0.998 0.882 1.000 2.668 0.612
0.136–1.003 0.514–3.525 1.321–4.751 0.613–2.067 0.928–1.285 0.618–1.094 0.993–1.003 0.993–1.003 0.766–1.015 1.000–1.000 1.117–6.372 0.282–1.331
0.051 0.545 0.005 0.703 0.290 0.045 0.594 0.498 0.080 0.434 0.027 0.215
OR 4 odds ratio; CI 4 confidence interval; OORx 4 other organ resection.
and mortality.22, 23 High age of older than 80 years old had a 4 per cent mortality rate,23 but on the other hand, some reports described that they did not find any difference among patients who were older or younger than 65 years. Intraoperative bleeding, malignant or nonmalignant, gender, age, and preoperative vascular complications were not associated with an increased risk factor for PF.24 The prevalence of obesity is increasing in patients undergoing general surgery.6 In the area of pancreatic surgery, a BMI over 25 kg/m2 was considered to be a risk factor for pancreatic fistula formation after distal pancreatectomy7; high BMI over 25 kg/m2 was considered to be a risk factor for increasing the risk of PF.7 It is also reported that BMI more than 30 kg/m2 was a risk factor leading to the intensive care unit after operation.24 BMI more than 25 kg/m2, soft remnant and fatty pancreas were also reported to be a risk factor of PF.25 Also, they were exposed to histologic scoring of pancreatic fibrosis and fatty infiltration and were found to have a correlation between decreased pancreatic fibrosis and increased fatty infiltration with increasing BMI. Overall, these reports suggested that high BMI induced a fatty change of pancreas tissue, disturbs the healing pancreas and jejunum anastomosis, and increases the risk of PF.24–26 In this study, we did not evaluate histological examination of resected specimens of the pancreas, but we scheduled to evaluate histological scoring of pancreatic fibrosis and fatty infiltration by the next step for patients with PF with BMI less than 20 kg/m2. Also, we calculated for patients with BMI less than 25 kg/m2; unfortunately, BMI more than 25 kg/m2 was not a significant risk factor for patients with PF in our series. Previous reports described that the fatty pancreas was a risk factor of PF after PD. Fatty infiltration of the pancreas graded according to a score derived from fibrosis suggested more frequent development of PF after
PD. The report described that a fatty pancreas was more frequent in the case of a soft pancreas and in patients with high BMI or who were older than 58 years.26 A fatty pancreas is correlated with obesity and also related to patient age. However, it is difficult to diagnose a fatty pancreas by diagnostic imaging before surgery. In this study, other factors including preoperative nourishment parameters (HbA1c, TC, TG, and albumin), tumor marker level (carcinoembryonic antigen and CA19-9), and procedures (PJ anastomosis, gastrojejunostomy, and gastrojejunostomy route) have been less effective or careful to explain the risk factor of PF. BMI greater than 20 kg/m2 and a soft pancreas were significantly associated with an increased incidence of PF. BMI gives an objective criteria to stratify the risk for PF preoperatively and plan appropriate measures for prevention and management of PF. REFERENCES
1. Cameron JL, Riall TS, Coleman J, Belcher KA. One thousand consecutive pancreaticoduodenectomies. Ann Surg 2006;244:10–5. 2. DeOliveira ML, Winter JM, Schafer M, et al. Assessment of complications after pancreatic surgery: a novel grading system applied to 633 patients undergoing pancreaticoduodenectomy. Ann Surg 2006;244:931–7; discussion 937–9. 3. Munoz-Bongrand N, Sauvanet A, Denys A, et al. Conservative management of pancreatic fistula after pancreaticoduodenectomy with pancreaticogastrostomy. J Am Coll Surg 2004;199:198–203. 4. de Castro SM, Busch OR, van Gulik TM, et al. Incidence and management of pancreatic leakage after pancreatoduodenectomy. Br J Surg 2005;92:1117–23. 5. Connor S, Alexakis N, Garden OJ, et al. Meta-analysis of the value of somatostatin and its analogues in reducing complications associated with pancreatic surgery. Br J Surg 2005;92:1059–67. 6. Pasulka PS, Bistrian BR, Benotti PN, Blackburn GL. The risks of surgery in obese patients. Ann Intern Med 1986;104:540–6. 7. Sledzianowski JF, Duffas JP, Muscari F, et al. Risk factors for mortality and intra-abdominal morbidity after distal pancreatectomy. Surgery 2005;137:180–5.
THE AMERICAN SURGEON
8. Oida T, Mimatsu K, Kawasaki A, et al. Vertical stomach reconstruction with pancreaticogastrostomy after modified subtotal-stomach-preserving pancreaticoduodenectomy for preventing delayed gastric emptying. Hepatogastroenterology 2009;56:565–7. 9. Kakita A, Takahashi T, Yoshida M, Furuta K. A simpler and more reliable technique of pancreatojejunal anastomosis. Surg Today 1996;26:532–5. 10. Satoi S, Toyokawa H, Yanagimoto H, et al. A new guideline to reduce postoperative morbidity after pancreaticoduodenectomy. Pancreas 2008;37:128–33. 11. Perkins NJ, Schisterman EF. The inconsistency of ‘optimal’ cutpoints obtained using two criteria based on the receiver operating characteristic curve. Am J Epidemiol 2006;163:670–5. 12. Akobeng AK. Understanding diagnostic tests 1: sensitivity, specificity and predictive values. Acta Paediatr 2007;96:338–41. 13. Bassi C, Dervenis C, Butturini G, et al. Postoperative pancreatic fistula: an international study group (ISGPF) definition. Surgery 2005;138:8–13. 14. WHO Expert Consultation. Appropriate body-mass index for Asian populations and its implications for policy and intervention strategies. Lancet 2004;363:157–63. 15. Rullier E, Laurent C, Garrelon JL, et al. Risk factors for anastomotic leakage after resection of rectal cancer. Br J Surg 1998;85:355–8. 16. Benoist S, Panis Y, Pannegeon V, et al. Predictive factors for perioperative blood transfusions in rectal resection for cancer: a multivariate analysis of a group of 212 patients. Surgery 2001; 129:433–9. 17. Reid-Lombardo KM, Farnell MB, Crippa S, et al. Pancreatic anastomotic leakage after pancreaticoduodenectomy in 1,507
patients: a report from the Pancreatic Anastomotic Leak Study Group. J Gastrointest Surg 2007;11):1451–8; discussion 1459. 18. Muscari F, Suc B, Kirzin S, et al. Risk factors for mortality and intra-abdominal complications after pancreatoduodenectomy: multivariate analysis in 300 patients. Surgery 2006;139:591–8. 19. Lin JW, Cameron JL, Yeo CJ, et al. Risk factors and outcomes in postpancreaticoduodenectomy pancreaticocutaneous fistula. J Gastrointest Surg 2004;8:951–9. 20. Yeo CJ. Does prophylactic octreotide benefit patients undergoing elective pancreatic resection? J Gastrointest Surg 1999;3: 223–4. 21. Pratt WB, Callery MP, Vollmer CM Jr. Risk prediction for development of pancreatic fistula using the ISGPF classification scheme. World J Surg 2008;32:419–28. 22. Miedema BW, Sarr MG, van Heerden JA, et al. Complications following pancreaticoduodenectomy. Current management. Arch Surg 1992;127:945–9; discussion 949–50. 23. Sohn TA, Yeo CJ, Cameron JL, et al. Should pancreaticoduodenectomy be performed in octogenarians? J Gastrointest Surg 1998;2:207–16. 24. Bentrem DJ, Yeh JJ, Brennan MF, et al. Predictors of intensive care unit admission and related outcome for patients after pancreaticoduodenectomy. J Gastrointest Surg 2005;9:1307– 12. 25. Gaujoux S, Cortes A, Couvelard A, et al. Fatty pancreas and increased body mass index are risk factors of pancreatic fistula after pancreaticoduodenectomy. Surgery 2010;148:15–23. 26. Mathur A, Pitt HA, Marine M, et al. Fatty pancreas: a factor in postoperative pancreatic fistula. Ann Surg 2007;246: 1058–64.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.