Sara van den Berg

190 Chapter 7 inflation of CMV-specific T-cells in humans was questioned recently [55]. Longitudinal studies in humans are very limited, and some report evidence for memory inflation [56, 57], while others do not [58, 59]. Our study allowed us to investigate how duration of CMV infection influences the T-cell memory pool in humans. We show that large polyfunctional CMV-specific T-cell responses and memory/effector CD4+ and CD8+ T-cell populations are already high shortly after CMV seroconversion. Thus, both the humoral immune response and the CD4+ and CD8+ T-cell pools do not require a long duration of CMV infection to develop and expand. We found weak evidence for memory inflation in the CD4+ T-cell pool, as duration of CMV infection correlated positively with CD4+ T EM and T EMRA numbers in short- term converters, but not in LT CMV+ individuals. Regardless of duration of CMV infection, we found a positive correlation between CMV seroconversion at older age and CD4+ T EM and T EMRA cell numbers, similar to what we observed for CMV-specific antibody levels. This may be explained either by the aged immune system requiring after delayed primary control a high number of CD4+ T EM and T EMRA cells to control latent CMV infection, or by enhanced stimulation of cells due to limited control of the virus. Taken together, our data contribute to the view that there is no evidence for a time dependent memory inflation during CMV infection in humans. Further longitudinal studies, possibly covering even longer periods of time might strengthen our results. We also assessed whether prolonged CMV infection might impair clinically relevant health outcomes. We did not find an association between frailty and CMV seroprevalence, or between frailty and duration of CMV infection, with frailty expressed by a frailty index score [34]. This is in line with some other papers [30], although several studies show a relation between CMV infection or CMV-antibody levels and frailty [11, 28, 29] or even the opposite relationship with a higher CMV seroprevalence in healthy people [31]. The lack of a relationship between frailty and CMV-seroprevalence or CMV-specific antibody levels in our study could be due to the heterogeneity between individuals with a high frailty index score, since various conditions and deficits (n=36) are included in the frailty score we used based on Rockwood criteria. Importantly, another study that used a Rockwood-based frailty index score also did not observe an effect of CMV-seropositivity on frailty [30]. Moreover, our study was performed in a relative younger population compared to the two other studies not showing a relation between CMV and frailty in people above 80 years of age that might be obscured by a survival bias. Also, the frail population in our study was overrepresented due to the stratified selection, which generally increases the chance of finding an association with frailty, emphasizing the negative relation we found. In sum, our study does not support the hypothesis that CMV causes ‘accelerated aging’ influencing general health. However, we found an association between frailty and age at first seropositive time point during follow up within short-term CMV+ individuals. In addition, a small sub-group of women who seroconverted recently (after round 5, n=3) had a steeper increase in frailty index than those who did not seroconvert that might suggest that people who become frail are more susceptible to CMV infection.