The authors expressed concerns that the study did not account for certain key covariates, such as septic shock, right heart failure, or hypoxemia, which are believed to influence the development of AKI.

We would like to clarify that sepsis, hypoxemia and vasopressors and inotropes use were included in the machine learning model.

Even further, we introduced a time-dependent analysis to account for covariates trajectories effect on mortality.

In other words, in addition to the baseline covariates, we took into account the changes over time in clinical severity of the syndrome (e.g., PaO2/FiO2 ratio, dead space, driving pressure, positive end expiratory pressure), and in the interventions undertaken to mitigate shock states (i.e., fluid balance and vasopressors use), see electronic supplementary material of the original article.

Read More