Mathematical coupling and the association between Kt/V and PCRn

As a statistician relatively new to the field of nephrology, I first met Fr ank Gotch at a series of Steering Committee meetings that were held to plan the pilot study that eventually led to the full-scale NIH HEMO trial. Duri ng the evenings, the discussion inevitably turned to kinetic modeling. Fran k would make elaborate presentations on subjects such as the prescription o f dialysis dose using double-pool kinetic models or the relative merits of blood-side versus dialysate-side kinetics. Initially I found most of this t o be incomprehensible, and must admit that I began to fear that my colleagu es and I at the data coordinating center had gotten ourselves into an esote ric endeavor that we had little hope of mastering. Fortunately Frank displayed great patience, and gradually the basic outline s of the role of urea kinetics in the study began to come into focus. It wa s some months later, during the intense period in which we developed the ki netic modeling programs for the study's database, that Frank's genius and d evotion became fully apparent. The HEMO study was the first large-scale att empt to prescribe dialysis dose using double-pool kinetics. Moreover, this methodology had to be implemented in a sufficiently practical and automated fashion to handle over 10,000 kinetic modeling sessions per year while mai ntaining delivered doses within the narrow target ranges of the study proto col. Time and again Frank displayed his depth of understanding both of the underlying mathematics and the practical issues of coordinating a large-sca le clinical trial. As delta began to come in from the early kinetic modelin g sessions bl the pilot study, Frank was routinely the first to discern imp lications for the study and work out the necessary adjustments in our proce dures. I will always be appreciative of Frank's patience as he directed me and others at the data center through the unfamiliar territory of urea kine tic modeling.