In the end, what clinical evidence supports IV albumin administration during severe sepsis or septic shock? The ALBIOS trial was specifically designed to verify whether the correction of hypoalbuminemia (ie, albumin replacement) could be beneficial during severe sepsis or septic shock. In the overall population, we observed no effect on survival. Nonetheless, we did observe improvements in hemodynamics during the early phase of treatment, with possible clinical relevance; these improvements included a slightly higher mean arterial pressure, a lower need for vasoactive drugs, and a lower daily net fluid balance as associated with albumin administration. Moreover, in a post hoc analysis of the subgroup of patients with septic shock, a significant 6.3% absolute reduction in 90-day mortality was observed in the albumin group, compared with the crystalloid group, suggesting a specific beneficial effect of albumin in this category of patients. Certainly, such an analysis is limited because it was not predefined. Methodologically, such statements hold true; clinically, however, the presence of shock within the general context of severe sepsis defines a well-characterized subset of patients, with a specific pathophysiology, clinical manifestation, and severity of the septic syndrome. Being a subgroup analysis, even if predefined, the study would have been underpowered to test such a hypothesis. Therefore, from a purely statistical point of view, a non-predefined analysis would not differ from a predefined analysis. The real concern should be whether such findings pertain to a data-dredging analysis; or, from a different perspective, whether there is a plausible biological and pathophysiologic rationale for accepting this hypothesis. Based on what I have discussed earlier, I think that the answer may be simple.