68-2

36/110

128Figure 6　Results of the receiver operating characteristic (ROC) analysis assessing the association between the numerical rating scale (NRS) pain score at Week 2-3 and the development of postherpetic neuralgia (PHN)The area under curve (AUC) and the cutoff point were 0.840 (p<0.00001) and 0.9 (sensitivity 80.0%, specificity 80.0%) for the NRS pain score at Week 2-3 predicting an increased risk for PHN.therapy later than 3 days both before and after PSM.Generally, the early initiation of antiviral therapy within 72 hours, and hopefully, within 48 hours, has been recommended for achieving better clinical outcomes, including rapider pain relief 5, 6, 10, 11, 16). Indeed, we found that compared with late visitors, early visitors tended to achieve better pain relief at Day 7 before PSM, and that they achieved signifi-cantly better pain relief at Day 7 after PSM. Because the early initiation of antiherpetic agent therapy thus could more effectively relieve acute HZ pain, FCV might exert pain-relieving effects as effective as AMNV in early visitors receiving antiherpetic agent therapy within 3 days after onset of HZ.On the other hand, AMNV was more effective than FCV in reducing acute HZ pain at Day 7 and Week 2-3 in late visitors receiving the therapy later than 3 days, both before and after PSM. FCV is converted by the liver enzyme to active penci-clovir that inhibits viral DNA polymerase in cells infected by the virus2, 17 18), whereas AMNV directly suppresses viral growth by inhibiting the activity of the helicase-primase complex required for cleavage of double strand DNA and synthesis of RNA primers, which is the initial stage of viral replication19, 20). Considering such pharmacological mechanisms of action quite different between AMNV and FCV, it seems plausible that AMNV has a faster onset of action and a more potent anti-viral effect6, 21), which might explain why AMNV exerted pain-relieving effects superior to FCV in late visitors in our study.Another major finding of this study was that pain intensities, not in the early phase of acute HZ at Day 0 or Day 3-4, but in the late phase at Day 7 and Week 2-3, especially at Week 2-3, were signifi-cantly associated with the development of PHN. Patients with any persisting pain at Week 2-3 was associated with a marked increase in a risk for PHN development, as indicated by a high odds ratio of 25.6. Many previous studies showed signifi-cant associations between initial pain intensities upon enrollment in studies and the development of PHN22-24). However, we could not find such associa-tions between initial pain intensities and PHN. Such discrepancies might result from different study designs, such as prospective randomized controlled studies vs. a retrospective observational study, and/or the enrollment of early visitors alone vs. DiscussionIt has been reported that the elderly over 60 years of age, female sex, and presences of a prodrome, severe rashes, severe acute pain, and/or immuno-compromised states are independent predictors of PHN1), 12-15). Because severe acute pain and/or rashes predispose to PHN, earlier and better controls of acute pain and/or rashes achieved with potent antiherpetic agents may help to reduce the inci-dence of PHN16). Based on this concept, we conducted this study to evaluate whether AMNV is superior to FCV in terms of acute pain control as primary clinical outcomes and prevention of PHN as secondary clinical outcomes.In the present study, we found that compared to FCV, AMNV was more effective in reducing pain at Day 7 and Week 2-3 in the total cohort both before and after PSM. By performing subgroup analyses, we found that the pain-relieving effects of AMNV and FCV were not different in early visi-tors receiving antiherpetic agent therapy within 3 days both before and after PSM, whereas AMNV was more effective in reducing acute HZ pain at Day 7 and Week 2-3 in late visitors receiving the