The thesis is based on seven publications in English and a review of the literature. The studies were carried out to contribute to the understanding of Hodgkin lymphoma epidemiology through descriptions of its occurrence and its association with Epstein-Barr virus (EBV) infection presenting as infectious mononucleosis. The investigations were supported by the Danish Cancer Society, the Swedish Cancer Society, the Danish Cancer Research Foundation, the Nordic Cancer Union, the Lundbeck Foundation, Plan Danmark, Danish National Research Foundation, Lily Benthine Lund's Foundation, Aase og Ejnar Danielsen's Foundation, Grosserer L. F. Foght's Foundation, the Leukaemia Reseach Fund, the Kay Kendall Leukaemia Fund, and the U.S. National Institutes of Health. The work was carried out in the period 1999-2010 during my employment at the Department of Epidemiology Research at Statens Serum Institut. The employed study designs included population-based incidence surveys of Hodgkin lymphoma in the Nordic countries and in Singapore, register-based cohort studies to characterise the pattern of cancer occurrence in patients with infectious mononucleosis and their first degree relatives, a register-based cohort and a population-based case-control study to characterise the association between infectious mononucleosis and Hodgkin lymphoma taking tumour EBV-status into consideration, and a case-series analysis to assess the association between HLA class I alleles and EBV-positive and EBV-negative Hodgkin lymphomas. Analyses of Nordic incidence data demonstrated that the occurrence of Hodgkin lymphoma had increased markedly younger adults in the period 1978-97, whereas it had decreased among older adults. In combination, these developments led to an accentuation of the younger adult Hodgkin lymphoma incidence peak, which has been a hallmark of Hodgkin lymphoma epidemiology in the Western hemisphere for more than a half century. The opposing incidence trends in younger and older adults are consistent with the prevailing hypothesis of aetiological heterogeneity between Hodgkin lymphomas in different age groups. In contrast to Western industrialised countries, absence of a younger adult incidence peak has been a characteristic of Hodgkin lymphoma epidemiology in developing and Asian populations. A survey of Hodgkin lymphoma occurrence in Singapore 1968-2002 revealed increasing incidence rates and the emergence of an incidence peak in younger adults. The appearance of a younger adult incidence peak in conjunction to socio-economic transition towards Western world lifestyle in Singapore is compatible with the suspicion that Hodgkin lymphoma in younger adults is associated with correlates of socioeconomic affluence in childhood, such as delayed exposure to childhood infectious agents. EBV can be demonstrated in the malignant cells in a subset of Hodgkin lymphomas and it has been speculated that the virus' presence and absence may distinguish between aetiologically separate Hodgkin lymphoma entities. This possibility was explored in five investigations characterising the association between infectious mononucleosis and Hodgkin lymphoma. In these studies, infectious mononucleosis was not accompanied by an increased risk of cancer in general, but specifically with an increased risk of Hodgkin lymphoma. The increased risk of Hodgkin lymphoma decreased with time since infectious mononucleosis and because of the typical adolescent age at infectious mononucleosis it was most prominent for Hodgkin lymphoma in younger adults. Supplementing studies provided little support for the notion that the observed association between Hodgkin lymphoma and infectious mononucleosis was explained by confounding or biases. Analyses stratified by Hodgkin lymphoma EBV status indicated that the increased risk after infectious mononucleosis was confined to the subset of Hodgkin lymphomas that harbour the virus in the malignant cells. The genetic analyses pointed to increased and decreased risk of EBV-positive Hodgkin lymphoma associated with HLA-A01 and HLA-A02 alleles, respectively. The increased risk of EBV-positive Hodgkin lymphoma after infectious mononucleosis was not explained by the two HLA class I alleles, but HLA-A*02 abrogated its effect. This led to an immunological model for EBV-positive Hodgkin lymphoma according to which the level of circulating EBV infected lymphocyte regulated by cytotoxic T-cell responses is a critical determinant of disease risk. Overall, the studies included in the thesis favour that EBV infection is causally associated with development of EBV-positive Hodgkin lymphoma. The circumstances under which the ubiquitous infection leads to lymphoma development must be explored in future studies, which should include analyses of gene-environment interactions. Meanwhile, the aetiology of EBV-negative Hodgkin lymphoma remains elusive. Possible clinical implications of the aetiological heterogeneity should also be considered and assessed.