Sara van den Berg

192 Chapter 7 individuals. D,E . Relationships are shown respectively between frailty and CMV-specific antibody levels at endpoint ( D ) and between frailty and fold change in CMV-specific antibody levels over 26 years ( E ). F, G. Difference in increase in frailty index of CMV- participants and those who seroconverted recently after measurement point T5 ( F ) and in recently converted women aged 60-65 years ( G ). Increase in frailty index: difference between frailty index as measured between T5 and T6. H, I. Comparison of cardiovascular disease (CVD) prevalence with CMV- specific antibody levels at endpoint ( H ) and fold change in 25 years ( I ) in these antibody levels (Prevalence of CVD is indicated in table 2). Symbols: Triangles represent women, circles represent men. Colors of symbols represent CMV-serostatus. Green: CMV-, blue: ST CMV+, purple: LT CMV+. As the effects of CMV might be related to specific chronic conditions and not to a general health state such as frailty we investigated the association with cardiovascular disease [45, 60]. Indeed, the prevalence of CVD was significantly higher in CMV-seropositive individuals. Furthermore, within CMV+ individuals CVD was associated with higher CMV-specific antibody levels at end point. A study in an “unhealthy” population above 90 years of age also reported a relationship between high CMV-specific antibody levels and the prevalence of CVD [13]. One of the mechanisms that could explain this association is that the lytic viral CMV lifecycle in endothelial cells induces vascular damage and contributes to CVD [61, 62] and in particular to atherosclerosis [63]. Alternatively, progressive endothelial damage in individuals with CVD and a pro-inflammatory environment could initiate inflammation leading to CMV reactivation [64]. We show that, age-related effects other than duration of CMV infection seem to contribute to the variation in CMV-specific immune responses at endpoint between individuals. We hypothesize that this variation between individuals is due to individual differences in the balance between virus and host factors and that these immune responses may in part be determined by viral reactivation, but for the larger part by the establishment of the latent CMV reservoir after primary infection ( Figure 6 ) . Due to immunosenescence at higher age, older individuals may be less able to control primary infection and viral spread. This may lead to enhanced activation of the immune system to enable control of (latent) CMV infection, regardless of duration of CMV infection. Other factors that could influence this balance include the route of viral transmission [65] and the viral inoculum during primary infection [24] as well as the state of the immune system at primary infection [53, 66]. A comparable model was previously proposed based on data from children, which suggested that the relatively immature state of the immune system in children leads to the establishment of a larger viral reservoir [53, 66], which may subsequently increase the chance of viral reactivation. The state of the immune system at primary CMV infection might thus affect the levels of life-long immunity to CMV ( Figure 6 ). Whether general health status, i.e. frailty, also enhances the establishment of the latent CMV reservoir and subsequent viral reactivation, remains to be investigated. In conclusion, our results indicate that age, regardless of duration of CMV infection, has a larger influence on the CMV-specific response than previously anticipated. High CMV- specific antibody levels should therefore not be interpreted as a measure of experienced viral reactivation or duration of CMV infection.