We evaluated the ability of perfusion index (PI) to predict vasopressor requirement during early resuscitation in patients with severe sepsis. All consecutive patients with clinically suspected severe sepsis as defined by the criteria of the American College of Chest Physicians/Society of Critical Care Medicine Consensus Conference were included. Perfusion variables included PI, arterial lactate level, central venous oxygen saturation, and the difference between central venous carbon dioxide and arterial carbon dioxide pressures, and were recorded before resuscitation and 6 h thereafter. We enrolled 36 patients with severe sepsis. Twenty-one patients required vasopressors, whereas 15 did not. The cut-off of the PI value for predicting vasopressor requirement was ≤0.3. This cut-off value had a sensitivity of 100% and a specificity of 93%; the area under the curve was 0.96 (95% confidence interval 0.8-0.99, P < 0.0001). The cut-off of the arterial lactate level for predicting vasopressor requirement was ≥1.8 mg dL. This cut-off value had a sensitivity of 82% and a specificity of 80%; the area under the curve was 0.84 (95% confidence interval 0.68-0.94, P < 0.0001). Other perfusion variables failed to predict vasopressor requirement in patients with severe sepsis. We concluded that PI and arterial lactate level are good predictors of vasopressor requirement during early resuscitation in patients with severe sepsis. Further studies are warranted to investigate whether monitoring PI during resuscitation improves the outcome of patients with septic shock.