Non-specific effects of vaccination against MeaslesMumps-Rubella on vaccine non-targeted infections in Denmark, Finland, Norway, and Sweden

Background
Vaccines have been found to have non-specific effects (NSE), affecting susceptibility towards other conditions than the vaccine targeted infections. In observational studies from high income countries, having the live vaccine against measles, mumps, and rubella (MMR) compared with non-live diphtheria- tetanusacellular pertussis-containing vaccines (DTaP) as the most recent vaccine has been associated with lower rates of vaccine-non-targeted infectious disease hospitalisations. However, existing studies have often been deemed inconclusive due to the risk of confounding in observational studies. Furthermore, differences in study designs hamper the comparability of results across settings.

Aim
To investigate if MMR holds potential for reducing the rates of infectious disease morbidity (infectious disease hospitalisations and antibiotic use) in Denmark, Finland, Norway, and Sweden (the Nordic countries). As a basis for undertaking these investigations, the secondary aims were to setup a Nordic data platform that facilitates analyses of register data across the Nordic countries and to identify and compare the rates of relevant measures of infectious disease morbidity across the Nordic countries.

Methods
First, the country specific data sources were investigated and compared for completeness and accuracy, whereafter the data sources were harmonized using a common data model to facilitate trans-Nordic investigations using identical statistical coding (Paper 1). Second, the rates of infectious disease hospitalisations (Paper 2) and antibiotic treatments (Paper 3) were compared across the Nordic countries. Finally, the potential for MMR to reduce the rates of infectious disease hospitalisations for non-targeted infections (Paper 4) and antibiotic treatments (Paper 5) among children below 2 years of age, was investigated in nationwide cohort studies with vaccination status as time varying exposures. The analyses were undertaken using Cox proportional hazards models with age as the underlying time scale. The analyses were conducted using covariate adjusted and inverse probability of treatment weighted models including numerous potential confounding factors such as indicators of socioeconomic status and health status.

Results
The recommended age of MMR vaccination was 12 months in Finland, 15 months in Denmark and Norway, and 18 months in Sweden. The rates of infectious disease hospital contacts and antibiotic treatments varied across countries. The differences across countries were greater for antibiotic treatments (the rates per 1000 person-years were 284 (Denmark), 429 (Finland), 184 (Norway), and 219 (Sweden)) and when including all inpatient, outpatient, and emergency room contacts combined (rates per 1000 person-years were 79.0 (Denmark), 87.1 (Finland), 62.1 (Norway), and 125.2 (Sweden)). The rates were more similar for inpatient hospitalisation with overnight stays (ranged from 13.0 to 19.3 per 1000 person years), which may reflect more similar handling of the more severe infections. We focused the investigations of NSE of MMR on infectious disease hospitalisations with overnight stays and all prescriptions with systemic antibiotics, respectively.
Summary estimates across the Nordic countries indicated that MMR was associated with 25% (16% to 35%) reduced rates of infectious disease hospitalisations with overnight stays and with 11% (7% to 15%) reduced rates of antibiotic treatments. The beneficial associations were observed in all countries but were greater in countries with steeper uptake of MMR around the age of recommended vaccination and higher MMR coverage. A negative control exposure analysis found that exposure to DTaP3 compared with DTaP2 was associated with around 20% reduced rates of both outcomes.

Discussion
We found MMR to be associated with lower rates of infectious disease hospitalisations and antibiotic treatments. The observed beneficial associations were greater for infectious disease hospitalisations compared with antibiotic treatments. All analyses were undertaken with excessive control for potential confounding factors and careful consideration of study design to enhance comparability between children with different vaccination statuses. However, in all countries, bias attributable to deviating from recommended MMR vaccination may account for at least some of the observed association and may possibly outweigh the beneficial association observed for antibiotic treatments. Bias related to deviating from vaccination recommendation is possibly stronger in countries with high and steep uptake of MMR resulting in a more selected group without MMR for comparison. A strength of the present investigations pertains to the multi country setup, which facilitates triangulation of results across settings with different underlying bias structures due to e.g. different vaccination coverages and coding practices.

Conclusion
The present thesis presents a comprehensive investigation of the potential of NSE of MMR for reducing non-targeted infectious diseases in high income countries. The results indicate a greater potential to prevent the more severe infections requiring hospitalisation whereas limited effect was seen for the milder infections handled with antibiotic treatments. The observed beneficial associations thus reflect a potential to prevent infectious diseases in populations with general good health, even when the majority already receives MMR, by encouraging timely vaccination. The variation in the recommended age at MMR across countries open for the potential to conduct randomized trials of varying MMR ages, to study both specific and non-specific effects of MMR, to define the optimal age of vaccination.