Abstract
Backgound: This study aimed to determine the usefulness of the Controlling Nutritional Status (CONUT) scorescore for predicting postoperative pancreatic fistula (POPF). Patients and Methods: Data from 108 consecutive pancreaticoduodenectomy cases performed at the Surgery Department of Iwakuni Clinical Center, from April 2008 to May 2018, were included. Preoperative patient data and postoperative complication data were collected. Results: Of the 108 patients (male=65; female=43; mean age=70 years), 41 (37.9%) had indication for pancreaticoduodenectomy due to pancreatic carcinoma. Grade B or higher POPF was diagnosed in 32 patients (29.6%). In the multivariate analysis, body mass index ≥22 kg/m2 [odds ratio (OR)=5.24; p=0.005], CONUT score ≥4 (OR=3.28; p=0.042), non-pancreatic carcinoma (OR=47.17; p=0.001), and a low computed tomographic contrast attenuation value (late/early ratio) (OR=4.39; p=0.029) were independent risk factors for POPF. Conclusion: Patients with high CONUT score are at high risk for POPF. Preoperative nutritional intervention such as immunonutrition might help reduce the POPF risk in these patients.
Currently, pancreaticoduodenectomy (PD) is the primary treatment for malignant tumours involving the pancreatic head, lower bile duct, and duodenal ampulla (1, 2). This procedure is technically difficult, highly invasive and is associated with high morbidity and mortality rates (3-5). The perioperative mortality rate is still up to 5% (6-8). The most important factor affecting morbidity and mortality after PD is the development of postoperative pancreatic fistula (POPF). According to recent studies, the incidence of POPF remains high, accounting for 11.4-64.3% of all PD cases (9-15). POPF is associated with delayed gastric emptying, intra-abdominal abscesses, surgical site infections, sepsis, and bleeding after PD (16-18). Several approaches may reduce the incidence of POPF. However, a definitive approach that prevents POPF is still not available.
The Controlling Nutritional Status (CONUT) score is an automatic tool used to assess nutritional status. It takes into account laboratory examination data including serum albumin level (indicating protein reserve), total cholesterol level (indicating calorie depletion), and total lymphocyte count (indicating loss of immune defence caused by immune malnutrition) (19). The CONUT score has been used to evaluate nutritional status objectively in patients with inflammatory diseases, chronic heart failure, and chronic liver diseases (20-22). Recently, the CONUT score was demonstrated to be a predictive or prognostic marker for patients with malignancies, including colorectal, oesophageal cancer, and hepatocellular carcinoma (23-26). However, to the best of our knowledge, the usefulness of the CONUT score to assess the risk of POPF after PD has not yet been determined. Therefore, this retrospective study aimed to assess whether the preoperative CONUT score might be a useful predictor of POPF.
Patients and Methods
Patient and data collection. We reviewed the data from 108 consecutive patients who underwent PD at the Department of Surgery of the Iwakuni Clinical Center from April 2008 to September 2018. Preoperative patient data collected were sex, age, hypertension, diabetes, alcohol consumption, smoking, body mass index (BMI), CONUT score, surgery indications, main pancreatic duct (MPD) diameter determined using preoperative computed tomography (CT) or magnetic resonance cholangiopancreatography, CT attenuation values [late/early (L/E) ratio] in the pancreatic body, blood transfusion, blood loss, operative time, and use of pancreaticojejunostomy technique. Neoadjuvant chemotherapy for patients with borderline resectable pancreatic cancer was performed from March 2015 to September 2018.
The Ethics Committee at the Iwakuni Clinical Center approved the study protocol (approval number: 0191). This study was performed in accordance with the protocols of the 1975 Declaration of Helsinki.
Preoperative calculation of the CONUT score and cutoff value. The CONUT score was calculated using the serum albumin level, total lymphocyte count, and total cholesterol level for each patient (19). Albumin concentrations ≥3.5, 3.0-3.49, 2.5-2.99, and <2.5 g/dl were scored as 0, 2, 4, and 6 points, respectively. Total lymphocyte counts ≥1,600, 1,200-1,599, 800-1,199, and <800/mm3 were scored as 0, 1, 2, and 3 points, respectively. Total cholesterol concentrations ≥180, 140-179, 100-139, and <100 mg/dl were scored as 0, 1, 2, and 3 points, respectively. The CONUT score was defined as the sum of these three sub-scores. Receiver operating characteristics (ROC) curves for these scores were analyzed for prediction of POPF by comparing the areas under the curves (AUCs). The cutoff value was considered optimal when the highest Youden index (sensitivity+specificity-1) was noted (26). Analysis showed that the optimal cutoff value for the CONUT score in POPF was 3 (sensitivity=50.00%; specificity=67.11%; AUC of ROC curve=0.604). Based on the cutoff value, the patients were categorized into two groups: score ≤3, low CONUT score; and score ≥4, high CONUT score (23, 24).
In addition, data regarding postoperative complications were collected for analysis considering the incidence of POPF. No patients were excluded.
Assessment of pancreatic firmness. Reportedly, soft pancreas is a risk factor for POPF (27-32). However, pancreatic firmness solely assessed by the surgeon during surgery may not be accurate. The distinction between the end of the soft area and the start of the firm area is obscure. Pancreatic fibrosis reduces the softness of the gland. Hashimoto et al. reported that the ratio of the mean pancreatic CT contrast attenuation value (hepatic to pancreatic phase; L/E ratio) upstream from the tumour can help in the assessment of the histologic degree of pancreatic fibrosis (33). Therefore, we used the L/E ratio at the pancreatic body to assess pancreatic firmness.
Operative procedure. All patients underwent subtotal, stomach-preserving PD via an open approach, and the degree of locoregional lymphadenectomy was determined according to the preoperative diagnosis. Surgical reconstruction was performed using a modification of Child's method. The proximal jejunal stump was passed through the retrocolic pathway, and pancreaticojejunostomy, biliojejunostomy, and gastrojejunostomy were subsequently performed. Pancreaticojejunostomy was performed using the modified Kakita anastomosis (n=47; April 2008-May 2013) or the modified Blumgart anastomosis method (n=50; June 2015-present) (34). During the procedure, plastic stents were inserted into the MPD for internal drainage at the surgeon's discretion. Two or three abdominal drains were placed either anteriorly or posteriorly to the pancreaticojejunostomy and hepaticojejunostomy anastomoses.
Classification and detailed definition of POPF. POPF was diagnosed and graded based on the International Study Group on Pancreatic Fistula classification (35). POPF was diagnosed when the amylase concentration in the drainage fluid on postoperative day 3 was more than three times the upper limit of its normal serum concentration. Grade A POPF is called a biochemical fistula and is defined as the measurable fluid output on or after postoperative day 3, with amylase content higher than three times the upper normal serum level. Grade A POPF has no clinical impact on the normal postoperative pathway. Clinically significant POPFs are classified as grades B and C. POPF with an elevated inflammatory response observed in blood examination and following the intravenous antibiotic administration was defined as grade B POPF caused by infection. POPF that required drain placement for >22 days and showed no elevated inflammatory response or did not require antibiotic administration was defined as grade B POPF caused by long-term drain placement. Whenever a major change in clinical management or deviation from the normal clinical pathway was required or organ failure occurs, the fistula was classified as grade C POPF. Latent POPF (36) was defined as POPF that initially lacked amylase-rich effluent but ultimately progressed to clinically relevant POPF.
Definition of operative mortality. Operative mortality was defined as any death, regardless of cause, occurring within 30 days after surgery in or out of hospital, and after 30 days during the same hospitalisation subsequent to the operation.
Statistical analysis. Statistical analyses were performed using the unpaired Student's t-test and the chi-squared test with Fisher's exact test. All variables were assessed using univariate analyses, and only those showing statistical significance (p<0.05) were evaluated using multivariate logistic analyses to determine the primary independent risk factors of POPF. Values of p<0.05 were considered statistically significant. Statistical analysis was undertaken using JMP version 9 software (SAS Institute, Cary, NC, USA).
Results
Clinical and preoperative characteristics of the 108 patients [65 males, 45 females; median age=70 years; interquartile range (IQR)=15-88 years] are summarized in Table I. The Median CONUT score was 2 (IQR=1-3; Figure 1). The low- and high-CONUT score groups included 87 (80.6%) and 21 (19.4%) patients, respectively. The indication for PD included pancreatic carcinoma in 41 (37.9%) patients. Malignant diseases, including pancreatic carcinoma, bile duct carcinoma, ampullary carcinoma, duodenal carcinoma, gallbladder carcinoma, and metastatic tumour occurred in 84 (77.8%) cases. For 10 patients with borderline resectable pancreatic cancer, neoadjuvant chemotherapy was performed. Mean postoperative hospital stay was 22.5 (12-103) days. A total of (54.7%) patients were diagnosed with POPF, with 27 (25.0%) classified as grade A, 28 (25.9%) classified as grade B, and 4 (3.7%) classified as grade C.
The operative mortality rate in the study population was 1.9% (2/108). One patient died of aspiration pneumonia and sepsis, and another died of abdominal bleeding, both associated with POPF.
The demographic and clinical variables of the two groups classified based on the CONUT score are shown in Table II. The high-CONUT score group comprised patients who were significantly older, and had lower serum albumin level, lower lymphocyte count, lower serum cholesterol level, and a higher incidence of grade B or higher POPF than those in the low-CONUT score group. The proportion of anticoagulant use was also significantly higher in the high-CONUT score group than that in the low-CONUT score group. The duration of hospital stay of the high-CONUT score group was significantly longer than of the low-CONUT score group. There was no significant difference in other factors between these two groups.
When the demographic and clinical variables were assessed using univariate analysis to determine their relationship with POPF, no statistical significance was observed for age, sex, hypertension, diabetes mellitus, smoking history, and anticoagulant use (Table III). POPF was significantly associated with BMI ≥22 kg/m2 (p=0.015), CONUT score ≥4 (p=0.044), non-pancreatic carcinoma (p<0.001), low L/E ratio in the pancreatic body (p=0.001), and operative time ≥500 min (p=0.043).
The significant preoperative risk factors of POPF identified using the univariate analysis were incorporated into logistic regression analysis. The results showed that BMI ≥22 kg/m2 [odds ratio (OR)=5.24; p=0.005], CONUT score ≥4 (OR=3.28; p=0.042), non-pancreatic carcinoma (OR=7.17; p=0.001), and low L/E ratio in the pancreatic body (OR=4.39; p=0.029) were independent risk factors for POPF (Table III).
The cutoff values for each of the factors were determined by their respective ROC curves predicting POPF. The cutoff value for the albumin level was 3.2 g/dl, that of the total lymphocyte count was 1,020/mm3 and that of the total cholesterol level was 204 mg/dl. In the univariate analysis for POPF, the albumin level was found to be a predictive factor (Table IV). The multivariate analysis showed the CONUT score to be superior to serum albumin, total lymphocyte count and total cholesterol for predicting POPF. This study suggested that the CONUT score is more useful for predicting POPF than the individual factors that comprise the CONUT score (Table V).
Discussion
The association between preoperative nutritional status and the outcomes of surgical interventions has been considerably researched (37, 38). Based on the results of such studies, a poor preoperative nutritional status was considered to correlate with the incidence of postoperative complications. A meta-analyses and systematic reviews suggest that nutritional intervention reduces a number of postoperative complications (39, 40).
In this study, the results of multivariate logistic regression analysis showed that CONUT score ≥4, BMI ≥22 kg/m2, non-pancreatic carcinoma, and low L/E ratio in the pancreatic body were the independent risk factors for POPF. Thus, we found a correlation between the CONUT score and POPF. Patients with a high CONUT score had a significantly higher incidence of POPF than those with a low CONUT score. To the best of our knowledge, this is the first report to investigate the relationship between the CONUT score and incidence of POPF. The CONUT score is easily determined using preoperative, blood examinations and is a useful tool to predict the incidence of POPF.
With regard to CONUT score parameters, the serum albumin level isa representative nutrition marker and used frequently to assess nutrition status for prediction of POPF (41). Hypoalbuminemia is often linked to poor tissue healing, reduced collagen synthesis at anastomoses, and impairment of cell-mediated immune response, such as macrophage activation and granuloma formation (42). Therefore, surgical site infection is commonly observed in hypoalbuminemic patients.
The total lymphocyte count is also an important marker of nutrition and immunity. Menges et al. revealed that lymphopenia caused by the systemic inflammatory response is characterized by significant depression of innate cellular immunity (43). A meta-analysis demonstrated that intervention with immune-enhancing nutrition increased the total lymphocyte count and reduced postoperative complications (44).
Studies suggested that a low serum cholesterol level correlated with morbidity and mortality after gastroenterological surgery (45, 46) but the reason for this remains unclear. A decrease in cholesterol level implies not only a calorie deficiency but also that cells are being deprived of an essential nutrient required to maintain metabolic and hormonal equilibrium and membrane integrity (47). Tissue fragility may explain why the cholesterol level is associated with POPF.
POPF (grades B and C) is the most common and challenging complication of PD and has the potential to trigger life-threatening, delayed, massive intra-abdominal haemorrhage and septicaemia. The early prediction of this complication may thus improve the postoperative monitoring of patients who are at high risk for POPF. The predictors of POPF have been extensively studied. Factors that have been reported to be related to POPF are male, old age, preoperative jaundice, intraoperative blood loss, low albumin level, high American Society of Anesthesiologists score, long operative time, soft pancreas, high BMI, small MPD diameter, and pancreaticojejunal anastomosis (14, 18, 48-50).
It may be that modifying a patient's CONUT score will reduce the incidence of POPF. To maintain or improve preoperative nutrition, several approaches are being investigated. However, studies with a large cohort aiming to establish the risk factors of postoperative morbidity did not include theserum cholesterol level as a variable (3, 51). The possible usefulness of cholesterol level as a predictive marker should be confirmed by future study.
Soft pancreas and increased BMI have already been widely accepted as a patient-related risk factor that predisposes to POPF (27-32). We assessed pancreatic firmness using the L/E ratio, which has been proposed to be associated with the texture of the pancreatic parenchyma. The L/E ratio positively correlated with pancreatic firmness, which reflected the histological degree of pancreatic fibrosis. Our data showed that the L/E ratio was significantly lower in patients with POPF. As the result, a soft pancreas was associated with POPF. The high incidence of POPF in patients with high BMI or a soft pancreas may lead to increased difficulty in exposing the pancreas during surgery owing to a higher volume of abdominal and peripancreatic fat, to a higher risk of damage to the pancreatic capsule during separation due to a soft and brittle pancreas, and to a higher risk of pancreatic leakage caused by damage to pancreatic tissue and the fine pancreatic ducts because of suturing and knotting during pancreaticojejunal anastomosis (52).
Non-pancreatic cancer, such as ampullary carcinoma, bile duct carcinoma, or intraductal papillary mucinous neoplasms, was found to be a risk factor of POPF because these diseases clearly reflect the characteristics of the pancreatic remnant, such as the soft texture of the pancreas, a thin pancreatic body, and a nonfibrotic pancreatic parenchyma, which greatly increase the risk of POPF (53).
This study had some limitations. Firstly, this was a retrospective, single-centre study; therefore, there may be potential selection bias in the enrollment of patients for PD. Secondly, the sample size was small. Compared to patients with benign disease, those with malignant disease have a very different presentation and time course. Thus, the mixture of diagnoses may be a confounding factor and potentially add bias to the study. Thirdly, although the CONUT score conventionally describes the four classes of undernutrition, we used other cutoff values reported in a previous study (23, 24). In this study, the cutoff value of the CONUT score associated with PF was determined using ROCs and was also 3 (AUC=0.61). Further studies are warranted to determine more adequate cutoff values of the CONUT score to predict the incidence of POPF. Finally, this study did not compare the efficiency of the CONUT score with that of other screening systems. Further studies are required to assess the efficacy of screening systems to evaluate patient status. CONUT score ≥4, BMI ≥22 kg/m2, non-pancreatic carcinoma, and low L/E ratio in the pancreatic body were the independent risk factors of PF after PD. The CONUT score is an effective tool for assessing the preoperative nutritional status and predicting the incidence of PF after PD.
Acknowledgements
The Authors would like to thank Enago (www.enago.jp) for English language review.
Footnotes
Authors' Contributions
MU, HA, SN, SN, YU, HK, YK, FT, TA, KK and KT designed the study. HA, MU and YK treated and observed the patients. MU prepared the article and performed the literature search. HA corrected and revised the article. All Authors read and approved the final article.
This article is freely accessible online.
Consent for Publication
Patients were not required to provide informed consent for the study because the analysis used anonymous data obtained after the patient agreed to treatment by written consent.
Conflicts of Interest
The Authors declare that they have no competing interests in regard to this study.
- Received March 9, 2020.
- Revision received April 10, 2020.
- Accepted April 13, 2020.
- Copyright© 2020, International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved