Using the Bayesian logistic regression model to determine the relationship of demographics and hyperaldosteronism

Aldosterone excess (Hyperaldosteronism) has been reported to be a common cause of resistant hypertension. High aldosterone is also linked to sleep apnea. Investigators have reported that primary aldosteronism (PA) has a prevalence of approximately 20% in patients with resistant hypertension. These studies have been based largely on office blood pressure (BP) measurements. Thus, the degree to which hyperaldosteronism contributes to resistance of 24-h blood pressure control is not known. Thus, an ambulatory investigation appeared to be in order. Two hundred fifty-one patients were referred to the University of Alabama at Birmingham (UAB) Hypertension Clinic for resistant hypertension and were enrolled over a 46-month period. Resistant hypertension was defined as uncontrolled hypertension (>140/90 mm Hg) as determined at 2 or more clinic visits, in spite of use of 3 or more antihypertensive medications at pharmacologically effective doses. Secondary causes of hypertension other than aldosteronism, such as renovascular hypertension, pheochromocytoma or Cushing's syndrome, had been excluded by laboratory analysis and/or radiological imaging as clinically indicated. Subjects with a history of congestive heart failure or chronic kidney disease (creatinine clearance <60 ml/min) were excluded from study participation. In the reporting of the results the investigators noted that there was no difference in mean office BP between aldosterone excess (H-Aldo) subjects and normal aldosterone status (N-Aldo) patients. Daytime, night-time, and 24-h systolic and diastolic BP were significantly higher in H-Aldo compared to N-Aldo males. Daytime, night-time, and 24-h systolic BP were significantly higher in H-Aldo compared to N-Aldo females. Multivariate analysis indicated a significant interaction between age and aldosterone status such that the effects of aldosterone on ambulatory BP levels were more pronounced with increasing age. The purpose of the present research is to pursue the secondary finding of demographics and its relationship to aldosterone excess using a Bayesian logistic model. Applying a Monte Carlo Markov Chain approach, the authors focused on both age and gender as possible correlates of hyperaldosteronism. Using normal priors for the age and gender parameters we note that both younger age (posterior OR=1.04) and males (posterior OR= 4.33) are more likely to experience aldosterone excess. The 95% posterior credible intervals did not contain unity. Posterior density profiles also indicated marked departure from posterior expected unit odds ratios. We demonstrate how the model performs under relevant clinical conditions. The conditions are all tested using a Bayesian statistical approach allowing for the robust testing of the model parameters under various stress conditions which we introduce into the model. The convergence of the parameters to stable values are seen in trace plots which follow the convergence patterns This allows for precise estimation for determining clinical conditions under which the logistic pattern will change. We give further numerical and graphical examples of our results.