Video-Assisted Thoracoscopic Surgery (VATS) serves as the bedrock of minimally invasive thoracic surgery (1). Its application in the diagnosis and treatment of pulmonary nodules has significantly reduced surgical trauma and accelerated patient postoperative recovery (2). Within this technical framework, uniportal VATS represents a further refined minimally invasive approach (3). By integrating the observation port with the operating port, it achieves less bodily tissue injury and superior postoperative pain control, leading to its widespread clinical adoption and application in anatomical pulmonary resections such as wedge resections and segmentectomies (4).

Despite continuous technological advancements, Prolonged Postoperative Chest Tube Drainage (PPCTD) remains a common clinical management challenge following VATS (5). Given its high incidence and direct impact on patient hospitalization duration and healthcare resource consumption, this metric is often regarded as a core endpoint for assessing the quality of perioperative outcomes (6). Abnormal prolongation of drainage time is not only a direct cause of patient discomfort and increased hospitalization costs but also an independent risk factor for complications such as pleural infection, persistent air leak, and atelectasis, severely constraining the effective implementation of the Enhanced Recovery After Surgery (ERAS) protocol in the field of thoracic surgery (7, 8).

Currently, most scholarly discussions on risk factors for prolonged postoperative drainage originate from retrospective studies on multiportal VATS or lobectomy (9). However, uniportal VATS pulmonary resection possesses unique characteristics in terms of surgical access, establishment of the operating triangle, and strategies for resolving instrument conflict (10, 11). For instance, the coincidence of the visual and operational axes may increase the technical difficulty of exposing deep-seated lesions and precisely placing stapler cartridges. The impact of these specific technical challenges on postoperative wound healing and pleural fluid production has not been fully elucidated or quantitatively evaluated. Furthermore, the predictive efficacy of patient-specific physiological reserves, such as preoperative impaired lung function and nutritional risk, as modifiable clinical variables, for drainage time in the context of this specific technique remains to be confirmed (12).

Therefore, to address the aforementioned research gaps, this study aims to systematically evaluate the clinical data of patients undergoing uniportal VATS pulmonary resection through a retrospective cohort analysis, with the goal of thoroughly identifying the independent risk factors leading to PPCTD. The conclusions of this study are expected to provide a solid evidence-based foundation for risk stratification, predictive medical intervention, and the development of personalized enhanced recovery protocols tailored to this specific surgical technique.

This study focused on identifying independent risk factors for PPCTD following uniportal VATS pulmonary resection. Through retrospective analysis of clinical data from 185 patients, we identified deep nodule resection, FEV1% < 70%, use of ≥4 stapler cartridges, intraoperative blood loss ≥250 mL, and preoperative anemia as independent predictors of this complication. These findings not only elucidate the multifactorial mechanism underlying prolonged drainage in this specific surgical approach but also provide crucial evidence for clinical risk stratification and individualized patient management. It is important to explicitly state that chest tube duration, as the primary endpoint, is inherently influenced by institutional protocols (detailed in Section 2.4), suction strategy, drain system characteristics, and ultimately, clinician decision-making regarding removal. The lack of universal standardization in these practices means that the identified associations could be influenced by these center-specific factors, which is an inherent limitation of observational studies in this domain.

The multivariate logistic regression analysis clearly identified five independent risk factors: deep nodule resection (OR = 6.625), preoperative FEV1% < 70% (OR = 5.590), use of ≥4 stapler cartridges (OR = 4.775), intraoperative blood loss ≥250 mL (OR = 4.064), and preoperative anemia (OR = 3.434). These findings not only confirm the combined decisive role of surgical technical complexity and patients' baseline physiological status in postoperative recovery, but also highlight the unique weighting of certain risk factors specific to the uniportal approach. Although the type of resection (lobectomy vs. sublobar resection) did not emerge as an independent risk factor in our multivariate model, we acknowledge that different resection extents inherently carry varying risks of air leak and pleural fluid production. The comparable risk profile observed in our cohort may reflect the standardized technical approach and drain management protocol applied in our uniportal VATS program. Nevertheless, clinicians should remain aware that extensive resections such as lobectomy may still pose higher intrinsic risks, and individualized assessment is warranted.

Deep nodule resection was confirmed as the most significant independent predictor. This strong association relates directly to the inherent technical challenges of managing deep-seated lesions during uniportal VATS (13). Compared to the multi-angular visualization provided by multiportal approaches, the single incision in uniportal VATS leads to frequent instrument conflict and complete overlap of the visual and operative axes (14). These limitations substantially compromise the exposure, precise dissection, and optimal placement of stapler cartridges for deep nodules. To achieve complete resection of nodules located >2 cm from the visceral pleura, surgeons typically need to perform more extensive parenchymal dissection (15). This not only increases the risk of unplanned pulmonary tissue tearing or contusion but also creates a larger potential surface for leakage (16). Furthermore, during deep stapling procedures, increased tissue thickness and tension may compromise ideal cartridge formation, consequently elevating the probability of complications such as persistent alveolar air leak or increased tissue fluid exudation, which can contribute to prolonged drainage (17).

The use of ≥4 stapler firings was identified as another crucial independent risk factor. The number of stapler cartridges serves as an objective quantitative indicator of surgical extent and anatomical complexity (18). Increased cartridge usage directly corresponds to longer pulmonary parenchymal cutting and staple lines, which undoubtedly expands the potential surface area for postoperative air leakage (19). From a biomechanical perspective, longer staple lines are subjected to more complex tension distribution during lung re-expansion, where localized stress concentration may elevate the risk of staple line disruption (20). From a tissue response standpoint, repeated stapling may induce more pronounced local tissue compression, ischemia, and inflammatory response along the elongated staple line, which could collectively impair the regional microcirculation and consequently delay wound healing and re-epithelialization (21, 22). These findings strongly suggest that, while maintaining oncological radicality and surgical safety, optimizing resection planning and enhancing stapling efficiency to minimize stapler cartridge usage may serve as an important technical strategy for reducing drainage duration (23).

Intraoperative blood loss ≥250 mL, serving as a direct reflection of surgical trauma magnitude, demonstrates a strong association with PPCTD consistent with the logical pathophysiology. Substantial bleeding typically accompanies more extensive tissue dissection, denser adhesionlysis, or unanticipated vascular injury (24, 25). This degree of trauma triggers more intense local and systemic inflammatory responses, releasing abundant inflammatory mediators. These mediators not only increase systemic capillary permeability, promoting transudation of vascular fluid components into interstitial spaces and the pleural cavity, but may also directly inhibit the reparative function of pleural mesothelial cells. These combined effects ultimately lead to increased postoperative pleural drainage volume and prolonged drainage duration (26, 27).

Regarding patient-related factors, the significant predictive value of preoperative FEV1% < 70% underscores the crucial role of baseline pulmonary function in postoperative recovery (28). Impaired lung function frequently coexists with underlying pathological conditions such as chronic obstructive pulmonary disease (COPD) (29). The mechanisms leading to PPCTD in these patients are multifactorial: First, diminished lung elastic recoil and respiratory muscle strength contribute to inadequate postoperative lung re-expansion and increased residual pleural dead space, thereby impairing physiological absorption of pleural fluid (30). Second, effective cough capability is essential for clearing respiratory secretions and promoting lung reexpansion; patients with poor pulmonary function typically exhibit compromised cough efficiency, facilitating pleural fluid retention (31, 32). Furthermore, chronic hypoxemia may inhibit fibroblast proliferation and collagen synthesis and deposition, consequently delaying mechanical healing of pleural wounds (33). Finally, such patients often manifest a persistent chronic inflammatory state with elevated serum levels of pro-inflammatory cytokines (e.g., IL-6, TNF-α), and surgical trauma exacerbates this response, leading to more pronounced inflammatory exudation (34).

It is important to note that intraoperative variables such as blood loss ≥250 mL and the use of ≥4 stapler cartridges, while identified as independent risk factors in our model, may also serve as robust indicators of underlying surgical complexity. Significant bleeding often accompanies challenging dissections (e.g., in cases with dense adhesions or deep-seated lesions), and a high number of stapler firings directly reflects the extent of parenchymal transection and the technical demands of the resection. Thus, these factors likely represent both measures of procedural complexity and contributors to prolonged drainage through distinct pathophysiological pathways (e.g., inflammatory exudation or increased staple line surface area). This distinction cautions against interpreting them solely as direct causal agents but underscores their high utility as integrated, objective markers for intraoperative risk stratification. From a clinical perspective, their value lies in alerting the surgical team to cases with higher technical demands, prompting enhanced vigilance in hemostasis, meticulous stapling technique, and anticipatory postoperative management.

This study represents the first identification of preoperative anemia as a non-negligible independent risk factor in the context of uniportal VATS pulmonary resection. Anemia signifies impaired blood oxygen-carrying capacity. Adequate tissue oxygen supply is essential for all phases of wound healing, including the inflammatory, proliferative, and remodeling stages. Pre-existing anemia creates a background “oxygen debt,” which surgical trauma further exacerbates by worsening the oxygen supply-demand imbalance. This condition may directly suppress fibroblast activity and collagen synthesis, consequently delaying the healing process at both the pleural surface and pulmonary parenchymal margins. Such delayed healing prevents the rapid restoration of equilibrium between pleural fluid production and absorption. This finding highlights the importance of preoperative identification and active correction of anemia as a crucial modifiable target for optimizing preoperative patient condition and improving postoperative outcomes.

The sensitivity analysis further enriched the clinical value of these risk factors. The combined prediction model demonstrated excellent discriminative performance (AUC = 0.892), significantly surpassing any single indicator, strongly supporting the comprehensive assessment of multiple risk factors in clinical practice rather than reliance on individual parameters. Among these, preoperative FEV1% (AUC = 0.812) and intraoperative blood loss (AUC = 0.801) exhibited outstanding predictive capability, suitable for preoperative evaluation and intraoperative risk warning respectively. The number of stapler cartridges used (AUC = 0.758), with its high sensitivity, serves as a practical tool for early postoperative screening of high-risk patients. Although deep nodule resection and preoperative anemia showed relatively limited individual predictive performance, their strong independence in the multivariate model warrants heightened clinical vigilance when present.

The identification of five binary predictors with relatively comparable effect sizes (β coefficients ranging from 1.234 to 1.891) facilitates the conceptual translation of our model into a simple, clinically actionable risk score for PPCTD. A practical implementation could involve two stages: (1) a preoperative score utilizing only preoperatively assessable variables (CT-estimated deep-seated nodule, FEV1% < 70%, and anemia) for initial risk counseling and targeted prehabilitation; and (2) a postoperatively updated score that incorporates intraoperative variables (blood loss ≥250 mL and use of ≥4 stapler cartridges) for final risk stratification and guidance of postoperative management. Given the similarity in effect sizes, two straightforward scoring methods are proposed based on our data: Option A (Simple Additive Score): Assign 1 point for the presence of each of the five predictors, yielding a total score of 0–5. Risk categories (e.g., low: 0–1, intermediate: 2–3, high: 4–5) could be defined empirically based on the observed event rates within our cohort. Option B (Weighted Integer Score): Assign points proportionally to the β coefficients (e.g., deep nodule resection: 4 points; FEV1% < 70%: 3 points; ≥4 cartridges: 3 points; blood loss ≥250 mL: 3 points; anemia: 2 points), yielding a total score of 0–15, with cut-offs calibrated to predicted probabilities. The β coefficients and event rates provided in this study serve as the foundation for such scores. Their formal derivation, optimal threshold determination, and validation of clinical utility represent a vital next step for prospective, multi-center studies.

This study has several limitations inherent to its design and setting. First, as a retrospective, single-center observational cohort analysis, it is susceptible to selection and information bias, and the associations identified cannot establish causality. Second, the predictive model, despite robust internal validation, lacks external validation in an independent cohort, which is necessary to confirm its generalizability to other institutions and patient populations. Third, our findings may be influenced by institution-specific perioperative protocols, particularly standardized chest tube management practices (e.g., criteria and timing for removal, use of suction), which could affect the primary outcome and limit direct comparability with centers using different protocols. Fourth, although our sample size is considerable, it remains insufficient for detailed subgroup analyses (e.g., by specific resection type or cancer stage) and for including all potential confounders, such as detailed analgesic regimens or dynamic drainage trends. Fifth, we developed a single integrative predictive model. While this model is most applicable for postoperative risk assessment, the construction of separate, phase-specific models (e.g., a purely preoperative model for counseling and an intraoperatively-updated model for management)—a methodologically appealing suggestion—was beyond the scope of this study. Future large-scale, multicenter prospective studies are needed to externally validate our model, control for center-specific practices, develop stage-specific predictive tools, and ultimately implement refined risk-stratified clinical pathways. Fifth, due to the retrospective nature of the study, detailed and standardized data on specific sublobar resection techniques (wedge vs. segmentectomy), exact lymphadenectomy templates, and comprehensive pathological subclassifications were not uniformly available for the entire cohort. Consequently, our analysis adjusted for the broad category of resection extent (lobectomy vs. sublobar) but could not fully account for the potential variability within these categories or the nuances of nodal dissection. It is important to recognize that unmeasured heterogeneity in resection extent and technique may act as a confounder and partly explain the observed associations with predictors such as stapler use and blood loss. Therefore, the identified risk factors should be interpreted as valid within the documentation constraints of this single-center dataset, and their independent contribution should be confirmed in future studies with more granular procedural data. Third, our findings may be influenced by institution-specific perioperative protocols, particularly standardized chest tube management practices (e.g., criteria and timing for removal, use of suction), which could affect the primary outcome and limit direct comparability with centers using different protocols. Furthermore, while we provided an evidence-based justification for the >7-day cutoff and performed a sensitivity analysis with an alternative threshold (>5 days), it should be acknowledged that definitions of prolonged drainage vary across institutions and studies, which may affect the comparability of risk estimates.

In summary, while deep nodule resection, use of ≥4 stapler cartridges, blood loss ≥250 mL, preoperative FEV1% < 70%, and anemia were identified as independent risk factors for PPCTD in our cohort, readers should be mindful that these associations are derived from a retrospective analysis with inherent data limitations, including unmeasured variation in specific resection techniques. These predictors provide a useful framework for risk assessment in settings with similar surgical documentation practices, but their generalizability and independent causal roles require validation in prospective studies with standardized, detailed procedural reporting.