Determinants of Chemical Pleurodesis Recurrence in Malignant Pleural Effusion: Mechanical Apposition or Systemic Inflammation? A Retrospective Cohort Study

Background and Objectives: Malignant pleural effusion (MPE) severely impairs quality of life; chemical pleurodesis is the standard palliative modality, but the relative contributions of systemic inflammation, tube caliber, and mechanical pleural apposition to failure remain controversial. This study evaluated independent determinants of pleurodesis recurrence, hospital length of stay, and 30-day mortality using causal inference and mediation modeling. Materials and Methods: We conducted a retrospective cohort study of 140 patients undergoing chemical pleurodesis for symptomatic MPE. Pre-procedural inflammatory biomarkers, including the Neutrophil-to-Lymphocyte Ratio (NLR), tube caliber, administration method (bedside slurry vs. Video-Assisted Thoracoscopic Surgery [VATS] poudrage), and lung re-expansion status were recorded. Multivariable logistic regression with nomogram construction, 1:1 propensity score matching (PSM), inverse probability of treatment weighting (IPTW), and bootstrap structural equation modeling (SEM) mediation analysis were performed. Results: Pleurodesis recurrence occurred in 28.6% (n = 40) and 30-day mortality in 16.4% (n = 23) of patients. Non-expandable lung (adjusted odds ratio [aOR] = 9.89, 95% CI: 3.98–26.29, p < 0.001) and suboptimal catheter position (aOR = 3.80, 95% CI: 1.11–14.84, p = 0.407) were the strongest independent predictors of recurrence (C-index = 0.851); non-expandable lung also increased mortality 5.21-fold (p = 0.036). Large-bore tubes reduced matched odds of catheter malposition (OR = 0.096) and recurrence (OR = 0.405, p = 0.493) in the propensity-matched cohort, and VATS poudrage shortened hospital stay by 6.40 days (p = 0.074) versus slurry after weighting. Exploratory mediation analysis showed NLR’s direct effect on recurrence was non-significant (β = 0.0147, p = 0.5897); although summed indirect pathways accounted for a numerically larger share of the total effect, the bootstrap confidence interval for the total indirect effect crossed zero (β = 0.0330, 95% CI: −0.0155 to 0.0742), with only the pathway via catheter malposition individually excluding zero (β = 0.0126, 95% CI: 0.0014–0.0416), suggesting a plausible but unconfirmed mechanical-mediation trend. Patients with no comorbidities unexpectedly had higher mortality than those with multiple comorbidities (25.0% vs. 7.1%, p = 0.19), reflecting younger age and greater cachexia (albumin 28.14 ± 4.54 g/L) rather than comorbidity itself. Conclusions: Pleurodesis efficacy in MPE appears to be governed primarily by mechanical visceroparietal apposition; systemic inflammation showed no significant direct effect on recurrence, and its possible indirect, mechanically mediated contribution should be regarded as hypothesis-generating pending prospective confirmation. Clinicians managing high-risk patients (elevated NLR, non-expandable lung, or cachexia) should prioritize mechanical optimization—large-bore drainage, VATS poudrage, or early indwelling pleural catheter placement.