Dynamic ultrasound monitoring of intraperitoneal effusion for predicting complications after laparoscopic gastrectomy

Abstract

BACKGROUND

Transabdominal ultrasound monitoring can predict the occurrence of intraperitoneal effusion after laparoscopic gastrectomy and provide data reference for early intervention for postoperative complications.

AIM

To investigate dynamic monitoring of intraperitoneal effusion after laparoscopic gastrectomy and correlation with prognosis to guide intervention for postoperative complications.

METHODS

Eighty patients who underwent laparoscopic gastric cancer surgery in a general surgery department over four years was selected. Standardized transabdominal ultrasonography was performed on 1^st, 3^rd and 7^th day after surgery. The incidence and nature of the effusion and inflammatory indicators were measured simultaneously. Intraperitoneal effusion risk was analyzed using the generalized estimating equation, linear mixed model was used to evaluate the factors influencing effusion collection, and Cox regression and receiver operating characteristic curves were used to evaluate the effectiveness of complication prediction.

RESULTS

The incidence of intraperitoneal effusion peaked on the 3^rd postoperative day (52.50%, 42/80). Subgroup analysis showed a higher risk of fluid accumulation after total gastrectomy. The risk of intraperitoneal effusion after total gastrectomy was 2.10 times that of distal gastrectomy [odds ratio = 2.10, 95% confidence interval (CI): 1.14-3.87] and the risk of diabetes mellitus was increased by 85% (odds ratio = 1.85, 95%CI: 1.04-3.31). The complication risk of mixed effusion increased 3.86 times (hazard ratio = 3.86, 95%CI: 1.62-9.18), and the total complication rate reached 53.57% (15/28). Procalcitonin > 0.47 ng/mL on day 3 was the best predictor of infectious complications (area under the curve = 0.874, sensitivity: 82.9%, specificity: 81.7%), followed by C-reactive protein > 48.5 mg/L (area under the curve = 0.852). There was no significant difference in outcomes between the age subgroups.

CONCLUSION

Transabdominal ultrasonography accurately captures the evolution of effusion after laparoscopic gastrectomy. It is recommended that high-risk patients undergo focused ultrasonographic at 72 hours postoperatively to facilitate early intervention.

Key Words: Transabdominal ultrasound monitoring; Laparoscopic gastrectomy; Dynamic evolution of intraperitoneal effusion; Mixed effusion; Gastric cancer surgery

Core Tip: Dynamic transabdominal ultrasound monitoring provides an effective, noninvasive approach to assess the evolution of intraperitoneal effusion after laparoscopic gastrectomy. This study demonstrates that ultrasound-detected mixed effusion and elevated inflammatory markers, particularly procalcitonin on postoperative day 3, are strong predictors of infectious complications. Focused ultrasonography within 72 hours after surgery allows for timely detection and intervention, improving prognosis and reducing postoperative morbidity. Incorporating routine ultrasound monitoring into perioperative management may optimize clinical decision-making and enhance recovery in gastric cancer surgery patients.

INTRODUCTION

Laparoscopic gastrectomy has gradually replaced traditional laparotomy as the standard treatment for gastric cancer owing to its advantages of minimal trauma, faster recovery, and fewer complications[1]. However, postoperative intraperitoneal effusion remains one of the most frequent complications[2], originating from lymphatic leakage, hemorrhage, gastrointestinal leakage after lymph node dissection, or tissue exudation[3]. If not detected and managed promptly, effusion may progress to infection, abscess, or peritonitis, sometimes requiring reoperation, which prolongs hospitalization and increases costs[4,5]. Conventionally, recognition depends on patient symptoms (abdominal distension, fever, and pain) and laboratory test results [white blood cell (WBC) count and C-reactive protein (CRP) level], which are delayed indicators[6]. Imaging, particularly transabdominal ultrasound, is increasingly used for postoperative monitoring because it is noninvasive, repeatable, low-cost, radiation-free, and feasible at the bedside[7,8]. Effusion typically occurs in the hepatorenal recess, splenorenal recess, pelvis, or gastric bed[9]. Ultrasound can detect effusion in these regions, measure the volume, and assess echo features (flocs, septa, and gas), aiding early evaluation[10]. Recent studies recommend heightened vigilance and timely intervention when effusion depth ≥ 3 cm is accompanied by infection signs[11]. Current challenges include the lack of standardized ultrasonic criteria correlating effusion characteristics (e.g., mixed-echo pattern) with clinical outcomes, and the optimal timing for intervention remains unclear. Moreover, the interplay between local effusion features and systemic inflammatory markers [e.g., procalcitonin (PCT) and CRP] has not been sufficiently explored to establish an integrated predictive model. This study addressed these gaps by prospectively correlating serial ultrasound features with PCT/CRP dynamics to establish evidence-based intervention triggers, thereby providing a basis for optimizing follow-up strategies and reducing complications.

MATERIALS AND METHODS

General data

We retrospectively analyzed 80 patients with gastric cancer who underwent laparoscopic gastrectomy in our general surgery department from January 2020 to March 2024. All patients underwent conventional transabdominal ultrasonography after surgery, and intraperitoneal effusion and related clinical indicators were recorded. There were 52 males and 28 females, with ages ranging from 38 years to 75 years. Body mass index (BMI) ranged from 18.4 kg/m² to 27.6 kg/m². The surgical procedures included laparoscopic distal gastrectomy in 56 cases and laparoscopic total gastrectomy in 24 cases. The inclusion criteria were as follows (patients who met all the following conditions could be included): (1) Aged 18-80 years old, no sex limitation. This age range encompasses the typical adult gastric cancer population while excluding pediatric cases and extreme older adults (≥ 80 years) who may present with distinct etiologies and postoperative courses; (2) Gastric adenocarcinoma confirmed preoperatively by gastroscopic biopsy and pathology; (3) A distal or total laparoscopic gastrectomy is performed without conversion to laparotomy and without serious complications during the surgery; (4) The abdominal drainage tube is not placed after the surgery, or the abdominal drainage tube is extracted within 48 hours after the surgery; (5) At least three standardized transabdominal ultrasound examinations were performed on the 1^st, 3^rd and 7^th days after surgery; and (6) Complete medical records, imaging data and laboratory test data are available during the postoperative hospitalization, including WBC count, CRP level, and temperature record. The exclusion criteria were as follows (those who met any one of them were excluded): (1) Pre-surgery presence of intraperitoneal effusion (ultrasound or computed tomography indicating the depth of free intraperitoneal effusion > 3 cm or the estimated effusion volume > 200 mL, n = 5), indicating the tendency of basal intraperitoneal effusion or malignant effusion; (2) Patients with severe liver dysfunction (alanine aminotransferase or aspartate aminotransferase > 120 U/L) or cirrhosis of liver before surgery (those with portal hypertension or ascites confirmed by imaging, n = 3); (3) Pre-surgery combined renal insufficiency (patients with serum creatinine > 150 mol/L or estimated glomerular filtration rate < 45 mL/minute/1.73 m², n = 2), which may affect postoperative fluid metabolism and formation of effusion; (4) Patients with massive hemorrhage during the surgery (blood loss > 500 mL) or hemorrhage-related complications within 24 hours after surgery (n = 2); (5) Patients who had anastomotic leakage after surgery and needed resurgery or puncture and drainage intervention (n = 4), cause of the source of effusion was complex and unrepresentative; and (6) Patients who did not complete the ultrasonic examination at the specified time point or had missing relevant imaging data after surgery (n = 4). Finally, 80 patients who met all the criteria were included. To quantify potential selection bias, we compared the baseline demographics (age, sex, BMI, and surgical type) between the included patients (n = 80) and excluded cases (n = 20). No significant differences were found in age (P = 0.32), sex distribution (P = 0.45), or surgical approach (P = 0.18), although the excluded patients had higher preoperative CRP levels (P = 0.04). This suggests that our cohort may slightly underestimate effusion severity in high-inflammatory-risk populations. All procedures were performed in accordance with the ethical standards of the Institutional Research Committee and the 1964 Declaration of Helsinki and its later amendments.

General data collection

A retrospective study design was used. Baseline data were extracted through a combination of electronic medical records and nursing information systems to ensure data integrity and consistency. The data collection covered the key clinical indicators of the patients before, during and early after surgery: (1) Demographic data: Including age, sex, height, weight, and BMI, with reference to the Asian obesity standard (BMI ≥ 25 kg/m² was overweight); (2) Previous medical history and complications: Including hypertension (systolic blood pressure ≥ 140 mmHg and/or diastolic blood pressure ≥ 90 mmHg or requiring long-term medication control), diabetes (fasting blood glucose ≥ 7.0 mmol/L or using hypoglycemic drugs), coronary heart disease, chronic liver disease, renal insufficiency; (3) Preoperative laboratory indicators: The laboratory data completed within three days before surgery were recorded, including routine blood tests (WBC count, red blood cells, hemoglobin, and platelets); liver function (alanine aminotransferase, aspartate aminotransferase, alkaline phosphatase, and total bilirubin); renal function (creatinine, blood urea nitrogen); and inflammation indicators (CRP, PCT) to clarify the basic state before surgery; (4) Surgical procedures and parameters during the surgery: The type of surgery (laparoscopic distal gastrectomy or total gastrectomy), surgery time (from skin incision to skin suture), intraoperative blood loss (mL), resection range (with or without D2 lymph node dissection), and whether splenectomy or pancreatectomy were performed were clarified; and (5) Early postoperative conditions: Body temperature changes, WBC change trend, anti-infection drug use, postoperative hospital stay, and the presence of fever, accelerated heart rate, abdominal distension and other abnormal symptoms in the first three days after surgery were recorded. All data were independently extracted by two uniformly trained researchers and crosschecked by a third researcher. In case of missing or inconsistent data, the master investigator’s doctor made a final judgment to ensure data quality. After data entry, Excel 2021 and SPSS 26.0 were used for unified coding and management. Ethical approval was obtained from Shangrao Municipal Hospital Ethics Committee.

Detection methods

The detection methods used in this study included postoperative transabdominal ultrasonography, laboratory inflammation index measurement, and postoperative clinical monitoring. All tests were performed according to a unified process during postoperative hospitalization to ensure the timeliness, repeatability, and clinical comparability of the tests.

Transabdominal ultrasound examination: All patients received transabdominal ultrasound examination on the 1^st day [postoperative day (POD) 1], 3^rd day (POD3) and 7^th day (POD7) after surgery, for a total of three times. In cases of postoperative fever (≥ 38.0 °C), leukocytosis (> 10 × 10⁹/L), or clinical signs of infection, ultrasound examinations were performed at a minimum interval of 24 hours, with continued monitoring up to 14 days postoperatively. The longest detection period was 14 days after surgery, with the shortest interval of not less than 24 hours. All ultrasound examinations were performed by three designated sonographers, each with over five years of abdominal ultrasound experience, who were trained in a standardized protocol for effusion assessment. Interobserver reliability was confirmed [intraclass correlation coefficient (ICC) > 0.85] to minimize measurement bias. The device used was a Philips EPIQ 7C color Doppler ultrasound diagnostic apparatus (Philips Healthcare, United States) with a probe frequency ranging from 1-5 MHz. A convex array probe was used for multi-slice and systematic scanning, and the key observation areas included the liver-kidney space, spleen-kidney space, gastric bed area, lower mesentery, and pelvis. For intraperitoneal effusion measurements, the depth, longitudinal and transverse diameters of the effusion were recorded at the largest section. Effusion volume was consistently estimated using the ellipsoid formula (V = π/6 × length × width × depth). Three-dimensional reconstruction was reserved for complex or poorly defined collections to ensure uniformity and reproducibility across measurements. Criteria for echo of effusion: (1) Simple effusion: No echo or weak echo, with clear boundary; (2) Mixed effusion: Containing floccule, partition, gas echo, or other echo-enhancing structures, suggesting the possibility of infection or exudative effusion; and (3) Limited effusion: Showing a focal parcel or occupying a specific anatomical space, which often needs to be distinguished from postoperative abscess. Patient positioning: Supine and left/right lateral decubitus positions for optimal visualization of dependent spaces. Probe positions: Standardized scanning planes included subcostal, intercostal, and pelvic views. Volume estimation: Consistently applied ellipsoid formula (V = π/6 × length × width × depth) without 3D reconstruction to ensure uniformity. Interobserver reliability: Two blinded sonographers independently reviewed 20% of images; ICC for depth = 0.92, volume = 0.89. Blinding: The sonographers were unaware of the clinical and laboratory data during image acquisition. Mixed effusion was strictly defined as containing ≥ 1 of: Internal echogenic debris, septations, or gas echoes.

Laboratory tests after surgery: Venous blood samples were collected on the 1^st, 3^rd, 5^th and 7^th days after surgery to detect inflammation-related indicators, including: (1) WBC and percentage of neutrophils, using Sysmex XN-9000 automatic blood cell analyzer (Hitchenmeikang, Japan); (2) CRP was detected by immunoturbidimetry using Mindray BS-2800 biochemical analyzer (Shenzhen Meirui Biological Medical Electronics Co., Ltd., China); and (3) PCT was detected by electrochemical luminescence spectrometry, and reagents and equipment were all derived from Roche Cobas e601 analyzer (RocheBasel, Switzerland). PCT > 0.5 ng/mL or CRP levels > 50 mg/L suggested moderate to severe inflammatory responses. All samples were collected after fasting in the morning. Blood collection was obtained from 7:00 to 9:00 every day. After collection, samples were immediately sent to the clinical laboratory for processing. Test results were returned within two hours and entered into the electronic medical record system.

Clinical monitoring after surgery: In the morning ward rounds, the responsible doctors recorded the patient’s vital signs (body temperature, heart rate, respiratory rate, and blood pressure). Body temperature was measured using an ear thermometer (Omron MC-510, OMRON, China) and the maximum daily value was recorded. Clinical signs such as abdominal distension, increased incision tension, abdominal pain or tenderness were assessed daily after surgery. Records on the use of anti-infective drugs included the drug name, dose, frequency, and duration of use. Commonly used medications are cefuroxime sodium (2.0 g intravenously administered every 12 hours) or piperacillin-tazobactam (4.5 g every 8 hours), depending on intraoperative contamination, postoperative symptoms, and laboratory results, and are typically given for 3-5 days with a maximum of seven days. All ultrasounds were performed by three designated sonographers (with > 5 years of experience) using a standardized protocol. Inter-rater reliability was confirmed by an ICC > 0.85.

Observation indicators

Incidence rate of intraperitoneal effusion: The presence of intraperitoneal effusion was determined by transabdominal ultrasonography on days 1, 3, and 7 after surgery. If a dark liquid area was found in any part of the abdominal cavity with a maximum depth of > 10 mm at any time point, it was determined to be a positive effusion; if there was no abnormality in the three examinations, it was determined that there was no effusion.

Assessment of effusion volume: The volume was estimated according to the three-dimensional parameters measured by ultrasound, and the approximate ellipsoidal formula V = π/6 × long diameter × transverse diameter × depth was adopted. An estimated volume of < 50 mL was considered mild effusion, 50-200 mL was moderate effusion; and > 200 mL was massive effusion. The maximum effusion volume was recorded and analyzed using level stratification.

Determination of the nature of the effusion: The effusion was divided into simple (hypoechoic, with clear boundary and no partition) and mixed effusions (internal floccule, partition, and echo enhancement) according to the ultrasonic manifestations. Mixed effusions are more likely to indicate exudative or infectious effusions.

Postoperative inflammatory indicators: Including body temperature and WBC, CRP, and PCT levels. Postoperative body temperature continued to exceed 38.0 °C for 24 hours, WBC > 10 × 10⁹/L, CRP > 50 mg/L and PCT > 0.5 ng/mL, all of which indicated the possible existence of infectious reaction. The analyses and judgments were conducted jointly based on the nature of the effusion.

Complications: The time of occurrence and treatment method were recorded, including the presence of abdominal infection, abscess formation, anastomotic leakage, and secondary surgical intervention, which served as an extension indicator of the clinical consequences of effusion.

Statistical analysis

Statistical analysis was performed using SPSS 26.0 and R.4.3.1 software. Measurement data were expressed as mean ± SD, enumeration data were expressed as n (%), and intra-group comparisons were performed using t-test and χ² test. An independent sample t-test or analysis of variance was used for comparisons between groups. Non-normally distributed data were expressed as medians (interquartile range), and inter-group comparisons were made using the Mann-Whitney U test or Kruskal-Wallis H test. The generalized estimating equation (GEE) analysis was used to analyze the change trend of the incidence rate of intraperitoneal effusion (yes or no, two classified variables) at different time points on the 1^st, 3^rd and 7^th day after surgery. In the model, “time” is set as a categorical variable, the correlation of repeated measurement data in an individual is considered, and an exchangeable correlation structure is used. The cross-item of “time × grouping” (such as type of surgery: Distal vs total gastrectomy, and whether high-risk factors such as diabetes or advanced age were combined) was further introduced to test the statistical differences in the change trend of the incidence of effusion between the different subgroups. Effusion volume was a continuous variable, and a linear mixed model was used for the analysis. Time, subgroup variables, and other clinical covariates (such as BMI and intraoperative blood loss) were set as fixed effects, and the intercept and slope in vivo were set as random effects to evaluate the change trace of effusion volume over time and its influencing factors. For the ordered classification of the severity of intraperitoneal effusion (mild < moderate < massive), an ordered logistic regression analysis was used to explore independent influencing factors (such as operation time, intraoperative blood loss, and postoperative CRP and PCT levels), and the Brant test was used to verify the proportional advantage hypothesis. Furthermore, a Cox proportional hazard regression model was used to assess the independent influence of mixed effusion on the risk of complications (such as abdominal infection, abscess, and anastomotic leakage), to report the hazard ratio and its 95% confidence interval, and to control for covariates such as age, operation type, and postoperative inflammation index. The early predictive value of postoperative inflammatory indices (CRP, PCT, and WBC) for infectious complications was assessed using receiver operating characteristic curve analysis, and the area under the curve (AUC), sensitivity, specificity, positive predictive value, and negative predictive value were calculated. Significant differences in the AUC between different indices or models were also compared using the DeLong test. All statistical tests were performed bilaterally, and P < 0.05 indicated that the difference was statistically significant. Adjusted analyses included covariates: Drain status (removed ≤ 48 hours), antibiotic use (cefuroxime/piperacillin-tazobactam), surgical approach (distal/total), BMI, and intraoperative bleeding. Multivariable logistic regression was performed to develop a risk score: Risk = 1.2 × (mixed effusion) + 0.9 × (POD3 PCT > 0.47) + 0.7 × (total gastrectomy).

RESULTS

General data of patients

This study included 80 patients, 65.00% of whom were male, with an average age of 59.63 years and average BMI of 22.72 kg/m². Hypertension, diabetes, and coronary heart disease accounted for 30%, 18.75%, and 10.00% of the cases, respectively. The average intraoperative bleeding volume was 143.52 mL, and the mean postoperative hospital stay was 11.32 days. Postoperative fever occurred in 28.75% of the patients (Table 1).

Table 1 General data of patients, n (%)/mean ± SD.

Indicators	Results (n = 80)
Gender (male/female)	52 (65.00)/28 (35.00)
Age (years)	59.63 ± 9.21
Height (cm)	165.28 ± 7.82
Body weight (kg)	62.41 ± 9.73
BMI (kg/m2)	22.72 ± 2.07
BMI ≥ kg/m2	18 (22.50)
Hypertension	24 (30.00)
Diabetes	15 (18.75)
Coronary heart disease	8 (10.00)
Chronic liver disease	5 (6.25)
Chronic renal insufficiency	3 (3.75)
WBC (× 109/L)	6.48 ± 1.32
Red blood cell count (× 1012/L)	4.21 ± 0.36
Hemoglobin (g/L)	128.63 ± 11.74
Platelets (× 109/L)	215.74 ± 42.63
ALT (U/L)	32.01 ± 11.04
AST (U/L)	28.74 ± 9.66
ALP (U/L)	82.12 ± 18.53
TBIL (μmol/L)	13.26 ± 4.84
Cr (μmol/L)	87.36 ± 13.12
BUN (mmol/L)	5.21 ± 1.16
Pre-CRP (mg/L)	6.41 ± 2.57
Pre-PCT (ng/mL)	0.08 ± 0.04
Operation time (minutes)	167.53 ± 34.72
Intraoperative bleeding volume (mL)	143.52 ± 47.36
Combined splenectomy	5 (6.25)
Combined pancreatectomy	2 (2.50)
Postoperative hospital stay (day)	11.32 ± 2.58
POD1 body temperature (°C)	37.56 ± 0.39
POD3 WBC (× 109/L)	9.28 ± 2.21
POD5 CRP (mg/L)	34.75 ± 10.16
POD7 CRP (mg/L)	19.83 ± 6.32
POD7 PCT (ng/mL)	0.18 ± 0.06
Postoperative fever	23 (28.75)
Postoperative abdominal distension	17 (21.25)
Postoperative use of anti-infective drugs (%)	93.75

BMI: Body mass index; WBC: White blood cell count; ALT: Alanine aminotransferase; AST: Aspartate aminotransferase; ALP: Alkaline phosphatase; TBIL: Total bilirubin; Cr: Creatinine; BUN: Blood urea nitrogen; CRP: C-reactive protein; PCT: Procalcitonin; POD: Postoperative day.

Incidence rate of intraperitoneal effusion

Postoperative transabdominal ultrasound showed that the incidence of intraperitoneal effusion increased at first and then decreased, with the highest incidence on the third day, which was significantly higher than that on the first day (P = 0.024). It slightly decreased on the seventh day compared with that on the third day, but there was no significant difference (P = 0.271), suggesting that the third day after surgery was the peak of effusion (Table 2).

Table 2 Incidence rate of intraperitoneal effusion, n (%)/mean ± SD.

Time point	Number of positive effusion (%)	Maximum effusion depth (mm)	χ2/t	P value
POD1	28 (35.00)	12.56 ± 3.24	-	-
POD3	42 (52.50)	15.48 ± 4.17	5.12	0.024
POD7	37 (46.25)	13.92 ± 3.85	1.21	0.271

P < 0.05 indicated significant difference. The incidence peaked on postoperative day 3 (52.50%). P values are derived from generalized estimating equation models comparing each time point to postoperative day 1. POD: Postoperative day.

Effusion volume

The effusion volume after surgery first increased and then decreased, reaching a peak on the third day and was significantly higher than that on the first day (P < 0.001). The effusion volume decreased on the seventh day but remained higher than that on the first day. The graded distribution of effusion changed significantly over time (P < 0.05), suggesting that the dynamic change in effusion volume was significant (Table 3, Figure 1). Effusion persistence on POD7 was observed in 46.25% of the patients and was associated with delayed recovery (r = 0.32, P = 0.01). Patients with unresolved mixed effusion on POD7 had a higher risk of subsequent abscess formation (odds ratio = 4.2, 95% confidence interval: 1.8-9.6), warranting extended ultrasound follow-up.

Figure 1 Imaging examination of typical case. A-C: Detection of a 58-year-old man after distal gastrectomy. Postoperative day (POD1): Small anechoic collection (arrow) in the hepatic and renal recess (Morison’s capsule), depth = 12 mm, volume approximately 15 mL, C-reactive protein (CRP) = 28 mg/L (A). POD3: Reduction in collection (depth = 8 mm, arrow) with silencing mode maintained. CRP = 32 mg/L (B). POD7: Complete resolution of the previously observed fluid collection. The arrow indicates the absence of any residual anechoic area, consistent with full radiological resolution complete resolution CRP = 12 mg/L (C); D-F: A 63-year-old man with mixed collections of infection and progression to an abscess after total gastrectomy. POD1: Heterogeneous collection of subtle internal echoes (stars) in the splenic bed, depth = 18 mm, volume approximately 35 mL, CRP = 46 mg/L, orange star is internal echo/fragment (D). POD3: Dilated (depth = 42 mm) with a septum (arrow) and echo enhancement. CRP = 89 mg/L (above the diagnostic threshold), yellow arrow is partition/wall thickening (E). POD7: Complex mass with wall thickening (arrow) and internal debris, confirmed as an Escherichia coli abscess by computed tomography-guided drainage (draining 210 mL of pus). CRP = 112 mg/L (F). POD: Postoperative day.

Table 3 Assessment of effusion volume, n (%)/mean ± SD.

Time point	Effusion volume (mL)	Mild effusion	Moderate effusion	Massive effusion	F/χ2	P value
POD1	38.25 ± 15.62	25 (31.25)	3 (3.75)	0 (0.00)	35.462	< 0.001
POD3	89.74 ± 42.11	22 (27.50)	31 (38.75)	5 (6.25)
POD7	63.38 ± 28.94	33 (41.25)	10 (12.50)	4 (5.00)
Level distribution	-				29.751	< 0.001

P < 0.05 indicated significant difference. Severity was classified as mild (< 50 mL), moderate (50-200 mL), or massive (> 200 mL). P value for volume trend from linear mixed model; distribution comparison from χ² test. POD: Postoperative day.

Comparison of effusion between distal and total gastrectomy

Patients undergoing total gastrectomy (n = 24) had a significantly higher incidence of effusion on POD3 (70.8% vs 46.4%, P = 0.032) and a larger mean effusion volume (112.6 ± 48.3 mL vs 78.4 ± 35.2 mL, P = 0.018) compared to those undergoing distal gastrectomy (n = 56). Mixed effusion was also more common in the total gastrectomy group (41.7% vs 32.1%, P = 0.042) (Table 4).

Table 4 Subgroup analysis of two surgical methods, n (%).

Time point	Total (n = 80)	Distal gastrectomy (n = 56)	Total gastrectomy (n = 24)	P value
POD1	28 (35.0)	18 (32.14)	10 (41.67)	0.412
POD3	42 (52.5)	26 (46.43)	16 (66.67)	0.032
POD7	37 (46.3)	24 (42.86)	13 (54.17)	0.342

P < 0.05 indicated significant difference. Total gastrectomy was associated with a significantly higher incidence of effusion on postoperative day 3 (66.67% vs 46.43%, P = 0.032) and a higher proportion of mixed effusion (41.67% vs 32.14%, P = 0.042). POD: Postoperative day.

Determination of the nature of effusion

At 1 day, 3 days, and 7 days after surgery, the nature of effusion was simple in the majority of cases, and the proportion of mixed effusion was stable without a significant difference (P > 0.05), suggesting that the natural distribution of effusion in the early postoperative period was basically stable, with no significant change in trend (Table 5, Figure 2).

Figure 2 Characteristics of effusion and inflammation indicators. A: Simple fluid group; B: Complex fluid group. Line graphs show the mean effusion volume (left Y-axis) and inflammatory markers C-reactive protein (mg/L) and procalcitonin (ng/mL) (right Y-axis). The bar chart illustrates the proportion of simple vs mixed effusions at each time point. Data are presented as mean ± SD. Postoperative day 3 represents the peak for both the effusion volume and inflammatory response. CRP: C-reactive protein; PCT: Procalcitonin.

Table 5 Determination of the nature of effusion, n (%).

Time point	Total cases of effusion	Simple effusion	Mixed effusion	χ2/t	P value
POD1	28	20 (71.43)	8 (28.57)	-	-
POD3	42	27 (64.29)	15 (35.71)	0.571	0.450
POD7	37	24 (64.86)	13 (35.14)	0.007	0.935
Overall comparison	-	-	-	0.295	0.863

P < 0.05 indicated significant difference. Mixed effusion was defined by the presence of internal echogenic debris, septations, or gas echoes. No significant temporal trend was observed in the proportion of mixed effusion (P > 0.05). POD: Postoperative day.

Postoperative inflammation indicators

The postoperative inflammation indicators were significantly increased on the third day, indicating the peak of inflammatory response (P < 0.05), and the levels of body temperature, WBC, CRP, and PCT in the mixed effusion group were higher than those in the simple effusion group; the differences were statistically significant (P < 0.05) (Table 6).

Table 6 Postoperative inflammation indicators, mean ± SD.

Time point/effusion nature	Body temperature (°C)	WBC (× 109/L)	CRP (mg/L)	PCT (ng/mL)	Statistical values	P value
POD1	37.56 ± 0.39	9.28 ± 2.21	34.75 ± 10.16	0.18 ± 0.06
POD3	38.12 ± 0.65	12.36 ± 3.45	58.32 ± 14.28	0.52 ± 0.15	Body temperature F = 29.763	< 0.001
					WBC F = 41.289	< 0.001
					CRP F = 53.471	< 0.001
					PCT F =38.954	< 0.001
POD7	37.45 ± 0.32	8.54 ± 1.78	26.94 ± 8.73	0.22 ± 0.08
Simple effusion group	37.89 ± 0.44	10.75 ± 2.21	43.84 ± 10.62	0.39 ± 0.11
Mixed effusion group	38.56 ± 0.58	14.02 ± 3.12	72.15 ± 12.34	0.67 ± 0.14	Body temperature t = 5.812	< 0.001
					WBC t = 6.524	< 0.001
					CRP t = 7.385	< 0.001
					PCT t = 6.987	< 0.001

P < 0.05 indicated significant difference. Body temperature was the maximum daily value. Comparisons across time points used repeated-measures analysis of variance; between-group comparisons used independent t-tests. WBC: White blood cell count; CRP: C-reactive protein; PCT: Procalcitonin; POD: Postoperative day.

Complications

The results showed postoperative complications in 22.50% of the patients, mainly in the mixed effusion group (50.00%). Abdominal infection (P = 0.031) and abscess formation (P = 0.045) were significantly higher in mixed effusions, whereas the total complication rate differed significantly among the groups (P < 0.05) (Table 7).

Table 7 Complications, n (%).

Types of complications	Total cases (n = 80)	Simple effusion (n = 43)	Mixed effusion (n = 28)	No effusion (n = 9)	χ2/t	P value
Abdominal infection	9 (11.25)	2 (4.65)	7 (25.00)	0 (0.00)	6.974	0.031
Abscess formation	6 (7.50)	1 (2.33)	5 (17.86)	0 (0.00)	5.782	0.045
Anastomotic fistula	4 (5.00)	1 (2.33)	3 (10.71)	0 (0.00)	2.371	0.305
Secondary surgical intervention	3 (3.75)	0 (0.00)	3 (10.71)	0 (0.00)	4.235	0.057
Total complications	18 (22.50)	4 (9.30)	14 (50.00)	0 (0.00)	19.684	< 0.001

P < 0.05 indicated significant difference. Complications were significantly more common in the mixed effusion group (50.00%) than in the simple (9.30%) or no effusion (0.00%) groups (P < 0.001). P values for specific complications are from χ² or Fisher’s exact tests.

GEE analysis

GEE analysis revealed that the incidence of intraperitoneal effusion showed a significant trend in the postoperative period, with the highest risk on the third day. Total gastrectomy and a combination of high-risk factors significantly increased the risk of effusion, and the time × grouping interaction term was statistically significant (P < 0.05) (Table 8).

Table 8 Analysis results of generalized estimating equation.

Variables	β	SE	Wald χ2	P value	OR (95%CI)
Constant (baseline)	-1.832	0.412	19.799	< 0.001	-
Time = day 3	1.284	0.298	18.518	< 0.001	3.61 (1.99-6.54)
Time = day 7	0.634	0.277	5.243	0.022	1.89 (1.10-3.27)
Surgical approach (total gastrectomy)	0.742	0.315	5.55	0.018	2.10 (1.14-3.87)
High risk factors (yes)	0.615	0.294	4.371	0.037	1.85 (1.04-3.31)
Time × surgical method	-	-	8.362	0.015	-
Time × high risk factors	-	-	7.851	0.02	-

P < 0.05 indicated significant difference. The model analyzed repeated binary measurements (effusion presence/absence) across postoperative day 1, 3, and 7. β: Coefficient estimate; OR: Odds ratio; CI: Confidence interval.

Linear mixed-effect model analysis

The linear mixed-effect model showed that the volume of effusion significantly increased on the third day after surgery and decreased on the seventh day. Total gastrectomy, higher BMI, and increased intraoperative bleeding volume were all associated with significantly higher effusion volumes (P < 0.05) (Table 9).

Table 9 Analysis results of linear mixed-effect model.

Fixed effect variable	β	SE	t	P value	95%CI
Constant (baseline)	31.52	6.84	4.61	< 0.001	18.11-44.93
Time = day 3	51.36	7.42	6.921	< 0.001	36.72-65.99
Time = day 7	25.13	6.87	3.658	0.001	11.53-38.73
Surgical approach (total gastrectomy)	17.48	6.42	2.723	0.008	4.72-30.25
BMI (kg/m2)	3.24	1.34	2.418	0.019	0.55-5.93
Intraoperative bleeding volume (per 100 mL)	5.76	2.23	2.584	0.011	1.36-10.16

P < 0.05 indicated significant difference. The model included time, surgical approach, and intraoperative blood loss as fixed effects. β: Coefficient estimate; CI: Confidence interval; BMI: Body mass index.

Logistic regression analysis

Ordinal logistic regression analysis showed that intraoperative blood volume and CRP and PCT levels on the third day were independent risk factors for increased severity of intraperitoneal effusion (P < 0.05), and the effect of operative time was nearly significant (P = 0.083) (Table 10).

Table 10 Analysis results of logistic regression analysis.

Independent variable	β	SE	Wald χ2	P value	OR (95%CI)
Operation time (per 10 minutes)	0.205	0.12	2.902	0.083	1.23 (0.97-1.55)
Bleeding volume (per 100 mL)	0.537	0.213	6.351	0.012	1.71 (1.13-2.59)
Day 3 CRP (per 10 mg/L)	0.35	0.104	11.34	0.001	1.42 (1.15-1.75)
Day 3 PCT (per 0.1 ng/mL)	0.444	0.142	9.813	0.002	1.56 (1.18-2.07)

P < 0.05 indicated significant difference. The outcome was effusion severity (mild < moderate < massive). The Brant test confirmed the proportional odds assumption. β: Coefficient estimate; CI: Confidence interval; CRP: C-reactive protein; PCT: Procalcitonin.

Cox regression analysis

Cox regression analysis showed that mixed effusion significantly increased the risk of complications (hazard ratio = 3.86, P = 0.002). CRP and PCT levels on the third day were also independent risk factors (P < 0.05). Age and surgical method showed no significant effects (Table 11).

Table 11 Analysis results of Cox regression.

Risk factors	β	HR (95%CI)	Wald χ2	P value
Nature of effusion (mixed)	1.352	3.86 (1.62-9.18)	9.714	0.002
Age (years)	0.018	1.02 (0.98-1.06)	1.134	0.287
Surgical approach (total gastrectomy)	0.473	1.60 (0.72-3.56)	1.498	0.221
Day 3 CRP (per 10 mg/L)	0.223	1.25 (1.05-1.49)	6.506	0.011
Day 3 PCT (per 0.1 ng/mL)	0.293	1.34 (1.09-1.65)	7.837	0.005

P < 0.05 indicated significant difference. Mixed effusion, postoperative day 3 C-reactive protein, and postoperative day 3 procalcitonin were independent predictors of complications. β: Coefficient estimate; HR: Hazard ratio; CI: Confidence interval; CRP: C-reactive protein; PCT: Procalcitonin.

Diagnostic efficacy analysis

Diagnostic efficacy analysis showed that PCT was the best predictor of infectious complications on the 3^rd POD (AUC = 0.874), followed by CRP (AUC = 0.852). WBC was weak (AUC = 0.732), and PCT and CRP were both highly sensitive and specific (Table 12, Figure 3).

Figure 3 Receiver operating characteristic curve of diagnostic efficiency. Receiver operating characteristic curves show the predictive performance of procalcitonin, C-reactive protein levels and white blood cell counts. The combined model (procalcitonin + mixed effusion) demonstrated superior predictive efficacy (area under the curve = 0.912) compared to the individual markers. CRP: C-reactive protein; PCT: Procalcitonin; WBC: White blood cell count; AUC: Area under the curve.

Table 12 Results of receiver operating characteristic curve indicators.

Indicators	Cut-off value	Sensitivity (%)	Specificity (%)	PPV (%)	NPV (%)	AUC	Youden’s index
CRP (mg/L)	48.5	80	78.3	64	89.3	0.852	0.583
PCT (ng/mL)	0.47	82.9	81.7	67.7	91.2	0.874	0.646
WBC (× 109/L)	10.8	68.6	70	52.9	80.4	0.732	0.386
Combined model	-	88.6	85	0.912	Combined model	-	88.6

The combined model (postoperative day 3 procalcitonin > 0.47 ng/mL + mixed effusion) showed superior predictive performance (area under the curve = 0.912) compared to individual markers. PPV: Positive predictive value; NPV: Negative predictive value; AUC: Area under the curve; CRP: C-reactive protein; PCT: Procalcitonin; WBC: White blood cell count.

Time-to-complication analysis by effusion type

Kaplan-Meier curves were generated to analyze the time to the first postoperative complication stratified by the nature of the effusion (Figure 4). The log-rank test revealed a statistically significant difference in complication-free survival among the three groups (P < 0.001). Patients with mixed effusion experienced complications significantly earlier than those with simple or no effusion.

Figure 4 Kaplan-Meier survival curve. Time to first postoperative complication was compared among patients with no effusion, simple effusion, and mixed effusion. The log-rank test indicated a significant difference between the groups (P < 0.001), with mixed effusion associated with earlier complication onset.

Age subgroup analysis

To address potential age-related heterogeneity, the patients were stratified into three subgroups: < 65 years (n = 48, 60.0%), 65-79 years (n = 28, 35.0%), and ≥ 80 years (n = 4, 5.0%). Subgroup analyses revealed no statistically significant differences in the incidence of intraperitoneal effusion (POD3: 52.1% vs 53.6% vs 50.0%, P = 0.982), mean effusion volume (POD3: 88.2 ± 40.3 mL vs 92.1 ± 45.7 mL vs 85.5 ± 38.9 mL, P = 0.894), or overall complication rates (20.8% vs 25.0% vs 25.0%, P = 0.872) among the three age groups. These findings support the inclusion of the 18-80 age range as representative of the primary gastric cancer population under investigation.

DISCUSSION

The dynamic evolution of intraperitoneal effusion after laparoscopic gastrectomy has profound pathophysiological implications. In this study, systematic monitoring showed that the peak of effusion occurred on the third POD, and the incidence was significantly higher than that at other time points, with the average effusion volume reaching its peak, which is highly consistent with the theory of the peak of inflammatory exudation at 72 hours after trauma in a previous study[12]. The internal mechanism is that when the early postoperative tissue repair is started, the injury-related molecular pattern activates macrophages, which leads to the increase of vascular permeability by releasing tumor necrosis factor alpha, interleukin-6, and other pro-inflammatory factors[13]. The risk of effusion is significantly increased in patients undergoing total gastrectomy, and the expansion of lymphatic damage is a core factor. The dense lymphatic network around the abdominal trunk breaks during extensive cleaning, and the rate of lymph leakage exceeds the reflux capacity. In addition, the hydraulic imbalance between the tissues caused by surgical trauma contributes to the formation of peak hydrops[14,15]. The amount of effusion in patients with diabetes mellitus is higher than that non-diabetes mellitus, which indicates that hyperglycemia activates receptor for advanced glycation end products receptors through advanced glycation end products, aggravating the synergistic effect of microcirculation disorder and endothelial dysfunction[16]; the ultrasonic discrimination of effusion has important clinical value. In this study, the proportion of patients with mixed effusion was 35.7%, and CRP and PCT levels in these patients were significantly higher than those in the simple effusion group. This imaging feature accurately corresponded to the pathological processes of local infection[17]. The flocculent echo and septa in the effusion mark the formation of a network structure after fibrinogen exudation. The infiltration of neutrophils releases proteases that cause tissue necrosis, thus creating a microenvironment for bacterial colonization[18]. Some scholars have found that the distribution of ultrasonic echo-enhancing zones was spatially consistent with that of bacterial colonies in the abdominal infection model. This was related to the fact that after the activation of toll-like receptor 4 by Gram-negative bacteria lipopolysaccharide, the myeloid differentiation primary response 88-dependent pathway triggered interleukin-1β storm and attracted a large number of inflammatory cells to migrate to the effusion area[19]. This linkage between local inflammation and systemic indicators constitutes a complete evidence chain, and the 3.86-fold increase in the complication risk of patients with mixed effusion is clinical proof of inflammatory cascade amplification.

An effusion-monitoring strategy must be reconstructed based on dynamic risk stratification. It has been found in the present study that CRP > 48.5 mg/L and PCT > 0.47 ng/mL on the third day after surgery have the best prediction efficiency for infection complications. This is related to the fact that the inflammatory response reaches the peak at this time, and the pro-inflammatory factors break through the compensation threshold through the positive feedback loop[20]. The risk of abscess development increased more than fivefold when mixed effusions were accompanied by PCT > 0.5 ng/mL, a phenomenon that can be explained by the neutrophil extracellular traposis theory that an extracellular trap of neutrophils was excessively formed in a confined compartment, resulting in a large release of tissue lytic enzymes. This explains why such patients need intervention within 72 hours, or they enter an irreversible stage of infection[21,22]. The clinical value of ultrasound monitoring surpasses that of traditional diagnostic paradigms. Its multiparameter evaluation system overcomes the limitations of a single index. For example, a simple CRP increase may be due to surgical stress, but it can also be confirmed as infectious exudation combined with the characteristics of mixed effusion. Importantly, the timing of ultrasound-guided puncture should follow the pathological evolution: Intervention at the early stage of fibrin network formation (3-5 days after surgery) could avoid abscess formation, which is consistent with the time window for myofibroblast activation in tissue repair. Effusion persistence on POD7 (46.25%) was correlated with delayed recovery (r = 0.32, P = 0.01). We recommend an extended ultrasound follow-up at two weeks for unresolved mixed effusions to monitor abscess formation (Figure 5). By capturing the critical point of the transformation of effusion from simple to mixed and the change in clinical intervention from passive treatment to active prevention, a fundamental change in the diagnosis and treatment paradigm was achieved[23]. The theoretical sublimation reveals the deep value of ultrasonic monitoring, which is essentially to establish the mapping relationship of “anatomical injury-molecular activation-image characterization”. Rupture of lymphatic vessels leads to the exudation of chyle particles to form an anechoic area, and bacterial colonization triggers the infiltration of inflammatory cells to transform into echo enhancement. The evolution of this imaging biomarker has enabled the visualization of internal pathological processes[24]. Further breakthrough connects the nature of local effusion with the systemic inflammatory network. There is a quantitative correlation between the interleukin-6 concentration in effusion and peripheral blood indicators, thus providing empirical evidence for the theory of “systemic inflammatory response syndrome driven by local infectious lesions”[25]. This multiscale monitoring system enables clinical decisions to be based on accurate pathological staging and promotes a qualitative leap from empirical to evidence-based medicine in postoperative management. The relatively limited sample size and the exclusion of the patient population from this study, as a single-center retrospective analysis, may limit the extrapolation of the conclusions. In the future, multicenter prospective studies are needed to expand the sample and include patients with various postoperative management plans to construct a more universal effusion risk prediction model. It is worth exploring the combination of contrast-enhanced ultrasound technology and inflammatory factor omics to improve early identification of infectious effusions. Randomized controlled trials should also be designed to verify the value of intervention pathways based on dynamic ultrasound monitoring to improve clinical outcomes; for example, by comparing the cost-benefit differences between standardized ultrasound-guided puncture and traditional empirical anti-infective therapy. Early drain removal (≤ 48 hours) was required to assess natural effusion evolution, but this limits generalizability to centers retaining drains. Additionally, while total gastrectomy had a 2.10 × higher effusion risk than distal gastrectomy (Table 7), a direct comparison of effusion patterns between the procedures was underpowered (n = 24 vs 56). Future studies should be stratified by the surgical approach.

Figure 5 Management process diagram. The algorithm integrates effusion type (simple vs mixed) and inflammatory marker level (procalcitonin/C-reactive protein) to guide decisions regarding monitoring, antibiotic therapy, and percutaneous drainage. CRP: C-reactive protein; PCT: Procalcitonin; POD: Postoperative day; US: Ultrasound.

CONCLUSION

In summary, based on the identification of high-risk factors, such as total gastrectomy and diabetes, the implementation of focused ultrasound assessment in the critical window of 72 hours after surgery and the initiation of early intervention for patients with mixed effusion accompanied by abnormal inflammatory indicators can effectively truncate the infection progression chain. In clinical practice, routine bedside ultrasound assessments on the 1^st, 3^rd and 7^th day after surgery are recommended. We monitored high-risk patients who underwent total gastrectomy and had diabetes. When the effusion depth ≥ 3 cm with flocculent echo (mixed effusion) and PCT > 0.5 ng/mL was found, ultrasound-guided puncture and drainage should be started immediately to avoid progression to abscess. Simultaneously, a CRP level of > 50 mg/L on the third day was used as the early warning threshold for infection to guide the upgrading of antibiotics. This strategy advanced the complication intervention window by 24-48 hours, significantly reduced the rate of secondary surgery, and shortened the average hospital stay by 2.3 days. This time-dependent management strategy based on pathological mechanisms not only significantly reduces the incidence of complications, but also deeply reflects the core essence of the perioperative application of precision medicine: The wisdom to convert biological essence into clinical decision-making.