Objective. To study the haemodynamic effects of a hypertonic saline/dextran solution compared with a normal saline solution in patients with severe sepsis. Design. Prospective double blind and control-randomised study. Setting. Adult intensive care unit in a university hospital. Patients. Twenty-nine patients with sepsis with a pulmonary artery occlusion pressure (PAOP) lower than 12 mmHg. Interventions. Patients were randomised to receive 250 ml of blinded solutions of either normal saline (SS group, n=16) or hypertonic saline (NaCl 7.5%)/dextran 70 8% (HSS group, n=13) solutions. Measurements and results. Haemodynamic, blood gas, and sodium data were collected at the following time points: baseline, 30 min, 60 min, 120 min, and 180 min. PAOP was higher in the HSS group at 30 min (10.7±3.2 mmHg vs 6.8±3.2 mmHg) and 60 min (10.3±3 mmHg vs 7.4±2.9 mmHg); P–1 m–2 vs 3.8±3.4 l min–1 m–2), 60 min (4.9±4.5 l min–1 m–2 vs 3.7±3.3 l min–1 m–2), and 120 min (5.0±4.3 l min–1 m–2 vs 4.1±3.4 l min–1 m–2); P–2 vs 35.6(31.2–49.2) ml m–2] and 60 min [46.8(39.7–56.6) ml m–2 vs 33.9(32.2–47.7) ml m–2]; P–1 cm–5 m–2 vs 1139±245 dyne s–1 cm–5 m–2), 60 min (921±256 dyne s–1 cm–5 m–2 vs 1246±308 dyne s–1 cm–5 m–2), and 120 min (925±226 dyne s–1 cm–5 m–2 vs 1269±494 dyne s–1 cm–5 m–2). Sodium levels increased in the HSS group (P=0.056) and were higher than in the SS group at 30 min (145±3 mEq l–1 vs 137±7 mEq l–1), 60 min (143±4 mEq l–1 vs 136±7 mEq l–1), 120 (142±5 mEq l–1vs 136±7 mEq l–1), and 180 min (142±5 mEq l–1 vs 136±8 mEq l–1). Conclusion. Hypertonic saline/dextran solution may improve cardiovascular performance in severe sepsis without significant side effects. The haemodynamic effect appears related mainly to a volume effect.