Risk of Hematologic Malignancy Following Infectious Mononucleosis Caused By Epstein-Barr Virus: A Danish and Swedish Observational Study over 40 Years

Objective: To estimate and compare the incidence of overall hematologic malignancy, inclusive of lymphomas, in subjects with EBV-IM diagnosed in the hospital setting (in- and outpatient care) and matched controls.

Methods: This was a non-interventional matched cohort study based on secondary use of individual-level data from Danish and Swedish nationwide registers, matching 1:10 on birth date (+/- 6 months) and sex. EBV-IM subjects were identified by ICD diagnosis codes in national patient registers, matched controls in national population registers, and malignancy events in national cancer registers and patient registers (Sweden only) from 1978 to 2021 in Denmark and 1988 to 2022 in Sweden. All patients with a diagnosis of IM were eligible for inclusion, with exclusion of patients with a diagnosis of cytomegalovirus (CMV)-related IM and/or a prior diagnosis of any malignancy. Incidence rates of overall as well as of subtypes of hematologic malignancy were calculated, and Cox proportional hazards models used to estimate hazard ratios (HR) for overall and subtypes of hematologic malignancy, comparing hospital EBV-IM cases to matched controls. A range of lag periods were introduced to take into consideration diagnostic delay and reverse causality, by shifting the start of follow-up to 2, 3, and 5 years after the date of hospital-diagnosed infection.

Results: A combined total of 85,874 subjects were identified as eligible for inclusion in Denmark and Sweden and included in the analysis. Baseline characteristics were generally comparable between EBV-IM subjects and matched controls in both countries. Mean (±SD) age at time of EBV-IM diagnosis was 18.2 years (±10.0) (Denmark) and 18.1 years (±10.2) (Sweden) and 47% (Denmark) and 50% (Sweden) of patients were female. We identified 205 (Denmark) and 248 (Sweden) hematologic malignancy events over 728,000 and 752,000 person-years in Denmark and Sweden, respectively, corresponding to incidence rates for any hematologic malignancy of 28.1 (95% CI: 24.3-32.0) per 100,000 person-years (Denmark) and 33.0 (28.9–37.1) (Sweden) among EBV-IM subjects vs 15.5 (14.6-16.4) (Denmark) and 13.3 (12.5–14.1) (Sweden) among matched controls. The corresponding age-sex-calendar-period-adjusted HR (95% CI) for the comparison of EBV-IM subjects with matched controls were 1.77 (1.53-2.06) (Denmark) and 2.49 (2.16–2.86) (Sweden). For sub-groups of hematologic malignancy, HRs (95% CI) were 1.85 (1.49-2.29) (Denmark) and 2.85 (2.37–3.43) (Sweden) for lymphomas, 2.60 (1.78–3.78) (Denmark) and 3.19 (2.41–4.23) (Sweden) for Hodgkin’s lymphoma, and 1.62 (1.25-2.10) (Denmark) and 2.60 (2.07–3.27) (Sweden) for non-Hodgkin’s lymphoma. For Burkitt’s lymphoma, a HR could not be estimated for Denmark due to too few events but was 2.99 (2.22–4.03) in Sweden. Inclusion of lag-periods of 2, 3 and 5 years resulted in lower risk estimates with less precision but that remained significant, with risk patterns consistent with main findings.

Conclusion: In this study using nationwide Danish and Swedish data spanning more than 40 years, a diagnosis of EBV-IM was associated with almost two times higher incidence of any hematologic malignancy compared to age and sex matched controls without hospital diagnosed EBV-IM. This association was strongest between EBV-IM infection and Hodgkin’s lymphoma in both countries. Understanding the long-term risk of malignancy occurring after EBV-IM in broad populations is crucial for improving patient outcomes.