Assessment of Preoperative Clinicophysiologic Findings as Risk Factors for Postoperative Pancreatic Fistula After Pancreaticoduodenectomy

A gradual decrease in the incidence of short-term complications after pancreaticoduodenectomy (PD) has been observed, and recent improvements in operative technique and perioperative management have resulted in an increase in the number of long-term survivors.¹ PD is an important method for the surgical treatment of tumors of the head and the periampullary parts of the pancreas. In recent years, the mortality rate after PD decreased to less than 5% in high-volume centers; however, the morbidity rate remains high between 30% and 50%.^2–5 One of the main reasons for the high morbidity rate after PD is postoperative pancreatic fistula (POPF).^6,7 POPF occurs in 5% to 30% of patients after PD,⁸ and its incidence has been observed to be fairly constant over the past 30 years even in high-volume centers.^9,10 The definition of POPF provided by the International Study Group on Pancreatic Fistula (ISGPF) has been helpful for the accurate comparison of different surgical experiences.¹¹ After this proposal of POPF, a standardized POPF definition was established: fluid output of any measurable volume via an operatively placed drain with amylase activity greater than 3 times the upper normal serum value. To update the definition of POPF by the ISGPF,¹² the grading of POPF had been changed to include the following 3 grades of severity: biochemical leak, grade B, and grade C. In a previous study, the risk factors of POPF were body mass index,^13–15 sex,^13–15 a narrow pancreatic duct width.^13,14,16 and pancreatic texture.^13,17 To date, only a few studies assessing preoperative clinicophysiologic findings as risk factors for POPF after PD have been performed; thus, we aimed to evaluate preoperative clinicophysiologic findings as risk factors for POPF after PD.

Materials and Methods

In total, 350 patients who underwent PD between 2007 and 2012 were examined retrospectively at Tokyo Women's Medical University. All patients were classified into 2 groups as follows: group A (no fistula/biochemical leak group, 289 patients) and group B (grade B/C of POPF group, 61 patients). Variables, including operative characteristics, length of stay in hospital, morbidity, mortality, and data regarding clinicophysiologic parameters, including sex, age, body mass index, biliary decompression, total bilirubin, albumin, creatinine, glycated hemoglobin (HbA1c), amylase, C-reactive protein, white blood cell (WBC) count, lymphocyte number, hemoglobin, platelet, carbohydrate antigen (CA)19-9, carcinoembryonic antigen (CEA), neutrophil/lymphocyte ratio (NLR), prognostic nutritional index (PNI), platelet/lymphocyte ratio (PLR), and C-reactive protein/albumin ratio (CAR), were collected and analyzed as predictors of POPF in univariate and multivariate analyses. All clinicophysiologic parameters were measured immediately before pancreatic resection under cholangitis and pancreatitis control.

Results

There were 213 male and 137 female patients, with a mean age was 65.4 years (range, 21–87 years). The enrolled patients had the following diseases: pancreatic cancer, 131 patients (37.4%); distal bile duct cancer 68 (19.4%); intraductal papillary mucinous neoplasm 58 (16.6%); papilla Vater cancer 42(12%); neuroendocrine neoplasm 12 (3.4%); serous cystic neoplasm 4 (1.1%); solid pseudopapillary neoplasm 3 (0.9%); and others 21 (6%) (Table1). Furthermore, operative procedures were pylorus-preserving PD, 280 (80%); subtotal stomach-preserving PD, 49 (14%); PD, 20 (5.7%); and duodenum-preserving pancreatic head resection 1 (0.3%; Table 1). Table 2 shows the preoperative clinicophysiologic findings of this study. The univariate analysis showed that sex (P = 0.047), amylase level (P = 0.032), PNI (P = 0.001), and CAR (P = 0.005) were independent risk factors for POPF, between group A and group B regarding preoperative clinicophysiologic findings (Table 3). In contrast, multivariate analysis showed that sex (P = 0.045) and PNI (P = 0.012) were independent risk factors for POPF between group A and group B regarding preoperative clinicophysiologic findings (Table 4).

Table 1 Patient characteristics

View Fullscreen

Table 2 Preoperative clinicophysiologic characteristics

View Fullscreen

Table 3 Univariate analyses between group A and group B on preoperative clinicophysiologic findings

View Fullscreen

Table 4 Multivariate analyses between group A and group B on preoperative clinicophysiologic findings

View Fullscreen

Discussion

PF has been reported to be a potentially lethal complication because it could result in delayed massive hemorrhage and septicemia after PD.^18–21 However, there is still no established procedure to prevent POPF after PD; thus, preoperative identification of patients at high risk for POPF is important for improving the morbidity and clinical outcome. Previous studies on preoperative clinicophysiologic findings for POPF are unavailable. In previous reports, preoperative serum-based inflammatory and nutritional indicators derived from clinicophysiologic findings, such as NLR, PLR, modified Glasgow prognostic score, PNI, and CAR, have been linked to prognosis in many types of cancer.^22–25 In this study, we evaluated preoperative clinicophysiologic findings as risk factors for POPF after PD and suggested PNI (≤48.64 U/mL) to be an effective biomarker for POPF.

As obtained in previous reports,^13–15 male sex was observed to be an independent POPF-related factor and was repeatedly revealed as a risk factor of POPF.²⁶ Mathur et al²⁷ reported that patients with POPF have significantly more pancreatic fat and less fibrosis than those without POPF. In addition, the presence of increased fat in the pancreas, along with a small nondilated duct, may be associated with decreased local blood flow and may increase the risk of perioperative pancreatitis. Dixon²⁸ described that men had significantly more fat within the abdominal cavity than women, even among people with similar total amounts of fat. This may be one of the causes of the significantly high incidence of POPF in men compared with women.

PNI had been considered a predictive factor of the prognosis and postoperative complication of various cancers.^29–31 PNI included a combination of the albumin and total lymphocyte count and was initially used to evaluate immunologic and nutritional status.³² Aida et al³³ revealed that preoperative immunonutrition, which modulates prostaglandin E2 production and T-cell differentiation, may prevent postoperative complications in patients after PD. Low PNI level was considered to have resulted from hypoalbuminemia and/or lymphocytopenia. Hypoalbuminemia is associated with poor tissue healing and impairment of immune responses related with macrophage activation and granuloma formation.^34,35 Lymphocytopenia has been known to decrease the antitumor effect induced by the cellular immunity of T lymphocytes.³⁶ An immunonutritional disorder causes a decline in albumin concentration and total lymphocyte count, including helper T cells, interleukins 2 and 3, and T cell blastogenic responses.³⁷ Kanda et al³⁸ reported an association of PNI with overall survival and postoperative complications, particularly pancreatic fistula, in patients with pancreatic cancer. They revealed that PNI less than 45 U/mL was a significant predictor of POPF in patients with pancreatic cancer, a result that is strongly similar to our finding of PNI ≤ 48.64 U/mL as a predictive risk factor for POPF. Therefore, PNI may be used as a biomarker of POPF after PD.

There are several limitations to this study. First, this study was retrospective in design; however, the PD cases were consecutive. Second, some of the parameters whose data could not be measured all round. Hence, additional multicenter investigations involving larger patient populations are needed before definitive conclusions can be drawn.

In conclusion, PNI (≤48.64 U/mL) and male sex are significant risk factors for POPF after PD, and low PNI level was suggested to be an effective biomarker for POPF after PD.