Predictive value of C-reactive protein, procalcitonin, and total bilirubin levels for pancreatic fistula after gastrectomy for gastric cancer

Abstract

BACKGROUND

Gastric cancer is the most common malignancy of the digestive system and surgical resection is the primary treatment. Advances in surgical technology have reduced the risk of complications after radical gastrectomy; however, post-surgical pancreatic fistula remain a serious issue. These fistulas can lead to abdominal infections, anastomotic leakage, increased costs, and pain; thus, early diagnosis and prevention are crucial for a better prognosis. Currently, C-reactive protein (CRP), procalcitonin (PCT), and total bilirubin (TBil) levels are used to predict post-operative infections and anastomotic leakage. However, their predictive value for pancreatic fistula after radical gastrectomy for gastric cancer remains unclear. The present study was conducted to determine their predictive value.

AIM

To determine the predictive value of CRP, PCT, and TBil levels for pancreatic fistula after gastric cancer surgery.

METHODS

In total, 158 patients who underwent radical gastrectomy for gastric cancer at our hospital between January 2019 and January 2023 were included. The patients were assigned to a pancreatic fistula group or a non-pancreatic fistula group. Multivariate logistic analysis was conducted to assess the factors influencing development of a fistula. Receiver operating characteristic (ROC) curves were used to determine the predictive value of serum CRP, PCT, and TBil levels on day 1 post-surgery.

RESULTS

On day 1 post-surgery, the CRP, PCT, and TBil levels were significantly higher in the pancreatic fistula group than in the non-pancreatic fistula group (P < 0.05). A higher fistula grade was associated with higher levels of the indices. Univariate analysis revealed significant differences in the presence of diabetes, hyperlipidemia, pancreatic injury, splenectomy, and the biomarker levels (P < 0.05). Logistic multivariate analysis identified diabetes, hyperlipidemia, pancreatic injury, CRP level, and PCT level as independent risk factors. ROC curves yielded predictive values for CRP, PCT, and TBil levels, with the PCT level having the highest area under the curve (AUC) of 0.80 [95% confidence interval (CI): 0.72-0.90]. Combined indicators improved the predictive value, with an AUC of 0.86 (95%CI: 0.78-0.93).

CONCLUSION

Elevated CRP, PCT, and TBil levels predict risk of pancreatic fistula post-gastrectomy for gastric cancer.

Key Words: Procalcitonin; C-reactive protein; Total bilirubin; Radical gastrectomy for gastric cancer; Pancreatic fistula; Predictive value

Core Tip: This study found that the serum C-reactive protein (CRP), procalcitonin (PCT), and total bilirubin (TBil) levels on day 1 post-surgery had a significant predictive value for pancreatic fistula post-radical gastrectomy; therefore, elevated serum CRP, PCT, and TBil levels in the early post-operative period could be used to indicate the presence of pancreatic fistula.

INTRODUCTION

Gastric cancer is a malignant tumor with the highest incidence rate in the digestive system. Surgical resection is the primary treatment for gastric cancer. With continuous improvements in surgical technology, the risk of complications post-radical gastrectomy has decreased. However, the risk of post-surgical pancreatic fistula, one of the most serious complications post-surgery, has not been eliminated[1,2]. Pancreatic fistula not only cause complications, such as abdominal infection and anastomotic leakage, but also increase pain during treatment and increase costs for patients[3]. Early diagnosis and proactive measures to effectively prevent pancreatic fistula are of great significance for improving prognosis. Currently, the C-reactive protein (CRP), procalcitonin (PCT), and total bilirubin (TBil) levels are used to predict post-operative infections and anastomotic leakage during abdominal surgery[4-6]. However, their value in the early prediction of pancreatic fistula post-radical gastrectomy for gastric cancer remains unclear. To investigate the predictive value of the serum CRP, PCT, and TBil levels for post-surgical pancreatic fistula, the following study was conducted.

MATERIALS AND METHODS

Patients

This study included 158 patients who underwent radical gastrectomy for gastric cancer at our hospital between January 2019 and January 2023. The cohort comprised 88 males and 70 females, with ages ranging from 36 years to 88 years and a mean ± SD age of 63.66 ± 15.14 years. The post-operative histopathological staging distribution was: 13 patients with stage I, 50 with stage II, 66 with stage III, and 29 with stage IV; 22 underwent proximal gastric cancer resection, 93 underwent distal gastric cancer radical resection, and 43 underwent radical total gastrectomy. Post-operative histopathology confirmed 79, 61, and 18 cases of high-differentiation, low-differentiation, and mucinous adenocarcinoma, respectively. Post-surgery, 31 patients developed pancreatic fistula, including 18 males and 13 females, aged 40-86 years, with a mean age of 67.39 ± 15.88 years. Post-operative histopathological staging was: Six patients with stage I, six with stage II, 11 with stage III, and eight with stage IV. Four patients had proximal gastric cancer resection, 18 had distal gastric cancer radical resection, and nine had radical total gastrectomy. Post-operative histopathology confirmed 16, 11, and four patients with high-differentiation, low-differentiation, and mucinous adenocarcinoma, respectively. The patients who developed a pancreatic fistula post-surgery were assigned to the pancreatic fistula group, whereas those who did not develop a pancreatic fistula after surgery were assigned to the non-pancreatic fistula group.

Inclusion criteria: (1) Diagnosed with gastric cancer based on clinical symptoms, imaging findings, and histopathological examination; and (2) Underwent radical gastrectomy.

Exclusion criteria: (1) Severe liver, heart, or kidney disease; (2) Cognitive impairment; (3) Poor compliance; and (4) Missing clinical data. The study was approved by the Ethics Committee of Western Theater Command Air Force Hospital (Chengdu, Sichuan Province, China). Informed consent was obtained from all patients.

Diagnostic criteria for post-surgical pancreatic fistula

The diagnosis of post-surgical pancreatic fistula was based on the relevant standards established by the International Study Group of Pancreatic Fistula[7]: Three days or more post-surgery, the amylase level in the drainage fluid of the percutaneous drainage tube was three times higher than the normal serum level (135 U/L). For patients who did not have a drainage tube inserted during surgery or those who had the tube removed within three days post-surgery, the presence of elevated body temperature (> 38 °C), elevated white blood cell count, peritonitis, and B-mode ultrasound findings of peri-pancreatic effusion were used to confirm the diagnosis. Post-surgical pancreatic fistulas were divided into three grades: (1) Grade A, postoperative drainage for three weeks, no infection, no special treatment required, and delayed extubation was sufficient; (2) Grade B, drainage for more than three weeks or infection, requiring adjustment of strategies, which may include dietary control, nutrition, antibiotics, and drainage tube adjustment; and (3) Grade C, surgical or interventional treatment required, and serious events may occur, requiring aggressive intervention, such as fasting, nutritional support, antibiotics, somatostatin, and possibly computed tomography/ultrasound-guided drainage, which may prolong hospitalization.

Serum index measurement on day 1 post-surgery

PCT, CRP, and TBil were used to predict post-operative abdominal infection and anastomotic fistula. Within 24 hours of surgery, peripheral venous blood was collected from patients who underwent radical gastrectomy for gastric cancer. Serum was separated at 3000 rpm, and the serum PCT, CRP and TBil levels were measured using chemiluminescence[8], immunoturbidimetry[9], and the vanadate oxidation method, respectively[10]. The reagent kits were purchased from Shanghai Jining Biotechnology Co., Ltd. and operated strictly in accordance with their relevant standards.

Analysis of related influencing factors of post-surgical pancreatic fistula

Based on clinical experience and relevant studies[11-13], factors that may affect the development of pancreatic fistulas after radical gastrectomy for gastric cancer were included. They included pre-operative complications (diabetes, hypertension, and hyperlipidemia), serum nutritional indicators (hemoglobin and albumin), surgical factors (including operation time, intra-operative blood loss, intra-operative pancreatic injury, retro-pancreatic lymph node dissection, and combined splenectomy), repeat abdominal surgery, histopathological factors (including tumor size, histopathological type, and histopathological stage), post-operative related factors (including the day 1 post-surgery serum amylase level in the drainage fluid and the day 1 post-surgery serum CRP, PCT and TBil levels). Univariate and multivariate logistic regression analyses were performed to identify the independent risk factors affecting the development of pancreatic fistulas following radical gastrectomy for gastric cancer. Among them, a hemoglobin level ≥ 110 g/L was considered to be normal, and an albumin level ≥ 35 g/L was considered to be normal.

Analysis of serum CRP, PCT, and TBil levels on day 1 post-surgery in predicting post-surgical pancreatic fistula

Receiver operating characteristic (ROC) curves were used to determine the predictive value of serum CRP, PCT, and TBil levels, as individual indices or combined, on day 1 post-surgery for the development of pancreatic fistula post-radical gastrectomy for gastric cancer.

Statistical analysis

The data were analyzed using statistical product and service solutions 19.0 statistical software. Measurement data are presented as mean ± SD and t-tests were used for comparisons of two groups or a one-way analysis of variance for multiple groups, with a subsequent least significant difference-t test for pairwise comparisons, if obvious differences were identified. Comparisons between the two groups were performed using the χ² test or continuity correction test. Variables that were statistically significant in the univariate analysis were further analyzed using multivariate logistic regression, and binary logistic regression was used to calculate the predictive value of the combined indices for post-surgical pancreatic fistula. ROC curves were plotted to assess the predictive value of individual or combined indices for pancreatic fistulas. Statistical significance was set at P < 0.05.

RESULTS

Comparisons of the serum CRP, PCT, and TBil levels between the two groups on day 1 post-surgery are summarized in Table 1. On day 1 post-surgery, the serum CRP, PCT, and TBil levels were significantly higher in the pancreatic fistula group compared to those in the non-pancreatic fistula group (P < 0.05).

Table 1 Comparison of serum C-reactive protein, procalcitonin, and total bilirubin levels on day 1 post-surgery between the two groups, mean ± SD.

Groups	n	PCT (μg/mL)	CRP (mg/L)	TBil (μmol/L)
Non-pancreatic fistula group	127	0.71 ± 0.11	106.69 ± 20.46	17.69 ± 4.13
Pancreatic fistula group	31	0.99 ± 0.33	173.14 ± 32.45	20.83 ± 4.18
t value		8.020	14.266	3.786
P value		0.000	0.000	0.000

CRP: C-reactive protein; PCT: Procalcitonin; TBil: Total bilirubin.

Comparison of the serum CRP, PCT, and TBil levels on day 1 post-surgery in patients with different grades of pancreatic fistulas are summarized in Table 2. A higher grade of pancreatic fistula was associated with higher serum CRP, PCT, and TBil levels on day 1 post-surgery with significant differences between the two groups (P < 0.05).

Table 2 Comparison of serum C-reactive protein, pro-calcitonin, and total bilirubin levels on day 1 post-surgery in patients with different grades of pancreatic fistula, mean ± SD.

Grade	n	PCT (μg/mL)	CRP (mg/L)	TBil (μmol/L)
Grade A	13	0.74 ± 0.14	145.69 ± 25.13	18.69 ± 1.66
Grade B	7	0.96 ± 0.161	176.58 ± 22.691	21.59 ± 1.141
Grade C	11	1.37 ± 0.171,2	191.56 ± 36.251,2	24.15 ± 1.371,2
F value		49.375	7.682	41.902
P value		< 0.001	0.002	< 0.001

¹P < 0.05 compared with grade A.

²P < 0.05 compared with grade B.

CRP: C-reactive protein; PCT: Procalcitonin; TBil: Total bilirubin.

The univariate analysis of factors affecting the development of pancreatic fistula after radical gastrectomy for gastric cancer is summarized in Table 3.

Table 3 Univariate analysis of factors affecting the development of a pancreatic fistula following radical gastrectomy for gastric cancer, mean ± SD.

Factors	Groups	Pancreatic fistula group (n = 31)	Non-pancreatic fistula group (n = 127)	t/χ2 value	P value
Age (years)	≥ 60	20	71	0.756	0.384
	< 60	11	56
Sex	Male	18	70	0.088	0.767
	Female	13	57
Diabetes	Yes	16	30	9.460	0.002
	No	15	97
Hypertension	Yes	6	21	0.140	0.708
	No	25	106
Hyperlipidemia	Yes	13	16	14.311	0.000
	No	18	111
Hemoglobin (g/L)	Normal	29	117	0.012	0.912
	Descend	2	10
Albumin (g/L)	Normal	29	121	0.004	0.949
	Descend	2	6
History of abdominal surgery	Yes	3	7	0.196	0.658
	No	28	120
Operative time (hour)	≤ 3	19	84	0.258	0.611
	> 3	12	43
Intraoperative pancreatic injury	Yes	14	16	17.177	0.000
	No	17	111
Combined splenectomy	Yes	4	2	4.785	0.029
	No	27	125
Combined splenectomy and partial pancreatectomy	Yes	2	2	0.832	0.362
	No	29	125
Retro-pancreatic lymph node dissection	Yes	19	60	1.967	0.161
	No	12	67
Scope of gastric resection	Proximal stomach	4	18	0.080	0.961
	Distal stomach	18	75
	Whole stomach	9	34
Post-operative pathological staging	I-II	12	51	0.022	0.883
	III-IV	19	76
Serum PCT level on day 1 post-surgery (g/mL)		0.99 ± 0.33	0.71 ± 0.11	8.020	0.000
Serum CRP level on day 1 post-surgery (mg/L)		173.14 ± 32.45	106.69 ± 20.46	14.266	0.000
Serum TBil level on day 1 post-surgery (μmol/L)		20.83 ± 4.18	17.69 ± 4.13	3.786	0.000

Hemoglobin levels are considered normal at ≥ 110 g/L; Albumin levels are considered normal at ≥ 35 g/L. CRP: C-reactive protein; PCT: Procalcitonin; TBil: Total bilirubin.

The results of the univariate analysis show differences in the presence of diabetes, hyperlipidemia, postoperative pancreatic injury, and combined splenectomy, and the serum CRP, PCT, and TBil levels on day 1 post-surgery between the two groups (P < 0.05).

The multivariate analysis of the factors affecting the development of pancreatic fistula after radical gastrectomy for gastric cancer is summarized in Table 4.

Table 4 Multivariate analysis of factors affecting the development of a pancreatic fistula following radical gastrectomy for gastric cancer.

Factors	β	SE	Wald χ2 value	OR	P value	95%CI
Diabetes	1.336	0.453	8.698	3.804	0.003	1.565-9.243
Hyperlipidemia	1.016	0.337	9.089	2.762	0.003	1.427-5.347
Intraoperative pancreatic injury	1.559	0.336	18.144	21.528	0.000	2.461-9.185
Serum PCT level on day 1 post-surgery	0.841	0.216	15.159	2.319	0.000	1.518-3.541
Serum CRP level on day 1 post-surgery	0.461	0.049	88.514	1.586	0.000	1.440-1.745

CRP: C-reactive protein; PCT: Procalcitonin; CI: Confidence interval; OR: Odds ratio; SE: Standard error.

The logistic multivariate analysis shows that the presence of diabetes, hyperlipidemia, and pancreatic injury during surgery, and the serum CRP and PCT levels on day 1 post-surgery were independent risk factors for pancreatic fistula after radical gastrectomy.

The predictive values of the serum CRP, PCT, and TBil levels for post-surgical pancreatic fistula on day 1 post-surgery are summarized in Table 5 and Figure 1.

Figure 1 Receiver operating characteristic curves of serum C-reactive protein, procalcitonin, and total bilirubin on day 1 post-surgery for predicting post-surgical pancreatic fistula. CRP: C-reactive protein; PCT: Procalcitonin; TBil: Total bilirubin.

Table 5 The predictive value of serum C-reactive protein, procalcitonin, and total bilirubin levels for post-surgical pancreatic fistula on day 1 post-surgery.

Indicators	Critical value	AUC	95%CI	P value	Sensitivity	Specificity
Serum PCT level on day 1 post-surgery	0.84	0.803	0.718-0.889	0.000	74.2	83.5
Serum CRP level on day 1 post-surgery	133.83	0.792	0.697-0.887	0.000	67.7	87.3
Serum TBil level on day 1 post-surgery	20.16	0.683	0.581-0.785	0.002	64.5	63.8
Combination of the three indices		0.857	0.784-0.929	0.000	87.1	76.4

CRP: C-reactive protein; PCT: Procalcitonin; TBil: Total bilirubin; AUC: Area under the curve; CI: Confidence interval.

ROC curve analysis showed that the serum PCT, CRP, and TBil levels on day 1 post-surgery predict post-operative pancreatic fistula. The area under the curve (AUC) of the PCT, CRP, TBil levels, and their combination for predicting post-operative pancreatic fistula were 0.80, 0.79, 0.68, and 0.86, respectively. The 95% confidence intervals were (0.72-0.89), (0.70-0.89), (0.58-0.79) and (0.78-0.93), respectively. The sensitivities and specificities for PCT, CRP, TBil, and the three combined indices were 74.2%, 67.7%, 64.5%, 87.1%, respectively, and 83.5%, 87.3%, 63.8%, 76.4%, respectively. The AUC and sensitivities of the three combined indices were higher than those of each individual index.

DISCUSSION

Pancreatic fistula refers to the phenomenon of rupture of the pancreatic duct and the outflow of pancreatic juice through non-physiological pathways. This is a common complication of radical gastrectomy for gastric cancer. It has been reported that the incidence of pancreatic fistula development post-radical gastrectomy for gastric cancer is between 10% and 20%[14]. Once a pancreatic fistula occurs, it increases the incidence of hemorrhage, infection, and treatment costs and may also cause death if not diagnosed and treated in a timely manner[15,16]. The symptoms and imaging manifestations of post-surgical pancreatic fistula and abdominal infection are similar and easily misdiagnosed. In addition, the diagnosis of pancreatic fistula is mainly based on the amylase level in the drainage fluid of the percutaneous drainage tube. However, this diagnosis is not applicable to patients who do not have a drainage tube placed during surgery. The combination of elevated body temperature (> 38 °C), abnormal white blood cell count, peritonitis, and ultrasound examination data can delay the diagnosis of pancreatic fistula, thus impeding timely intervention.

Serum CRP is a sensitive inflammatory marker that participates in various pathophysiological processes. It is a relatively stable and common acute-phase response protein used in clinical practice[17]. The PCT level is a common indicator of bacterial and fungal infections[18]. Studies have reported that an elevated serum PCT level on day 2 and day 5 post-pancreaticoduodenectomy have better positive and negative predictive values for diagnosing post-surgical pancreatic fistula, and in combination with the CRP level has a better predictive value for post-surgical pancreatic fistula[19,20]. The present study has demonstrated that patients who underwent radical gastrectomy for gastric cancer and developed pancreatic fistula post-surgery had significantly higher serum CRP and PCT levels on day 1 post-surgery than those who did not develop pancreatic fistula. In addition, a higher grade pancreatic fistula is associated with higher serum CRP and PCT levels on day 1 post-surgery. This suggests that the serum CRP and PCT levels on day one post-surgery are of great significance in reflecting the risk and severity of pancreatic fistula post-surgery. This is important because pancreatic fistula is a serious post-operative complication that can lead to electrolyte imbalance, infection, hemorrhage, organ failure, and other complications. The leakage of pancreatic juice into the abdominal cavity triggers widespread inflammation, prompting the release of inflammatory factors such as PCT and CRP; PCT is a sensitive indicator of inflammation, which increases during surgery and infection, and further increases when a pancreatic fistula is present due to abdominal inflammation; CRP is an acute reactive protein that rises rapidly during infection and inflammation, and also significantly increases in patients with pancreatic fistula. The higher the grade of pancreatic fistula, the more severe the inflammation in the abdominal cavity, with associated higher PCT and CRP levels. Relevant studies have also found that the PCT and CRP levels in patients with pancreatic fistula are higher than those in patients without pancreatic fistula[21,22].

TBil is a breakdown product of hemoglobin and an antioxidant produced autonomously in humans. An excessively high TBil level indicates abnormal bile metabolism, leading to increased pressure in the biliary system, causing biliary inflammation and further spreading to the pancreas, which is prone to pancreatic fistula[23]. In this study, the serum TBil level in 31 patients who developed pancreatic fistula post-surgery was significantly higher than for those in the non-pancreatic fistula group on day 1 post-surgery. As the pancreatic fistula grade increased, the serum TBil level on day 1 post-surgery was elevated, indicating that the serum TBil level on day 1 post-surgery is valuable for predicting post-surgical pancreatic fistula and the severity of the pancreatic fistula. This may be because a high TBil level may also affect the microcirculation of the pancreas, damage the microvasculature of the pancreas, leading to ischemia and hypoxia of the pancreatic tissue, which in turn can cause pancreatitis or pancreatic fistula. Also, microcirculation disorders may also affect the ability of the pancreas to repair and regenerate, making it more difficult for the pancreas to recover after damage, thereby increasing the risk of pancreatic fistula. Li et al[24] also reported that a high TBil level is a risk factor for pancreatic fistula.

In addition, this study screened 18 factors related to the occurrence of pancreatic fistula post-surgery and conducted a single-factor analysis, which revealed that the presence of diabetes, hyperlipidemia, post-operative pancreatic injury, and combined splenectomy, and the serum CRP, PCT, and TBil levels on day 1 post-surgery differed. Multivariate analysis confirmed that the presence of diabetes, hyperlipidemia, and intraoperative pancreatic injury, and the serum CRP and PCT levels on day 1 post-surgery were independent risk factors for pancreatic fistula after radical gastrectomy. The ROC curve analysis showed that the serum CRP, PCT, and TBil levels, individually, on day 1 post-radical gastrectomy for gastric cancer, had predictive value for post-operative pancreatic fistula, but the combination of the three indices had a higher predictive value.

This study had limitations as it was conducted at a single-center with a small sample size and a limited research design. The serum indices were not measured at multiple time points, such as pre-operatively, intra-operatively, or post-operatively.

CONCLUSION

In summary, the serum CRP, PCT, and TBil levels on day 1 post-surgery have predictive value for the development of pancreatic fistula following radical gastrectomy for gastric cancer, thus elevated serum levels of these indices in the early post-operative period can serve to predict the risk of pancreatic fistula.