There has been great interest recently in better understanding how an episode of acute kidney injury (AKI) affects risk of development or acceleration of chronic kidney disease. This area of epidemiology research presents several methodological challenges that have not been sufficiently discussed in the literature. These are related to the current consensus definitions of AKI; the determination of 'baseline' renal function before the AKI episode; and the possibility that observed associations between AKI and future adverse events are confounded by differences in the severity of baseline chronic kidney disease. In this study, we discuss several potential solutions to these problems. Concerted efforts to address these methodological issues will propel research in this field to a higher level.