Frontiers in Endocrinology (Jan 2025)

A clinical decision model for failed adrenal vein sampling in primary aldosteronism

  • Sophie N. M. ter Haar,
  • Sofie J. van Goor,
  • Eleonora P. M. Corssmit,
  • Arian R. van Erkel,
  • Bartholomeus E. P. B. Ballieux,
  • Olaf M. Dekkers,
  • Olaf M. Dekkers,
  • Michiel F. Nijhoff

DOI
https://doi.org/10.3389/fendo.2024.1497787
Journal volume & issue
Vol. 15

Abstract

Read online

ObjectivePrimary aldosteronism (PA) is a common cause of secondary hypertension with unilateral and bilateral subtypes requiring different treatments. Adrenal vein sampling (AVS) is the gold standard for subtype differentiation but can be unsuccessful by challenging right adrenal vein anatomy. This study aimed to develop a clinical decision model using only measurements from the left adrenal vein (LAV) and peripheral blood (IVC) to differentiate between PA subtypes.MethodsThe retrospective cohort study included 54 PA patients who underwent bilaterally successful AVS. The main objective was to determine optimal cut-off values for the LAV/IVC index, using ROC analysis for subtype prediction. The predictive value of this index was assessed with the Area Under the Curve (AUC). The Youden index calculated cut-off values, targeting a specificity >90% for PA subtype differentiation.ResultsThe cohort, averaging 48.5 ± 9.5 years in age, comprised 21 women and 33 men, among whom 26 presented with unilateral and 28 with bilateral disease. LAV/IVC values <1.2 indicated unilateral right-sided disease (specificity 91%, sensitivity 96%, AUC 0.98, 95% confidence interval (CI) 0.95-1.0), values 1.2-2.4 suggested bilateral disease (sensitivity 93%, specificity 64%, AUC 0.85, CI 0.73-0.96), whereas values ≥4.4 predicted unilateral left-sided disease (specificity 93%, sensitivity 60%, AUC 0.85, CI 0.73-0.96). Published literature aligns with our results on cut-off values.ConclusionsUtilizing the LAV/IVC index, over 70% of unsuccessful AVS procedures due to failed right adrenal cannulation could be interpreted with over 90% certainty regarding the PA subtype, preventing unnecessary resampling and aiding in determining the preferred treatment.

Keywords