Hypertension resolution after adrenalectomy for primary hyperaldosteronism: Which is the best predictive model?

Background: We aimed to compare the predictive performance of three distinct clinical models purported to predict the resolution of aldosteronoma-associated hypertension after adrenalectomy. Methods: A tri-institutional database of aldosteronoma patients who underwent adrenalectomy between 2004 and 2019 was retrospectively reviewed. The three models of interest incorporate various preoperative clinical factors, such as age and sex. The predictive accuracy, as measured by area under the curve of receiver operator characteristic, was estimated. Receiver operator characteristic was evaluated across the whole cohort, then stratified by treatment location. Results: A total of 200 patients were included (91 American, 109 French). The clinicodemographic variables between groups were similar; the French cohort had a lower mean body mass index (P = .02). The overall complete clinical resolution of hypertension after adrenalectomy for the entire data set was 45.5% (n = 91). The regression coefficients in the Utsumi et al (2014) Japanese model produced a superior overall area under the curve (0.78, 95% confidence interval [CI] [0.71-0.84]). This model also performed best when the cohort was stratified by treatment location (French area under the curve = 0.74, 95% CI [0.64-0.83], US area under the curve = 0.82, 95% CI [0.72-0.91]). Conclusion: When comparing three predictive models of aldosteronoma-associated hypertension resolution after adrenalectomy, the Utsumi et al model demonstrated the highest predictive validity across all cohorts. Counseling based on this model regarding probability of cure is recommended. (C) 2020 Elsevier Inc. All rights reserved.

Automated summary

The study compared three clinical models for predicting resolution of aldosteronoma-associated hypertension after adrenalectomy, finding that the Utsumi et al model demonstrated the highest predictive validity across all cohorts and treatment locations.