Real-life use of vasopressors and inotropes in cardiogenic shock—observation is necessarily ‘theory-laden’

We commend the attempt of Tarvasmaki et al. [1] to identify the mortality risk associated with individual vaso-active agents used in cardiogenic shock. However, despite their rigorous statistical analysis, we recommend caution in interpreting these results. Propensity score matching accounts for prior bias in the choice of vaso-active agents, but this choice of agent is often deeply engrained in individual clinician dogma which even propensity score matching may not uncover.