Early valve surgery for infective endocarditis, that is surgery performed during the course of antibiotic therapy, is widely used. However, this intervention has never been proven effective in any clinical trial. Five recently published observational studies, using propensity analysis, analyzed the relationship between surgery and mortality. But they gave conflicting results, mainly because of their methodological limitations. Thus, to date, there is no reliable estimation of early valvular surgery impact on mortality. In addition, this example allows for discussing the interest of propensity analyses and to emphasizing their limitations.