**Background:** Heart failure involves complex pathophysiological changes including pulmonary congestion, reduced cardiac output, neurohumoral activation, and renal dysfunction. Each of these factors may independently affect the hemogram (white blood cell count, hemoglobin level, and platelet count), but their confounding relationships make it difficult to isolate individual effects. Recent evidence that the lungs produce approximately 50% of total platelets—about 10 million per hour—raises the question of whether pulmonary congestion in heart failure could impair platelet biosynthesis. This study aimed to disentangle the relationships between four key heart failure indicators (LVEDP representing intracardiac pressure, CI representing blood flow, plasma noradrenaline representing neurohumoral activation, and eGFR representing renal function) and three hemogram parameters (WBC, Hb, platelets) using structural equation modeling to minimize confounding.
**Methods:** A total of 345 patients admitted to The Jikei University Hospital who underwent left heart catheterization with Swan-Ganz catheterization for heart failure evaluation between 1 January 2015 and 31 December 2020 were included. Exclusion criteria included acute coronary syndrome, acute heart failure with Swan-Ganz catheterization only, dialysis-dependent chronic renal failure, contrast allergy or steroid pretreatment, hematological diseases, and pheochromocytoma. Hemodynamic indices (LVEDP, CI) were measured during catheterization, and blood samples collected simultaneously were analyzed for plasma noradrenaline concentration, WBC count, Hb level, platelet count, and eGFR. Statistical analyses included simple and multiple regression analyses, followed by structural equation modeling using IBM SPSS AMOS Version 25. Bayesian inference was additionally performed to re-evaluate SEM results, with posterior distributions visualized as two-dimensional contour plots where black represented 95%, dark gray 90%, and light gray 50% confidence intervals.
**Key Results:** The cohort had a mean age of 66.0 ± 13.2 years, 73% male, mean LVEF 45.1 ± 15.1%, mean LVEDP 14.2 ± 6.9 mmHg, mean CI 2.77 ± 0.76 L/min/m², mean plasma noradrenaline 371.1 ± 217.4 pg/mL, and mean eGFR 59.6 ± 19.5 mL/min/1.73 m². In simple regression analyses, plasma noradrenaline showed significant positive relationships with WBC (P = 0.003) and platelet count (P = 0.001), and a negative relationship with Hb (P = 0.038). CI had a significant negative relationship with Hb (P < 0.001). eGFR had a significant positive relationship with Hb (P < 0.001). LVEDP had a significant negative relationship with platelet count (P = 0.020). Multiple regression analyses confirmed these findings: noradrenaline positively associated with WBC (standardized coefficient 0.165, P = 0.003) and platelets (0.198, P < 0.001), and negatively with Hb (−0.091, P < 0.001); CI negatively associated with Hb (−0.262, P < 0.001); eGFR positively associated with Hb (0.271, P < 0.001); LVEDP negatively associated with platelets (−0.129, P = 0.016). SEM results were largely consistent: LVEDP showed a significant negative relationship with platelet count (standardized estimate −0.129, P = 0.015); CI showed a significant negative relationship with Hb (standardized estimate −0.263, P < 0.001); noradrenaline showed significant positive relationships with WBC (standardized estimate 0.165, P = 0.003) and platelet count (standardized estimate 0.198, P < 0.001); eGFR showed a significant positive relationship with Hb (standardized estimate 0.274, P < 0.001). Notably, in SEM, the relationship between noradrenaline and Hb did not reach significance (P = 0.079). Bayesian inference confirmed that LVEDP had a strong negative effect on platelet count (posterior distribution entirely in the negative range), while its effects on Hb and WBC straddled zero, indicating weak or absent relationships.
**Clinical Implications:** This study provides the first clinical evidence linking elevated LVEDP specifically to decreased platelet count in heart failure patients, independent of other hemodynamic and neurohumoral factors. The specificity of this relationship—LVEDP did not correlate with WBC or Hb—supports the hypothesis that pulmonary congestion mechanically interferes with platelet biosynthesis in the lungs. If confirmed, platelet count could serve as a simple, readily available clinical marker of pulmonary congestion severity. The finding that lower CI was associated with higher Hb suggests a compensatory erythropoietin-mediated response to tissue hypoxia, potentially involving HIF pathway activation. The positive association between noradrenaline and both WBC and platelet counts reflects sympathetic nervous system effects on hematopoiesis, likely mediated through β2- and α-adrenergic receptors respectively. The positive eGFR-Hb relationship is consistent with known mechanisms of renal anemia. Limitations include the cross-sectional design (precluding causal inference), the single-center setting, potential unmeasured confounders (e.g., thrombopoietin, liver congestion), and the inherent subjectivity in path diagram construction for SEM. Future research should investigate the mechanistic link between pulmonary capillary hemodynamics and platelet production, and explore whether platelet count monitoring can guide decongestive therapy in heart failure.