Reevaluating the Use of the Nationwide Inpatient Sample to Identify Incident Cases of Atrial Fibrillation After Aortic Valve Replacement—Reply

In Reply We thank Malik et al for their interest in our article¹ and their comments. We agree with the authors regarding the difficulty of differentiating between incident and prevalent atrial fibrillation (AF) in the Nationwide Inpatient Sample (NIS) database. We noted this in the article’s Limitations section.¹ Furthermore, we also stated in our Methods section¹ that the reported rates of AF from the NIS assumed that all secondary discharge field disease coding was for new diagnoses of AF. We realized this assumption or approach may not be completely accurate, but it has been previously published using the NIS database.^2,3 We were cognizant of the limitations with the NIS database that Malik et al highlighted, so we attempted to combat them using the additional analyses involving a validation cohort from the New York state inpatient database. This database has a unique patient identifier, and prior data have confirmed correct estimates of incident AF.⁴