APPLICATION OF A NEW PHASE-SPECIFIC MODEL IN THE EVALUATION OF HERPES ZOSTER PAIN

H. Weiss (1), J. Gnann (1), M. Wood (2), P. Fiddian (3), J. Crooks (4), S.J. Soong (1), *R.J. Whitley (1)
(1) University of Alabama at Birmingham, Alabama, USA; (2) Birmingham Heartlands Hospital, England; (3) Private Consultant; (4) Glaxo Wellcome Research and Development, England

Pain associated with herpes zoster is the most debilitating complication of the disease. Three phases of pain are acute, sub-acute, and chronic, with the latter two phases comprising post-herpetic neuralgia (PHN). A phase-specific model was applied to assess pain in patients participating in trials of antiviral therapies for herpes zoster. Estimates of transition times (i.e. change-points) by phase were calculated. Hazard rates for each phase were estimated and compared to determine rates of pain resolution. Comparison of rates of pain resolution between treatment groups and other covariates were performed. 1,138 patients (>50 year old) receiving either Valaciclovir (VACV) or acyclovir (ACV) were evaluated. Estimates of change-points from acute to sub-acute and from sub-acute to the chronic phase were 24 days (+ 3) and 110 days (+ 12). Statistically significant differences in rates of pain resolution occurred from acute to chronic phase (p<0.001) and from the sub-acute to chronic phase (p<0.001). Treatment comparisons at each specific phase indicated a significant difference in the effect of VACV vs. ACV on pain resolution during the sub-acute phase (p=0.008) but not during the acute or chronic phases. A second cohort of 1,124 patients (<18 year old) receiving VACV showed similar estimates of transition times (25 and 116 days). Statistically significant differences in rates of pain resolution occurred across all three phases. Younger patients exhibited faster rates of pain resolution during the acute phase. The phase-specific analysis of herpes zoster pain provides an improved methodology for the analysis of zoster-associated pain. The model can be used in the planning of future zoster studies. For example, sample size power estimates indicate about 200-250 patients per treatment group are needed to detect a 50% difference in rates of pain resolution during each of the sub-acute and chronic phases.

Corresponding Author: R.J. Whitley, M.D., Professor of Pediatrics, Microbiology and Medicine, Department of Pediatrics, Children's Hospital, 1600 7th Ave. South, Suite 616, Birmingham, AL 35233, USA