In reply

We would like to thank Drs Mullins and Theodoro for their comments and interest in on our recent article.¹ There are several statements, however, that need to be addressed.

The authors question the internal validity of the study design. Internal validity errors concern instances in which a study design fails to address the narrowest of research questions. We feel the basic study question was answered and clearly stated in the introduction.

As to 鈥渙bvious selection bias鈥� of studying only those with suspected AI, we would agree and this was addressed in our study limitations. However, we were not evaluating cosyntropin unresponsiveness but rather the presence of clinically relevant AI that warranted intervention. In fact, cosyntropin unresponsiveness (in the absence of hemodynamic issues) has been shown not to correlate with outcome.² Therefore, it appears relevant that we selected those patients in whom AI was suspected (those who will actually require treatment for AI).