T. Nicholas,1 B. Binneman2; 1Pfizer, Groton, CT, 2Pfizer, Cambridge, MA

BACKGROUND: The Hamilton Anxiety Ratings Scale (HAM-A) has been used to study Generalized anxiety disorder (GAD) for many years. A myriad of drug classes have been approved for treatment including alpha-2-delta (a2d), benzodiazepine (bz), and SNRI/SSRI pharmacologies. The time course of effect has not been well characterized.
METHODS: Study level data were combined from both internal and external sources to assess the longitudinal nature of HAM-A in clinical trials for GAD. The final data set was comprised of 279 records over 22 studies, in which there were 15 treatments tested in either a monotherapy or adjunct therapy. Data were analyzed using NONMEM 7.3 and R 3.02.
RESULTS: The structural model consisted of a placebo component and an active treatment component. The placebo and active treatment were considered additive. The placebo effect was characterized by an exponential rate constant to a maximal effect (PBOmax). Between study variability was explored on the PBOmax. The effects of baseline HAM-A and mono vs. adjunctive therapy were tested on PBOmax. Baseline was found to be significant in this model, while adjunctive therapy was not. Treatment effect was tested as both an offset to the placebo and also with an exponential time component (if the data warranted). Baseline and type (mono vs adjunct therapy) were assessed as covariates on maximal drug effect (Emax). Both were found to be significant.
CONCLUSION: The placebo time course is moderately well estimated. Time to ~90% of placebo effect stead-state is approximately 6.5 weeks. The bz and a2d classes (and other non-SNRI) were found to reach steady state effect very quickly (consistent with prior analyses). Given the additive nature between placebo and treatment, the placebo adjusted effect will be similar whether it is a 4-, 6-, or 8-week study.