Life course epidemiology and infectious diseases

Abstract

There has been a traditional view that divided epidemiology into infectious and chronic diseases. Since we now know that at least 15% of cancers worldwide are caused by infections, 1 that infections frequently have a natural history lasting decades and that the same epidemiological methods can be applied to both infectious and non-infectious diseases, this view can be considered purely historical. However, the nature of infectious diseases has implications for the application of epidemiological approaches including the specific example of life course epidemiology. Whilst the definition of this is still debated, it clearly refers to the way in which influences at different stages of life affect and modify risks of disease. Life course epidemiology particularly focuses on disease in adult life and so that is what we discuss here. We are not going to consider childhood or adolescent diseases. In addition it recognizes two types of influence on risk: (1) a simple accumulation of risks and (2) ‘programming’. 2 The latter is the way in which an influence modifies a biological system so that it responds differently to subsequent exposures. The epidemiological paradigm for this has been David Barker’s hypothesis that intra-uterine nutrition determines physiological responses in the cardiovascular system. 3 When we come to talk about infections we have to deal with a few semantics. Susceptibility means that the subject is vulnerable to infection if they are exposed. Infants may not be susceptible because of transferred immunity from their mother. Many, but not all, infections result in a loss of susceptibility due to immune memory. Exposure is traditionally used to define exposure to the infectious agent. We will use it in that context here and use the term ‘influence’ for non-agent exposures. There is a critical difference between infection and disease. The proportion of individuals infected who develop disease varies from 0% (as with infection with the hepatitis G virus) to virtually 100% (with measles and the human immunodeficiency virus [HIV]) depending on the infection. However, for many infections the relationship is strongly dependent on the age at which infection occurs—an obvious entree for life course epidemiology. Thus Epstein-Barr virus infection in childhood hardly causes any symptoms whereas in adolescence it results in glandular fever. Severity and rate of disease progression are also influenced by age at infection. Hepatitis A becomes an increasingly severe disease the older the subject. We return to this point below. Finally, we use the term ‘reactivation’ for persistent infections that produce disease after a prolonged period of latency following a first disease episode. Varicella zoster virus remains latent for years following chickenpox until around a fifth of infected people develop shingles. The simple model in Figure 1 shows how influences may affect various stages of the process. This emphasises proximate biological factors. Clearly there are more distal factors that may be more easily measured. For example, exposure to the infectious agent is highly dependent on behaviour—mixing patterns with other young children may determine exposure to childhood infections and sexual activity modulates risk of sexually transmitted infections. The timing from susceptibility to disability, death or recovery is incredibly variable. With many infections that do not persist, symptoms of disease occur, if they are going to, within days—at most months—after initial infection. Agents, such as hepatitis B and C viruses, that persist and replicate within the body may not lead to disease (in this case primary liver cancer) until 50 years after infection. Once infectious disease is recognized the natural history is as variable as that of non-communicable diseases. The reasons for all this variation are the influences at each stage.