Fig. 2: Exposures indicating or impacting VZV reactivation are associated with a modified risk of dementia.
From: Varicella-zoster virus reactivation and the risk of dementia
a, Experiencing two or more episodes of HZ is associated with a substantially increased dementia risk (cumulative hazard, y axis) up to 9 years after the second HZ event (x axis), compared to experiencing a single episode of HZ. b,c, Suppression of VZV reactivation through exposure to the (b) ZVL and (c) RZV is associated with a substantially reduced risk of dementia compared with exposure to the elective PPSV23 used as control at 3 and 5 years postindex date (PPSV23 is used by a similar population as HZ vaccines29 but does not confer protection against HZ). Results presented here are consistent with results obtained using IPTW and OW (Supplementary Figs. 7 and 8 and Supplementary Tables 21 and 22); see Methods for more details. The postmatching cohort characteristics for comparisons shown in this figure are presented in Supplementary Table 23. The curves show the Nelson–Aalen estimates of the cumulative hazard function (y axis) over the follow-up period (x axis) with a 95% CI band (shaded areas around each curve) for each cohort being compared. Cumulative hazards at specific time points are compared between the cohorts based on RR using a two-sided chi-squared statistic with 1 degree of freedom without adjusting for multiple comparisons. Superscript letter ‘a’ indicates censored, cumulative number of individuals who were lost to follow-up or had died until each time point; superscript letter ‘b’ indicates cumulative numbers of individuals with the event of interest (dementia diagnosis) until each time point; those numbers are to be interpreted relative to the cohort size under observation as indicated by ‘At risk’. NMD, normalized mean difference; RR, relative risk.
