Abstract
Purpose:
For acceptance of dark adaptation as a clinical trial endpoint in studies of age-related macular degeneration (AMD), it is necessary to identify a single parameter that describes the speed of rod-mediated recovery. In addition, the parameter must be calculated using the same algorithm for every study participant. To fulfill these requirements, we developed the rod intercept. This poster describes the rod intercept algorithm and compares it to conventional nonlinear regression modeling.
Methods:
Representative dark adaptation curves were selected for the four dark adaptation phenotypes encountered in AMD studies. Dark adaptation speed was estimated using the rod intercept algorithm and multiple nonlinear regression models. The resulting parameters were compared and assessed against the requirements for a single summary parameter and a uniform methodology.
Results:
A case series of eight dark adaptation curves was analyzed, two each for the four dark adaptation phenotypes. The rod intercept algorithm successfully estimated dark adaption speed in all cases (8/8). None of the nonlinear regression models was able to estimate dark adaptation speed for more than two cases (2/8). To estimate dark adaptation speed for all eight cases using nonlinear regression, four different models were required. The common practice in general research of selectively deleting data segments to make a single model work for all dark adaptation phenotypes is considered data adulteration in clinical trials and is not acceptable.
Conclusions:
The rod intercept algorithm provides a single, uniform methodology for estimation of dark adaptation speed. This satisfies the need in clinical studies of AMD to use the same methodology to analyze the full range of dark adaptation phenotypes encountered. Furthermore, the rod intercept is robust to changes in the dark adaptation phenotype of an individual over time.
Keywords: 412 age-related macular degeneration •
468 clinical research methodology