Impact of pleural lavage fluid volume on perioperative outcomes in non-small cell lung cancer patients undergoing video-assisted thoracoscopic lobectomy: a randomized controlled trial.

Introduction
Pleural lavage is a standard procedure in lung cancer resection, playing a crucial role in removing tumor cells shed during surgery and enabling cytological examination, especially in lavages performed before chest closure (1-3). Recent evidence confirmed pleural lavage cytology (PLC) positivity as a strong predictor of early recurrence (4). Furthermore, current international staging guidelines stipulate that in the latest TNM classification, patients undergoing lung resection for carcinoma with positive PLC should be considered as having a microscopically incomplete resection (R1) (5). Therefore, pleural lavage has gained increasing recognition for its role in tumor cell clearance, postoperative prognostic prediction, and guiding treatment decisions in lung cancer surgery.
Pleural lavage volume critically influences perioperative outcomes through bidirectional physiological modulation. While optimized volumes reduce pleural viscosity to enhance drainage and remove inflammatory mediators, insufficient volumes risk incomplete clearance of tumor cells and debris, potentially increasing postoperative infections. Conversely, excessive lavage may disrupt pleural homeostasis, elevate non-infectious fever incidence, and prolong surgical/drainage durations while wasting resources. These volume-dependent effects consequently impact surgical duration, pneumonia rates, drainage requirements, and hospitalization length as key perioperative parameters. In clinical practice, the volume of pleural lavage fluid varies widely, with reported volumes ranging from 20 to 5,000 mL (6-24). Specific volumes used prior to chest closure include 20 mL (6,16), 40 mL (7), 50 mL (9, 10), 100 mL (11-15), 300 mL (21,22) and even as much as 5,000 mL (23). Despite the considerable variation in lavage volumes, there remains no consensus or standardized guideline for determining the optimal volume for pleural lavage (25).
This randomized controlled trial (RCT) elucidates the dose-dependent effects of pleural lavage volumes on multidimensional perioperative outcomes in non-small cell lung cancer (NSCLC) patients undergoing uniportal video-assisted thoracoscopic surgery (VATS) lobectomy and mediastinal lymph node dissection (MLND). This work aims to establish evidence-based procedural guidelines for clinical practice and improve patient prognosis. We present this article in accordance with the CONSORT reporting checklist (26) (available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2025-1059/rc).

Methods

Study design
This participant-blinded, parallel-group RCT was conducted in the Department of Thoracic Surgery, West China Hospital. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by West China Hospital’s Biomedical Research Ethics Committee (No. 536 [2022]) and informed consent was obtained from all individual participants. Minor protocol deviations from the original design published earlier (27) are detailed in the revised protocol (available upon request).

Participant enrollment
Eligibility criteria: (I) age 18–75 years; (II) planned VATS lobectomy and MLND. Concomitant resections were performed when indicated; (III) American Society of Anesthesiologists (ASA) I–II; (IV) signed informed consent. Exclusion criteria: (I) smoking cessation <2 weeks preoperatively; (II) preoperative pleural effusion (>500 mL, or persistent, accompanied by prominent symptoms such as dyspnea, chest pain, and others); (III) pregnancy/breastfeeding; (IV) mental illness; (V) comorbidities (intestinal/hematologic disorders, myocardial ischemia); (VI) neoadjuvant therapy; (VII) preoperative antibiotic use.
A secondary confirmation for eligibility was performed intraoperatively by assessing the feasibility for standardized intervention of pleural lavage. Intraoperative exclusion criteria: (I) surgical scope change, bleeding >500 mL, thoracotomy conversion, or cardiac arrest; (II) final pathology indicating non-NSCLC (benign/SCLC); (III) reoperation before discharge.
Clinical tumor-node-metastasis (cTNM) staging was determined based on preoperative imaging studies, while pathological TNM (pTNM) staging was ascertained from the final histopathological report of the resected specimens, in accordance with the 8th edition of the American Joint Committee on Cancer (AJCC)/Union for International Cancer Control (UICC) TNM classification system for lung cancer (28).

Randomization and masking
Patients were randomized (1:1) to receive pleural lavage with 1,000 or 250 mL using computer-generated numbers. Allocation concealment was ensured via sequentially numbered sealed opaque envelopes opened after enrollment. The principal investigator assigned groups according to the drawn numbers, with strict adherence to protocol by research assistants. This single-blind trial blinded participants only; investigators and the project manager were unblinded. Emergency unblinding required participant withdrawal and documentation of the process.

Procedures
All patients underwent general anesthesia with double-lumen intubation. Uniportal thoracoscopic lobectomy and MLND were performed. Prior to chest closure, hemostasis was ensured, followed by pleural lavage using 1,000 or 250 mL of 0.9% normal saline at 38–40 ℃. A single 20F chest tube (Yangzhou Hanjiang Huafi Medical Devices) was inserted apically before closure and connected to a water-seal drainage system. No external suction was applied, relying solely on gravity for drainage. To maintain tube patency and minimize fluid retention, the chest tube was manually stripped periodically by nursing staff. The chest tube was removed when the postoperative drainage volume was less than 200 mL over 24 hours and a chest X-ray confirmed complete lung expansion. Postoperative care was standardized for all patients and administered by clinical staff blinded to group assignment. This encompassed unified protocols for antibiotics, analgesia, anticoagulation, hydration, and nutrition. A standardized antibiotic prophylactic regimen consisted of intravenous cefuroxime (750 mg per dose). The initial dose was administered within 0.5–2 hours before skin incision. A second intraoperative dose was given if the surgical duration exceeded 3 hours or the estimated blood loss was greater than 1,500 mL. This was followed by postoperative doses administered every 8 hours. A multimodal analgesia protocol was implemented, which included local wound infiltration with 20 mL of 0.5% ropivacaine hydrochloride at the conclusion of surgery, followed by a patient-controlled intravenous analgesia (PCIA) pump initiated upon emergence from anesthesia and maintained for 48 hours. The PCIA solution contained fentanyl, tramadol, and tropisetron in 100 mL normal saline, configured with a background infusion of 2 mL/h, a bolus dose of 2 mL, and a 15-minute lockout period. Pain was routinely assessed using the Visual Analog Scale (VAS), and intravenous flurbiprofen axetil (50 mg) was administered as rescue analgesia for VAS scores ≥4. For anticoagulation, subcutaneous low-molecular-weight heparin (0.4 mg) was administered daily from the day of surgery until discharge, unless contraindicated by bleeding. All patients were encouraged to mobilize within 24 hours postoperatively (29).

Outcome assessment
The primary outcome of this study was the overall fever rate from surgery to discharge. Secondary outcomes included: (I) drainage parameters, including total drainage volume and duration; (II) incidence of other postoperative complications, such as pneumonia; (III) operative duration, postoperative length of stay (LOS), and total hospitalization costs. Total hospitalization costs were calculated from the hospital perspective and included all direct medical expenses related to the surgical admission, such as fees for the operation, anesthesia, perioperative medications, consumables, and room occupancy. For international readability, costs originally recorded in Chinese Yuan (CNY) were converted to U.S. dollars (USD) using the average exchange rate for the study period (April 2023–December 2024), which was approximately 1 USD =7.10 CNY. Postoperative pneumonia was defined as the presence of new or progressive pulmonary infiltrates on chest radiography or computed tomography, accompanied by at least two of the following clinical criteria: (I) fever (body temperature ≥38.0 ℃); (II) leukocytosis (white blood cell count >10×109/L) or leukopenia (white blood cell count <4×109/L); and (III) purulent respiratory secretions. Fever was defined using dual thresholds: a lower threshold of ≥37.3 ℃ (defining low-grade fever as 37.3–38 ℃) and a higher threshold of ≥38 ℃ (indicating high-grade fever), allowing for a better assessment of mild and more significant temperature elevations during the recovery process. Temperature data was collected by trained nursing staff using axillary mercury thermometers from the day of surgery to postoperative day 3 (POD0 to POD3). From POD1 to POD3, it was measured four times daily (morning, noon, afternoon, and evening).

Sample size
As the first study investigating pleural lavage volume effects after VATS lobectomy and MLND, no prior data guided sample size estimation. Power calculations were performed using a Chi-squared test for the overall postoperative fever rate in each group, with an estimated effect size of |Cohen’s h| =0.30, α=0.05, β=0.20 (PASS 15.0). The required sample size was 350 participants, providing an actual power of approximately 80.13%. Considering 10% dropout rate, 400 participants (200 each group) were targeted.

Statistical analysis
Continuous variables were presented as mean ± standard deviation (SD) for normally distributed data and as median (Q1, Q3) for non-normally distributed data, while categorical variables were expressed as counts and percentages. Statistical analyses included Student’s t-tests for normally distributed continuous variables, Mann-Whitney U tests for non-normally distributed continuous variables, and Chi-squared tests for categorical variables to compare baseline and perioperative outcomes. Effect sizes were calculated using the corresponding Cohen’s h, Cohen’s d, and r for each comparison of perioperative outcomes. P<0.05 was considered statistically significant. Analyses were performed using R 4.4.1 and SPSS 25.0, with statisticians blinded to group assignments.

Results

Baseline characteristics
From April 2023 to December 2024, 415 patients were screened (Figure 1), with 406 randomized and 9 excluded (6 declined surgery, 1 received neoadjuvant therapy, 2 used preoperative antibiotics). Among these randomized, 393 received allocated interventions and 13 were excluded (9 intraoperative pathology exclusions, 3 unsuitable for lavage, 1 thoracotomy conversion). Final analysis included 389 participants (1,000 vs. 250 mL =195:194), excluding 4 (3 missing data, 1 reoperation pre-discharge).
The baseline demographic and clinical characteristics, as well as tumor location, size, cTNM, clinical stage, pathological type, pTNM and pathological were comparable between the two groups (Table 1). And no evident difference in preoperative body temperature was observed between the groups (1,000 vs. 250 mL =36.50 vs. 36.40 ℃, P=0.37). Additionally, prophylactic antibiotics were administered preoperatively to 194 patients (99.49%) in the 1,000 mL group and 192 patients (98.97%) in the 250 mL group, with no significant difference between the two groups (P>0.99). Furthermore, the duration of prophylactic antibiotic use was identical in both groups, with a median duration of 2 days in each group (P>0.99).

Postoperative fever
Using a fever threshold of ≥37.3 ℃, a total of 78 patients developed fever during hospitalization, with an incidence of 23.08% in the 1,000 mL lavage group and 17.01% in the 250 mL group [P=0.17, |Cohen’s h| =0.152, 95% confidence interval (CI): 0.073–0.231]. When the threshold was raised to ≥38 ℃, the fever rates were 5.13% and 4.12% in the 1,000 and 250 mL groups, respectively (P=0.82, |Cohen’s h| =0.048, 95% CI: 0.006–0.090). For the daily fever rates, although the 1,000 mL lavage group generally exhibited higher compared to the 250 mL group, particularly within the low-grade fever range (37.3–38 ℃), these differences were not statistically significant as well (Table 2).
Both groups showed a similar diurnal fever pattern (low morning levels, rising daytime temperatures peaking in the evening, then declining overnight) (Figure 2). However, fever progression differed temporally: the 1,000 mL group peaked on POD1 evening, while the 250 mL group peaked 24 hours later (POD2 evening). As the postoperative period progressed, the fever rates of both groups showed a convergent trend, with nearly identical values recorded by the end of POD3.

Other perioperative outcomes
Regarding postoperative complications, pneumonia occurred in six cases, with similar incidence rates in the 1,000 mL (1.54%, n=3) and 250 mL (1.55%, n=3) lavage groups (P>0.99). Chylothorax was reported in three patients, one in the 1,000 mL group and two in the 250 mL group. Additionally, the 1,000 mL group had three unique cases of hydropneumothorax, wound exudation, and pulmonary embolism, none of which occurred in the 250 mL group (Table 3).
Minor clinical differences in surgery duration, drainage volume, and cost were observed between the two groups (Table 3). Compared to the 1,000 mL lavage group, the 250 mL group had a slightly shorter surgery duration (91.00 vs. 87.50 minutes), marginally lower drainage volume (430.00 vs. 400.00 mL), and reduced median cost ($6,697.14 vs. $6,434.25). However, none of these differences reached statistical significance (P=0.21, P=0.17, P=0.35, respectively). Additionally, postoperative pleural drainage duration and LOS were comparable between the groups, both with a median of 3.00 and 4.00 days, respectively.

Discussion
This RCT assessed perioperative outcomes of different pleural lavage volumes (1,000 vs. 250 mL) in NSCLC patients undergoing VATS lobectomy and MLND. The 1,000 mL group demonstrated earlier fever escalation and a non-significantly higher overall fever rate. Postoperative complications (e.g., pneumonia) were comparable between groups. The 250 mL group showed non-significant trends toward shorter surgery duration, reduced drainage, and lower hospitalization costs.
In this study, pleural lavage was performed before chest closure rather than immediate post-thoracotomy. The reasons are as follows: At this stage, lavage effectively removed residual blood, inflammatory exudates, and exfoliated tumor cells, thereby reducing the risks of postoperative infection, pleural effusion, and tumor cell implantation. Furthermore, since surgical manipulation may lead to tumor cell exfoliation (30), performing lavage at the end of the procedure, combined with cytological analysis, allowed for a more accurate evaluation of postoperative recurrence risk (10) and informed subsequent treatment decisions. Additionally, from a research perspective, this method ensured the validity of study results, providing a clearer assessment of the impact of different lavage volumes on postoperative outcomes. Therefore, this timing offers greater clinical relevance and scientific rigor, making it the preferred choice.
The literature defined postoperative fever as >38 ℃ for two consecutive days or >39 ℃ on any one day after surgery (31). However, due to widespread prophylactic antibiotic use in China, few patients met these criteria despite excluding preoperative antibiotic recipients. Thus, we defined fever as ≥37.3 ℃ (a lower threshold) or ≥38 ℃ (a higher threshold) on any POD. At ≥37.3 ℃, the 1,000 mL group had a clinically higher but non-significant fever rate versus the 250 mL group. At ≥38 ℃, group differences diminished, suggesting that patients in the 1,000 mL group were more likely to develop postoperative low-grade fever. Temporal analysis revealed earlier and faster fever escalation in the 1,000 mL group, with the difference being most pronounced between POD0 and the evening of POD1. Since early fever (≤48 hours) typically reflects non-infectious systemic inflammation (32), while infectious fever (e.g., pneumonia) occurs post-POD3 and exceeds 38 ℃ (31), we hypothesize that larger lavage volumes induce mechanical and chemical thoracic irritation, disrupting homeostasis and amplifying inflammatory responses, thereby driving earlier low-grade fever. Conversely, lower-volume lavage preserved physiological balance, enhancing fluid clearance and reducing low-grade fever incidence. This hypothesis aligns with known pathways where surgical trauma activates cytokine cascades [e.g., interleukin-6 (IL-6), tumor necrosis factor-alpha (TNF-α)] that drive non-infectious pyrexia within 48 hours postoperatively (32). Comparative analysis of postoperative pneumonia incidence revealed no statistically significant difference between the two groups, with rates consistent with previously reported incidences following VATS lobectomy (1.54%, n=6/389), which also supports this hypothesis (33). It should be noted that the median duration of postoperative antibiotic use in this study was 2 days. Although this practice deviates from strict prophylactic principles, it reflects the routine protocol of our center. Nevertheless, the identical antibiotic regimens used in both groups ensure the validity of intergroup comparisons. Moreover, the nearly universal administration of prophylactic antibiotics (99% in both groups) likely suppressed culture-positive infections, thereby masking later-onset infection-related fevers (e.g., after POD3). Therefore, the observed fevers—particularly the early low-grade fevers that differed in timing between groups—are more likely attributable to a non-infectious, volume-related inflammatory response rather than to infectious complications. Inflammatory markers may explain the temperature change. For instance, elevated IL-6 in pleural fluid correlates with the severity of postoperative fever, and lavage volume may modulate this response through dilution or physical stimulation (34). However, the absence of biomarker analysis, such as IL-6 and C-reactive protein (CRP) in lavage/drainage fluid, limits mechanistic insights. Future studies could prioritize correlating lavage volumes with inflammatory dynamics, measuring inflammatory markers such as IL-6, CRP, and procalcitonin in drainage fluid at 24 and 48 hours postoperatively to establish causal links between volume, cytokine release, and fever patterns.
Considering other perioperative outcomes, the 250 mL lavage protocol demonstrated comparable or superior clinical benefits, including shorter operative duration, reduced drainage volume, and lower hospitalization costs. Although the observed reduction in total hospitalization cost with the 250 mL volume was not statistically significant, it remains practically relevant. This cost efficiency—driven by lower fluid consumption and potentially shorter operative time—aligns well with the principles of value-based care. In high-volume surgical centers, adopting a low-volume lavage strategy could lead to substantial annual savings without compromising clinical outcomes, thereby supporting more efficient resource allocation. Notably, two cases of hydropneumothorax and wound exudation in the 1,000 mL group (potentially volume-related) were observed, though intergroup differences lacked statistical significance. These findings necessitate further investigation to elucidate the association between lavage volume and such complications.
Additionally, the present study did not investigate whether different volumes of lavage fluid could meet the requirements for intraoperative PLC. However, a previous multicenter, retrospective study compared the differences in cytological positivity rates using various volumes of lavage fluid (35). In this study, PLC was performed by immediately washing the thoracic cavity with 20–500 mL of physiological saline after thoracotomy, and 10–20 mL of the specimen was collected for cytological examination. Most institutions used ≤100 mL of lavage fluid, with two institutions using 500 mL. The analysis showed no statistically significant difference in PLC positivity rates across institutions (P=0.21). Thus, prior studies recommended limiting lavage to ≤100 mL. In contrast, our pre-closure lavage aimed not only to collect cytology specimens but also to remove thoracic residual blood, exudates, and tumor cells. Insufficient volumes may fail to achieve adequate cleansing. Therefore, determining the optimal lavage volume before chest closure to achieve a balance between effective specimen collection and clearance of residual material is essential. Our findings indicate that 250 mL achieves an effective cleansing outcome comparable efficacy to that of 1,000 mL, as supported by comparable rates of serious infectious complications and a potential reduction in inflammation-driven low-grade fever. The critical, yet unresolved question is whether this volume also maintains diagnostic cytological accuracy. While 250 mL may offer a better balance between effective lavage and minimized dilution than 1,000 mL, it might still exceed the optimal volume for cellular yield. Therefore, the cytological adequacy of 250 mL cannot be determined from our data and requires direct evaluation. We recommend that future studies include standardized cytological analysis to compare cellular yield and positivity rates between volumes, establishing a volume that meets both therapeutic and diagnostic purposes is essential for evidence-based guidelines.
There are several limitations to this study. First, the observed temporal shift in the fever peak (POD1 evening in the 1,000 mL group vs. POD2 evening in the 250 mL group) is consistent with a volume-dependent difference in inflammatory activation, which could be mediated by cytokines such as IL-6 or CRP. However, this mechanism remains speculative in the absence of biomarker data, radiologic and clinical parameters. Moreover, the absence of cytological analysis of the lavage fluid precludes any assessment of whether the reduced volume (250 mL) compromises the diagnostic yield for tumor cells, which is a critical aspect of pleural lavage. Furthermore, the single-blind design may have introduced potential for investigator bias in outcome assessment and limits the generalizability of our findings to other institutions with different patient populations or surgical practices. Generalizability is also constrained by the underrepresentation of patients with severe pleural adhesions and the use of a fixed-volume protocol. As this RCT did not tailor lavage volume to individual factors—such as thoracic cavity size, tumor extent, or surgical complexity—caution is advised when extrapolating these results to patients with extensive tumors or dense adhesions. Future clinical practice should consider adapting lavage volumes to patient-specific surgical conditions.

Conclusions
The current study evaluates the impact of lavage fluid volume on postoperative outcomes, particularly fever rate. Our analysis suggests that 250 mL pleural lavage demonstrated comparable efficacy to 1,000 mL in controlling postoperative fever and complications, while showing trends toward reduced resource utilization (shorter surgery duration, lower drainage volume and cost). Future research should focus on validating the cytological efficacy of reduced volumes and elucidating the underlying inflammatory mechanisms to establish comprehensive, evidence-based guidelines for personalized lavage strategies.

Supplementary
The article’s supplementary files as