The Authors Reply.

The Authors Reply

We thank Drs. Matthieu and Jerome for their interest[1] in our study[2] and agree that the interpretation of our composite outcome should not be extrapolated to isolated AKI progression. A major goal of our study was to examine whether AKI biomarkers provide overall prognostic information in patients with AKI. We included death in our composite outcome based on its strong association with AKI[3,4] and our hypothesis that proximate effects of evolving renal dysfunction in a prevalent AKI population are more likely to contribute to mortality than be ‘unrelated’. We agree that in certain situations, including unrelated deaths may hamper prognostic accuracy for specific events. The data in our study do not support this concern. For deaths unrelated to kidney dysfunction to decrease power, biomarker concentration would need to be similar to the non-event group. In contrast, we observed that uL-FABP was significantly elevated for each component of our composite (Table 2).[2]

We appreciate the authors highlighting the potential utility of competing risk analysis to describe the association between biomarkers and discrete events. While potentially providing more granular information on renal-specific outcomes, this approach may cause loss of power if deaths are related to renal injury. Elevated urine biomarker concentrations in patients who died may also reflect severe renal injury that has not had the opportunity to manifest. We agree that competing risk analyses will be a useful additional tool in assessing prognostic accuracy in future studies,[5,6] particularly when the relationship between death and other outcomes of interest are less well-established.

On Behalf of the Authors,

Sharidan Parr