The safety of non-selective -blockers (NSBB) has been questioned in refractory ascites (RA). We studied the effects of NSBB on cardiac systolic function, systemic hemodynamics, and renal perfusion pressure (RPP) and function in patients with diuretic-responsive (DRA) and RA.Prospective pre-post repeated-measures study in cirrhotic patients, 18 with DRA and 20 with RA on NSBB for variceal bleeding prophylaxis. The following were measured at baseline and 4-weeks after propranolol: systolic function by the ejection intraventricular pressure difference (EIVPD), hepatic venous pressure gradient (HVPG), cardiopulmonary pressures, RPP, and sympathetic activation.EIVPD was elevated at baseline (RA: 4.5 [2.8-5.7] and DRA: 4.2 [3.1-5.7] mmHg; normal: 2.4-3.6 mmHg) and directly related to the severity of vasodilation and sympathetic activation. NSBB led to similar reductions in heart rate and HVPG in both groups. NSBB reduced EIPVD in RA but not in DRA (-20% vs. -2%, p<0.01). In RA, the reduction in EIPVD by NSBB correlated with the severity of vasodilation and with higher plasma nitric oxide, norepinephrine and IL-6 (r>0.40, all p<0.05). NSBB reduced RPP in both groups, but impaired renal function only in RA. Reduction in EIPVD by NSBB correlated with decreases in RPP and eGFR (r>0.40, all p<0.01). After NSBB, RPP dropped below the threshold of renal flow autoregulation in 11 of the 20 (55%) RA patients, including the 4 fulfilling HRS-AKI criteria.Renal perfusion and function depend critically of systolic function and sympathetic hyperactivation in RA. NSBB blunt the sympathetic overdrive on cardiac function, hamper cardiac output, lower RPP below the critical threshold and impair renal function.-blockade should be cautious or even avoided in RA.