Setting: In the period 1950 to 1980 the risk of tuberculous infection in the Netherlands declined more steeply than tuberculosis incidence. This study aimed at determining whether this might be explained by preferential transmission within age groups.
Methods: Using restriction fragment length polymorphism (RFLP) typing on all Mycobacterium tuberculosis isolates in the Netherlands from 1993 to 1996, clusters with identical fingerprints were identified. The correlation between the ages of people in clusters of two Dutch patients was determined.
Results: The mean difference in age between the two people, in 81 clusters of two, was 13.9 years, while the mean age difference between all possible pairs of individuals in this data set was 25.5 years. Fisher's intraclass correlation coefficient was 0.62 (95% confidence interval [CI] 0.46-0.74).
Conclusion: It is concluded that sources of tuberculosis may preferentially transmit infection to people close to their own age. As the average age of cases has increased in the period 1950-1980, sources may have become less likely to infect children in whom the risk of infection has been measured. The annual risk of infection measured in children and young adults in countries with low levels of tuberculosis may not apply to older members of the population.