Abstract‘Immunosenescence’ is an imprecise term used to describe deleterious age‐associated changes to immune parameters observed in all mammals studied so far. Primarily anecdotal evidence implies that failing immunity is responsible for the increased incidence and severity of infectious disease in old people. However, there is a serious dearth of accurate hard data concerning the actual cause of death in the elderly and the contribution thereto of the multitude of age‐associated alterations measured in the immune system. Cross‐sectional studies comparing those currently young with those currently old reveal a large number of differences in the distribution of immune cell types in the blood, and to some extent the functional integrity of those cells. Many of these parameters differ markedly between individuals infected with CMV and uninfected people, regardless of infection with other persistent herpesviruses. The adaptive arm of immunity appears to be more seriously affected than the innate arm, particularly the T lymphocytes. However, cross‐sectional studies suffer the disadvantage that like is not being compared with like, because the conditions applied during the entire life course of the currently elderly were different from those applied now to the young. These differences in environment, nutrition, pathology and possibly genetics, rather than merely age, may be expected to influence the parameters studied. Moreover, pathogen exposure of the currently elderly was also different from contemporary exposure, probably including CMV. Some of the problems associated with cross‐sectional studies can be overcome by performing longitudinal studies, as pointed out in an earlier analysis of the Baltimore Longitudinal Ageing study looking at lymphocyte numbers. However, longitudinal studies are challenging in humans. Nonetheless, the pioneering Swedish OCTO/NONA studies of the very elderly which for the first time included a range of immune parameters, have identified a set of immune parameters predicting mortality at 2, 4 and 6 year follow‐up; CMV infection makes a material contribution to this so‐called ‘immune risk profile (IRP)’. Whether the IRP is informative in younger individuals and the mechanism of the CMV effect is discussed in this review. Copyright © 2008 John Wiley & Sons, Ltd.