SAT-544 Primary Aldosteronism in Intracerebral Hemorrhage. Not Intracerebral Hemorrhage in Primary Aldosteronism

Background: Patients with primary aldosteronism (PA) are more prone to cardiovascular complications including intracerebral hemorrhage (ICH) than those with essential hypertension. But how about PA in ICH patients - not ICH in PA patients? Since ICH patients with PA are at high risk for recurrent hemorrhage and other complications, diagnosis of PA is important even after the occurrence of ICH. Here we aimed to clarify the clinical features of PA in ICH patients at the largest scale ever and to evaluate the efficacy of PA screening with machine learning. Methods: Out of 181 ICH patients admitted to our hospital between June 2016 and February 2017, 126 patients of hypertensive ICH were enrolled in this study. Plasma aldosterone concentration (PAC), plasma renin activity (PRA) and other hormones were measured in the morning two days after admission. Results: After leaving out those who had been taking medications which could intervene with PAC or PRA on admission, nine patients were positive for PA screening (PAC/PRA ratio >200 and PAC >120 pg/mL) and 53 were negative. Age (68.6 vs 68.5 y), sex (male 66.7 vs 69.8 %) and blood pressure (172/97 vs 177/100 mmHg) were similar between these two groups. Bleeding volume (14.6 vs 16.2 mL) was also similar, but the lesion was more common in thalamus rather than putamen in PA positive group. Serum potassium was slightly lower in positive group (3.6 vs 3.9 mmol/L; P=0.096) on admission, and the difference became more evident two days later (3.7 vs 4.0 mmol/L; P=0.040). There were no differences in other hormones including cortisol and catecholamine. As for prognosis, PA positive patients had more severe motor or cognitive impairments. Dimension reduction procedure using t-SNE certainly divided these patients into clusters compatible with PA screening tests. Further, we extrapolated this result to 21 patients who had been taking medications which could lower PAC/PRA ratio such as ACE inhibitors, ARB or diuretics and were excluded from the analysis above. K-nearest neighbor method revealed that even in those who had been taking PAC/PRA lowering medications, PAC/PRA ratio >160 could be regarded as positive for PA screening. Discussion: This is the largest study ever that investigated the clinical features of PA in ICH patients. Contrary to expectations, ICH patients with PA were not necessarily younger than those with essential hypertension. But they were more likely to have severe outcomes even though blood pressure and bleeding volume were similar. This may be partly because of higher aldosterone. The difference in serum potassium was masked on admission probably due to increased sympathetic activity. But about two days later, when its activity peaked out, lower serum potassium in PA positive group became more evident. This can support the efficacy of PA screening tests even in ICH patients, so PA screening should not be awaited just because they have developed ICH.