Sara van den Berg

206 Chapter 8 to influenza virus in vitro in healthy individuals, and revealed either a negative association [22] or a positive association [23]. Taken together, we conclude that there is no unequivocal evidence that CMV infection leads to an impaired immune response to influenza, one of the most relevant and most frequently used models for heterologous infections in humans. As the CMV-enhanced ageing theory seems to persevere and influenza in older adults remains an important matter, more studies into CMV and the immune response to influenza will probably be performed. It remains important to interpret small-scale studies, especially with reports of only one influenza strain or one antibody outcome, such as the response rate, with caution. For future studies, we recommend to at least always report the geometric mean titer (GMT) pre- and post-vaccination (or the ratio, GM ratio) and confidence intervals, in line with European Medicines Agency-guidelines. This enables objective comparison between studies, and prevents to some extent influences of pre-existing immunity that may act as confounders. Furthermore, statistically adjusted analysis for the confounders age and pre- existing immunity are always needed. Beyond influenza The impact of CMV-enhanced ageing of the immune system on public health, and on the economy, may be most profound for the immune response to influenza, and influenza remains a leading cause of infectious morbidity, hospitalization and death in older adults. Also, the influenza vaccine is the number one vaccine recommended in Europe and the United States, next to the pertussis and varicella zoster vaccines, for older adults. Thus, clarifying the role of CMV on influenza infection and vaccination may help to identify possible vaccine optimization strategies that could have large impact. On the other hand, antibody responses to influenza vaccination are quite complicated to investigate. They constitute mostly a recall response, established by relatively frequent, natural exposure to the influenza virus and by possible previous seasonal vaccinations. Due to relatively large cross-reactivity among strains, also pre-existing immunity of other influenza virus types may influence the subsequent seasonal influenza vaccine response. Therefore, other models might be more suitable to test the principle of CMV-enhanced ageing. Beyond influenza, only a handful of studies have been performed in humans, with contradicting results. In a study in adult patients with antineutrophil cytoplasmic antibody–associated vasculitis, it was shown that expansion of CD4 + CD28 null T-cells, most likely induced by CMV reactivation, was associated with a reduced antibody response to pneumococcal vaccination [24]. In contrast, in healthy young children CMV infection was associated with an increased differentiation state of T-cells without hampering the CD8 + T-cell response to staphylococcal enterotoxin B (SEB) and the antibody response to measles[25]. In addition, in adults [26] and in our cohort of healthy to frail older adults ( chapter 7 , data not shown), the response to SEB was not negatively associated with CMV infection. Thus, results on heterologous challenges other than influenza, also indicate no clear evidence for a negative association of CMV infection and the immune response to a heterologous infection.