A prediction model for primary aldosteronism when the salt loading test is inconclusive

Abstract

Objective: To develop a prediction model to confirm or exclude primary aldosteronism (PA) in patients with an inconclusive salt loading test (SLT). Context: Diagnosis in patients with a suspicion of PA can be confirmed using an SLT. In case of inconclusive test results the decision about how to manage the patient is usually based on contextual clinical data. Design: We included a retrospective cohort of 276 patients in the final analysis. Methods: All patients underwent an SLT between 2005 and 2016 in our university medical center. The SLT was inconclusive (post-infusion aldosterone levels 140–280 pmol/L) in 115 patients. An expert panel then used contextual clinical data to diagnose PA in 45 of them. Together with 101 patients with a positive SLT this resulted in a total of 146 patients with PA. A total of 11 variables were used in a multivariable logistic regression analysis. We assessed internal validity by bootstrapping techniques. Results: The following variables were independently associated with PA: more intense potassium supplementation, lower plasma potassium concentration, lower plasma renin concentration before SLT and higher plasma aldosterone concentration after SLT. The resulting prediction model had a sensitivity of 84.4% and a specificity of 94.3% in patients with an inconclusive SLT. The positive and negative predictive values were 90.5 and 90.4%, respectively. Conclusions: We developed a prediction model for the diagnosis of PA in patients with an inconclusive SLT that results in a diagnosis that was in high agreement with that of an expert panel.

Keywords

• adrenal
• primary aldosteronism
• inconclusive salt loading test
• logistic regression
• prediction model