Invasive left atrial pressure at transseptal puncture correlates with proBNP but adds limited prognostic information

Background

Left atrial pressure (LAP) reflects LV filling pressure and pulmonary venous congestion. Although LAP is routinely recorded during transseptal puncture (TSP) for pulmonary vein isolation (PVI), its value for risk stratification and phenotyping in atrial fibrillation (AF) remains insufficiently defined.

Methods

We performed a two-center retrospective analysis of single shot pulmonary vein isolation procedures under moderate sedation. LAP was measured immediately after TSP. All patients were eligible irrespective of age, sex or comorbidities. Prior left atrial ablation procedures were an exclusion criterion. We summarized clinical, echocardiographic, and laboratory data and related mean LAP (LAPm) to covariates using Spearman rank correlations. A multivariable linear regression (complete-case subset) assessed independent associations of LAPm with AF recurrence (yes/no), proBNP, presenting rhythm, and AF duration.

Results

We included 546 patients (48% female; age 66.7 ± 11.4 years; BMI 28.5 ± 4.7 kg/m²; proBNP 1034 ± 1366 pg/mL; LVEF 55.8 ± 8.7%; LAVI 41.9 ± 12.7 mL/m²; left atrial area 24.8 ± 5.4 cm²). AF type was paroxysmal in 41% and persistent in 59%. Mean LAPm was 10.2 ± 6.2 mmHg, and 15% had LAPm > 15 mmHg (left atrial hypertension). At transseptal puncture, 46% were in sinus rhythm and 54% in AF. Comorbidities included coronary artery disease in 23.5%, hypertension in 75.2%, diabetes in 13.3%, and prior stroke/TIA in 8% of patients. The median CHA2DS2-VASc score was 3.

Spearman analysis showed a moderate correlation between LAPm and proBNP (ρ = 0.497, p < 0.001). Correlations were weaker for presenting rhythm (ρ = 0.201, p < 0.001) and AF duration (ρ = 0.198, p < 0.001), and small for AF recurrence (ρ = 0.118, p = 0.006), CHA2DS2-VASc (ρ = 0.119, p = 0.005), and PR interval (ρ = 0.145, p = 0.016). No significant correlations were observed with hypertension, diabetes, prior stroke/TIA, renal function, LVEF, or LAVI.

In the multivariable model (complete-case subset), overall fit was significant (F(4,110) = 5.59, p < 0.001; R² = 0.169; adjusted R² = 0.139). proBNP remained positively associated with LAPm (B = 0.001, SE < 0.001, β = 0.250, p = 0.015). AF recurrence (B = −0.082, SE = 0.783, β = −0.010, p = 0.917) and AF duration (B = −0.348, SE = 1.131, β = −0.042, p = 0.759) were not associated. Presenting rhythm showed a positive, non-significant trend (B = 2.127, SE = 1.137, β = 0.259, p = 0.064).

Conclusions

In patients undergoing first-time PVI, LAPm correlates most strongly with proBNP whereas AF recurrence and AF duration show no independent association. In this predominantly non-severely impaired cohort, the incremental prognostic value of invasively measured LAP for outcome prediction appears limited. Biomarker-integrated or longitudinal models may be required to enhance risk stratification.Study Population